Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 L...
756 downloads
2776 Views
353MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Front of Book > Editors
Editors William E. Brant MD, FACR Professor of Radiology, Director ThoracoAbdominal Division, Department of Radiology, Virginia Health System, Charlottesville, Virginia
University
Clyde A. Helms MD Professor of Radiology and Surgery, Chief Division of Musculoskeletal Radiology, Department of Radiology, Duke University Medical Center, Durham, North Carolina
Secondary
Editors
Lisa McAllister Acquisitions Editor Rebecca Barroso Developmental Editor Kerry Barrett Managing Editor Angela Panetta Marketing Manager Nicole Project
Walz Manager
Benjamin Senior
Rivera
Manufacturing
Manager
Risa Clow Design Coordinator
2
of
Wanda Espana Cover Designer TechBooks Production Services Maple-Vail Printer
Contributors Ramsey D. Badawi PhD Assistant Professor of Radiology University of California, Davis; Senior PET Physicist, Department of Radiology, University of California, Davis Medical Center, Sacramento,
California
Jerome A. Barakos MD Assistant Clinical Professor Department of Radiology, Neuroradiology Section, University of California, San Francisco; Director of Neuroimaging, Department of Radiology, California Pacific Medical Center, San Francisco, California Robert M. Barr MD Mecklenburg Radiology Associates, P.A. Presbyterian Hospital, Charlotte, North Carolina Bijan Bijan MD Assistant Professor of Diagnostic Radiology and Nuclear Medicine Department of Cross-Sectional Body Imaging/MRI and Nuclear Medicine/PET Divisions, University of California, Davis Medical Center, Sacramento, California; Director of Radiology, MRI Department, Elk Grove Diagnostic Imaging–MRI Medical Center, Elk Grove, California
3
Peter W. Blue MD Professor of Radiology University of South Carolina School of Medicine, Columbia, South Carolina; Chief, Nuclear Medicine Service, Moncrief Army Community Hospital, Fort Jackson, South Carolina William E. Brant MD, FACR Professor of Radiology, Director ThoracoAbdominal Division, Department Virginia
Health
System,
Charlottesville,
of
Radiology,
University
of
Virginia
Jerrold T. Bushberg PhD, DABMP Clinical Professor Department of Radiology, University of California, Davis, School of Medicine, Sacramento, California Marc G. Cote DO, FACP Chief Department of Radiology; Nuclear Medicine Consultant to the Office of the Army Surgeon General, Madigan Army Medical Center, Tacoma,
Washington
Raymond S. Dougherty MD Associate Clinical Professor, Residency Program Director, Diagnostic Radiologist Division of Abdominal Imaging, Department of Radiology, University of California, Davis Medical Center, Sacramento, California Erik H.L. Gaensler MD Associate Clinical Professor Department of Radiology, University of California, San Francisco, San Francisco, California Alisa D. Gean MD Professor of Radiology, Neurology, and Neurosurgery
4
University of California, San Francisco; Chief of Neuroradiology, San Francisco General Hospital, San Francisco, California Michael F. Hartshorne MD Professor of Radiology University of New Mexico School of Medicine; Staff Radiologist, University of New Mexico Hospital, Albuquerque, New Mexico Clyde A. Helms MD Professor of Radiology and Surgery, Chief Division of Musculoskeletal Radiology, Department of Radiology, Duke University Medical Center, Durham, North Carolina Timothy J. Higgins MD Clinical Instructor of Radiology University of Vermant College of Medicine; Resident in Radiology, Fletcher Allen Health Care, Burlington, Vermant Susan D. John MD, FACR Professor and Chair, Diagnostic and Interventional Imaging University of Texas Health Science Center; Chair, Imaging Services, Memorial Hermann Children's Hospital, Houston, Texas Jeffrey S. Klein MD Professor of Radiology University of Vermont College of Medicine; Chief of Thoracic Radiology, Fletcher Allen Health Care, Burlington, Vermont Kelly K. Koeller MD, FACR Associate Professor of Radiology Mayo
Clinic,
Rochester,
Minnesota
Christopher M. Kramer MD Professor of Radiology and Medicine, Associate Chief Cardiovascular Division, Department of Medicine, University of
5
Virginia
Health
System,
Charlottesville,
Virginia
Linda A. Kroger MS Radiation Safety Officer University of California, Davis Health System, Sacramento, California Tuong H. Le MD, PhD Assistant Clinical Professor Department of Radiology, San Francisco General Hospital, University of California, San Francisco, California Todd E. Lempert MD Chief Diagnostic Mission
and
Viejo,
Interventional
Neuroradiology,
Mission
Hospital,
California
Huong T. Le-Petross MD, FRCPS Assistant Professor of Radiology, Radiologist Breast Imaging Section, The University of Texas M.D. Anderson Cancer Center, Houston, Texas David H. Lewis MD Associate Professor of Radiology University of Washington School of Medicine; Director of Nuclear Medicine,
Harborview
Medical
Center,
Seattle,
Washington
Karen K. Lindfors MD, MPH Professor of Clinical Radiology, Chief of Breast Imaging University of California, Davis School of Medicine, Sacramento, California Charito Love MD Research Scientist Division of Nuclear Medicine, Long Island Jewish Medical Center, New Hyde Park, New York
6
Vivek Manchanda MD Chief Resident Nuclear Medicine Division, Department of Radiology, University of Washington,
Seattle,
Washington
Mike McBiles MD Staff Radiologist Saratoga Hospital, Saratoga Springs, New York Michael J. Miller Jr. MD Associate in Department of Interventional Radiology Duke University Medical Center, Durham, North Carolina Santiago Miró BSc, MD Assistant Professor of Radiology Université Laval, Quebec, QC, Canada; Attending Radiologist, Cardio-thoracic Section Hôpital Laval, Quebec, QC, Canada Walter L. Olsen MD Assistant Clinical Professor of Radiology University of California, San Diego; Radiologist, San Diego Diagnostic Radiology Medical Group, San Diego, California Christopher J. Palestro MD Professor of Nuclear Medicine and Radiology Albert Einstein College of Medicine, Bronx, New York; Chief, Division of Nuclear Medicine, Long Island Jewish Medical Center, New Hyde Park, New York Howard A. Rowley MA, MD Chief of Neuroradiology, Joseph Sackett Professor of Radiology, Associate Professor of Radiology, Neurology, and Neurosurgery University of Wisconsin, Madison, Wisconsin
7
David J. Seidenwurm MD Chair Diagnostic Division; Radiological Neuroradiologist, Medical Center,
Associates
Diagnostic Imaging and Sacramento, California
of
Sacramento,
Radiation
Oncology,
Sutter
David K. Shelton MD Professor Nuclear Medicine and Radiology; Chief, Nuclear Medicine, University of California, Davis Medical Center, Sacramento, California Tony P. Smith MD Professor of Radiology, Chief Division of Interventional Radiology, Department of Radiology, Duke University Medical Center, Durham, North Carolina Leonard E. Swischuk MD Chairman and Professor of Radiology University of Texas Medical Branch, Galveston, Texas James H. Timmons MD, PhD Radiology
Consultants
PLC,
Director
Nuclear Medicine and PET, Department of Radiology, Battle Creek Health Systems, Battle Creek, Michigan Rhonda A. Wyatt MD Chief of Nuclear Medicine Department of Radiology; MD Imaging, Redding, California
8
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Front of Book > Dedication
Dedication I dedicate this book to my wife, Barbara, my true companion and the love of my life and to my expanding family: Dan and Lindsay, Ryan and Tiffany, Jonathan Allen, Rachel and Rob, Jake and Dilara, Dan and Debra, and our precious grandchildren, Evan Edward, Finley Matthew, and Danielle Marion. W E B
To my wife, Nancy Marie Major, and to our son, Austin Michael. C A H
9
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Front of Book > Foreword
Foreword Congratulations to Drs. Brant and Helms for updating their classic and outstanding textbook, Fundamentals of Diagnostic Radiology, in this third edition. For many years, this textbook has served as a basic overview of diagnostic imaging for medical students, radiology residents, and others in the medical field. The third edition comes in two different formats: a single hardcover edition as previously published and a new four-volume soft-cover edition that can be carried around easily as one rotates on different services. The third edition retains many of the previous authors; however, chapters on chest, bones, and joints; the gastrointestinal tract; ultrasound; breast imaging; and pediatrics all have major revisions, and new authors have written chapters on cardiovascular and vascular/interventional radiology as well as on nuclear medicine. The third edition continues the strong legacy of past editions, emphasizing the fundamentals of diagnostic imaging. Two new chapters on cardiac imaging and PET-CT are appropriately added to update these new imaging topics. Basic chapters have also been expanded to include topics such as virtual colonoscopy, the use of CT in the evaluation of renal stones, CT of abdominal trauma, and pulmonary complications of bone marrow transplants. To complete newer topics the authors have included global infectious diseases such as SARS and anthrax. This complete textbook will continue to be basic reading for diagnostic radiology residents, serving as a review as they prepare for written and oral boards. It is also a great textbook for medical students and those practitioners who want a concise reference
10
textbook of diagnostic imaging for their offices. I commend Drs. Brant and Helms for this updated third edition of their outstanding textbook, Fundamentals of Diagnostic Imaging, as did Drs. Keats and Ravin before me in their forewords to the previous editions. One of an academician's rewards is watching one's previous residents/coworkers excel in academic radiology. I have personally followed Drs. Brant and Helms as they rose through academics, Bill first as one of my residents and Clyde as a teacher and coworker. Thanks to these two superb teachers of radiology for their outstanding textbook. Alfred B. Watson Jr. MD, MPH
11
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Front of Book > Preface to the Third Edition
Preface to the Third Edition We are happy and proud to offer a third edition of our Fundamentals of Diagnostic Radiology text. We cherish the fact that so many radiology residents around the country continue to use our text as we had originally intended—as a first read for beginning residents and as a concise but comprehensive review text for preparation for American Board of Radiology written and oral examinations. We are particularly pleased that most of our authors from our original text published in 1994 have returned to contribute to this edition. In 1992 as we conceived the notion of writing this text we deliberately chose authors who were very junior faculty or fresh out of their fellowship training to identify what is fundamental and important to the student of radiology. Now more than a decade later these same authors are leaders in the teaching of radiology, serving as department chairs, division chiefs, professors, and senior radiologists. All have kept pace with what is fundamental and important in learning the rapidly progressing specialty of diagnostic imaging. Dr. Erik Gaensler has returned to organize his expert group of neuroradiologists writing the Neuroradiology section. These include Kelly Koeller, the former chairman of Radiologic Pathology at the Armed Forces Institute of Pathology and currently associate professor of radiology at the Mayo Clinic; Alisa Gean, a leading expert in brain trauma imaging; and Howard Rowley, the Chief of Neuroradiology at the University of Wisconsin. Dr. Susan John, the primary author of our pediatric radiology chapters, is now the Chair of the Department of Diagnostic and Interventional Imaging at the University of Texas–Houston Medical School. Susan is the first woman to chair a major department at her institution. Dr. David K.
12
Shelton, Professor of Radiology and Chief of the Nuclear Medicine division at the University of California–Davis was essential in reorganizing and authoring the Nuclear Radiology section. Without his commitment and fine work we would have been in deep trouble completing this revision. He made a major contribution as the section editor for Cardiac Radiology as well. Jeffrey Klein again contributed the entire chest radiology section, covering the topics so vital to diagnostic radiology. Tony Smith and Michael Miller have provided a stellar rewrite of the entire Vascular and Interventional Radiology
section.
Our new edition features a revised organization to keep pace with the dynamics of the American Board of Radiology examinations. In addition to our introductory chapter, we have the 11 sections matching the categories of the oral board examination. Cardiac Radiology and Vascular and Interventional Radiology are now separate sections. We have added a chapter dedicated to the expanding use of Cardiac MR. We have added a chapter on Positron Emission Tomography and PET-CT to match the rapid expansion of this combined imaging modality. All chapters feature updated and added illustrations to match our reputation as pictorialists. We have added descriptions of renal stone CT, abdominal trauma CT, CT and MR urography, the imaging of anthrax and severe acute respiratory distress syndrome (SARS), the pulmonary complications of bone marrow transplant, breast MR, and many more topics of current interest. We have striven in every case to keep to our goals of being fundamental and comprehensive while concise. Growth in the size of our book parallels growth in the scope of our field. We have learned that many residents with shoulders strained by its weight have taken our book to their local copy shop and had it rebound into manageable soft-cover sections. Our publisher has agreed to offer the current edition with a four-volume soft cover option in addition to our original single-volume hardcover text. We offer our thanks to the many individuals in addition to our authors who have contributed mightily to the text's completion. Kerry Barrett, Managing Editor, and Lisa McAllister, Executive Editor, at Lippincott Williams & Wilkins, have guided this edition from
13
inception to completion. Barbara Stabb and her team at TechBooks in York, Pennsylvania, provided outstanding book production services. Most of all we thank the many residents, fellows, medical students, and faculty colleagues with whom we have worked and who have challenged us as teachers and radiologists. William E. Brant MD, FACR Clyde A. Helms MD
14
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Front of Book > Abbreviations
Abbreviations List
of
Universal
Abbreviations
AIDS Acquired
immunodeficiency
syndrome
ARDS Acute
respiratory
distress
syndrome
CNS Central
nervous
system
CT Computed
tomography
CSF Cerebrospinal
fluid
FDG PET Fluorodeoxyglucose
positron
emission
tomography
GI Gastrointestinal GRE Gradient recalled echo (MR sequence) HIV Human
immunodeficiency
virus
HRCT High
resolution
computed
tomography
15
(Lungs)
IV Intravenous LA Left
atrium
LV Left
ventricle
MR Magnetic
resonance
imaging
PA Pulmonary
artery
PET Positron
emission
tomography
emission
tomography-computed
PET-CT Positron
tomography
SPECT Single-photon
emission
computed
tomography
STIR Short TI inversion recovery (MR sequence) RA Right
atrium
RV Right
ventricle
T1WI T1-weighted
image
T2WI T2-weighted
image
TE Time of echo (for MR sequences) TR Time of repetition (for MR sequences)
16
US Ultrasound
17
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section I - Basic Principles > Chapter 1 Diagnostic Imaging Methods
Chapter 1 Diagnostic
Imaging
Methods
William E. Brant Diagnostic radiology is a dynamic specialty that continues to undergo rapid change with ongoing advancements in technology. Not only has the number of imaging methods increased, but each method continues to undergo improvement and refinement of its use in medical diagnosis. This chapter reviews the basics of the major diagnostic imaging methods and provides the basic principles of image interpretation for each method. Contrast agents commonly used in diagnostic radiology are also discussed. The basics of nuclear radiology are discussed in later chapters.
CONVENTIONAL
RADIOGRAPHY
Conventional radiographic examination of the human body dates back to the genesis of diagnostic radiology in 1895, when Wilhelm Roentgen produced the first x-ray film image of his wife's hand. Conventional radiography remains fundamental to the practice of diagnostic imaging.
Image
Generation
X-rays are a form of radiant energy that is similar in many ways to visible light. X-rays differ from visible light in that they have a very short wavelength and are able to penetrate many substances that are opaque to light. The x-ray beam is produced by bombarding a tungsten target with an electron beam within an x-ray tube (1) .
18
Film
Radiography
Conventional film radiography utilizes a screen-film system within a film cassette as the x-ray detector. As x-rays pass through the human body, they are attenuated by interaction with body tissues (absorption and scatter) and produce an image pattern on film that is recognizable as human anatomy. X-rays that are transmitted through the patient bombard a fluorescent particle–coated screen within the film cassette, causing a photochemical interaction that emits light rays, which expose photographic film within the cassette (Fig. 1.1). The film is removed from the cassette and developed by an automated chemical film processor. The final product is an x-ray image of the patient's anatomy on a film (Fig. 1.2) . Computed radiography (CR) is a filmless system that eliminates chemical processing and provides digital radiographic images. CR substitutes a phosphor imaging plate for the film screen cassette (2) . The same gantry, x-ray tube, and exposure control systems used in conventional radiography are utilized for CR. The phosphor-coated imaging plate interacts with x-rays transmitted through the patient to capture a latent image. The phosphor plate is placed within a reading device that scans the plate with a helium-neon laser, emitting light, which is then captured by a photomultiplier tube and processed into a digital image. The digital image is transferred to a computerized picture archiving and communication system (PACS). The PACS stores and transmits digital images via computer networks to give physicians and health care providers in many locations simultaneous instant access to the diagnostic images. Digital radiography (DR) provides a film-free and cassette-free system for capturing x-ray images in digital format. DR substitutes a fixed electronic detector or charge-coupled device for the film screen cassette or phosphor imaging plate. Direct readout detectors produce an immediate DR image. Most DR detectors are installed in a fixed gantry, limiting the ability of the system to obtain P.4 P.5
19
images portably at the patient's bedside. CR is generally used for that purpose in a digital imaging department. Direct digital image capture is particularly useful for angiography, because it provides rapid digital image subtraction, and for fluoroscopy, because it captures video images with low, continuous levels of radiation.
FIGURE 1.1. X-ray Film Cassette. Diagram demonstrating a sheet of x-ray film between two fluorescent screens within a lightproof cassette.
20
FIGURE 1.2. Conventional Radiography. A. Diagram of an x-ray tube producing x-rays that pass through the patient and expose the radiographic film. For digital radiography, a phosphor imaging plate or fixed electronic detector takes the place of the film cassette. B . Supine AP radiograph of the abdomen reveals the patient's anatomy because anatomic structures differ in their capacity to attenuate x-rays that pass through the patient. The stomach (S) and duodenum (d) are visualized because air in the lumen is of different radiographic density than the soft tissues that surround the GI tract. The right kidney (between short straight arrows), edge of the liver (long straight arrow), edge of the spleen (open arrow), and the left psoas muscle (curved arrow) are visualized because fat outlines the soft tissue density of these structures. The bones of the spine, pelvis, and hips are clearly seen through the soft tissues because of their high radiographic density.
21
Conventional tomography provides radiographic images of slices of a living patient. This is done by simultaneously moving both the x-ray tube and the x-ray detector around a pivot point centered in the patient in the plane of the anatomic structures to be studied (Fig. 1.3) . Structures above and below the focal plane are blurred by the motion of the tube and detector. Objects within the focal plane are visualized with improved detail as a result of the blurring of the overlying and underlying structures. The motion of the x-ray tube and detector may be linear, circular, elliptic, spiral, or hypocycloid. Tomography is a useful adjunct to conventional radiographs in situations in which improved detail is needed for diagnosis. However, with the proliferation and wide availability of cross-sectional imaging, the use of conventional tomography is currently quite limited. Fluoroscopy enables real-time radiographic visualization of moving anatomic structures. A continuous x-ray beam passes through the patient and falls onto a fluorescing screen (Fig. 1.4). The faint light pattern emitted by the fluorescing screen is amplified electronically by an image intensifier, and the image is displayed on a television monitor and recorded digitally as a single image or series of images for real-time viewing (i.e., a movie or “cine-fluoroscopy―). Fluoroscopy is extremely useful to evaluate motion such as GI peristalsis, movement of the diaphragm with respiration, and cardiac action. Fluoroscopy is also used to perform and monitor continuously radiographic procedures, such as barium studies and catheter placements. Video and static fluoroscopic images are routinely stored in digital format on a PACS.
22
FIGURE
1.3. Conventional
Tomography. In this technique, the
x-ray tube and film simultaneously move about a pivot point at the level of the desired focal plane. Anatomic structures within the focal plane remain in sharp focus, whereas the structures above and below the focal plane are blurred by the motion of the tube and film.
FIGURE 1.4. Fluoroscopy. This diagram of a fluoroscopic unit illustrates the x-ray tube located beneath the patient examination
23
table and the fluorescing screen with the image intensifier positioned above the patient. Amplification of the faint fluorescing image by the image intensifier allows the radiation exposure to the patient to be kept at low levels during fluoroscopy. The real-time fluoroscopic images are viewed on a television monitor and may be recorded on videotape. Radiographs are obtained by digital image capture or by placing a film cassette between the patient and the image intensifier and exposing the image receptor with a brief pulse of radiation.
Conventional
angiography involves the opacification of blood vessels
by intravascular injection of iodinated contrast agents. Conventional arteriography uses small flexible catheters that are placed in the arterial system, usually via puncture of the femoral artery in the groin. With the use of fluoroscopy for guidance, catheters of various sizes and shapes can be manipulated selectively into virtually every major artery. Contrast injection is performed by hand or by mechanical injector and is accompanied by timed rapid-sequence filming or digital computer acquisition (DR) of the fluoroscopic image. The result is a timed series of images depicting contrast flow through the injected artery and the tissues that the artery supplies. Conventional venography is performed by contrast injection of veins via distal puncture or selective catheterization.
Naming
Radiographic
Views
Most radiographic views are named on the basis of the way that the xray beam passes through the patient. A posteroanterior (PA) chest radiograph is one in which the x-ray beam passes through the back of the patient and exits through the front of the patient to expose an xray detector positioned against the patient's chest. An anteroposterior (AP) chest radiograph is exposed by an x-ray beam passing through the patient from front produced by passing a caudad) direction, with additionally named by
to back. A craniocaudad (CC) mammogram is beam through the breast in a vertical (cranial to the patient standing or sitting. Views are identifying the
24
P.6 position of the patient. Erect, supine, or prone views may be specified. A right lateral decubitus view of the chest is exposed with a horizontal x-ray beam passing through the chest of a patient lying on his or her right side. Radiographs taken during fluoroscopy are named on the basis of the patient's position relative to the fluoroscopic table, because the x-ray tube is positioned beneath the table. A right posterior oblique (RPO) view is obtained with the patient lying with the right side of his or her back against the table and the left side elevated away from the table. The x-ray beam generated by the x-ray tube located beneath the table passes through the patient to the x-ray cassette or detector located above the patient.
Principles
of
Interpretation
Conventional radiographs demonstrate five basic radiographic densities: air, fat, soft tissue, bone, and metal (or x-ray contrast agents). Air attenuates very little of the x-ray beam, allowing nearly the full force of the beam to blacken the image. Bone, metal, and radiographic contrast agents attenuate a large proportion of the x-ray beam, allowing very little radiation through to blacken the image. Thus, bone, metallic objects, and structures opacified by x-ray contrast agents appear white on radiographs. Fat and soft tissues attenuate intermediate amounts of the x-ray beam, resulting in proportional degrees of image blackening (shades of gray). Thick structures attenuate more radiation than thin structures of the same composition. Anatomic structures are seen on radiographs when they are outlined in whole or in part by tissues of different x-ray attenuation. Air in the lung outlines pulmonary vascular structures, producing a detailed pattern of the lung parenchyma (Fig. 1.5). Fat within the abdomen outlines the margins of the liver, spleen, and kidneys, allowing their visualization (Fig. 1.2B). The high density of bones enables visualization of bone details through overlying soft tissues. Metallic objects such as surgical clips are usually clearly seen because they highly attenuate the x-ray beam. Radiographic contrast agents are suspensions of iodine and barium compounds that highly attenuate the x-ray beam and are used to outline anatomic structures. Disease
25
states may obscure normally visualized anatomic structures by silhouetting their outline. For example, pneumonia in the right middle lobe of the lung replaces air in the alveoli with fluid and silhouettes the right heart border (Fig. 1.6). Nancy Major provides an excellent text on the basics of radiographic interpretation (3) .
FIGURE 1.5. Erect PA Chest Radiograph. The pulmonary arteries (white open arrow) are seen in the lung because the vessels are outlined by air in alveoli. The left cardiac border (long arrow) is crisply defined by the adjacent air-filled lung. The left main bronchus (curved arrow) is seen because its air-filled lumen is surrounded by soft tissue of the mediastinum. An air-fluid level (open black arrow) in the stomach confirms the erect position of the patient during exposure of the radiograph.
26
FIGURE
1.6. Right Middle Lobe and Left Lower Lobe
Pneumonia. PA erect chest radiograph demonstrates pneumonia (P) in the right middle lobe replacing air density in the lung with soft tissue density and silhouetting the right heart border. The dome of the right hemidiaphragm (black arrow) is defined by air in the normal right lower lobe and remains visible through the right middle lobe infiltrate. The left heart border (white arrow), defined by air in the lingula, remains well defined despite infiltrate in the left lower lobe.
CROSS-SECTIONAL
IMAGING
TECHNIQUES
CT, MR, and US are techniques that produce cross-sectional images of the body. All three interrogate a three-dimensional volume or slice of patient tissue to produce a two-dimensional image. The resulting image is made up of a matrix of picture elements (pixels), each of which P.7 represents a volume element (voxel) of patient tissue. The tissue composition of the voxel is averaged (volume averaged) for display as a pixel. CT and MR assign a numeric value to each picture element in
27
the matrix. The matrix of picture elements that make up each image is usually between 128 × 256 (32,768 pixels) and 560 × 560 (313,600 pixels) and is determined by the specified acquisition parameters
(Fig. 1.7) .
FIGURE 1.7. Image Matrix. A. Magnified CT image of a pulmonary nodule (N). The pixels that make up the image are evident as tiny squares within the image. The window width is set at 2,000 H with a window level of 600 H to accentuate visualization of the white soft tissue nodule on a background of gray, air-filled lung. B . Diagram of the matrix that constitutes the CT image. A pixel from the air-filled lung with a calculated CT number of 524 H is gray, whereas a pixel from the soft tissue nodule with a calculated CT number of +46 H is white.
To produce an anatomic image, shades of gray are assigned to ranges of pixel values. For example, 16 shades of gray may be divided over a window width of 320 pixel values (Fig. 1.8). Groups of 20 pixel values are each assigned one of the 16 gray shades. The middle gray shade is assigned to the pixel values centered on a selected window level. Pixels with values greater than the upper limit of the window width are displayed white, and pixels with values less than the lower limit of the
28
window width are displayed black. To analyze optimally all of the anatomic information of any particular slice, the image is viewed at different window-width and window-level settings, which are optimized for bone, air-filled lung, soft tissue, and so forth (Fig. 1.9) . The digital images obtained by CT, MR, and US examination are ideal for storage and access on PACS. Current PACSs allow a broad range of image manipulation during viewing and interpretation of images. Among the features that can be used are interactive alterations in window width and window level, magnification, fusing of images from different modalities, reformatting serial images in different anatomic planes, creation of three-dimensional reconstructions, and marking of key images that summarize major findings.
Computed
Tomography
CT uses a computer to reconstruct mathematically a cross-sectional image of the body from measurements of x-ray transmission through thin slices of patient tissue. CT displays each imaged slice separately, without the superimposition of blurred structures that is seen with conventional tomography. A narrow, well-collimated beam of x-rays is generated on one side of the patient (Fig. 1.10). The x-ray beam is attenuated by absorption and scatter as it passes through the patient. Sensitive detectors on the opposite side of the patient measure x-ray transmission through the slice. These measurements are systematically repeated many times from different directions while the x-ray tube is pulsed as it rotates 360° around the patient. CT numbers are assigned to each pixel in the image by a computer algorithm that uses as data these measurements of transmitted x-rays. CT pixel numbers are proportional to the difference in average x-ray attenuation of the tissue within the voxel and that of water. A Hounsfield unit (H) scale, named for Sir Godfrey N. Hounsfield, the inventor of CT, is used. Water is assigned a value of 0 H, with the scale extending from 1,024 H for air to +3,000 to 4,000 H for P.8 P.9 very dense bone. H units are not absolute values but, rather, are
29
relative values that may vary from one CT system to another. In general, bone is +400 to +1,000 H, soft tissue is +40 to +80 H, fat is -60 to -100 H, lung tissue is -400 to -600 H, and air is -1,000 H.
FIGURE 1.8. Gray Scale. A CT image of the abdomen includes a gray scale along its left edge. Each individual pixel in the CT image is assigned a shade of gray, depending on its calculated CT number (H unit) and the window width and window level selected by the CT operator. Pure white (arrowhead) and pure black (arrow) are at the top and bottom of the gray scale. Along the right side of the CT image is a centimeter scale that can be used to measure the size of objects in the image. R indicates the patient's right side, and L indicates the patient's left side. Cross-sectional images in the transverse plane are routinely viewed from “below,― as if standing at the patient's feet. This orientation allows easy correlation with plain-film radiographs, which are routinely viewed as if facing the patient with the patient's right side to the viewer's left. This patient has an abscess (A) in the liver.
30
FIGURE
1.9. CT Windows. A. A CT image of the upper abdomen
photographed with “soft tissue windows― (window width = 482 H, window level = 14 H) portrays a thoracic vertebra (arrows) entirely white with no bone detail. B . The same CT image rephotographed with “bone windows― (window width = 2,000 H, window level = 400 H) demonstrates destructive changes in the vertebral body (arrows) owing to metastatic lung carcinoma.
31
FIGURE 1.10. Computed Tomography. Diagram of a CT scanner. The patient (P) is placed on an examination couch within the core of the CT unit. An x-ray tube rotates 360° around the patient, producing pulses of radiation that pass through the patient. Transmitted x-rays are detected by a circumferential bank of radiation detectors. X-ray transmission data are sent to a computer, which uses an assigned algorithm to calculate the matrix of CT numbers used to produce the anatomic cross-sectional image. With the helical CT scan technique, the patient couch moves the patient continuously through the rotating x-ray beam. In MDCT, multiple image slices are obtained simultaneously as the patient is moved through the scanner.
Voxel dimensions are determined by the computer algorithm chosen for reconstruction and the thickness of the scanned slice. Most CT units allow slice thickness specifications between 0.5 and 10 mm. Data for an individual slice, 360° tube rotation, are routinely acquired in 1 s or less. The advantages of CT compared with MR include rapid scan acquisition, superior bone detail, and demonstration of calcifications. CT scanning is generally limited to the axial plane; however, images may be reformatted in sagittal, coronal, or oblique planes or as three-
32
dimensional
images.
Conventional CT (nonhelical) obtains image data one slice at a time (4). The patient holds his or her breath, a slice is taken, the patient breathes, the table moves, and the sequence is repeated. This technique requires at least two to three times the total scanning time of helical CT for any given patient scan volume, making optimization of scanning during maximum contrast more difficult. Minor changes in lung volume with each breath-hold may make substantial changes in the chest and abdomen anatomy scanned, resulting in “skip― areas. More recent conventional scanners can simulate helical scanning by the “cluster― technique. Several sequential scans are obtained during a single breath-hold. Helical CT, also called spiral CT, is performed by moving the patient table at a constant speed through the CT gantry while scanning continuously with an x-ray tube rotating around the patient. A continuous volume of image data is acquired during a single breathhold. This technique dramatically improves the speed of image acquisition, enables scanning during optimal contrast opacification, and eliminates artifacts caused by misregistration and variations in patient breathing. The entire liver may be scanned in a single breath-hold; the entire abdomen and pelvis, in two to three breath-holds, all during the first 60 to 90 seconds of intravenous contrast administration. Volume acquisition enables retrospective reconstruction of multiple overlapping slices, improving visualization of small lesions and making high-detail three-dimensional
CT
angiography
possible
(Fig. 1.11) (5) .
Multidetector helical CT (MDCT) is the latest technical advance in CT imaging; it utilizes the principles of the helical scanner but incorporates multiple rows of detector rings (6). This allows the acquisition of multiple slices per tube rotation, increasing the area of the patient that can be covered in a given time by the x-ray beam. Available systems have moved from two slices to 64 slices, which covers 40 mm of patient length for each 1-second or less rotation of the tube. Prototype 256-detector scanners are being developed. The key advantage of MDCT is speed. MDCT is five to eight times faster than single-slice helical CT. For body scanning, 1-mm slices can be
33
obtained creating isotropic voxels (1 × 1 × 1 mm), allowing image reconstruction in any anatomic plane without loss of resolution (4) . Broad area coverage allows for high-detail CT angiography and “virtual― CT colonoscopy and bronchoscopy. A disadvantage of MDCT is radiation dose, which can be three to five times higher with MDCT than with single-slice CT.
FIGURE 1.11. CT Angiogram. A three-dimensional, shaded surface display, angiogram image of the aorta and its branches was created from a series of axial plane MDCT images obtained during rapid bolus IV contrast agent administration. Contrast enhancement greatly increases the CT numbers of the arteries and kidneys and allows removal of structures with lower CT density from the image by “thresholding.― Only pixels with CT numbers higher than a specified threshold value are displayed. Computer algorithms create a “virtual― three-dimensional
34
image from data provided by many overlapping axial slices. The three-dimensional image can be rotated and viewed from any desired angle. “Shading,― simulating light cast from a remote light source, enhances the three-dimensional visual effect.
Radiation Dose in CT As the diagnostic capability of CT expands dramatically, so does its utilization. Unfortunately, the technologic advances in CT carry a price of increased radiation exposure to each patient imaged (7,8). CT now accounts for more than 40% of all radiation exposure to patients from diagnostic imaging. There may be as many as 65 million CT examinations performed each year in the United States and as many as 260 million CTs performed yearly worldwide. Many (up to 11%) of these examinations are performed on infants and children, who are more susceptible to the adverse effects of radiation. These considerations mandate a responsibility for the P.10 radiologist and the ordering physician to limit CT to definitive indications; provide dose-efficient CT imaging protocols; offer alternative imaging techniques, especially for young children, who are at the greatest risk from radiation; work with manufacturers to limit radiation dose; and educate patients and health care providers about the potential risks of low-dose radiation.
Contrast
Administration
in
CT
Intravenous iodine-based contrast agents are administered in CT to enhance density differences between lesions and surrounding parenchyma, to demonstrate vascular anatomy and vessel patency, and to characterize lesions by their patterns of contrast enhancement. Optimal use of intravenous contrast depends upon the anatomy, physiology, and pathology of the organ of interest. In the brain, the normal blood-brain barrier of tight neural capillary endothelial junctions prevents access of contrast into the neural extravascular space. Defects in the blood-brain barrier associated with tumors,
35
stroke, infection, and other lesions enable contrast accumulation within abnormal tissue, improving its visibility. In nonneural tissues, the capillary endothelium has loose junctions, enabling free access of contrast into the extravascular space. Contrast administration and the timing of CT scanning must be carefully planned to optimize differences in enhancement patterns between lesions and normal tissues. For example, most liver tumors are supplied predominantly by the hepatic artery, whereas the liver parenchyma is supplied predominantly by the portal vein (about 70%), with a lesser contribution from the hepatic artery (about 30%). Contrast given by bolus injection into a peripheral arm vein will arrive earliest in the hepatic artery and enhance (that is, increase the CT density of) many tumors to a greater extent than the liver parenchyma. Maximal enhancement of the liver parenchyma is delayed 1 to 2 minutes until the contrast has circulated through the intestinal tract and returned to the liver via the portal vein. Differentiation of tumor and parenchyma by contrast enhancement can thus be maximized by administration of an IV bolus of contrast and by performing rapid CT scanning of the liver in the first 1 to 2 minutes following contrast administration. Helical CT is ideal for this early and rapid scanning of the liver. Oral or rectal contrast is generally required to opacify the bowel for CT scans of the abdomen and pelvis. Bowel without intraluminal contrast may be difficult to differentiate from tumors, lymph nodes, and hematomas.
CT
Artifacts
Artifacts are components of the image that do not faithfully reproduce actual anatomic structures because of distortion, addition, or deletion of information. Artifacts degrade the image and may cause errors in diagnosis (9) . Volume averaging is present in every CT image and must always be considered in image interpretation. The displayed two-dimensional image is created from data obtained and averaged from a threedimensional volume of patient tissue. Slices above and below the image that is being interpreted must be examined for sources of volume averaging that may be misinterpreted as pathology.
36
FIGURE
1.12. Beam-Hardening
Artifact. A CT image of the
abdomen is severely degraded by a beam-hardening artifact that produces dark streaks across the lower half of the image. The artifact was caused by marked attenuation of the x-ray beam by the patient's arms, which were kept at his sides owing to injury.
A
beam-hardening
artifact results from greater attenuation of low-
energy x-ray photons than high-energy x-ray photons as they pass through tissue. The mean energy of the x-ray beam is increased (the beam is “hardened―), resulting in less attenuation at the end of the beam than at its beginning. Beam-hardening errors are seen as areas or streaks of low density (Fig. 1.12) extending from structures of high x-ray attenuation, such as the petrous bones, shoulders, and hips. A motion
artifact results when structures move to different positions
during image acquisition. Motion occurs as a result of voluntary or involuntary patient movement, breathing, heartbeat, vessel pulsation, or peristalsis. Motion is demonstrated in the image as prominent streaks from high- to low-density interfaces or as blurred or duplicated images (Fig. 1.13) .
37
Streak artifacts emanate from high-density sharp-edged objects, such as vascular clips and dental fillings (Fig. 1.14). Reconstruction algorithms cannot handle the extreme differences in x-ray attenuation between very dense objects and adjacent tissue.
Principles
of
CT
Interpretation
Like all imaging analysis, CT interpretation is based on an organized and comprehensive approach. CT images are viewed in sequential anatomic order, with each slice examined with reference to slices above and below. The radiologist must seek to develop a threedimensional concept of the anatomy and pathology displayed. The study must be interpreted with reference to the scan parameters, slice thickness and P.11 spacing, administration of contrast, and artifacts. Images are oriented so that the observer is looking at the patient from below. The patient's right side is oriented on the left side of the image. Optimal bone detail is viewed at “bone windows,― generally a window width of 2,000 H and a window level of 400 to 600 H. Lungs are viewed at “lung windows,― with a window width of 1,000 to 2,000 H and window levels of about 500 to 600 H. Soft tissues are examined at a window width of 400 to 500 H and window level 20 to 40 H. Narrow windows (width of 100 to 150 H, level of 70 to 80 H) increase image contrast and aid in the detection of subtle liver and spleen lesions. PACS workstation viewing allows the interpreter to actively change window width and level settings to optimize visualization of anatomic structures.
38
FIGURE
1.13. Motion
Artifact. Breathing motion during image
acquisition duplicates the margin (arrow) of the spleen, simulating a subcapsular hematoma in this patient, who was imaged because of abdominal trauma.
39
FIGURE 1.14. Streak Artifact. Shotgun pellets produce a severe streak artifact on this CT image.
Magnetic
Resonance
Imaging
MR is a technique that produces tomographic images by means of magnetic fields and radio waves. Although CT evaluates only a single tissue parameter—x-ray attenuation—MR analyzes multiple tissue characteristics, including hydrogen (proton) density, T1 and T2 relaxation times of tissue, and blood flow within tissue. The soft tissue contrast provided by MR is substantially better than for any other imaging modality. Differences in the density of protons available to contribute to the MR signal discriminate one tissue from another. Most tissues can be differentiated by significant differences in their characteristic T1 and T2 relaxation times. T1 and T2 are features of the three-dimensional molecular environment that surrounds each proton in the tissue imaged. T1 is a measure of a proton's ability to exchange energy with its surrounding chemical matrix. It is a measure of how quickly a tissue can become magnetized. T2 conveys how quickly a given tissue loses its magnetization. Blood flow has a complex effect on the MR signal that may decrease or increase signal intensity within blood vessels. The complex physics of MR is beyond the scope of this book. In the simplest terms, MR is based on the ability of a small number of protons within the body to absorb and emit radio wave energy when the body is placed within a strong magnetic field. Different tissues absorb and release radio wave energy at different, detectable, and characteristic rates. MR scans are obtained by placing the patient in a static magnetic field 0.02 to 4 tesla (T) in strength, depending on the particular MR unit used. Low–field strength systems ( Table of Contents > Section II - Neuroradiology > Chapter 3 Craniofacial Trauma
Chapter 3 Craniofacial
Trauma
Robert M. Barr Alisa D. Gean Tuong H. Le
HEAD
TRAUMA
Imaging
Strategy
CT Imaging of acute head trauma is performed to detect treatable lesions before secondary neurologic damage occurs. Currently, this is best performed by CT for several reasons: it is quick, widely available, and highly accurate in the detection of acute intra-axial and extra-axial hemorrhage, as well as skull, temporal bone, facial, and orbital fractures. Monitoring equipment is easily accommodated. CT images must be reviewed using multiple windows. A narrow window width is used to evaluate the brain, whereas a slightly wider window width is used to exaggerate contrast between extra-axial collections and the adjacent skull, and a very wide window is used to evaluate the skull itself (see Figs. 3.1, 3.6). Contiguous 5-mm sections through the brain provide sufficient detail and can be obtained with modern scanners in less than 15 minutes. Thinner sections are used to evaluate the orbits, facial skeleton, and skull base. IV contrast media is not used in the acute setting because it
139
may mimic or mask underlying hemorrhage. When CT is performed in unconscious patients with severe head injury, it may be wise to include routine coverage of the craniocervical junction. A study by Link et al. (1) found that 18% of these patients had fractures of C1, C2, or the occipital condyles and that roughly half of all fractures were missed by plain radiographs. M R has traditionally been less desirable than CT in the acute setting because of the longer examination times, difficulty in managing life support and other monitoring equipment, and inferior demonstration of bone detail. MR, however, has been shown to be comparable or superior to CT in the detection of acute epidural and subdural hematomas and nonhemorrhagic brain injury (2,3). MR is also more sensitive to brainstem injury and, especially with fluid-attentuated inversion recovery (FLAIR) and gradient-echo pulse sequences, to acute, subacute, and chronic hemorrhage (4,5). Diffusion-weighted and diffusion tensor imaging have improved detection of both acute and chronic neuronal injury (6, 7, 8, 9). In the majority of cases, MR is the modality of choice for patients with subacute and chronic head injury and is recommended for patients with acute head trauma when neurologic findings are unexplained by CT. MR is also more accurate in predicting long-term prognosis. With the development of parallel imaging, faster sequences, improved monitoring equipment, and greater scanner availability, MR will continue to play an increasing role in the evaluation of acute head trauma.
Skull
Films
Unfortunately, plain films continue to be used in evaluating patients with acute head trauma, despite abundant evidence that they are not helpful (1 0, 1 1, 1 2). Patients who are judged to be at low risk for intracranial injury on the basis of a careful history and physical examination should be observed, and patients at high risk should be imaged by CT. Plain films virtually never demonstrate significant findings in the low-risk group and are inadequate to characterize or exclude intracranial injury in the high-risk group. Further, the absence of skull fractures on plain films clearly does not exclude
140
significant intracranial injury. In fact, in one large autopsy series of patients with fatal head injuries, only 75% had skull fractures (1 3) . The decision to obtain a head CT in the setting of trauma must be based on clinical grounds. Skull films are poor P.56 predictors of significant intracranial pathology and should not be used either to prevent or encourage further diagnostic evaluation.
Scalp
Injury
When interpreting CT scans for head trauma, it is helpful to begin by examining the extracranial structures for evidence of scalp injury or radiopaque foreign bodies. Scalp soft tissue swelling is often the only reliable evidence of the site of impact. The subgaleal hematoma is the most common manifestation of scalp injury and can be recognized on CT or MR as focal soft tissue swelling of the scalp located beneath the subcutaneous fibrofatty tissue and above the temporalis muscle and calvarium.
Skull
Fractures
Nondisplaced linear fractures of the calvarium are the most common type of skull fracture. They may be difficult to detect on CT scans, especially when the fracture plane is parallel to the plane of section. Fortunately, isolated linear skull fractures do not require treatment. Surgical management is usually indicated for depressed and compound skull fractures, both of which are seen better on CT scans than on plain films (Fig. 3.1). Depressed fractures are frequently associated with an underlying contusion. Intracranial air (“pneumocephalus―) may be seen with compound skull fractures or fractures involving the paranasal sinuses. Thin-section CT using a bone algorithm is the best method to evaluate fractures in critical areas, such as the skull base, orbit, or facial bones. Thin sections can also be helpful to evaluate the degree of comminution and depression of bone fragments.
141
FIGURE 3.1. Depressed Skull Fracture. A. Axial CT scan demonstrates a right parietal depressed skull fracture with overlying soft tissue swelling. The fracture is well seen when a wide window is used to enhance contrast between bone and soft tissue. B . The narrower window demonstrates excellent contrast between gray and white matter but fails to show the fracture. A small extra-axial hematoma is seen in the right parietal area.
Temporal
Bone
Fractures
Thin-section, high-resolution CT scanning has led to a dramatic improvement in the ability to detect and characterize temporal bone fractures. Patients with fractures of the temporal bone may present with deafness, facial nerve palsies, vertigo, dizziness, or nystagmus. Clinical symptoms are often masked in the presence of other serious injuries. Physical signs of temporal bone fracture include hemotympanum, CSF otorrhea, and ecchymosis over the mastoid process (“Battle's sign―). Temporal bone fractures may be first suspected on standard head CT scans performed to exclude intracranial injury. Findings such as opacification of the mastoid air
142
cells, fluid in the middle ear cavity, pneumocephalus, or occasionally pneumolabyrinth, should raise the suspicion of a temporal bone fracture. Optimal evaluation of a suspected temporal bone fracture requires thin-section (1 to 1.5 mm) axial and direct coronal CT imaging using a bone algorithm. With multidetector CT, thinnersection axial imaging can be performed, and coronal reformats may be adequate for interpretation. Fractures of the temporal bone can be classified as longitudinal or transverse depending on their orientation relative to the long axis of the petrous bone. If the fracture parallels the long axis of the petrous pyramid, it is termed a “longitudinal― fracture; fractures perpendicular to the long axis of the petrous bone are termed “transverse― also occur.
fractures.
“Mixed―
fracture
types
The longitudinal temporal bone fracture (Fig. 3.2) represents 70% to 90% of temporal bone fractures (1 4). It P.57 results from a blow to the side of the head. Complications include conductive hearing loss, dislocation or fracture of the ossicles, and CSF otorhinorrhea. Facial nerve palsy may occur, but it is often delayed and incomplete. Sensorineural hearing loss is uncommon.
143
FIGURE
3.2. Longitudinal
Temporal
Bone
Fracture. Axial CT
scan shows a longitudinal left temporal bone fracture (arrowheads) with opacification of the mastoid air cells. Diastasis of the left lambdoid suture (open arrow) and fractures of the sphenoid sinus (curved
arrow) and left lateral orbital wall (arrow)
are also present. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins;
1994:63.)
The transverse temporal bone fracture usually results from a blow to the occiput or frontal region. Complications are usually more severe and include sensorineural hearing loss, severe vertigo, nystagmus, and perilymphatic fistula. Facial palsy is seen in 30% to 50% of these cases and is often complete (1 4). Transverse fractures may also involve the carotid canal or jugular foramen, causing injury to the carotid artery or jugular vein. Mixed or complex temporal bone fractures represent approximately 10% of temporal bone fractures. They involve a combination of
144
fracture planes and generally follow severe crushing blows to the skull. Patients with mixed temporal bone fractures have a high incidence of associated intracranial injury.
Classification
of
Head
Injury
Traumatic head injury can be divided into primary and secondary forms. Primary lesions are those that occur as a direct result of a blow to the head. Secondary lesions occur as a consequence of primary lesions, usually as a result of mass effect or vascular compromise. Secondary lesions are often preventable, whereas primary injuries, by definition, have already occurred by the time the patient arrives in the emergency department. Primary lesions include epidural, subdural, subarachnoid, and intraventricular hemorrhage, as well as diffuse axonal injury (DAI), cortical contusions, intracerebral hematomas, and subcortical gray matter injury. Direct injury to the cerebral vasculature is another type of primary lesion. Secondary lesions include cerebral swelling, brain herniation, hydrocephalus, ischemia or infarction, CSF leak, leptomeningeal cyst, and encephalomalacia. Brainstem injury, which is also divided into primary and secondary forms, is discussed later in this chapter.
Primary
Head
Injury:
Extra-axial
Epidural hematomas are usually arterial in origin and often result from a skull fracture that disrupts the middle meningeal artery. The developing hematoma strips the dura from the inner table of the skull, forming an ovoid mass that displaces the adjacent brain. They may occur from stretching or tearing of meningeal arteries without an associated fracture, especially in children. Overall, skull fractures are seen in 85% to 95% of cases. In approximately a third of patients with an epidural hematoma, neurologic deterioration occurs after a lucid interval (1 5) . Most epidural hematomas are temporal or temporoparietal in
145
location, although frontal and occipital hematomas can also occur. Venous epidural hematomas are less common than arterial epidurals and tend to occur at the vertex, posterior fossa, or anterior aspect of the middle cranial fossa. Venous epidural hematomas usually occur as a result of disrupted dural venous sinuses. On CT, acute epidural hematomas appear as well-defined, highattenuation lenticular or biconvex extra-axial collections (Fig. 3.3) . Associated mass effect with sulcal effacement and midline shift is frequently seen. Bone windows usually demonstrate an overlying linear skull fracture. Because epidural hematomas exist in the potential space between the dura and inner table of the skull, they usually will not cross cranial sutures, where the periosteal layer of the dura is firmly attached (Fig. 3.4). Near the vertex, however, the periosteum forms the outer wall of the sagittal sinus and is less tightly adherent to the sagittal suture. Therefore, vertex epidurals, which are usually of venous origin from disruption of the sagittal sinus, can cross the midline. Occasionally, an acute epidural hematoma will appear heterogeneous, containing irregular areas of lower attenuation. This finding may indicate active extravasation of fresh unclotted blood into the collection and warrants immediate surgical
attention.
Subdural hematomas are typically venous in origin, resulting from stretching or tearing of cortical veins that traverse the subdural space en route to the dural sinuses. They may also result from disruption of penetrating branches of P.58 superficial cerebral arteries. Because the inner dural layer and arachnoid are not as firmly attached as the structures that make up the epidural space, the subdural hematoma typically extends over a much larger area than the epidural hematoma. Patients with a subdural hematoma commonly present after acute deceleration injury from a motor vehicle accident or fall. The same mechanism can cause cortical contusions and DAI, which are frequently seen in association with acute subdural hematomas.
146
FIGURE 3.3. Epidural Hematoma. Axial CT scan demonstrates a biconvex, high-attenuation, extra-axial collection causing mass effect on the right frontal lobe and mild midline shift (subfalcial herniation). Note how the epidural hematoma does not extend beyond the right coronal suture.
147
FIGURE 3.4. Epidural Versus Subdural Hematoma. Axial diagram of the brain surface in the frontal region demonstrates the characteristic locations of the epidural hematoma (EDH) compared with the subdural hematoma (SDH). Note how the EDH is located above the outer dural layer and the SDH is located beneath the inner dural layer. Only the EDH can cross the falx cerebri. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:76.)
148
FIGURE
3.5. Left Subdural and Right Epidural Hematomas.
Axial CT scan demonstrates a crescent-shaped high-attenuation collection extending along the entire left hemisphere consistent with a subdural hematoma (arrowheads). Compare the appearance with that of a small epidural hematoma seen on the right (arrow), where overlying scalp soft tissue swelling is also present. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:120.)
On axial CT, acute subdural hematomas appear as crescent-shaped extra-axial collections of high attenuation (Fig. 3.5). Small subdural hematomas may be masked by adjacent cortical bone when viewed on a narrow window width but will be apparent with an intermediate window width (Fig. 3.6). Most subdural hematomas are supratentorial, located along the convexity. They are also frequently
149
seen along the falx and tentorium. Because dural reflections form the falx cerebri and tentorium, subdural collections will not cross these structures (see Fig. 3.4). Unlike epidural hematomas, subdural hematomas can cross sutural margins and, in fact, are frequently seen layering along the entire hemispheric convexity from the anterior falx to P.59 the posterior falx. Diffuse swelling of the underlying hemisphere is common with subdural hematomas. Because of this, there may be more mass effect than would be expected by the size of the collection, and there may be little or no reduction in midline shift after evacuation of a hemispheric subdural hematoma.
FIGURE 3.6. Subdural Hematoma Seen on Intermediate Window Only. A small right temporal subdural hematoma is masked on the CT that used a narrow window (A) but is clearly seen (B) (arrowheads) with an intermediate (or subdural) window.
The CT appearance of subdural hematomas changes with time. The
150
density of an acute subdural hematoma initially increases because of clot retraction. By the time most acute subdural hematomas are imaged, the collection is hyperdense, measuring 50 to 60 H, relative to normal brain, which measures 18 to 30 H. The density will then progressively decrease as protein degradation occurs within the hematoma. Occasionally, acute subdural blood may be isodense or hypodense in patients with severe anemia or active extravasation (“hyperacute― subdural hematoma). Rebleeding during evolution of a subdural hematoma causes a heterogeneous appearance from the mixture of fresh blood and partially liquefied hematoma (Fig. 3.7). A sediment level or “hematocrit effect― may be seen either from rebleeding or in patients with clotting disorders (Fig. 3.8). Chronic subdural hematomas have low attenuation values, similar to those of CSF (Fig. 3.9). On noncontrast CT scans, it can be difficult to distinguish them from prominent subarachnoid space secondary to cerebral atrophy. Contrast enhancement can help by demonstrating an enhancing capsule or displaced
cortical
veins.
During the transition from acute to chronic subdural hematomas, an isodense phase occurs, usually between several days and 3 weeks after the acute event. Although the subdural hematoma itself is less conspicuous during this isodense phase, there are indirect signs on a noncontrast CT scan that should lead to the correct diagnosis. These include effacement of sulci, displacement of cortex with white matter “buckling,― and midline shift (Fig. 3.10) . The MR appearance of subdural hematomas depends on the biochemical state of hemoglobin, which varies with the age of the hematoma (see Chapter 4). Acute subdural hematomas are isointense to brain on T1WIs and hypointense on T2WIs. MR is particularly helpful during the subacute phase, when the subdural hematoma may be isodense or hypodense on CT scans. T1WIs will demonstrate high signal intensity caused by the presence of methemoglobin in the subdural collection. This high signal clearly distinguishes subdural hematomas from most nonhemorrhagic fluid collections. MR also reveals that subacute subdural hematomas frequently have a lentiform or biconvex appearance when seen in the
151
coronal plane (Fig. 3.11), rather than the crescent-shaped appearance that is characteristic on axial CT scans. The multiplanar capability of MR scanning is helpful in identifying small convexities and vertex hematomas that might not be detected on axial CT scans because of the similar attenuation of the adjacent bone. Subarachnoid
hemorrhage is common in head injury but is rarely
large enough to cause a significant mass effect. It results from the disruption of small subarachnoid vessels or direct extension into the subarachnoid space by a contusion or hematoma. On CT, subarachnoid
hemorrhage P.60 P.61
appears as linear areas of high attenuation within the cisterns and sulci (Fig. 3.12). Subarachnoid collections along the convexity or tentorium can be differentiated from subdural hematomas by their extension into adjacent sulci. Occasionally, the only finding is apparent effacement of sulci when the sulci are filled with small amounts of blood. In patients who are found unconscious after an unwitnessed event, detection of subarachnoid hemorrhage may indicate a ruptured aneurysm, rather than trauma, as the primary cause. In such cases, contrast-enhanced CT angiography and/or conventional catheter angiography needs to be considered.
152
FIGURE
3.7. Acute and Chronic Subdural Hematoma. Axial
CT scan demonstrates the heterogeneous appearance of superimposed acute and chronic subdural hematomas. The higher-attenuation material (open arrow) represents fresh bleeding into a chronic, low-attenuation subdural hematoma (closed arrow). Layering of acute blood products is seen in the posterior aspect of the collection (arrowhead). Midline shift or “subfalcial herniation― is also present, evidenced by displacement of the right lateral ventricle (asterisk) across the midline.
153
FIGURE
3.8. Subdural
Hematomas
With
Hematocrit
Effect.
A CT scan (A) and T2-WI (B) in two different patients show large left hemispheric subdural hematomas with fluid–fluid levels, known as the hematocrit effect. This appearance can be seen in patients with clotting disorders or in patients with rebleeding into an older subdural collection. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:89, 95.)
154
FIGURE
3.9. Chronic
Subdural
Hematoma.
Contrast-enhanced
CT scan shows a large water-density left subdural collection consistent with a chronic subdural hematoma. There is considerable mass effect with midline shift. Displaced cortical veins can be seen along the brain surface (arrowheads). (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:96.)
155
FIGURE
3.10. Subacute Subdural Hematoma on CT.
Noncontrast CT scan shows an isodense left subdural hematoma with displacement of the underlying cortex (arrows), compression of the lateral ventricle, and mild midline shift.
Hyperacute subarachnoid hemorrhage may be more difficult to detect on conventional MR than it is on CT scans, because it can be isointense to brain parenchyma on T1W and T2W images. However, FLAIR has been shown to be more sensitive than CT in detecting acute subarachnoid hemorrhage in an animal model, especially when a high volume (1 to 2 mL) is present (5). Subacute subarachnoid hemorrhage may be better appreciated on MR because of its high signal intensity at a time when the blood is isointense to CSF on CT.
156
Chronic hemorrhage on MR scans may show hemosiderin staining in the subarachnoid space, which appears as areas of markedly decreased signal intensity on T1- and T2-weighted sequences (“superficial hemosiderosis―). Subarachnoid hemorrhage may lead to subsequent hydrocephalus by impaired CSF resorption at the level of the arachnoid villi.
FIGURE 3.11. Subacute Subdural Hematoma on MR. Noncontrast coronal T1WI shows a well-defined, uniform, hyperintense extra-axial collection (asterisk) with associated mass effect on the left cerebral hemisphere. This represents a subacute subdural hematoma. The increased signal intensity on a T1-weighted sequence is attributable to methemoglobin. Subdural hematomas can appear crescent-shaped in the axial plane and biconvex in the coronal plane.
157
Intraventricular hemorrhage is commonly seen in patients with head injuries and can occur by several mechanisms. First, it can result from rotationally induced tearing of subependymal veins on the surface of the ventricles. Another mechanism is by direct extension of a parenchymal hematoma into the ventricular system. Third, intraventricular blood can result from retrograde flow of subarachnoid hemorrhage into the ventricular system through the fourth ventricular outflow foramina. Patients with intraventricular hemorrhage are at risk for subsequent hydrocephalus by obstruction, at the level of either the aqueduct or the arachnoid villi. On CT, intraventricular hemorrhage appears as hyperdense material, layering dependently within the ventricular system (see Fig. 3.17b) . Tiny collections of increased density layering in the occipital horns may be the only clue to intraventricular hemorrhage.
FIGURE 3.12. Subarachnoid Hemorrhage. Noncontrast axial CT scans in two different patients demonstrate high-attenuation material within the sulci (A) and right sylvian fissure (B) consistent with subarachnoid hemorrhage. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia:
158
Lippincott Williams & Wilkins; 1994:130, 131.)
P.62
Primary
Head
Injury:
Intra-axial
Diffuse axonal injury (DAI) is one of the most common types of primary neuronal injury in patients with severe head trauma. As the name implies, DAI is characterized by widespread disruption of axons that occurs at the time of an acceleration or deceleration injury. The affected areas of the brain may be distant from the site of direct impact; in fact, direct impact is not necessary to cause this type of injury. The incidence of DAI was likely underestimated until recently because of the difficulty in visualizing these lesions on existing imaging studies as well as on histologic specimens. DAI is much better seen by MR than CT. This factor accounts to a large degree for the increased success of MR at explaining neurologic deficits after trauma and in predicting long-term outcome. Though MR has improved the detection of DAI in patients who have suffered head trauma, the incidence of this form of injury is probably still underestimated.
Newer
imaging
methods,
such
as
diffusion-weighted
and diffusion tensor imaging with three-dimensional (3D) tractography, have shown potential in improving the detection of white matter injury in both acute and chronic DAI (6, 7, 8, 9, 1 6) . Patients with DAI are most commonly injured in high-speed motor vehicle accidents. These lesions have not been seen as a consequence of simple falls, such as when a patient falls from the standing position. Loss of consciousness typically starts immediately after the injury and is more severe than in patients with cortical contusions
or
hematomas.
CT findings in DAI can be subtle or absent. Only approximately 20% of lesions contain sufficient hemorrhage to be visible on CT scans, accounting for the low sensitivity of this modality. Most common is the finding of small, petechial hemorrhages at the gray–white
159
junction of the cerebral hemispheres or corpus callosum (Fig. 3.13) . Ill-defined areas of decreased attenuation on CT may occasionally be seen with nonhemorrhagic lesions. On MR, nonhemorrhagic DAI lesions appear as small foci of increased signal on T2WIs (T2 prolongation) within the white matter (Fig. 3.14). The lesions tend to be multiple, with as many as 15 to 20 lesions seen in patients with severe head injury. If seen on T1WIs, they appear as subtle areas of decreased intensity. Petechial hemorrhage causes a central hypointensity on T2WIs and hyperintensity on T1WIs within a few days as a result of intracellular methemoglobin. The conspicuity of DAI on MR diminishes over weeks to months as the damaged axons degenerate and the edema resolves. Residual findings might include nonspecific atrophy or hemosiderin staining, which can persist for years and is especially obvious on gradient-echo sequences (Fig. 3.15) .
160
FIGURE 3.13. The CT Appearance of DAI. Noncontrast CT scan shows punctuate, high-attenuation foci with surrounding edema in the left frontal and parietal white matter consistent with hemorrhagic DAI. Additional lesions could be seen at other levels.
FIGURE 3.14. The MR Appearance of Acute DAI. Protondensity (left) and T2-weighted (right) MR images show several adjacent foci of high signal, representing DAI in the right frontal parasagittal white matter. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:225.)
P.63 DAI is seen in characteristic locations that correlate with the severity of the trauma. Patients with the mildest forms of injury have lesions
161
confined to the frontal and temporal white matter, near the gray–white junction. The lesions typically involve the parasagittal regions of the frontal lobes and periventricular regions of the temporal lobes. Patients with more severe trauma have DAI involving lobar white matter as well as the corpus callosum, especially the posterior body and splenium (Fig. 3.16). The corpus callosum accounts for approximately 20% of all DAI lesions (1 5). Initially thought to be caused by direct impact from the falx, injury to the corpus callosum, as shown by experimental work (1 7), is most commonly caused by rotational shear forces, like all forms of DAI. The corpus callosum may be particularly susceptible to DAI because the falx prevents displacement of the cerebral hemispheres. DAI of the corpus callosum is almost always seen in association with lesions in the lobar white matter. DAI in the most severe cases involves the dorsolateral aspect of the midbrain and upper pons, in addition to the lobar white matter and corpus callosum (see “Brainstem Injury―) . Cortical contusions are areas of focal brain injury primarily involving superficial gray matter. Patients with cortical contusions are much less likely to have loss of consciousness at the time of injury than are patients with DAI. Contusions are also associated with a better prognosis than P.64 DAI. They are very common in patients with severe head trauma and are usually well seen on CT scans. Contusions characteristically occur near bony protuberances of the skull and skull base. They tend to be multiple and bilateral and are more commonly hemorrhagic than DAI. Common sites are the temporal lobes above the petrous bone or posterior to the greater sphenoid wing, and the frontal lobes above the cribriform plate, the planum sphenoidale, and the lesser sphenoid wing (Fig. 3.17a). Less than 10% of lesions involve the cerebellum (1 8). Contusions can also occur at the margins of depressed skull fractures.
162
FIGURE 3.15. The Appearance of Acute DAI on DiffusionWeighted MR. (A) Combined diffusion-weighted (DW) and (B) apparent diffusion coefficient (ADC) MR images from a patient who fell down nine steps show a focus of high signal on the combined DW image (white arrow) and dark signal on the ADC image (black arrow) within the splenium of the corpus callosum. Note that the extent of ADC abnormality is smaller than signal abnormality on the combined DW image. The reduced ADC represents the true area of acute cytotoxic injury, while the bright signal area on the combined DW image also has contribution from vasogenic edema (T2 prolongation). The T2 abnormality that appears on the combined DW image without the corresponding ADC abnormality has been termed ‘T2 shinethrough.’ (C) The T2 abnormality can be appreciated on
163
the spin-echo T2WI (black arrowhead). (D) This patient also has findings of hemorrhagic DAI involving the peripheral gray–white junction in the right frontal lobe (white arrowhead) . (Reprinted with permission from Le TH, et al. Diffusion tensor imaging with three-dimensional fiber tractography of traumatic axonal shearing injury: an imaging correlate for the posterior callosal “disconnection― syndrome: Neurosurgery 2005;56(1):E195–201.)
Case
Report.
The CT appearance of cortical contusions characteristically varies with the age of the lesion. Many nonhemorrhagic lesions are initially poorly seen but become more obvious during the first week because of associated edema. Hemorrhagic lesions are seen as foci of high attenuation within superficial gray matter (Fig. 3.17b). These may be surrounded by larger areas of low attenuation secondary to surrounding edema. During the first week, the characteristic CT pattern of mixed areas of hypodensity and hyperdensity (“salt and pepper― pattern) becomes more apparent. Occasionally, surgical decompression of the contused brain is required to alleviate severe mass effect. Areas of prior contusion can often be recognized as foci of encephalomalacia within the same characteristic locations just described. On MR, contusions appear as poorly marginated areas of increased signal on proton density and T2-weighted sequences. They are recognized because of their characteristic distribution in the frontal and temporal lobes and often have a “gyral― morphology. Hemorrhage causes heterogeneous signal intensity that varies depending on the age of the lesion (Fig. 3.18). Hemosiderin staining from hemorrhage of any cause leads to markedly decreased signal intensity on a T2WI, especially at higher field strengths. This signal loss can persist indefinitely as a marker of prior hemorrhage.
Intracerebral
Hematoma
Occasionally, intraparenchymal hemorrhage is seen that is not necessarily associated with cortical contusion but rather represents
164
shear-induced hemorrhage from the rupture of small intraparenchymal blood vessels. This lesion is known simply as an intracerebral hematoma. Intracerebral hematomas tend to have less surrounding edema than cortical contusions because they represent bleeding into areas of relatively normal brain. Most intracerebral hematomas are located in the frontotemporal white matter, although they have also been described in the basal ganglia. They are often associated with skull fractures and other primary neuronal lesions, including contusions and DAI. In the absence of other significant lesions, patients with intracerebral hematomas can remain lucid after their injury. When symptoms develop, they commonly result from the mass effect associated with an expanding hematoma. Intracerebral hematomas can also present late secondary to delayed hemorrhage, which is another cause of clinical deterioration during the first several days after head trauma (Fig. 3.19) . Subcortical gray matter injury is an uncommon manifestation of primary intra-axial injury and is seen as multiple petechial hemorrhages primarily affecting the basal ganglia and thalamus. These represent microscopic perivascular collections of blood that may result from disruption of multiple small perforating vessels. P.65 P.66 P.67 These lesions are typically seen following severe head trauma.
165
FIGURE
3.16. The MR Appearance of Chronic DAI. Coronal
gradient-echo images in a patient with a history of prior severe head trauma demonstrate numerous hypointense foci in a distribution characteristic of DAI, including the gray–white junction (open arrow), corpus callosum (closed
arrow), and
cerebral peduncle (arrowhead). Evidence of remote hemorrhage is especially conspicuous on gradient-echo sequences. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia:
Lippincott
Williams
&
166
Wilkins;
1994:235.)
FIGURE
3.17. The MR and CT Appearance of Cortical
Contusion. A. Sagittal T1WI demonstrates multiple peripheral areas of increased signal intensity involving the inferior frontal (open arrow), anterior temporal (closed arrow), and superior frontal lobes (arrowhead) consistent with subacute hemorrhage from cortical contusion. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:151.) B . Noncontrast CT scan reveals highattenuation lesions involving the bilateral inferior frontal (open arrow) and anterior temporal (closed arrow) gray matter, consistent with hemorrhagic cortical contusions. The patient also has high-attenuation fluid within the lateral ventricles (arrowhead), consistent with intraventricular hemorrhage, and diffuse high-attenuation fluid within the bilateral subarachnoid spaces of the temporal lobe (broken arrow), consistent with subarachnoid hemorrhages.
167
FIGURE 3.18. Intracerebral Hematoma. A. Axial CT scan demonstrates a high-attenuation mass within the right temporal lobe. B . The corresponding T1WI demonstrates a central region of isointensity consistent with acute hemorrhage (deoxyhemoglobin). The surrounding high signal intensity rim represents the conversion to methemoglobin that begins to form at the periphery of a hematoma. High signal in the inferior right frontal lobe (curved arrow) represents an associated frontal contusion. A small amount of subdural blood is also present bilaterally and is hyperintense (arrowheads) .
168
FIGURE
3.19. Delayed
Hemorrhage. Admission CT scan (left)
shows a small right frontal hematoma without significant mass effect (open arrow). Left parietal soft tissue swelling indicates the site of impact (closed
arrow). A follow-up CT scan (right) was
performed when the patient's clinical condition deteriorated, and it demonstrates a marked increase in the size of the hematoma with increased edema, mass effect, and compression of the ipsilateral
frontal
horn.
169
FIGURE
3.20. The MR Appearance of Carotid and Vertebral
Artery Dissection. A . T1WI with fat suppression demonstrates an acute dissection of the right internal carotid artery with surrounding intramural hematoma which appears as high T1 signal (arrow). B . Image from the same patient also demonstrates crescentic high T1 signal of the left vertebral artery (arrow), again representing an intramural hematoma from an acute dissection.
Vascular
injuries as causes of intra-axial and extra-axial
hematomas were discussed previously. Other types of traumatic vascular injury include arterial dissection or occlusion, pseudoaneurysm formation, and the acquired arteriovenous fistula. Arterial injury commonly accompanies fractures of the base of the skull. The internal carotid is the most often injured artery, especially at sites of fixation. These include its entrance to the carotid canal at the base of the petrous bone and its exit from the cavernous sinus below the anterior clinoid process. MR findings of vascular injury include the presence of an intramural
170
hematoma (best seen on a T1WI with fat suppression, Fig. 3.20) or intimal flap with dissection, or the absence of normal vascular flow void with occlusion. An associated parenchymal infarction might also be seen. There is a potential role for MR angiography in evaluating patients with suspected traumatic vascular injury. Conventional angiograms are usually needed to confirm and delineate dissections and may also show spasm or pseudoaneurysm formation in injuries to the vessel wall. A carotid
cavernous
fistula (CCF) is a communication between the
cavernous portion of the internal carotid artery and the surrounding venous plexus. The lesion typically follows a full-thickness arterial injury, resulting in venous engorgement of the cavernous sinus and its draining tributaries (e.g., the ipsilateral superior ophthalmic vein and inferior petrosal sinus). Findings may be bilateral, because venous channels connect the cavernous sinuses. A CCF most often results from severe head injury. Skull base fractures, especially those involving the sphenoid bone, indicate patients at increased risk for associated cavernous carotid injury. The CCF may also result from ruptured cavernous carotid aneurysms. On MR, the CCF may manifest as an enlarged superior ophthalmic vein, a cavernous sinus, and petrosal sinus flow voids. There may be evidence of proptosis, swelling of the preseptal soft tissues, and enlargement of the extraocular
musculature.
Diagnosis
usually
requires
selective
carotid
angiography with rapid filming to demonstrate the site of communication (Fig. 3.21). On occasion, patients present with findings weeks or months after the initial trauma. Dural fistulas are also associated with trauma. For example, they may be caused by laceration of the middle meningeal artery with resultant formation of a fistula connecting the meningeal artery to the meningeal vein. Drainage via meningeal veins prevents formation of an epidural hematoma. Patients may be asymptomatic or present with nonspecific complaints, including tinnitus.
Mechanisms
of
Primary
Head
Injuries
Early research suggested that head injuries could be explained by
171
areas of parenchymal compression and rarefaction caused by direct impact. Many authors still use the terms coup and contrecoup to describe intracranial lesions that characteristically occur on and opposite to, respectively, the side of a blow to the head. However, Gentry and others have questioned the use of these terms, which they P.68 feel incorrectly imply that neuronal injury is caused by compression and rarefaction strains subsequent to direct impact. Gennarelli et al. have shown in a primate model that all major types of intra-axial lesions, as well as subdural hematomas, can be produced purely by rotational acceleration of the head without direct impact. Only skull fractures and epidural hematomas require a physical blow to the head. Rotational acceleration causes damage by shear forces, rather than by compression-rarefaction strain. Compression-rarefaction strain is not felt to play a significant role in most head injuries.
172
FIGURE 3.21. Carotid Cavernous Fistula. A. CT scan shows fullness in the right cavernous sinus (open arrow) and right proptosis, with swelling of the extraocular muscles (closed arrows) and preseptal soft tissues (arrowheads). B . Internal carotid angiogram in a different patient shows abnormal opacification of the cavernous sinus (open arrow) and jugular vein (closed arrow) during the arterial phase. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:349.)
The character of the accelerational force influences the type of injury
173
produced. Cortical contusions and intracranial hematosis are more severe when the period of acceleration or deceleration is very short, whereas DAI and gliding contusions are associated with a longer acceleration or deceleration injury. Thus, DAI is more common in motor vehicle accidents, while contusions and hematomas are more frequent in falls.
Secondary
Head
Injury
Diffuse cerebral swelling is a common manifestation of head trauma. It may occur either because of an increase in cerebral blood volume (hyperemia) or an increase in tissue fluid content (cerebral edema). Both conditions lead to generalized mass effect, with effacement of sulci, suprasellar and quadrigeminal plate cisterns, and compression of the ventricular system. Effacement of the brainstem cisterns indicates severe mass effect and may herald impending transtentorial herniation. P.69 Cerebral swelling from hyperemia is most commonly seen in children and adolescents. The pathogenesis is poorly understood but appears to be the result of loss of normal cerebral autoregulation. Hyperemia is recognized on CT as ill-defined mass effect, effacement of sulci, and normal attenuation of brain. Acute subdural hematomas are often associated with unilateral swelling of the ipsilateral hemisphere. Diffuse cerebral edema occurs secondary to tissue hypoxia. Because of the increase in tissue fluid, edema causes decreased attenuation on CT images, with loss of gray–white differentiation. The cerebellum and brainstem are usually spared and may appear hyperdense relative to the cerebral hemispheres (Fig. 3.22). Often, the falx and cerebral vessels appear dense, mimicking acute subarachnoid hemorrhage. Focal areas of edema are frequently seen in association with cortical contusions and may contribute significantly to mass effect.
Brain
Herniation
174
Several forms of herniation are seen secondary to mass effect produced by primary intracranial injury. These are not specific for head trauma and can be seen secondary to mass effect produced by other causes as well, including intracranial hemorrhage, infarction, or neoplasm (Fig. 3.23) .
FIGURE 3.22. Diffuse Cerebral Edema. Noncontrast CT scan in an infant with diffuse cerebral edema following strangulation. There is a diffuse decrease in attenuation of the cerebral hemispheres with loss of gray–white differentiation. Sparing of the brainstem and cerebellum causes these structures to appear dense relative to the rest of the brain. Subdural hematomas are noted overlying the tentorium (arrows) .
Subfalcial
herniation, in which the cingulate gyrus is displaced across
175
the midline under the falx cerebri, is the most common form of brain herniation (see Fig. 3.7). Compression of the adjacent lateral ventricle may be seen on CT scans, as well as enlargement of the contralateral ventricle from obstruction at the level of the foramen of Monro. Both anterior cerebral arteries (ACAs) may be displaced to the contralateral side. These patients are at risk of ACA infarction in the distribution of the callosomarginal branch of the ACA, where it becomes trapped against the falx. Uncal
herniation, in which the medial aspect of the temporal lobe is
displaced medially over the free margin of the tentorium, is also common (Fig. 3.23). Uncal herniation causes focal effacement of the ambient cistern and the lateral aspect of the suprasellar cistern. Rarely, displacement of the brainstem causes compression of the contralateral cerebral peduncle against the tentorial margin, resulting in peduncular hemorrhage or infarction. The focal impression on the cerebral peduncle is known as Kernohan's notch. Mass effect on the third cranial nerve and compression of the contralateral cerebral peduncle cause a recognizable clinical syndrome characterized by a blown pupil with ipsilateral hemiparesis.
Transtentorial
Herniation
The brain can herniate either downward or upward across the tentorium. Descending transtentorial herniation is recognized by effacement of the suprasellar and perimesencephalic cisterns. Pineal calcification, usually seen at about the same level as calcified choroid plexus in the trigones of the lateral ventricles, is displaced inferiorly. Large posterior fossa hematomas can cause ascending transtentorial herniation, in which the vermis and portions of the cerebellar hemispheres can herniate through the tentorial incisura. This is much less common than descending transtentorial herniation. Posterior fossa hematomas can also cause herniation of the cerebellar tonsils downward through the foramen magnum. Finally, external herniation can occur in which swelling or mass effect causes the brain to herniate through a calvarial defect. This can be posttraumatic or occur at the time of craniotomy and prevent closure of the skull flap.
176
Hydrocephalus can occur after subarachnoid or intraventricular hemorrhage as a result of either impaired CSF reabsorption at the level of the arachnoid granulations or obstruction at the level of the aqueduct or fourth ventricular outflow foramina. Mass effect from cerebral swelling or an adjacent hematoma can also cause hydrocephalus by compression of the aqueduct or outflow foramina of the fourth ventricle. Asymmetric lateral ventricular dilatation can be produced by compression of the foramen of Monro.
Ischemia
or
Infarction
Posttraumatic ischemia or infarction can result from raised intracranial pressure, embolization from a vascular dissection, or direct mass effect P.70 on cerebral vasculature from brain herniation or an overlying extraaxial collection. In addition, patients may suffer diffuse ischemic damage from acute reduction in cerebral blood flow or from hypoxemia secondary to respiratory arrest or status epilepticus. Patterns of infarction from focal mass effect include anterior cerebral artery infarction from subfalcial herniation, posterior cerebral artery infarction from uncal herniation, and posterior inferior communicating
artery
infarction
from
tonsillar
herniation.
Ischemia
or infarction secondary to globally reduced cerebral perfusion tends to occur in characteristic “watershed zones― and is not specific for trauma (see Chapter 4) .
177
FIGURE
3.23. Brain Herniation. A. Diagram of the major types
of brain herniation. 1, subfalcial herniation; 2, uncal herniation; 3, descending transtentorial herniation; 4, external herniation; 5, tonsillar herniation. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:264.) B . Uncal herniation. Contrast-enhanced CT scan shows compression of the left aspect of the brainstem, displacement of the left posterior cerebral artery (PCA) (arrowheads), and effacement of the ambient and crural cisterns. The temporal horns of the lateral ventricles are dilated, indicating obstructive hydrocephalus. Compression of the PCA during uncal herniation can lead to a PCA infarct. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:273.)
CSF leak requires a dural tear and can occur after calvarial or skull base fractures. CSF rhinorrhea occurs subsequent to fractures in which a communication develops between the subarachnoid space and the paranasal sinuses or middle ear cavity. CSF otorrhea occurs when a communication between the subarachnoid space and middle ear occurs in association with disruption of the tympanic membrane.
178
CSF leaks can be difficult to localize and can lead to recurrent meningeal infection. Radionuclide cisternography is highly sensitive for the presence of CSF extravasation; however, CT scanning with intrathecal contrast is required for detailed anatomic localization of the defect (Fig. 3.24) . Leptomeningeal
cyst or “growing fracture― is caused by a
traumatic tear in the dura, which allows an outpouching of arachnoid to occur at the site of a suture or skull fracture. This leads to progressive, slow widening of the skull defect or suture, presumably as a result of CSF pulsations. The leptomeningeal cyst appears as a lytic skull defect on CT or plain skull films, which can enlarge over time.
Encephalomalacia Focal encephalomalacia consists of tissue loss with surrounding gliosis and is a frequent manifestation of remote head injury. It may be asymptomatic or serve as a potential seizure focus. CT demonstrates fairly well-defined areas of low attenuation, with volume loss. There may be dilation of adjacent portions of the ventricular system (Fig. 3.25). Encephalomalacia will follow CSF signal on MR sequences, except for gliosis, which appears as increased signal intensity on both proton-density and T2-weighted images. The appearance of encephalomalacia is not specific for posttraumatic injury, but the locations are characteristic: anteroinferior frontal and temporal lobes. Focal volume loss along the white matter tracts associated with cell death is known as wallerian degeneration and may be seen on CT and especially MR studies.
Brainstem
Injury
Primary The most common form of primary brainstem injury is DAI, which affects the dorsolateral aspect of the midbrain and upper pons (Fig. 3.26). The superior cerebellar peduncles and the medial lemnisci are
179
P.71 P.72 particularly vulnerable. Both the location and lack of sufficient amounts of hemorrhage make this lesion difficult to diagnose on CT scans. Brainstem DAI is nearly always seen in association with lesions of the frontal or temporal white matter and corpus callosum. This distinguishes brainstem DAI from a rare form of primary injury caused by direct impact of the free margin of the tentorium on the brainstem. Primary brainstem injury may also occur in the form of multiple petechial hemorrhages in the periaqueductal regions of the rostral brainstem (see previous discussion on subcortical gray matter injury). They are not associated with DAI, although they occur in a similar distribution. This form of injury represents disruption of penetrating brainstem blood vessels by shear strain and carries a grim prognosis.
FIGURE
3.24. CSF Leak. A. Coronal CT image with bone window
of the paranasal sinuses in a patient with chronic sinusitis and CSF leak demonstrates periosteal mucosal thickening of the sphenoid sinuses. A defect of the left superolateral wall of the sphenoid sinus is seen (arrow). B . Follow-up imaging with intrathecal contrast demonstrates contrast extravasation into the left sphenoid sinus through the defect of the left superolateral
180
wall (arrow). The exact cause of the bony defect in this patient is unknown.
FIGURE
3.25. Posttraumatic
Encephalomalacia. Admission
(A) and follow-up (B) scans in a patient with severe head trauma show the interval development of left frontal (open arrow) and right posterior temporal (closed arrow) encephalomalacia in the same locations as the initial intracerebral hematomas. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:507.)
181
FIGURE
3.26. Brainstem DAI. A. Noncontrast CT scan shows a
punctate focus of increased attenuation representing focal hemorrhage from DAI of the brainstem (arrow). Note the characteristic location in the dorsolateral aspect of the brainstem. B . T2WI in a different patient shows a hyperintense lesion in a similar location.
An extremely rare form of indirect primary brainstem injury is the pontomedullary separation or rent. As the name implies, this represents a tear in the ventral surface of the brainstem at the junction of the pons and medulla. There is a spectrum of severity ranging from a small tear to complete avulsion of the brainstem. Pontomedullary separation can occur without associated diffuse cerebral injury. This lesion is usually fatal. Secondary brainstem injury includes infarction, hemorrhage, or compression of the brainstem as a result of adjacent or systemic pathology. Brainstem infarction from hypotension-induced cerebral hypoperfusion is usually seen in conjunction with supratentorial ischemic injury. The brainstem may be relatively spared in hypoxic injury. Mechanical compression of the brainstem usually occurs in the setting of uncal herniation. There may be visible displacement or a
182
change in the overall shape of the brainstem as a result of the mass effect. Neurologic injury caused by brainstem compression may be reversible in the absence of intrinsic brainstem lesions. Brainstem lesions that occur as a result of downward herniation, hypoxia, or ischemia usually involve the ventral or ventrolateral aspect of the brainstem, in contrast to primary brainstem lesions, which are most common in the dorsolateral aspect of the brainstem. A characteristic secondary brainstem lesion is the Duret hemorrhage. This is a midline hematoma in the tegmentum of the rostral pons and midbrain seen in association with descending transtentorial herniation. It is believed to result from stretching or tearing of penetrating arteries as the brainstem is caudally displaced (Fig. 3.27). The brainstem infarct is another type of secondary brainstem injury that typically occurs in the central tegmentum of the pons and midbrain.
Penetrating
Trauma
Unlike blunt head trauma, in which diffuse injury often occurs secondary to acceleration-induced shear strain, in penetrating injury the damage is defined by the trajectory of the object. Penetrating sharp objects such as knives or glass cause tissue laceration along their course, with resultant bleeding or infarction from vascular injury. Plain films or CT can be used to confirm and localize radiopaque P.73 intracranial foreign bodies. Leaded glass and metal are hyperdense on CT scans, whereas wood is hypodense.
183
FIGURE
3.27. Duret
Hemorrhage. Noncontrast CT scan
performed 24 hours after severe head trauma shows a midline pontine hemorrhage. This type of secondary brainstem injury, known as the Duret hemorrhage, occurs in association with downward transtentorial herniation and can be distinguished from most primary brainstem injuries by its midline location (compare with Fig. 3.26). (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:282.)
Gunshot wounds are among the most common causes of penetrating head trauma. They can cause the type of injuries seen in nonpenetrating trauma as well, because significant blunt force occurs from the bullet's impact on the skull. Metallic foreign bodies such as bullet fragments often cause a significant streak artifact, which can
184
obscure underlying injury. Tilting the CT gantry to change the plane of section helps minimize this artifact. The entry and exit sites can often be distinguished by the direction of beveling of the calvarial defect or from the pattern of calvarial fracture. The bullet path can often be recognized on CT as a linear hemorrhagic strip (Fig. 3.28) . Gunshot wounds in which the bullet crosses the midline or in which small fragments are seen displaced from the main bullet are associated with a poorer prognosis. Additional complications of penetrating injury are caused by associated skull fractures and dural lacerations with resultant pneumocephalus, CSF leaks, and infection. Fragments of bone, skin, or hair that may be driven intracranially also increase the risk of subsequent
abscess
Predicting Trauma
formation.
Outcome
after
Acute
Head
The Glasgow coma scale (GCS), which stratifies patients with acute head trauma based on clinical findings, including level of consciousness, brainstem reflexes, and response to pain, helps standardize assessment of the severity of injury (Table 3.1). Mild head injury refers to a GCS score of 13 to 15, moderate head injury refers to a GCS score of 9 to 12, and severe head injury is defined as a GCS score of 8 or below. Although there is a direct correlation between the initial mortality, the GCS outcome. Likewise, identifying injuries
GCS score and subsequent morbidity and is limited in its ability to predict long-term CT findings, although they are valuable in requiring acute intervention, do not correlate well
with prognosis. There is growing evidence, however, that MR will be helpful in determining a patient's prognosis after severe head injury (9,1 8,1 9). This reflects the advantage of MR over CT in detecting brainstem injury and DAI. MR studies have shown good correlation between initial GCS and the number and distribution of DAI lesions. Numerous DAI lesions and the presence of DAI in the corpus callosum or brainstem are associated with more severe clinical findings and low initial scores on the GCS. Perhaps more important is
185
the finding that the number of DAI lesions and the presence of brainstem injury or corpus callosum DAI are associated with poor long-term outcome (1 8). The number of cortical contusions is not related to outcome, except in cases with significant mass effect. There is also a poor correlation between the presence of an isolated epidural or subdural hematoma and long-term outcome, unless transtentorial herniation is also present.
Child
Abuse
Nonaccidental trauma accounts for at least 80% of deaths from head trauma in children younger than 2 years of age (2 0). It is important to consider the possibility of child abuse and to recognize the characteristic features of these suspected cases. Skull fractures represent the second most common skeletal injury in child abuse (the most common is long bone fracture). They are only found in approximately 50% of children with intracranial injuries from abuse (2 1,2 2). In patients with suspected intracranial injury, CT should be the initial imaging study. Skull films are rarely indicated, except perhaps for documentation of cranial injury in neurologically intact children with suspected child abuse. Subdural hematomas are the most commonly recognized intracranial complication from child abuse. The association of subdural hematomas and retinal hemorrhages in children with metaphyseal long bone fractures was described as “whiplash shaken injury― by Caffey in 1946 (2 3). The mechanism was thought to be one of violent shaking, P.74 with generation of rotational and shear forces intracranially because of the weak neck musculature. The mechanism might include impact against a soft object such as a mattress, which has been shown experimentally (2 4) to increase the forces produced into the range that could cause coma, subdural hematomas, and primary brain injury, leading to the term shaken impact injury.
186
FIGURE
3.28. Gunshot Wound. A. Noncontrast CT scan shows
hemorrhage delineating the bullet's path in this despondent southpaw. There is associated intraventricular and subarachnoid hemorrhage as well as pneumocephalus and a right subdural hematoma. B . Bone window shows the typical beveled entry site (arrow) and scattered bullet fragments along the trajectory. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:193.)
Subdural hematomas in child abuse often are found in the posterior interhemispheric fissure. These are seen on CT as hyperdense collections with a flat medial border along the falx and an irregular convex lateral border. Subdural hematomas may also be found along the convexity, over the tentorial surface, at the skull base, or in the posterior fossa (see Fig. 3.22). Occasionally, low-density extra-axial fluid collections are seen in infants without any clear precipitating trauma or infection. These most often represent dilated CSF spaces, known as “benign enlargement of the subarachnoid space of infancy,― but can mimic chronic subdural hematomas. They occur in neurologically intact infants 3 to 6 months of age who present with
187
enlarging head circumference. In this setting, they require no treatment and usually regress by age 2. An old term for this condition, “external hydrocephalus,― has been abandoned by many because it fails to convey the benign nature of the condition. Epidural hematomas are not frequently seen in child abuse.
TABLE 3.1 The Glasgow Coma Scale
Eye
Opening
4 -
Best
Motor
Best
Verbal
6 - obeys
5 - oriented
3 - to voice
5 - localizes
4 - confused
2 - to pain
4 - withdraws
3 - inappropriate words
1 - none
3 - abnormal
2 - incomprehensible
flexion
words
2 - extensure
1 - nothing
spontaneous
posturing
1 - flaccid
The total score is the sum of the scores in each category.
The most common intra-axial manifestation of head injury related to child abuse is diffuse brain swelling. The initial swelling is believed to be caused by vasodilation associated with loss of autoregulation. At this stage, the injury may be reversible, despite dramatic findings on
188
CT. CT scans show global effacement of the subarachnoid space and compressed ventricles. As the brain becomes edematous, the normal attenuation of gray and white matter may appear indistinguishable or even reversed. The cerebral hemispheres will demonstrate diffusely decreased attenuation. The brainstem, cerebellum, and possibly deep gray matter structures may be spared (see Fig. 3.22). Cerebral edema in the setting of shaking injury can also occur secondary to respiratory depression, apnea, and hypoxia. The other manifestations of intra-axial injury previously P.75 described in this chapter may also be seen in child abuse, including DAI and brainstem injury. Cortical contusions occur but are considered less common, possibly because the inner surface of the skull is relatively smooth in children. In infants, head trauma may lead to tears at the gray–white junction, especially in the frontal and temporal lobes.
FIGURE 3.29. Subacute and Chronic Interhemispheric Subdural Hematomas. Midline sagittal and parasagittal T1WIs in a child demonstrate a low signal intensity chronic subdural hematoma (arrowheads) and superimposed high signal intensity
189
subacute hematoma (arrow). The presence of intracranial injury of different ages is strong presumptive evidence of child abuse. The appearance is not pathognomonic for child abuse, however, because subdural hematomas do have a propensity to rebleed.
Multiple injuries of various ages also strongly suggest child abuse. Chronic sequelae of head injury in children include chronic subdural collections (which may occasionally calcify), global cerebral atrophy, and encephalomalacia. Although CT is the modality of choice for the evaluation of acute head injury in children, MR can help identify subdural collections of various ages or hemosiderin deposits from prior hemorrhages (Fig. 3.29). The ability of MR to identify these remote intracranial hemorrhages makes it an important tool in the evaluation of suspected child abuse. In some centers, it has been proposed as a necessary complement to the skeletal series. MR is also recommended when patients are clinically stable after head injury to help determine the full extent of injury and prognosis.
FACIAL Imaging Plain
TRAUMA Strategy
Films
Many facial fractures can often be diagnosed by plain films alone and need no further imaging. Four views are usually adequate in the plain-film evaluation of acute facial trauma: the Caldwell view, a shallow Waters view, a cross-table lateral view, and a submental vertex view. When patients are acutely injured and unable to undergo upright imaging, the Caldwell and Waters views can be obtained supine in the anteroposterior projection. Films obtained in the posteroanterior projection provide better bone detail and less magnification and may be helpful if the initial films are difficult to interpret. The lateral and submental vertex views are both obtained with a horizontal beam, thus enabling the detection of air–fluid
190
levels. C T is indicated when the clinical or plain-film findings suggest complex facial fractures or complications such as extraocular muscle entrapment or optic nerve impingement. Patients with facial fractures frequently have concurrent intracranial injury, especially victims of motor vehicle accidents. Imaging of the potential intracranial injury takes precedence in the acute management of these patients. If CT of the facial bones is required in patients suspected of concurrent intracranial injury, it is usually performed after CT imaging of the brain or delayed several days until the patient is clinically stable. Either 1-mm or overlapping 3-mm sections are usually obtained through the facial bones in the axial plane using a bone algorithm. Depending on pitch and rotational speed, the overlapping sections can be reconstructed to thinner sections. The field of view should extend from the orbital roof to the superior alveolar ridge. The frontal sinus or P.76 maxillary dentition can be included if fractures are suspected in these areas. The mandible should be included when maxillary alveolar or palatal fractures are seen because of the high incidence of associated mandibular fractures in this setting. A standard algorithm with soft tissue windows can be used to evaluate potential nonosseous injury, especially in the orbits. If there is no concern for a cervical spine injury, patients can also undergo scans in the direct coronal plane for better visualization of the orbital floors, palate, and floor of the anterior cranial fossa. Coronal reformations of axial or helical acquisitions may be used when patients are unable to tolerate direct coronal scanning. Contrast is unnecessary except in the rare circumstance in which vascular injury is being considered. Occasionally, 3D reconstruction may be used for planning preoperative repair of displaced or comminuted facial fractures (Fig. 3.30) .
191
FIGURE
3.30. Three-Dimensional
Preoperative
Reconstruction
for
Planning. A . Preoperative 3D reconstruction from
a facial CT demonstrates right mandibular condylar (arrow) and comminuted parasymphyseal (arrowhead) fractures. B . Postoperative 3D reconstructed image shows interval plate fixation of both fractures (arrows) .
MR The facial bones are difficult to visualize on MR scanning because they and the adjacent aerated sinuses are relatively void of signal. CT is the preferred modality for cross-sectional evaluation of facial injuries, primarily because it provides excellent bone detail. MR may be useful for injuries to orbital contents, including the optic nerve, the globe, and the extraocular muscles. It is also useful for assessing potential vascular complications, such as arterial dissections, pseudoaneurysms, and arteriovenous fistulas, and it is the best way to evaluate trauma to the temporomandibular joint. Angiography may be indicated when clinical or radiographic evidence suggests a vascular injury. Vascular injuries are more frequent with
192
penetrating trauma, such as that occurring from gunshot or stab wounds. Fractures that extend through the carotid canal also predispose to vascular injury and may require angiographic evaluation.
Soft
Tissue
Findings
Indirect signs of facial injury on plain films can help provide objective evidence of trauma, localize the site of impact, and direct attention to areas of potential bony injury. Soft tissue swelling is the most commonly seen plain-film finding in facial trauma. It may help localize the site of impact but does not necessarily indicate associated facial fractures or other more severe injury. Paranasal sinus opacification suggests the presence of an associated fracture, particularly when air–fluid levels are seen. Fluid levels are most commonly seen in the maxillary sinus but may also be seen in the frontal or sphenoid sinuses. The ethmoids may become opacified with acute hemorrhage but are less likely to demonstrate fluid levels on plain films, probably because they contain internal septa. Air in the soft tissues is also suggestive of associated fractures, depending on location. Orbital emphysema is most commonly caused by fracture of the thin medial orbital wall. Orbital floor blowout fractures can also cause orbital emphysema (Fig. 3.31) . Occasionally, facial films reveal important findings unrelated to fracture of the facial bones. For example, the films should be scrutinized for the presence of foreign bodies that may not be clinically apparent. The craniocervical junction and upper cervical spine should be examined when included on the film. Nasopharyngeal and prevertebral soft tissue swelling can indicate hemorrhage from P.77 cervical or skull base fractures. Pneumocephalus or depressed skull fractures are also occasionally seen. Rarely, a shift of pineal calcification can be detected, indicating the presence of intracranial mass effect. Although plain films are usually no longer indicated for evaluation of head trauma, it still pays to remain alert to indirect
193
manifestations of head trauma when reviewing facial films.
FIGURE
3.31. Orbital Injuries. A. Orbital emphysema on plain
film. Air in the left orbit can be seen outlining the optic nerve (arrow) in this shallow Waters view. An ipsilateral orbital floor fracture is also evident (arrowheads). B . Orbital floor blowout fracture on CT scan. Direct coronal CT scan from the same patient shows a depressed left orbital floor fracture (arrowheads) with opacification of the ipsilateral maxillary sinus. Orbital air can be seen outlining the optic nerve. A subtle medial wall fracture is also present (arrow), which likely accounts for the large amount of orbital emphysema in this case.
Nasal
Fractures
Nasal bone fractures are the most common fractures of the facial skeleton. They can occur as an isolated injury or in association with other facial fractures. Nasal trauma frequently results in a depressed fracture of one of the paired nasal bones, without associated ethmoidal injury. An anterior blow can fracture both nasal bones as well as the nasal septum. Associated fractures of the frontal process of the maxilla can be seen. Cartilaginous nasal injury cannot be diagnosed radiographically.
194
Nasal fractures are usually clinically evident and do not require radiologic diagnosis. Films of the nasal bone may document injury but are generally not useful for patient management and are often unnecessary. Fractures of the nasal bone may be transverse or longitudinal. Longitudinal fractures can be confused with the nasomaxillary suture and nasociliary grooves, which have the same orientation. Transverse fractures of the nasal bone are more common and are easily detected because they are oriented perpendicular to the normal suture line. When films are obtained, remember to look for fractures of the anterior nasal spine of the maxilla, which may be associated with nasal fractures. One potentially serious injury that can be suggested on plain films or CT is a septal hematoma. Trauma to the septal cartilage may lead to hematoma formation between the perichondrium and cartilage, which can cause cartilage necrosis by disrupting the vascular supply. An organized hematoma can also cause difficulty in breathing and may predispose to septal abscess formation.
Maxillary
and
Paranasal
Sinus
Fractures
Fracture of the maxillary alveolus is the most common isolated maxillary fracture. It frequently results from a blow to the chin that drives the teeth of the mandible into the maxillary dental arch. These fractures are usually demonstrated by dental films or panoramic (Panorex) radiographs, but can be seen on CT if the scan is extended inferior to the level of the palate. Associated fractures of the mandible are common with this form of injury, as predicted by the mechanism. Fractures of the palatine process of the maxilla and horizontal plate of the palatine bone commonly occur in the sagittal plane near the midline (Fig. 3.32). Palatine fractures may also be seen in association with complex fractures of the midface. The most common isolated sinus fracture involves the anterolateral wall of the maxillary antrum. The fracture may be seen directly or may be suspected by the finding of a maxillary sinus fluid level in the
195
setting of acute trauma. Isolated frontal sinus fractures can also occur and may be more serious if they extend intracranially. Frontal sinus fractures may be linear or comminuted and depressed. Open (compound) frontal sinus fractures involve the P.78 posterior sinus wall (Fig. 3.33). These can lead to CSF rhinorrhea and recurrent meningitis or intracerebral abscess formation. Pneumocephalus may be seen in association with these fractures. Fractures of the medial wall and superior rim of the orbit frequently involve the frontal sinus.
FIGURE
3.32. Palate
Fracture. Axial CT scan demonstrates a
nondisplaced right palatine fracture in the characteristic parasagittal location (arrow). (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:439.)
Fractures of the sphenoid sinus are often seen in association with fractures of the orbital roof, nasoethmoid complex, midface, or temporal bone. Nondisplaced sphenoid sinus fractures may be subtle on CT. Angiography should be considered if there is a suspicion of
196
associated vascular injury involving the cavernous portion of the internal carotid artery.
Orbital
Trauma
Fractures The orbit is involved in a number of facial fractures, including the tripod, Le Fort, and nasoethmoidal complex fractures. Isolated orbital wall fractures usually involve either the medial wall or orbital floor. Medial wall fractures are detected on plain films by the presence of orbital emphysema and opacification of the adjacent ethmoid air cells. Medial wall fractures can be directly visualized well with axial or coronal CT scans. Bone displacement is usually minimal, and muscle entrapment is unusual.
FIGURE 3.33. “Open― Frontal Sinus Fracture. Noncontrast CT scan demonstrates a severely comminuted fracture involving both walls of the frontal sinus (open fracture). The frontal sinus is opacified and subcutaneous air is present (arrow). Open fractures are prone to CSF leakage and meningitis or intracerebral abscess formation. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:46.)
197
Orbital floor fractures are usually linear when seen in association with other facial fractures. These are rarely associated with entrapment. Comminuted orbital floor fractures, or blowout fractures, may be seen as an isolated injury and result from a direct blow to the eye. Intraorbital pressure is acutely increased and relieved by fracture through the orbital floor (Fig. 3.34). The orbital rim remains intact in pure blowout fractures. Blowout fractures P.79 are often associated with herniation of orbital contents through the fracture. When the inferior rectus muscle is compromised, patients will experience persistent vertical diplopia. Mild or transient diplopia can occur simply as a result of periorbital edema or hemorrhage. Rarely, fragments from an orbital floor fracture buckle upward into the orbit, an injury referred to as a “blow-in― fracture.
FIGURE
3.34. Diagram of Orbital Floor Blowout Fracture.
Sudden increase in intraocular pressure from a direct blow to the eye can lead to a comminuted fracture of the orbital floor, with herniation of orbital contents into the maxillary sinus. A fluid level in the sinus is often seen acutely secondary to bleeding.
198
(Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:478.)
FIGURE
3.35. Orbital Floor Blowout Fracture on Plain Film.
Waters view shows the major findings associated with an orbital floor blowout injury: disruption of the orbital floor (arrowheads) , soft tissue mass in the superior aspect of the maxillary sinus (open arrow), and a maxillary sinus fluid level (closed
arrow) .
(Reprinted with permission from Gean AD. Imaging of Head Trauma, Philadelphia: Lippincott Williams & Wilkins; 1994:478.)
Plain-film findings suggestive of orbital floor blowout fractures include orbital emphysema, a fluid level in the ipsilateral maxillary sinus, indistinct orbital floor on Waters view, and soft tissue representing prolapsed orbital contents in the superior aspect of the maxillary sinus (Fig. 3.35). A bony spicule may be seen in the antrum, representing the inferiorly displaced fracture fragment. Blowout P.80 fractures are best seen on direct coronal CT images (see Fig. 3.31b) .
199
These should be obtained with the patient lying prone. In the supine position, fluid and debris in the maxillary antrum will layer against the orbital floor and could obscure soft tissue that is herniated through
the
fracture.
Soft
Tissue
Injury
Penetrating foreign bodies such as bullets, metal fragments, glass, or other sharp objects account for a significant amount of traumatic injury to the orbit. Thin-section CT is the method of choice for confirming the presence and localization of foreign bodies (Fig. 3.36). CT can usually clearly define the relationship of bone fragments or foreign bodies to critical structures such as the optic nerve, globe, or extraocular muscles (Fig. 3.37). MR carries a potential risk of further injury by causing motion of intraocular ferromagnetic metal. Traumatic optic neuropathy is seen in a significant number of patients with severe head trauma and occasionally occurs in patients with relatively minor deceleration injury. Damage may be maximal initially, with unilateral blindness or decreased acuity, or may worsen in the first few days after the injury. When delayed worsening occurs, secondary optic nerve compression from edema or hemorrhage in the optic nerve sheath should be considered. Imaging studies, particularly CT scans, are indicated to detect fractures through the optic canal or orbital apex. Rarely, displaced fractures are responsible for direct injury to the optic nerve sheath. More commonly, these fractures are nondisplaced but serve as evidence of severe stress transmitted to the orbital apex. Primary optic nerve injury may occur as a result of deceleration strain causing damage to the delicate meningeal vessels or direct neural disruption. Secondary optic nerve injury may occur as a result of swelling of the optic nerve within the rigid bony canal, with subsequent mechanical compression and
vascular
compromise.
200
FIGURE 3.36. Intraocular Metallic Foreign Body. Axial (A) and coronal (B) CT scans confirm the presence of a metallic foreign body in the left globe.
201
FIGURE
3.37. Lateral Orbital Wall Fracture With
Impingement of Lateral Rectus Muscle. Noncontrast CT scan precisely localizes the site and degree of impingement on the right lateral rectus muscle in this patient with a comminuted fracture
involving
the
zygomaticofrontal
suture.
Fractures of the Zygoma The zygoma, or “cheekbone,― is one of the most common sites of injury in fractures that involve multiple facial bones. Zygomatic arch fractures may occur as an isolated finding or as part of a zygomaticomaxillary complex (“tripod,― “quadripod,― or “trimalar―) fracture. Comminution and depression are frequently seen with zygomatic arch fractures. On plain films, the zygomatic arch is best evaluated on the submental vertex view (Fig. 3.38). Deformity of the arch is a frequent finding in populations with a high incidence of facial trauma, and clinical examination may be required to differentiate acute from chronic injury. Zygomaticomaxillary complex fractures usually result from a blow to the face. The zygoma articulates with the frontal, maxillary,
202
sphenoid, and temporal bones. Fractures are somewhat variable but typically involve the zygomatic arch, zygomaticofrontal suture, infraorbital rim, orbital floor, lateral wall of the maxillary sinus, and lateral wall of the orbit. Injury to the infraorbital nerve is P.81 common secondary to fracture of the infraorbital rim at the infraorbital foramen. Diastasis of the zygomaticofrontal suture may injure the lateral canthal ligament or suspensory ligaments of the globe. Many of the fractures associated with this injury can be seen on both plain films and CT scans (Fig. 3.39). Associated findings on plain films include opacification of the ipsilateral maxillary antrum and posterior displacement of the body of the zygoma on the submental vertex view with overlying soft-tissue swelling.
FIGURE 3.38. Right Zygomatic Arch Fracture. Submental vertex view shows a comminuted, depressed right zygomatic arch fracture (black arrow). Soft tissue swelling anterior to the body of the zygoma is also seen (white arrow). (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:448.)
203
Fractures of the Midface (Le Fort Fractures) Complex fractures of the facial bones are frequently classified according to the method of Le Fort, who developed his theory by inflicting facial trauma on cadavers and analyzing the results. He described three general patterns of fractures that differ in location of the fracture plane across the face (Fig. 3.40) (2 5). The three Le Fort fractures initially described are bilateral processes. All involve the pterygoid plates, which help anchor the facial bones to the skull. Although there is great variability in complex facial fractures, and the classic Le Fort injuries are rarely seen in their pure form, they remain a convenient way to categorize and describe basic patterns of injury. Frequently, similar patterns of injury are seen on one side only and are known as “hemi–Le Forts.― Combinations also occur, such as a Le Fort I pattern on one side and a Le Fort II pattern on the other. Le Fort I or “floating palate― fracture is a horizontal fracture through the maxillary sinuses. It extends through the nasal septum and walls of the maxillary sinuses into the inferior aspect of the pterygoid plates. The fracture plane is parallel to the plane of axial CT images but is recognized by the fracture of all walls of both maxillary sinuses (Fig. 3.41). It is well seen in the coronal plane. There may be an associated midpalatal or maxillary split fracture. The Le Fort I fracture is more often seen in the pure form than either the Le Fort II or Le Fort III fractures. It occasionally may be accompanied by a unilateral zygomaticomaxillary complex fracture.
204
FIGURE 3.39. Zygomaticomaxillary Complex Fracture. A. Plain film shows diastasis of the left zygomaticofrontal suture (open arrow) and disruption of the orbital floor (closed arrow) . An associated zygomatic arch fracture was seen on submental vertex view (not shown). B . A CT scan in a different patient shows comminuted left zygomatic arch fracture (curved arrows) , with fractures of the anterior and posterolateral walls of the maxillary sinus (arrowheads). Associated signs of acute injury include soft tissue swelling and partial opacification of the maxillary sinus. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:452.)
205
Le Fort II or “pyramidal― fracture describes a fracture through the medial orbital and lateral maxillary walls. It begins at the bridge of the nose and extends in a pyramidal fashion through the nasal septum; frontal process of the maxilla; medial wall of the orbit; inferior orbital rim; superior, lateral, and posterior walls of the maxillary antrum; and midportion of the pterygoid plates. The zygomatic arch and lateral orbital walls are left intact. The Le Fort II is usually associated with posterior displacement P.82 of the facial bones, resulting in a “dish-face― deformity and malocclusion. The infraorbital nerve is frequently injured. Le Fort II fractures are rarely seen in the pure form.
FIGURE 3.40. Diagram of Le Fort Fractures. Frontal (A) and lateral (B) projections demonstrate the patterns of facial fractures as originally described by Le Fort. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:454.)
Le Fort III fracture, or “craniofacial dysjunction,― is a horizontally oriented fracture through the orbits. It begins near the nasofrontal suture and extends posteriorly to involve the nasal septum, medial and lateral orbital walls, zygomatic arch, and base
206
(superior aspect) of the pterygoid plates. Patients with a Le Fort III fracture also have dish-face deformity and malocclusion. Injury to the infraorbital nerve is less commonly seen with Le Fort III than with Le Fort II fractures. A recognizable feature on plain films is the elongated appearance of the orbits on Waters and Caldwell views.
FIGURE
3.41. Le Fort I Fracture. Axial CT scan demonstrates
comminuted fractures involving all walls of both maxillary sinuses, with associated fractures through the pterygoid plates (black arrows). Both nasolacrimal ducts are also disrupted (white arrows). Both maxillary antra are completely opacified. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:456.)
When interpreting CT scans obtained for facial trauma, it is probably best to describe the specific bones that are fractured on either side of the face. When appropriate, the Le Fort injury that best describes the distribution of fractures may also be used to categorize complex fractures.
207
Nasoethmoidal
Fractures
Nasoethmoidal complex injuries describe the constellation of findings seen as a result of a blow to the midface between the eyes. This term encompasses a wide variety of different fracture complexes P.83 that are best described by listing the specific fractures seen on CT scans. These injuries may include fractures of the lamina papyracea; inferior, medial, and supraorbital rims; frontal or ethmoid sinuses; orbital roofs; nasal bone and frontal process of the maxilla; and sphenoid bone (Fig. 3.42). These fractures have also been called orbitoethmoid or nasoethmoid–orbital fractures because of the importance of the often associated orbital injuries. There may be associated fractures of the skull base and clivus. Other findings include orbital and intracranial air, opacification of the ethmoid and frontal sinuses, and depression of the midface. Nasoethmoidal fractures can be suspected on plain films when the lateral view shows posterior displacement of nasion. Thin-section CT helps evaluate the extent of the injury and helps localize bony fragments that might encroach on the optic nerve or canal.
208
FIGURE 3.42. Nasoethmoidal Complex Fracture. Axial CT scan demonstrates a depressed fracture involving the root of the nose (curved arrow) and anterior ethmoids. Bilateral fractures of the medial orbital walls are also present (arrowheads) with bilateral orbital emphysema.
Complications of nasoethmoidal complex fractures depend on the location and extent of injury. Patients with fractures involving the floor of the anterior cranial fossa are prone to develop CSF leaks because of the high frequency of associated dural lacerations. The olfactory nerves are frequently injured when fractures extend to the cribriform plate. As mentioned earlier, orbital injuries are often seen as a component of nasoethmoid fractures. The globes or optic nerves may be damaged by displaced medial orbital wall fracture fragments.
FIGURE 3.43. Panorex Radiograph With Bilateral Mandibular Fractures. Fractures of the left mandibular angle (extending into the root of a molar tooth) and right horizontal ramus are both clearly seen on single panoramic film. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:431.)
Mandibular
fractures are extremely common in patients with
209
maxillofacial injury. Plain films are used in P.84 the initial evaluation of patients with suspected mandibular injury. The mandibular series includes posteroanterior, lateral, Towne, and bilateral oblique projections. CT or Panorex films can also be used to evaluate mandibular injury (Fig. 3.43) . Mandibular fractures can be considered either simple or compound. Simple fractures are most common in the ramus and condyle and do not communicate externally or with the mouth. Compound fractures are those that communicate internally through a tooth socket or externally through a laceration (Fig. 3.44). Fractures of the body of the mandible are almost always compound fractures. Pathologic mandibular fractures can occur at sites of infection or neoplasm. Mandibular fractures are frequently multiple or bilateral, and such fractures often involve the condyle (Fig. 3.45). Subcondylar fractures may be recognized on plain films by the “cortical ring― sign, a well-corticated density seen above the condylar neck on lateral views because of the horizontal axis of the fragment. A common pattern of injury is a unilateral condylar fracture with a contralateral fracture of the mandibular angle. The mandibular angle is also the most common site of isolated injury. Fractures of the ramus and coronoid processes are rare. Fractures through the symphysis or parasymphyseal region are common but difficult to diagnose on plain films because of the obliquity of the fracture plane. P.85 Fractures involving the dentoalveolar complex are also often missed on mandibular series and require intraoral dental films or CT for evaluation. Bilateral fractures through the mandibular body or comminuted fractures can lead to airway obstruction from posterior displacement of the tongue and free mandibular fragment.
210
FIGURE
3.44. Compound Fracture of the Mandible. Oblique
view of the mandible demonstrates a posterior ramus fracture extending through the adjacent tooth socket (open arrow). A contralateral fracture of the horizontal ramus is also present (closed
arrow). (Reprinted with permission from Gean AD.
Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:467.)
211
FIGURE 3.45. Mandibular Condylar Fracture. A. Plain film (Towne projection) shows a displaced right subcondylar fracture (open arrow). B . Axial CT in a different patient shows a right condylar fracture (open arrow) and an associated parasymphyseal fracture (closed arrow). The latter fracture is easily missed on plain films because of the oblique fracture plane. (Reprinted with permission from Gean AD. Imaging of Head Trauma. Philadelphia: Lippincott Williams & Wilkins; 1994:464.)
SUGGESTED
READINGS
Intracranial
Injury
Eelkema EA, Hecht ST, Horton JA. Head trauma. In: Latchaw RE, ed. MR and CT Imaging of the Head, Neck, and Spine. 2nd ed. St. Louis: CV Mosby, 1991:203–265. Gean AD. Imaging of Head Trauma. New York: Raven Press, 1994.
212
Cranial and Skull Base Injury Davidson HC. Imaging of the temporal bone. Neuroimaging Clin North Am 2004;14(4):721–760. Holland BA, Brant-Zawadzki M. High-resolution CT of temporal bone trauma. AJNR Am J Neuroradiol 1984;5:291–295.
Head Trauma in Child Abuse Merten DF, Radkowski MA, Leonidas JC. The abused child: a radiological
reappraisal.
Radiology
1983;146:377–381.
Petitti N, Williams DW III. CT and MR imaging of nonaccidental pediatric head trauma. Acad Radiol 1998;5(3):215–223. Sato Y, Smith WL. Head injury in child abuse. Neuroimaging Clin North
Facial
Am
1991;1:475–492.
Trauma
DelBalso AM, Hall RE. Mandibular and dentoalveolar fractures. Neuroimaging
Clin
North
Am
1991;1:285–303.
Kassel EE, Gruss JS. Imaging of midfacial fractures. Neuroimaging Clin North Am 1991;1:259–283. Som PM, Brandwein MS. Sinonasal facial fractures and postoperative findings. In: Som PM, Curtin HD, eds. Head and Neck Imaging. 4th ed. St. Louis: CV Mosby, 2003:374–438.
REFERENCES
213
1. Link TM, Schuierer G, Hufendiek A, et al. Substantial head trauma: value of routine CT examination of the cervicocranium. Radiology 1995;196:741–745. 2. Gentry LR, Godersky JC, Thompson B, et al. Prospective comparative study of intermediate-field MR and CT in the evaluation of closed head trauma. AJR Am J Roentgenol 1988;150:673–682. 3. Orrison WW, Gentry LR, Stimac GK, et al. Blinded comparison of cranial CT and MR in closed head injury evaluation. AJNR Am J Neuroradiol
1995;15:351–356.
4. Noguchi K, Ogawa T, Seto H, et al. Subacute and chronic subarachnoid hemorrhage: diagnosis with inversion-recovery MR imaging. Radiology 1997;203(1):257–262.
fluid-attenuated
5. Woodcock RJ Jr, Short J, Do HM, et al. Imaging of acute subarachnoid hemorrhage with a fluid-attenuated inversion recovery sequence in an animal model: comparison with noncontrast-enhanced CT. AJNR Am J Neuroradiol 2001;22(9):1698–1703. 6. Alsop DC, Murai H, Detre JA, McIntosh TK, Smith DH. Detection of acute pathologic changes following experimental traumatic brain injury using diffusion-weighted magnetic resonance imaging. J Neurotrauma 1996;13:515–521. 7. Liu AY, Maldjian JA, Bagley, LJ, et al. Traumatic brain injury: diffusion-weighted MR imaging findings. AJNR Am J Neuroradiol 1999;20:1636–1641. 8. Arfanakis K, Haughton VM, Carew JD, et al. Diffusion tensor MR imaging in diffuse axonal injury. AJNR Am J Neuroradiol
214
2002;23(5):794–802. 9. Huisman TA, Schwamm LH, Schaefer PW, et al. Diffusion tensor imaging as potential biomarker of white matter injury in diffuse axonal injury. AJNR Am J Neuroradiol 2004;25(3):370–376. 10. Masters SJ. Evaluation of head trauma: efficacy of skull films. AJR Am J Roentgenol 1980;135:539–547. 11. Bell RS, Loop JW. The utility and futility of radiographic skull examination for trauma. N Engl J Med 1971;284:236–239. 12. Hackney DB. Skull radiography in the evaluation of acute head trauma: a survey of current practice. Radiology 1991;181:711–714. 13. Adams JH. Pathology of nonmissile head injury. Neuroimaging Clin North Am 1991;1:397–410. 14. Gentry LR. Temporal bone trauma. Neuroimaging Clin North Am 1991;1:319–340. 15. Gentry LR. Imaging of closed head injury. Radiology 1994;191:1–17. 16. Le TH, Mukherjee P, Henry RG, et al. Diffusion tensor imaging with three-dimensional fiber tractography of traumatic axonal shearing injury: an imaging correlate for the posterior callosal “disconnection― syndrome. Neurosurgery 2005;56(1):E195–201. 17. Gennarelli TA, Thibault LE, Adams JH, et al. Diffuse axonal injury and traumatic coma in the primate. In: Dacey RG Jr, Winn HR, Rimel RW, et al, eds. Trauma of the Central Nervous System.
215
New York: Raven Press, 1985:169–193. 18. Gentry LR. Head trauma. In: Atlas SW, ed. Magnetic Resonance Imaging of the Brain and Spine. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 1996;611–647. 19. Shanmuganathan K, Gullapallo RP, Mirvis SE, et al. Whole brain apparent diffusion coeffiecient in traumatic brain injury: correlation with Glasgow Scale score. AJNR Am J Neuroradiol 2004;25(4):539–544. 20. Bruce DA, Zimmerman RA. Shaken impact syndrome. Pediatr Ann 1989;18:482–494. 21. Merten DF, Osborne DRS, Radkowski AM. Craniocerebral trauma in the child abuse syndrome: radiological observations. Pediatr Radiol 1984;14:272–277. 22. Zimmerman RA, Bilaniuk LT. Pediatric head trauma. Neuroimaging Clin North Am 1994;4:349–366. 23. Caffey J. Multiple fractures in the long bones of infants suffering from chronic subdural hematoma. AJR Am J Roentgenol 1946;56:163–173. 24. Duhaime AC, Gennarelli TA, Thibault LE, et al. The shaken baby syndrome. A clinical, pathological, and biomechanical study. J Neurosurg 1987;66:409–415. 25. Le Fort R. Etude experimental sur les fractures de la machoire superieure, parts I, II, III. Rev Chir (Paris) 1901;23:208–227.
216
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section II - Neuroradiology > Chapter 4 Cerebrovascular Disease
Chapter 4 Cerebrovascular
Disease
Howard A. Rowley Stroke is a clinical term applied to any abrupt nontraumatic brain insult—literally “a blow from an unseen hand.― Strokes are caused by either brain infarction (75%) or hemorrhage (25%) and must be distinguished from other conditions that cause abrupt neurologic deficits. Infarction is a permanent injury that occurs when tissue perfusion is decreased long enough to cause necrosis, typically as the result of occlusion of the feeding artery. Transient ischemic
attacks (TIAs) are defined as transient neurologic
symptoms or signs lasting less than 24 hours, which may serve as a “warning sign― of an infarction occurring in the next few weeks or months. TIAs are often caused by temporary occlusion of a feeding artery. Hemorrhage is seen when blood ruptures through the arterial wall, spilling into the surrounding parenchyma, subarachnoid space, or ventricles. Stroke is the third leading cause of death in the United States and a major source of long-term disability among survivors. The approach to treatment of ischemic stroke has been largely preventative or supportive in the past, but approval of IV thrombolysis for acute stroke and neuroprotective drug development have made rapid imaging and intervention a critical part of stroke management. The patient with hemorrhage may harbor an aneurysm, vascular malformation, or other condition, and each condition involves important differences in treatment options. The radiologist plays a
217
critical role in the triage and evaluation of all stroke patients. Selection of the proper imaging technique, recognition of early ischemic changes, differentiation of stroke from other brain disorders, and recognition of important stroke subtypes can have a significant impact on therapy and outcome. This chapter reviews the pathophysiology of stroke, the time course of findings on CT and MR, patterns of arterial and venous occlusions, and overall radiologic approach to evaluation of the stroke patient.
ISCHEMIC
STROKE
Etiology Despite our best clinical efforts, no clear source is ever identified in up to a quarter of patients with brain infarction. Among those with an established mechanism, about two thirds of infarcts are caused by thrombi and one third are caused by emboli. Thrombi are formed at sites of abnormal vascular endothelium, typically over an area of atherosclerotic plaque or ulcer. A large-artery thrombosis in the neck may or may not cause distal infarction, depending on the time course of occlusion and available collateral supply. Small-vessel thrombi frequently occur in end arteries of the brain, accounting for about one fifth of infarcts (“lacunes―). Emboli may arise from the heart, aortic arch, carotid arteries, or vertebral arteries, causing infarction by distal migration and occlusion. There is obviously overlap between the thrombotic and embolic groups, since the majority of emboli begin as thrombi somewhere more P.87 proximal in the cardiovascular tree (hence the practical term “thromboembolic disease―). Vasculitis, vasospasm, coagulopathies, global hypoperfusion, and venous thrombosis
each
account for 5% or fewer of acute strokes, but are important to recognize because of differing treatment and prognosis. A patient's age, medical history, and type of stroke seen will help establish the major etiologic considerations (Table 4.1 ).
218
Congenital heart Cardiac emboli Atherosclerosis
disease
Blood dyscrasias Atherosclerosis Cardiac emboli Meningitis Drug abuse Coagulopathy Arterial dissection Arterial dissection Amyloid Trauma Coagulopathy Vasculitis ECMO Vasculitis Venous Venous
thrombosis thrombosis
Venous thrombosis ECMO, extracorporeal Pediatric
Young
membrane Adult
oxygenation.
Elderly
TABLE 4.1 Differential Diagnosis of Ischemic Stroke by Age
Pathophysiologic Changes
Basis
Brain Metabolism Vulnerability
and
for
Imaging
Selective
Neurons lead a precarious life. The brain consumes 20% of the total cardiac output to maintain its minute-to-minute delivery of glucose and oxygen. Because the brain holds no significant long-term energy
219
stores (e.g., glycogen, fat), disruption of blood flow for even a few minutes will lead to neuronal death. The extent of injury depends on both the duration and degree of ischemia. Minor reduction in perfusion is initially compensated for by increased extraction of substrate, but injury becomes inevitable below a critical flow threshold (10 to 20 mL/100 g tissue/min vs. normal 55 mL/100 g/min). Certain cell types and neuroanatomic regions show selective vulnerability to ischemic injury. Gray matter normally receives 3 to 4 times more blood flow than white matter and is therefore more likely to suffer under conditions of oligemia. Some subsets of neurons (e.g., cerebellar Purkinje cells, hippocampal CA-1 neurons) are injured more readily than others, possibly because of their greater concentrations of receptors for excitatory amino acids. The slower-metabolizing capillary endothelial cells and white matter oligodendrocytes are more resistant to ischemia than gray matter but will also die when deprived of nutrients. Cells served by penetrating end arteries or those residing in the watershed zone between major territories have no alternate route for perfusion and are therefore more prone to infarction. Damage will likely be more severe in a patient with an incomplete circle of Willis than in one with a complete arterial collateral pathway.
Imaging
Findings
in
Acute
Ischemia
Ischemia causes a cascade of cellular level events leading to the gross pathologic changes detected in clinical imaging. Failure of membrane pumps permits efflux of potassium ions (K+ ) and simultaneous influx of calcium ions (Ca2 + ), sodium ions (Na + ), and water. This leads to cellular (“cytotoxic―) edema, observed clinically as increased water content in the affected region. Changes in brain water are key to understanding signs of infarction by CT and MR. Even a small increase in water content causes characteristic decreased attenuation on CT, low signal on T1WIs, and high signal on T2- and diffusion-weighted MR. This edema peaks 3 to 7 days postinfarction and is maximal in the gray matter. A smaller component of vasogenic edema also develops as the more resistant
220
capillary endothelial cells lose integrity. (In contrast, tumorassociated edema is primarily vasogenic and preferentially affects the white matter; see Chapter 5 .) Careful inspection of CT and MR images captured within minutes to a few hours after vessel occlusion can give clues to ischemic injury, even before gross tissue edema or mass effect are seen. These “hyperacute― signs primarily relate to morphologic changes in the vessels rather than density or signal changes in the parenchyma. On CT, the actual thrombus may occasionally be seen in larger intracranial branches, resulting in the “hyperdense artery sign― (Fig. 4.1 ). On MR, the normal black signal of flowing blood within the lumen (“flow void―) is immediately lost and may be replaced by abnormal signal representing clotting or slow flow (Fig. 4.2 ). Loss of the flow void is best seen acutely in the large vessels (carotid siphon, vertebrobasilar vessels, middle cerebral artery [MCA] branches). Dissolution of clot and improved collateral flow may occur within the first few days, leading to re-establishment of flow void on follow-up MR exams.
Acute MCA Ischemia on CT: Insular Ribbon and
Lentiform
Nucleus
Edema
CT scans done within 6 hours of MCA occlusion will commonly exhibit the “insular ribbon sign,― a subtle but important blurring of the gray–white layers of the insula caused by early edema (Fig. 4.3 ). Early edema may also be proximal middle cerebral artery sign). MR exams in the first few gray–white borders and slight to undergo infarction. However,
most conspicuous in the putamen in occlusions (lentiform nucleus edema hours may show a similar loss of crowding of sulci in areas destined the most sensitive imaging
sequence for detection of brain ischemia is diffusion-weighted MR imaging (DWI), which may turn positive minutes after infarction begins, well before a CT would show even subtle signs. Hyperintense signal on DWIs (“light-bulb sign―) precedes T2 hyperintensity, which typically develops at 6 to 12 hours postictal. (Fig. 4.4 ).
221
CT
Screening
for
Thrombolysis
Careful but rapid interpretation of CT scans is particularly important in patients who are candidates for thrombolytic drug P.88 P.89 P.90 treatment (e.g., tissue plasminogen activator [tPA]). Administration of intravenous tPA within 3 hours of stroke onset has been reported to improve neurologic outcome, provided that rigid inclusion and exclusion treatment criteria are met. A screening CT is examined to exclude patients with brain hemorrhage, masses, or other structural abnormalities that contraindicate thrombolysis. Patients with extensive edema on their initial CT scan may be at particularly high risk for reperfusion hemorrhage, so these patients should be excluded
from
thrombolytic
treatment.
Although
universal
guidelines
are not agreed upon, patients with edema affecting more than one third of the MCA territory should generally be excluded. More subtle baseline changes, such as an isolated insular ribbon sign or limited lentiform nucleus edema alone, are not considered contraindications for thrombolysis. Current work suggests that perfusion-sensitive CT and MR techniques may also prove useful in identifying ischemic but still salvageable tissue (ischemic penumbra) to help guide selection of patients for acute treatment beyond 3 hours. The treatment window of opportunity may also widen beyond 3 hours as intraarterial interventions and neuroprotective drugs are introduced in clinical use.
222
FIGURE 4.1. Hyperdense Artery Sign and Early Edema on CT. Three hours postocclusion, high density is seen in the proximal right middle cerebral artery (MCA) (arrows ). Acute thrombus fills the lumen. Low attenuation and loss of gray–white distinction are seen in the insula (asterisk ), posterior putamen (P), and majority of the MCA cortex (within dashed lines).
223
FIGURE
4.2. Loss of Flow Void. A. Six hours after right internal
carotid occlusion, there is a loss of vascular flow voids in the internal carotid and middle cerebral artery branches (arrows ) compared with the patent left side (open arrows ). Hyperintensity is developing in the right posterior sylvian region, indicative of early edema on this T2WI. B . A section below shows complete occlusion of the right internal carotid artery in its cavernous segment (arrowhead ), with normal flow void preserved on the left (arrow ). An older lacune in the pons is also seen.
224
FIGURE
4.3. Insular Ribbon Sign. A. A noncontrast CT done 4
hours after right middle cerebral artery (MCA) occlusion shows decreased attenuation and loss of gray–white borders in the right insular region (arrows ). B . Diagram of the insula in transverse and coronal planes. The insular cortex, claustrum, and extreme capsule are infarcted due to occlusion of the MCA (arrow ) beyond the lateral lenticulostriate vessels. (From Truwit CL, et al. Loss of the insular ribbon: another early CT sign of acute middle cerebral artery infarction.
Radiology
1990;176:801–806;
225
used
with
permission.)
FIGURE 4.4. Edema in Early Ischemia. This patient was found unresponsive with unknown time of symptom onset. Edema is detected as high signal intensity and mild sulcal effacement in the left middle cerebral artery territory on T2W transverse images. Hyperintensity on the diffusion-weighted image (DWI) and hypointensity on the apparent diffusion coefficient (ADC) map are characteristic of cytotoxic edema in acute ischemia. Note preferential gray matter involvement during early ischemia. These images suggest the stroke is approximately 4 to 8 hours old.
226
FIGURE
4.5. Diffusion-Perfusion
Mismatch
in
Acute
Ischemia.
This 86-year-old woman with a history of atrial fibrillation developed sudden right hemiplegia and aphasia. The noncontrast CT shows subtle low attenuation in the left putamen, insula, and sylvian cortex (arrows ). On the T2WI, the cortical gray matter shows mild edema, confirmed to represent cytotoxic edema on the diffusion-weighted image (DWI) and the apparent diffusion coefficient (ADC). Fluidattenuated inversion recovery (FLAIR) shows cortical edema and stasis in the left middle cerebral artery. Perfusion-weighted images (mean transit time [MTT] and cerebral blood volume [CBV]) show a larger area at risk, extending into the parietal lobe (MTT defect in white dashes; DWI lesion superimposed in black dashes). The hypoperfused tissue not yet infarcted is considered tissue at risk, or the ischemic penumbra. Diffusion lesions tend to “grow into― severe surrounding perfusion lesions if untreated. The follow-up CT (CT-F/U) shows extension of infarction into the penumbral tissue identified by MTT.
Diffusion-Weighted
MR
in
227
Acute
Ischemia
DWI uses a novel form of MR tissue contrast to noninvasively detect ischemic changes within minutes of stroke onset. DWIs are acquired by applying a strong gradient pair that sensitizes the images to microscopic (brownian) water motion. Brain water diffusion rates fall rapidly during acute ischemia, recovering to normal over days or weeks in infarcted tissues. Because random water motion is slowed down in areas of acute ischemia, the early infarct stands out as bright signal on DWIs, compared to dark signal (dephasing) in the normal areas. Acute stroke patients may show clear DWI changes hours before any abnormality can be seen on spin-echo T2WIs (Fig. 4.5 ). This can also be a useful way to distinguish new ischemic areas (high signal on DWIs) from older lesions (normal or low signal on DWIs). Via a series of different diffusion gradient strengths, the process may also be quantified in an apparent diffusion coefficient (ADC). The ADC reflects “pure― diffusion behavior, free of any underlying T2 contributions (“shine through― or “dark through―). DWI acquisition is facilitated using echo-planar MR systems, with their inherently faster, stronger gradients and rapid digitization equipment. P.91
Fluid-Attenuated Inversion (FLAIR) in Ischemia
Recovery
FLAIR allows heavy T2 weighting of the parenchyma while simultaneously suppressing free water signal from the CSF. These techniques increase the conspicuity of T2 changes in ischemia. FLAIR is not inherently better than T2 MR for early detection of ischemia, but it may be particularly helpful in detecting small lesions in the cortex and for exclusion of acute subarachnoid hemorrhage.
Subacute
and
Chronic
Ischemia
In the subacute phase, edema leads to mass effect, ranging from slight sulcal effacement to marked midline shift with brain herniation, depending on the size and location of infarct. These changes peak at 3 to 7 days, with progressive brain softening
228
(encephalomalacia) ensuing thereafter. One potential imaging pitfall, the “fogging effect,― may be encountered on CTs obtained during the second week after infarction, while edema and mass effect are subsiding. At this stage, decrease in edema and accumulation of proteins from cell lysis balance one another such that brain morphology and density in the injured region can appear nearly normal by CT. Fogging effects are much less of a problem on MR because of its greater tissue sensitivity, particularly when contrast is used (Fig. 4.6 ). Edema or mass effect that persists beyond 1 month effectively rules out simple ischemia and should raise the possibility of recurrent infarction or an underlying tumor.
FIGURE 4.6. Fogging Effect in Subacute Infarction. As edema and mass effect subside, but before development of atrophy, infarcts may be inconspicuous on unenhanced CT or MR. A . The T2WI is essentially normal in the occipital regions 13 days after right posterior cerebral artery infarction. B . T1WI after gadolinium administration shows enhancement of the infarcted deep right occipital cortex (arrow. )
229
In the weeks and months following infarction, macrophages remove dead tissue, leaving a small amount of gliotic scar and encephalomalacia behind. CSF takes up the space previously occupied by brain. The affected corticospinal tract atrophies (wallerian degeneration), leading to a shrunken appearance of the ipsilateral cerebral peduncle. If hemorrhage accompanied the infarct, hemosiderin may be seen grossly or detected as signal hypointensity by T2WIs. Widening of adjacent sulci and “ex vacuo― dilatation of the ventricle occurs adjacent to the infarcted area (Fig. 4.7 ).
Hemorrhagic Infarction
Transformation
of
Reperfusion into infarcted capillary beds may lead secondarily to gross or microscopic hemorrhage, seen in up to half of infarcts. In most cases this takes the form of microscopic leakage (diapedesis) of red blood cells, but on rare occasions a frank hematoma will form. Physical disruption of the capillary endothelial cells, loss of P.92 vascular autoregulation, and anticoagulation or use of thrombolytics may all contribute to the development of these hemorrhages. Patients may develop headaches at the time of bleeding but commonly have no new symptoms, presumably because the hemorrhage occurs within brain areas that are already dead or dysfunctional. Hemorrhagic infarction is confined to the territory of the infarcted vessel, whereas primary hemorrhage does not necessarily respect vascular boundaries. Intraventricular extension is uncommonly seen with hemorrhagic transformation and should raise the possibility of another process (such as hypertensive bleed or a ruptured arteriovenous malformation [AVM]).
230
FIGURE 4.7. Chronic Infarction. Cystic encephalomalacia is present in the right middle cerebral artery territory on a MR of a 7month-old with neonatal infarction. Note cystic changes approaching CSF on all sequences, including diffusion-weighted (DWI) and apparent diffusion coefficient (ADC), with minimal gliosis. There is volume loss with widening of the ipsilateral ventricle (ex vacuo dilatation).
231
FIGURE
4.8. Petechial Hemorrhage and Gyral Enhancement in
Subacute Infarction. A. Precontrast T1WI shows mild effacement of sulci in the right middle cerebral artery territory. A few subtle areas of bright signal intensity scattered along the cortex indicate areas of petechial hemorrhage or laminar necrosis (arrows ). B . Postcontrast T1WI demonstrates marked gyral enhancement, a hallmark of subacute infarction.
The peak time for hemorrhagic transformation is at about 1 to 2 weeks postinfarction. It is usually manifested as a serpiginous line of petechial blood following the gyral contours of the infarcted cortex. These dots of hemorrhage are often patchy and discontinuous. On CT a faint line of high attenuation is observed, and on MR bright signal is seen along the affected gyrus on the unenhanced T1WI because of methemoglobin (Fig. 4.8 .A). (Alternate explanations for this bright signal have been offered, P.93 including laminar necrosis or calcification related to infarction; the practical point is to recognize this appearance as a feature of ischemia.) The petechial gyral pattern is not seen in primary brain
232
hemorrhage and can be helpful in confirming the underlying ischemic etiology of a suspicious lesion. This is considered a normal part of the evolution of an infarct. Management in the presence of petechial hemorrhage is controversial, but many neurologists continue anticoagulation if there is a well-documented embolic source. More extensive hemorrhagic transformation of the infarcted tissue may lead to the formation of a gross parenchymal hematoma. Here, the blood does not conform to a gyrus and may form a clot that is indistinguishable from a primary hematoma. Large cortical infarcts are at somewhat higher risk for this type of change, compared with limited cortical or subcortical lesions. Catastrophic hemorrhagic transformation can also follow thrombolysis, particularly when treatment is delayed or the baseline CT shows extensive edema. In contrast to the petechial gyral transformation described above, gross parenchymal hematomas tend to occur earlier and are more commonly associated with clinical deterioration. Confluent hematomas seen on infarct follow-up studies should be reported promptly since anticoagulation therapy is contraindicated, even when the finding is incidental.
Use of Contrast in Ischemic Stroke CT
Contrast
A noncontrast CT remains the radiologic exam of choice for emergency assessment of suspected acute stroke An unenhanced study is necessary to help triage the patient. It serves to rule out hemorrhage, may define patterns and extent of ischemic injury, shows areas of abnormal vascular calcification (e.g., giant aneurysms), and excludes mass lesions. This is important firstline information needed by the clinician faced with determining the need for lumbar puncture, vascular surgery, anticoagulation, thrombolysis, cardiac evaluation, or other therapies. However, all
233
acute stroke CTs should be reviewed on a scanning console or a picture archiving and communication system, because the unenhanced study may rarely show the need for intravenous contrast. A nonstroke lesion such as a tumor, abscess, or an isodense subdural hematoma might be suspected on the noncontrast exam and then be shown to better advantage with contrast. Older studies had suggested that contrast is contraindicated in brain infarction. They cited a slightly increased risk of seizures and other untoward CNS effects, presumably caused by a toxic effect of the contrast as it leaks through the abnormal blood-brain barrier. Most of these data, however, were based on studies that used ionic contrast media. Recent CT protocols have safely used contrast, not only to exclude tumor or infection but also to evaluate vessels (CT angiography) and blood delivery (CT perfusion). An intact blood-brain barrier normally excludes contrast from the brain. Leakage of macromolecular contrast agents through damaged vessels leads to local accumulation of iodine, seen as high attenuation (enhancement) of infarcted parenchyma. Breakdown of the blood-brain barrier underlies both hemorrhagic transformation and contrast enhancement of infarctions. Not surprisingly, then, these processes are seen at roughly the same time and often in combination. As with petechial gyral hemorrhage, a gyral pattern of enhancement (by CT or MR) is highly specific evidence of an underlying infarction. CT-detected enhancement of infarcted brain parenchyma typically begins at about 1 week, peaks at 7 to 14 days, often assumes a gyral pattern, and is less commonly observed in subcortical regions (Fig. 4.8 .B). Enhancement is seen in about half of patients during the 1st week and in about two thirds of patients between weeks 1 and 4. As gliosis ensues and the blood-brain barrier is repaired, enhancement fades and then resolves by 3 months.
MR
Contrast
Most of the comments regarding the strategy, pathophysiology, and enhancement patterns for CT also generally hold true for contrast in
234
MR. Intravenous gadolinium contrast agents are very well tolerated by stroke patients and may give valuable information not readily available from the noncontrast MR. Stasis of gadolinium within vessels or leakage of contrast through an abnormal blood-brain barrier will shorten T1 relaxation of adjacent protons, leading to hyperintensity (enhancement) on T1WIs. As with CT, a noncontrast MR sequence is mandatory before contrast is given, since enhancement and subacute blood both appear hyperintense on T1WIs. (This will be discussed in the “Hemorrhage― section.) An intravenous bolus of contrast may also be captured dynamically using rapid imaging techniques to produce a family of perfusionweighted images to help identify ischemic regions. Intravascular enhancement on MR is commonly seen in the infarcted territory during the first week. This may be caused by slow flow or vasodilatation leading to stasis of gadolinium, likely in both arteries and veins. The intravascular enhancement pattern may be detected within minutes of vessel occlusion, is seen in a majority of cortical infarcts at 1 to 3 days, and resolves by 10 days. The proximal trunks of more distally occluded arteries and leptomeningeal cortical channels are most prominently involved (Fig. 4.9 ). The area of vascular enhancement may extend beyond the T2 hyperintensity, possibly indicating recruitment of collateral supply at the ischemic border. Meningeal enhancement, which attends meningitis, and dural enhancement seen postoperatively can superficially resemble intravascular enhancement, but the distinction should be obvious on clinical grounds. MR intravascular enhancement helps identify early strokes, indicates ongoing slow flow, and has no obvious CT counterpart.
235
FIGURE
4.9. Intravascular Stasis and Enhancement in Acute
Infarction.
Postcontrast
T1
and
fluid-attenuated
inversion-recovery
(FLAIR) images in acute left middle cerebral artery (MCA) infarction. Mild sulcal effacement and prominent enhancement of sylvian branches of the MCA (arrows ) are evident on T1. As seen here, FLAIR can show similar vascular signs of stasis, either before or after contrast. Intravascular enhancement is typically seen only during the first 10 days after stroke.
P.94 MR parenchymal enhancement occurs in a similar pattern to that seen on CT (and with the same time course seen by nuclear medicine infarct scans of the past). It may occur as early as day 1, but more typically begins after the first week, a time when intravascular enhancement is waning (Fig. 4.10 ). Reperfusion after thrombolysis can lead to early enhancement. Virtually all cortical infarcts enhance on MR at 2 weeks. Elster has summarized this in his “Rule of Threes―: MR parenchymal enhancement peaks at 3 days to 3 weeks and resolves by 3 months. The imaging time courses for CT and MR examinations in brain infarction are summarized in Table 4.2 .
Pattern
Recognition
in
236
Ischemic
Stroke
Familiarity with the major vascular territories can help distinguish between infarction and other pathologic processes. The clinical time course and localization should be consistent with the imaging findings, and all should correspond to a known vascular distribution. Stroke localization is not necessarily synonymous with “focal.― An ischemic event may cause a pattern of damage that is diffuse (hypoxic-ischemic injury); multifocal (vasculitis, emboli); or focal (single embolism or thrombus). The vessels causing stroke may be large or small, and they may be on either the arterial or venous side. There is no such thing as a “funny― stroke; if it doesn’t fit a vascular territory, then the differential diagnosis changes (Fig. 4.11 ). The relation of vascular anatomy to functional neuroanatomy is at the heart of clinical/radiologic correlation in stroke. Classically, strokes and TIAs are divided into anterior (carotid territory) or posterior (vertebrobasilar territory) events. Patients with anterior circulation ischemia have been shown to benefit from carotid endarterectomy when the carotid is narrowed by at least 70% compared to its normal diameter. Surgery has not been proven beneficial for patients with lesser degrees of carotid stenosis or for those with posterior territory TIAs, who therefore usually receive medical therapy (e.g., anticoagulation). Ischemia in the carotid territory may cause visual changes, P.95 P.96 aphasia, or sensorimotor deficits caused by retinal, cortical, or subcortical damage. Vertebrobasilar strokes are more likely to cause syncope, ataxia, cranial nerve findings, homonymous visual field deficits, and facial symptoms opposite those of the body. A given deficit can be predicted from the known functional topography of the cortex and its connections through the internal capsule (Fig. 4.12 ). Minutes No changes Absent flow void Arterial enhancement
(days
1–10)
237
DWI: high signal 2–6 hours Hyperdense artery
sign
Brain swelling (T1) Insular ribbon sign Subtle T2 hyperintensity 6–12 hours Sulcal effacement T2
hyperintensity
±Decreased 12–24
attenuation
hours
Decreased attenuation T1 hypointensity 3–7 days Maximal Maximal
swelling swelling
3–21 days Gyral enhancement (peak: 7–14 days) Gyral enhancement (peak: 3–21 days)
Petechial
methemoglobin
30–90 days Encephalomalacia Encephalomalacia Loss of enhancement Loss of enhancement Resolution of petechial blood Resolution of petechial blood DWI, diffusion-weighted imaging. Time
CT
MR
238
TABLE 4.2 Imaging Time Course After Brain Infarction
239
FIGURE 4.10. Evolution of Petechial Hemorrhage and Parenchymal Enhancement. Precontrast and postcontrast T1WIs in left sylvian cortical infarction. The acute studies ( A and B ) show nonhemorrhagic swelling (straight arrows ) with prominent cortical enhancement (B , curved arrows ). At the 2-month follow-up ( C and D ) petechial hemorrhage (open arrows ) and decreasing parenchymal enhancement (D , curved arrows ) are seen. Parenchymal enhancement resolved by 3 months.
FIGURE 4.11. Glioblastoma Mimicking a Stroke. A. T2W axial section shows edema primarily in the right middle cerebral artery territory, but with additional involvement of the medial temporal lobe, thalamus, and periatrial regions. B . Postcontrast coronal T1WI shows patchy, nodular areas of enhancement in the basal ganglia and periventricular regions (arrows ). Even with a strong clinical history for strokelike onset, the nonvascular distribution and atypical enhancement pattern effectively exclude underlying infarction. When in doubt, follow-up imaging studies will usually clarify the diagnosis.
240
The patterns of injury observed after occlusion of large arteries in the anterior and posterior circulations, small arteries in any region, and of the dural venous channels are reviewed in turn.
Anterior Internal
(Carotid) Carotid
Circulation
Artery
(ICA)
Thromboembolic disease in the ICA may cause TIAs or infarction in its MCA or anterior cerebral artery (ACA) branches or in the watershed zone between them. Embolic occlusion of the ophthalmic branch of the ICA may cause transient monocular blindness (amaurosis fugax). Observation of any of these patterns should prompt imaging of the carotid arteries. The extent and distribution of ischemia observed depend on the time course of occlusion, degree of oligemia, and available collateral supply. Complete carotid occlusions are occasionally found in asymptomatic patients with a well-developed
collateral
supply.
Atherosclerotic disease near the carotid bifurcation is responsible for the majority of ischemic events in the ICA territory. Arterial dissection, trauma, fibromuscular dysplasia, tumor encasement, prior neck radiotherapy, and connective tissue diseases may also cause significant carotid narrowing (Fig. 4.13 ). Hemodynamic effects begin to be seen when there is >80% reduction in area or >60% decrease in diameter. Lesions causing less severe narrowing may nonetheless become symptomatic when they serve as a nidus for thrombus formation or are unmasked by hypotension. Studies have shown a clear benefit of endarterectomy in symptomatic patients with >70% stenosis but not for those with Section II - Neuroradiology > Chapter 6 Central Nervous System Infections
Chapter 6 Central Nervous Infections
System
Walter L. Olsen CNS infections commonly require evaluation by radiologists. Because these infections often have dire neurologic consequences, early diagnosis and management are crucial. CT and MR have significantly aided this effort. For example, prior to CT, pyogenic abscesses of the brain carried a 30% to 70% mortality rate. The mortality rate has dropped to less than 5% in recent years, largely because of the ability of cross-sectional imaging to accurately diagnose the abscess and monitor the efficacy of treatment. MR is usually the imaging modality of choice in the evaluation of CNS infection, because of its improved sensitivity. However, because both CT and MR are highly accurate, the choice of modality often depends on the clinical situation. Gravely ill patients are often better evaluated by CT, which is faster, less susceptible to patient motion artifact, and permits closer patient monitoring. MR is generally preferable in the clinically stable patient.
PARENCHYMAL Pyogenic
INFECTIONS
Cerebritis
and
Abscess
Pyogenic infections of the brain may develop by direct extension following trauma, surgery, sinusitis, dental infections, or
394
otomastoiditis. Hematogenous infections occur even more frequently, especially in patients with lung infections, endocarditis, or congenital heart disease. Anaerobic bacteria are the most common organisms overall. Infection with Staphylococcus aureus is common after surgery or trauma. Gram-negative rod, pneumococcal, streptococcal, listerial, nocardial, and actinomycotic infections also occur with some frequency. With infections resulting from hematogenous spread, the frontal and parietal lobes (middle cerebral artery distribution) are most commonly involved, with the abscess centered at the gray–white junction. The frontal lobes are most commonly affected with spread of sinus infections. The temporal lobe or cerebellum is involved in patients with spread from otomastoiditis. Clinical symptoms in patients with pyogenic brain infections may be mild or severe. Usually there is headache. There may be varying degrees of lethargy, obtundation, nausea, vomiting, and fever. Fever is absent more than 50% of the time. Meningeal signs are present in only 30% of patients. Focal neurologic deficits, papilledema, nuchal rigidity, and seizures often develop rapidly, over the course of a few days. This is in distinction to tumors, where these symptoms usually develop more slowly. There is often, but not invariably, an elevated white blood cell count. CSF findings are often nonspecific and are usually not obtained because of the risk of lumbar puncture in the setting of a brain mass. Pathologically, there are four stages of evolution of a brain abscess, which correlate with the imaging findings.
Early
Cerebritis
Within the first few days of infection, the infected portion of brain is swollen and edematous. Areas of necrosis are filled with polymorphonuclear leukocytes, lymphocytes, and plasma cells. Organisms are present in both the center and the periphery of the lesion, which has ill-defined margins. CT scans may be normal P.157 or show an area of low density (Fig. 6.1A). There may be mild mass effect and patchy areas of enhancement within the lesion. On MR the
395
lesion shows increased signal on proton density images, fluidattenuated inversion recovery (FLAIR) images, and T2WIs, with low intensity or isointensity on T1WIs (Figs. 6.1B, C). Enhancement with gadolinium is inconstant at this stage. Use of high-dose (0.3mmoL/kg) gadolinium and/or magnetization transfer will increase the likelihood of detecting enhancement. A ring of enhancement is not present at this stage, P.158 distinguishing it from the later three stages. Unfortunately, these imaging features are nonspecific and can be seen with tumors or infarcts. The clinical features are therefore most important in making the correct diagnosis. If the diagnosis can be made at this stage, nonsurgical treatment with antibiotics is often effective.
396
FIGURE 6.1. Early Cerebritis. A. Contrast CT scan shows a subtle area of decreased density in the left frontal lobe (arrowhead). B . T2WI obtained the next day shows high signal intensity in the left frontal lobe and left frontal sinusitis. C . Gadolinium-enhanced T1WI shows low signal intensity without enhancement, consistent with early cerebritis. D . Two weeks later, a T1W, gadolinium-enhanced scan shows a ring-enhancing
397
abscess.
Late cerebritis occurs within 1 or 2 weeks of infection. Central necrosis is increased, with fewer organisms detected pathologically. There is vascular proliferation at the periphery of the lesion, with more inflammatory cells, which represents the brain's effort to contain the infection. Not surprisingly, this results in thick, irregular contrast enhancement at the edges of the lesion on imaging studies (Fig. 6.2). Centrally, there is increased signal on FLAIR images and T2WIs. Diffusion-weighted imaging (DWI) may show some increased signal within the center of the lesion. Vasogenic edema is seen outside the enhancing rim at this stage as well. Delayed scans may show central filling in with contrast. No discrete, low-signal capsule is evident on T2WIs, in distinction to some mature abscesses. This stage can also be treated effectively with antibiotic therapy, but distinguishing late cerebritis from an early abscess or tumor can be difficult, and surgery is often performed.
Early
Capsule
Within 2 weeks, the infection is walled off as a capsule of collagen and reticulin forms in the inflammatory, vascular margin of the infection. Macrophages, phagocytes, and neutrophils are also present in the capsule. The necrotic center contains very few organisms. Contrast-enhanced CT and MR scans show a well defined rim of enhancement (Fig. 6.1D). The rim tends to be low in signal on T2WIs. Centrally there is necrosis (low density on CT, low signal on T1WIs, and high signal on intermediate images, FLAIR images, and T2WIs). Prominent surrounding vasogenic edema is usually present. There is increased signal centrally on DWI, which is very helpful in distinguishing abscesses from necrotic tumors, which are not usually bright on diffusion.
398
FIGURE
6.2. Late
Cerebritis. This contrast-enhanced CT scan
demonstrates irregular enhancement peripherally and low density centrally. There is surrounding low-density vasogenic edema. This is typical of the late cerebritis stage of pyogenic infection.
Late
Capsule
In the late capsule stage, the rim of enhancement becomes even better defined and thin, reflecting more complete collagen in the abscess wall (Fig. 6.3). Multiloculation is common. The capsule often exhibits characteristic MR features that are helpful diagnostically at this stage. On T1WIs, the capsule is usually isointense or hyperintense to white to white matter (Fig. paramagnetic T1 and hematoma evolution
matter, and on T2WIs it is usually hypointense 6.4A, B) These signal characteristics suggest T2 shortening, similar to that seen in (see Chapter 4). However, hemorrhage is not
399
often found pathologically, and the paramagnetic effects may be secondary to the presence of free radicals produced by macrophages in the capsule. In any case, the MR appearance of the capsule is fairly specific for abscess. There is marked increased signal centrally on DWI, which is a very helpful imaging feature (Fig. 6.4C). The inner aspect of the enhancing capsule is often (about 50% of the time) thinner than the peripheral aspect (Figs. 6.3C, 6.4D). This reflects decreased blood supply and fibroblast migration centrally compared with cortically. This thin medial rim predisposes to intraventricular rupture, with resulting ependymitis/ventriculitis (Fig. 6.3C). CT or MR scans reveal enhancement of the ependymal lining of the ventricle and sometimes abnormal density/signal intensity of the intraventricular CSF. A solitary abscess is usually treated surgically. Often, stereotactic needle aspiration, followed by antibiotic therapy, is performed, especially if the abscess is in an eloquent area of the brain. If there is significant mass effect, or if the lesion is in a relatively “safe― area, a formal drainage or resection is performed. If there are multiple abscesses, or if the patient is at high surgical risk, antibiotic therapy alone is used. Imaging studies should be performed frequently (about once a week) to monitor the efficacy of treatment and to assess for complications such as ventriculitis, infarction,
or
hydrocephalus.
The differential diagnosis of pyogenic abscess includes tumor and resolving hematoma. The clinical features, combined with the appearance of central high signal on DWI, thin enhancing rim (thinnest medially), much edema, and paramagnetic effects in the capsule (with no blood products centrally), should strongly suggest a brain abscess.
400
FIGURE
6.3. Multiple Pyogenic Abscesses. A. T2WI reveals a
right parietal mass lesion with high signal intensity centrally and low signal intensity peripherally within the capsule. There is surrounding high-signal-intensity edema. Two smaller high signal lesions are present on the left. B . Gadolinium-enhanced T1WI shows thin, smooth enhancement of all three lesions. C . More inferiorly, the contrast-enhanced T1WI reveals a fourth abscess
401
that has extended into the atrium of the left lateral ventricle (arrowheads). The enhancement pattern and intraventricular extension favor the diagnosis of abscess over tumor. These lesions proved to be abscesses that cultured anaerobic streptococci. (Case courtesy of Dr. Vincent Burke, Atherton, California.)
P.159
Septic
Embolus
Infections that begin with a septic embolus may not have the typical appearance of an abscess. The embolus frequently causes an infarct that dominates the imaging findings. Depending on the size of the embolus, there may be a small rounded area of enhancement or a larger, wedge-shaped cortical infarct. As with other embolic infarcts, hemorrhage may occur. Because the nonviable, infarcted tissue has a poor blood supply, a typical capsule may not form. A thicker, more irregular ring of enhancement that persists within an area of infarction should suggest the diagnosis. Septic emboli may lead to mycotic aneurysm formation, which can result in intraparenchymal or subarachnoid
hemorrhage.
Listeria monocytogenes is an anaerobic gram-positive bacillus that primarily infects immunocompromised patients. The organism can cause meningitis, meningoencephalitis, and abscesses. Listerial rhombencephalitis involves the brainstem and cerebellum. MR scans reveal areas of abnormal signal and enhancement in the brainstem and cerebellar white matter tracts. The appearance is similar to that of acute disseminated encephalomyelitis (see “Viral Infections―). Occasionally, otherwise healthy patients may develop this infection.
402
FIGURE
6.4. Pyogenic
Cerebral
Abscess. This case illustrates
most of the classic features of a cerebral abscess. A . T1W sagittal scan shows high signal in the rim of the abscess as a result of paramagnetic T1 shortening. B . T2WI shows low signal in the rim from T2 shortening with high signal centrally and significant surrounding edema. C . Diffusion-weighted scan shows increased signal centrally, a characteristic feature of abscesses
403
that is usually not seen with necrotic tumors. D . Postcontrast T1WI shows enhancement of the rim that is thinnest medially, as is often the case with abscesses.
P.160
Mycobacterial
Infections
The most common form of CNS mycobacterial infection is tuberculous meningitis, which will be discussed later in this chapter. Focal mycobacterial infection of the brain occurs in two forms: tuberculoma and abscess. A tuberculoma is a granuloma with central caseous necrosis. A tuberculous abscess has characteristics similar to those of a pyogenic abscess, but it develops in patients with impaired T-cell immunity.
Tuberculoma In the early 1900s, one third of all brain mass lesions in England were tuberculomas. Because of P.161 improved prevention and treatment, these lesions are now unusual in industrialized countries. In developing areas of the world, however, tuberculomas account for 15% to 30% of brain masses. Approximately 5% to 10% of patients with tuberculosis develop CNS disease. There is a predilection for the extremes of age: children and the elderly. The infection spreads to the brain hematogenously from the lungs. In developed countries, tuberculomas usually result from reactivation of quiescent disease, although only 50% of patients have a known history of previous tuberculosis. Most lesions in adults are supratentorial, involving the frontal or parietal lobes. Sixty percent of tuberculomas in children are in the posterior fossa, usually the cerebellum. Multiple lesions are common. Most tuberculomas are not associated with tuberculous meningitis. Clinical features include headache, seizures, papilledema, and focal neurologic signs. Fever is seen only rarely. The CSF is almost always abnormal, showing elevated protein and reduced glucose. An abnormal chest radiograph
404
is present in up to 50% of patients. These lesions can be treated medically if there are characteristic clinical and imaging features. Surgery is often performed when the diagnosis is in doubt, for medical treatment failures, and for large lesions. Noncontrast CT scans show one or more isodense or slightly hyperdense nodules or small mass lesions. Multiple lesions are present about 50% of the time. The center of the tuberculoma is usually denser than the fluidlike center of a pyogenic abscess because of caseous necrosis. A “target― appearance, with a central calcification surrounded by rim enhancement, is an uncommon but helpful finding, strongly suggesting the diagnosis. Calcification is present in fewer than 5% of cases at the initial diagnosis but is commonly seen with treatment as the lesions resolve. With MR, tuberculomas may be high or low in signal intensity with T2WIs, depending upon the size of the lesion and the water content of the caseous necrosis. (Fig. 6.5A) The wall of the tuberculoma is often low in signal on T2WIs. There is significant enhancement after gadolinium administration, with a solid nodular or thick ring-shaped appearance (Fig. 6.5B). There may or may not be increased signal on DWI, unlike bacterial infections, which usually show restricted diffusion (Fig. 6.5C). Surrounding edema is often mild. The differential diagnosis includes tumor, pyogenic abscess, fungal and parasitic infections, and sarcoidosis. Tuberculous
abscess is a rare complication seen primarily in
immunocompromised
patients.
Abnormal
T-cell P.162
function prevents the normal host response of tuberculoma formation with caseous necrosis. Symptoms develop more rapidly than with tuberculomas. The imaging features are similar to that seen with pyogenic abscesses. The lesions are often large and multiloculated, in distinction to tuberculomas. Prominent edema and mass effect also distinguish tuberculous abscess from tuberculoma. mycobacterial infections are also more common in immunocompromised patients.
405
Atypical
FIGURE 6.5. Multiple Tuberculomas. A. Fluid-attenuated inversion recovery (FLAIR) axial scan shows multiple small areas of increased signal without significant edema bilaterally. B . Postcontrast T1W axial scan shows multiple small enhancing nodules. C . There is no abnormal restricted diffusion on this DWI axial scan. The appearance of tuberculomas on diffusionweighted imaging (DWI) is variable, unlike pyogenic infections, which almost always show restricted diffusion (bright on DWI).
Fungal
Infections
Fungal infections of the CNS can be grouped into endemic and cosmopolitan categories. Endemic fungal infections are geographically restricted. They can occur in both immunocompetent and immunosuppressed patients. Cosmopolitan fungal infections occur worldwide, usually in immunosuppressed patients, infants, the elderly or chronically ill, with the exception of cryptococcosis, which
406
also occurs in patients with normal immunity.
Endemic
Fungal
Infections
The most common endemic fungal infections in the United States are coccidioidomycosis, North American blastomycosis, and histoplasmosis. These infections usually manifest as granulomatous meningitis, as will be discussed. Focal parenchymal lesions are unusual. CNS involvement is a manifestation of disseminated infection, with hematogenous spread, usually from pulmonary disease. Coccidioidomycosis occurs in the southwestern United States. The spores are inhaled, with outbreaks occurring after groundbreaking for construction projects. Most infected patients are asymptomatic or have mild respiratory symptoms. Less than 1% of patients develop disseminated
infection
and
meningitis.
Focal
parenchymal
granulomas
are rare. Blastomycosis occurs in the Ohio and Mississippi River valleys. CNS involvement occurs in 6% to 33% of disseminated cases. Meningitis is the most frequent presentation, but parenchymal abscesses and granulomas occur more frequently than with coccidioidomycosis. Epidural granulomas and abscesses also occur in the head and spine, usually from direct extension from bone infection. Up to 40% of focal brain lesions are multiple. Histoplasmosis is usually a benign, asymptomatic infection, occurring in the Midwest and southern United States. Dissemination is unusual, and only a small percent of disseminated cases involve the CNS. Meningitis is most common, but multiple or solitary granulomas may occur. Abscesses are unusual. As seen with CT or MR, most fungal granulomas are small and show solid or thick rim enhancement (Fig. 6.6). Fungal abscesses (as sometimes seen with blastomycosis) have an appearance similar to that of the pyogenic abscesses that were described earlier. Meningeal enhancement from meningitis is a common accompanying feature. Hydrocephalus is also common, especially with coccidioidomycosis.
407
FIGURE
6.6. Histoplasmosis
Granuloma. This patient had
disseminated histoplasmosis with several lesions in the brain and spine. This contrast-enhanced CT scan shows a solidly enhancing lesion near the atrium of the right lateral ventricle (arrowhead) . Most fungal granulomas are small and show either solid or thick rim enhancement. (Case courtesy of Dr. J. R. Jinkins, San Antonio, Texas.)
Cosmopolitan
Fungal
Infections
The most common cosmopolitan fungal infections are cryptococcosis, aspergillosis, mucormycosis, and candidiasis. These infections also usually present as meningitis, but focal parenchymal lesions are fairly common.
408
Aspergillosis involves the CNS in 60% to 70% of patients with disseminated disease. The infection may arise from hematogenous spread or by direct extension from an infected paranasal sinus, leading to meningitis or meningoencephalitis. Parenchymal disease usually takes the form of an abscess. Granulomas are unusual. The abscesses are often multiple and show irregular ring enhancement (Fig. 6.7). Subcortical or cortical infarcts and hemorrhage from blood vessel invasion may occur. The mortality rate with invasive intracerebral aspergillosis is greater than 85%.
Mucormycosis Mucor invades the brain usually by direct extension from the sinuses, nose, or oral cavity, but hematogenous spread also occurs. Almost all patients are diabetic or otherwise immunocompromised. The mortality rate in treated diabetic patients is 65% to 75% and is P.163 worse in immunocompromised patients. Like aspergillosis, mucormycosis tends to invade blood vessels. Imaging studies in patients with CNS mucormycosis will reveal single or multiple mass lesions with varying degrees of peripheral enhancement. The amount of enhancement depends upon the compromised host's ability to fight the infection. Surrounding edema is variable in amount. Smaller lesions will show a solid enhancement pattern. The lesions are often in the base of the brain, adjacent to diseased sinuses. Infarcts, intraaxial or extra-axial hemorrhage, and meningeal enhancement can be seen with CT or MR. A lesion with peripheral enhancement, cortical sparing, and a nonvascular distribution is more likely to be a mucormycotic abscess than an infarct, but often it is difficult to distinguish the two.
409
FIGURE CT (A),
6.7. Disseminated Aspergillosis. intermediate-weighted MR (B), and
Contrast-enhanced gadolinium-enhanced
T 1 W (C) scans show a large necrotic mass in the right frontal lobe and several smaller lesions in the left hemisphere. The right frontal lobe lesion was surgically drained and aspergillosis was found. The patient was a poorly controlled diabetic.
410
Candidiasis usually causes meningitis, but granulomas and small abscesses may occur. Spread to the CNS is usually hematogenous from the lungs or GI. In cases of CNS candidiasis, meningeal enhancement or multiple small enhancing granulomas or microabscesses are usually seen. Infarcts, hydrocephalus, and large abscesses may also be identified. Cryptococcosis is the most frequently reported CNS fungal infection. It occurs in patients with normal immune P.164 function in about 50% of cases. This is also an extremely common infection in patients with AIDS, as will be discussed later in this chapter. Infection of the CNS occurs via hematogenous spread from the lungs. Serologies and CSF studies are valuable in making the diagnosis: about 90% of patients have antigen in the CSF and/or antibody in the serum. The usual manifestation is meningitis. Granulomas can occur and are usually multiple. Abscesses are less common. CT scans in patients with cryptococcosis are usually normal, reflecting relatively mild meningeal involvement in most cases. Mass lesions are seen in about 10% of cases. Cryptococcomas are shown as small, usually multiple, solid-enhancing, peripheral parenchymal nodules. Calcifications within a granuloma are occasionally seen. With the improved sensitivity of MR, meningeal and parenchymal lesions are seen more frequently than with CT. Leptomeningeal nodules are often only seen on T1W, contrastenhanced MR as multiple tiny enhancing lesions near the basal cisternae and sulci. Diffuse meningeal enhancement is unusual. Granulomas may show either solid or ring enhancement. Another characteristic cryptococcal lesion is the gelatinous pseudocyst. This is a cystic lesion, usually in the basal ganglia, representing enlarged Virchow-Robin spaces filled with the organism. These lesions are usually found only in immunocompromised patients (see Fig. 6.33). Viewed on CT, gelatinous pseudocysts are smooth, round, low-density masses in the basal ganglia which show no contrast enhancement. They are usually better seen with MR than with CT, as lesions are nearly isointense with CSF on all sequences
411
that do not enhance.
Parasitic
Infections
Parasitic infections are common throughout much of the developing world but are relatively uncommon in the industrialized nations. The most common infections likely to be encountered in the United States are cysticercosis, echinococcosis, toxoplasmosis, and rarely amebiasis. CNS involvement in malaria, trypanosomiasis, paragonimiasis, sparganosis, and schistosomiasis is rarely encountered in the United States and will not be discussed. Cysticercosis is caused by the larvae of the pork tapeworm Taenia solium. Infestation occurs via the fecal-oral route. When larvae are ingested, intestinal disease results, and eggs are released into the bowel stream. The life cycle can be completed if the eggs are ingested by the pig but not by humans. In this situation, the eggs form oncospheres (primary larvae), which hatch in the intestine and are hematogenously distributed throughout the body and form cysticerci (secondary larvae). The cysticerci cannot develop further in humans and they eventually die. Cysticerci that reach the CNS may infest the parenchyma, meninges, ventricles, or spine. This disease is fairly frequently encountered in the southwestern United States in Latin American immigrants. Seizures occur in more than 90% of patients. Cysticercosis is the most common cause of seizures in Latin America. Encephalitic symptoms are also common. Treatment is with anticysticercus drugs such as praziquantel and albendazole. Parenchymal cysticercosis is the most common type. Early in the infestation, during the tissue invasion stage, CT or MR scans show edema and/or nodular enhancement. Later, the viable cysts appear as small (usually 1 cm or less), solitary or multiple rounded lesions that are low density on CT and isointense to CSF on MR (Fig. 6.8) . The lesions are usually peripherally distributed near the gray–white junction or in the gray matter. A small marginal nodule representing the scolex is sometimes seen (Figs. 6.8B and 6.9). There is usually no enhancement or edema at this vesicular stage. When the cyst dies, the fluid within it leaks into the surrounding brain, causing
412
inflammation. This produces clinical symptoms of an acute encephalitis, which may be severe, depending on the number of lesions. Imaging studies now reveal ring-enhancing lesions with surrounding edema (Fig. 6.9). The cyst fluid is of increased density on CT and increased signal compared with CSF on T1W and T2W MR. As the dead cyst degenerates, it becomes smaller, showing nodular enhancement, and then calcifies. CT scans at this late stage show small, peripheral calcifications, with no edema or enhancement (Fig. 6.10). With MR, the calcifications are best seen on T2W or T2*W GRE images, but they are better demonstrated by CT. Imaging with CT or MR is useful in staging and monitoring treatment. Once the cyst has degenerated, further drug therapy is not warranted. Intraventricular cysticercosis is similar to the parenchymal variety in pathogenesis and appearance (Fig. 6.11). The cysts are usually isodense and isointense to CSF, making them difficult to visualize. MR is superior to CT for imaging, as subtle signal changes and lack of CSF pulsations within the cyst makes them more visible. Enhancement may or may not be present, depending on the stage of disease, similar to the parenchymal form. The cysts may obstruct the foramen of Monro, the third ventricle, or the cerebral aqueduct, resulting in hydrocephalus. If acute hydrocephalus occurs, death may rapidly ensue. Ventriculitis occurs if the cyst ruptures. Meningeal infestation is known as meningobasal (because the basal cisterns are most frequently involved) or racemose (Latin for “clusters―) cysticercosis. The cysts lack a scolex but may grow by proliferation of the cyst wall. The cysts may grow in grapelike clusters (Figs. 6.12, 6.13) or conform to the shape of the cistern. No mural nodules or calcifications are seen. CT scans show CSF-density cysts in the basal cisterns. MR reveals cysts that are isointense with CSF, often with mural enhancement or diffuse meningeal enhancement. Hydrocephalus is commonly observed. Spinal cysticercosis is usually intradural, but can be either intramedullary
or
extramedullary.
Intramedullary P.165
lesions are best seen with MR as solid or ring-enhancing cord lesions,
413
similar to that seen in the brain parenchyma. Extramedullary cysts are analogous to the racemose form and, like most spinal pathology, are also best evaluated with MR.
FIGURE
6.8. Cysticercosis.
A. T2WI shows a right frontal lesion
isointense with CSF (arrowhead). There is no surrounding edema, indicating that this is early in the course of the disease. Three smaller lesions are present posteriorly. B . The T1W parasagittal image in the same patient shows two cysticercal cysts that are isointense with CSF. A scolex is visible in one of the cysts (arrowhead) .
Echinococcosis, also known as hydatid disease, occurs in South America, Africa, Central Europe, the Middle East, and rarely in the southwestern United States. The etiologic agent is the dog tapeworm, and humans are intermediate hosts. Hydatid cysts are most frequently present in the lung and liver, but the brain is involved in 1% to 4% of cases. The cysts are usually solitary, unilocular, large, round, and smoothly marginated. They are most often supratentorial, in the middle cerebral artery territory. There may rarely be mural calcification. With CT, the fluid within the cyst is usually isodense
414
with CSF. There is usually no surrounding edema or abnormal contrast enhancement, unless the cyst has ruptured, leading to an inflammatory reaction. With MR, the lesions are usually nearly isointense with CSF. Toxoplasmosis is caused by the protozoa Toxoplasma gondii, which occurs worldwide. The disease may be either congenital or acquired. The acquired form is seen primarily in immunocompromised patients and is very common in AIDS patients, as will be discussed later. The congenital form results when a pregnant woman eats poorly cooked meat or is infected by a cat. A diffuse encephalitis of the fetal brain ensues, usually causing severe destruction. The infant is usually born with microcephaly (Fig. 6.14), chorioretinitis, and mental retardation. Imaging studies reveal atrophy, dilated ventricles, and calcifications. The calcifications occur in the periventricular white matter, basal ganglia and cerebral hemispheres. This is in distinction to congenital cytomegalovirus (CMV), in which the calcifications are usually periventricular Amebic
only.
meningoencephalitis is sometimes seen in the southern
United States. The amebae enter the nasal cavity of patients swimming in infested freshwater ponds or pools. There is direct extension through the cribriform plate to the brain. Severe meningoencephalitis results and is usually fatal. Imaging studies often underestimate the severity of the disease. Early in the infection, there may be meningeal and/or gray matter enhancement. Later, there is diffuse cerebral edema. There are a few reports of single or multiple, ring-enhancing or solid-enhancing lesions with surrounding edema in patients with amebic brain abscesses. Amebic abscesses are more common in immunosuppressed patients.
Spirochete
Infections
Neurosyphilis develops in about 5% of patients who are not treated for the primary infection. Involvement of the CNS usually occurs in the secondary or tertiary stages. Because of effective antibiotic therapy, the disease is rare. However, there has been a significant increase in incidence since the AIDS epidemic. Neurosyphilis is more
415
likely to develop in HIV-infected patients, and the neurologic symptoms occur after a shorter latency period than P.166 in uninfected patients. Patients with neurosyphilis are usually asymptomatic. Symptomatic patients may have an aseptic meningitis, tabes dorsalis, general paresis, or meningovascular disease. Imaging studies are usually normal in patients with tabes dorsalis, but rarely gummas are found. These usually appear as small enhancing nodules at the surface of the brain, with adjacent meningeal enhancement. Meningovascular syphilis presents as an acute stroke syndrome or a subacute illness with a variety of symptoms. Pathologically, there is thickening of the meninges and a medium to large vessel arteritis. Imaging studies reveal small infarcts of the basal ganglia, white matter, cerebral cortex, or cerebellum (Fig. 6.15A). The infarcts may exhibit patchy or gyriform enhancement, which is best seen with MR. Meningeal enhancement is unusual, but cranial nerve enhancement in patients with syphilitic cranial neuritis has been described. Angiography in patients with meningovascular neurosyphilis reveals multiple segmental constrictions and/or occlusions of large and medium arteries, including the distal internal carotid, anterior cerebral, middle cerebral, posterior cerebral, and distal basilar arteries (Fig. 6.15B) .
416
FIGURE 6.9. Cysticercosis. The contrast-enhanced CT scan (A), T2W scan (B), and the contrast-enhanced T1W scan (C) all show a cystic lesion in the left frontal lobe. The rim of the cyst enhances with contrast and there is surrounding edema (large arrowheads), indicating that the cyst has died and that fluid has leaked out, inciting an inflammatory response. The scolex is visible (small arrowheads) .
Lyme
disease is a multisystem spirochete infection caused by
417
Borrelia burgdorferi. It is found worldwide in deer, mice, raccoons, and birds. It is spread to humans via ticks, especially the deer tick. The disease occurs most frequently on the East Coast, but may occur anywhere in the United States. The disease begins as a flulike illness, P.167 with a rash and an expanding skin lesion at the tick bite site. In a small percentage of patients, cardiac, arthritic, or neurologic symptoms develop. Neurologic abnormalities are found in 10% to 15% of patients. A variety of symptoms, including peripheral and cranial neuropathies, radiculopathies, myelopathies, encephalitis, meningitis, pain syndromes, cognitive disorders, and movement disorders, have been reported. Treatment with antibiotics and corticosteroids may have variable results. MR is the modality of choice for imaging these patients. In patients with cranial neuritis, MR scans may show thick, enhancing cranial nerves. Cranial nerves III to VIII can be involved, with the facial nerve most commonly affected. In patients with parenchymal CNS Lyme disease, MR scans show multiple small white matter lesions, similar to that seen with multiple sclerosis. The lesions can be found in the supratentorial and infratentorial white matter tracts. The lesions often enhance with contrast in a nodular or ring pattern, depending on the size. There may be meningeal enhancement. The differential diagnosis includes multiple sclerosis and other demyelinating processes.
418
FIGURE
6.10. Late-Stage
Cysticercosis. This unenhanced CT
scan shows multiple calcifications in the gray matter and gray–white junction, which are typical of old cysticercosis.
Viral
Infections
The most common viral infections of the CNS include CMV, herpes simplex, varicella zoster, and HIV. Rubella was once a devastating fetal viral infection P.168 but is now uncommon because of widespread immunization.
419
FIGURE 6.11. Intraventricular Cysticercosis. Intermediateweighted axial (A) and gadolinium-enhanced T1W coronal (B) scans show a cystic mass in the frontal horn of the right lateral ventricle (large arrowheads). The lesion is of slightly increased signal intensity compared with CSF in the ventricle. The scolex is of high signal intensity in the posterior aspect of the cyst in (A). There is also a small parenchymal lesion in the left basal ganglia (small arrowhead) .
420
FIGURE
6.12. Subarachnoid
(Racemose)
Cysticercosis.
There are multiple cysts in a grapelike cluster in the basal cisterns on this parasagittal T1WI (arrowheads). These cysts lack a scolex but grow by proliferation of the cyst wall.
Cytomegalovirus is a DNA virus in the herpesvirus family. It causes symptomatic CNS disease primarily through congenital transmission. Maternal CMV infection results in transplacental transmission to the fetus in 30% to 50% of cases and symptomatic disease in 5%. Symptomatic neonates may have hepatosplenomegaly, jaundice, cerebral involvement (psychomotor retardation), chorioretinitis, and deafness. Mental retardation and deafness are present in 20% of cases. The intracranial manifestations of congenital CMV infection largely depend upon the time of infection during gestation. Infection in the first trimester results in necrosis in the germinal matrix, resulting in migrational anomalies. MR scans may show a spectrum of abnormalities, including agyria, polymicrogyria, and focal P.169 cortical dysplasia. Delayed myelination and cerebellar hypoplasia are also common findings. Patients infected later during gestation may have a normal gyral pattern, but delayed myelination and periventricular white matter lesions are often seen. The most
421
common finding with CT scanning is periventricular calcification (Fig. 6.16). These are detected better with CT than MR. There are usually no calcifications of the basal ganglia or cortices, as is seen in congenital toxoplasmosis. Parenchymal atrophy and ventriculomegaly are common. The disease has been diagnosed in utero with obstetric US. Periventricular hyperechoic calcifications, preceded by hypoechoic periventricular ringlike zones, are characteristic findings. CMV infection in adults is unusual, except in immunosuppressed patients. This will be discussed in the AIDS section.
FIGURE
6.13. Subarachnoid
(Racemose)
Cysticercosis. T1W
axial (A) and gadolinium-enhanced T1W sagittal (B) scans show multiple nonenhancing cysts (arrowheads) in the left sylvian fissure, callosal sulcus, and cingulate sulcus. The corpus callosum is markedly distorted by the cysts.
422
FIGURE
6.14. Microcephaly. A lateral skull film shows a small
cranial vault, as can be seen in congenital toxoplasmosis or other TORCH (Toxoplasmosis; Other, which includes syphilis; Rubella; Cytomegalovirus; and Herpes) infections.
Herpes
simplex encephalitis occurs most frequently in neonates.
The infant is usually infected during descent through the birth canal when the mother has genital (type 2) herpes. Occasionally there is transplacental transmission before delivery, but this usually results in spontaneous abortion. The infection causes a severe encephalitis, which is either fatal or has severe neurologic consequences. The patient usually presents with seizures in the second to fourth week of life. If the patient survives, varying degrees of microcephaly, mental P.170 retardation,
microphthalmia,
enlarged
423
ventricles,
intracranial
calcifications, and multicystic encephalomalacia may occur. Early in the course of the encephalitis, CT scans may reveal diffuse brain swelling or bilateral patchy areas of decreased density in the cerebral white matter and cortex, with relative sparing of the basal ganglia, thalami, and posterior fossa structures (Fig. 6.17A). Cranial sonography will show areas of increased echogenicity corresponding to these low-density zones. The low-density lesions progress to areas of necrosis, which are sometimes hemorrhagic and may eventually calcify. Multicystic encephalomalacia is the end result. Increased density in the cortical gray matter is characteristic during this late stage (Fig. 6.17B). With MR there is decreased gray–white matter contrast early in the infection, reflecting gray matter edema. Later, there is decreased signal on T2WIs within the thinned cortical gray matter.
FIGURE 6.15. Meningovascular Syphilis. A. Contrastenhanced CT scan reveals a small infarct in the left striate nucleus in this 21-year-old man with meningovascular syphilis. B . A left internal carotid arteriogram in the frontal projection on
424
another patient with meningovascular syphilis shows occlusion of the left anterior cerebral artery (small arrowhead) and narrowing of branches of the left middle cerebral artery (large arrowhead) . Both patients improved with penicillin therapy.
FIGURE 6.16. Congenital Cytomegalovirus Infection. A noncontrast CT scan shows multiple periventricular high density calcifications. The calcifications in congenital cytomegalovirus infection tend to be periventricular only, as in this case. With congenital toxoplasmosis, calcifications may be found throughout the brain.
In adults, herpes simplex infection may cause encephalitis or cranial
425
neuritis. The infection usually is secondary to reactivation of latent herpes simplex type 1. Patients with herpes encephalitis present with the gradual onset of personality changes, dysphasia, and focal neurologic deficits. Seizures and coma may occur. An inconstant but characteristic electroencephalographic finding is a localized spiked and slow wave pattern. Early diagnosis is crucial, because there is a greater than 70% mortality rate in untreated patients. Unfortunately, CSF studies are often negative. Treatment is with acyclovir, which significantly reduces mortality, but many survivors have permanent deficits. CT scans show a poorly defined area of decreased density in one or both temporal lobes (Figs. 6.18, 6.19A). The predilection for the temporal lobes is because the virus is usually latent within the gasserian ganglion. The frontal lobes may also be involved. The insular cortex is often involved, but the adjacent putamen is usually spared. There is usually swelling with mass effect. Streaky enhancement is variable. The CT findings are not usually seen before the fifth day of symptoms. With MR, the findings may be identified somewhat sooner. There is a nonspecific pattern of increased signal on the intermediate, FLAIR, and T2W images in the temporal and/or frontal lobe(s) with sparing of the putamen. This increased signal is best seen on a FLAIR sequence (Fig. 6.19B, C). Increased signal on DWI has been reported. Early on, meningeal enhancement may be seen. Later, there may be parenchymal enhancement or evidence of hemorrhage. The differential diagnosis includes middle cerebral artery infarct (which will often involve the putamen, unlike herpes), early bacterial cerebritis, and other types of viral encephalitis. Varicella zoster rarely causes an encephalitis that is similar to that caused by herpes simplex. Another unusual manifestation is the syndrome of herpes zoster ophthalmicus and delayed contralateral hemiparesis caused by cerebral angiitis. In this syndrome, there are cerebral infarcts resulting from large and medium vessel angiitis on the same side as ophthalmic zoster skin manifestations. Imaging studies show typical infarcts, and angiography shows segmental areas of narrowing and/or beading of the arteries. Herpes zoster may also cause cranial neuritis, which may involve any of the cranial nerves. The most commonly involved is the facial nerve, resulting in
426
the Ramsay Hunt syndrome. Clinically, there is ear pain and facial paralysis, accompanied by a vesicular eruption about the ear. CT scans are usually normal, but MR may reveal increased contrast enhancement of the facial nerve. Acute disseminated encephalomyelitis (ADEM) is an acute demyelinating disease that occurs after a viral infection, following a vaccination, or sometimes spontaneously. It probably has an autoimmune basis; organisms are not isolated from brain tissue. Symptoms develop acutely: fever, headache, and meningeal signs. Seizures, focal neurologic deficits, stupor, and coma may develop. The mortality rate is 10% to 20%. When treated early with steroids, most patients make a full recovery. T2WIs and FLAIR imaging are much more sensitive than CT in identifying lesions of increased signal intensity in the white matter (Fig. 6.20). The brainstem, cerebellum, and basal ganglia are often involved. Optic neuritis is also common. The P.171 lesions are usually multiple, but few in number. Sometimes there is solid or ring enhancement of the lesions. The appearance is similar to that seen in multiple sclerosis, but with a monophasic clinical course. The lesions regress with successful treatment, correlating with clinical improvement. Acute hemorrhagic leukoencephalitis is a severe variant of ADEM that is often fatal. Pathologically, there is perivascular hemorrhagic necrosis, primarily in the centrum semiovale. The major imaging feature is a rapid progression of white matter lesions over the course of a few days.
427
FIGURE
6.17. Neonatal Herpes. A. There is low density and
swelling in the right temporal lobe, and to a lesser extent in the frontal and left temporal lobes, on the noncontrast CT scan of this 2-week-old child with herpes type 2 infection in the acute stage. B . Three weeks later, the noncontrast CT scan on this same infant reveals multiple areas of cystic encephalomalacia and widespread gray matter calcification, which are typical of late-stage
neonatal
herpes
infection.
Subacute sclerosing panencephalitis is caused by a variant of the measles virus. It typically occurs in children and young adults who had measles before age 2, after a 6- to 10-year asymptomatic period. The disease causes progressive dementia, seizures, and paralysis, leading to death. There is no treatment. Imaging studies initially reveal focal lesions in the gray matter and subcortical white matter. Later, there are periventricular white matter lesions that may enhance. In the late stages there is usually profound cortical atrophy.
428
Progressive multifocal leukoencephalopathy (PML) is a demyelinating disease caused by a papova virus (the JC virus—not to be confused with the Creutzfeldt-Jakob agent, which is a prion). It occurs only in immunosuppressed patients and has an increased incidence in patients with AIDS, as will be discussed. Encephalitis can be caused by a variety of viruses not already discussed, including flaviviruses (arboviruses), Epstein-Barr virus, mumps, measles, rubella, and enteroviruses. Rickettsia (Rocky Mountain spotted fever) and mycoplasma pneumoniae may also produce encephalitis. In the United States, St. Louis, California, western equine, and eastern equine encephalitis are caused by flavivirus infection. Japanese encephalitis is a flaviviral encephalitis found in Asia. Bilateral thalamic and basal ganglia lesions are typical of this disease. A related virus causes West Nile fever, which has an increasing incidence in the United States. This is a mosquito-borne infection that results in a meningoencephalitis of variable clinical severity. As with Japanese encephalitis, MR shows increased signal on T2WIs in the thalami and basal ganglia (Fig. 6.21). Rasmussen encephalitis is a devastating disease of childhood, most likely of viral etiology. There are intractable seizures and progressive neurologic deficits. The disease usually affects one cerebral hemisphere. Imaging studies show severe atrophy of the involved hemisphere. PET scans show the hemisphere to be hypometabolic. Creutzfeldt-Jakob
disease of the sporadic type (sCJD) is caused by
a transmissible protein called a prion or P.172 slow virus. Clinically, there is rapidly progressive dementia, ataxia, and myoclonus, leading to death. Conventional MR scans are often normal early in the disease. DWI frequently demonstrates increased signal in the cerebral cortex and basal ganglia in early cases. Atrophy and increased signal in the cortex and basal ganglia on FLAIR and T2W images develop as the disease progresses.
429
FIGURE
6.18. Adult
Herpes
Encephalitis. Both temporal lobes
are of low density and appear swollen, especially on the right, on this contrast-enhanced CT scan. The appearance is similar to cerebral infarcts, but the clinical presentation is usually different.
Variant Creutzfeldt-Jakob disease (vCJD) is linked to bovine spongiform encephalopathy in cows. Most cases have occurred in the United Kingdom. This prion infection is transmitted to humans who eat infected beef. The clinical features are similar to sCJD. MR scans usually show characteristic findings of increased T2 signal in the pulvinar (posterior) nuclei of the thalamus.
430
EXTRA-AXIAL
INFECTIONS
Meningitis Meningitis can be caused by bacteria, mycobacteria, fungi, parasites, and viruses. Bacterial meningitis is caused by Haemophilus influenzae (in children), Neisseria meningitidis (in teens and young adults), and Streptococcus pneumoniae (in older adults) in over 80% of cases. Escherichia coli, group B streptococcus, and Listeria monocytogenesoccur commonly in neonates. The bacteria most commonly enter the meninges during a systemic bacteremia but can spread directly from infected sinuses or after surgery or trauma. Patients present with a relatively acute onset of fever, a stiff neck, and headache, followed by a decline in mental status. CSF studies are usually diagnostic, and imaging studies are generally not required. CT scans may be performed in the emergency setting in acutely comatose patients or in patients with a nonspecific headache but are usually normal (Fig 6.22A). The inflammatory exudate caused by the meningitis may occasionally produce high density within the subarachnoid spaces, similar to that seen in subarachnoid hemorrhage. The increased density is often more pronounced in the peripheral sulci than in the basal cisterns, unlike most cases of aneurysmal bleeding. Diffuse cerebral edema is sometimes seen (Fig. 6.22B). If contrast is given, there may or may not be meningeal enhancement. CT and MR are more often used later in the course of meningitis, when there are suspected complications such as hydrocephalus, cerebritis/abscess, ventriculitis, and venous or arterial infarctions. The hydrocephalus that may develop is usually of the communicating type, reflecting decreased function of the arachnoid villi in absorbing CSF. Subdural effusions may be seen in infants, especially with H influenza meningitis. With CT and MR, subdural effusions appear as thin collections along the surface of the brain that are isodense/isointense with CSF (Fig. 6.23). These sterile effusions can be identified with cranial sonography in infants. Echogenic sulci, ventriculomegaly, and abnormal parenchymal echogenicity have also
431
been reported in infants with bacterial meningitis, imaged with US (Fig. 6.22C) . Tuberculous
meningitis is the most common form of CNS
tuberculosis. It is usually caused by Mycobacterium tuberculosis, but atypical mycobacteria, such as M avium-intracellulare can cause meningitis in AIDS patients. Tuberculous meningitis has a predilection for infants and children but is seen in all age groups. The disease spreads to the meninges hematogenously from the lungs, but the chest radiograph is normal in 40% to 75% of patients. The tuberculin skin test is also frequently negative. Clinically there is usually a subacute or insidious onset of headache, malaise, weakness, apathy, or focal neurologic findings. Imaging studies will show enhancing, thickened meninges, especially near the base of the brain (Fig. 6.24), unlike bacterial meningitis, where the peripheral meninges are more often involved. The often marked thickening of the meninges also distinguishes tuberculous and other granulomatous meningitides from pyogenic meningitis. The thick exudate in the basal cisterns may extend into the Virchow-Robin spaces, causing a vasculitis. This frequently leads to infarcts, which are better detected with MR than with CT. P.173 Communicating
hydrocephalus
is
another
complication.
432
relatively
common
FIGURE 6.19. Type 1 Herpes Encephalitis. The contrastenhanced CT scan (A) on this 8-year-old boy with decreased level of consciousness reveals subtle low density in the right temporal lobe (arrowheads). T2W fast–fluid-attenuated
433
inversion recovery scans (B, C) performed on the same day show prominent areas of high signal intensity in both temporal lobes with sparing of the putamina. This case illustrates why MR is the imaging modality of choice when herpes encephalitis is suspected.
The differential diagnosis of tuberculous meningitis includes fungal meningitis, racemose cysticercosis, sarcoidosis, and carcinomatous meningitis. Fungal meningitis usually causes thick meningeal enhancement in the basal cisterns, as with tuberculosis (Fig. 6.25). Enhancement is variable with cryptococcosis, depending on the immune status of the patient. Hydrocephalus is common, but infarcts and extension of fungal meningitis into the brain occur less frequently than with tuberculous or pyogenic meningitis (except in cases of aspergillosis and
mucormycosis).
Racemose cysticercosis may show thick meningeal enhancement, but cystic lesions in the cisterns are also frequently found (see Figs 6.12, 6.13) .
434
FIGURE 6.20. Acute Disseminated Encephalomyelitis (ADEM). The T2W fast–fluid-attenuated inversion recovery (FLAIR) scans (A, B) show multiple areas of high signal intensity in the cerebral white matter and midbrain. FLAIR sequences are extremely sensitive for detecting white matter lesions. This 8year-old child recovered fully after steroid therapy.
P.174 Sarcoidosis involves the CNS in up to 14% of patients at autopsy, but causes neurologic symptoms only rarely. It primarily affects the leptomeninges, so that abnormal meningeal enhancement is seen with CT or MR. Focal parenchymal enhancing mass lesions or nonenhancing small white matter lesions may also be seen. Viral meningitis is caused most commonly by the enteroviruses but can be caused by mumps, togaviruses, herpes simplex, lymphocytic choriomeningitis virus, and HIV. Most patients do not require treatment and neurologic deficits are uncommon. Imaging studies are typically normal.
Subdural
and
Epidural
Infections
Extra-axial pyogenic infections may also involve the epidural or subdural spaces. An epidural abscess may be caused by penetrating injuries, surgery, sinusitis, mastoiditis, orbital infection, or, rarely, hematogenous spread. CT scans show an inwardly convex, extraaxial collection with increased density compared to CSF (Fig. 6.26) . The inner margin usually enhances with contrast. There may be adjacent sinusitis or skull abnormalities. MR is more sensitive in demonstrating these lesions and can do so in multiple planes. The strong dural attachments prevent rapid expansion of epidural abscesses. However, a subdural empyema may spread rapidly throughout the subdural space and is acutely life threatening. Cortical venous thrombosis resulting in venous infarcts is a common result of these infections. Subdural empyemas are caused by the same conditions that produce epidural abscesses. Both CT and MR
435
can demonstrate these subdural collections, which usually have enhancing inner margins (Fig. 6.27A, B). MR is more sensitive in showing smaller lesions because of the problem of partial volume averaging with the calvarium on CT. MR is also better at detecting venous thrombosis and venous infarcts. Subdural empyemas show increased signal on DWI, distinguishing them from subdural effusions (Fig. 6.27C) . Mild, smooth dural, or meningeal enhancement may be seen after brain surgery and in patients with a ventriculostomy tube, especially with MR (Fig. 6.28). The enhancement can persist for years and should be considered benign in this clinical setting. It most likely reflects a chemical meningitis resulting from perioperative hemorrhage. Intracranial hypotension from a spontaneous or iatrogenic CSF leak also results in smooth dural thickening and enhancement.
ACQUIRED SYNDROME
IMMUNODEFICIENCY
The CNS is a common site of involvement in patients with AIDS. The incidence of CNS involvement has decreased since the introduction of highly active antiretroviral therapy (HAART), yet up to two thirds of AIDS
patients P.175
develop some kind of CNS disease. A variety of infections and neoplasms may be diagnosed in these patients. The most common infections include HIV encephalopathy; toxoplasmosis, cryptococcosis, and other fungal infections; CMV and Herpes meningoencephalitis; mycobacterial infection; PML; and meningovascular syphilis. Primary CNS lymphoma is by far the most common tumor, but metastatic lymphoma, gliomas, and rarely Kaposi sarcoma may also occur.
436
FIGURE 6.21. West Nile Fever. A. Fluid-attenuated inversion recovery scan in this 7-year-old child with lethargy shows marked increased signal in the thalami bilaterally. B . Diffusionweighted imaging sequence shows some increased signal, but most of the thalami do not demonstrate restricted diffusion. C . There is no abnormal contrast enhancement on the postcontrast
437
T1W scan. CSF studies were positive for the virus that causes West Nile Fever.
HIV
Encephalopathy
The etiologic agent in AIDS, HIV is neurotropic, infecting the brain in up to 90% of patients at autopsy. Clinical symptoms of brain involvement by HIV occur in a minority of these patients. Pathologically, HIV infection results in vacuolation of the white matter, with areas of demyelination and multinucleated giant cells. The centrum semiovale is involved most severely, but all white matter tracts, including the brainstem and cerebellum, may be affected. The cortical gray matter is P.176 P.177 usually spared. Clinically, patients with HIV encephalitis may develop a subcortical dementia with cognitive, behavioral, and motor deterioration. This is known as AIDS dementia complex (ADC), which occurs in about 7% to 15% of AIDS patients. Infants and children with HIV encephalitis exhibit loss of developmental milestones, apathy, failure of brain growth, and spastic paraparesis. This is the most common form of CNS disease in pediatric patients with AIDS, in whom opportunistic infections and CNS tumors are unusual.
438
FIGURE
6.22. Bacterial
Meningitis.
439
A. The initial contrast-
enhanced CT scan on this 3-month-old boy is normal. B . A contrast-enhanced CT scan obtained one day later shows marked brain swelling with focal areas of low density representing edema or ischemia in the frontal and occipital lobes. C . One month later, an intracranial US shows ventriculomegaly from marked cortical atrophy resulting from widespread cortical destruction.
FIGURE 6.23. Subdural Effusion. A contrast-enhanced CT scan on this 6-year-old with Haemophilus influenzae meningitis reveals a subdural collection nearly isodense with CSF (arrowheads). Subdural effusions are common with H influenzae meningitis. There is also enlargement of the lateral and third ventricles because of communicating hydrocephalus, which is a common complication of meningitis.
440
Diffuse atrophy is the most common manifestation of HIV infection of the brain on neuroimaging studies (Fig. 6.29). This is largely central atrophy, reflecting the predominant white matter involvement. White matter lesions are also commonly seen in patients with ADC. MR is significantly more sensitive than CT for detecting these abnormalities. A diffuse pattern of increased signal in the deep white matter or multiple small punctate white matter lesions on T2WIs are the most common findings. The punctate lesions do not correlate well with symptoms. The lesions do not exhibit mass effect or abnormal contrast enhancement. The most severe cases of HIV encephalopathy show extensive bilateral areas of abnormal signal throughout the periventricular white matter, brainstem, and cerebellum (Fig. 6.30). A decreased nacetyl aspartate (NAA) peak is often found in the affected white matter with MR spectroscopy. Such severe involvement usually correlates with symptoms of ADC. The clinical and imaging abnormalities often respond to treatment with HAART. In infants and children with HIV infection, atrophy is the most common observation, followed by calcifications in the basal ganglia. White matter calcifications and low-density lesions are also sometimes seen.
441
FIGURE 6.24. Tuberculous Meningitis. Contrast-enhanced CT scan shows marked abnormal contrast enhancement in the left sylvian fissure, interhemispheric fissure, ambient cistern, and along the tentorium. This thick, irregular enhancement in the basal cisterns is typical of a pachymeningitis such as tuberculosis or fungal meningitis. CT scans in patients with bacterial meningitis are usually normal or may reveal subtle increased density or enhancement in the peripheral sulci.
Toxoplasmosis is the most common opportunistic CNS infection in AIDS, occurring in about 13% to 33% of patients with CNS complications. It occurs in patients with CD4 counts below 100 cells/mm3 . T gondii is an obligate intracellular protozoan that is ubiquitous throughout the world, causing subclinical or mild infection
442
in a large percentage of the population. In AIDS, CNS toxoplasmosis results from reactivation of the previously acquired infection. A necrotizing encephalitis usually results, with the formation of thinwalled abscesses. Patients present clinically with headache, fever, lethargy, decreased level of consciousness, and focal signs, which initially can be P.178 confused with the subacute encephalitis of HIV infection. Neuroimaging studies are therefore crucial in patient management.
FIGURE
6.25. Coccidioidomycosis
Meningitis.
Contrast-
enhanced, T1W axial (A) and coronal (B) scans reveal abnormal enhancement of the meninges in the basal cisterns (arrowheads) .
443
FIGURE
6.26. Epidural
Abscess. This patient had a penetrating
injury to the frontal bone several months prior to this contrastenhanced CT scan. There is medium-density pus extending through the calvarial defect into the epidural space. The inner margin enhances markedly. Surgical clips are present near the frontal horn of the right lateral ventricle.
The typical appearance of CNS toxoplasmosis is that of multiple enhancing mass lesions with surrounding vasogenic edema (Figs. 6.31, 6.32). The lesions are usually relatively small—between 1 and 4 cm in diameter. Larger lesions usually exhibit ring enhancement, while smaller lesions are usually solid. The lesions are usually of increased signal on precontrast T2WIs. Unlike bacterial abscesses, toxoplasmosis lesions are not high in signal on DWI. The basal
444
ganglia are a favored site, but white matter and cortical lesions are also common. The main differential consideration is primary CNS lymphoma, which will be discussed later. A clinical and imaging response to antitoxoplasmosis antibiotics will usually distinguish between toxoplasmosis and lymphoma in most cases (Fig. 6.32) . Biopsy is usually reserved for atypical cases or when there is no response to antibiotics. Other infections or tumors may occasionally mimic toxoplasmosis but are unusual. Fungal, mycobacterial, and amebic abscesses have been described. Bacterial abscesses are rare in AIDS patients.
Fungal
Meningitis
Although fungal abscesses and granulomas are unusual, fungal meningitis is a common complication of AIDS, occurring in 5% to 15% of patients. Cryptococcosis is the most common fungal infection. Meningitis is usually mild because of the diminished inflammatory response of the immunocompromised host. Therefore, there is usually little or no enhancement of the meninges, and imaging studies are usually normal. The P.179 P.180 P.181 diagnosis is made when there are elevated cryptococcal antigen titers in the serum and CSF.
445
FIGURE 6.27. Subdural Empyema. A. The T2WI of this 8-yearold patient shows a thin subdural fluid collection of increased
446
signal along the left cerebral hemisphere with mass effect. B . Postcontrast T1WI shows low-signal left subdural fluid with dural enhancement. C . Diffusion-weighted scan shows increased signal in the fluid, indicating an empyema, not a sterile subdural effusion, which would be dark on this sequence.
FIGURE 6.28. Benign Postoperative Meningeal Enhancement. Several years after brain surgery, this T1W, contrast-enhanced MR scan reveals smooth but definitely abnormal enhancement of the dura (small arrowheads). There were no signs of infection or tumor recurrence. A ventricular shunt tube is seen on the right side (large arrowheads) .
447
FIGURE 6.29. AIDS-Related Atrophy. A noncontrast CT scan reveals enlarged ventricles and sulci in this 24-year-old patient with AIDS. This is the most common abnormality found on brain imaging of patients with AIDS. It often correlates with the AIDS dementia complex.
448
FIGURE 6.30. HIV Encephalopathy. This young patient clinically had the AIDS dementia complex. The T2WI shows widespread abnormal high signal in the periventricular white matter.
449
FIGURE
6.31. Toxoplasmosis. A contrast-enhanced CT scan
reveals bilateral ring-enhancing lesions in the basal ganglia of this patient with AIDS. There is marked surrounding low-density edema. The basal ganglia are a common site for toxoplasmosis.
450
FIGURE
6.32. Toxoplasmosis.
A. A contrast-enhanced CT scan
shows a large right basal ganglia–enhancing mass and several other small enhancing lesions (arrows). The small size and multiplicity of the lesions favor toxoplasmosis over lymphoma. B . Following 2 weeks of antibiotic therapy, the contrast-enhanced CT scan reveals complete resolution of the lesions, typical for toxoplasmosis.
As already mentioned, cryptococcosis may sometimes present as dilated Virchow-Robin spaces filled with cryptococcus organisms, known as gelatinous pseudocysts. These cysts appear as rounded, smoothly marginated lesions in the basal ganglia that are nearly isodense and isointense to CSF (Fig. 6.33). There is no enhancement following contrast administration, which distinguishes these from toxoplasmosis. Enhancing cryptococcomas are rare.
lesions
Progressive multifocal leukoencephalopathy is an infection of immunosuppressed patients caused by reactivation of a papova virus (the JC virus). The incidence of PML in AIDS patients is
451
approximately 8%. It can also occur in other immunosuppressed patients, such as transplant recipients and leukemics, but it does not occur in patients with normal immunity. The infection causes demyelination and necrosis, primarily involving white matter. Clinical symptoms include changes in mental status, blindness, aphasia, hemiparesis, ataxia, and other focal findings. There is a progressive course to death within months, although treatment with HAART significantly prolongs survival. In non-AIDS immunosuppressed patients, PML has a predilection for the occipital lobes, but in AIDS patients, any part of the brain may be involved. MR reveals focal lesions of increased signal on FLAIR and T2W images and decreased signal on T1WIs within the subcortical and deep white matter (Fig. 6.34). CT shows white matter lesions of decreased density. The lesions may be solitary or multifocal. Mass effect and contrast enhancement are almost always absent, which are important distinguishing features. Rarely, both gray and white matter or the basal ganglia are involved, simulating an infarct. The main differential diagnosis in the setting of AIDS is that of HIV encephalitis. Unlike PML, HIV encephalitis is usually more diffuse and less intense on T2WIs, and does not extend to the gray–white junction.
Viral
Infection
CMV infection is a common CNS infection in AIDS patients pathologically but does not usually result in frank tissue necrosis and is usually subclinical. There are many cases of pathologically proven CMV brain infection with normal CT and MR scans. CMV meningoencephalitis is occasionally imaged as areas of increased signal on T2WIs in the periventricular white matter. Subependymal contrast enhancement, if present, is a valuable diagnostic sign. Rarely, CMV will present as a ring-enhancing mass. Herpes virus and varicella virus infections are also only occasionally imaged. In AIDS patients, these viral infections often have a more benign clinical course and imaging appearance because of a diminished immune response.
452
FIGURE
6.33. Cryptococcosis
and
Toxoplasmosis.
A. T2WI
reveals multiple rounded lesions that are isointense to CSF in the basal ganglia (small arrowheads). There is no surrounding edema. Darker lesions with surrounding edema are present in the right frontal and left occipital areas (large arrowheads). B . A contrast-enhanced T1WI again reveals the basal ganglia lesions to be isointense with CSF (small arrowheads). There is no contrast enhancement. The appearance of these lesions is typical of gelatinous pseudocysts of cryptococcosis. These lesions represent dilated Virchow-Robin spaces filled with cryptococcus organisms. The right frontal and left occipital lesions do enhance with contrast (large arrowheads), as is typical of toxoplasmosis.
453
FIGURE 6.34. Progressive Multifocal Leukoencephalopathy (PML). A. There is an area of abnormal high signal intensity in the right corona radiata on this T2WI. There is no significant mass effect. B . On the contrast-enhanced T1WI, the lesion is of low signal intensity (arrowhead), and does not enhance. These are typical features of PML, which was proven with biopsy in this patient with AIDS. Incidentally, a left temporal arachnoid cyst can be noted.
P.182 P.183 Intracranial mycobacterial infections occur in a small percentage of AIDS patients. Most of these patients are intravenous drug abusers with pulmonary tuberculosis. Chest radiographs are positive in about 65% of cases. There is a very high mortality rate (nearly 80%) in these patients. Most patients present with meningitis. Imaging studies in these patients reveal communicating hydrocephalus and/or meningeal enhancement. Tuberculomas occur in about 25% of patients with HIV-related CNS tuberculosis, but tuberculous abscesses are less common. Tuberculomas are usually
454
smaller and have less edema than tuberculous abscesses. Primary CNS lymphoma is by far the most common intracranial tumor in AIDS. Up to 6% of AIDS patients will develop this tumor. It is the main differential diagnostic consideration along with toxoplasmosis when a mass lesion is found in patients with AIDS. Patients present with symptoms of a space-occupying lesion, as with toxoplasmosis. Solitary or multiple enhancing mass lesions are found with neuroimaging studies (Fig. 6.35). The lesions are usually centrally located within the deep white matter or basal ganglia, but cortical lesions also occur. There may be subependymal spread or extension across the corpus callosum, which do not usually occur with toxoplasmosis. With MR imaging, there is variable signal intensity, with areas of low or high signal on T2WIs and isosignal or low signal on T1WIs. With CT, the lesions are often isodense with gray matter. The lesions almost always enhance with contrast in either a ring or solid pattern. The imaging appearance is often indistinguishable from that of toxoplasmosis. The main distinguishing features are size and number. Toxoplasmosis is more frequently multiple, and the lesions are usually smaller than with lymphoma. Isointensity with white matter on T2WIs and diffuse, homogeneous contrast enhancement favor lymphoma. High signal on T2WIs (often with a low signal rim) and ring enhancement following contrast administration
favor
toxoplasmosis.
MR
spectroscopy
shows
increased choline and decreased NAA with lymphoma, while toxoplasmosis shows decreased choline and NAA with increased lipid and lactate. Toxoplasmosis is also more common than lymphoma and responds to antibiotic therapy.
455
FIGURE 6.35. Primary CNS Lymphoma. There are two solidly enhancing mass lesions with surrounding edema on this CT scan of a patient with AIDS. The relatively large size and solid enhancement pattern are more suggestive of lymphoma than toxoplasmosis, as was proven in this case.
Suggested
Readings
Barkovich AJ, Lindan CE. Congenital cytomegalovirus infection of the brain: imaging analysis and embryologic considerations. AJNR Am J Neuroradiol 1994;15:703–715. Brightbill TC, Ihmeidan IH, Donovan Post MJ, Berger JR, Katz DA.
456
Neurosyphilis in HIV-positive and HIV-negative neuroimaging findings. AJNR Am J Neuroradiol 1995;16:703–711.
patients:
Collie DA, Summers DM, Ironside JW, et al. Diagnosing variant Creutzfeldt-Jakob disease with the pulvinar sign: MR imaging findings in 86 neuropathologically confirmed cases. AJNR Am J Neuroradiol 2003;24:1560–1569. Dumas JL, Visy JM, Belin C, Gaston A, Goldlust D, Dumas M. Parenchymal neurocysticercosis: follow-up and staging by MRI. Neuroradiology
1997;39:12–18.
Lai PH, Ho JT, Chen WL, et al. Brain abscess and necrotic brain tumor: discrimination with proton MR spectroscopy and diffusionweighted imaging. AJNR Am J Neuroradiol 2002;23:1369–1377. Lim CCT, Sitoh YY, Hui F, et al. Nipah viral encephalitis or Japanese encephalitis? MR findings in a new zoonotic disease. AJNR Am J Neuroradiol 2000;21:455–461. Mader I, Stock KW, Ettlin T, Probst A. Acute disseminated encephalomyelitis: MR and CT features. AJNR Am J Neuroradiol 1996;17:104–109. Mishra AM, Gupta RK, Jaggi RS, et al. Role of diffusion-weighted imaging and in vivo proton magnetic resonance spectroscopy in the differential diagnosis of ring-enhancing intracranial cystic mass lesions. J Comput Assist Tomogr 2004;28:540–547. Rosas H, Wippold FJ II. West Nile virus: case report with MR imaging findings. AJNR Am J Neuroradiol 2003;24:1376–1378. Sibtain NA, Chinn RJS. Imaging of the central nervous system in
457
HIV
infection.
Imaging
2002;14:48–59.
Stadnik TW, Demaerel P, Luypaert RR, et al. Imaging tutorial: differential diagnosis of bright lesions on diffusion-weighted MR images [erratum Radiographics 2003;23:686]. Radiographics 2003;23:e7. Thurnher MM, Schindler EG, Thurnher SA, Pernerstorfer-Schon H, Kleibl-Popov C, Rieger A. Highly active antiretroviral therapy for patients with AIDS dementia complex: effect on MR imaging findings and clinical course. AJNR Am J Neuroradiol 2000;21:670–678. Tien RD, Chu PK, Hesselink JR, Duberg A, Wiley C. Intracranial cryptococcosis in immunocompromised patients: CT and MR findings in 29 cases. AJNR Am J Neuroradiol 1991;12:283–289. Ukisu R, Kushihashi T, Kitanosono T, et al. Serial diffusionweighted MRI of Creutzfeldt-Jakob disease. AJR Am J Radiol 2005;184:560–566. Whiteman M, Espinoza L, Donovan Post MJ, Bell MD, Falcone S. Central nervous system tuberculosis in HIV-infected patients: clinical and radiographic findings. AJNR Am J Neuroradiol 1995;16:1319–1327. Wong AM, Zimmerman RA, Simon EM, Pollock AN, Bilaniuk LT. Diffusion-weighted MR imaging of subdural empyemas in children. AJNR Am J Neuroradiol 2004;25:1016–1021.
458
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section II - Neuroradiology > Chapter 7 White Matter and Neurodegenerative Diseases
Chapter 7 White Matter and Neurodegenerative
Diseases
Jerome A. Barakos In contrast to gray matter, which contains neuronal cell bodies, white matter is composed of the long processes of these neurons. The axonal processes are wrapped by myelin sheaths, and it is the lipid composition of these sheaths for which white matter is named. In this chapter, a host of diseases characterized by the involvement of white matter is described. This is followed by a discussion of hydrocephalus
and
neurodegenerative
disorders.
The marked sensitivity of T2WIs allows white matter lesions to be readily detected. The difficulty that confronts the radiologist is that a wide gamut of diseases may involve the white matter, and these lesions are often nonspecific in nature. An understanding of these white matter diseases, their clinical features, and parenchymal patterns of involvement is important in enabling the radiologist to generate a useful differential diagnostic list. Cerebral white matter diseases are classified into two broad categories: demyelinating and dysmyelinating. Demyelination is an acquired disorder that affects normal myelin. The vast majority of white matter diseases, especially in the adult, fall into this category and are the principal focus of this chapter. In contrast, dysmyelination is an inherited disorder affecting the formation or maintenance of myelin, and thus is typically encountered in the
459
pediatric population. Dysmyelination is rare and is discussed later in this chapter.
DEMYELINATING
DISEASES
Demyelinating disease can be divided into four main categories based on etiology: (1 ) primary, (2 ) ischemic, (3 ) infectious, and (4 ) toxic and metabolic (Table 7.1 ).
Primary
Demyelination
Multiple sclerosis (MS) is the classic example of a primary demyelinating disease. MS is a disease characterized by immune dysfunction in the production of abnormal immunoglobulins and T cells, which are activated against myelin and mediate the damage associated with the disease. MS is a chronic, relapsing, often disabling disease affecting more than a quarter of a million people in the United States alone. The age of onset is between 20 and 40 years, with only 10% of cases presenting in individuals older than 50. There is a female predominance of almost two to one. Although several environmental factors have been associated with MS, such as higher geographic latitudes and upper socioeconomic status, the etiology of MS remains unclear. Establishing a diagnosis of MS is challenging, because no specific examination, laboratory test, or physical finding is unequivocally diagnostic or pathognomonic of this disorder. Making a diagnosis of MS is portentous, as there are significant implications on many aspects of one's life, including eligibility for health insurance. However, P.185 establishing the diagnosis is important because promising therapies are available, including β-interferon and antineoplastic drugs. These agents suppress the activity of the T cells, B cells, and macrophages that are thought to lead the attack on the myelin sheath. Primary demyelination Multiple sclerosis
460
Ischemic demyelination Deep white matter infarcts Lacunar infarcts Vasculitis (including sarcoidosis Dissection Thromboembolic infarcts
and
lupus)
Migrainous ischemia Moyamoya disease Postanoxia Infection-related demyelination Progressive multifocal leukoencephalopathy HIV encephalopathy Acute disseminated encephalomyelitis Subacute
sclerosing
panencephalitis
Lyme disease Neurosyphilis Toxic
and
metabolic
demyelination
Central pontine myelinolysis Marchiafava-Bignami disease Wernicke-Korsakoff syndrome Radiation injury Necrotizing leukoencephalopathy Dysmyelination
(inherited
white
matter
disease)
Metachromatic leukodystrophy Adrenal leukodystrophy Leigh disease Alexander disease TABLE 7.1 Classification of White Matter Diseases The classic clinical definition of MS is multiple CNS lesions separated in both time and space. Patients may present with virtually any neurologic deficit, but they most commonly present with limb weakness, paresthesia, vertigo, and visual or urinary disturbances. Important characteristics of MS symptoms are their multiplicity and tendency to vary over time. The clinical course of MS is characterized by unpredictable relapses and remissions of
461
symptoms. The diagnosis can be supported with clinical studies, which include visual, somatosensory, or motor-evoked potentials and analysis of CSF for oligoclonal banding, immunoglobulin G index, and presence of myelin basic protein. Histopathologically, active MS lesions represent areas of selective destruction of myelin sheaths and perivenular inflammation, with relative sparing of the underlying axons. These lesions may occur throughout the white matter of the CNS, including the spinal cord. The inflammatory demyelination interrupts nerve conduction and nerve function, producing the symptoms of MS. Note that histopathologically, the inflammation is a key differentiating feature between MS and other white matter conditions, such as osmotic myelinolysis (central pontine and extrapontine myelinolysis) and posterior reversible encephalopathy syndrome (PRE), which lack inflammatory changes. MR is the most sensitive indicator in the detection of MS plaques, but imaging findings alone should never be considered diagnostic. In clinically confirmed cases of MS, MR typically demonstrates lesions in more than 90% of P.186 cases. This compares with far less than 50% for CT and 70% to 85% for laboratory tests such as brainstem-evoked potentials and CSF oligoclonal bands. Nevertheless, the ultimate diagnosis rests with the careful combination of clinical symptoms, history, and clinical testing, including MR imaging.
462
FIGURE
7.1. Multiple
Sclerosis. T2WI of a 26-year-old woman
with MS demonstrates a cluster of periventricular white matter lesions. These lesions are ovoid, and many are perpendicular to the long axis of the ventricles (perivenular in location, referred to as Dawson's fingers) (arrows ). Although the periventricular lesions are very suggestive of MS, these lesions are nonspecific and must be correlated with clinical examination and other clinical studies (visual, somatosensory, or motor-evoked potentials, and analysis of CSF for oligoclonal banding and immunoglobulin G index) before confirming a diagnosis of MS. These lesions may be indistinguishable from other demyelinating conditions, such as acute disseminated encephalomyelitis, Lyme disease, and autoimmune/connective disorders such as systemic lupus erythematosus.
463
tissue
A variety of T2WI techniques have been described for optimizing the detection of white matter lesions, including conventional spin-echo (SE) imaging, fast SE (FSE), short tau inversion recovery (STIR), and fluid-attenuated inversion recovery (FLAIR) sequences. As the name suggests, FLAIR imaging has the advantage of providing heavy T2 weighting while suppressing signal from CSF. As such, FLAIR images provide improved lesion conspicuity of periventricular lesions, which may be obscured by the bright signal of CSF on SE or FSE T2WIs. Comparative studies have demonstrated that FLAIR imaging provides the best visualization of supratentorial white matter lesions. However, the FLAIR sequence may have mild limitations when imaging the posterior fossa and spine, partly because of pulsation artifacts. MS plaques are typically round or ovoid, with a periventricular or subcortical location (Fig. 7.1 ). Lesions are bright on T2WIs, reflecting active inflammation or chronic scarring, and only a fraction of MS plaques will demonstrate contrast enhancement. Lesions that enhance are thought to reflect new lesions with active demyelination and disruption of the blood–brain barrier (Fig. 7.2 ). In older lesions, without residual inflammatory reaction, abnormal high signal on T2WIs persists, reflecting residual scarring. Within the CNS, cells can mount only a limited response to neuronal injury. This scarring typically manifests as a focal proliferation of astroglia at the site of injury, termed gliosis. In severe cases of MS, actual loss of neuronal tissue may occur and the white matter lesions may actually have dark signal on T1WIs, often referred to as the “dark lesions― of MS. These lesions are prognostically significant, since they reflect actual loss of underlying neuronal tissue rather than simple demyelination. Additionally, in chronic cases of MS, there is diffuse loss of deep cerebral white matter, with associated thinning of the corpus callosum and potential ex vacuo ventriculomegaly. MS lesions are nonspecific, and many of the diseases and conditions discussed in this chapter may have a similar appearance. Patients with migraines are especially challenging, because both their symptoms and imaging findings may closely mimic those of MS. A
464
pattern that is suggestive of MS is one of periventricular lesions that are ovoid and aligned perpendicular to the long axis of the ventricles. This pattern is the result of the alignment of the lesions along the perivenular spaces. Additional characteristic features include lesions along the callosal septal interface, as well as lesions that are confluent in nature and greater than 6 mm in diameter with a periventricular location (Fig. 7.3 ). In addition to the periventricular white matter, the cerebellar and cerebral peduncles as well as the corpus callosum, medulla, and spinal cord can be involved in MS (Fig. 7.4 ). Ischemic changes are rare in these locations; as a result, if periventricular lesions are accompanied by lesions in any of these areas, this dramatically increases the specificity for the diagnosis of MS. For example, because ischemic changes rarely involve the medulla and cerebellar/cerebral peduncles, the presence of posterior fossa lesions is a useful differential diagnostic factor in suggesting P.187 MS. This is particularly important in patients older than 50 years, because it is difficult to decide whether multifocal white matter lesions are the result of ischemia or a demyelinating process. Additional concepts for making this distinction are discussed in the next
subsection.
465
FIGURE
7.2. Multiple
Sclerosis
With
Lesion
Enhancement.
Images of a 28-year-old woman with a recent flareup in clinical symptoms. Axial fast spin-echo T2W (A) and gadolinium-enhanced T 1 W (B) images reveal interval development (compared to prior MR exam performed 6 months earlier) of a new contrast-enhancing lesion within the right brachium pontis (arrow ). The larger left anterior pontine lesion (curved
arrow ) is unchanged when compared
to the earlier exam and fails to enhance. Contrast enhancement is reflective of actively demyelinating lesions and can be used to assess disease activity.
466
FIGURE 7.3. Multiple Sclerosis With Callosal–Septal Involvement. Sagittal fluid-attenuated inversion recovery images show lesions located along the ventricular ependymal surface (arrows ) as well as along the callosal–septal interface (open arrows ), which are very characteristic for MS. The callosal–septal interface refers to the region where the septum pellucidum contacts the undersurface of the corpus callosum.
MS lesions may also present as a large, conglomerate, deep white matter mass that can be mistaken for a neoplasm (Fig. 7.5 ). A characteristic finding in these conglomerate MS plaques is that they often demonstrate a peripheral crescentic rim of contrast enhancement,
which P.188 P.189
represents the advancing region of active demyelination. Detecting this pattern of enhancement, and searching carefully for other more characteristic periventricular lesions, are helpful in distinguishing a giant MS plaque from a neoplasm.
467
FIGURE 7.4. Multiple Sclerosis With Brainstem Involvement. MR images of a 31-year-old male with a history of right-sided weakness and sensory changes. Axial proton density–weighted image (A) and coronal fluid-attenuated inversion recovery image (B) reveal a lesion involving the left corticospinal tracts at the level of the midbrain (cerebral peduncle) (arrow ). Lesions within the medulla and cerebellar/cerebral peduncles are quite characteristic for MS and serve to support this diagnosis when supratentorial lesions appear nonspecific in nature.
468
FIGURE 7.5. Tumefactive Multiple Sclerosis. Images from a 32year-old woman presenting with transient bouts of left hemiparesis, as well as depression and fatigue. Proton density weighted image (PDWI) (A) , T2WI (B) , and postgadolinium T1WI (C) reveal a large right parietal mass with a peripheral rim of enhancement (arrow ). This lesion could easily be mistaken for a neoplasm or
469
progressive multifocal leukoencephalopathy and undergo biopsy. The sagittal T1WI (D) demonstrates the presence of a characteristic periependymal lesion (arrow ), suggesting the diagnosis of MS. These are the “dark lesions― of MS, which are of greater concern than simple demyelinating plaques, because they represent actual neuronal loss. The diagnosis of MS was confirmed with additional clinical testing, including evoked potentials and CSF oligoclonal bands.
The spinal cord may also be involved with MS, and whenever a focal abnormality of the spinal cord is detected, a demyelinating MS plaque must be in the differential diagnosis. Demyelinating plaques may have mild mass effect as well as contrast enhancement, thus mimicking a neoplasm. The majority of spinal cord MS lesions (70% to 80%) will have associated plaques in the brain. In the setting of a cord lesion, performing an MR scan of the head may confirm the diagnosis, thus avoiding a spinal cord biopsy (see Chapter 10 ).
Ischemic
Demyelination
Although MR imaging is extremely sensitive in the detection of white matter lesions, a major difficulty in arriving at a diagnosis is that white matter lesions are often nonspecific. Thus, distinguishing MS lesions from other white matter lesions can be difficult. The most commonly encountered white matter lesions are ischemic in origin.
Age-Related
Demyelination
Small-vessel ischemic changes within the deep cerebral white matter are seen with such frequency in the older population (>60 years) that they are considered a normal part of aging. This represents an arteriosclerotic vasculopathy of the penetrating cerebral arteries. The deep white matter is more susceptible to ischemic injury than gray matter, because it is supplied by long, small-caliber penetrating end arteries, without significant collateral supply. In contrast, cortical gray matter, as well as parts of the brainstem such as the midbrain and medulla, have robust collateral blood supply, thus
470
minimizing the risk of ischemia. The deep penetrating vessels supplying the white matter become narrowed by arteriosclerosis and lipohyalin deposits. The result is the formation of small ischemic lesions, primarily involving the deep cerebral and periventricular white matter as well as the basal ganglia (Fig. 7.6 ). The cortex, subcortical “U― fibers, central corpus callosum, medulla, midbrain, and cerebellar peduncles are usually spared because of their dual blood supply, which decreases their vulnerability to hypoperfusion. As previously described, if lesions are identified in these locations, a cause other than ischemia should be entertained. Histologically, areas of infarction demonstrate axonal atrophy with diminished myelin. Early neuropathologists noted the areas of paleness associated with these changes and coined the term “myelin pallor.― These white matter changes have received many names over the years, including leukoaraiosis, microangiopathic leukoencephalopathy, and subcortical arteriosclerotic encephalopathy. None of these terms are very satisfying, as they do not accurately reflect all the changes observed histologically and overstate the clinical significance of these lesions. A more appropriate term may simply be “age-related white matter changes.― These small ischemic white matter lesions are often asymptomatic, and clinical correlation is always required before a diagnosis of subcortical arteriosclerotic encephalopathy or multi-infarct dementia (Binswanger disease) is made. The white matter infarcts just described differ from lacunar infarcts. Lacunae refer to small infarcts (5 to 10 ganglia, typically the upper two and deep white matter infarcts result of disease involving the
mm) occurring within the basal thirds of the putamina. Both lacunar have similar etiologies and are the deep penetrating arteries.
471
FIGURE
7.6. Ischemic
Demyelination. This 72-year-old woman
presented with forgetfulness. Axial fast spin-echo T2WI reveals diffuse patchy lesions throughout the subcortical and deep white matter. These lesions are in keeping with ischemic demyelination of the deep white matter, with several old lacunar infarcts of the basal ganglia (arrow ). Note the ex vacuo ventriculomegaly resulting from loss of deep cerebral white matter.
Differentiating white matter lesions related to ischemic changes from MS lesions can be difficult, especially in the older patient. This is important because 10% of patients who present with MS are older than 50 years of age. P.190 Clinical testing and history are helpful. Additionally, deep white matter infarcts tend to spare the subcortical arcuate fibers and the corpus callosum, both of which can be involved with MS. Involvement of the callosal–septal interface is quite specific for
472
MS.
FIGURE
7.7. Antiphospholipid
Antibody
Syndrome. This 32-
year-old woman presented with headaches and a history of several miscarriages. T2WIs demonstrated scattered focal subcortical and deep white matter lesions. Although these lesions are nonspecific, serum
testing
revealed
elevated
circulating
pathogenic
immunoglobulins/antibodies specifically targeting DNA and other nuclear constituents collectively termed antibodies to nuclear antigens, e.g., lupus anticoagulants and anticardiolipin antibodies. This represents an immune complex disease referred to as antiphospholipid antibody syndrome.
Nonspecific punctuate white matter lesions (small bright lesions on T2WIs) are more prominent in any patient with a vasculopathy, whether related to atherosclerosis (age, hypertension, diabetes, hyperlipidemia, coronary artery disease); hypercoagulable conditions; or vasculitis (lupus, sarcoid, polyarteritis nodosa, Behçet syndrome). In younger individuals with punctuate white matter lesions, hypercoagulable states, as well as embolic and
473
vasculitic etiologies, figure prominently (Figs. 7.7 , 7.8 , 7.9 ). Hypercoagulable conditions include a diverse set of diseases with the common theme of increased risk of microvascular thrombotic disease. Serum testing can be used to evaluate for the presence of these disease conditions, which include homocystinemia, antiphospholipid syndrome, Factor V Leiden, prothrombin gene mutation, and deficiencies of natural proteins that prevent clotting (the anticoagulant proteins such as antithrombin, protein C, and protein S deficiencies). A classic case presentation is that of a young adult female with prior miscarriages presenting with headaches/migraines and ischemic white matter changes. These findings are suggestive of antiphospholipid syndrome (a.k.a. phospholipid antibody syndrome), where circulating antiphospholipid antibodies (cardiolipin or lupus anticoagulant antibodies) lead to a hypercoagulable state with resultant white matter and ischemic changes. In the young adult population presenting with small white matter lesions, in addition to hypercoagulable conditions and migrainous ischemia, consider cardiogenic embolic etiologies. An echocardiogram plays an important role in the evaluation of a potential patent foramen ovale P.191 or valvular vegetation. In many normal children and young adults, subcortical lesions and periventricular hyperintensities are common; they are reported to be present in these locations in 6% and 74%, respectively, of the young normal population. Commonly these punctuate foci of white matter T2 hyperintensity will have no known etiology despite evaluation for all the conditions outlined above. In this setting, these lesions are simply the gliotic residue of a remote unspecified
insult,
usually
an
immune-mediated
474
postviral
condition.
FIGURE
7.8. Lupus
Cerebritis. Image from a 38-year-old woman
presenting with cognitive deficits and history of a connective tissue disorder. The T1WI demonstrates numerous dark periventricular lesions with striking loss of deep white matter and associated ex vacuo ventriculomegaly. These dark lesions represent underlying axonal loss with neuronal dropout, reflecting a more severe stage of white matter disease. These findings are characteristic of any severe or long-standing white matter disease such as chronic MS, or as in this case, chronic lupus cerebritis.
475
FIGURE 7.9. Moyamoya Disease. Six-year-old boy presents with episodes of focal motor weakness. T2WI (not shown) showed multiple scattered subcortical white matter T2 hyperintensities. MR angiography (A) and conventional angiography (B) reveal marked stenosis of the supraclinoid internal carotid vasculature (open arrow ), with a dramatic proliferation of tiny collateral vessels (arrows ) presenting as a “puff of smoke― (the literal Japanese translation of moyamoya ). The cause of this vascular disorder is unknown but can be treated with various external to internal vascular
bypass
surgeries
such
as
encephaloduroarteriosynangiosis.
MR angiography plays a useful role in assessing the patency of these shunts once surgically completed.
Ependymitis granularis is a normal anatomic finding that may mimic pathology. Ependymitis granularis consists of an area of high signal on a T2WI along the tips of the frontal horns (Fig. 7.10 ). These foci of signal range in width from several millimeters to a centimeter. Histologic studies of this subependymal area reveal a loose network of axons with low myelin count. This porous ependyma allows transependymal flow of CSF, resulting in a focal area of T2 prolongation. Unfortunately, this entity has been given a name that sounds more like a disease entity than a simple histologic observation. Similarly, with the use of FLAIR imaging, a region of periventricular T2 hyperintensity can be noted about the ventricular trigones as a normal finding. With age, prominent periventricular T2 hyperintensity may be noted along the entire length of the lateral
476
ventricles as a normal finding, and this may be referred to as senescent periventricular hyperintensity. Prominent
perivascular
spaces can also mimic deep white matter
or lacunar infarcts. As blood vessels penetrate into the brain parenchyma, they are enveloped by CSF and a thin sheath of pia. These CSF-filled perivascular clefts are called Virchow-Robin spaces and present as punctate foci of high signal on T2WIs (Fig. 7.11 ). They are typically located in the centrum semiovale (high cerebral hemispheric white matter) and the lower basal ganglia at the level of the anterior commissure, where the lenticulostriate arteries enter the brain parenchyma. These perivascular spaces are typically 1 to 2 mm in diameter but can be considerably larger. They can be seen as a normal variant at any age but become more prominent with increasing age as atrophy occurs. An important means for differentiating a periventricular space from a parenchymal lesion is the use of the proton density–weighted (first-echo T2W) or FLAIR images. On the proton density–weighted sequence, CSF has similar signal intensity as white matter. A perivascular space is composed of CSF and will parallel CSF signal intensity on all sequences (i.e., isointense to brain parenchyma on proton density sequences). In contrast, ischemic lesions, unless cavitated with cystic change, will be bright on the proton density sequence as a result of the presence of associated gliosis. Both a deep infarct and a perivascular space will be bright on the secondecho T2WI, but only the infarct will remain bright on the first-echo image. Similarly, on a FLAIR image, because fluid signal is attenuated, only true parenchymal lesions with gliosis will yield abnormal signal. On occasion, however, a small amount of persistent T2 hyperintensity can be associated with P.192 perivascular spaces on the proton density or FLAIR sequences. An additional differentiating feature between giant perivascular spaces and lacunae is location. Lacunar infarcts tend to occur in the upper two thirds of the corpus striatum because they reflect end-arteriole infarcts in the distal vascular distribution. In contrast, periventricular spaces are typically smaller, bilateral, and often
477
symmetric within the inferior third of the striatum, where the vessels enter the anterior perforated substance.
FIGURE
7.10. Ependymitis
Granularis
(Normal
Finding).
A. and
B . Axial fluid-attenuated inversion recovery images in a 42-year-old man presenting with headaches. The periventricular hyperintensity noted about the tips of the frontal and occipital ventricular horns is a normal finding (arrows ). These areas of periependymal hyperintensity may be exacerbated by any process that results in underlying white matter disease. Note the circular artifact located within the left basal ganglia; it is related to magnetic susceptibility artifact from the patient's orthodontic braces (curved arrow ). One should be aware of artifacts that may mimic pathologic lesions, especially flow and magnetic susceptibility artifacts that can give rise to lesions that are not necessarily contiguous to the cause of the artifact. B . Incidental note is made of a small focus of subcortical hyperintensity along the left temporoparietal lobe related to a site of posttraumatic gliosis (open arrow ).
478
FIGURE 7.11. Virchow-Robin Spaces. Small punctuate foci of water signal are noted within the centrum semiovale (A) and basal ganglia (B) , consistent with perivascular (PV) spaces. These spaces penetrate the brain parenchyma and reflect PV extensions of the pia
479
mater that accompany the arteries entering and the veins emerging from the cerebral cortex. These PV spaces are almost imperceptible on the proton density–weighted image (C) , which help confirm their identity as water, rather than white matter ischemic gliotic lesions. Although PV spaces are typically 1 to 2 mm in diameter, they can be considerably larger. Large PV spaces (about 0.5 to 1 cm) are occasionally noted within the caudal aspect of the basal ganglia and referred to as giant PV spaces. Coronal T1WI (D) and fast spin-echo T2WI (E) in a 38-year-old man demonstrate wellrounded, left-sided cysts along the course of the lenticulostriate arteries as they enter the basal ganglia through the anterior perforated substance (arrow ). An old cavitated lacunar infarction may have a similar appearance but would be distinctly unusual in the inferior portion of the striatum. Note that lacunar infarcts are the result of vessel occlusion and thus occur along the distal extent of the lenticulostriate arteries; therefore, they tend to be located more superiorly within the basal ganglia. Additionally, lacunar infarcts may have associated gliotic T2 hyperintensity on proton density and fluid-attenuated inversion recovery images, a finding not seen with giant PV spaces.
P.193 P.194
Infection-related
Demyelination
Various infectious agents may result in white matter disease, either directly or indirectly, and most commonly are viral. Some of the more common agents are described here. For further discussion of virus-induced white matter pathology, see Chapter 6 . Progressive multifocal leukoencephalopathy (PML) is seen with increasing frequency because of the growing number of AIDS patients. PML represents a reactivation of a latent JC polyoma virus. This opportunistic infection is usually seen in severely immunocompromised patients with very low T-cell counts, particularly individuals with AIDS, lymphoma, organ transplantation,
480
and disseminated malignancies. The JC virus infects oligodendrocytes, which are the axonal support cells that generate the myelin sheath. As a result, damage to the oligodendrocytes results in widespread demyelination. PML typically involves the deep cerebral white matter, with subcortical U-fiber involvement, but spares the cortex and deep gray matter (Fig. 7.12 ). Lesions are characterized by a lack of mass effect, contrast enhancement, and hemorrhage and are typically located in the parietooccipital region. These lesions progress rapidly and coalesce into larger confluent asymmetric areas. Although most lesions involve supratentorial white matter, gray matter and infratentorial involvement (cerebellum and brainstem) are not uncommon. PML is relentlessly progressive, with death typically ensuing within several months from the time of initial diagnosis.
HIV
Encephalopathy
HIV involvement of the brain presents as a subacute encephalitis, referred to as AIDS dementia complex or diffuse HIV encephalopathy . This is characterized clinically by a progressive dementia without focal neurologic signs. HIV encephalopathy does not appear to be the result of a direct infection of the neurons or macroglia (i.e., CNS support cells, astrocytes, oligodendrocytes). Instead, the active HIV infection develops in the microglia (brain macrophages). The cytokines and excitatory compounds that are produced as a result of this infection have a toxic effect on adjacent neurons. HIV encephalopathy most often results in mild cerebral atrophy without a focal abnormality. Occasionally, HIV encephalopathy causes focal or diffuse white matter hyperintensities on T2WIs. Typically, HIV white matter involvement presents as subtle, diffuse T2 hyperintensity that often is bilateral and relatively symmetric (Fig. 7.13 ). This supratentorial white matter signal abnormality is ill defined and often involves a large area, in contrast to the dense lesions that are characteristic of PML. HIV encephalopathy can also present with more P.195 focal punctate lesions. HIV lesions do not demonstrate contrast
481
enhancement.
FIGURE
7.12. Progressive
Multifocal
Leukoencephalopathy. A
32-year-old HIV-positive man presents with cognitive deterioration and weakness. Proton density weighted image (PDWI) (A) and T2WI (B) reveal large confluent areas of T2 hyperintensity in the subcortical white matter of the parietooccipital lobes (arrows ). Characteristic features of this demyelinating process include minimal mass effect, despite the large size of these patchy white matter lesions, and essentially no contrast enhancement or hemorrhage. A very low T-cell count reflecting an immunocompromised status is key to the diagnosis. In an immunocompetent patient, differential diagnostic considerations would include posterior reversible encephalopathy syndrome, which can have an identical imaging appearance.
482
FIGURE
7.13. HIV
Encephalopathy. T2WI of a 27-year-old man
demonstrates diffuse hazy hyperintensity of the deep cerebral white matter (arrows ), as well as cortical atrophy. Note how this hazy T2 hyperintensity differs from the dense confluent lesions of progressive multifocal leukoencephalopathy shown in Figure 7.12 .
Demyelination may also occur as an indirect result of a viral infection. Specifically, demyelination may follow a viral illness, the result of a virus-induced autoimmune response to white matter. Acute disseminated encephalomyelitis (ADEM), a postinfectious and postvaccinal encephalomyelitis, typically occurs after a viral illness or vaccination, with measles, rubella, varicella, and mumps being the most common agents. This condition is considered an
483
immune-mediated inflammatory demyelinating disease, but sometimes it has no recognized antecedent infection or inciting malady. It is theorized that the body's antiviral immune reaction cross-reacts with myelin sheaths, resulting in an acute, aggressive form of demyelination. This unintended antiviral response against myelin is a result of shared molecular homology between viral proteins and normal human CNS proteins. Recall that oligodendrocytes are responsible for the formation and maintenance of the myelin sheaths, and their damage results in demyelination. Demyelinating lesions associated with ADEM typically begin approximately 2 weeks after a viral infection with the abrupt clinical onset of neurologic symptoms, which include decreased levels of consciousness varying from lethargy to coma; convulsions; multifocal neurologic symptoms such as hemiparesis, paraparesis, and tetraparesis; cranial nerve palsies; movement disorders; and seizures. In the majority of cases, there is spontaneous resolution of symptoms, but permanent sequelae can be seen in up to 25% of patients, with some even progressing to death. Although ADEM occurs most commonly in children, persons of any age can be affected. Lesions primarily involve white matter, but gray matter may also be affected. MR imaging demonstrates multifocal or confluent white matter lesions similar to those of MS (Fig. 7.14 ). A differential feature is that ADEM is a monophasic illness, unlike MS, which has a remitting and relapsing course. This is a feature often useful in differentiating ADEM from MS. Specifically, if the majority of the identified white matter lesions enhance, this suggests a monophasic demyelinating process (i.e., ADEM). Subacute
sclerosing
panencephalitis represents a reactivated,
slowly progressive infection caused by the measles virus. Children between the ages of 5 and 12 years who have had measles, usually before the age of 3, are typically affected. MR demonstrates patchy areas of periventricular demyelination as well as lesions of the basal ganglia. The disease course is variable and may be rapidly progressive or protracted.
484
Herpes encephalitis is the most common fatal encephalitis. Although this condition is also discussed in Chapter 6 , its importance warrants repetition. The form of herpes encephalitis that we will discuss occurs in children and adults and is caused by herpes simplex virus (HSV) type 1 (oral herpes); this is in contrast to neonatal herpes encephalitis, which is caused by herpes simplex virus 2 (genital herpes). Presenting symptomology may be nonspecific, such as headache, fever, mental deterioration, and seizures. As a result of this variable clinical presentation, diagnosis may be difficult. This emphasizes the crucial role of the radiologist in entertaining this diagnosis when appropriate imaging findings are noted. Antiviral treatment is simple and effective, but failure to treat yields 100% mortality. Although the diagnosis may be confirmed by polymerase chain reaction detection of herpes DNA in CSF, therapy must be instituted prior to the return of this test result. HSV type 1 has a particular predilection for the limbic system, with localization of infection to temporal lobes, insular cortex, subfrontal area, and cingulate gyri (Fig. 7.15 ). The limbic system is responsible for integration of emotion, memory, and complex behavior, and involvement of these structures accounts for some of the behavioral symptoms at presentation. Imaging reveals primarily T2 hyperintensity of the involved cortex and subcortical structures presenting as an encephalitis with variable contrast P.196 enhancement.
Initially,
herpes
encephalitis
is
usually
unilateral;
however, sequential bilateral involvement is highly suggestive of the disease. Histopathologically, herpes infection is a fulminant necrotizing meningoencephalitis associated with edema, necrosis, hemorrhage, and eventually encephalomalacia. As a result, hemorrhage within the area of involved parenchyma is strongly supportive of this diagnosis.
485
FIGURE 7.14. Acute Disseminated Encephalomyelitis. T2WIs (A, B) and postgadolinium T1WI (C) in a 7-year-old boy who presented with deteriorating mental status 10 days following viral gastroenteritis. Imaging reveals multiple patchy subcortical white
486
matter lesions as well as involvement of deep gray matter structures, including the corpus striatum (lentiform nucleus plus caudate nucleus) and the thalamus (arrows ). Following the administration of gadolinium-DTPA, numerous punctate foci of enhancement are noted consistent with an acute demyelinating process. The enhancement of most lesions is suggestive of a monophasic demyelinating process. The patient improved after treatment with steroids.
Toxic
and
Metabolic
Demyelination
Central pontine myelinolysis (CPM) is a disorder that results in characteristic demyelination of the central pons. This is most commonly seen in patients with electrolyte abnormalities, particularly involving hyponatremia, that are rapidly corrected, giving rise to the term “osmotic demyelination syndrome.― This condition occurs most commonly P.197 in children and alcoholics with malnutrition. Occasionally, cases have been associated with diabetes, leukemia, transplant recipients, chronically debilitated patients, and others with conditions resulting in chronic malnutrition. The clinical course is classically described as biphasic, beginning with a generalized encephalopathy caused by the hyponatremia, which usually transiently improves following initial correction of sodium. This is followed by a second neurologic syndrome, which occurs 2 to 3 days following correction or overcorrection of hyponatremia caused by myelinolysis. This latter phase is classically characterized by a rapidly evolving corticospinal syndrome with quadriplegia, acute changes in mental status, and a “locked-in― state in which the patient is mute, unable to move, and occasionally comatose. Patients tend to be extremely ill and often have a very poor prognosis.
487
FIGURE 7.15. Herpes Encephalitis. T2WI of a 31-year-old man who presented with behavioral disturbance and new-onset seizures. MR demonstrates diffuse hyperintensity of the right insular cortex, and the adjacent orbitofrontal and temporal lobes are characteristic for herpes encephalitis (arrows ). The radiologist must have a low threshold for considering this diagnosis when signal abnormality of the temporal lobes, insular cortex, or cingulate gyrus is noted, as failure of treatment results in 100% mortality.
The pathophysiology of CPM relates to a disturbance in the physiologic balance of osmoles in the brain. Oligodendroglial cells are most susceptible to CPM-related osmotic stresses, with the distribution of CPM changes paralleling the distribution of oligodendroglial cells within the central pons, thalamus, globus
488
pallidus, putamen, lateral geniculate body, and other extrapontine sites. The mechanism of myelinolysis remains to be completely elucidated, but it appears to be distinct from a demyelinating process like that of MS, in which an inflammatory response predominates. CPM is characterized by intramyelinitic splitting, vacuolization, and rupture of myelin sheaths, presumably because of osmotic effects. However, there is preservation of neurons and axons. Note that there is no inflammatory reaction associated with osmotic demyelination, differentiating this process from MS, which is characterized by marked perivascular inflammation. MR characteristically demonstrates abnormal high signal on T2WI, corresponding to the regions of central pontine demyelination (Fig. 7.16 ). Additionally, extrapontine sites of involvement have been described in this condition, including the white matter of the cerebellum, thalamus, globus pallidus, putamen, and lateral geniculate body, giving rise to the term extrapontine myelinolysis. Posterior
reversible
encephalopathy
syndrome (PRE) is a
condition characterized by signal changes within the brain parenchyma, primarily involving the posterior vascular distribution. This condition has also been referred to as reversible posterior leukoencephalopathy syndrome . Patients present with headache, seizures, visual changes, and altered mental status, with MR revealing symmetric areas of bilateral subcortical and cortical vasogenic edema within the parietooccipital lobes (Fig. 7.17 ). The leading theory regarding the etiology of this condition is a temporary failure of the autoregulatory capabilities of the cerebral vessels, leading to hyperperfusion, breakdown of the blood–brain barrier, and consequent vasogenic edema, but no acute ischemic changes. Autoregulation maintains a constant blood flow to the brain, despite systemic blood pressure alterations, but this can be overcome at a “breakthrough― point, at which point the increased systemic blood pressure is transmitted to the brain, resulting in brain hyperperfusion. This increased perfusion pressure is sufficient to overcome the blood–brain barrier, allowing extravasation of fluid, macromolecules, and even red blood cells into the brain parenchyma. The preferential involvement of the parietal and
489
occipital lobes is thought to be related to the relatively poor sympathetic innervation of the posterior circulation. A very diverse set of conditions leads to this characteristic clinical and radiologic presentation, including treatment with cyclosporin A or tacrolimus (FK506), acute renal failure/uremia, hemolytic uremic syndrome, eclampsia, thrombotic thrombocytopenia purpura, and treatment with a wide variety of chemotherapeutic agents, including interferon. This suggests a final common etiologic pathway involving either endothelial injury, elevated blood P.198 pressure, or a combination of these factors. Associated clinical conditions presumably contribute to this physiologic effect by cytotoxic effects on the vascular endothelium (endotoxins), causing increasing capillary permeability that allows this process to occur at near normal blood pressures, or by inducing or exacerbating hypertension. Hypertension is commonly associated with PRE but may be relatively mild and is not universally present, especially in the setting of immunosuppression. Note that this condition is not always reversible and may occasionally result in hemorrhagic infarctions.
490
FIGURE
7.16. Central
Pontine
Myelinolysis
(CPM). A 52-year-
old alcoholic was admitted with a serum sodium of 110 mEq/mL. After rapid normalization of sodium, the patient became comatose.
491
The T2WI (A) demonstrates well-defined intense high signal within the basis pontis (arrows ). Often, the central corticospinal tracts remain preserved, giving rise to this characteristic appearance of two rounded areas of spared central pontine tracts. Note that CPM should be differentiated from ischemic demyelination, as both conditions may have T2 hyperintensity within the basis pontis. B . and C. A 72-year-old man presented with progressive confusion and known vasculopathy related to longstanding hypertension and diabetes, without evidence of osmotic or electrolyte disturbance. The T2WIs reveal diffuse patchy T2 hyperintensity of the basis pontis (arrows ). Given the clinical history, this finding is consistent with small vessel ischemic changes within the pons rather than CPM. Statistically speaking, hyperintensity within the pons will be more often related to ischemic demyelination than CPM, simply because of the relative frequency of ischemic pathology. However, clinical history will allow easy differentiation between these conditions.
Marchiafava-Bignami disease is a rare form of demyelination seen most frequently in alcoholics. This condition was first described in Italian red wine drinkers, but it has since been reported with other types of alcohol use as well as in nonalcoholics. The disease is characterized by demyelination involving the central fibers (medial zone) of the corpus callosum, although other white matter tracts may be involved, including the anterior and posterior commissures, the centrum semiovale, and the middle cerebral peduncles. This is felt to reflect a form of osmotic demyelination, as discussed earlier in extrapontine myelinolysis. P.199 Onset is usually insidious, with the most common symptom being nonspecific dementia.
492
FIGURE 7.17. Posterior Reversible Encephalopathy Syndrome (PRES). A 43-year-old transplant patient who was being treated with cyclosporine presented with visual disturbances and confusion. Axial (A) and coronal (B) T2WIs reveal patchy areas of cortically based signal abnormality within the parietooccipital lobes (arrows ), corresponding to the posterior vascular distribution. These findings are in keeping with dysfunction of vascular permeability, the result of a combination of endothelial toxicity and elevated blood pressure. Both clinical symptoms and imaging findings resolved after the cyclosporine
doses
were
reduced.
Wernicke encephalopathy and Korsakoff syndrome are metabolic disorders caused by thiamine (B1 vitamin) deficiency secondary to poor oral intake in severe chronic alcoholics (most common association), hematologic malignancies, or recurrent vomiting in pregnant patients. In fact, this condition may occur in many different non-alcohol-related pathologic conditions that share the common denominator of malnutrition. In general, there is a good clinical response to thiamine administration. Classically, Wernicke encephalopathy is characterized by the clinical triad of acute onset of ocular movement abnormalities, ataxia, and confusion. Korsakoff,
493
a Russian psychiatrist, described the disturbance of memory in longterm alcoholics. Therefore, if persistent learning and memory deficits are present in patients with Wernicke encephalopathy, the symptom
complex
is
termed
Wernicke-Korsakoff
syndrome.
In the acute stage of this disease, MR may reveal T2 hyperintensity or contrast enhancement of the mamillary bodies, basal ganglia, thalamus, and brainstem, with periaqueductal involvement. In contrast, the chronic stage may show atrophy of the mamillary bodies, midbrain tegmentum, as well as dilatation of the third ventricle. Except for the mamillary body involvement, these findings are very similar to Leigh disease, which supports the notion that enzymatic deregulation in Leigh disease is tied in some fashion to thiamine
metabolism.
Radiation
Leukoencephalitis
Radiation may result in damage to the white matter secondary to a radiation-induced vasculopathy. Radiation leukoencephalitis usually follows a cumulative dose in excess of 40 Gy delivered to the brain and occurs 6 to 9 months after treatment. Findings consist of areas of abnormal high signal on T2WIs, typically involving confluent areas of white matter extending to involve the subcortical U fibers in the distribution of the irradiated brain (Fig. 7.18 ). Note that this represents an indirect effect of radiation on the brain and results from an arteritis (endothelial hypertrophy, medial hyalinization, and fibrosis) involving small arteries and arterioles.
Radiation
Necrosis
and
Radiation
Arteritis
In contrast to the rather benign nature of radiation leukoencephalitis, radiation necrosis and radiation arteritis are major hazards related to CNS radiation. Both of these radiation effects are strongly dose related and are less P.200 commonly seen today because of greater fractionation of CNS radiation doses. Radiation necrosis may occur several weeks to years after radiation, but it most commonly occurs between 6 and
494
24 months after radiation. Radiation necrosis is rarely noted at less than 6 months after treatment unless gamma knife is employed (Fig. 7.19 ). Note that gamma knife is an ablative procedure designed to destroy targeted tissue and thus may more easily incite frank radiation necrosis. This is in contrast to radiation therapy, which is not ablative in nature. Radiation necrosis can be progressive and fatal. Radiation necrosis typically presents as an enhancing lesion with mass effect and ring enhancement or as multiple foci of enhancement, mimicking recurrent neoplasm. Radiation may also induce telangiectasia within the radiation field, which may appear similar to cryptic vascular malformations.
FIGURE 7.18. Radiation Leukoencephalopathy. MR of a 62-yearold woman obtained 1 year after whole brain radiation for metastatic breast carcinoma to the brain shows a delayed neurologic sequelae of radiotherapy. Coronal fluid-attenuated inversion recovery image reveals confluent areas of high signal involving the periventricular white matter (arrows ). This finding may be associated with loss of deep cerebral white matter with concomitant ex vacuo ventriculomegaly, as noted in this case. Although this condition may
495
result in some degree of neurocognitive deficits, this patient was entirely asymptomatic and was simply returning for a routine followup examination.
Radiation necrosis is found most commonly in or near the irradiated tumor bed, but it sometimes is more remote from the tumor bed. It is theorized that the partially injured brain parenchyma within and adjacent to the tumor bed is more susceptible to radiation injury, thus accounting for the distribution of radiation necrosis. After resection of a brain neoplasm and subsequent radiation therapy, it can be very difficult to differentiate tumor recurrence from radiation-associated necrosis, because both conditions may continue to grow and demonstrate imaging features characteristic of neoplasm, i.e., lesion growth, irregular ring enhancement, edema, and mass effect (Fig. 7.20 ). If during serial scanning a lesion within the treated tumor bed stabilizes and regresses, this is obviously radiation necrosis, but if the lesion progresses, differentiation between tumor and radiation necrosis is difficult. PET and MR spectroscopy (MRS) are valuable in distinguishing between tumor recurrence and radiation necrosis. With PET scanning, a short-lived radioactive isotope (e.g., 18 F fluorodeoxyglucose) that decays by emitting a positron, is combined with glucose, a metabolically active molecule. This tracer mimics glucose and is taken up and retained by tissues with higher than normal metabolic activity, such as tumor recurrence. This is in contrast to radiation necrosis, which is not metabolically
active
(Fig. 7.21 ).
Proton (hydrogen) MRS imaging characterizes the metabolite profiles of tumoral and nontumoral brain lesions. This biochemical information helps distinguish areas of tumor recurrence from areas of radiation necrosis. Major brain metabolites include choline (Cho), creatine (Cr), and n-acetylaspartate (NAA) (located at 3.2, 3.0, and 2.0 ppm, respectively). Choline reflects cellular density and proliferation, and is often elevated with tumor. Creatine is a normal cellular metabolite and is often stable in a variety of disease conditions. Thus creatine is often used as a denominator in calculating choline and NAA ratios (Cho/Cr and NAA/Cr), which
496
corrects for individual variation and allows for comparison between individual subjects. NAA is a neuronal marker and reflects neuronal density. Loss of the NAA signal is consistent with neuronal loss or damage, which can be seen in a wide variety of disease conditions, including radiation necrosis and even MS. Large vessels included within the radiation port may undergo radiation-induced endothelial hypertrophy, medial hyalinization, and fibrosis. The net result is a progressive vascular narrowing that may be obliterative in nature. This often involves the cavernous and supraclinoid portions of the carotid arteries in children who have undergone irradiation of the parasellar region for treatment of tumors, for example, craniopharyngiomas or optic and hypothalamic gliomas. The near complete obliteration of the supraclinoid carotid arteries results in cerebral and striatal ischemic changes. Occasionally, there may be a compensatory proliferation of lenticulostriate collaterals. When performing angiography, these collateral vessels present with a blush, which in Japan has been referred to as Moyamoya , meaning “puff of smoke.― Moyamoya disease classically refers to a supraclinoid obliterative arteriopathy that occurs primarily in children and is idiopathic in nature (Fig. 7.8 ). When methotrexate chemotherapy (intrathecal or systemic) is administered in combination with CNS radiation, P.201 P.202 these agents may have a synergistic effect in causing marked white matter abnormalities. It is theorized that low-dose radiation alters the blood–brain barrier, allowing increased penetration of methotrexate to neurotoxic levels. This has been noted most frequently in children being treated for leukemia, and two specific conditions have been described. The first has been called mineralizing microangiopathy , which is seen in up to one third of these children. This results in diffuse destructive changes to the brain characterized by symmetric corticomedullary junction and basal ganglia calcifications. There is also diffuse signal abnormality throughout the white matter. A more serious but less common
497
complication of combined radiation and methotrexate therapy is called necrotizing leukoencephalopathy. This process results in widespread damage to the white matter, consisting of demyelination, necrosis, and gliosis. MR reveals large, diffuse, confluent areas of white matter signal abnormality with cortical sparing. Clinically, these children may have symptoms ranging from slight reductions in cognitive function to progressive dementia, seizures, hemiplegia, and coma.
498
FIGURE 7.19. Acute Radiation Necrosis. A. Pretreatment axial proton density–weighted image in a 37-year-old woman with a deep temporoparietal arteriovascular malformation (arrows ). B . and
499
C . Less than 6 months after treatment with gamma knife radiation, the patient returned with marked vasogenic edema and contrast enhancement, consistent with radiation necrosis. Note that without clinical history, these imaging findings are indistinguishable from a neoplastic or infectious process. D . MR spectroscopy of the lesion reveals marked elevation of lactate and lipids (0.9 to 1.3 ppm), with reductions of all other major metabolites (choline, creatine, and Nacetylaspartate).
FIGURE 7.20. Radiation Necrosis. This 47-year-old man presented at 6 months (A) and 8 months (B) after resection and irradiation of a high right frontoparietal glioma. Coronal gadolinium-
500
enhanced T1WI reveals interval appearance (A) and progression (B) of a ring-enhancing mass lesion within the operative bed (arrows ). Despite this ominous appearance, this lesion revealed no radioisotope uptake on 18 F-2-fluoro-2-d-deoxyglucose PET. C . MR spectroscopy (MRS) of the lesion reveals marked elevation of lactate and lipids (0.9 to 1.3 ppm) with reduction in all other major metabolites (choline, creatine, and N-acetylaspartate). Both PET and MRS confirm the diagnosis of radiation necrosis. Serial MR scanning performed at 3-month intervals revealed a slowly regressing lesion that resolved by the 24-month follow-up study.
DYSMYELINATING
DISEASES
The disease processes that have been described up until this point are demyelinating, as they represent the destruction of normal myelin. In contrast, the dysmyelinating conditions, also referred to as leukodystrophies, are disorders in which myelin is abnormally formed or P.203 cannot be maintained in its normal state because of an inherited enzymatic or metabolic disorder. Although most of these conditions are not treatable, establishing a diagnosis is valuable in providing a prognosis and enables parental genetic counseling. These conditions are characterized by the progressive destruction of myelin owing to the accumulation of various catabolites, depending on the specific enzyme deficiency. Children often present clinically with progressive mental and motor deterioration. Radiographically, these diseases present with diffuse white matter lesions that are very similar to one another; however, some distinguishing features do exist (Table 7.2 ). The radiologist may play an important role in the diagnosis of these conditions, because astute interpretation of abnormal imaging findings may allow them to be the first physician to suggest the possibility of a metabolic disease. Factors that are helpful in differentiation between the leukodystrophies include the age of onset and the pattern of white matter involvement. Ultimately, serum biochemical and enzymatic analyses allow a specific diagnosis
501
to be made. Dysmyelinating diseases are rather uncommon, and we will focus on a few of the classic conditions. Metachromatic leukodystrophy Normal Infantile form: 1–2 Juvenile form: 5–7 Diffusely affected None Adrenoleukodystrophy Normal 5–10 Symmetric occipital and splenium of corpus callosum None Leigh
disease
Normal 5 mm below the foramen magnum). Although patients may be asymptomatic, alterations of CSF dynamics at the level of the foramen magnum are believed to give rise to cervical spinal cord syrinx in some patients (Fig. 8.30). For this reason, the
574
exact number of millimeters of downward extension of the cerebellar tonsils is less important than the degree of crowding of the foramen magnum. Chiari II malformations are associated with cord syrinx (Fig. 8.31) .
Cystic Lesions of the Posterior Fossa A simple working classification of cystic posterior fossa malformations is to consider them within the spectrum of DandyWalker malformations.
Dandy-Walker
Malformations
In distinction to Chiari malformations, Dandy-Walker malformations are characterized by a large posterior fossa with a high tentorial insertion. The posterior fossa is filled by a cystically P.233 dilated fourth ventricle that exerts mass effect (Fig. 8.32) . Hypoplasia or absence of the cerebellar vermis and cerebellar hemispheres are associated findings. Hydrocephalus is also common in this disorder, as is callosal hypogenesis. From this definition follow the less severe entities within the spectrum: Dandy-Walker variant and mega cisterna magna, both of which may have a normal position of the torcula.
575
FIGURE 8.29. Myelomeningocele. Sagittal T1WI of a myelomeningocele in a patient with Chiari II malformation. The myelomeningocele sac (arrow) contains neural tissue and meninges and protrudes through the dysraphic posterior elements of the lumbar spine.
576
FIGURE
8.30. Chiari I Malformation: Syrinx. Altered CSF flow
dynamics around the foramen magnum from tonsillar ectopia (arrowhead) in this Chiari I malformation patient have resulted in cervical cord syrinx (arrows) .
577
FIGURE 8.31. Chiari II Malformation: Syrinx. Sagittal T1WIs show features of Chiari II malformation with syrinx formation in the cervical cord (arrowheads). Note the more crowded appearance of the posterior fossa as compared with the Chiari I malformation shown in Fig. 8.30. The cisterna magna is obliterated, and the fourth ventricle (arrow) is effaced.
Dandy-Walker
variant shows a normal-sized posterior fossa and
hypoplasia or absence of the vermis and cerebellar hemispheres but no significant mass effect. Mega cisterna magna shows a normalsized posterior fossa and relatively normal cerebellar hemispheres and vermis. It is characterized by a prominent cisterna magna CSF space without mass effect. Mega cisterna magna exerts no mass effect, in distinction from retrocerebellar arachnoid cysts and epidermoid neoplasms. These mass lesions cause an inward convex bowing of brain tissue at their interface, and long-standing masses will cause smooth erosion of the inner table of the skull (see Chapter 5) .
578
THE
PHAKOMATOSES
The phakomatoses are a group of hereditary syndromes of the neuroectodermal system that are grouped together because they all share neurologic and cutaneous manifestations. Neurofibromatosis is divided into type 1 (NF-1, known as von Recklinghausen manifestations,
disease), with characteristic and neurofibromatosis
neurocutaneous P.234
type 2 (NF-2), which has very few, if any neurocutaneous lesions (Table 8.5). Note that NF-1 is one of the most common inherited CNS disorders, where as NF-2 is about one tenth as common.
TABLE 8.5 Neurofibromatosis: Type 1 (NF-1) Versus Type 2 (NF-2)
NF-1 (von Recklinghausen Syndrome)
N F- 2*
Epidemiology
Incidence
1 in 4,000
1 in 50,000
Age
Childhood
Young adult
17
2 2*
at
presentation
Affected chromosome
CNS findings
579
Vestibular schwannomas
No
Yes*
Meningiomas
No
Yes
Spinal
No
Yes
Yes
No
Yes
No
ectasia
Yes
No
Optic & other
Yes
Yes
dysplasia
Yes
No
Thinning long bone cortex (ribbon ribs)
Yes
No
Yes
No
Typically six or more
Rare
glial
tumors
Altered signal in white
matter
Altered signal in basal ganglia
Dural
gliomas
Skeletal
findings
Sphenoid
Other
findings
Plexiform neurofibromas
Café au lait spots
580
Iris (Lisch
hamartomas nodules)
Vascular
stenoses
Yes
No
Yes
No
*For NF-2, use the number 2 as your mnemonic: NF-2 patients typically have 2 (bilateral) acoustic schwannomas and an abnormal chromosome 22.
581
FIGURE 8.32. Dandy-Walker Malformation. malformation characteristically demonstrates a
A. Dandy-Walker cystic dilated
fourth ventricle and vermian agenesis, as demonstrated on this sagittal T1WI. The posterior fossa is enlarged, as seen by a high torcular insertion (arrow). B . Axial T1WI shows the cystically dilated fourth ventricle (arrow). C . This case demonstrates some of the features of Dandy-Walker malformation, but there is also asymmetric hypoplasia of the right cerebellar hemisphere.
582
Neurofibromatosis type 1 is characterized by skin lesions (café au lait spots), neurofibromas, and ocular findings. The findings in the brain consist of tumors and nonneoplastic lesions in the white matter and globus pallidus. NF-1 is associated with a high incidence of gliomas, with the optic pathways most commonly affected. A typical tumor causes fusiform enlargement of the optic nerve (Fig. 8.33) . However, the chiasm, optic tracts, and optic radiations can also become involved. Some optic pathway gliomas are intensely enhancing, but this is not a predictor of histologic grade, a topic covered in detail in Chapter 5. Parenchymal involvement along optic pathways is seen P.235 P.236 as hyperintense signal on T2WIs. Parenchymal gliomas may also occur in the cerebellum, brainstem, and cerebrum (Fig. 8.34) .
FIGURE
8.33. Neurofibromatosis:
583
Optic
Glioma.
A. Optic
nerve gliomas in a neurofibromatosis type I case are fusiform enlargements of the nerve on thin-section coronal T1WIs (arrows). B . Axial T2WI at the mid-orbit level shows the fusiform enlargement of the right optic nerve (arrow) .
FIGURE 8.34. Neurofibromatosis: Glioma. Axial T2WI shows an exophytic tectal glioma (arrow) in this patient with neurofibromatosis
type
I.
584
FIGURE
8.35. Neurofibromatosis:
Nonneoplastic
Lesions.
T2WI of patient with neurofibromatosis type I shows multifocal area of hyperintense signal in both middle cerebellar peduncles (arrowheads) .
585
FIGURE
8.36. Neurofibromatosis:
Nonneoplastic
Lesions.
A.
Hyperintense foci on T2WI are seen in the basal ganglia in this patient with neurofibromatosis type 1 (arrows). B . The lesions are seen to regress on a 4-year follow-up scan.
Hyperintense foci in the deep cerebral and cerebellar white matter are commonly seen on T2WIs. The etiology of these lesions is uncertain but is suggestive of benign myelin vacuolization. These lesions wax and wane when analyzed over serial scans, do not cause mass effect, and do not enhance (Figs. 8.35, 8.36). In general, these lesions tend to regress with increasing age. Lesion progression in a child more than 10 years old warrants close follow-up to rule out neoplastic transformation. Significant enlargement, new mass effect, and gadolinium enhancement may herald degeneration into gliomas. The globus pallidi may also exhibit abnormal hyperintense signal on both T1WIs and T2WIs. NF-1 has other manifestations. These include plexiform neurofibromas (ropelike masses of neural tissue in subcutaneous soft
586
tissues); vascular lesions (aneurysms, vascular ectasias, stenoses, moyamoya disease); spinal lesions (neurofibromas, meningoceles, scoliosis); and osseous lesions (sphenoid bone/lambdoid suture dysplasia,
pseudoarthrosis,
rib
abnormalities).
Neurofibromatosis type 2 differs considerably from NF-1. The classic feature of NF-2 is bilateral acoustic P.237 neuromas (the modern term for which is vestibular schwannoma; see Chapter 5). These patients often have meningiomas and schwannomas involving other cranial nerves, with cranial nerve V the second most common site for schwannomas in patients with NF-2 (Fig. 8.37). NF-2 also has spinal manifestations, which include meningiomas, ependymomas, and nerve sheath tumors (both schwannomas
and
neurofibromas).
FIGURE 8.37. Neurofibromatosis Type 2. Axial T1W post–gadolinium administration image in a patient with
587
neurofibromatosis type 2 shows bilateral enhancing acoustic schwannomas of the eighth cranial nerves (arrows) .
FIGURE 8.38. Tuberous Sclerosis. A. An axial slice from a CT scan in a patient with tuberous sclerosis shows a single calcified ependymal tuber (arrow). B . T2WI shows the same ependymal tuber seen on the CT scan. The tuber is seen as low-signalintensity nodule (arrow). The subcortical tubers are also well demonstrated
(arrowheads) .
Tuberous sclerosis is another distinctive neurocutaneous disorder. The skin lesions are adenoma sebaceum and ash-leaf spots. Brain lesions consist of subependymal hamartomas and cortical tubers (Fig. 8.38). Some of the P.238
588
subependymal nodules near the foramen of Monro can enlarge, cause mass effect, and invade brain tissue. Locally aggressive nodules are called subependymal “giant cell astrocytomas,― are typically located at the foramina of Monro, and may lead to hydrocephalus (Fig. 8.39) .
FIGURE 8.39. Tuberous Sclerosis. A. In another case of tuberous sclerosis, the sagittal T1WI demonstrates a giant cell astrocytoma arising in the region of the foramen of Monro (arrow). B . Axial T2WIs of a giant cell astrocytoma with concomitant hydrocephalus. The lesion (arrows) is heterogeneous signal
intensity.
589
FIGURE 8.40. Sturge-Weber Syndrome. Axial T2WI illustrates the dilated ependymal veins seen in Sturge-Weber syndrome. The veins (arrow) are serpentine flow voids bordering the ependymal surface of the ventricle.
Both cortical and subependymal lesions can undergo age-dependent calcification. Subependymal nodules represent hamartomas, and before calcification they tend to parallel white matter signal on MR images. Subependymal nodules are distinct from heterotopic gray matter in both their signal characteristics and tendency to calcify. Calcified nodules may be isointense or hyperintense on T1WIs. Enhancement of nodules at the foramen of Monro does not determine malignant transformation to a giant cell astrocytoma; rather, look for brain invasion to make this distinction. Cortical tubers are usually hypointense on T1WIs and hyperintense on T2WIs, and they may
590
calcify. Sturge-Weber Syndrome, or encephalotrigeminal angiomatosis, features angiomatous lesions of the skin and meninges. The facial lesion (a skin angioma called a port-wine nevus) appears in the ophthalmic division of the fifth cranial nerve. The pathologic entity seen in the brain is pial angiomatosis. These pial angiomas undergo age-dependent calcification and appear on CT scans as gyral cortical calcifications. The pial angiomatosis results in chronic ischemia of the gray matter, leading to gyral atrophy and underlying gliosis. Another sequela of pial angiomatosis is alteration of normal superficial cortical venous drainage with concomitant enlargement of deep and subependymal veins (Fig. 8.40). These dilated subependymal veins can mimic arteriovenous malformations. Gadolinium enhancement can reveal the full extent of pial angiomatosis and is helpful in cases where calcific atrophic changes have not yet occurred (Fig. 8.41). Young children may show subtle hypointensity of the underlying white matter on T2WIs without calcification of the cortex. Ipsilateral choroid plexus hypertrophy is another feature of this entity (Fig. 8.42). The gradient-recalled echo technique should be used on MR images to accentuate the presence of calcium (Fig. 8.43) .
591
FIGURE
8.41. Sturge-Weber
Syndrome. A young patient with
Sturge-Weber syndrome with before (left) and after (right) gadolinium T1WI demonstrating the full extent of pial angiomatosis (arrow). Gadolinium may be particularly useful in younger patients who do not yet demonstrate cortical calcifications.
FIGURE 8.42. Sturge-Weber Syndrome. Sagittal T1WI of a patient with Sturge-Weber syndrome shows ipsilateral choroid plexus hypertrophy (arrow) .
P.239 Von Hippel-Lindau Syndrome is an inherited disorder consisting of retinal angiomas and cerebellar and spinal hemangioblastomas. Hemangioblastomas are considered benign neoplasms (see Chapter 5), and the presence of multifocal spinal cord nodules near the piaarachnoid surface represents multicentric tumors, not drop metastasis.
592
Characteristic
features
of
cerebellar
hemangioblastomas
include
a
well-circumscribed cystic lesion with an enhancing mural nodule. Other appearances include solid tumors, solid masses with central cysts, and a cyst alone (Figs. 8.44, 8.45). Another helpful finding is a large blood vessel leading to the nodule. The small multifocal P.240 hemangioblastoma nodules are seen near the pial surface of the cerebellum or spinal cord.
FIGURE
8.43. Sturge-Weber
Syndrome.
Gradient-recalled
echo image accentuates the magnetic susceptibility artifact associated with the cortical calcifications (arrow) of SturgeWeber syndrome.
593
FIGURE
8.44. Von
Hippel-Lindau
Syndrome. Axial T1WIs in a
patient with von Hippel-Lindau syndrome. The left image is pre–gadolinium injection and the right image is post–gadolinium injection. Some of the lesions are cystic, and one of the enhancing foci (arrow) was seen to represent a mural nodule on a lower image.
594
FIGURE
8.45. Von
Hippel-Lindau
Syndrome. Von Hippel-
Lindau syndrome has a variable appearance, as demonstrated in this case of a solid enhancing mass in the fourth ventricle. A central speck of hypointensity represents a blood vessel (arrow) .
Although they are considered benign neoplasms, recurrence rates of up to 25% are reported. These vascular lesions are prone to sudden spontaneous hemorrhage. Gadolinium-enhanced MR imaging is the examination of choice for preoperative evaluation. Other associations with von Hippel-Lindau syndrome include renal cell carcinoma, along with angiomas of the liver and kidney.
Suggested
Readings
Barkovich AJ. Pediatric Neuroimaging. 4th ed. Philadelphia: Lippincott Williams & Wilkins, 2005.
595
Barkovich AJ. MR and CT evaluation of profound neonatal and infantile asphyxia. AJNR Am J Neuroradiol 1992;13:959–972. Barkovich AJ, Raybaud CJ. Malformations of cortical development. Neuroimaging Clin North Am 2004;41:401–423. Barkovich AJ, Hajnal BL, Vigneron D, et al. Prediction of neuromotor outcome in perinatal asphyxia: evaluation of MR scoring systems. AJNR Am J Neuroradiol 1998;19:143–149. Brodsky MC, Glasier CM. Optic nerve hypoplasia: clinical significance of associated central nervous system abnormalities on magnetic resonance imaging. Arch Ophthalmol 1993;111:66–74. Glass, RB, Fernbach SK, Norton KI, et al. The infant skull: a vault of
information.
Radiographics
2004;24:507–522.
Hunt RW, Neil JJ, Coleman LT, et al. Apparent diffusion coefficient in the posterior limb of the internal capsule predicts outcome after
perinatal
asphyxia.
Pediatrics
2004;114:999–1002.
Inder TE, Huppi PS, Warfield S, et al. Periventricular white matter injury in the premature neonate is followed by reduced cerebral cortex gray matter volume at term. Ann Neurol 1999;46:755–760. McLone DG, Dias MS. The Chiari II malformation: cause and impact. Childs Nerv Syst 2003;19:540–550. Ment LR, Bada HS, Barnes P, et al. Practice parameter: neuroimaging
of
the
neonate.
Neurology
2002;58:1726–1738.
Nelson MD Jr, Maher K, Gilles FH. A different approach to cysts of
596
the
posterior
fossa.
Pediatr
Radiol
2004;34:720–732.
Simon EM, et al. The middle interhemispheric variant of holoprosencephaly. AJNR Am J Neuroradiol 2002;23:151–156. Smirniotopoulos JG, Murphy FM. The phakomatoses. AJNR Am J Neuroradiol 1992;13:725–746. Truwit CL, Lempert TE. Pediatric Neuroimaging: A Casebook Approach. Piedmont, CA: DPS Press, 1991.
597
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section II - Neuroradiology > Chapter 9 Head and Neck Imaging
Chapter 9 Head and Neck Imaging Jerome A. Barakos Head and neck is a term used collectively to describe the extracranial structures, including the sinonasal cavity, skull base, pharynx, oral cavity, larynx, neck, orbit, and temporal bone. The head and neck region encompasses a tremendous spectrum of tissues in a compact space, with almost every organ system represented, including the digestive, respiratory, nervous, osseous, and vascular systems. Because of this anatomic complexity, the head and neck region is an area
approached
with
considerable
trepidation.
However,
accurate
assessment of this area can be accomplished by understanding both the normal anatomy and the scope of pathologic entities that may occur. We will begin our discussion by considering lesions of the paranasal sinuses and nasal cavity. This will be followed by a review of the skull base, the deep spaces of the neck, the lymph nodes, the orbits, and finally congenital head and neck lesions.
Imaging
Methods
Both multislice helical CT and MR can provide exquisite imaging of the normal and pathologic anatomy of the head and neck. Although each modality has advantages and disadvantages, the decision on whether to use CT versus MR for each individual case is often based on considering which technique the patient is more likely to tolerate. For example, if a patient has difficulty handling their oral secretions
598
because of prior head and neck surgery, particularly following tracheotomy or partial glossectomy, they may have significant hardship lying still for the time required for MR scanning. In such cases, the rapid imaging time of multislice helical CT is more likely to yield a study unmarred by motion artifact. Because calcification is better depicted with CT, this is the modality of choice when looking for obstructing salivary ductal calculi (sialoliths) or for the detection of fractures. In contrast, MR provides outstanding sensitivity for the discrimination of soft tissues and often better demonstrates the full extent of pathology. At the same time, the superior tissue contrast discrimination of MR allows for enhanced diagnostic specificity. The direct multiplanar capability of MR may also provide for improved evaluation of pathologic entities. For example, because of the horizontal orientation of the palate, floor of the mouth, and skull base, sagittal and coronal imaging are invaluable in optimally assessing these areas.
Positron
Emission
Tomography
(PET)
The advent of PET imaging has had a profound effect on the evaluation of head and neck malignancies. In combination with either MR or CT imaging, PET has greatly increased the sensitivity and specificity in the evaluation of primary as well as recurrent malignancies. PET is a functional imaging modality based upon the distribution of a glucose analogue radioisotope (18F fluorodeoxyglucose). Pathologic conditions that have an affinity for glucose will take up this P.242 isotope at a greater rate than normal surrounding tissues and thus be identifiable as areas of abnormality (Fig. 9.1) .
599
FIGURE 9.1. (Color Plates) Positron Emission Tomography (PET). A . Axial T1WI of a 77-year-old man who presented with a large right hypopharyngeal malignancy (solid arrow) and extensive ipsilateral necrotic adenopathy (not shown). As expected, histology reflected a squamous cell carcinoma, the most common malignancy arising from the head and neck mucosal surfaces. PET (B) and fused MR/PET (C) scans confirm the abnormal radioisotope accumulation within the hypopharyngeal malignancy (solid arrows). However, the PET study demonstrates an abnormality of the right sternocleidomastoid muscle that was not suspected on MR imaging (open arrows). This proved to be a diffuse infiltrative metastatic focus of the right sternocleidomastoid muscle. This case demonstrates the value and increased specificity of combining functional/physiologic imaging (PET) with morphologic imaging (MR).
Lesions found on PET scan are characterized by a standardized uptake value (SUV). The SUV refers to the relative radioactivity of a particular lesion when standardized to the injection dose and adjusted for body weight. As a result, the SUV is an absolute value that can be compared from patient to patient and exam to exam. In general, an SUV of greater than 3.0 is considered pathologic, but there are many caveats. A wide variety of nonmalignant conditions may give rise to an elevated SUV, most notably infection and
600
postoperative changes. Additionally, some neoplasms have poor glucose affinity, resulting in a low SUV. PET alone may be highly sensitive, but it is not very specific. The true benefit of PET is realized when its physiologic/functional information is combined with the high-spatial-resolution morphologic information of CT and MR. In summary, combining PET findings with CT and/or MR results in a marked increase in specificity, making this combination a powerful diagnostic tool.
PARANASAL
SINUSES
AND
NASAL
CAVITY Sinusitis Inflammatory disease is the most common pathology involving the paranasal sinuses and nasal cavity. Mild mucosal thickening, primarily within the maxillary and ethmoid sinuses, is common, even in asymptomatic individuals. In contrast, acute sinusitis is characterized by the presence of air–fluid levels or foamyappearing sinus secretions and is typically caused by a viral upper respiratory tract infection (Fig. 9.2). In chronic sinusitis, changes include mucoperiosteal thickening as well as osseous thickening of the sinus walls. Soft tissue findings suggestive of sinusitis are best detected on T2WIs, as they are often high in signal. An exception is chronic sinus secretions that have become so desiccated that they yield no signal and may mimic an aerated sinus. These sinus concretions and the bony wall thickening associated with chronic sinusitis are most easily appreciated on CT. Endoscopic sinonasal surgery, used for the evaluation and treatment of inflammatory sinonasal disease, is being performed with increasing frequency. Direct coronal sinus CT provides exquisite definition of sinonasal anatomy and provides pre-endoscopic sinus assessment (Fig. 9.3). Knowledge of the anatomy of the lateral wall of the nasal cavity and routes of mucociliary drainage of the paranasal sinuses is critical to understanding patterns of inflammatory sinonasal disease. A major area of mucociliary drainage is the middle meatus, known as
601
the ostiomeatal unit. It is important to note that disease limited to the infundibulum of the maxillary ostium will result in isolated obstruction of the maxillary sinus. In contrast, a lesion located in the hiatus semilunaris (middle meatus) results in combined obstruction of the ipsilateral maxillary sinus, anterior and middle ethmoid air cells, and the frontal sinus. This combined pattern of sinonasal disease has been described as the “ostiomeatal pattern― of obstruction. This pattern is significant because it indicates that one's P.243 attention should be directed to identifying the offending lesion within the hiatus semilunaris, rather than simply describing the presence of diffuse
sinus
FIGURE
disease.
9.2. Acute Sinusitis With Cavernous Sinus
602
Thrombosis. A. Axial T2WI of a 27-year-old man who presented with a rapidly progressive sinusitis. The right sphenoid and ethmoid sinuses are opacified, with an air–fluid level within the left sphenoid sinus (arrow). Sphenoid sinusitis is of great clinical concern as it may easily extend intracranially owing to the presence of valveless veins. Coronal T1WI (B) and axial fatsaturated T1WI with gadolinium contrast (C). The patient's clinical condition deteriorated rapidly as the infection extended into the cavernous sinus, with resultant cavernous sinus and left superior ophthalmic vein thrombosis. The thrombosis is characterized by the marked sinuses (arrows, B), while frank thrombus left superior ophthalmic vein (arrows, C).
cavernous sinus enlargement of the is visualized within the Differential diagnostic
conditions would include a carotid-cavernous fistula and TolosaHunt syndrome (an idiopathic nongranulomatous inflammatory condition of the cavernous sinus). Note the parenchymal abscess forming along the right middle cerebral artery cistern (black arrows) .
Several common complications are associated with sinusitis, including inflammatory polyps, mucous retention cysts, mucoceles, and most importantly cavernous sinus thrombosis.
Inflammatory
Polyps
Chronic inflammation leads to mucosal hyperplasia, which results in mucosal redundancy and polyp formation. Most often these polyps blend imperceptibly with the mucoperiosteal thickening and cannot be clearly differentiated. When an antral polyp expands to the point where it prolapses through the sinus ostium, it is referred to as an antrochoanal polyp. Although these polyps may not be associated with chronic sinusitis, they are similar to inflammatory polyps in that they represent areas of reactive mucosal thickening. Their characteristic appearance is that of a soft tissue mass extending from the maxillary sinus to fill the ipsilateral nasal cavity and nasopharynx. Often, the ostium of the maxillary sinus will be
603
enlarged secondary to the mass effect of the polyp. The importance in recognizing such a lesion is that if it is surgically snared as if it were a nasal polyp, without regard for its antral stalk, it will recur. Mucous retention cysts simply represent obstructed mucous glands within the mucosal lining. These lesions have a characteristic rounded appearance, measuring one to several centimeters in diameter, with the maxillary sinus being most commonly involved. These lesions are commonly recognized in asymptomatic individuals. Mucocele is similar to a retention cyst, but instead of a single mucous gland becoming obstructed, the entire sinus is obstructed. This typically occurs because of a mass obstructing the draining sinus ostium. The characteristic feature of a mucocele is frank expansion of the sinus with associated sinus wall bony thinning and remodeling. The frontal sinus is the sinus most commonly affected, but any sinus may be involved (Fig. 9.4). If the mucocele becomes P.244 infected, it demonstrates peripheral enhancement and is referred to as a mucopyocele.
604
FIGURE
9.3. Ostiomeatal
Unit
(OMU). Line drawing in coronal
plane demonstrates the anatomy of the OMU. Lines with arrows show the normal route of mucociliary clearance. Infundibular (dashed line) and OMU (solid line) patterns of obstruction are shown. Coronal CT far surpasses plain sinus films in evaluating problems of the OMU for potential relief through endoscopic surgery. B, ethmoid bulla; M, maxillary sinus; u, uncinate process; mt, middle turbinate; mm, middle meatus; im, inferior meatus; it, inferior turbinate; S, nasal septum. (Reprinted with permission. From Babbel RW, Harnsberger HR, Sonkens J, Hunt S. Recurring patterns of inflammatory sinonasal disease demonstrated on screening sinus CT. AJNR Am J Neuroradiol 1992;13:903–912.)
Inverting
Papilloma
605
A variety of papillomas occur within the nasal cavity, but most attention has focused on the inverting papilloma. These papillomas are named based on their histologic appearance. In this condition, the neoplastic nasal epithelium inverts and grows into the underlying mucosa. These papillomas are not believed to be associated with allergy or chronic infection because they are almost invariably unilateral
in
location.
Inverting P.245
papillomas occur exclusively on the lateral nasal wall, centered on the hiatus semilunaris. Because of their increased association with squamous cell carcinoma, it is recommended that these lesions be surgically resected with wide mucosal margins.
FIGURE 9.4. Sinus Mucocele. A. Coronal T1WI. B . Axial T2WI. Patient presented with proptosis, resulting from mass effect from an ethmoid sinus mucocele (arrows). A mucocele results from chronic obstruction of a paranasal sinus that becomes blocked and converted into a fluid-filled cyst. Over time this lesion may expand, eroding bone and resulting in proptosis. Differential diagnostic considerations would include a dermoid cyst, which would be characterized by the presence of fat (see Fig. 9.36) .
Juvenile nasopharyngeal angiofibromas are typically seen in male adolescents presenting with epistaxis. The tumor arises from
606
fibrovascular stroma of the nasal wall adjacent to the sphenopalatine foramen. This is a benign tumor that can be very locally aggressive. In an adolescent male presenting with a nasal mass and epistaxis, it is important to have a high clinical suspicion for this lesion, because life-threatening hemorrhage may result if a biopsy or limited resection is attempted. The tumor characteristically fills the nasopharynx and bows the posterior wall of the maxillary sinus forward. In fact, the retromaxillary pterygopalatine fossa location is a hallmark feature that should elicit this diagnosis for consideration. Juvenile nasopharyngeal angiofibromas enhance markedly with contrast administration, differentiating them from the rarer lymphangioma. Preoperatively, interventional radiology may play a role in embolization of these lesions, making them less vascular and facilitating
surgical
resection.
Malignancies The tissues within the paranasal sinuses and nasal cavity that give rise to malignancies include squamous epithelium, lymphoid tissue, and minor salivary glands. The corresponding malignancies are therefore squamous cell carcinoma, lymphoma, and minor salivary tumors. Because the entire upper aerodigestive tract is lined with squamous epithelium, it follows that squamous cell carcinoma is the most common malignancy (80% to 90%) of not only the paranasal sinuses and nasal cavity, but of the entire head and neck. Squamous cell carcinoma of the sinuses is often clinically silent until it is quite advanced. Early symptoms are usually related to obstructive sinusitis. Imaging findings consist of an opacified sinus with associated bony wall destruction. These findings are nonspecific and do not allow differentiation from non-Hodgkin lymphoma or a minor salivary gland malignancy. The presence of constitutional symptoms with prominent head and neck or systemic adenopathy suggests lymphoma, particularly in a child or young adult. Minor salivary glands are dispersed throughout the upper aerodigestive tract but are most highly concentrated in the palate. Any of these minor salivary glands found throughout the head and neck may give rise to salivary neoplasms. In contrast to parotid
607
gland salivary neoplasms, the majority of which are benign, most minor salivary neoplasms are malignant. The most common salivary malignancies include adenoid cystic carcinoma, adenocarcinoma, and mucoepidermoid
carcinoma.
An esthesioneuroblastoma is an additional malignancy that should be mentioned when describing lesions of the nasal cavity. The esthesioneuroblastoma is a tumor that arises from the neurosensory receptor cells of the olfactory nerve and mucosa. Thus, this lesion may originate anywhere from the cribriform plate to the turbinates. This tumor is often quite destructive by the time of diagnosis and is found high within the nasal vault (Fig. 9.5). Involvement of the cribriform plate with extension into the anterior cranial fossa is not uncommon with esthesioneuroblastoma and should suggest this lesion.
FIGURE
9.5. Esthesioneuroblastoma.
608
Coronal
fat-suppressed
postgadolinium T1WI. A large destructive mass (M) in the nasal cavity extends through the cribriform plate into the anterior cranial fossa (arrows). This degree of frank bony destruction is unusual for squamous cell carcinoma and lymphoma, but characteristic for esthesioneuroblastoma.
In assessing the size and extent of sinonasal cavity pathology, it is often difficult to differentiate the offending lesion from associated obstructed sinus secretions. In such instances, heavily T2W sequences are of value, because in general, sinus secretions will be brighter than the malignancy, which is often isointense with respect to muscle.
SKULL
BASE
The skull base extends from the nose anteriorly to the occipital protuberance posteriorly and is composed of five bones: the ethmoid, sphenoid, occipital, temporal, and frontal bones. The skull base contains many foramina through which both vessels and nerves pass. Because
the P.246
skull base has an undulating surface with a horizontal orientation, coronal or sagittal images are valuable in its evaluation.
Tumors of the Skull Base Tumors may arise that are intrinsic to the skull base. Additionally, an extrinsic lesion may extend to involve the skull base either from above or below. Any lesion from the paranasal sinuses and nasal cavity already described may extend to involve the skull base. Other lesions that may extend to involve the skull base include paragangliomas, neural sheath tumors (schwannoma and neurofibroma), and meningiomas. Although various primary malignant neoplasms of the skull base are described in the following, most malignant lesions of the skull base are metastatic in origin. Primary
malignant
neoplasms are relatively uncommon,
609
comprising only about 2% to 3% of skull base tumors. The three most common primary malignant tumors are chordoma, chondrosarcoma, and osteogenic sarcoma. Chordoma is a bone neoplasm that arises from remnants of the primitive notochord. Classically, this lesion will present as a destructive midline mass centered in the clivus. These tumors may be found anywhere along the craniospinal axis; typically 35% of lesions involve the clivus, 50% the sacrum, and 15% the vertebral bodies. Radiographically, this lesion is characterized as a midline destructive bony lesion with predilection for the sphenooccipital synchondrosis. On a sagittal image, the sphenooccipital synchondrosis is occasionally seen as a horizontal line in the midclivus, midway between sella and basion (tip of clivus). Chondrosarcomas are malignant tumors that develop from cartilage. Because the skull base is preformed in cartilage, there is a predilection for chondrosarcoma to involve the skull base. A preferred site of origin is parasellar in location, at the petroclival junction. Osteogenic
sarcoma is typically the result of prior radiation
therapy or malignant transformation of Paget disease. Although a central destructive clival lesion is characteristic for chordoma and a paraclival destructive bony lesion is suggestive of chondrosarcoma, our differential diagnostic list includes several other bony lesions. The skull base, like any bone, may be affected by metastases, myeloma, plasmacytoma, fibrous dysplasia, and Paget disease. As with any bony lesion, CT helps to differentiate among these diagnostic possibilities. For example, fibrous dysplasia will reveal a smooth, ground-glass appearance on CT, while Paget disease will demonstrate trabecular coarsening, and neither of these conditions will reveal bony destruction. Lesions of the jugular foramen are most commonly paragangliomas and are discussed under the heading “Carotid Space.― These patients commonly present with pulsatile tinnitus and a conductive hearing loss. These tumors are best initially evaluated with CT. If extension into the jugular fossa is identified, MR is valuable in defining the full extent of the lesion. CT often demonstrates “moth-eaten― destruction of the bone surrounding the jugular fossa, with MR revealing the typical heterogeneous “salt and
610
pepper― signal related to numerous flow voids. Malignant tumors are often indistinguishable from paragangliomas on CT, but most fail to demonstrate flow voids on MR. Other lesions of the jugular fossa include schwannomas (arising from cranial nerves IX to XI) and meningiomas. These lesions cause a smooth expansion of the jugular foramen with marked enhancement. Additionally, schwannomas may demonstrate
cystic
Temporal
components.
Bone
Although a thorough discussion of the temporal bone is beyond the scope of this chapter, we will focus on some highlights. The most common diseases involving the temporal bone are inflammatory in nature and include cholesteatomas. Eustachian tube dysfunction with resultant decreased intratympanic pressure is believed to be the principal defect responsible for inflammatory disease of the middle ear and mastoid. Cholesteatoma is an epidermoid cyst composed of desquamating stratified squamous epithelium. These cysts enlarge because of the progressive accumulation of epithelial debris within their lumen. They may be either congenital (2%) or acquired (98%). Congenital cholesteatomas originate from epithelial rests within or adjacent to the temporal bone. Acquired cholesteatomas originate from the stratified squamous epithelium of the tympanic membrane. These begin as localized tympanic membrane retraction pockets. The diagnosis of a cholesteatoma is based on the detection of a soft tissue mass within the middle ear cavity, typically with associated bony erosion. The superior portion of the tympanic membrane (pars flaccida) retracts easily and is the most common site for formation of an acquired cholesteatoma. Cholesteatomas arising in this area originate within the Prussak space (superior recess of the tympanic membrane), which is located medial to the pars flaccida between the scutum and the neck of the malleus. Thus, a finding of soft tissue in this region with subtle erosion of the scutum and medial displacement of the ossicles is characteristic of a cholesteatoma. Note that when fluid or inflammatory pathology is present, such as with otitis media, these changes cannot be differentiated from
611
cholesteatoma because they have similar densities. Although most cholesteatomas can be easily diagnosed otoscopically, the clinician cannot judge the size and full extent of the lesion. As a result, CT plays an important role in determining the size of the lesion, as well as the status of the ossicles, the labyrinth, the tegmen, and the facial nerve. MR has a limited role in the evaluation of erosive lesions of the temporal bone, because the lack of landmarks does not allow localization of the process and it P.247 gives no information concerning the status of the ossicles and other bony structures. Cholesterol granuloma, also known as giant cholesterol cyst, is a type of granulation tissue that may involve the petrous apex. These lesions represent petrous apex air cells that have become partially obstructed and are filled with cholesterol debris and hemorrhagic fluid. Because of their hemorrhagic components, these lesions are characterized by high signal on both T1WIs and T2WIs. When faced with an opacified petrous apex, differential diagnostic considerations include: retained fluid secretions (parallels signal intensity of fluid, dark T1, bright T2, and no enhancement); petrous apicitis (parallels signal intensity of abscess, dark T1, bright T2, and ring enhancement); and nonaerated petrous apex (parallels signal intensity of fatty bone marrow, bright T1, dark T2, and no enhancement).
SUPRAHYOID
HEAD
AND
NECK
When a patient presents with a head and neck mass, the age of presentation is an important consideration when establishing a differential diagnostic list. In the pediatric age group, the majority of lesions (>90%) will be benign and consist of a variety of congenital or inflammatory entities (see “Congenital Lesions―). If a malignancy is encountered, it will most likely be a lymphoma (e.g., Burkitt lymphoma if rapid growth is noted) or rhabdomyosarcoma. In sharp contrast, when an adult presents with a head and neck mass (excluding thyroid lesions), the vast majority of lesions (>90%) will
612
be malignant (Fig. 9.6). In the younger adult (20 to 40 years), the most common malignancy will be lymphoma (Fig. 9.7), and in adults older than 40 years, the most common neck mass will be nodal metastases. The suprahyoid head and neck is traditionally divided into compartments that include the nasopharynx, oropharynx, and oral cavity. An understanding of the division between these spaces is essential to accurately determine and describe the full extent of mucosal lesions. The term nasopharynx is frequently misused as a nonspecific term to describe any area in the upper aerodigestive tract. In fact, the nasopharynx refers to a very specific portion of the pharynx. The nasopharynx lies above the oropharynx and is divided from the oropharynx by a horizontal line drawn along the hard and soft palates. Posteriorly the nasopharynx is bounded by the pharyngeal constrictor muscles, and anteriorly it is bounded by the nasal cavity at the nasal choana (paired funnel-shaped openings between the nasal cavity and the nasopharynx). Below the hard palate lie the oral cavity and oropharynx. These two areas are divided by a ring of structures that includes the circumvallate papillae (located along the P.248 posterior aspect of the tongue), the tonsillar pillars, and the soft palate.
613
FIGURE
9.6. Squamous Cell Carcinoma: Cystic Nodal
Metastasis.
A. Sagittal T1WI. B . Axial postgadolinium fat-
suppressed T1WI. This 43-year-old patient was referred for a “branchial cleft cyst.― Patient had a 6-month history of a right-sided neck mass that would swell during upper respiratory tract infections. Images reveal a multiseptated cystic lesion (c) in the right jugular nodal chain. On biopsy, this proved to be a squamous cell, cystic nodal metastasis. Although this lesion may appear similar to a branchial cleft cyst, the presence of multiple additional nodes (n) is unusual. A branchial cleft cyst may exhibit a thickened wall with septations, depending on current or previous infections. S, submandibular glands. Note that the jugulodigastric node is easily identified by its characteristic location; situated immediately posterior to the submandibular gland.
614
FIGURE
9.7. (Color Plates) Non-Hodgkin Lymphoma of the
Masticator
Space. A . Axial T1WI. B . Corresponding fused
MR/PET scan. A 21-year-old man presented with a diffusely infiltrating mass of the right masticator space (arrows). Although MR reveals a diffusely infiltrating lesion, PET demonstrates that only a focal portion of the lesions is glucose avid (arrow). This helped direct the biopsy to the most metabolically active portion of the lesion, increasing the likelihood that a pathologic diagnosis could be made. Much of the remaining infiltrative change of the masticator space proved to be reactive in nature, without frank tumor. This case highlights the improved specificity that results from combining morphologic imaging (MR/CT) with physiologic (PET) imaging.
These
traditional
compartments
(nasopharynx,
oropharynx,
and
oral
cavity) are important for describing the spread of superficial, mucosa-based lesions. In contrast to this division, multiple facial planes divide the deep head and neck into spaces that form true compartments. It is important to realize that these deep spaces are unrelated to the traditional division of the head and neck and traverse the neck without regard to the traditional divisions. Therefore, when describing deep head and neck lesions, the
615
traditional pharyngeal subdivisions are of limited value. Most radiologists have adapted a spatial approach to the head and neck, described as follows and popularized by Harnsberger. The deep anatomy of the head and neck is subdivided by layers of the deep cervical fascia into the following spaces: (1) superficial mucosal, (2) parapharyngeal, (3) carotid, (4) parotid, (5) masticator, (6) retropharyngeal, and (7) prevertebral. When evaluating a patient with pathology in the deep head and neck, it is important to determine within which space the pathology lies. Because only a limited number of structures are located within each compartment, these are the structures from which pathology will arise. Therefore, only specific pathology will be found within these separate fascial spaces, markedly limiting the differential diagnosis. For example, the principal structures within the parotid space are the parotid gland and parotid lymph nodes. Consequently, if a parotid space mass is identified, the diagnosis is primarily limited to either a parotid tumor or nodal disease. Each of these seven spaces will be reviewed in detail (Table 9.1). Note that although this spatial division is popular with radiologists, surgeons and otolaryngologists occasionally use different terms, e.g., “retrostyloid space― instead of “carotid
space.―
Superficial
Mucosal
Space
The superficial mucosal space includes all structures on the airway side of the pharyngobasilar fascia. The principal constituent of this space is the mucosa of the upper aerodigestive tract, which consists of squamous epithelium, submucosal lymphatics, and hundreds of minor salivary glands. The pharyngobasilar fascia represents the superior aponeurosis of the superior pharyngeal constrictor muscle, which inserts into the skull base. This tough fascia separates the mucosal space from the surrounding parapharyngeal space. Lesions originating within the superficial mucosal space may invade deep to the mucosal surface, resulting first in lateral displacement and then obliteration of the parapharyngeal space. However, P.249
616
many early lesions that begin within the mucosal space present as only mild mucosal irregularities or asymmetries (Fig. 9.8). This space is easily evaluated by the clinician and thus the radiologist should have a low threshold for suggesting the presence of abnormalities within this space. In children, there is frequently prominent adenoidal tissue that fills the nasopharynx. Even in adults, following an upper respiratory infection, prominent mucosal tissue may be noted; this is of little concern as long as there is no invasion of deep facial places and no associated adenopathy (Fig. 9.9) .
TABLE 9.1 Deep Compartments of the Head and Neck
Compartment Mucosal
Contents
Pathology
Squamous
Nasopharyngeal
mucosa Lymphoid tissue (adenoids, lingual
carcinoma Squamous cell carcinoma
tonsils) Minor salivary glands
Lymphoma Minor salivary gland tumors Juvenile angiofibroma Rhabdomyosarcoma
Parapharyngeal
Parotid
Fat Trigeminal (V3)
nerve
Minor salivary gland tumor Lipoma
Internal maxillary artery Ascending pharyngeal artery
Cellulitis/abscess Schwannoma
Parotid
Salivary
gland
617
gland
Intraparotid lymph nodes Facial nerve (VII)
tumors Metastatic adenopathy
External carotid artery Retromandibular
Lymphoma Parotid cysts
vein
Carotid
Cranial nerves IX–XII Sympathetic
Schwannoma Neurofibroma Paraganglionoma
nerves Jugular
chain
Metastatic adenopathy
artery vein
Lymphoma Cellulitis/abscess Meningioma
nodes Carotid Jugular
Masticator
Muscles of
Odontogenic
mastication
abscess
Ramus and body of mandible
Osteomyelitis Direct spread of
Inferior
squamous
alveolar
nerve
cell
carcinoma Lymphoma Minor salivary tumor Sarcoma of muscle or bone
Retropharyngeal
Lymph
nodes
Metastatic
(lateral and medial retropharyngeal)
adenopathy
Fat
Lymphoma
618
Prevertebral
Cervical vertebrae Prevertebral
Abscess/cellulitis Osseous metastases Chordoma
muscles Paraspinal muscles
Osteomyelitis Cellulitis Abscess
Phrenic
nerve
For further discussion see Harnsberger, et al. Diagnostic Imaging: Head and Neck. Salt Lake City, UT: AMIRSYS, 2004.
Benign
Lesions
The most common benign lesions arising in the mucosal space are Tornwaldt cysts and lesions related to minor salivary gland tissue. Tornwaldt cysts are found in the midline and have high intensity on T2WIs (Fig. 9.10). They are believed to be remnants of notochordal tissue aberrantly located in the nasopharynx and have an incidence of approximately 1% to 2% in normal patients. Lesions arising from minor salivary glands include retention cysts and benign neoplasms. Retention cysts represent obstructed glands similar to those found within the paranasal sinuses. The most common benign neoplasm is the benign mixed-cell tumor (pleomorphic adenoma). Both of these lesions present as well-circumscribed, rounded lesions that have high signal intensity on T2WIs.
619
FIGURE
9.8. Squamous
Cell
Carcinoma. Axial postgadolinium
fat-suppressed T1WI through the level of the nasopharynx. Contrast-enhancing soft tissue fills the right fossa of Rosenmüller (arrows). Although this lesion does not obviously invade the underlying parapharyngeal tissues, submandibular nodal metastases were present. This example underscores the point that even mild asymmetries of the mucosal space may represent a malignancy, and careful correlation with physical examination should be suggested by the radiologist.
P.250
Malignant
Lesions
The most common malignant neoplasms of the mucosal space are
620
squamous cell carcinoma, non-Hodgkin lymphoma, and minor salivary gland malignancies; of these, squamous cell carcinoma is by far the most common. Unfortunately, these malignancies all appear similar on CT and MR. Initially there is mass effect, often associated with lateral compression or obliteration of the parapharyngeal space, followed by invasion of the skull base. An early triad of radiographic findings consists of (1) superficial nasopharyngeal mucosal asymmetry, (2) ipsilateral retropharyngeal adenopathy, and (3) mastoid opacification. Mastoid opacification is an important early warning sign (Fig. 9.11). Mastoid opacification is easily detected on T2WIs and suggests potential dysfunction of the eustachian tube, frequently the result of tumor infiltration of the tensor veli palatini muscles. This finding directs the radiologist to carefully evaluate the mucosa of the nasopharynx. Note that both the nasopharynx and the mastoid air cells are included on every head CT and MR scan, and these areas should not be overlooked on routine head imaging. Fat-suppressed, fast spin-echo (FSE) T2, and contrast-enhanced imaging are useful in detecting and defining the extent of pathology. Additionally, these sequences allow the detection of subtle perineural spread of neoplasms, particularly along cranial nerves extending into the skull base. This is particularly important with adenoid cystic carcinoma, which has a marked propensity for perineural spread and is the most common minor salivary gland malignancy (Fig. 9.12) . Squamous
cell
carcinoma is the most common malignancy of the
upper aerodigestive tract. However, a particular variant of squamous cell carcinoma occurs within the nasopharynx and is termed nasopharyngeal carcinoma. Nasopharyngeal carcinoma has several unique histologic features that distinguish it from squamous cell carcinoma. Although squamous cell carcinoma is common in the Caucasian population, nasopharyngeal carcinoma is not, with an incidence of about 1 in 100,000 people per year. This is in contrast to rates that are 20 times higher in Asia, particularly in southern regions of China. Although smoking and alcohol abuse are often associated with squamous cell carcinoma, they have no causal association with nasopharyngeal carcinoma. However, both environmental and genetic factors do appear to play a role in the
621
genesis of nasopharyngeal carcinoma. Specifically, immunoglobulin-A antibodies to the Epstein-Barr virus have been associated with nasopharyngeal carcinoma. Lymphoma involving the mucosa cannot be differentiated by imaging from squamous cell or minor salivary gland carcinoma. However, non-Hodgkin lymphoma frequently has systemic manifestations, with extranodal and extralymphatic sites of involvement that are atypical for these other malignancies. Thus, the presence of a mucosal mass in association with bulky supraclavicular and mediastinal adenopathy as well as splenomegaly would be suggestive of lymphoma.
Parapharyngeal
Space
The parapharyngeal space is a triangular, fat-filled compartment that extends from the skull base to the submandibular gland region. It is located at the center of the surrounding spaces and is compressed or infiltrated in a characteristic fashion by masses originating from the various spaces. The primary importance of the parapharyngeal space is that it serves as an important landmark of mass effect in the deep face. When a lesion occurs in any of the four surrounding spaces, there will be characteristic impressions on the parapharyngeal fat space, which will suggest the space of tumor origin. The parapharyngeal space is surrounded by the carotid space posteriorly, the parotid space laterally, the masticator space anteriorly, and the superficial mucosal space medially. Therefore, the parapharyngeal space will be compressed on its medial surface by masses originating P.251 P.252 from the mucosal surface, displaced anteriorly by carotid sheath masses, displaced medially by parotid masses, and displaced posteriorly and medially by masses within the masticator space. Thus, by assessing the location and displacement pattern of the parapharyngeal space, one can assign a space of origin to a deep facial mass (Fig. 9.13) .
622
FIGURE
9.9. Adenoidal
Hypertrophy. Axial first-echo and
second-echo T2 images in a 5-year-old child. Prominent adenoidal tissue (open arrows) fills the nasopharynx, expanding the fossa of Rosenmüller bilaterally. Additionally, lateral pharyngeal nodes (arrowheads) are clearly visualized. These findings are typical for a child. The age of the patient and the symmetry are in keeping with normal findings.
623
FIGURE 9.10. Tornwaldt Cyst. Axial first-echo and secondecho T2WIs. A high-signal-intensity lesion appears in the superficial mucosa (arrows). This midline location is characteristic of a Tornwaldt cyst, a remnant of the primitive notochord, and is found in 1% to 2% of the normal population.
624
FIGURE
9.11. Nasopharyngeal
Malignancy. Axial
postgadolinium T1WI. The triad of nasopharyngeal malignancy consists of (1) mucosal mass (double white arrows) of the lateral nasopharynx (fossa of Rosenmüller), (2) lateral retropharyngeal nodes (arrowheads), and (3) mastoid opacification (white arrow). Mastoid opacification is the result of dysfunction of the eustachian tube, and should always prompt search for the offending nasopharyngeal mass.
Carotid
Space
Masses of the carotid space deviate the parapharyngeal space anteriorly and will separate or anteriorly displace the carotid and jugular vein. They sometimes displace the styloid process anteriorly,
625
which narrows the stylomandibular notch (the space between the styloid process and the mandible). This is a characteristic feature that distinguishes these lesions from deep parotid space lesions, which widen the stylomandibular notch.
Pseudomasses When evaluating carotid space tumors, there are several pseudomasses of the carotid space that must be taken into account. These pseudomasses are vascular variants that may be mistaken for masses both clinically and radiographically. Asymmetry of the internal jugular veins is the most common variation in the vascular anatomy of the neck. Marked asymmetry between the size of the left and right jugular veins is common, with the right vein typically being the larger of the two. Additionally, the jugular veins may demonstrate considerable variability in the degree of signal within their lumina, ranging from bright to signal void. The intraluminal bright regions should not be mistaken for thrombosis. It is important to follow the signal on serial images to visualize the tubular nature, thus confirming that the signal represents vasculature; otherwise it may easily be mistaken for adenopathy. Tortuosity of the carotid artery may present as a submucosal pulsatile mass in the pharynx. This variation, which is frequently seen in the elderly, is easily detected on CT or MR and obviates the need for further diagnostic workup unless a posttraumatic aneurysm is suspected.
Tumors Most carotid space masses are benign neoplasms that arise from nerves located within the carotid sheath. The most common lesions are paragangliomas (also called chemodectomas) and nerve sheath tumors such as schwannomas and neurofibromas. Paragangliomas are vascular tumors that arise from neural crest cell derivatives. These lesions are named according to the nerves from which they arise and their location of origin. When arising from the carotid body, at the carotid bifurcation, paragangliomas are called carotid body tumors (Fig. 9.14). Paragangliomas may also arise from the ganglion
626
of the vagus nerve (glomus vagale tumors), along the jugular ganglion of the vagus nerve (glomus jugulare tumors), and around the Arnold and Jacobson nerves in the middle ear (glomus tympanicum tumors). Despite the use of different names, the imaging features and histology remain the same. Clinically, patients with paragangliomas present with a painless, slowly progressive neck mass that may be pulsatile with an associated bruit. Because these lesions are located within the carotid sheath, there are often associated slowly progressive cranial neuropathies (cranial nerves IX to XII) (Fig. 9.15). Paragangliomas are often multiple (5% to 10%) and, in familial cases, are multiple 25% to 33% of the time. Therefore, if a lesion is detected, it is essential to look for others. Angiographically, paragangliomas are very vascular, with a strong blush in the capillary phase. Treatment often consists of surgical resection. Interventional radiology plays an important role in permitting preoperative embolization, thus reducing blood loss during surgery. On CT and MR scanning, paragangliomas and neuromas are both densely enhancing and are typically indistinguishable. In contrast, on MR, paragangliomas are characterized by multiple flow voids and prominent enhancement, but neuromas usually do not demonstrate flow voids and can be cystic (Fig. 9.16). These features reflect the typically more vascular nature of paragangliomas. Note that these findings are not pathognomonic for paragangliomas, because very vascular schwannomas may also, on occasion, have associated flow voids.
627
FIGURE
9.12. Recurrent
Adenoid
Cystic
Carcinoma. Coronal
postgadolinium fat-suppressed T1WIs of a 50-year-old patient after resection of nasal septum and turbinates for adenoid cystic carcinoma. A recurrent mass (M) extends into the right pterygopalatine fossa (arrows). Adenoid cystic carcinoma has a marked propensity for perineural spread, which allows the tumor to extend rapidly into noncontiguous spaces. Once the tumor enters the pterygopalatine fossa, it may extend into the orbit via the inferior orbital foramen, into the cavernous sinus via the foramen rotundum, and into the infratemporal fossa via the pterygomaxillary fissure. Once in the cavernous sinus, the tumor can travel back along the cisternal portion of the trigeminal nerve into the brain stem.
628
FIGURE
9.13. Parotid
Benign
Mixed-Cell
Adenoma
(Pleomorphic Adenoma). Axial T1WI through the level of the oropharynx. A mass (M) displaces the parapharyngeal space medially (open white arrows) and the masticator space anteriorly (white arrow). The stylomandibular notch from the carotid space to the mandible (curved arrows) is widened, characteristic of a deep lobe parotid lesion. Conversely, a lesion originating from the carotid space would result in narrowing of the stylomandibular notch. The lesion is sharply demarcated from the normal parotid tissue (P).
P.253 Schwannomas are encapsulated tumors that arise from nerve sheath coverings and do not infiltrate the substance of the nerve. Within the carotid space, schwannomas often arise from the vagus nerve and present as benign neck masses. Schwannomas may occasionally show cystic change and necrosis. In contrast to schwannomas, neurofibromas are not encapsulated and usually occur as multiple lesions that permeate the substance of the nerve fibers.
629
Lymph nodes are a common source of pathology within the carotid space. In fact, the principal malignancy of the carotid space is squamous cell nodal metastasis (Fig. 9.17). The deep cervical jugular nodal chain is located within the carotid space and serves as the final common efferent pathway of lymphatic drainage from the head and neck. As such, any pathology of the head and neck (metastases, lymphoma, infection, benign hyperplasia) will typically involve the jugular nodal chain and be found within the carotid space.
Parotid
Space
Masses arising from the deep lobe of the parotid gland will deviate the parapharyngeal space medially. Unlike carotid space masses, deep parotid masses push the styloid process and carotid vessels posteriorly. This results in P.254 characteristic widening of the stylomastoid foramen (Fig. 9.13). The structures within the parotid space that may give rise to pathology include the parotid gland and lymph nodes. The parotid gland is the only salivary gland with lymph nodes contained within its capsule. This reflects the embryogenesis of the parotid gland, the late encapsulation of which results in the presence of 10 to 20 nodes within the gland parenchyma (Fig. 9.18). Consequently, pathology of the parotid space includes salivary gland tumors and nodal disease.
630
FIGURE
9.14. Carotid Body Tumor. Sagittal T1WI (A), axial
postgadolinium fat-suppressed T1WI (B), and angiogram (C). A vascular mass located between the carotid bifurcation, with splaying of the internal and external carotid arteries, is characteristic of a carotid body tumor (c). The multiple flow voids in the axial image supports the diagnosis of a paraganglioma. The angiogram confirms the marked vascularity of this lesion. Angiography is helpful in providing preoperative embolization, which makes the lesion less vascular and easier to remove surgically.
631
Parotid
Tumors
Most parotid tumors are benign (80%), and most of these are benign mixed-cell tumors (pleomorphic adenomas). The second most common benign salivary gland tumor is the Warthin tumor. Malignant tumors, which account for 20% of all parotid lesions, include adenocystic carcinoma, adenocarcinoma, squamous cell carcinoma, and mucoepidermoid carcinoma. MR and CT imaging cannot with certainty differentiate benign from malignant disease. Both may present as well-circumscribed lesions. Tumor homogeneity, indistinct margins, and signal intensity are poor predictors of histology. Nevertheless, benign pleomorphic adenomas are typically well circumscribed and very bright on T2WIs and demonstrate heterogeneous enhancement (Fig. 9.19). Both CT and MR are useful in portraying the relationship of a tumor to surrounding normal anatomy and can demonstrate P.255 the location and extent of a parotid mass before biopsy. A feature predictive of malignancy is infiltration into deep neck structures, such as the masticator or parapharyngeal space. Clinical involvement of the facial nerve is another ominous finding suggestive of malignancy.
632
FIGURE 9.15. Glomus Jugulare Tumor. A. Axial contrastenhanced CT. Fatty atrophy of the right tongue (hypoglossal nerve palsy) (black arrows) and patulousness of the right
633
oropharynx (vagus nerve palsy) (white arrow) are evident. Dysfunction of multiple lower cranial nerves suggests involvement of the skull base, where cranial nerves IX through XII arise in close proximity. B . Axial T1WIs with fat suppression, pregadolinium (left) and postgadolinium (right). A contrastenhancing mass (arrows) extending through the right jugular foramen into the posterior fossa is indicative of glomus jugulare tumor.
The presence of multiple lesions within the parotid space may be seen with several conditions, including either inflammatory or malignant adenopathy. Another possibility is the Warthin tumor (benign salivary gland tumor), which is multiple 10% of the time and more common in men. Parotid cysts have been seen in collagen vascular disease (Sjögren syndrome) and also described in patients with AIDS (Fig. 9.20). These parotid cysts, also known as lymphoepithelial cysts, are believed to be the result of partial obstruction of the terminal ducts by surrounding lymphocytic infiltration.
Masticator
Space
The masticator space is formed by a superficial layer of the deep cervical fascia that surrounds the muscles of mastication and the mandible. It extends from the angle of the P.256 mandible superiorly to the skull base and over the temporalis muscle. The muscles of mastication include the temporalis, the medial and lateral pterygoid, and the masseter. In addition, branches of the trigeminal nerve and the internal maxillary artery are located within this space. Masses in the masticator space displace the parapharyngeal
space
medially
and
634
posteriorly.
FIGURE 9.16. Schwannoma. Axial T2WI through the floor of the mouth. The patient presented with a painless neck mass. A homogeneous mass (S) displaces the carotid space anteriorly (black arrow) and the parotid space (P) laterally (open arrow) . Anterior displacement of the carotid artery is characteristic of a carotid space mass. The lack of associated flow voids suggests that this lesion is a nerve sheath tumor, i.e., schwannoma of the vagus nerve, as opposed to a paraganglioma. High signal within the right retromandibular vein (white arrow) is a result of partial compression. Normal flow void is seen in the opposite retromandibular
vein.
635
Most masses of the masticator space are infectious in origin. They usually result from either dental caries or dental extraction. A mass will often surround the mandible and may extend superiorly along the temporalis muscle. Additionally, pseudotumors of the masticator space are common and include accessory parotid glands as well as marked muscle hypertrophy resulting from bruxism. Occasionally, an accessory parotid gland may occur along the anterior surface of the masseter muscle and can be mistaken for a mass. Asymmetry of the muscles of mastication may result from unilateral atrophy, owing to compromise of the mandibular division of the fifth cranial nerve (V3). This is most commonly seen in patients with head and neck neoplasms with perineural extension along the trigeminal nerve. Primary malignancies of the masticator space are very uncommon. Malignancies of this space most often result from the extension of oropharyngeal or tongue base squamous cell carcinoma to involve the muscles of mastication (Fig. 9.21). In addition, tumor or infection from oropharyngeal or nasopharyngeal lesions may spread along the third division of the fifth cranial nerve, allowing the tumor to ascend through the foramen ovale into the cavernous sinus (Fig. 9.22). From this location, a tumor may extend posteriorly along the cisternal portion of the trigeminal nerve to the brainstem. Primary malignancies of the masticator space include sarcomas arising from muscle, chondroid, or nerve elements. In addition, sarcomas of the bone such as osteosarcoma and Ewing sarcoma may be seen. NonHodgkin lymphoma will occasionally involve the mandible or extraosseous soft tissues of the masticator space (Fig. 9.7) .
Retropharyngeal
Space
The retropharyngeal space is a potential space that lies posterior to the superficial mucosal space and pharyngeal constrictor muscles and anterior to the prevertebral space. A mass within this space results in characteristic posterior displacement of the prevertebral muscles. The fascial planes in this area are complex but can be considered as forming a single compartment for simplicity. This space is significant
636
because it serves as a potential conduit for the spread of tumor or infection from the pharynx to the mediastinum (Fig. 9.23). In contrast to the carotid and parotid spaces, in which inflammatory disease and metastases account for a minority of lesions, most lesions of the retropharyngeal space are a result of infection or nodal malignancy. This space is most often involved with nodal malignancy because of lymphoma or metastatic head and neck squamous cell carcinoma. These tumors frequently affect the retropharyngeal nodes, which are divided into a medial and lateral group. The lateral retropharyngeal nodes, also known as nodes of Rouviere, are normal when seen in younger patients but must be viewed with suspicion in individuals older than 30 years. In addition, head and neck infections may sometimes extend into the retropharyngeal space via lymphatics. Because the retropharyngeal space may serve as a conduit, spreading infection into the mediastinum, this space has also been referred to as the “danger space.― Neck infections are most often the result of tonsillitis, dental disease, trauma, endocarditis,
and P.257
systemic infections such as tuberculosis. With the advent of antibiotics, infections occur much less commonly but are often seen in immunosuppressed patients. On routine T1WIs and T2WIs it can be difficult to differentiate an abscess from cellulitis, as both can be isointense to muscle on T1 and hyperintense on T2. Gadolinium is of value in making this differentiation, as an abscess will demonstrate a rim of contrast enhancement surrounding a liquefied center.
637
FIGURE
9.17. Carotid Space Mass (Jugular Chain
Adenopathy). This 68-year-old man presented with a left neck mass. A . Coronal T2WI reveals left deep jugular chain lymph nodes (arrows), with foci of high signal, indicating necrosis. The presence of abnormal adenopathy must initiate a search for the primary lesion, which is most commonly a mucosa-based squamous cell carcinoma. B . Axial T2WI reveals the primary lesion, consisting of a tonsillar squamous cell carcinoma (arrows) .
638
FIGURE 9.18. Metastatic Lymph Nodes Within the Parotid Gland Capsule. This 79-year-old man presented with left parotid swelling. Coronal T1WI reveals several enlarged nodes within the left parotid gland (arrows). The parotid gland serves as the drainage pathway for the posterior auricular scalp and is characterized by its fatty signal intensity. The finding of parotid nodes initiated a search for ipsilateral pathology, which revealed a retroauricular scalp angiosarcoma.
Prevertebral
Space
The prevertebral space is formed by the prevertebral fascia, which surrounds the prevertebral muscles. Masses of the prevertebral space displace the prevertebral muscles anteriorly. This allows prevertebral lesions to be easily differentiated from retropharyngeal processes, which will displace these muscles posteriorly. The structures that give rise to most pathologies in this space are the cervical vertebral bodies. Any process that involves the vertebral
639
bodies, such as tumor (metastasis, chordoma, etc.) or osteomyelitis, may extend anteriorly to involve this space.
Trans-spatial
Diseases
Occasionally, masses may not be localized to one of the spaces described above. Such masses are often secondary to lesions involving anatomic structures that normally traverse spaces of the head and neck, e.g., lymphatics, nerves, and vessels. Examples include the following three categories: (1) lymphatic masses (lymphangioma); (2) neural masses (neurofibroma, schwannoma, perineural spread of tumor); and (3) vascular masses (hemangioma). Differentiation between these subtypes can occasionally be made by virtue of signal intensity characteristics. For instance, neurofibromas may have a characteristic low-intensity center on T1 and often involve more than one P.258 peripheral nerve. This is distinctly different from both lymphatic and vascular masses. Lymphangiomas and hemangiomas are congenital abnormalities that look quite similar on MR. Both entities have increased signal intensity on T2WIs and are infiltrative. Hemangiomas may have phleboliths, which may be easily detected on CT (Fig. 9.24). Lymphangiomas tend to have heterogeneous signal intensity with evidence of blood degradation products. Both entities should be considered in a patient with a history of chronic facial swelling and who shows CT or MR evidence of an infiltrative process that traverses several spaces.
640
FIGURE 9.19. Benign Pleomorphic Adenoma. A. Axial T1WI. B . Axial fast spin-echo fat-saturated T2WI. C . Postgadolinium fat-saturated T1WI. The patient presents with a wellcircumscribed parotid mass (arrow), which is bright on T2WI and demonstrates heterogeneous contrast enhancement. These imaging features are consistent with a benign pleomorphic ademona, which is the most common parotid lesion, accounting for 80% of all benign parotid tumors.
641
Perineural
Disease
Perineural spread of disease allows tumor or infection to gain access into noncontiguous spaces of the head and neck. The complex system of cranial nerves coursing through the skull base serves as a conduit for the spread of tumor and infection. Fungal infections (Fig. 9.22) , squamous cell carcinoma, and adenoid cystic carcinoma have a particular proclivity for perineural spread of disease, which serves as a hallmark of these diseases. If a patient with a known head and neck primary neoplasm or immunocompromised status (susceptible to fungal infections) presents with facial numbness or dysesthesias, this is highly suggestive of perineural spread of P.259 disease, and careful attention must be paid to imaging of the cranial nerves of the skull base (Fig. 9.25) .
FIGURE 9.20. Benign Lymphoepithelial Cysts in Sjögren Syndrome. Axial T2WI. The 27-year-old woman presented with
642
parotid swelling and complaints of dry eyes and mouth and was diagnosed with Sjögren syndrome, a chronic autoimmune disorder. MR reveals innumerable tiny parotid cysts, reflecting the lymphocytic infiltration of the exocrine glands, which causes lymphatic obstruction and cyst formation (arrows). Parotid cysts (benign lymphoepithelial cysts) can be seen in a variety of conditions
LYMPH
with
lymphocytic
infiltration,
including
AIDS.
NODES
Once a primary neoplasm of the head and neck is detected, the assessment of lymph nodes is a vital part of tumor staging. The presence of a single ipsilateral malignant node reduces the patient's expected survival by 50%, with extracapsular nodal extension reducing survival by an additional 25%. Thus, the detection of nodal disease is critical for both prognosis and therapy. CT, MR, and PET all play a vital role in the staging of head and neck neoplasms, because clinically, it is difficult to determine the full size of the primary neoplasm and its associated nodal extension. At least 15% of malignant nodes are clinically occult because of their deep location (e.g., retropharyngeal nodes) and thus are not palpable by the clinician. The overall error rate in assessing the presence of adenopathy by palpation is between 25% and 33%. Thus, PET combined with either CT or MR is vital in obtaining the most accurate pretreatment planning information. There are at least 10 major lymph node groups in the head and neck. Knowledge of the location of these cervical lymph node chains and the usual modes of spread of head and neck disease is essential for successful analysis of CT and MR scans. We will focus on the principal lymph node group of the neck: the internal jugular chain. The internal jugular nodal chain serves as the final common afferent pathway for lymphatic drainage of the entire head and neck. This nodal chain follows the oblique course of the jugular vein beneath and adjacent to the anterior border of the sternocleidomastoid muscle. The jugulodigastric node is the highest node of the internal
643
jugular chain. It is located where the posterior belly of the digastric muscle crosses this chain, near the level of the hyoid bone. The jugulodigastric lymph node is immediately posterior to the submandibular gland and provides lymphatic drainage from the tonsil, oral cavity, pharynx, and submandibular nodes.
FIGURE
9.21. Squamous Cell Carcinoma of the Tongue. Axial
postgadolinium fat-suppressed T1WI through the level of the oropharynx. A left tongue base squamous cell carcinoma extends posteriorly along the oropharyngeal wall into the masticator space. Malignancies of the masticator space are most frequently the result of the direct posterior extension of oropharyngeal squamous cell carcinoma. In this example, the left half of the tongue (arrows) is diffusely enhancing, indicating denervation myositis, due to involvement of the hypoglossal nerve. Later, the muscles will atrophy and become replaced by fat. mp, medial
644
pterygoid muscle; lp, lateral pterygoid muscle.
The jugulodigastric node and submandibular nodes may normally measure up to 1.5 cm in diameter; in contrast, all other nodes of the head and neck are considered abnormal if larger than 1.0 cm. When an enlarged node P.260 is encountered on CT or MR, differentiation between a benign reactive node and a malignant one can be difficult. Several features that suggest malignancy are (1) peripheral nodal enhancement with central necrosis, (2) extracapsular spread with infiltration of adjacent tissues, and (3) a matted conglomerate mass of nodes. Nodal size itself is a less reliable indicator of malignancy, but it is used because the other more reliable differentiating features are frequently not present. If size criteria alone are used, approximately 70% of enlarged nodes are secondary to metastatic disease and 30% are caused by benign reactive hyperplasia. Note that the features described as characteristic for malignancy are the same as those for infection, and the two cannot be differentiated by imaging. Fortunately, this distinction is often easily made clinically.
645
FIGURE
9.22. Perineural Spread of Disease: Mucormycosis
Infection. A 32-year-old man presented in diabetic ketoacidosis, with left facial numbness. Perineural spread of disease is noted extending from the anterior cheek all the way to the cavernous sinus and brainstem. Perineural spread of a neoplasm, such as adenoid cystic carcinoma or squamous carcinoma, would have an identical imaging appearance. A . Axial postgadolinium fatsuppressed T1WI (left) and T2WI (right) through the level of the nasopharynx. Soft tissue infiltration involves the left malleolar soft tissues, and extends along the maxillary division of the trigeminal nerve (V2) (arrows) into the cavernous sinus. From the cavernous sinus, contrast-enhancing tissue extends along the cisternal portion of the trigeminal nerve (open arrows) to the
646
brainstem. B . Coronal T1WIs, pregadolinium (left) and postgadolinium (right) with fat suppression. Contrast enhancement is seen filling the cavernous sinus and extending through the foramen ovale (arrow) into the masticator space along the mandibular division of the trigeminal nerve (V3).
PET scanning plays a vital role in the staging of any head and neck malignancy. Because metastatic nodes, regardless of size, are typically very glucose avid, PET provides exquisite sensitivity and specificity in the detection of cervical metastatic nodal disease. A lymph node that appears normal by size criteria on MR or CT may in fact be malignant if hot on PET scan (Fig. 9.26). The converse is also true; an enlarged lymph node on MR or CT may in fact be benign reactive in nature, if cold on PET. Lymph nodes can be accurately detected with either multislice helical CT or MR, and the decision regarding which technique to use should be based upon the imaging the patient is most likely to tolerate. Head and neck oncology patients often have respiratory and swallowing issues that prevent them from keeping sufficiently still for satisfactory MR scans. In contrast, multislice CT provides for rapid thin-section imaging of the neck with minimal motion artifact. With MR imaging, lymph nodes are well visualized on fat-suppressed FSE T2WIs, as well as precontrast T1WIs and postcontrast fat-suppressed T1WIs. Normal lymph nodes demonstrate homogeneous signal P.261 intensity, whether on precontrast or postcontrast T1WIs or T2WIs. Any heterogeneity in signal, especially in the presence of cystic change or necrosis, is consistent with metastatic disease (Fig. 9.6) . Note that a fatty central hilus is a normal finding. Shape is also a differentiating feature, as a rounded shape suggests neoplastic nodal infiltration with associated nodal expansion. In contrast, if a node is enlarged but maintains its normal reniform configuration, it more likely reflects benign reactive change rather than metastatic disease.
647
FIGURE
9.23. Retropharyngeal
Abscess. Axial postcontrast CT
through the level of the larynx (A) and the upper mediastinum (B). A large fluid collection (A) extends from the retropharyngeal space into the upper mediastinum. The posterior displacement of the prevertebral muscles (m) (arrows) identifies this collection as being retropharyngeal as opposed to prevertebral.
648
ORBIT Both CT and MR are valuable for imaging of the orbit; each has distinct merits. When evaluating for calcification, such as in retinoblastoma in a child with leukocoria or for bony fracture following trauma, CT is the modality of choice. MR, on the other hand, with its multiplanar capability and superior soft tissue discrimination, has proven to be of tremendous value in orbital imaging. For most orbital abnormalities, including evaluation of the visual pathways, MR is the procedure of choice. Knowledge of the contents of the various orbital spaces provides insight into the naturally occurring lesions that develop within each area. The retrobulbar space contains both the extraconal and the intraconal spaces, which are separated by the muscle cone or “annulus of Zinn.― This muscle cone is formed by the extraocular muscles (superior, inferior, medial, and lateral rectus; superior oblique; and levator palpebrae superior) and a fibrous septum. Together these structures form a cone with its base at the posterior of the globe and its apex at the superior orbital fissure. When identifying an intraconal lesion, an essential issue is whether the lesion arises from the optic nerve sheath complex or is extrinsic to it. The optic nerve sheath complex is composed of the optic nerve and the surrounding perioptic nerve sheath. The optic nerve is an extension of the brain enveloped by CSF and leptomeninges, which form the optic nerve sheath. Therefore, the CSF space that envelops the optic nerve is continuous with the intracranial subarachnoid space. If a lesion arises from the optic nerve sheath complex, the most common lesion is either an optic nerve glioma or optic sheath meningioma. Optic nerve glioma is the most common tumor of the optic nerve and typically occurs during the first decade of life (Fig. 9.27). There is a high association with neurofibromatosis type 1, particularly when there is bilateral optic nerve involvement. Histologically, these lesions are low-grade pilocytic astrocytomas. The characteristic imaging finding is that of enlargement of the optic nerve sheath complex. The enlarged sheath complex may be tubular, fusiform, or
649
eccentric with kinking. Some optic nerve gliomas have extensive associated thickening of the perioptic meninges. Histologically, this reflects peritumoral-reactive meningeal change, which has been termed arachnoidal hyperplasia or gliomatosis. This finding is often seen in patients with neurofibromatosis. Optic
sheath
meningiomas arise from hemangioendothelial cells of
the arachnoid layer of the optic nerve sheath. These lesions assume a circular configuration and grow in a linear fashion along the optic nerve. Optic sheath meningiomas demonstrate a characteristic “tram track― pattern of linear contrast enhancement, because the nerve sheath enhances, rather than the nerve itself. MR easily displays any tumor extension along the optic nerve sheath through the orbital apex (Fig. 9.28). In contrast to optic nerve gliomas, meningiomas may invade and grow through the dura, resulting in an irregular and asymmetric appearance. Additionally, optic sheath meningiomas may be extensively calcified, whereas optic nerve gliomas rarely have any calcification. In patients with sarcoidosis, P.262 leukemia or lymphoma, cellular infiltrates may deposit within the perioptic nerve sheath CSF space. In such cases, contrast enhancement of the perioptic nerve sheath space may mimic the “tram track― appearance of a nerve sheath meningioma. An important differential diagnostic consideration for enhancement of the optic nerve sheath is optic neuritis. In contrast to the conditions just mentioned, which demonstrate enhancement of the optic nerve sheath (i.e., peripheral optic nerve enhancement), optic neuritis demonstrates abnormal T2 hyperintensity and contrast enhancement as a result of inflammation of the optic nerve itself (Fig. 9.29). Optic neuritis presents with an acute visual deficit, often described as “blurring― of vision, and can be the first sign of multiple sclerosis (MS). Approximately 20% of patients with MS initially present with an episode of optic neuritis. In fact, of patients with isolated optic neuritis, approximately 50% eventually are diagnosed with MS.
650
FIGURE
9.24. Hemangioma. Patient presented with a facial
mass which demonstrates high signal on T2WI (A) with punctuate foci of signal void (arrows). On CT (B), these foci of low T2 signal prove to be phleboliths (arrows), which is essentially pathognomonic of the diagnosis of hemangioma. In another patient with a similar clinical presentation, a T2WI (C) reveals a multilobulated and multiseptated high-signal-intensity lesion with imaging characteristics also consistent with hemangioma (arrows). Lymphangiomas may be indistinguishable from this lesion, but often have fluid–fluid levels related to hemorrhage.
651
FIGURE 9.25. Perineural Spread of Tumor. This 18-year-old woman was recent postresection of a right cheek melanoma with clean histologic margins. However, a CT was performed following persistent maxillary division paresthesias. Coronal plane image reveals abnormal enlargement of the maxillary nerve within the infraorbital canal (solid arrow), which extended back to the pterygopalatine fossa. Focal nerve biopsy revealed perineural spread of melanoma. Compare to normal infraorbital nerve and canal (open arrow) .
Vascular
Lesions
A variety of vascular lesions may develop in the orbit. The four lesions we will consider include capillary hemangioma, lymphangioma, cavernous hemangioma, and varix. These lesions are readily distinguished by a combination of imaging and clinical findings, including the patient's age (see Table 9.2). Capillary hemangiomas P.263
652
develop in infants (younger than 1 year) and are diagnosed within the first weeks of life. Although these lesions may grow rapidly in size, they typically plateau during the first year or two then regress spontaneously. On imaging studies, a capillary hemangioma appears as an infiltrative soft tissue complex, often with multiple vascular flow voids. In contrast, lymphangiomas are one of the most common orbital tumors of childhood and occur in an older group of children (3 to 15 years). Lymphangiomas are characterized by their propensity to bleed, and they often contain blood degradation products. An acute hemorrhage may result in marked expansion of the lesion with sudden proptosis (Fig. 9.30). MR reveals a multiloculated, lobular mass with characteristic signal heterogeneity caused by blood degradation products (Fig. 9.31). The older age of presentation, combined with the characteristic heterogeneous signal related to blood products, allows differentiation from the capillary hemangiomas. Cavernous hemangiomas are one of the most common orbital masses in adults. In contrast to the other vascular lesions of the orbit, hemangiomas are characterized as a sharply circumscribed, rounded mass (Fig. 9.32). These lesions demonstrate diffuse enhancement, sometimes with a mottled pattern. The venous varix is an enormously dilated vein that is characterized by its marked change in size with the Valsalva maneuver.
653
FIGURE
9.26. Metastatic Lymph Node That Appeared
Normal Based on MR Size Criteria. This 68-year-old man presented with a right nasopharyngeal carcinoma, which appeared localized without significant cervical adenopathy. However, a PET scan (A) revealed a focus of abnormal uptake (arrow) in a lymph node, which appeared normal on MR. Note incidental uptake in the lingual salivary glands and tongue base lymphoidal tissue on the PET scan (open arrows). B . This lymph node (arrow) appears morphologically benign on MR, but was proven metastatic disease on modified radical neck dissection. V, vertebral
Superior
body.
ophthalmic
vein is well visualized on MR studies.
Pathology includes thrombosis and enlargement. Thrombosis often occurs in conjunction with cavernous sinus thrombosis and presents as loss of the normal flow void, with signal intensity related to the age of the thrombus. Enlargement of the superior ophthalmic vein may also be seen with cavernous carotid fistulas (Fig. 9.33) . Cavernous carotid fistulas represent direct or indirect communication between the internal carotid artery and the venous cavernous sinus. These are either spontaneous or posttraumatic, and patients may present with pulsating exophthalmos and bruit. Pseudotumor
and
lymphoma are two important orbital lesions that
may present with similar imaging findings. Idiopathic inflammatory pseudotumor is a poorly characterized condition that results from an inflammatory lymphocytic infiltrate. This is the most common cause of an intraorbital mass lesion in the adult population. Pseudotumor is often rapidly developing and presents with painful proptosis, chemosis, and ophthalmoplegia. In contrast, lymphoma tends to present with painless proptosis. Lymphoma is the third most common adult orbital mass lesion, following pseudotumor and cavernous hemangioma. On imaging studies, both lymphoma and pseudotumor appear as diffusely infiltrating lesions capable of involving and extending into any retrobulbar structures (Fig. 9.34). Several reports have suggested that T2 shortening of the tumor (dark on T2) is
654
suggestive of pseudotumor. Nevertheless, these two entities frequently
the
distinction
between P.264
remains very difficult clinically, radiographically, and even histopathologically.
FIGURE
9.27. Optic Nerve Glioma. Axial
pregadolinium (A)
and postgadolinium (B) fat-suppressed T1WIs through the orbits. A large mass involves the right optic nerve. Following the administration of contrast, the enlarged optic nerve (arrows) is visible coursing through markedly thickened optic sheath soft
655
tissue. This soft tissue represents arachnoidal hyperplasia, a finding associated with optic gliomas in patients with neurofibromatosis.
It is reported that a trial dose of steroids may be valuable in differentiating these two entities. Steroids are reported to have a lasting effect, eliminating a pseudotumor lesion. However, the cytolytic effect of steroids on lymphoma may also have a similar but short-lived response that may initially be confounding. Additionally, when a diffusely infiltrative mass is encountered in a young child anywhere in the head and neck region, including the orbits, rhabdomyosarcoma Thyroid
should
be
a
consideration.
ophthalmopathy (Graves disease) is a common lesion and
is the most frequent cause of unilateral or bilateral proptosis in adults. This condition is the result of an inflammatory infiltration of the orbital muscles and orbital connective tissues. Most patients will have clinical or laboratory evidence of hyperthyroidism, but 10% will not; these are referred to as “euthyroid ophthalmopathy.― Imaging findings consist of enlargement of the extraocular muscles with sparing of the tendinous attachments to the globe (Fig. 9.35) . This is in contrast to pseudotumor, which typically involves the muscle attachments to the globe. The muscles involved, in decreasing order of frequency, are the inferior, medial, superior, and lateral rectus (pneumonic “I’m slow― reminds one of the order of muscle involvement and the typical orbital symptoms of Grave disease, namely lid lag and limitation in orbital movement). Eighty percent of patients have bilateral muscle involvement. In some cases of thyroid ophthalmopathy, the extraocular muscles may be normal, and exophthalmos is the result of increased retrobulbar fat.
656
FIGURE
9.28. Optic
Sheath
Meningioma. Axial postgadolinium
fat-suppression T1WI through the orbits. “Tram track― enhancement involves the left optic nerve sheath (long arrow) , and a tumor (short arrows) extends into the middle cranial fossa. The “tram track― enhancement and the dural tail within the middle cranial fossa are characteristic of a meningioma.
Lacrimal
Gland
657
The extraconal space primarily contains fat and the lacrimal gland. However, many lesions P.265 involving the extraconal space are the result of tumor or inflammation extending from surrounding structures. These may include most of the lesions described earlier, as well as sinus-related inflammation. In contrast, lesions arising from within the extraconal space are primarily lacrimal in origin. Lesions of the lacrimal gland are very nonspecific, but can be divided into inflammatory types (e.g., sarcoidosis, Sjögren syndrome) and neoplastic types. Neoplasms of the lacrimal gland include epithelial and lymphoid tumors. Epithelial tumors are any of the lesions that arise from the salivary glands, such as benign mixed-cell tumor or adenoid cystic carcinoma. Lymphoid tumors include lymphoma and pseudotumor. Although none of these lesions have specific imaging findings, dermoid is one lesion that does have a characteristic finding, consisting of a fat–fluid level (Fig. 9.36) .
FIGURE 9.29. Optic Neuritis. This 23-year-old woman presented with bilateral loss of vision. Coronal fat-saturated T1WI postgadolinium enhancement of both (arrows), which often Optic neuritis reflects
administration reveals marked optic nerves characteristic of optic neuritis heralds the onset of multiple sclerosis. a demyelinating often related to multiple
658
sclerosis; however, other etiologies include demyelination or inflammation secondary to infections related to sinusitis, tuberculosis, and viral agents such as herpes and cytomegalovirus, or as a complication of radiation therapy.
FIGURE
9.30. Lymphangioma. Axial T2WI reveals a cystic
retrobulbar lesion (arrows) with a hematocrit effect (serum layered above red blood cells). Hemorrhage into a lesion is a characteristic feature of lymphangiomas and may be responsible for the rapid development of proptosis.
659
TABLE 9.2 Vascular Orbital Lesions
Lesion
Age
Imaging Features
Morphology
Capillary hemangioma
Table of Contents > Section II - Neuroradiology > Chapter 10 Nondegenerative Diseases of the Spine
Chapter
10
Nondegenerative Spine
Diseases
of
the
Erik H.L. Gaensler This chapter focuses on nondegenerative diseases of the spinal cord, meninges, and paraspinous soft tissues and is divided into sections covering inflammation, infection, neoplasms, vascular diseases, congenital malformations, and trauma (1 , 2 , 3 , 4 , 5 ). The spine is composed of vertebrae, which house the spinal cord and proximal spinal nerves, and thereby represents a “border zone― between the CNS and the musculoskeletal system (this is true politically as well as anatomically—with both neurosurgeons and orthopaedic surgeons claiming the spine as their province). Disk degeneration and spinal stenosis are covered in Chapter 11 . Primary osseous tumors involving the vertebrae are covered in Section X.
Common
Clinical
Syndromes
The clinical syndromes produced by degenerative disease and nondegenerative disease can be indistinguishable. Patients with spine disorders present with focal or diffuse back pain, radiculopathy, or myelopathy. Focal back pain without neurologic compromise or fever is not usually an emergency and is an epidemic in our society, with tremendous implications in terms of lost productivity. Focal back pain can be caused by a wide variety of both
677
degenerative and nondegenerative processes. In the low back, the causes most commonly are orthopaedic, such as muscle and ligament strain, facet joint disease, or discogenic disease that does not compromise the nerve roots. However, vertebral metastases or infectious discitis may also cause focal back pain. Because degenerative disease of the spine is far more common than nondegenerative disease, nondegenerative processes may initially be overlooked, with disastrous consequences. Therefore, a good clinical history that specifically addresses any previous cancers or ongoing fevers and chills is crucial in raising the suspicion for a nondegenerative process. When history and physical findings are nonspecific, as often is the case, imaging procedures become central to the diagnosis. In patients with spinal neurologic findings, an attempt should be made to distinguish between the clinical syndromes of myelopathy and radiculopathy, because they differ in significant respects, including degree of urgency. Important distinctions between radiculopathy and myelopathy are summarized in Table 10.1 . Myelopathy results from compromise of the spinal cord itself, owing to mechanical compression, intrinsic lesions, or inflammatory processes loosely grouped under the term “myelitis.― Classic symptoms include bladder and bowel incontinence, spasticity, weakness, and ataxia. With cord compression, a clear motor or sensory spinal cord “level― may develop, and knowledge of this level is helpful in focusing the imaging examination. However, the lesion may be several vertebral bodies higher than the apparent dermatomal sensory level, particularly in the thoracic P.272 region. Myelopathy often presents without a clear sensory level, and complete screening of the cord from the cervicomedullary junction to the conus may be required. Cause Spinal cord compromise Spinal nerve compromise Typical disease processes
678
Extramedullary disease: cord compression caused by epidural mass effect Cervical spinal stenosis Intramedullary disease: tumor, inflammation, AVMs, Osteophytic spurring (especially cervical spine) Disk herniations Lumbar spinal stenosis Extramedullary and paraspinous compromising nerve roots
tumors
and
SDAVFs
inflammatory
processes
Neurologic findings Ataxia Bowel and bladder incontinence Babinski sign Weakness and diminished reflexes in specific muscle groups; dermatomal sensory deficits Accuracy of clinical localization Often poor; lesion may be several levels higher than anticipated Usually quite good Urgency for imaging (of acute presentations) High—little recovery expected with deficits untreated >24 hr Low—short delay for conservative treatment usually entails little risk Preferred
imaging
modality
M R has no substitute as the initial screening exam M R , although C T with intrathecal contrast is still excellent, particularly in cervical spine AVM, arteriovenous malformation;
SDAVF,
spinal
dural
arteriovenous
fistula. Myelopathy
Radiculopathy
TABLE 10.1 Myelopathy Versus Radiculopathy The spinal cord, like the brain, has limited healing powers. In fact, the spinal cord in many respects is less tolerant of injury than the brain. A small benign mass, such as a 2-cm epidural hematoma or meningioma, may permanently damage the cord because of the
679
small diameter of the spinal canal. However, a similar-sized mass may be asymptomatic within the voluminous calvaria. The “plasticity― of the brain, whereby remaining cortex can assume the function of injured areas through a complex network of redundant neurons, is well documented, particularly in younger patients. The spinal cord, which consists mostly of long linear axonal tracts, has far less plasticity. After 24 hours of acute cord compression, little hope may remain for significant recovery of function. Therefore, acute myelopathy is an emergency in which the radiologist should do everything to facilitate prompt imaging. Radiculopathy generally results from impingement of the spinal nerves, either within the spinal canal, lateral recess, or neural foramen, or along the extraforaminal course of the nerve. This compromise, typically because of mass effect, results in specific dermatomal sensory deficits and/or muscle group weakness. These are outlined in any neurology or physical diagnosis text and are worth knowing. The most common causes of pain and neurologic deficit are disk herniations and spinal stenosis and, in the cervical spine, uncovertebral joint spurring. Of course, malignant and infectious processes compromise spinal nerves but overall are less common. The peripheral nervous system, unlike the CNS, has significant ability to withstand injury and to regenerate. Therefore, pure radicular symptoms, although at times they are excruciatingly painful, rarely represent a surgical emergency. Extensive epidural neoplasms and infections may present with mixed myelopathic and radicular signs. These patients must be imaged with the urgency of a pure cord syndrome.
Imaging
Methods
Plain radiographs of the spine were once the initial test in every spine evaluation, but with newer techniques, this is no longer logical or cost effective. Radiographs continue to be useful for ruling out trauma to the vertebral column and other acute screening settings. Plain films and fluoroscopy are indispensable for correct localization in the operating room. Radiographs have a great deal of useful information to offer when evaluating degenerative processes,
680
particularly with extensive osteophyte formation in the cervical spine. Flexion and extension plain films used to be the only dynamic imaging technique for assessment of spine stability. MR now also can be done in flexion and extension, which can be useful in evaluating cord compression that is positional (see Fig. 10.8 ). In nondegenerative disease, careful attention should be paid to the integrity of the vertebral bodies and pedicles, which are frequent sites of metastases. However, plain films cannot detect early infiltrative changes in the marrow space, which are easily seen on MR. The classic radiographic findings of widened interpedicular distance with tumors and midline bony spurs with diastematomyelia are rarely seen except on board examinations.
Myelography The indications for plain film myelography alone are limited. Myelography today is almost always done in conjunction with CT (see
following).
Indications P.273
include complex postoperative cases and patients in whom MR is contraindicated because of MR-incompatible implanted devices. Ionic contrast agents are absolutely contraindicated for myelography because they can result in severe inflammation, seizures, arachnoiditis, and even death. Always personally inspect the vial of contrast you are using, and fill the syringe yourself. The recommended dosage of nonionic contrast in adults depends on the region to be studied, the size of the patient, and the size of the thecal sac. A convenient and conservative rule of thumb in adults is not to exceed 3 g of intrathecal iodine, which works out to 17 mL of 180 mg/mL, 12.5 mL of 240 mg/mL, or 10 mL of 300 mg/mL (these are three standard concentrations). In general, lumbar myelography should be performed using contrast media with a concentration between 180 and 240 mg/mL, and cervical and/or thoracic myelography should be performed with concentrations of 200 to 300 mg/mL. The smaller the area of the subarachnoid space, the denser the contrast must be for good plain films. Plain films and
681
fluoroscopic spot films, however, are becoming increasingly superfluous with the dramatic improvements in multiplanar CT reconstructions. Myelography begins with a lumbar puncture, with the patient in prone position under fluoroscopy. The preferred puncture site depends on the clinical findings, but is usually the midlumbar region, inferior to the posterior elements of L2 or L3. This injection level will avoid most disk herniations and spinal stenosis, which are usually worse at lower levels, and the conus, which in adults lies between the T12–L1 and L1–L2 disk spaces. Care should be taken to place the needle near the midline to reduce the chances of an extraarachnoid injection or spearing of an exiting nerve root. Contrast should be injected only after spontaneous CSF backflow is established. The complications of poor needle placement include subdural and epidural injection. Examples of these complications are well illustrated in older neuroradiology textbooks. If there is doubt as to where the contrast is going, stop, take frontal and lateral spot films, and examine them carefully. Avoid any air bubbles in the tubing system, as they can cause filling defects, which are easily confused with drop metastases. If tumor or infection is suspected, collect adequate CSF for chemistry, cultures, and cytology if this has not already been done. For routine degenerative cases, CSF examination has not proved worthwhile. C1–C2 punctures are rarely required and are inherently more dangerous than lumbar injection, as direct injury to the cord or a low-lying posterior inferior cerebellar artery loop can occur. The puncture is best done under lateral fluoroscopy, with the needle placed in the posterior third of the spinal canal between C1 and C2. Classic indications include known blocks caudally or the need for dense opacification of the cervical and upper thoracic spinal canal for plain films. Today, one of the rare good reasons for a C1–C2 puncture would be complete spine block in the midthoracic region identified by lumbar myelography, with the need to define the upper extent of the block—in a patient with a pacemaker (thus precluding MR). If the pacemaker were not an issue, MR would be the study of choice. MR is far quicker, more comfortable, and, most important,
682
safer for the patient. Even if there is no technical complication with a myelogram, patients with spine block can deteriorate from the subtle fluid and pressure shifts that inevitably accompany needle placement in the subarachnoid space, a syndrome known as “spinal coning.― The multiple steps in the evaluation of spine block by plain film myelography followed by CT are shown in Fig. 10.1 . Contrast this with the simplicity and elegance of MR, as shown in Fig. 10.2 . In oncologic cases, MR has the additional benefit of excellent evaluation of the marrow space—which is not available with CT. Space-occupying lesions of the spinal canal are classified according to their location as intramedullary, intradural-extramedullary, or extradural. This distinction can be made on myelography, as well as on CT and MR, and is critical in formulating a differential diagnosis. Intramedullary lesions are usually confined to the spinal cord itself but may be exophytic. Extramedullary lesions are by definition outside the cord, but may be either intradural or extradural. A summary of the radiologic appearance and differential diagnosis for each lesion location is outlined in Table 10.2 . Remember that the lesion must be seen in at least two (and preferably three) 90° orthogonal planes, since large intradural lesions may simulate an extradural mass on any single view. Similarly, bilateral extradural disease can flatten the cord, increasing its apparent anteroposterior dimension in sagittal view and thereby giving the false impression of an intramedullary mass (Fig. 10.3 ). Correlation with axial imaging is invaluable in this regard. Also remember that lateral lesions, such as lateral disk herniations, may be completely missed by myelography. In almost all cases today, a CT is performed after myelography.
CT The decline of plain film myelography for degenerative disease was initially because of CT, especially CT with intrathecal contrast, which is superior to myelography in diagnostic accuracy. However, CT has largely been replaced by MR for most screening examinations of the spine, except for acute trauma. Low-dose CT myelography remains
683
the gold standard in cases where the limits of the thecal sac or nerve root sleeves need to be precisely defined, such as in complex postoperative states. Small leptomeningeal (drop) metastases can be identified (see Fig. 10.34 ); however, MR with gadolinium has replaced CT myelography as the initial screening examination for drop metastases (see Figs 10.33 , 10.35 ). CT is far less effective than MR in depicting intramedullary diseases of the spinal cord, such as primary tumors, myelitis, and syringohydromyelia. For example, P.274 P.275 a nonexpansile multiple sclerosis (MS) plaque will escape detection on any imaging examination except MR.
684
FIGURE
10.1. Acute
Cord
Compression. Middle-aged patient with
acute myelopathy and midthoracic back pain, worked up the “oldfashioned― way, as the patient had a pacemaker. A . Plain film obtained in the emergency department shows compression fracture of a midthoracic vertebra (arrow ). B . Lumbar myelogram shows complete block to contrast in the midthoracic vertebrae (arrows ). A portable C-arm fluoroscope then had to be obtained to do a C1–C2 puncture, followed by a cervical and upper thoracic myelogram (not shown). C . Upper thoracic CT myelogram images show gradual effacement of the subarachnoid space (arrow ), which disappears at site of the block (arrowheads ) D . Sagittal reconstruction enables assessment of the entire process in a single image, showing cord compression centered around an abnormal disk space (arrow ), consistent with infection, which was proven at laminectomy. Note the gradual effacement of the subarachnoid space (arrowheads ).
M R has done for the spinal canal what CT did for the calvaria,
685
allowing for the first time a noninvasive “look inside.― Therefore, it is the examination of choice for any disorder of the spine resulting in myelopathy. The key to the success of MR has been its superior soft tissue contrast (including the ability to evaluate the marrow compartment), multiplanar capabilities, noninvasiveness, and high sensitivity to gadolinium enhancement. MR scanning techniques for the spine continue to improve, and with the wide variety of imaging systems available, it makes little sense to recommend specific protocols in a general text. A few general guidelines follow. Surface coils are an absolute must to obtain an adequate signal-to-noise ratio in most systems. Motion-suppression techniques, such as anterior radiofrequency saturation bands, gradient moment nulling, and cardiac/respiratory gating, are critical to reduce motion artifact. Fast spin-echo (FSE) sequences have replaced conventional spin echo for spine work, with great time saving and little cost when only degenerative disease is present. The FSE technique, however, is poor for marrow evaluation, but this can be overcome by using fat saturation with the T2WI, a technique widely used in musculoskeletal MR to search for marrow edema. Short time inversion recovery (STIR) probably offers the highest sensitivity for marrow space edema. Fast inversion recovery techniques compete with T2 FSE with fat saturation as the optimal marrow-screening
exam
(see Figs 10.37 , 10.43 ).
Gradient-echo images are poor for marrow space evaluation because of susceptibility effects from the bony trabeculae and are of little utility in evaluating nondegenerative spinal disease, except when searching for blood P.276 breakdown products (see Fig. 10.65 ). Ultra thin–section imaging ( Table of Contents > Section II - Neuroradiology > Chapter 11 Lumbar Spine: Disk Disease and Stenosis
Chapter
11
Lumbar Spine: Disk Disease and Stenosis Clyde A. Helms
Imaging
Methods
Imaging the lumbar spine for disk disease and stenosis has evolved in the past 20 years from predominantly myelography-oriented examinations to plain CT and MR examinations. Multiple studies have shown that myelography is not as accurate as CT or MR (1, 2, 3), yet myelography continues to be performed. Little justification exists for using a lumbar myelogram to determine disk disease or stenosis in this era. Although few differences between CT and MR have been noted concerning diagnostic accuracy in the lumbar spine, MR will give more information and a more complete anatomic depiction than will CT. For example, MR can determine whether a disk is degenerated by showing loss of signal on T2WIs (Fig. 11.1). CT cannot provide this information. Whether or not this is useful information remains to be proven. To achieve a high degree of accuracy, the proper imaging protocols must be observed. With CT scans, thin-section (3- to 5-mm) axial images should be obtained from the midbody of L3 to the midbody of S1 in a contiguous manner, i.e., no skip areas or gaps should be present (Fig. 11.2). One of the leading causes of failed back surgery
790
is missed free fragments. Skip areas will often allow a free fragment to remain undiagnosed. Angling the gantry parallel to the endplates is not necessary, and image reformations are not helpful in the routine evaluation of disk disease and stenosis. The MR imaging protocol is similar to that of CT, in that thin-section axial images should be obtained from the midbody of L3 to the midbody of S1 (Fig. 11.3). Angling of the plane of imaging to be parallel to the endplates is not necessary, and contiguous images without skip areas are considered mandatory. Even though sagittal images will be obtained, free fragments and areas of stenosis P.323 P.324 are often seen on axial images to better advantage than on sagittal images. Other entities that can be overlooked if gaps are present in the axial imaging protocol include conjoined nerve roots, pars defects (spondylolysis), and lateral recess stenosis. These entities occur cephalad or caudal to the vertebral body, away from the disk level; thus, axial images limited to the disk level will not show them, and they may not be conspicuous on the sagittal images.
791
FIGURE
11.1. Desiccated
Disk. A sagittal T2WI (TR, 4,000; TE,
102) shows the L2-L3 and L3-L4 disks to be abnormally low in signal, indicating disk dessication and degeneration. Compare with the normal L1-L2 disk (arrow), which has high signal.
792
FIGURE
11.2. Inadequate
Technique:
Skip
Areas. This MR
scout film has cursors placed through the disk spaces. This allows large gaps or skip areas that can result in missed free fragments of disks.
793
FIGURE 11.3. Proper MR Technique. This MR scout, with cursors placed contiguously from the body of L3 to S1, allows complete coverage of the lower lumbar spine in the axial plane.
794
FIGURE 11.4. Disk Protrusions. Axial images show focal (A) (arrows) and broad-based (B) disk protrusions (arrows). Because both are showing impression of the thecal sac, they could each cause
symptoms.
Both T1WIs (or proton density–weighted images) and T2WI (or T2*WI) should be obtained in the sagittal and the axial planes. Attempting to shorten the study by foregoing one of the sequences is not recommended.
Disk
Disease
Disk
Protrusions
Terminology plays a large role in how radiologists describe disk bulges or protrusions. Since the advent of CT in the 1970s, disk bulges have been described by their morphology. A broad-based disk bulge has been said to be a bulging annulus fibrosus, and a focal disk bulge is a herniated nucleus pulposus. These interpretations are no more than 90% accurate. More significantly, most surgeons are not concerned with the name applied to a disk bulge; they do not treat a bulging annulus differently than a herniated nucleus pulposus. They
795
treat the patient's symptoms and have to decide whether the disk bulge is responsible for those symptoms. Most surgeons are satisfied with the terms “bulge― or “protrusion.― Up to 50% of the asymptomatic population have disk protrusions (4); hence, evidence of a disk bulge on CT or MR does not mean it is clinically significant. Both CT and MR have a high degree of accuracy in delineating disk protrusions and showing whether neural tissue is impressed (Fig. 11.4). MR can also show whether annular fibers of the disk are disrupted by noting high signal on T2WIs, which disrupts the annulus. This has been termed a “high intensity zone― (HIZ) (Fig. 11.5). Although CT cannot be used to diagnose annular tears clinicians treat them the same way they treat protrusions (annular fibers intact).
Free
Fragments
A type of disk protrusion that is critical to diagnose is the free fragment or sequestration. Missing free fragments is one of the most common causes of failed back surgery (5). The preoperative diagnosis of a free fragment contraindicates chymopapain, percutaneous discectomy, and, for many surgeons, microdiscectomy. At the very least, the presence of a free fragment means the surgeon must explore more cephalad or caudally during the surgery to remove the free fragment. Because free fragments can be very difficult to diagnose clinically, imaging is critical in the evaluation of the spine for any patient contemplating surgery. At times it can be difficult to ascertain whether a disk that has extruded is still attached to the parent disk or is really “free.― If disk material is above or below the level of the disk space, whether it is attached really does not matter. Chymopapain and percutaneous discectomy would still be contraindicated, and many surgeons would not perform or, at the very least, would modify a microdiscectomy. The key element is recognizing that disk material is present away from the level of the disk space.
796
FIGURE
11.5. Annular
Tear. This sagittal fast spin-echo T2
image shows a focus of increased signal (arrow) in the annulus, which is called a high intensity zone (HIZ) and indicates an annular tear.
Free fragments are diagnosed on CT by the presence of soft tissue density with a higher attenuation value than that of the thecal sac,
797
which is located away from the disk space. A conjoined root (a normal variant of two roots exiting the thecal sac together; seen in 1% to 3% of the population [6]) (Fig. 11.6) or a Tarlov cyst (a normal variant referring to a dilated nerve root sleeve) can have a similar appearance to a free fragment, but these will have P.325 attenuation values similar to that of the thecal sac. A conjoined root has a characteristic appearance on MR (Fig. 11.6C). Free fragments are diagnosed on MR by noting disk material that has moved away from the disk space (Fig. 11.7). Free fragments migrate either cephalad or caudally, with no documented preference (7). It is imperative to obtain contiguous axial images without large skip areas or gaps when imaging with both CT and MR to avoid missing free fragments.
FIGURE
11.6. Conjoined Root and Free Fragment. A. A soft
798
tissue mass is seen in the right L5 lateral recess, which has CT attenuation values that are identical to those of the thecal sac. This is a conjoined nerve root. B . In the same patient, a soft tissue mass is present in the left S1 lateral recess (arrow), which has a density greater than that of the adjacent thecal sac. This is a free fragment. C . An axial proton density–weighted MR image shows a mass in the right lateral recess that had signal characteristics identical to that of the thecal sac on all sequences. This is a conjoined nerve root. A free fragment would not have signal identical to that of the thecal sac on all sequences.
Lateral
Disks
Disks will occasionally protrude in a lateral direction, causing the nerve root that has already exited the central canal to be stretched (Fig. 11.8). Although not common ( Table of Contents > Section III - Pulmonary > Chapter 12 Methods of Examination, Normal Anatomy, and Radiographic Findings of Chest Disease
Chapter
12
Methods of Examination, Normal Anatomy, and Radiographic Findings of Chest Disease Santiago
Miró
Jeffrey S. Klein There are many imaging techniques available to the radiologist for the evaluation of thoracic disease (1). The decision about which imaging procedures to perform depends upon many factors, the most important of which are the availability of various modalities and the type of information sought. Although conventional radiographs of the chest still constitute 25% to 35% of the volume of any general radiology department, there has been a steady decline in favor of CT, despite the considerable increase in radiation to the patient. The recent years have seen near disappearance of diagnostic thoracic vascular interventions, thanks to CT and MR. The recent advent of multichannel, parallel MR imaging might allow for gradual replacement of CT for thoracic vascular diagnostics. Although the imaging algorithm for specific problems may seem relatively straightforward, medical judgment should be preferred. For example, a thin-section CT showing a suspicious solitary pulmonary nodule might be followed directly by a thoracotomy, or rather, in selected patients, by transthoracic needle biopsy. This type of flexible approach will often streamline the diagnostic workup and ultimately
817
lead to better patient care.
Imaging
Modalities
Conventional
Chest
Radiography
Posteroanterior (PA) and lateral chest radiographs are the mainstays of thoracic imaging. Conventional radiographs should be performed as the initial imaging study in all patients with thoracic disease. These films are obtained in most radiology departments on a dedicated chest unit capable of obtaining radiographs with a focusto-film distance of 6 feet, a high kilovoltage-potential (140-kVp) technique, a grid to reduce scatter, and a phototimer to control the length of exposure (2) . The recognition of proper radiographic technique on frontal radiographs involves assessment of four basic features: penetration, rotation, inspiration, and motion. Proper penetration is present when there is faint visualization of the intervertebral disk spaces of the thoracic spine and P.336 discrete branching vessels can be identified through the cardiac shadow and the diaphragms. Rotation is assessed by noting the relationship between a vertical line drawn midway between the medial cortical margins of the clavicular heads and one drawn vertically through the spinous processes of the thoracic vertebrae. Superimposition of these lines (the former in the midline anteriorly and the latter in the midline posteriorly) indicates a properly positioned, nonrotated patient. An appropriate deep inspiration in a normal individual is present when the apex of the right hemidiaphragm is visible below the tenth posterior rib. Finally, the cardiac margin, diaphragm, and pulmonary vessels should be sharply marginated in a completely still patient who has suspended respiration during the radiographic exposure (Fig. 12.1) .
818
FIGURE 12.1. Normal PA (A) and Lateral (B) Radiographs of the Chest.
Portable
Radiography
Portable anteroposterior (AP) radiographs are obtained when patients cannot be safely mobilized (3). Portable radiographs help monitor a patient's cardiopulmonary status; assess the position of various monitoring and life support tubes, lines, and catheters; and detect complications related to the use of these devices. There are technical and patient-related compromises as well as inherent physiologic changes with portable bedside radiography. The limited maximal kilovoltage potential of portable units requires longer exposures to penetrate cardiomediastinal structures, which results in greater motion artifact. Because critically ill patients are difficult to position for portable radiographs, the patient is often rotated. Inaccuracies in directing the x-ray beam perpendicular to the patient lead to kyphotic or lordotic radiographs. The short focus-
819
to-film distance (typically 40 inches) and AP technique result in magnification of intrathoracic structures. For instance, the apparent cardiac diameter increases by 15% to 20%, bringing the upper limit of normal for the cardiothoracic ratio from 50% on a PA radiograph to 57% on an AP. Physiologically, the supine position of critically ill patients elevates the diaphragm, thus compressing lower lobes and decreasing lung volumes. The normal gravitational effect evens out the blood flow between upper and lower zones in supine patients, which makes assessment of pulmonary venous hypertension difficult. The increase in systemic venous return to the heart widening of the upper mediastinum or “vascular gravitational layering of free-flowing fluid may hide Similarly, a pneumothorax may be difficult to detect
produces a pedicle.― The small effusions. because free
intrapleural air rises to a nondependent position, producing a subtle anteromedial or inferior radiolucency. A device called the inclinometer has been developed that accurately records the position of the bedridden patient from supine to completely upright. This device, which clips onto the portable film cassette, gives an accurate estimate of the patient's position at the time of the radiograph, which helps assess the distribution of pulmonary blood flow, pleural effusions, and pneumothorax. P.337
Digital
or
Analog?
The main advantages of digital chest radiography are superior contrast resolution and the availability of the image on any computer monitor through a PACS (Picture Archiving and Communication System). Contrast levels and windows can be adjusted to enhance visualization of various regions in the chest or compensate partly for faulty exposure. Although digital images have poorer spatial resolution than their analog counterparts, these benefits render the system appealing.
Special A lateral
Techniques decubitus radiograph is obtained with a horizontal x-ray
820
beam while the patient lies in the decubitus position. It is used to detect small effusions, to characterize free-flowing effusions on the decubitus side, or to detect a small pneumothorax on the contralateral side. As little as 5 mL of fluid (Fig. 12.2) or 15 mL of air can be demonstrated by this view. Normally, the downside diaphragm assumes a higher position than the upside one. Air trapping can be demonstrated in the dependent lung in patients with a check valve bronchial obstruction who are unable to cooperate for inspiratory/expiratory radiographs or chest fluoroscopy. An expiratory radiographobtained at residual volume (end of maximal forced expiration) can detect focal or diffuse air trapping and eases detection of a small pneumothorax. In the absence of a direct communication between the pleura and the bronchi, the volume of air in the pleural space remains stable, whereas the volume of air in the lung parenchyma decreases. Because the lung is also displaced away from the chest wall, the visceral pleural line becomes more visible.
FIGURE Pleural
12.2. Lateral Decubitus Film for the Detection of Effusion. An upright radiograph (A) in a patient
recovering from pulmonary edema shows blunting of both lateral costophrenic sulci. A left lateral decubitus film (B) demonstrates
821
free-flowing effusion laterally on the down side (solid arrows) and within the lateral aspect of an incomplete oblique fissure (open arrows). Note the clearing of fluid from the right lateral costophrenic sulcus with the patient in the opposite decubitus position, with fluid layering medially along the mediastinal pleural surface.
An apical lordotic view improves visualization of the lung apices, which are obscured on routine PA radiographs by the clavicles and first costochondral junctions. Caudocephalad angulation of the tube projects these anterior bony structures superiorly, providing an unimpeded view of the apices. This view enhances the visualization of middle lobe atelectasis by placing the inferiorly displaced minor fissure in tangent with the x-ray beam and by increasing the AP thickness of the atelectatic middle lobe. Chest
fluoroscopy is used mainly to assess chest dynamics on
patients with suspected diaphragmatic paralysis. Although it has been widely abandoned to the benefit of CT, fluoroscopy can still often bring the same answers as CT at a fraction of the radiation exposure: evaluation of a nodular opacity seen on only one view, evaluation of apparent pseudotumor images caused by vertebral lamina, osteophytes, vertebral transverse processes, healed rib fractures, skin lesions, nipples, or other external objects. P.338
CT and HRCT Thoracic chest CTs can be acquired either in an incremental “stop, acquire, and go― mode, such as for HRCTs, or in a helical mode, whereby acquisition occurs while the patient translates through the gantry on the CT scan table. The latter allows single breath-hold scans with optimal contrast enhancement, without the respiratory misregistration inherent in incremental scanning. Multidetector scanners now allow for full chest coverage with collimation as narrow as 1.0 mm in approximately 20 seconds. Cardiac gating on such scanners can eliminate pulsation artifacts, for example, in the
822
ascending aorta, and also allows for diagnostic evaluation of the heart, at a threefold or fourfold increase in radiation dose (4). Scans without contrast are usually performed for evaluation or follow-up of parenchymal disease. Iodinated contrast material is administered for mediastinal mass or cancer staging evaluation, systemic or pulmonary arterial evaluation, or for cardiac studies. The field of view for image reconstruction is determined by measuring the widest transverse diameter, as seen on the CT scout view. An edge-enhancing computer reconstruction algorithm (“bone― or “sharp― algorithm) improves the spatial resolution of parenchymal structures and is used for all types of thoracic CT scans. Most frequently, the image is reconstructed in a 512 × 512 matrix size. Matrix sizes up to 1,024 × 1,024 are now available, but studies would be needed to assess whether there is any diagnostic benefit to this fourfold increase in image size. Although images can still be filmed using a laser camera, PACS workstation viewing offers the possibility to modify window width (WW) and window level (WL) as needed. Routine settings for CT display of mediastinal structures are WW = 400 and WL = 40 and for the lungs are WW = 1,500 and WL = –700. HRCT technique involves incremental thinly collimated scans (1.0 to 1.5 mm) obtained at evenly spaced intervals through the thorax for the evaluation of diffuse bronchial or parenchymal lung disease. Image acquisition time is limited to minimize the effects of respiratory and cardiac motion. Expiratory HRCT scans are useful for the detection of air trapping in patients with small airways disease. Normal and abnormal HRCT findings are reviewed in Chapter 17. The volume of data of a helical CT is acquired with a thickness (collimation) of 0.5 to 10 mm, and the user can then determine the reconstruction interval, which is chosen according to the amount of desired overlap. For example, a helical scan covering 25 cm with a 2.0-mm collimation can be reconstructed with a 2.0-mm interval, yielding 125 contiguous images with no overlap, but could also be reconstructed at a 1.25-mm interval, yielding 200 images, each of which overlaps the following image by 0.75 mm (4) .
823
The major advantages of CT are its superior contrast resolution and cross-sectional display format. Superior contrast resolution allows for the differentiation of calcium, soft tissue, and fat within lung nodules or mediastinal structures. Intravenous enhancement improves contrast within structures or masses, as well as within blood vessels (e.g., pulmonary emboli, aortic dissection). The cross-sectional display eliminates the superimposition of structures and allows visualization of parenchymal nodules as small as 2 mm.
TABLE 12.1 Indications for Thoracic CT
Indication
Example
Evaluation of an abnormality conventional
identified
Densitometry of a solitary on
radiographs
pulmonary
nodule
Localization and characterization of a hilar or mediastinal
Staging of lung cancer
mass
Assessment of extent of the primary tumor and the relationship of the tumor to the pleura, chest wall, airways, and mediastinum
Detection of hilar and mediastinal lymph node enlargement
Detection of occult
Extrathoracic
pulmonary
a propensity to metastasize to the lung (osteogenic sarcoma, breast and renal cell carcinoma).
metastases
824
malignancies
with
Detection nodes
of
Distinction
mediastinal
of
Lymphoma, Infections
empyema
metastases
Contrast-enhanced C T can
from lung abscess
usually distinguish a peripheral lung abscess from loculated empyema
Detection of central
Angio-C T with high injection
pulmonary
rate, thin collimation, and precise contrast bolus timing
Detection
embolism
and
evaluation
Detection and localization of
of aortic disease:
extent, including aortic branch
aneurysm, intramural
involvement
aortitis,
dissection, hematoma,
trauma
The clinical indications for thoracic CT will vary among institutions. The indications for thoracic CT and HRCT are shown in Tables
12.1
and 12.2.
MR As MR usage expands, studies must be tailored to the individual patient. Morphologic studies usually require only spin-echo T1W and T2W sequences in the axial plane. Coronal and sagittal planes are used in selected cases. Mass evaluation might benefit from fatsuppressed sequences such as STIR, or from gadolinium-enhanced sequences. Angiographic acquisitions are most often performed with GRE volumetric acquisitions. Cardiac sequences benefit from cardiacgated
balanced
steady-state
free
precession P.339
(SSFP) techniques. Respiratory motion is minimized by performing
825
rapid single breath-hold acquisitions or by using respiratory compensation techniques. The latest generation of multichannel scanners with parallel imaging and faster gradients show promise in evaluation of embolic disease, without the radiation cost of multidetector CTs (5) .
TABLE 12.2 Indications for Thoracic HRCT
Indication Solitary
pulmonary
Example Breath-hold
volumetric
exam
nodule
with thin collimation for accurate density determination without respiratory misregistration
Detection of lung disease in a patient with pulmonary symptoms or
Emphysema Extrinsic allergic alveolitis Small airways disease
abnormal pulmonary function studies and a
Immunocompromised
patient
normal or equivocal chest film
Evaluation of diffusely abnormal chest film
A baseline for evaluation of patients with chronic
Cystic fibrosis Sarcoidosis
diffuse infiltrative lung disease for follow-up changes with therapy
Interstitial lung disease Histiocytosis X Adult respiratory distress syndrome
To
Bronchoscopy versus VATS or
determine
approach
826
(type and location) of biopsy
needle
biopsy
VATS = video assisted thoracic surgery
The major advantages of MR are the superior contrast resolution between tumor and fat, the ability to characterize tissues based on T1 and T2 relaxation times, the ability to scan in direct sagittal and coronal planes, and the lack of need for intravenous iodinated contrast (6). In addition, the ability to obtain images along the long axis of the aorta and the advent of cine-MR techniques have made MR the primary modality for the imaging of most congenital and acquired thoracic vascular disorders. Direct coronal scans are of benefit in imaging regions that lie within the axial plane and are therefore difficult to depict on CT. For this reason, superior sulcus tumors, subcarinal and aortopulmonary window lesions, and certain hilar masses are better depicted by MR than CT. MR is superior to CT in the diagnosis of chest wall or mediastinal invasion because of the high contrast between tumor and chest wall fat and musculature and tumor and mediastinal fat, respectively. The characterization of tissues by their T1 and T2 relaxation times allows for the diagnosis of fluid-filled cysts, hemorrhage, and hematoma formation. The ability to distinguish tumor from fibrosis, based on their T1 and T2 relaxation times, has proven particularly useful in the follow-up of patients irradiated for Hodgkin disease. MR is currently unable to distinguish benign masses from malignant masses or lymph nodes.
827
TABLE 12.3 Indications for MR of the Thorax
Evaluation of aortic disease in stable patients: Dissection, aneurysm, intramural hematoma, aortitis Assessment of superior sulcus tumors Evaluation of mediastinal, vascular, and chest wall invasion of lung cancer Staging of lung cancer patients unable to receive intravenous iodinated contrast Evaluation of posterior mediastinal masses
The major disadvantages of thoracic MR scanning are the limited spatial resolution, the inability to detect calcium, and the difficulties in imaging the pulmonary parenchyma. MR is also more timeconsuming and expensive than CT. These factors, along with the ability of CT to provide superior or equivalent information in most situations, have limited the use of thoracic MR for most noncardiovascular thoracic disorders. The primary indications for thoracic MR are listed in Table 12.3.
PET PET utilizing fluorodeoxyglucose (FDG) is an imaging modality based on the metabolic activity of neoplastic and inflammatory tissues and therefore can be considered complementary to the anatomic information provided by chest radiography and CT (1). The role of PET in oncologic diagnosis and staging has developed gradually over the past decade. There is a growing published experience of wholebody PET in the evaluation of patients with malignancy, particularly bronchogenic carcinoma, and of thoracic PET for the evaluation of the solitary pulmonary nodule.
US Transthoracic US is now commonly used for the detection,
828
characterization, and sampling of pleural, peripheral parenchymal, and mediastinal lesions (see Chapter 39). The aspiration of small pleural effusions visualized on real-time US is preferable to blind thoracentesis. Similarly, sampling of visible pleural masses in patients with malignant effusions can diminish the number of negative pleural biopsies. The aspiration of pleural-based masses and abscesses can be safely performed by US-guided needle placement into the lesion through the point of contact between the mass and pleura. Large anterior mediastinal masses that have a broad area of contact with the parasternal chest wall may be biopsied without transgressing the lung. Real-time US can also confirm phrenic nerve paralysis without the use of ionizing radiation. It also easily detects P.340 subpulmonic and subphrenic fluid collections, which may cause diaphragmatic
elevation.
Ventilation/Perfusion
Lung
Scanning
The nuclear medicine examinations utilized in the evaluation of noncardiac
thoracic
disease
are
ventilation/perfusion
(V/Q)
lung
scintigraphy (see Chapter 56) and gallium scintigraphy. V/Q scanning is used almost exclusively for the diagnosis of pulmonary embolism, although quantitative VQ imaging may be useful in the planning of bullectomy, lung volume reduction surgery for emphysema, and lung transplantation. Gallium-67 scanning of the chest is used in the detection of pulmonary infection (e.g., Pneumocystis carinii pneumonia in a patient with a normal radiograph) or inflammation (e.g., disease activity in idiopathic pulmonary fibrosis) and in the evaluation of suspected sarcoidosis. Diagnostic arteriography has mainly been replaced by angio-CT. Pulmonary angiograms are only performed in cases where angio-CT is suboptimal or equivocal. Thanks to the newer scanners and to the improvement of threedimensional (3D) rendering tools, thoracic aortography has also been largely replaced by CT, MR, or US. On occasion, an equivocal
829
diagnosis of an aortic laceration following blunt chest trauma can be resolved with this technique. Inflammatory changes of infectious aortitis are also better imaged with MR or CT. Active bleeding through a bronchial artery is still best addressed by bronchial arteriography, as an active bleeding site is often difficult to pinpoint. When massive or recurrent hemoptysis occur, most commonly from bronchiectasis, neoplasm, or mycetoma, arteriography and embolization can be performed in the same setting. Transthoracic needle biopsy guided by CT, fluoroscopy, or US is a diagnostic technique utilized in selected patients with pulmonary, pleural, or mediastinal lesions (7) . Percutaneous catheter drainage of intrathoracic air or fluid collections, performed by imaging-guided placement of small-bore multihole catheters, is used for the treatment of empyema, pneumothorax, malignant fluid collections (3) .
NORMAL
LUNG
Tracheobronchial
pleural
effusion,
and
other
intrathoracic
ANATOMY Tree (Fig.
12.3)
The trachea is a hollow cylinder composed of a series of C-shaped cartilaginous rings. The rings are completed posteriorly by a flat band of muscle and connective tissue called the posterior tracheal membrane. The tracheal mucosa consists of pseudostratified, ciliated columnar epithelium, which contains scattered neuroendocrine (APUD) cells. The submucosa contains cartilage, smooth muscle, and seromucous glands. The left lateral wall of the distal trachea is indented by the transverse portion of the aortic arch. The trachea is approximately 12 cm long in adults, with an upper limit of normal coronal tracheal diameter of 25 mm in men and 21 mm in women. In cross section, the trachea is oval or horseshoeshaped, with a coronal-to-sagittal diameter ratio of 0.6:1.0. A narrowing of the coronal diameter producing a coronal/sagittal ratio
830
of 1 cm in diameter within this radiolucent region is an accurate indicator of unilateral or bilateral hilar mass. Occasionally, the silhouetting of the anterior wall of the LLL bronchus, recognized as a concave anterior curvilinear structure contiguous with the anterior aspect of the LUL bronchus, allows lateralization of a mass to the left lower hilum (Fig. 12.33). The added opacity of a mass within the normally radiolucent inferior hilar window produces an oval opacity to the composite hilar shadow on lateral radiographs. On a lateral radiograph, enlargement of pulmonary arteries is assessed by measuring the left descending pulmonary artery as it arches over the left mainstem/LUL bronchus at a 2:00 position (Fig. 12.19B) . Helical CT is the most sensitive method of detecting and localizing enlarged hilar (bronchopulmonary) lymph nodes and masses. Although contrast enhancement is almost never necessary to assess mediastinal nodes, it simplifies identification of enlarged vascular structures or nonenhancing hilar nodes (defined as nodes that exceed 10 mm in short-axis diameter) or masses. Hilar masses are seen on axial or coronal spin-echo MR as round masses of low or intermediate signal intensity, in distinction to the signal void of flowing blood within the hilar vessels or of air in the bronchi. Coronal MR may be superior to CT in the detection of enlarged hilar lymph nodes because it displays the hilar vessels, which are oriented in the cephalocaudad direction, in length rather than in cross section. Displacement or distortion of the hilar vessels provides indirect evidence of hilar disease. Tumor invasion of a branch of the pulmonary artery or vein
926
within the hilum produces a filling defect within the vessel on contrast-enhanced CT or intraluminal signal on MR. The density characteristics of hilar masses on CT can help provide important information P.381 for differential diagnosis; for example, a round, cystic hilar mass with imperceptible walls in an asymptomatic young person is typical of a bronchogenic cyst.
FIGURE 12.33. Hilar Mass Within the Inferior Hilar Window. A . A cone-down view of a lateral radiograph in a patient subsequently found to have a plasmacytoma of the left hilum shows a mass (large arrows) within the inferior hilar window obliterating the anterior wall of the left lower lobe bronchus (small arrows). B . A CT scan through the lower hila confirms the presence of a left hilar mass (asterisk) .
927
Enlarged hilar lymph nodes can be detected by CT without the use of intravenous contrast. A detailed knowledge of the normal hilar vascular and bronchial anatomy, as seen on CT, is necessary for the identification of subtle hilar contour abnormalities. In those portions of the hilum where lung directly contacts a wall of a bronchus, thickening or lobulation of the normal thin linear shadow of the bronchial wall indicates hilar abnormality. This is particularly well seen where the RLL and LLL contact the posterior walls of the bronchus intermedius and the LUL bronchus, respectively (Fig. 12.34). Lymph node enlargement in these regions is obscured on frontal radiographs by the overlying cardiac and hilar vascular shadows. CT is more sensitive than plain radiographs or MR for the detection of soft tissue masses within lobar or proximal segmental bronchi. In most patients with an endobronchial mass, a large extraluminal component produces a radiographically visible hilar soft tissue mass and obstructive atelectasis.
FIGURE 12.34. Enlarged Hilar Nodes on CT. Enhanced CT in a patient with biopsy-proven sarcoidosis demonstrates bilateral hilar lymph node enlargement (arrows) .
P.382
928
Enlarged hilar lymph nodes may have different appearances on CT. Enlargement of discrete lymph nodes, most commonly seen in sarcoidosis, appears as multiple distinct round masses (Fig. 12.34) . When tumor or an inflammatory process extends through the nodal capsule to involve contiguous nodes, a single large mass of confluent lymph nodes is produced that may be difficult to distinguish from a primary hilar bronchogenic carcinoma. This latter appearance is most often seen in hilar nodal metastases from small cell carcinoma of the lung or lymphoma (see Fig. 15.10). As in enlargement of mediastinal lymph nodes, the CT density of enlarged hilar nodes can provide clues to the diagnosis (see Table 13.5) . An abnormally small hilum indicates a diminution in the size of the right or left pulmonary artery.
Pleural
Effusion
The radiographic appearance of pleural effusions depends upon the amount of fluid present, the patient's position during the radiographic examination, and the presence or absence of adhesions between the visceral and parietal pleura. Small amounts of pleural fluid initially collect between the lower lobe and diaphragm in a subpulmonic location. As more fluid accumulates, it spills into the posterior and lateral costophrenic sulci. A moderate amount of pleural fluid (>175 mL) in the erect patient will have a characteristic appearance on the frontal radiograph, with a homogeneous lower zone opacity seen in the lateral costophrenic sulcus with a concave interface toward the lung. This concave margin, known as a pleural meniscus, appears higher laterally than medially on frontal radiographs because the lateral aspect of the effusion, which surrounds the costal surface of the lung, is tangent to the frontal xray beam. Similarly, the meniscus of pleural fluid as seen on lateral radiographs peaks anteriorly and posteriorly (Fig. 12.35) (2 7) .
929
FIGURE
12.35. Pleural Effusion on Chest Radiographs.
Posteroanterior (A) and lateral (B) chest radiographs demonstrate the typical meniscoid appearance (arrows) in a patient with a left pleural effusion resulting from mediastinal Hodgkin
lymphoma.
In patients with suspected pleural effusion, a lateral decubitus film with the affected side down is the most sensitive technique to detect small amounts of fluid. With this technique, pleural fluid collections as small as 5 mL may be seen layering between the lung and lateral chest wall. While a moderate-size, free-flowing collection should be obvious on upright radiographs, a large pleural effusion can cause passive atelectasis of the entire lung, producing an opaque hemithorax. It may be difficult to distinguish the latter condition from collapse of an entire lung. While produce contralateral mediastinal shift, pleural effusion will show shift toward patients, CT or US may be necessary collapsed lung.
a massive effusion should a collapsed lung without the opaque side. In some to distinguish pleural fluid from
CT is quite sensitive in the detection of free pleural fluid. On axial
930
scans, pleural fluid layers posteriorly with a characteristic meniscoid appearance and has a CT attenuation value of 0 to 20 H. Small effusions may be difficult to P.383 differentiate from pleural thickening, fibrosis, or dependent atelectasis, and decubitus scans are useful in making this distinction. The pleural and peritoneal spaces are oriented in the axial plane at the level of the diaphragm. This may cause some difficulty in localizing the fluid to one or both spaces. Fluid in either the pleural or peritoneal space can displace the liver and spleen medially, away from the chest wall. A key to distinguishing ascites from pleural fluid on axial CT scans is to observe the relationship of the fluid to the diaphragmatic crus. Pleural fluid in the posterior costophrenic sulcus will lie posteromedial to the diaphragm and displace the crus laterally. In contrast, peritoneal fluid lies within the confines of the diaphragm and therefore will displace the crus medially. Another useful distinguishing feature is the quality of the interface of the fluid with the liver or spleen. Intraperitoneal fluid will show a distinct, sharp interface with the liver and spleen as it directly contacts these organs, whereas pleural effusions will have a hazy, indistinct interface with these viscera because of the interposed hemidiaphragms. Because the peritoneal space does not extend posterior to the bare area of the liver, right-sided fluid extending posteromedially must be pleural. A large effusion will allow the inferior edge of the adjacent atelectatic lower lobe to float in the fluid, creating a curvilinear opacity that can be misinterpreted as the diaphragm separating pleural fluid from ascites. This “pseudodiaphragm― is recognized as a broad band that does not extend far laterally or anteriorly and is contiguous superiorly with an atelectatic lung containing air bronchograms (Fig. 12.36). US is particularly useful in detecting free flowing pleural effusions, which are usually seen as anechoic collections at the base of the pleural space surrounding atelectatic lung (see Chapter 39) . Pleural fluid may become loculated between the pleural layers to produce an appearance indistinguishable from that of a pleural mass. Fluid loculated within the costal pleural layers appears as a vertically
931
oriented elliptical opacity with a broad area of contact with the chest wall, producing a sharp, convex interface with the lung when viewed in tangent. CT is commonly utilized to detect and localize loculated pleural fluid collections. The characteristic finding is a sharply marginated lenticular mass of fluid attenuation conforming to the concavity of the chest wall that forms obtuse angles at its edges and compresses and displaces the subjacent lung. Multiple fluid locules can mimic pleural metastases or malignant mesothelioma radiographically; CT or US can confirm the fluid characteristics of these
pleural
“masses.―
Pleural fluid may extend into the interlobar fissures, producing characteristic findings. Free fluid within the minor fissure is usually seen as smooth, symmetric thickening on a frontal radiograph. Fluid within the major fissure is normally not visible on frontal radiographs, as the fissures are viewed en face. An exception is fluid extending into the lateral aspect of an incomplete major fissure, which produces a curvilinear density extending from the inferolateral to the superomedial aspect of the lung. Fluid loculated between the leaves of visceral pleura within an interlobar fissure results in an elliptic opacity oriented along the length of the fissure. These loculated collections of pleural fluid are termed “pseudotumors― and are most often seen within the minor fissure on frontal radiographs in patients with congestive heart failure. The tendency for these opacities to disappear rapidly with diuresis has led to the term “vanishing lung tumor.― Although a characteristic appearance on plain radiographs is usually sufficient for diagnosis, the CT demonstration of a localized fluid collection in the expected location of the major or minor fissure is confirmatory. An uncommon appearance of pleural effusion is seen when fluid accumulates between the lower lobe and diaphragm and is termed a subpulmonic effusion. While small amounts of pleural fluid normally accumulate in this location, it is uncommon for larger effusions to remain subpulmonic without spilling into the posterior and lateral costophrenic sulci. A subpulmonic effusion may be difficult to appreciate on upright chest radiographs, because the fluid collection mimics an elevated hemidiaphragm. Clues to its presence on frontal
932
radiographs include: apparent and new elevation of the diaphragm, lateral peaking of the hemidiaphragm that is accentuated on expiration, a minor fissure that is close to the diaphragm (right-sided effusions), and an increased separation of the gastric air bubble from the base of the lung (left-sided effusions). Despite the atypical subpulmonic accumulation of fluid with the patient upright, the effusion will layer dependently on lateral decubitus radiographs (Fig. 12.37) . The radiographic detection of pleural effusion in the supine patient can be difficult because fluid accumulates in a dependent location posteriorly. The most common finding is a hazy opacification of the affected hemithorax with obscuration of the hemidiaphragm and blunting of the lateral costophrenic angle. Fluid extending over the apex of the lung may produce a soft tissue cap with a concave interface inferiorly, while medial fluid may cause an apparent mediastinal widening.
Pneumothorax The classic radiographic finding of pneumothorax on upright chest films is visualization of the visceral pleura as a curvilinear line that parallels the chest wall, separating the partially collapsed lung centrally from pleural air peripherally (Fig. 12.38). An expiratory radiograph aids in the detection of a small pneumothorax by increasing the volume of intrapleural air relative to lung, thereby displacing the visceral pleural reflection away from the chest wall and by exaggerating the differences in density of pneumothorax (black) to lung (gray) at the end of expiration. In a small P.384 percentage of patients, a pneumothorax will be visible only on a lateral or decubitus film or a frontal radiograph obtained in full inspiration. This suggests that when there is a strong clinical suspicion of pneumothorax and the frontal expiratory radiograph is normal, a lateral or inspiratory film may be beneficial for proper diagnosis.
933
934
FIGURE 12.36. Subpulmonic Pleural Effusion on CT. A. A CT scan through the lower chest shows fluid surrounding an enhancing broad curvilinear structure (asterisks). The fluid creates an ill-defined interface with the liver (arrows). B . A scan 1 cm more cephalad shows that the curvilinear density represents the tip of an atelectatic right lower lobe containing air bronchograms (arrows). C . More inferiorly, the crus of the diaphragm (dotted structure) is displaced laterally by posteromedial pleural fluid.
The detection of a pneumothorax is difficult when chest films are obtained in the supine position. Approximately 30% of pneumothoraces imaged on supine radiographs go undetected. Because many portable radiographs are obtained with the patient supine, the recognition of a pneumothorax on a supine film is particularly important in the critically ill patient, who is at high risk from iatrogenic trauma or barotrauma. In a supine patient, the most nondependent portion of the pleural space is anterior or anteromedial. Small pneumothoraces will initially collect in these regions and will fail to produce a visible pleural line. The affected hemithorax may appear hyperlucent. Anteromedial air may sharpen the borders of mediastinal soft tissue structures, resulting in improved visualization of the P.385 cardiac margin and aortic knob. The lateral costophrenic sulcus may appear abnormally deep and hyperlucent, a finding known as the “deep sulcus― sign. Visualization of the anterior costophrenic sulcus owing to air anteriorly and inferiorly produces the “double diaphragm― sign, as the dome and anterior portions of the diaphragm are outlined by lung and pleural air, respectively. When an anterior pneumothorax is suspected on a supine radiograph, an upright film, lateral decubitus film with the affected side up, or CT scan should be obtained (Fig. 12.39) .
935
FIGURE
12.37. Bilateral Subpulmonic Pleural Effusions. A.
An upright posteroanterior radiograph in a 41-year-old woman with ascites demonstrates apparent elevation of both hemidiaphragms. Right (B) and left (C) decubitus films demonstrate dependent layering of the subpulmonic pleural fluid (arrows) .
Subpulmonic pneumothoraces are rare. Radiographically, a localized area of hyperlucency is seen inferiorly, with the visceral pleural line paralleling the hemidiaphragm. Loculated pneumothoraces develop as the result of adhesions between visceral and parietal pleura and may be found anywhere in the pleural space. CT is often necessary for diagnosis.
936
Several entities produce a curvilinear line or interface or hyperlucency on chest radiographs and must be distinguished from a pneumothorax. Skin folds resulting from the compression of redundant skin by the radiographic cassette can produce a curvilinear interface that simulates the visceral pleural line. A skin fold produces an edge or interface with atmospheric air, in distinction to the visceral pleural line seen in a pneumothorax. The interface produced by a skin fold rarely continues over the lung apex and is often seen to extend beyond the chest wall. Pulmonary vascular opacities may be followed peripheral to the skin fold interface. Bullae may simulate pneumothorax by producing localized or unilateral hyperlucency. They are marginated by thin curvilinear walls that are concave rather than convex to the chest wall. The distinction of pneumothorax from bullous disease may be difficult but is usually evident by the clinical presentation. However, since this distinction has important therapeutic implications, certain patients may require CT. CT is more sensitive than conventional radiographs in the detection of pneumothorax because of its cross-sectional nature and superior contrast resolution. The CT demonstration of linear parenchymal bands of tissue traversing large avascular areas helps distinguish bullae from loculated pneumothoraces. CT may be used to P.386 detect and drain loculated pneumothoraces in critically ill patients.
937
FIGURE 12.38. The Visceral Pleural Line in Pneumothorax. A cone-down view of an upright posteroanterior radiograph in a patient with a spontaneous pneumothorax demonstrates a curvilinear visceral pleural line (solid arrows) separating the lung medially from the chest wall laterally. Note the presence of thinwalled cysts (open arrows) from Coccidioides infection, which are most likely responsible for the pneumothorax.
Localized
Pleural
Thickening
938
Localized pleural thickening is seen as a flat, smooth, slightly raised soft tissue opacity extending over one or two intercostal spaces that displaces the lung from the innermost cortical margin of the ribs when viewed in tangent. Localized pleural thickening viewed en face is usually undetectable radiographically because the lesion does not significantly attenuate the x-ray beam and does not present a raised edge to be recognized as a distinct opacity. An exception is the presence of pleural calcification, which can usually be recognized as discrete thin linear or curvilinear calcific opacities paralleling the inner surface of the ribs when viewed end-on or as geographic areas of increased density with round or lobulated borders when viewed en face. Focal areas of pleural fibrosis are best appreciated on conventional and high-resolution CT scans, where they are easily distinguished from deposits of subpleural fat by their density. There are two additional radiographic findings that mimic the appearance of focal pleural thickening. The apical cap is a curvilinear subpleural opacity Table of Contents > Section III - Pulmonary > Chapter 13 Mediastinum and Hila
Chapter
13
Mediastinum
and
Hila
Jeffrey S. Klein This chapter will review the radiologic approach to mediastinal masses, diffuse mediastinal disease, and hilar abnormalities.
MEDIASTINAL Localized
mediastinal
MASSES abnormalities
are
common
diagnostic
challenges for the radiologist. Patients with mediastinal masses tend to present in one of two fashions: with symptoms related to local mass effect or invasion of adjacent mediastinal structures (stridor in a patient with thyroid goiter), or incidentally with an abnormality on a routine chest radiograph. Occasionally, a mediastinal mass is discovered in the course of an evaluation for known malignancy (e.g., a patient with non-Hodgkin lymphoma) or for a condition such as myasthenia gravis, in which there is an association with thymoma. Multidetector-row CT (MDCT) and MR are the primary cross-sectional modalities used to evaluate mediastinal masses, with PET useful to assess response of mediastinal tumors to therapy, particularly lymphoma, and to distinguish residual or recurrent tumor from fibrosis (Table 13.1) . For the purposes of the following discussion, the divided into superior (thoracic inlet) and inferior the inferior mediastinum subdivided into anterior, posterior compartments, as described in Chapter
946
mediastinum is components, with middle, and 12.
Thoracic
Inlet
Masses
The thoracic inlet is the region of the upper thorax marginated by the first rib and represents the junction between the neck and thorax. Masses in this region commonly present as neck masses or with symptoms of upper airway obstruction resulting from tracheal compression. Thyroid masses, lymphomatous nodes, and lymphangiomas are the most common thoracic inlet masses (Table 13.2) .
Thyroid
Masses
In a small percentage of patients with a cervical thyroid goiter, a thyroid carcinoma, or an enlarged gland from thyroiditis, extension of the thyroid through the thoracic inlet into the superior mediastinum may occur. These lesions are usually discovered as incidental findings on chest radiographs; a minority of patients will present with complaints of dyspnea or dysphagia as a result of tracheal or esophageal compression by the mass. Thyroid goiters arising from the lower pole of the thyroid or the thyroid isthmus can enter the superior mediastinum anterior to the trachea (80% of cases) or to the right and posterolateral to the trachea (20% of cases). On chest radiographs, an anterosuperior mediastinal mass typically deviates the trachea laterally and either posteriorly (anterior masses) or anteriorly (posterior masses). Coarse, clumped calcifications are common in thyroid goiters. Radioiodine studies should be performed as the initial imaging procedure, although falsenegative results do occur. CT usually shows characteristic findings: (1) well-defined margins, (2) continuity of the mass with the cervical thyroid, (3) coarse calcifications, (4) cystic or necrotic areas, (5) baseline high CT attenuation (because of intrinsic iodine content), and (6) intense enhancement (>25 H) as a result of the hypervascularity of most thyroid masses and prolonged enhancement (resulting from active uptake of iodine from contrast media) following
intravenous
947
P.390 contrast administration (Fig. 13.1) (1). MR is useful in depicting the longitudinal extension of thyroid goiters without the use of intravenous
contrast.
TABLE 13.1 Utility of MDCT, MR, and PET in the Evaluation of Mediastinal Masses
Indication for Study
Modalities
Confirming the presence of a
MDCT = M R
mass versus tortuous vascular structures
Localization of mass to
MDCT = M R
anterior, middle, or posterior compartment
Suspected aneurysm vascular anomaly
Tissue
or
characterization
MDCT = M R
of
mass
Detection of fluid
MDCT = M R = US (for anterior masses or periesophageal masses)
Detection
CT
of
calcium
Distinction of tumor from fibrosis
PET>M R>C T
948
Relationship structures
to
adjacent
Vascular
invasion
MDCT = M R
Tracheal
involvement
MDCT > M R
Involvement of spinal canal
Thoracic
inlet
Contraindication
M R > MDCT
lesions
M R = MDCT
to
M R > MDCT
iodinated
contrast
Percutaneous biopsy mediastinal mass
of
CT US for anterior mediastinal masses
MDCT,
multidetector-row C T.
Parathyroid
Masses
In approximately 2% of patients, the parathyroid glands fail to separate from the thymus in the neck and descend with the gland into the anterosuperior mediastinum. These glands can be found near the thoracic inlet in or about the thymus. This becomes important in the small percentage of patients with persistent clinical and biochemical evidence of hyperparathyroidism following routine neck exploration and parathyroidectomy. Most of these ectopic parathyroidlesions are small (1.0 cm in their short axis diameter. Although CT is unable to distinguish between benign inflammatory nodes and those involved by malignancy based upon size criteria alone, CT can provide useful information about the internal density of the nodes (Table 13.5) .
TABLE 13.5 Density of Mediastinal/Hilar Nodes on CT
Calcification
Central
Mycobacteria Fungus
Peripheral
Silicosis
(eggshell)
Sarcoidosis
Hypervascular
Carcinoid tumor/small cell carcinoma Kaposi sarcoma Metastases Renal cell carcinoma Thyroid Castleman
Necrosis
carcinoma disease
Mycobacteria Fungus Metastases
968
Squamous cell Seminoma Lymphoma
carcinoma
A standardized classification system for hilar and mediastinal lymph nodes has recently been advanced by the American Thoracic Society (Fig. 13.8) (7). This scheme correlates with easily identifiable CT and anatomic landmarks and is most important when reporting lymph node enlargement in patients with bronchogenic carcinoma. MR is as sensitive as CT in detecting enlarged mediastinal lymph nodes. Advantages of MR include the absence of iodinated contrast, easy distinction between vascular and soft tissue structures, exquisite contrast resolution between mediastinal nodes and fat on T1W sequences, and the ability to image in the direct coronal or sagittal plane. The latter feature is an advantage in those mediastinal regions that parallel the axial plane (subcarinal space, aortopulmonary window) and therefore tend to suffer from partial volume averaging effects on CT. The major disadvantages of MR at present are the inability to detect nodal calcification and limited spatial resolution; the latter can result in an inability to distinguish between a group of normal size nodes and a single enlarged node, thereby leading to false-positive results. In addition to the detection and characterization of enlarged mediastinal nodes, CT can help guide diagnostic nodal tissue sampling. This is usually most helpful in the setting of suspected bronchogenic carcinoma, where accurate staging of mediastinal nodal disease is important for prognostic purposes and treatment planning. The recognition of enlarged subcarinal or pretracheal nodes on CT may suggest biopsy via transcarinal Wang needle or mediastinoscopy, respectively. As mentioned above, mediastinal lymph node enlargement is common in Hodgkin disease and NHL. Lymphoma accounts for 20% of all mediastinal neoplasms in adults, and most patients with intrathoracic lymphoma have concomitant extrathoracic disease. In most patients, the nodal enlargement is bilateral but asymmetric. Nodular sclerosing
969
Hodgkin disease commonly results in lymph node enlargement, predominantly within the anterior mediastinum and thymus. Isolated posterior nodal enlargement is usually seen only in patients with NHL. Leukemia, particularly the T-lymphocytic varieties, can cause intrathoracic lymph node enlargement. The lymph node enlargement is usually confined to the middle mediastinal and hilar nodes. The most common source of metastases to middle mediastinal nodes is bronchogenic carcinoma. In the majority of patients, symptoms or plain radiographic findings suggest the presence of a primary tumor in the lung. In a small percentage of patients, particularly those with small cell carcinoma, the primary carcinoma may be inconspicuous or invisible on plain radiographs, with nodal metastases being the only visible abnormality. Lymph node enlargement is often unilateral on the side of the visible pulmonary or hilar abnormality. Paratracheal and aorticopulmonary nodes are most commonly involved. Since the accuracy of CT in predicting the presence or absence of mediastinal lymph node metastases is approximately 60% to 70%, PET—and in particular integrated CT/PET—should be performed in most patients with bronchogenic carcinoma. A more thorough discussion of mediastinal nodal involvement in bronchogenic carcinoma may be found in Chapter 15. Lymph node metastases from extrathoracic malignancies can result in mediastinal node enlargement, either with or without concomitant pulmonary metastases. These mediastinal nodal metastases may result from inferior extension of neck masses (thyroid carcinoma, head and neck tumors); extension along lymphatic channels from below the diaphragm (testicular or renal cell carcinoma, GI malignancies); or hematogenous extension (breast carcinoma, melanoma, Kaposi sarcoma) (8) . Mediastinal lymph node enlargement is very common in patients with sarcoidosis, occurring in 60% to 90% of patients at some stage of their disease. Nodal enlargement is typically bilateral and symmetric and involves the hila as well as the mediastinum (Fig. 13.9); this usually
970
P.399 allows for differentiation of sarcoidosis from lymphoma and metastatic disease. In sarcoidosis, the enlarged nodes produce a lobulated appearance on chest radiographs and CT, because the enlarged nodes do not coalesce. This is in contrast to lymphoma and nodal metastases, in which the intranodal tumor extends through the nodal capsule to form conglomerate enlarged nodal masses. Right and left paratracheal lymph nodes are typically involved; anterior or posterior mediastinal nodal enlargement has been described with greater frequency recently, probably as a P.400 result of the improved sensitivity of CT for detecting nodal involvement in these regions.
971
FIGURE 13.8. American Thoracic Society Nodal Stations. Ao, aorta; PA, pulmonary artery. From Mountain CF, Dresler CM. Regional lymph node classification for lung cancer staging. Chest
972
1997;111:1718–1723;
reprinted
with
permission.
FIGURE 13.9. Lymphadenopathy in Sarcoidosis. Posteroanterior radiograph in a 56-year-old woman with sarcoidosis reveals discrete hilar, paratracheal, aortopulmonary window lymphadenopathy.
and
A variety of infections, most commonly histoplasmosis, coccidioidomycosis, cryptococcosis, and tuberculosis, can cause mediastinal nodal enlargement (Fig. 13.10). Typically these patients have parenchymal opacities on chest radiographs, but isolated lymph node enlargement may be seen, particularly in children and young adults. Bacterial infections such as anthrax, bubonic plague, and tularemia are uncommon causes of lymph node enlargement. Typically, there will be symptoms and signs of acute infection, and chest radiographs will show evidence of pneumonia. Bacterial lung abscesses also may be associated with reactive lymph node
973
enlargement. Hilar and mediastinal lymph nodes may be enlarged in patients with measles pneumonia and infectious mononucleosis.
FIGURE
13.10. Tuberculous
Lymphadenopathy. Contrast-
enhanced CT at the level of the tracheal carina demonstrates enlarged precarinal and left peribronchial lymph nodes with central necrosis and peripheral enhancement. Material obtained by
mediastinoscopy
revealed Mycobacterium
tuberculosis.
Angiofollicular lymph node hyperplasia (Castleman disease) is characterized by enlargement of hilar and mediastinal lymph nodes, predominantly in the middle and posterior mediastinal compartments. In the more common hyaline vascular type, the disease is localized to one lymph node region and presents as an asymptomatic mediastinal soft tissue mass. Histologically, there is replacement of normal nodal architecture with multiple germinal centers and multiple small vessels with hyalinized walls that course perpendicularly toward the germinal centers to give a characteristic “lollipop―
974
appearance on light microscopy. The vascular nature of these masses accounts for the intense enhancement seen on contrast-enhanced CT or angiography. Calcification within these masses has been described. These lesions are cured by resection. Angioimmunoblastic lymphadenopathy is a rare disorder seen in older adults; it is characterized by constitutional symptoms, lymphadenopathy, hepatosplenomegaly, and skin rash. Hemolytic anemia and hypergammaglobulinemia may be seen. Histologically, the enlarged nodes contain a chronic inflammatory infiltrate and are hypervascular. Chest radiographs and CT show hilar and mediastinal lymph node enlargement that are indistinguishable from other etiologies. As with Castleman disease, the vascular nature of the involved lymph nodes accounts for the contrast enhancement seen on CT. These patients manifest signs of immunodeficiency similar to those associated with AIDS, with one third developing high-grade lymphoma and many succumbing to opportunistic infections such as Pneumocystis
carinii pneumonia and cytomegalovirus inclusion
disease. Foregut and mesothelial cysts are common mediastinal lesions that typically present as asymptomatic masses on routine chest radiographs in young adults. CT and MR show findings characteristic of the cystic nature of these lesions. Congenital bronchogenic cysts result from anomalous budding of the tracheobronchial tree during development. To be characterized as bronchogenic in origin, the wall of the cyst must be lined by a respiratory epithelium with pseudostratified columnar cells and contain seromucous glands; some may contain cartilage and smooth muscle within their walls. It is often difficult to distinguish between bronchogenic and enteric cysts based on their location and pathologic appearance; the term foregut cyst has been used to describe those lesions that cannot be specifically characterized. The majority of bronchogenic cysts P.401 (80% to 90%) arise within the mediastinum in the vicinity of the tracheal carina. Most mediastinal lesions are asymptomatic;
975
occasionally, compression of the tracheobronchial tree or esophagus may produce dyspnea, wheezing, or dysphagia. Rarely, mediastinal cysts become secondarily infected after communication with the airway or esophagus, or they cause symptomatic compression after rapid enlargement following hemorrhage. Bronchogenic cysts are seen as soft tissue masses in the subcarinal or right paratracheal space on frontal chest radiographs; less common sites of involvement include the hilum, posterior mediastinum, and periesophageal region. They appear as a single smooth, round, or elliptic mass; a minority are lobulated in contour. CT is the method of choice for the diagnosis of a mediastinal cyst. If a well-defined, thin-walled mass of fluid density (0 to 10 H) is seen that fails to enhance following intravenous contrast administration, it can be assumed to represent a benign cyst (Fig. 13.11) (9). High CT numbers (>40 H) suggesting a solid mass can be seen when the cyst is filled with mucoid material, milk of calcium, or blood. Calcification of the cyst wall has been described but is uncommon. MR shows characteristic low signal intensity on T1WIs and high signal intensity on T2WIs. The presence of proteinaceous material within the cyst will shorten T1 relaxation times, yielding high signal intensity on T1WIs. In many patients, resection is required for definitive diagnosis. Both transbronchoscopic and percutaneous needle aspiration and drainage have been used successfully for the diagnosis and treatment of these lesions.
976
FIGURE
13.11. Bronchogenic
Cyst.
Unenhanced (A) and
enhanced (B) CT scans in a 38-year-old man demonstrate a smooth, low-attenuation paratracheal mass (arrows) that fails to enhance, consistent with a bronchogenic cyst.
Pericardial
cysts arise from the parietal pericardium and contain clear
serous fluid surrounded by a layer of mesothelial cells. Most often, they arise in the anterior cardiophrenic angles, with right-sided lesions being twice as common as left-sided lesions; approximately 20% arise more superiorly within the mediastinum. These lesions usually present as incidental asymptomatic round or oval masses in the cardiophrenic angle (Fig. 13.12). Their pliable nature can be demonstrated with a change in patient position. CT typically shows a unilocular cystic mass P.402 adjacent to the heart; MR or US via a subxiphoid approach shows findings characteristic of a simple cyst. As with bronchogenic cysts, there have been reports of cysts with high attenuation on CT that on resection are found to be filled with proteinaceous or mucoid material.
977
FIGURE 13.12. Pericardial Cyst. Enhanced CT scan through heart shows a smooth, sharply marginated, low-attenuation mass (arrow) in the right cardiophrenic angle, consistent with a pericardial cyst.
Tracheal and central bronchial masses commonly produce upper airway symptoms with obstructive pneumonitis and atelectasis and rarely present as asymptomatic mediastinal masses. Occasionally, central airway masses present as radiographic abnormalities when they distort the tracheal air column or mediastinal contour. These masses are discussed in Chapter 18. Diaphragmatic hernias, which may present as pericardiac masses, are discussed in Chapter 19.
Vascular
Lesions
Congenital or acquired anomalies of the heart and great vessels are common middle mediastinal masses and are discussed in Chapter 14.
978
Neurogenic
Lesions
Rarely, a neurofibroma arising from the phrenic nerve may present as a middle mediastinal juxtacardiac mass.
Posterior Neurogenic
Mediastinal
Masses
Tumors (Table 13.6). Posterior mediastinal masses
arising from neural elements are classified by their tissue of origin. Three groups have been recognized: (1) tumors arising from intercostal nerves (neurofibroma, schwannoma); (2) sympathetic ganglia (ganglioneuroma, ganglioneuroblastoma, and neuroblastoma); and (3) paraganglionic cells (chemodectoma, pheochromocytoma). Tumors in each of these three groups may be benign or malignant neoplasms (5). Although neurogenic tumors can occur at any age, they are most common in young patients. Neuroblastoma and ganglioneuroma are most common in children, whereas neurofibroma and schwannoma affect adults more frequently. Histologically, both neurofibroma and schwannoma are comprised of spindle cells that arise from the Schwann cell. While neurofibroma is an encapsulated tumor that contains interspersed neurons, schwannoma is not encapsulated and contains no neuronal elements. Both tumors are more common in patients with neurofibromatosis. Multiple lesions in the mediastinum, particularly bilateral apicoposterior
masses,
are
virtually
diagnostic
of
neurofibromatosis.
A small percentage of schwannomas (10%) are locally invasive (malignant schwannoma). Radiographically, intercostal nerve tumors appear as round or oval paravertebral soft tissue masses. CT shows a smooth or lobulated paraspinal soft tissue mass, which may erode the adjacent vertebral body or rib. CT demonstration of tumor extension from the paravertebral space into the spinal canal via an enlarged intervertebral foramen is characteristic of a “dumbbell― neurofibroma. MR is the modality of choice for imaging a suspected neurofibroma. In addition to the occasional demonstration of both
979
intra–and extra–spinal canal components, MR shows typical high signal intensity on T2WIs.
of
neurofibromas
TABLE 13.6 Posterior Mediastinal Masses
Neurogenic
tumors
Peripheral
(intercostal)
nerves
Neurofibroma Schwannoma Sympathetic ganglia Ganglioneuroma Ganglioneuroblastoma Neuroblastoma Paraganglion
cells
Chemodectoma Pheochromocytoma
Esophageal
lesions
Duplication (enteric) Diverticulum Neoplasm Leiomyoma Squamous cell Esophageal
cyst
carcinoma
dilatation
Achalasia Scleroderma Peptic stricture Carcinoma Paraesophageal varices Hiatal hernia Sliding Paraesophageal
Foregut
cysts
Enteric Neurenteric
980
Vertebral
lesion
Trauma Paraspinal Infection
hematoma
Paraspinal abscess Tuberculosis Staphylococcus Tumor Metastases (bronchogenic, renal cell carcinoma)
breast,
Multiple myeloma Lymphoma Degenerative disease (osteophytosis) Extramedullary hematopoiesis
Lateral thoracic meningocele
Pancreatic pseudocyst
Tumors that arise from the sympathetic ganglia represent a continuum from the histologically benign ganglioneuroma found in adolescents and young adults to the highly malignant neuroblastoma seen almost exclusively in children under the age of 5. These tumors generally present as elongated, vertically oriented paravertebral soft tissue masses with a broad area of contact with the posterior mediastinum (Figs. 13.13, 13.14). These findings may help distinguish these lesions from neurofibromas, which P.403 P.404 usually maintain an acute angle with the vertebral column and posterior mediastinum and therefore tend to show sharp superior and inferior margins on lateral chest radiographs. Large masses may erode vertebral bodies or ribs. Calcification, seen in up to 25% of cases, is a helpful diagnostic feature of these tumors but does not
981
help distinguish benign from malignant neoplasms. Because these tumors often produce catecholamines, urinary levels of vanillylmandelic acid or metanephrines, which are byproducts of catecholamine metabolism, may be elevated. Prognosis depends upon the histologic features of the tumor and the patient's age and extent of disease at the time of diagnosis.
FIGURE
13.13. Neurofibroma.
A. Frontal chest radiograph
shows a left upper mediastinal mass (arrow). B . Contrastenhanced CT confirms the presence of a left paravertebral soft tissue mass (arrow). Surgical resection confirmed a neurofibroma.
982
FIGURE
13.14. Ganglioneuroma.
A.
Posteroanterior
radiograph in a 15-year-old woman reveals an oval, vertically oriented, right-sided mediastinal mass (arrows). B . Contrastenhanced CT shows a low-attenuation posterior mediastinal mass (arrow) with calcification. This was surgically proven to be ganglioneuroma.
Paragangliomas are tumors that arise in the aorticopulmonary paraganglia of the middle mediastinum or the aorticosympathetic ganglia of the posterior mediastinum. They are divided into nonfunctioning neoplasms (chemodectomas), which occur almost exclusively in or about the aortopulmonary window, and functioning neoplasms (pheochromocytomas), which are found in the posterior sympathetic chain or in or about the heart or pericardium. Approximately 2% of all pheochromocytomas arise in the mediastinum. The posterior mediastinum is the site of fewer than 25% of mediastinal paragangliomas, with the majority arising in the anterior or middle mediastinum. Radiographically, these tumors are indistinguishable from other neurogenic tumors. However, most patients have hypertension and biochemical evidence of excess
983
catecholamine production. CT and angiography demonstrate hypervascular masses; radionuclide iodine-131-metaiodobenzylguanidine (MIBG) scanning is diagnostic in functioning tumors.
Esophageal
Lesions
Because most of the intrathoracic esophagus is intimately associated with the thoracic spine and descending thoracic aorta, lesions in the middle or distal third of the esophagus may present as posterior mediastinal masses. Common presenting symptoms include dysphagia and aspiration pneumonia, although many patients are asymptomatic. The majority of esophageal neoplasms, excluding lesions that arise at the esophagogastric junction, are squamous cell carcinomas. Unlike benign neoplasms of the posterior mediastinum, these lesions, when seen on chest radiographs, are rarely asymptomatic. Typically these patients have a history of dysphagia and significant weight loss. Difficulty in detecting asymptomatic lesions and the absence of a serosa account for the advanced stage of most esophageal carcinoma at presentation and a 5-year survival rate of less than 20%. Most patients with esophageal carcinoma have abnormal plain radiographic findings, including an abnormal azygoesophageal interface, widening of the mediastinum (resulting from the tumor itself or a dilated esophagus proximal to the obstructing lesion), abnormal thickening of the tracheoesophageal stripe, and tracheal deviation and compression. The diagnosis is usually made on barium esophagram and confirmed by endoscopic biopsy. CT scanning has proved accurate for staging esophageal carcinoma: findings include an intraluminal mass; thickening of the esophageal wall; loss of fat planes between the esophagus and adjacent mediastinal structures (usually the trachea, with upper esophageal lesions, and the descending aorta, with lower esophageal lesions); and evidence of nodal and distant metastases. Several
benign
esophageal
neoplasms,
including
leiomyoma,
fibroma,
and lipoma, can present as smooth, solitary mediastinal masses
984
projecting laterally from the posterior mediastinum on frontal chest radiographs. They generally involve the lower third of the esophagus from the level of the subcarinal space to the esophageal hiatus. Initial evaluation is with barium studies, which show a smooth, broad-based mass forming obtuse margins with the esophageal wall. CT demonstrates a smooth, well-defined soft tissue mass adjacent to the esophagus without obstruction. The absence of esophageal dilatation above the mass helps distinguish benign tumors from carcinoma. Pulsion diverticula arising at the cervicothoracic esophageal junction or distal esophagus are false diverticula representing mucosal outpouchings through defects in the muscular layer of the esophagus. A large proximal pulsion diverticulum (Zenker) may extend through the thoracic inlet and appear as a retroesophageal superior mediastinal mass containing an air–fluid level on upright chest radiographs. A distal pulsion diverticulum appears as a juxtadiaphragmatic mass with an air–fluid level projecting to the right of midline. Barium swallow is diagnostic. A dilated esophagus resulting from functional (achalasia, scleroderma) or anatomic (stricture, carcinoma) obstruction may produce a mass that courses vertically over the length of the mediastinum, projecting toward the right side on frontal chest radiographs. An air–fluid level on upright films is usually present. A completely air-filled, dilated esophagus appears as a thin curvilinear line along the medial right thorax, because the right lateral wall of the esophagus is outlined by intraluminal air medially and the right lung laterally. Barium study or CT will confirm the diagnosis of a dilated esophagus; determination of the cause of obstruction often requires endoscopy or esophageal manometry. Esophageal varices may produce a round or lobulated retrocardiac mass in patients with portal hypertension. The diagnosis is usually made by endoscopic recognition of submucosal varices involving the distal esophagus. The varices are readily recognized on contrast CT, MR, or portal venography. A common cause of a mass in the posteroinferior mediastinum is a
985
hiatal hernia. This results from a separation of the superior margins of the diaphragmatic crura P.405 and stretching of the phrenicoesophageal ligament. The stomach is by far the most common structure in the hernia sac; the gastric cardia (sliding hernia) or fundus (paraesophageal hernia) may be involved. Rarely, omental fat, ascitic fluid, or a pancreatic pseudocyst herniates through the esophageal hiatus into the mediastinum. The characteristic location at the esophageal hiatus and the presence of a rounded density containing an air or air–fluid level on upright films are diagnostic. Barium swallow or a CT scan will confirm the diagnosis (see Fig. 19.25) .
Enteric/Neurenteric
Cysts
Enteric cysts are fluid-filled masses lined by enteric epithelium. Esophageal cysts usually arise intramurally or immediately adjacent to the esophagus. When an enteric cyst has a persistent communication with the spinal canal (canal of Kovalevski) and is associated with congenital defects of the thoracic spine (anterior spina bifida, hemivertebrae, or butterfly vertebrae), it is termed a neurenteric cyst. CT or MR can confirm the cystic nature of these masses (Fig. 13.15). If the cyst communicates with the GI tract, it may contain air or an air–fluid level or opacify with contrast during an upper GI series.
Vertebral
Abnormalities
A variety of conditions that affect the thoracic spine may manifest as posterior mediastinal masses. These lesions typically produce lateral deviation of the paraspinal reflection on frontal radiographs. Often, the bony origin of these lesions is not obvious on initial examination, making distinction from neurogenic tumors and other posterior mediastinal
masses
difficult.
986
FIGURE
13.15. Esophageal
Duplication
Cyst. Enhanced CT in
an 18-year-old man with a posterior mediastinal mass on chest radiography (not shown) demonstrates a low-attenuation right paraesophageal duplication
mass
(arrow), consistent with an esophageal
cyst.
Neoplastic, infectious, metabolic, traumatic, or degenerative processes of the thoracic spine may produce a paraspinal mass by one of four mechanisms: (1) expansion of vertebral body or posterior elements (multiple myeloma, aneurysmal bone cyst); (2) extraosseous extension of infection, tumor, or marrow elements (infectious spondylitis, metastatic carcinoma, extramedullary hematopoiesis, respectively); (3) pathologic fracture and paraspinal hematoma formation (any destructive neoplastic or inflammatory process, trauma); or (4) protrusion of degenerative osteophytes. Neoplastic processes are usually easily identified by expansion and destruction of vertebral bodies, with sparing of intervertebral disks.
987
Bronchogenic, breast, or renal cell carcinoma are the most common primary sites of thoracic spinal metastases. Infectious spondylitis is distinguished from neoplastic processes by the presence of a paravertebral mass centered at the point of maximal bone destruction. In patients with a paravertebral abscess secondary to tuberculosis or bacterial infection, narrowing of the adjacent disk space and destruction of vertebral endplates are important clues to the diagnosis. Extramedullary hematopoiesis is seen almost exclusively in conditions associated with ineffective production or excessive destruction of erythrocytes, such as thalassemia major, congenital spherocytosis, and sickle cell anemia. It is recognized by noting expansion of the medullary space and cyst formation within long bones, ribs, and vertebral bodies, with associated lobulated paraspinal soft tissue masses. These masses, which represent hyperplastic bone marrow that has extruded from the vertebral bodies and posterior ribs, are typically seen in the lower thoracic and upper lumbar region. Traumatic injuries to the thoracic spine are usually obvious from the patient's history and recognition of spine fracture on conventional and CT studies of the spine. Degenerative disk disease may produce a localized paraspinal mass on frontal radiographs. Well-penetrated films will show the characteristic inferolaterally projecting osteophytes at the level of the mass, which are most commonly right-sided because of the inhibitory effect of the pulsating descending aorta on left-sided osteophyte formation. Lateral
thoracic
meningoceles represent an anomalous herniation
of the spinal meninges through an intervertebral foramen, resulting in a paravertebral soft tissue mass. Most meningoceles are discovered in middle-aged patients as asymptomatic masses. They are slightly more common on the right, and are multiple in 10% of cases. There is a high association between lateral thoracic meningoceles and neurofibromatosis. A meningocele is the most common posterior mediastinal mass in P.406 patients with neurofibromatosis; conversely, approximately two thirds of patients with meningoceles have neurofibromatosis. Chest radiographs typically reveal a round, well-defined paraspinal mass
988
that is indistinguishable from a neurofibroma. Additional clues to the diagnosis include rib erosion, enlargement of the adjacent neural foramen, vertebral anomalies, or kyphoscoliosis. When a lateral meningocele is associated with kyphoscoliosis, it is usually found at the apex of the scoliotic curve on the convex side. MR demonstration of a herniated subarachnoid space is the diagnostic technique of choice; conventional or CT myelography, which demonstrates filling of the meningocele with contrast, is reserved for equivocal cases.
989
FIGURE 13.16. Pancreatic Pseudocyst as Posterior Mediastinal Mass. A. Portable chest radiograph in a 62-year-old man with an episode of severe pancreatitis 7 months earlier shows a posteroinferior mediastinal mass (arrows). B . Unenhanced CT through the lower chest shows a thick-walled cystic posterior mediastinal mass. C . Scan through the upper abdomen shows communication of the abdominal and thoracic components of the pseudocyst (arrows) via the esophageal hiatus.
Miscellaneous
Conditions
A pancreatic pseudocyst rarely produces a posterior mediastinal mass by extending cephalad from the retroperitoneum through the esophageal or aortic hiatus of the diaphragm. The diagnosis relies on CT demonstration of continuity of a predominantly cystic mass with its retroperitoneal portion (Fig. 13.16) . P.407 The presence of a left pleural effusion is a further clue to the diagnosis. Hernias through the foramen of Bochdalek, which produce a posterior mediastinal mass, are discussed in Chapter 19. Rarely, malignant lymph node enlargement may produce a recognizable paraspinal mass. This is most often seen in NHL and metastatic lung cancer; other mediastinal or extrathoracic sites of involvement are invariably present. Despite the advances in detection and characterization of mediastinal masses with cross-sectional imaging, most patients will require tissue sampling for definitive diagnosis. However, the radiologist can use the information provided by CT or MR to help limit the differential diagnosis and thereby guide the appropriate evaluation and treatment. In a large percentage of cases, when tissue sampling is required, it can be accomplished by CT- or US-guided transthoracic biopsy.
990
DIFFUSE
MEDIASTINAL
DISEASE
The differential diagnosis of diffuse widening of the mediastinum is reviewed
in Table 13.7.
TABLE 13.7 Diffuse Mediastinal Widening
Smooth
Mediastinal lipomatosis Malignant infiltration Lymphoma Small cell carcinoma Adenocarcinoma Mediastinal hemorrhage Arterial bleeding Traumatic
aortic
laceration Aneurysmal Venous
vessel
rupture
bleeding
SVC/right atrial Mediastinitis Acute
arch/great
laceration
(suppurative)
Chronic (sclerosing) Histoplasmosis Tuberculosis Idiopathic
Lobulated
Lymph node enlargement (see Table 13.4) Thymic mass (see Table 13.3) Germ cell neoplasm (see Table 13.3) Vascular lesions Tortuosity of great vessels SVC occlusion (dilated venous Malignancy Sclerosing mediastinitis Catheter-induced thrombosis
991
collaterals)
Neurofibromatosis
SVC, superior vena cava.
Mediastinal infection is an uncommon condition that may be divided into acute and chronic forms based upon etiology, clinical features, and radiologic findings. The distinction between acute and chronic infection is important because there are considerable differences in treatment and prognosis. Acute
mediastinitis is caused by bacterial infection that most often
develops following esophageal perforation or is a complication of cardiothoracic surgery. Esophageal perforation may complicate esophageal instrumentation (e.g., endoscopy, biopsy, dilatation, or stent placement); penetrating chest trauma; esophageal carcinoma; foreign body or corrosive ingestion; or vomiting. Spontaneous esophageal perforation following prolonged vomiting is termed Boerhaave syndrome. In this condition, a vertical tear occurs along the left posterolateral wall of the distal esophagus, just above the esophagogastric junction, leading to signs and symptoms of acute mediastinitis. Less commonly, acute mediastinitis may develop from intramediastinal extension of infection in the neck, retropharyngeal space, lungs, pleural space, pericardium, or spine. The clinical presentation of acute mediastinitis is usually dramatic and is characterized by severe retrosternal chest pain, fever, chills, and dysphagia, often accompanied by evidence of septic shock. Physical examination may reveal findings associated with pneumomediastinum, with subcutaneous emphysema in the neck and an apical, systolic crunching sound on chest auscultation (Hamman sign). The most common chest radiographic findings are widening of the superior mediastinum in 66% of patients and pleural effusion in 50% of patients. Specific findings such as mediastinal air or air–fluid levels are less common. When mediastinitis occurs in association with Boerhaave syndrome, pneumoperitoneum and left
992
hydropneumothorax may be seen. When esophageal perforation is suspected, an esophagram should be performed to detect leakage of contrast into the mediastinum and to localize the exact site of perforation. In a patient not at risk for aspiration, a water-soluble contrast agent is administered initially. Once gross contrast extravasation has been excluded, barium is then given for superior radiographic detail. The sensitivity of the esophagram for detecting contrast leakage is highest when the study is obtained within 24 hours of the perforation. MDCT is the radiologic study of choice for the diagnosis of acute mediastinitis (1 0). CT findings include extra luminal gas, bulging of the mediastinal contours, and focal or diffuse soft tissue infiltration of mediastinal fat. Localized fluid collections suggest focal abscess formation. Associated findings include mediastinal venous thrombosis, pneumothorax, pleural effusion abscess, and vertebral osteomyelitis.
or
empyema,
subphrenic
P.408 While the clinical and radiographic diagnosis of mediastinitis is often straightforward, it may be difficult in postoperative patients who have undergone recent median sternotomy. In these patients, infiltration of mediastinal fat and focal air or fluid collections may be normal findings on postoperative CT scans performed days to weeks following the removal of intraoperatively placed mediastinal drains. In such patients, the progression of findings on follow-up CT scans will correctly identify the majority of those with postoperative mediastinal
infection.
The prognosis for patients with acute mediastinitis varies with the underlying etiology and the extent of mediastinal involvement at the time of diagnosis. Esophageal perforation is associated with the poorest outcome, with a mortality approaching 50%. A delay in diagnosis and treatment of the mediastinal infection of greater than 24 hours is associated with a significant increase in overall morbidity and mortality. In addition to its sensitivity in the diagnosis of mediastinitis, CT can be used to guide treatment and predict outcome. Those patients with
993
evidence of extensive mediastinal infection, seen on CT as diffuse infiltration of the mediastinal fat without evidence of abscess formation, have a mortality approaching 50%. In contrast, patients with discrete mediastinal abscesses that are amenable to surgical or percutaneous drainage, or with small localized abscesses that are amenable to antibiotic therapy alone, have a more favorable prognosis. In addition, patients with mediastinal abscesses and contiguous empyema or subphrenic abscess may respond favorably to drainage of these extramediastinal collections.
Chronic Sclerosing Mediastinitis
(Fibrosing)
The hallmarks of chronic sclerosing mediastinitis are chronic inflammatory changes and mediastinal fibrosis. The most common cause of this rare condition is granulomatous infection, usually secondary to Histoplasma capsulatum. Tuberculous infection, radiation therapy, and drugs (methysergide) are less common causes. Idiopathic mediastinal fibrosis, which is probably an autoimmune process, is related to fibrosis in other regions, including the retroperitoneum, intraorbital fat, and thyroid gland. Several theories have been advanced to explain the pathogenesis of sclerosing mediastinitis owing to histoplasmosis. The most widely accepted theory suggests that affected patients develop an idiosyncratic hypersensitivity response to a fungal antigen that “leaks― from infected mediastinal lymph nodes. Clinically, this condition occurs in adults and presents with a variety of symptoms, depending upon the extent of fibrosis and the mediastinal structures compromised by the fibrotic process. The superior vena cava (SVC) is the most commonly affected structure, with involvement in over 75% of symptomatic patients. The SVC syndrome manifests with headache, epistaxis, cyanosis, jugular venous distention, and edema of the face, neck, and upper extremities. The most serious and potentially fatal manifestation of sclerosing mediastinitis is obstruction of the central pulmonary veins, which produces pulmonary edema that may mimic severe mitral
994
stenosis. Patients with involvement of the tracheobronchial tree may have cough, dyspnea, wheezing, hemoptysis, and obstructive pneumonitis. Dysphagia or hematemesis can be seen with esophageal involvement. Less commonly, pulmonary arterial hypertension and cor pulmonale can develop from narrowing of the pulmonary arteries. The most common finding noted on chest radiographs is asymmetric lobulated widening of the upper mediastinum, most often on the right. When the process is secondary to granulomatous infection, enlarged calcified lymph nodes may be seen. Narrowing of the tracheobronchial tree may be evident. The sequelae of vascular involvement may be seen, including oligemia from pulmonary arterial compression or venous hypertension and pulmonary edema from involvement of the central pulmonary veins. Postobstructive atelectasis or consolidation may also be seen. CT is the modality of choice for the diagnosis and assessment of chronic sclerosing mediastinitis. Enlarged lymph nodes with calcification are the most common finding (Fig. 13.17). The fibrotic infiltration of the mediastinal fat that is characteristic of this condition is seen as abnormal soft tissue density replacing the normal mediastinal fat with obliteration of the normal mediastinal interfaces. CT delineates the degree of involvement of the mediastinal vessels, trachea, and central bronchi. In patients with significant SVC involvement, collateral venous channels within the mediastinum and chest wall are well demonstrated. MR is superior to CT for the assessment of vascular involvement. The ability to examine the mediastinal vessels in both the axial and coronal planes without the need for intravenous contrast helps detect vascular compromise. A significant disadvantage of MR is its inability to detect nodal calcification, a finding that is key to the diagnosis. For this reason, MR is most often utilized as an adjunct to CT when findings of vascular involvement are equivocal. A definitive diagnosis of chronic sclerosing mediastinitis and the establishment of the underlying etiology are difficult. Skin tests for histoplasmosis and tuberculosis may add additional information but are usually not helpful. The precise diagnosis, and more important
995
the distinction from infiltrating malignancy, usually requires biopsy.
Mediastinal
Hemorrhage
Injury to mediastinal vessels resulting from blunt or penetrating thoracic trauma is the most common cause of mediastinal hemorrhage. Blunt chest trauma most often occurs in the setting of a motor P.409 vehicle accident, when rapid deceleration and thoracic cage compression produce shearing effects at the aortic isthmus. Iatrogenic trauma, usually from attempts at central line placement, can also cause mediastinal hemorrhage. Spontaneous hemorrhage may develop in patients with a coagulopathy, or with aortic rupture from aneurysm or dissection. Chronic hemodialysis, radiation vasculitis, and bleeding into a mediastinal mass are rare causes of mediastinal
hemorrhage.
996
FIGURE 13.17. Sclerosing Mediastinitis From Histoplasmosis. A. Posteroanterior chest film in an asymptomatic 68-year-old man shows lobulated widening of the upper mediastinum. B . Contrast-enhanced CT reveals marked dilatation of the left superior intercostal vein (arrows), highattenuation material in and around the superior vena cava, and numerous collaterals within the mediastinal fat. C . A noncontrast scan at approximately the same level reveals mediastinal calcification obliterating the superior vena cava. The patient was a former resident of Ohio where histoplasmosis is endemic.
In the nontraumatic setting, the symptoms and signs of mediastinal hemorrhage are often mild or absent. The patient may complain of retrosternal chest pain radiating toward the back. Rarely, SVC
997
compression may result in the SVC syndrome. Extension of blood from the mediastinum superiorly into the retropharyngeal space may result in neck stiffness, odynophagia, or stridor. The main radiographic finding in mediastinal hemorrhage of any cause is a focal or diffuse widening of the mediastinum that obscures the normal mediastinal contours (1 1). In mediastinal hemorrhage, the mediastinum develops a flat or slightly convex outward contour, unlike the round, lobulated, or irregular contour seen with enlarged lymph nodes or a localized mediastinal mass. Blood extending from the mediastinum into the pleural or extrapleural space produces a free-flowing effusion or a loculated extrapleural collection, respectively. Rarely, extension of blood into the lungs via the bronchovascular interstitium produces interstitial opacities that mimic pulmonary edema. Serial radiographs may show rapid changes in mediastinal or pleural fluid collections in patients with persistent hemorrhage. CT demonstrates abnormal soft tissue within the mediastinum that obliterates the normal interfaces between the mediastinal fat, the vessels, and the airways (Fig. 13.18). Freshly clotted blood is high in attenuation and is usually easily appreciated on helical CT. CT is also superior to plain radiography in P.410 demonstrating the extramediastinal extent of hemorrhage and is useful in demonstrating associated thoracic injuries in patients following blunt chest trauma.
998
FIGURE
13.18. Mediastinal
Hematoma
From
Ruptured
Thoracic Aortic Aneurysm. A. Portable chest radiograph in an 83-year-old woman with chest pain shows marked mediastinal widening. B . Contrast-enhanced CT demonstrates aneurysmal dilatation of the descending aorta, with active extravasation (arrow) into a large mediastinal hematoma. The patient was not a surgical candidate and expired shortly after the study.
Mediastinal
lipomatosis is a benign, asymptomatic condition
characterized by excessive deposition of fat in the mediastinum. Predisposing conditions include obesity, Cushing disease, and corticosteroid therapy. However, this entity is unassociated with identifiable conditions in approximately 50% of patients. On conventional radiographs, the most common finding is smooth, symmetric widening of the superior mediastinum. If the amount of fat deposition is marked, the mediastinum may show lobulated margins. Unlike mediastinal tumor infiltration or hemorrhage, which usually cause tracheal deviation or narrowing, the trachea remains at midline in mediastinal lipomatosis. Fat may also accumulate in the paraspinal regions, chest wall, and cardiophrenic angles; the latter produces enlargement of the epipericardial fat pads that is a clue to the proper diagnosis.
999
CT provides a definitive diagnosis by demonstrating abundant, homogeneous, unencapsulated fat that bulges the mediastinal contours (Fig. 13.19). Displacement or compression of mediastinal structures, particularly the trachea, is notable by its absence. Heterogeneity within the fat suggests other primary or superimposed conditions, such as neoplastic infiltration, infection, hemorrhage, or fibrosis. Multiple symmetric lipomatosis is a rare entity that resembles simple mediastinal lipomatosis radiographically. The distinction between these two conditions is made by the distribution of abnormal fat and mass effect on mediastinal structures. In multiple symmetric lipomatosis, the cardiophrenic angles, paraspinal areas, and the anterior mediastinum are spared; periscapular lipomas may also be seen. The trachea is often compressed or displaced by fat in patients with this condition, whereas this is not seen in simple lipomatosis.
Malignancy Malignant involvement of the mediastinum is typically seen as discrete masses or lymph node enlargement. Rarely, diffuse soft tissue infiltration of the mediastinal fat may occur, either alone or in association with focal lesions. Plain radiographs are nonspecific, usually demonstrating mediastinal widening. CT shows soft tissue infiltration of the normal mediastinal fat and obliteration of the normal tissue planes. This pattern is most common with extracapsular spread of lymphoma or small cell carcinoma of the lung. The latter disease has a high propensity to invade mediastinal structures and therefore may present with symptoms of airway obstruction or SVC syndrome. Pneumomediastinum is the presence of extraluminal gas within the mediastinum. Possible sources of such gas include the lungs, trachea, central bronchi, esophagus, and extension of gas from the neck or abdomen (Table 13.8) (Fig. 13.20) (1 2) . Air from the lungs is the most common source of pneumomediastinum. The mechanism of pneumomediastinum formation involves a sudden rise in intrathoracic and intra-alveolar
1000
pressure that leads to alveolar rupture. The P.411 extra-alveolar air first collects within the bronchovascular interstitium and then dissects centrally to the hilum and mediastinum (the Macklin effect). Less commonly, the air may dissect peripherally toward the subpleural interstitium and rupture through the visceral pleura to produce a pneumothorax.
FIGURE
13.19. Mediastinal
Lipomatosis.
A. Frontal chest
radiograph shows a widened superior mediastinum, particularly on the right (arrow). B . Unenhanced CT at the level of the aortic arch shows abundant mediastinal fat, responsible for the mediastinal widening.
Pneumomediastinum most commonly complicates mechanical ventilation in patients with ARDS, because the combination of positive pressure ventilation and abnormally stiff lungs predisposes to alveolar rupture. Spontaneous pneumomediastinum can occur with deep inspiratory or Valsalva maneuvers during strenuous exercise, childbirth, weightlifting, and inhalation of drugs such as marijuana, nitrous oxide, and crack cocaine. Patients with asthma are prone to pneumomediastinum; this is related to the airways obstruction that
1001
characterizes this disease. Prolonged vomiting from any cause may lead to intrathoracic pressures that are sufficiently high to produce pneumomediastinum. In patients with diabetic ketoacidosis, the increased respiratory effort that accompanies attempts at correcting the underlying metabolic acidosis can lead to pneumomediastinum. Blunt chest trauma can result in pneumomediastinum as a result of an abrupt increase in intra-alveolar pressure and shearing forces affecting the alveolar walls.
TABLE
Intrathoracic source
13.8
Pneumomediastinum
Alveoli Valsalva
maneuver
Positive pressure Esophagus Boerhaave
ventilation
syndrome
Endoscopic interventions dilatation, sclerotherapy)
(biopsy,
Carcinoma Tracheobronchial tree Bronchial stump dehiscence Tracheobronchial
laceration
Fistula formation Tracheal/esophageal malignancy Infection (tuberculosis, histoplasmosis)
Extrathoracic
Recent
sternotomy/thoracotomy
source
Pneumoperitoneum/pneumoretroperitoneum Subcutaneous emphysema in neck Stab wound Laryngeal fracture
Pneumomediastinum arising from the tracheobronchial tree or esophagus usually is a result of traumatic disruption of these
1002
structures. The marked shearing forces that develop with blunt trauma may lead to fracture of the trachea or mainstem bronchi. Penetrating trauma to the tracheobronchial tree is usually iatrogenic and may follow endotracheal intubation, bronchoscopy, or tracheostomy. Rarely, neoplasms or inflammatory lesions (e.g., tuberculosis) may erode through the tracheal wall and into the peritracheal fat. Esophageal rupture is most often spontaneous, usually in the setting of severe, prolonged vomiting (Boerhaave syndrome). In addition to pneumomediastinum, a left hydropneumothorax
and P.412
pneumoperitoneum may be present in this condition. Spontaneous esophageal rupture may occur during childbirth, during a severe asthmatic episode, or with blunt chest trauma. Endoscopic procedures, stent placement, esophageal dilatation, corrosive ingestion, and carcinoma may lead to esophageal perforation. Mediastinal gas may be produced by bacterial organisms in acute mediastinitis.
1003
FIGURE 13.20. Pneumomediastinum. Cone-down view of a patient with spontaneous pneumomediastinum shows vertically oriented lucencies (arrowheads) outlining the aorta, esophagus, and heart and extending into the thoracic inlet superiorly.
Air within the soft tissues of the neck from penetrating trauma or laryngeal fracture may lead to pneumomediastinum by extending inferiorly through the retropharyngeal and prevertebral spaces, or
1004
along the sheaths of the great vessels. Deep space infections in the neck can spread along the same fascial planes and lead to mediastinitis. The term Ludwig angina describes the substernal chest pain caused by the intramediastinal extension of such infections. Rarely, pneumomediastinum develops as air dissects superiorly from the retroperitoneum through the aortic hiatus or from the peritoneal cavity along the internal mammary vascular sheaths. The symptoms associated with pneumomediastinum vary with the underlying etiology, extent of mediastinal air, and presence of mediastinitis. Mediastinal air without infection is generally asymptomatic and does not require treatment. In some patients with spontaneous pneumomediastinum, there may be substernal, pleuritic-type chest pain of sudden onset that can be related to a specific inciting incident, such as vomiting or the Valsalva maneuver. Dyspnea may be present. In adults, mediastinal air under pressure usually escapes into the neck, producing crepitus over the neck, supraclavicular regions, and chest wall. Rarely, mediastinal air under pressure may produce a tension pneumomediastinum in which the clinical findings are those of cardiac tamponade. Patients with mediastinitis and pneumomediastinum are usually seriously ill with chest pain, high fevers, dyspnea, and signs of sepsis. The radiographic findings of pneumomediastinum are reviewed in Chapter 1 2.
THE
HILA
Hilar abnormalities are first appreciated on conventional posteroanterior and lateral chest radiographs. CT and MR are used to confirm and characterize hilar masses or to detect subradiographic involvement of the hila; the latter most often in patients with bronchogenic carcinoma.
Unilateral
Hilar
Enlargement
Malignancy (Table 13.9). A hilar mass usually represents bronchogenic carcinoma or confluent lymph node metastases (Fig. 13.21). Unilateral hilar enlargement may be the presenting
1005
radiographic feature of squamous cell carcinoma, where the hilar mass represents the central extension of an endobronchial tumor from its origin within a segmental bronchus. Concomitant hilar lymph node involvement may contribute to hilar enlargement in some of these patients. Approximately 20% of patients with squamous cell carcinoma have a hilar mass on chest radiograph. In contrast, patients with adenocarcinoma and large cell carcinoma more commonly present with a peripheral pulmonary nodule or mass. In many patients, the hilar mass may be obscured by adjacent lung collapse
or
obstructive
pneumonitis.
Unilateral hilar enlargement resulting from metastatic lymph node involvement is most often seen in small cell carcinoma. The propensity of this tumor for early invasion of the bronchial submucosa and peribronchial lymphatics accounts for the high incidence of widespread hematogenous and hilar and mediastinal lymph node metastases P.413 at the time of diagnosis. Plain film evidence of enlarged hilar lymph nodes resulting from metastases from adenocarcinoma of lung or large cell carcinoma are seen in only 10% to 15% of patients. Contrast-enhanced CT or MR is more sensitive for detecting enlarged hilar nodes and should be performed in all patients to guide further staging procedures and for proper preoperative or treatment planning.
TABLE 13.9 Unilateral Hilar Enlargement
Lymph node enlargement
Malignancy
Bronchogenic
carcinoma
Lymph node metastases Bronchogenic carcinoma Head and neck malignancy
1006
Squamous cell carcinoma of skin, larynx Thyroid carcinoma Breast carcinoma Melanoma Genitourinary malignancy Renal cell Testicular Lymphoma
Infection
carcinoma neoplasm
Tuberculosis Histoplasmosis Coccidioidomycosis Pneumonic plague Tularemia Anaerobic lung Measles
abscess
Mononucleosis
Pulmonary artery enlargement
Valvular pulmonic stenosis Pulmonary artery aneurysm Infection Tuberculosis aneurysm)
(Rasmussen
Left-to-right shunts Patent ductus arteriosis Atrial and ventricular septal defects Arteritis (see below) Tetralogy of Fallot Central pulmonary embolus Chronic thromboembolic disease Pulmonary arteritis Behc¸et disease Hughes-Stovin syndrome Takayasu
1007
arteritis
Cyst
Bronchogenic
cyst
Metastases to hilar and mediastinal lymph nodes from extrathoracic malignancies are uncommon, occurring in approximately 2% of patients. The malignancies that are most often associated with intrathoracic nodal metastases are genitourinary (renal and testicular); head and neck (skin, larynx, and thyroid); breast; and melanoma (7). In renal cell carcinoma and seminoma, lymphatic spread of tumor to retroperitoneal nodes and up the thoracic duct to the posterior mediastinum is the mode of spread to thoracic nodes. Although there is no direct communication between the thoracic duct and anterior mediastinal lymph nodes, reflux of tumor emboli through incompetent valves may allow tumor spread to hilar, paratracheal, and intraparenchymal lymphatics. Head and neck tumors reach the mediastinum via lymphatic spread from cervical lymph nodes. Intrathoracic nodal metastases from breast carcinoma are often seen late in the course of disease, often years after the initial diagnosis. Malignant melanoma is the extrathoracic neoplasm with the highest incidence of intrathoracic nodal metastases; patients with nodal disease will almost invariably have radiographic evidence of
parenchymal
metastases.
Although 75% of patients presenting with Hodgkin lymphoma have evidence of intrathoracic lymph node enlargement, isolated unilateral hilar lymph node enlargement is uncommon. The thoracic manifestations in NHL differ in primary pulmonary lymphoma, versus lymphoma that primarily involves extrathoracic sites with secondary pulmonary involvement. Thoracic involvement in primary pulmonary lymphoma is largely limited to parenchymal and pleural disease, whereas secondary pulmonary lymphoma generally manifests as intrathoracic lymph node enlargement, with 35% showing hilar or middle mediastinal lymph node enlargement and some presenting as an isolated finding.
Infection
1008
Unilateral hilar or mediastinal lymph node enlargement is a characteristic feature in primary pulmonary tuberculosis in distinction to postprimary tuberculosis; an exception is the severely immunocompromised patient with AIDS. Isolated lymph node enlargement as a manifestation of primary tuberculosis is more common in children than in adults. There is almost always concomitant parenchymal disease in immunocompetent patients with lymph node enlargement. Fungal infections such as histoplasmosis and coccidioidomycosis may present with hilar lymph node enlargement, typically associated with patchy or lobar airspace consolidation in the ipsilateral lung. A variety of bacterial infections have been associated with unilateral hilar lymph node enlargement, including plague, tularemia, and anaerobic lung abscess. A characteristic finding in patients with pneumonic plague is the detection on unenhanced CT of increased attenuation within hilar and mediastinal nodes that drain regions of parenchymal involvement owing to intranodal hemorrhage. Tularemia (Francisella
tularensis)
causes parenchymal consolidation in association with hilar lymph node enlargement and pleural effusion. The viral infections most commonly associated with hilar lymph node enlargement are infectious mononucleosis and measles pneumonia. The thorax is infrequently involved in mononucleosis, but hilar lymph node P.414 enlargement is the most common manifestation of intrathoracic disease. Lymph node enlargement may accompany the reticular interstitial opacities of typical measles pneumonia, or it may be associated with nodular, segmental, or lobar opacities and pleural effusion in atypical measles pneumonia.
1009
FIGURE 13.21. Hilar Nodal Metastases From Melanoma. A. Posteroanterior radiograph in a patient with melanoma shows left hilar enlargement (arrow). B . Enhanced CT scan shows enlarged left hilar lymph nodes (arrows) from metastatic disease.
FIGURE 13.22. Unilateral Hilar Enlargement From Idiopathic Dilatation of the Pulmonary Artery. A. Scout view
1010
from chest CT shows abnormal convexity in the region of the main pulmonary artery (arrow). Note thoracic scoliosis. B . Enhanced CT scan shows dilated main pulmonary artery with normal right and left pulmonary arteries. Physical examination and echocardiogram showed no evidence of pulmonic valve disease.
Pulmonary
Artery
Enlargement
Although unilateral hilar enlargement is most often the result of a mass or enlarged lymph nodes, abnormal enlargement of the right or left pulmonary artery may cause hilar prominence (Fig. 13.22) . Vascular disorders that produce unilateral pulmonary artery enlargement include poststenotic P.415 dilatation from valvular or postvalvular pulmonic stenosis, pulmonary artery aneurysms, and distension of the pulmonary artery by thrombus or tumor. Patients with congenital valvular pulmonic stenosis may develop poststenotic dilatation or aneurysms of the main and left pulmonary arteries from the jet effect of blood upon these vessels. Rarely, stenoses resulting from pulmonary artery vasculitis, congenital rubella, or Williams syndrome may lead to poststenotic dilatation of a pulmonary artery. Aneurysms of the central pulmonary arteries are usually associated with congenital heart disease, such as pulmonic stenosis and left-to-right shunts from ventricular septal defect and patent ductus arteriosis. Rare vasculitides such as Behçet disease and Hughes-Stovins syndrome may present with pulmonary artery aneurysms. A large pulmonary embolus lodging in the proximal portion of a pulmonary artery may cause proximal dilatation. Obviously, these patients are symptomatic and will show characteristic findings on perfusion lung scan, helical CT,
and
pulmonary
arteriography.
Bronchogenic cyst is an uncommon cause of a hilar mass. CT and MR will show a round, smooth, thin-walled cyst, usually found in an asymptomatic young adult. Because the hilum is an unusual location
1011
for a bronchogenic cyst, and distinction from a necrotic tumor or lymph node mass cannot be made radiographically, these lesions should be biopsied or removed.
Bilateral
Hilar
Enlargement
Bilateral hilar enlargement is the result of enlargement of either the hilar lymph nodes or the central pulmonary arteries (Table 13.10) .
Malignancy The malignancies producing bilateral hilar lymph node enlargement are similar to those producing unilateral enlargement. In distinction to unilateral nodal enlargement, metastases are uncommon causes of bilateral hilar nodal enlargement. The most frequent solid tumors producing bilateral hilar disease are small cell carcinoma of the lung and
malignant
melanoma.
TABLE 13.10 Bilateral Hilar Enlargement
Lymph node
Malignancy
enlargement
Infection
(see Table 13.2) (see Table 13.2)
Inflammatory disease Sarcoidosis Berylliosis Angioimmunoblastic lymphadenopathy Inhalational disease Silicosis
Pulmonary artery enlargement
Pulmonary arterial hypertension Left-to-right intracardiac shunt High output state Anemia Thyrotoxicosis
1012
Cystic
fibrosis
Bilateral hilar lymph node involvement by lymphoma is more common in Hodgkin disease than NHL. Hilar involvement is virtually never seen without concomitant anterior mediastinal nodal enlargement in Hodgkin disease, whereas NHL may produce isolated hilar disease. The most common chest radiographic manifestation of leukemic involvement of the thorax is hilar and mediastinal lymph node enlargement; it is seen in up to 25% of patients. Lymph node enlargement is much more common in the lymphocytic than the myelogenous form, particularly in chronic lymphocytic leukemia.
Infection Mediastinal and hilar lymph node enlargement from infection is most often seen in tuberculous and fungal infection with histoplasmosis and coccidioidomycosis. In these diseases, the lymph node enlargement may be unilateral or bilateral. With bilateral disease, the enlargement is asymmetric in distinction to sarcoidosis, which is typically symmetric. Bacterial infection from Bacillus anthracis (anthrax) and Yersinia pestis (plague) may produce bilateral hilar enlargement. In anthrax infection, the lymph node enlargement is often associated with patchy airspace opacities in the lower lobes. The bubonic form of plague may produce marked hilar and mediastinal adenopathy without pneumonia. Recurrent bacterial infection complicating cystic fibrosis is often associated with bilateral hilar lymph node enlargement, and distinction from pulmonary artery enlargement owing to pulmonary hypertension may be difficult. Sarcoidosis is associated with bilateral hilar lymph node enlargement in 80% of patients. Most of these patients have concomitant paratracheal lymph node enlargement, and nearly half have concomitant radiographic parenchymal disease. The pattern of lymph node involvement in sarcoidosis has been termed the 1-2-3 sign, with 1 = right paratracheal, 2 = right hilar, and 3 = left hilar lymph node enlargement (Fig. 13.9; see Fig. 17.25). The enlarged
1013
nodes produce symmetric, lobulated hilar masses on plain film, since the enlarged nodes remain separate. In 20% of patients, the involved lymph nodes will calcify; usually the calcifications are punctate in appearance, but occasionally peripheral “eggshell― calcification is seen. In some patients, the involved nodes can be seen to enhance after contrast administration on CT. In the majority of patients, the enlarged nodes resolve within 2 years of discovery; in a small percentage, the nodes remain enlarged for many years.
Berylliosis
and
Silicosis
The hilar and mediastinal lymph node enlargement of chronic berylliosis is radiographically indistinguishable from that of sarcoidosis. Similarly, silicosis can produce hilar and mediastinal lymph node enlargement; eggshell calcification of hilar nodes is highly suggestive of this entity, although P.416 peripheral nodal calcification may also be seen with sarcoidosis, histoplasmosis, or amyloidosis.
TABLE 13.11 Small Hilum (Hila)
Unilateral
Absence or hypoplasia of the pulmonary artery Hypoplastic or hypogenetic lung Swyer-James syndrome Lobar atelectasis Lobar resection Compression/invasion Cyst
of
the
pulmonary
Neoplasm Fibrosing mediastinitis
Bilateral
Emphysema Obstruction to pulmonary flow Fibrosing mediastinitis
1014
artery
Tetralogy of Fallot Valvular pulmonic stenosis Ebstein anomaly
Bilateral pulmonary artery enlargement is seen with increased flow or increased resistance in the pulmonary circulation. The conditions associated with bilateral are reviewed in Chapter 14.
Small
pulmonary
arterial
enlargement
Hila
Bilaterally small hila (Table 13.11) can be seen in some adults with severe pulmonary overinflation from emphysema or in those with diminished pulmonary blood flow due to congenital pulmonary outflow obstruction (tetralogy of Fallot, Ebstein anomaly). The most common causes of a small hilum are atelectasis and resection of a portion of lung, which leave a small residual hilar artery supplying the remaining lobe or lobes. Hypoplasia of the pulmonary artery, often with associated abnormalities of the ipsilateral lung (hypogenetic lung syndrome, Swyer-James syndrome), is another cause of a small hilum. Less commonly, invasion of the proximal pulmonary artery by mediastinal tumor, or obstruction of the pulmonary artery on account of fibrosing mediastinitis, can produce a diminutive hilar shadow. In any patient in whom a small hilum is a new radiographic finding, a CT scan should be performed to assess the mediastinum for central obstructing lesions. The left hilum can appear small in patients in whom the hilar shadow is obscured by the upper left heart margin or by fat in the region of the aortopulmonic interface. In these cases, the lateral radiograph will usually show a left pulmonary artery of normal size.
REFERENCES 1. Glazer GM, Axel L, Moss AA. CT diagnosis of mediastinal
1015
thyroid. AJR Am J Roentgenol 1992;138:495–498. 2. Tomiyama N, Johkoh T, Mihara N, et al. Using the World Health Organization classification of thymic epithelial neoplasms to describe CT findings. AJR Am J Roentgenol 2002;179:881–886. 3. Uffmann M, Schaefer-Prokop C. Radiological diagnostics of Hodgkin and non-Hodgkin lymphomas of the thorax. Radiologe 2004;44:444–456. 4. Zinzani PL, Magagnoli M, Chierchietti F, et al. The role of positron emission tomography (PET) in the management of lymphoma patients. Ann Oncol 1999;10:1181–1184. 5. Strollo DC, Rosado-de-Christenson ML, Jett JR. Primary mediastinal tumors. Part 1: tumors of the anterior mediastinum. Chest 1997; 112(2):511–522. 6. Strollo DC, Rosado-de-Christenson ML, Jett JR. Primary mediastinal tumors. Part 2: tumors of the middle and posterior mediastinum. Chest 1997;112(5):1344–1357. 7. Mountain CF, Dresler CM. Regional lymph node classification for lung cancer. Chest 1997;111:1718–1723. 8. McLoud TC, Kalisher L, Stark P, et al. Intrathoracic lymph node metastases from extrathoracic neoplasm. AJR Am J Roentgenol 1978;131:403–407. 9. McAdams HP, Kirejczyk WM, Rosado-de-Christenson ML, Matsumoto S. Bronchogenic cyst: imaging features with clinical and histopathologic correlation. Radiology 2000;217:441–446. 10. Carrol CL, Jeffrey RB, Federle MP, et al. CT evaluation of
1016
mediastinal infections. J Comput Assist Tomogr 1989;11:449–454. 11. Woodring JH, Loh FK, Kryscio RJ. Mediastinal hemorrhage: an evaluation of radiographic manifestations. Radiology 1984;151: 15–21. 12. Gray JM, Hanson GC. Mediastinal emphysema: etiology, diagnosis,
and
treatment.
Thorax
1017
1966;21:325–332.
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section III - Pulmonary > Chapter 14 Pulmonary Vascular Disease
Chapter
14
Pulmonary
Vascular
Disease
Jeffrey S. Klein
Pulmonary Basic
Edema
Principles
Under normal conditions, the interstitial space of the lung is kept dry by pulmonary lymphatics located within the axial and peripheral interstitium of the lung. The lymphatics drain the small amounts of transudated fluid that enters the interstitial spaces as an ultrafiltrate of plasma. Because there are no lymphatic structures immediately within the alveolar walls (parenchymal interstitium), filtered interstitial fluid is drawn to the lymphatics by a pressure gradient from the alveolar interstitium to the axial and peripheral interstitium. When the rate of fluid accumulation in the interstitium exceeds the lymphatic drainage capabilities of the lung, fluid accumulates first within the interstitial space. As the amount of extravascular fluid increases, fluid accumulates in the corners of the alveolar spaces. Progressive fluid accumulation eventually produces flooding of the alveolar spaces, resulting in airspace pulmonary edema. While interstitial edema may leave the gas-exchanging properties of the lung unaffected, flooding of the alveolar spaces leads to impaired oxygen and carbon dioxide exchange. Excess fluid accumulation in the lung is caused by one of three basic mechanisms. The most common mechanism involves a change in the
1018
normal Starling forces that govern fluid movement in the lung. Because normal fluid movement is determined by the differences in hydrostatic and oncotic pressure between the pulmonary capillaries and surrounding alveolar interstitium, an imbalance in these forces may lead to pulmonary edema. Pulmonary edema is most commonly the result of increased capillary hydrostatic pressure from left heart failure, although diminished plasma oncotic pressure or diminished interstitial hydrostatic pressure may be contributing factors. Another mechanism involves an absence or obstruction of the normal pulmonary lymphatics, which leads to the excess accumulation of interstitial fluid. Third, a wide variety of disorders can injure the capillary endothelium and alveolar epithelium, producing an increase in capillary permeability that allows protein-rich fluid to escape from the capillaries and into the pulmonary interstitium. Radiographic findings in pulmonary edema can be divided into interstitial and airspace components. The radiographic appearance of interstitial pulmonary edema is attributed to thickening of the various components of the interstitial spaces by fluid (1). Thickening of the axial interstitium results in the loss of definition of the intrapulmonary vascular shadows, peribronchial cuffing, and tram tracking. Edema within alveolar septa is not discernible as discrete opacities but produces a ground-glass opacification in the perihilar and lower lung zones—regions in which fluid tends to accumulate in the early phases of edema. Involvement of peripheral and subpleural interstitial structures produces Kerley lines and subpleural edema. Kerley A and B lines represent thickening of central connective tissue septa and peripheral interlobular septa, respectively, while Kerley C lines represent a network of thickened interlobular septa (Fig. 14.1) . Subpleural edema is the accumulation of fluid within the innermost (interstitial) layer of the visceral pleura and is best seen on the lateral radiograph as smooth thickening of the interlobar fissures. The radiographic changes of interstitial pulmonary edema may progress to those of airspace edema or, if successfully treated, resolve within 12 to 24 hours. Airspace pulmonary edema develops when progressive fluid accumulates in the interstitial spaces and spills into
1019
P.418 the alveoli. The chest radiograph typically shows symmetric bilateral airspace opacities that are confluent and predominate in the mid and lower lung zones. Airspace nodules and the findings of interstitial edema (Kerley B lines and subpleural edema) may be seen peripherally. An uncommon form of airspace pulmonary edema, seen most commonly in left heart failure or renal failure, is the “bat's wing― or “butterfly― distribution of disease. In this situation, the airspace opacification is sharply confined to the central, parahilar regions of lung, with sparing of the peripheral or subpleural regions (Fig. 14.2). The reason for this distribution of edema is unknown. As with interstitial edema, the airspace opacities of alveolar edema tend to change rapidly, often within hours. The differential diagnosis of diffuse airspace opacities has been reviewed (see Table 12.9) .
FIGURE 14.1. Interstitial Pulmonary Edema Caused by Cardiac Disease. Posteroanterior (A) and lateral (B) chest films in a 65-year-old man with an anterior wall myocardial infarction shows bilateral linear opacities (Kerley A, B, and C lines) representing interstitial pulmonary edema. Note the prominence of upper lobe vessels, indicating concomitant pulmonary venous
1020
hypertension.
Although not commonly used to evaluate the patient with pulmonary edema, CT and HRCT demonstrate fairly characteristic findings in this disorder (2). Thickening of subpleural, septal, and bronchovascular structures are well depicted on HRCT. Mild parenchymal edema produces a ground-glass pattern around the hila (Fig. 14.3). Early alveolar edema is seen as centrilobular airspace nodules surrounding the arteries within the lobular core, while severe alveolar edema produces dense perihilar airspace opacification. Cardiomegaly, pulmonary venous distention, and pleural effusions are associated findings in cardiogenic or fluid overload edema.
FIGURE 14.2. Perihilar “Bat's Wing― Pulmonary Edema. Frontal chest radiograph in a 32-year-old man with dilated cardiomyopathy reveals dense bilateral perihilar airspace opacification
resulting
from
pulmonary
1021
edema.
FIGURE 14.3. Thin-Section CT of Interstitial Pulmonary Edema. Thin-section CT in a 28-year-old woman with postpartum cardiomyopathy shows characteristic findings of interstitial edema: interlobular septal thickening, peribronchial cuffing, and ground-glass and airspace opacities. Note the associated findings of congestive heart failure, with pulmonary vascular engorgement and bilateral pleural effusions.
P.419
Atypical
Radiographic
Appearances
Several conditions may give rise to atypical radiographic appearances of pulmonary by gravity, it posteriorly or or decubitus
edema. Because the distribution of edema is affected is not surprising that edema fluid accumulates unilaterally in patients maintaining a prolonged supine position, respectively. The diagnosis of unilateral edema
is suggested by typical radiographic and clinical findings of pulmonary edema in one lung that resolve rapidly or redistribute with changes in patient positioning. Another cause of asymmetric or unilateral pulmonary edema is an interruption in the blood supply to
1022
one lung. This may be seen in pulmonary artery hypoplasia or in an acquired obstruction to pulmonary arterial blood flow, such as central pulmonary embolus or extrinsic compression of the pulmonary artery from tumor or fibrosis. In these conditions, the lung with diminished pulmonary blood flow is “protected― from the transudation of fluid and the development of pulmonary edema. Bronchogenic carcinoma, lymphoma, or other causes of unilateral lymph node enlargement can impede normal lymphatic drainage and predispose to unilateral pulmonary edema. Similarly, unilateral pulmonary venous obstruction from tumor or fibrosing mediastinitis will predispose to edema on the affected side. Unilateral pulmonary edema may develop in the lung that is re-expanded by the rapid evacuation of a large pleural fluid collection or pneumothorax. This is known as re-expansion subsequent
pulmonary
edema and is discussed in a
section.
Alveolar pulmonary edema localized to the right upper lung may be seen in patients with severe mitral regurgitation. Edema formation is likely the result of preferential regurgitant flow of blood into the right upper lobe pulmonary vein across the superiorly and posteriorly oriented mitral valve. These patients will usually have typical radiographic findings of interstitial edema elsewhere in the lungs (Fig. 14.4) .
1023
FIGURE
14.4. Interstitial and Right Upper Lung Pulmonary
Edema. In a 64-year-old woman with unstable angina and mitral regurgitation on echocardiography, posteroanterior chest radiograph shows diffuse interstitial pulmonary edema with localized right upper lobe airspace opacification, the latter secondary to ischemic mitral valvular dysfunction.
Patients with pulmonary emphysema may have unusual appearances of alveolar edema. Areas of bullae, most commonly in the apical portions of the lungs, are spared from the development of alveolar edema because the pulmonary blood flow to these regions has already been obliterated by the emphysematous process. These emphysematous regions within adjacent areas of airspace opacification can simulate cavity formation and may be difficult to distinguish radiographically from necrotizing pneumonia or pneumatocele formation. Comparison with previous radiographs and correlation with the clinical course will aid in the proper diagnosis. Hydrostatic pulmonary edema (normal capillary permeability) is the most common form of pulmonary edema. Patients with acute or
1024
chronic renal failure may develop pulmonary edema because of increased pulmonary capillary hydrostatic pressure. The elevated hydrostatic pressure is caused by a combination of hypervolemia and LV dysfunction, with resultant pulmonary venous and capillary hypertension. Volume overload without renal failure may also produce pulmonary edema by a hydrostatic mechanism. Decreased capillary oncotic pressure, present in patients with hypoalbuminemia secondary to the nephrotic syndrome or liver failure, is not considered to be an independent risk factor for the development of pulmonary edema but is a cofactor in several conditions.
TABLE 14.1 Causes of Pulmonary Venous Hypertension and Pulmonary Edema
Obstruction
to L V outflow
Aortic
coarctation
Aortic stenosis Hypoplastic left syndrome
heart
L V failure
Mitral valve disease
Mitral Mitral
stenosis insufficiency
L A myxoma
Cor
triatriatum
Obstruction
of
pulmonary
veins
Central
pulmonary
veins
Fibrosing mediastinitis Pulmonary vein stenosis
1025
Pulmonary
Intrapulmonary
veins
venous
Pulmonary disease
thrombosis
venoocclusive
P.420 Hydrostatic pulmonary edema is usually caused by pulmonary venous hypertension secondary to congestive heart failure. Thus, identification of the radiographic findings of pulmonary venous hypertension and pulmonary edema will provide the diagnosis. The majority of these patients will have LV failure or mitral valve disease. A list of the causes of mechanical or functional obstruction to pulmonary venous return is found in Table 14.1. The radiographic findings of pulmonary venous hypertension are enlargement of pulmonary veins and redistribution of pulmonary blood flow to the upper lung zones (1). Pulmonary venous enlargement is seen as progressive dilatation of horizontally oriented pulmonary veins on serial chest radiographs. The redistribution of pulmonary blood flow results from lower zone pulmonary venous constriction and increased resistance to lower zone blood flow, with resultant preferential flow through upper lobe vessels. Therefore, in pulmonary venous hypertension in the upright patient, the upper zone vessels are as large as or larger in diameter than the lower zone vessels. This is the opposite of the normal appearance, in which the lower zone vessels are larger than the upper zone vessels as the result of the normal gravitational effects on pulmonary blood flow. It should be noted that there are conditions other than pulmonary venous hypertension in which there is distention of upper zone pulmonary vessels, including left-right shunts and basilar lung disease. The association of upper zone vascular prominence with findings of LV failure (cardiomegaly, pulmonary edema, and pleural effusion) usually allows for the correct diagnosis. The sequence of events following the development of pulmonary venous hypertension has been studied in patients with acute cardiac decompensation following myocardial infarction. Several studies have
1026
correlated the radiographic findings of pulmonary venous hypertension in the erect patient with measurements of pulmonary capillary wedge pressure (PCWP) using flow-directed balloon occlusion (i.e., Swan-Ganz) catheters. When PCWP is normal (8 to 12 mm Hg), the chest radiograph is normal. Mild elevation of PCWP (12 to 18 mm Hg) produces constriction of lower lobe vessels and enlargement of upper lobe vessels. Progressive elevation of PCWP (19 to 25 mm Hg) leads to the findings of interstitial pulmonary edema: loss of vascular definition, peribronchial cuffing, and Kerley lines (Fig. 14.1). PCWP above 25 mm Hg produces alveolar filling with radiographic findings of bilateral airspace opacities in the perihilar and lower lung zones. The causes of pulmonary venous hypertension may be divided radiographically into those associated with a normal heart size and those associated with cardiomegaly. A severe, long-standing obstruction to LV outflow (e.g., aortic stenosis) is usually associated with a normal heart size unless the LV has failed. Chronic left ventricular failure is invariably associated with cardiomegaly, although acute LV decompensation, as in acute myocardial infarction or acute aortic regurgitation, may show a normal heart size. Obstruction or incompetence at the level of the mitral valve (e.g., mitral stenosis or regurgitation, left atrial myxoma) may only show left atrial enlargement without ventricular dilatation. Obstruction of the central pulmonary veins (i.e., fibrosing mediastinitis, pulmonary vein thrombosis) is usually associated with the radiographic findings of pulmonary venous hypertension and pulmonary edema with a normal heart size. Intrapulmonary venous obstruction (i.e., pulmonary venoocclusive disease) may show only pulmonary edema, but often the diagnosis is delayed until pulmonary arterial hypertension (PAH) has developed (Table 14.1) .
Increased
Capillary
Permeability
Edema
Rapidly progressive respiratory compromise caused by leakage of protein-rich edema fluid into the lung, resulting from damage to the pulmonary microcirculation, may develop as a complication of a variety of systemic conditions. When respiratory failure develops as a
1027
result of this condition and is associated with increased lung stiffness (noncompliance) it is termed ARDS (for acute respiratory distress syndrome) (3). The edema associated with this syndrome is called lung injury or increased capillary permeability edema, as compared to the normal alveolocapillary permeability of hydrostatic edema. A long list of pulmonary and nonpulmonary disorders have been associated with increased-permeability edema (Table 14.2); the most common are shock, severe trauma, burns, sepsis, narcotic overdose, and pancreatitis. Although the precise pathogenesis of capillary permeability edema has yet to be completely elucidated, current evidence suggests that recruitment and activation of neutrophils in the lung with release of enzymes and oxygen radicals are key factors in the development of capillary endothelial damage. The pathologic changes associated with ARDS are those of diffuse alveolar damage and are common to all P.421 patients regardless of the underlying etiology. Within 12 to 24 hours following the initial insult (stage 1 ARDS), damage to capillary endothelium produces engorged capillaries and proteinaceous interstitial edema. Within the first week (stage 2), the injury to type 1 pneumocytes leads to the flooding of alveoli with edema fluid and proteinaceous and cellular debris, which form hyaline membranes lining the distal airways and alveoli. In stage 3 ARDS, type 2 pneumocytes proliferate in an attempt to reline the denuded alveolar surfaces, and fibroblastic tissue proliferates within the airspaces. This fibroblastic tissue may resolve and leave minimal scarring or—particularly in those with severe disease and long-standing oxygen requirements—result in extensive interstitial fibrosis.
TABLE 14.2 Etiologies of Increased Permeability Pulmonary Edema
Septicemia
Gram-negative
1028
bacteria
Shock
Major
surgery
Burns
Acute
pancreatitis
Disseminated coagulation
intravascular
Drugs
Narcotics Heroin Crack cocaine Aspirin
Inhalation of noxious
Nitrogen
fumes
disease)
dioxide
(silo-filler's
Hydrocarbons
Smoke Chlorine Phosgene
Aspiration of fluid
Fresh or salt water near drowning Gastric fluid aspiration (Mendelson syndrome)
Fat
embolism
1029
Amniotic
fluid
embolism
Radiographically, ARDS follows a predictable pattern. Chest radiographs become abnormal by 12 to 24 hours following the onset of dyspnea and demonstrate patchy peripheral airspace opacities (Fig. 14.5) (2). These opacities coalesce over the next several days to produce confluent bilateral airspace opacities with air bronchograms. Radiographic improvement in the opacities may be seen within the first week, but this is often caused by the effects of increasing positive pressure ventilation rather than true histologic improvement. After 1 week, the airspace opacities gradually give way to a coarse reticulonodular pattern that may resolve over the course of several months or remain unchanged, in which case the pattern represents irreversible pulmonary fibrosis (i.e., honeycombing). Pneumonia complicating ARDS is difficult to diagnose radiographically, but it should be suspected when a focal area of airspace opacification or a significant pleural effusion develops during the course of the disease. Likewise, the superimposition of LV failure may be impossible to recognize but is suggested by rapid clinical and radiographic deterioration associated with changes in measured PCWP and edema fluid protein content. Pneumomediastinum and pneumothorax may result as a complication of positive pressure ventilation to stiff lungs and should be sought on portable chest radiographs.
1030
FIGURE
14.5. Increased Permeability (Lung Injury) Edema
in ARDS. Portable chest radiograph in a 46-year-old woman with severe pancreatitis and respiratory failure reveals bilateral airspace opacification with a somewhat peripheral distribution, representing diffuse alveolar damage and permeability edema.
Radiographic Distinction of Hydrostatic From Increased Capillary Permeability Edema Beyond identifying the presence of pulmonary edema, the ability to distinguish between types of pulmonary edema has significant diagnostic and therapeutic import. Measurements of PCWP and transbronchial sampling of pulmonary edema fluid are techniques that accurately distinguish hydrostatic from increased capillary permeability edema. In hydrostatic edema, PCWP measurements are elevated and a protein-poor transudative edema fluid is present,
1031
while in increased-permeability edema, there is a normal PCWP and proteinaceous edema fluid is seen. Milne and colleagues have described the findings on the chest radiograph that can be used to distinguish cardiac and overhydration edema from increased capillary permeability edema (3). In pulmonary edema associated with chronic cardiac failure, the heart is usually enlarged and displays an inverted (redistributed) pulmonary blood flow pattern. The distribution of edema is even from central to peripheral over the lower lung zones. The vascular pedicle, which represents the mediastinal width at the level of the superior vena cava and left subclavian artery, is widened P.422 (>53 mm on posteroanterior [PA] radiograph), reflecting increased circulating blood volume. Lung volumes are diminished because of decreased pulmonary compliance from edema. Peribronchial Kerley lines, and pleural effusions represent interstitial and intrapleural
transudation
of
fluid,
cuffing,
respectively.
Overhydration or renal failure edema has some features in common with chronic cardiac failure and may be indistinguishable radiographically. Capillary permeability edema can sometimes be distinguished from hydrostatic edema. A more peripheral distribution of edema with a normal heart size and normal vascular pedicle width, the latter indicating normal circulating blood volume, are findings typical of increased capillary permeability edema. It should be noted that some factors may render radiographic distinction of types of pulmonary edema difficult. Radiographs of supine patients will make evaluation of pulmonary blood flow distribution and vascular pedicle width difficult. The presence of severe alveolar edema will obscure underlying vascular markings. Many patients with capillary permeability edema will be overhydrated in attempts to maintain circulating blood volume, producing complex radiographic findings. Neurogenic
pulmonary
edema following head trauma, seizure, or
increased intracranial pressure is a complex phenomenon that appears to involve both hydrostatic and increased permeability mechanisms. Massive sympathetic discharge from the brain in these
1032
conditions produces systemic vasoconstriction and increased venous return, with resultant increase in LV diastolic pressure and hydrostatic pulmonary edema. However, the finding of protein-rich edema fluid and normal PCWP in some patients suggests that increased permeability may be a contributing factor. High-altitude
pulmonary
edema develops in certain individuals
after rapid ascent to altitudes above 3,500 m. Edema typically develops within 48 to 72 hours of ascent and appears to reflect a varied individual response to hypoxemia, in which scattered areas of pulmonary arterial spasm result in transient PAH. This produces an overflow of blood at high pressure to uninvolved areas, resulting in damage to the capillary endothelium and increased permeability edema that is patchy in distribution. Rapid resolution usually occurs within 24 to 48 hours after the administration of supplemental oxygen or a return to sea level.
Re-expansion
Pulmonary
Edema
Rapid re-expansion of a lung after collapse lasting greater than 48 hours may result in the development of unilateral pulmonary edema. Marked increases in negative pleural pressure following pleural tube placement, impaired pulmonary lymphatic drainage following prolonged lung collapse, and ischemia-induced surfactant deficiency resulting in the need for high negative pleural pressure to re-expand the collapsed lung are proposed mechanisms. Recent evidence points toward prolonged collapse producing ischemia and hypoxemia within the lung, which promotes anaerobic metabolism and formation of free radicals. Reperfusion of the lung upon re-expansion then leads to lung injury and permeability edema. Gradual re-expansion of the lung by slow removal of pleural air or fluid over a 24- to 48-hour period and administration of supplemental oxygen helps limit the incidence and severity of this complication.
Acute
Upper
Airway
Obstruction
Pulmonary edema may be seen during or immediately after treatment of acute upper airway obstruction. The proposed mechanism involves
1033
the creation of markedly negative intrathoracic pressure by attempts to inspire against an extrathoracic airway obstruction, producing transudation of fluid into the lung. There are no distinguishing radiographic
features.
Amniotic
Fluid
Embolism
A severe and often fatal form of pulmonary edema may develop in a pregnant woman when amniotic fluid gains access to the systemic circulation during labor. There is an association of this entity with fetal distress and demise, because the mucin within fetal meconium plays a key role in the pathogenesis of this disorder. Embolic obstruction of the pulmonary vasculature by mucin and fetal squames within the amniotic fluid leads to sudden PAH and cor pulmonale with decreased cardiac output and pulmonary edema. An anaphylactoid reaction and disseminated intravascular coagulopathy (DIC) from factors within the amniotic fluid contribute to the shock state. Radiographically,
there
are
bilateral
confluent
airspace
opacities
indistinguishable from pulmonary edema of other etiologies. In severe cases, there is secondary enlargement of the central pulmonary arteries and right heart as a manifestation of cor pulmonale. The diagnosis can be confirmed by identification of fetal squames and mucin in blood samples obtained from indwelling pulmonary
Fat
artery
catheters.
Embolism
The embolization of marrow fat to the lung is a common complication occurring 24 to 72 hours after the fracture of a long bone (e.g., femur). Within the lung, the fat is hydrolyzed to its component fatty acids, causing increased pulmonary capillary permeability and hemorrhagic pulmonary edema. Radiographically and on CT, confluent ground-glass and airspace opacities are seen (Fig. 14.6) . The diagnosis is made by recognizing findings of systemic fat embolism (petechial rash, CNS depression) and pulmonary changes in the appropriate time period following trauma. Most patients have a mild course with minimal respiratory compromise, while a minority
1034
will develop progressive respiratory failure leading to death.
PULMONARY
HEMORRHAGE
Hemorrhage or hemorrhagic edema of the lung can result from trauma; bleeding diathesis; infections (invasive mucormycosis, Pseudomonas, influenza); drugs pulmonary embolism; fat
aspergillosis, (penicillamine); P.423
embolism; ARDS; and autoimmune diseases (Table 14.3) (4). The autoimmune diseases that can cause pulmonary hemorrhage include Goodpasture syndrome, idiopathic pulmonary hemorrhage, Wegener granulomatosis, systemic lupus erythematosus, rheumatoid arthritis, and polyarteritis nodosa.
FIGURE 14.6. Fat Embolism Producing Permeability Edema. CT in an 18-year-old man with dyspnea and hypoxemia 48 hours after intramedullary rod placement for a femoral fracture shows asymmetric ground-glass and airspace opacities with small left pleural effusion.
Goodpasture syndrome is an autoimmune disease characterized by damage to the alveolar and renal glomerular basement membranes
1035
by a cytotoxic antibody. The antibody is directed primarily against renal glomerular basement membrane and cross-reacts with alveolar basement membrane to produce the renal injury and pulmonary hemorrhage characteristic of this disorder. Young adult men are most commonly affected and present with cough, hemoptysis, dyspnea, and fatigue. The pulmonary complaints usually precede clinical evidence of renal failure. Chest films show bilateral coalescent airspace opacities that are radiographically indistinguishable from those of pulmonary edema (Fig. 14.7). Within several days, the airspace opacities resolve, giving rise to reticular opacities in the same distribution. Complete radiographic resolution is seen within 2 weeks, except in those with recurrent episodes of hemorrhage, in whom the reticular opacities persist and represent pulmonary fibrosis. The diagnosis is made by immunofluorescent studies of renal or lung tissue, which show a smooth wavy line of fluorescent staining along the basement membrane. The overall prognosis is poor, although the use of immunosuppressive drugs and plasmapheresis has
improved
survival.
TABLE 14.3 Causes of Pulmonary Hemorrhage
Spontaneous
Thrombocytopenia Hemophilia Anticoagulant
Trauma
Embolic
Pulmonary
disease
Vasculitis
therapy
contusion
Pulmonary embolism Fat embolism
Autoimmune Goodpasture syndrome Idiopathic pulmonary hemorrhage Wegener granulomatosis
1036
Infectious Gram-negative Influenza
bacteria
Aspergillosis Mucormycosis
Drugs
Penicillamine
FIGURE 14.7. Pulmonary Hemorrhage in Goodpasture Syndrome. Posteroanterior chest film in a patient with Goodpasture syndrome shows asymmetric bilateral airspace disease presenting intra-alveolar blood.
Idiopathic
Pulmonary
Hemorrhage
The pulmonary manifestations of idiopathic pulmonary hemorrhage
1037
are clinically and radiographically indistinguishable from those of Goodpasture syndrome. In distinction to Goodpasture syndrome, this disorder is most common in children, with an equal sex distribution. The diagnosis is one of exclusion and is suggested when pulmonary hemorrhage and anemia are found in a patient with normal renal function and urinalysis and an absence of antiglomerular basement membrane
antibodies.
Other Wegener
Vasculitides granulomatosis,
systemic
lupus
erythematosus,
rheumatoid
arthritis, and polyarteritis nodosa are autoimmune disorders associated with a systemic immune complex vasculitis (5). The development of pulmonary hemorrhage in these diseases is secondary to small vessel pulmonary arteritis and capillaritis, which result in spontaneous hemorrhage. The pulmonary manifestations of these diseases are discussed in subsequent sections.
FIGURE 14.8. CT of Pulmonary Hemorrhage. Thin-section CT in a 46-year-old woman with hemoptysis 1 day after initiation of prophylactic
anticoagulation
for
radiofrequency
1038
ablation
of
atrial
fibrillation shows asymmetric bilateral perihilar opacities, reflecting alveolar hemorrhage.
ground-glass
P.424 Differentiation of pulmonary hemorrhage from pulmonary edema or pneumonia may be difficult, particularly because many causes of pulmonary edema and pneumonia may have a significant hemorrhagic component. The rapid development of airspace opacities associated with a dropping hematocrit and hemoptysis should suggest the diagnosis (Fig. 14.8). Hemoptysis, however, is not always present. Associated renal disease, hematuria, or findings of a collagen vascular disorder or systemic vasculitis may provide additional clues. The distinction of pulmonary hemorrhage from pneumonia is made by the absence of fever or purulent sputum and the finding of a normal or elevated carbon monoxide–diffusing capacity. This latter determination is directly related to the volume of gas-exchanging intravascular and extravascular intrapulmonary red blood cells and is therefore elevated in pulmonary hemorrhage or hemorrhagic edema but decreased in pneumonia. The presence of hemosiderin-laden macrophages in sputum, bronchoalveolar lavage fluid, or tissue specimens is evidence of chronic or recurrent intrapulmonary hemorrhage. A rapid radiographic improvement of the airspace opacities in pulmonary hemorrhage is common and may aid in
diagnosis.
PULMONARY
EMBOLISM
Pulmonary embolism (PE) is a common cause of acute chest symptoms. While it is associated with significant morbidity and mortality, treatment with anticoagulation can significantly reduce the likelihood of recurrent emboli that might result in chronic thromboembolic pulmonary hypertension or death. Since anticoagulation has associated morbidity, particularly in elderly and debilitated patients, an accurate diagnosis of the presence or absence of PE is necessary.
1039
The radiologist plays a central role in the diagnostic evaluation of the patient with suspected PE. This section will briefly review the aspects of patient evaluation not related to imaging and then detail the various imaging modalities available to the radiologist. A practical algorithm that serves as a useful guide to the workup of each patient with suspected PE will be provided.
Clinical
and
Laboratory
Findings
The majority of patients with PE have a variety of symptoms, including dyspnea (84%), pleuritic chest pain (74%), anxiety (59%), and cough (53%). However, in certain groups of patients at high risk for PE, asymptomatic embolization is known to occur. Physical examination may reveal tachypnea (respiratory rate >16/min), rales, and a prominent pulmonary component of the second heart sound. Unfortunately, these findings are entirely nonspecific. Only a minority of patients presenting to an emergency department with pleuritic chest pain will be found to have PE. The main laboratory test obtained in patients with suspected pulmonary embolism is a plasma D-dimer level. D-dimer is a degradation product of fibrin and is a very sensitive indicator of the presence of venous thrombosis. Enzyme-linked immunosorbent D-dimer measurements have a sensitivity for deep venous
assay
thrombosis (DVT) of 98% to 100%, and therefore a normal value will effectively exclude the possibility of DVT and PE, particularly when the clinical probability for PE is low.
Radiologic
Evaluation
A number of imaging techniques are routinely employed in the evaluation of the patient with suspected PE. These include the chest radiograph, ventilation/perfusion (V/Q) lung scintigraphy, helical CT angiography, and conventional pulmonary angiography. Noninvasive methods of imaging DVTs include compression and Doppler ultrasound of the legs, lower-extremity indirect CT venography, and magnetic resonance venography of the extremities and pelvis. The relatively noninvasive nature and high accuracy of these techniques
1040
to diagnose DVT and an increasing familiarity with their performance and interpretation among radiologists has led to their widespread use in the workup of PE. A practical algorithm for the radiologic evaluation of PE is shown in Fig. 14.9. Chest radiography is the first examination obtained in all patients with suspected PE. Although the majority of patients with PE will have abnormal radiographs, a significant percentage of patients will have normal chest radiographs. The radiographic findings include cardiac, pulmonary arterial, parenchymal, pleural, and diaphragmatic changes (6) . Cardiac, or more precisely right heart enlargement, is an uncommon finding seen with massive or extensive PE producing cor pulmonale. Enlargement of the central P.425 pulmonary arteries from PAH may also be seen but is more commonly a late sequela of chronic thromboembolic disease. The most common radiographic findings in PE without infarction are localized peripheral oligemia with or without distended proximal vessels (Westermark sign) and peripheral airspace opacification or linear atelectasis. The airspace
opacification
represents
localized
pulmonary
hemorrhage
produced by bronchial and pulmonary venous collateral flow to the obstructed region and is seen with peripheral but not central emboli. Volume loss in the lower lung from adhesive atelectasis caused by ischemic injury to type 2 pneumocytes and secondary surfactant deficiency may produce diaphragmatic elevation and the development of linear atelectasis.
1041
FIGURE
14.9. Algorithm for the Radiologic Evaluation of
Pulmonary
Thromboembolism. PE, pulmonary embolism.
Less than 10% of all PEs result in lung infarction. Collateral bronchial arterial and retrograde pulmonary venous flow prevent infarction in most patients. The distinction between embolism without and with infarction is usually impossible radiographically and is of limited importance, as treatment is identical. Infarction from embolism occurs with greater frequency in patients with underlying heart failure because of their limited collateral bronchial arterial flow to the ischemic region. In PEs with infarction, the cardiac, pulmonary arterial, and peripheral vascular changes are indistinguishable from those seen in embolism without infarction. Radiographic features that suggest infarction include the presence of a small pleural effusion and the development of a pleura-based wedge-shaped opacity (Hampton hump). This opacity, typically found in the posterior or lateral costophrenic sulcus of the lung, is wedge-shaped, homogeneous, and lacks air bronchograms. The blunted apex of the wedge points toward the occluded feeding vessel, while the base is against the pleural surface. This wedge-shaped opacity is often
1042
obscured by surrounding areas of hemorrhage in the early phases following infarction and becomes more obvious with time as the peripheral areas of hemorrhage resolve. A distinction between PE with and without infarction is usually made by noting changes in the radiographic opacities with time. In embolism without infarction, the airspace opacities should resolve completely within 7 to 10 days, while infarcts resolve over the course of several weeks or months and usually leave a residual linear parenchymal scar and/or localized pleural thickening. None of the aforementioned radiographic findings, either alone or in combination, are useful in making a firm diagnosis of PE. Conversely, a completely normal radiograph may be seen in up to 40% of patients with emboli. The prime utility of the chest radiograph in the evaluation of PE is in the detection of conditions that mimic PE clinically, such as pneumonia or pneumothorax, and as an aid to the interpretation of the ventilation/perfusion lung scan.
Ventilation/Perfusion Scintigraphy
(V/Q)
Lung
The intravenous administration of macroaggregates of albumin radiolabeled with technetium (Tc-99m) has given the radiologist a noninvasive method of assessing the patency of the pulmonary circulation. The sensitivity of this technique allows for the confident exclusion of PE when a technically adequate perfusion scan is normal. The addition of ventilation scanning increases the specificity of an abnormal perfusion scan and is always performed in conjunction with the perfusion scan when possible. Perfusion lung scanning is performed by intravenous injection of 5 mCi of Tc-99m macroaggregated albumin P.426 with the patient supine (see Chapter 56). Images are then obtained in eight projections: anteroposterior, PA, right and left lateral, and right and left anterior and posterior oblique views. If perfusion abnormalities are present, a ventilation scan utilizing krypton-81m, xenon-133, or aerosolized Tc-99m diethylenetriamine pentaacetic
1043
acid (DTPA) is then performed. The use of krypton-81m and Tc-99m DTPA allows for comparable oblique projections identical to the perfusion scan. Perfusion defects can then be characterized as ventilation/perfusion matches (absent ventilation/absent or mismatches (normal ventilation/absent perfusion). Ventilation/perfusion mismatch is the hallmark of PE.
perfusion)
Although V/Q scanning is commonly used in the evaluation of the patient with suspected PE, there are limitations to its utility for the diagnosis of PE. First, only a minority of patients (27% in the Prospective Investigation of Pulmonary Embolism Diagnosis [PIOPED] study) undergoing V/Q studies will have either a normal or highprobability study, a result that clinicians can confidently rely upon to guide treatment decisions (5). Second, there is significant interobserver variability in the interpretation of V/Q studies. Finally, there are few well-constructed prospective studies evaluating the accuracy of various patterns of V/Q abnormality in predicting the likelihood of PE. Several diagnostic schemes have been proposed to assign a probability of PE (as determined by pulmonary angiography) given specific combinations of ventilation, perfusion, and concurrent chest radiographic findings. The V/Q scan interpretation categories published with the results of the PIOPED study have become the standard for radiologists interpreting V/Q studies. A normal V/Q scan effectively excludes PE because of the high sensitivity of the test. A high-probability scan, particularly in a patient with a strong clinical suspicion for embolic disease, can be confidently treated for PE. Patients with intermediate or indeterminate (because of extensive obstructive lung disease) probability scans have a 30% to 40% incidence of PE. Likewise, those with a low-probability V/Q scan and a high clinical suspicion for PE should have further noninvasive imaging of the deep venous system or pulmonary arteries. See Chapter 56 for an expanded discussion of pulmonary scintigraphy. Despite its limitations, V/Q scanning can provide useful information and remains a useful noninvasive screening modality for detecting PE. Although uncommon, a normal perfusion study excludes
1044
embolism, whereas a high-probability V/Q study, in the appropriate clinical setting, allows for a confident enough diagnosis of PE to initiate anticoagulant therapy. Currently, its role in the evaluation of PE is primarily limited to those patients with a high likelihood of having a diagnostic result (i.e., normal or high probability); such patients are generally young individuals with normal chest radiographs and no history of chronic obstructive pulmonary disease.
Helical
CT
Pulmonary
Angiography
Dynamic CT angiography of the pulmonary arteries (CTPA) using multidetector CT (MDCT) has proven accurate in the detection of PE (7). Contiguous or overlapping 1- to 2-mm scans through the entire thorax during injection of 80 to 120 mL of 300 to 350 mg I(iodine)/mL nonionic contrast injected through an 18-gauge or larger intravenous catheter allows routine dense opacification of second- and third-order subsegmental pulmonary arteries. Scans must be interpreted on workstations in a paging or cine mode to allow efficient review and accurate interpretation of the large data sets produced by the current 16- to 64-channel MDCT scanners. Emboli are recognized as intraluminal filling defects (Fig. 14.10) or nonopacified vessels with a convex filling toward the proximal lumen. Secondary findings that can be seen on CT include peripheral oligemia (Westermark sign), pleura-based wedge-shaped consolidation reflecting peripheral hemorrhage or infarct (Fig. 14.11), linear atelectasis, and pleural effusion. The detection of a high-attenuation thrombus in the pulmonary arteries on unenhanced CT in patients with PE has been rarely described. Common diagnostic pitfalls in the detection of PE on CTPA include motion artifact, streak artifact from dense contrast or catheters, partial volume averaging of obliquely oriented vessels, prominent hilar lymphoid tissue, poorly opacified pulmonary veins, mucus-filled bronchi, and regional areas of increased pulmonary arterial resistance from consolidation or atelectasis, all of which can simulate intraluminal arterial filling defects. At present MDCT is widely considered the first-line diagnostic
1045
modality for the evaluation of suspected PE. P.427 Confident detection of a discrete intraluminal filling defect is highly specific for PE. Conversely, multiple studies have shown that the negative predictive value of a good-quality CTPA for PE is greater than 95%. For these reasons, only those patients at high risk for significant morbidity or mortality from recurrent PE (i.e., patients with severe chronic obstructive pulmonary disease or cor pulmonale) should be considered for conventional angiography following a negative or inconclusive CT study; the latter occurs in approximately 5% of patients referred for CTPA, a percentage similar to that of nondiagnostic
pulmonary
arteriograms.
FIGURE 14.10. Pulmonary Embolism on CT Angiography. CT pulmonary angiogram in a 47-year-old man demonstrates filling defects (arrowheads) within the middle lobe and both lower lobe pulmonary arteries, representing pulmonary emboli.
1046
FIGURE 14.11. Pulmonary Infarct (Hampton Hump). Enhanced CT scan in a patient with pulmonary infarction shows a smooth, eccentrically cavitated, pleura-based mass (arrowhead) representing an infarct, with the thrombus visible as a filling defect (arrow) in the lateral segment middle lobe pulmonary artery. An associated right pleural effusion (e) is a common finding in pulmonary infarction.
Although the ability to detect small emboli has improved significantly with MDCT (Fig. 14.12), the main limitation of CTPA remains the reliable detection of small (subsegmental) emboli, although the
1047
frequency and clinical significance of such emboli is the subject of significant debate. In addition to the detection of emboli, up to two thirds of patients with acute chest symptoms who are studied with CTPA to exclude PE have an alternative diagnosis suggested by findings detected on CT, something not possible with techniques that only evaluate the pulmonary vasculature such as perfusion scintigraphy,
MR
angiography,
and
conventional
angiography.
Pulmonary angiography has traditionally been considered to be the gold standard in the diagnosis of PE (8). Digital subtraction angiography is the technique selectively utilized when a definitive diagnosis of PE or DVT cannot be achieved by less invasive means. This study, which requires right heart and pulmonary arterial catheterization with selective injection of nonionic contrast, can be performed safely in a majority of patients. The accuracy of pulmonary arteriography in the diagnosis of PE is high. Based upon clinical follow-up of patients with negative studies, the sensitivity of pulmonary angiography is 98% to 99%, although as with CTPA, the accuracy for the detection of subsegmental PE is closer to 66%. PE is diagnosed on pulmonary angiography when an intraluminal filling defect or the trailing end of an occluding thrombus is outlined by contrast (Fig. 14.13). Secondary signs, including a prolonged arterial phase, diminished peripheral perfusion, and delay in the venous phase, are nonspecific and are not used to diagnose PE. Once a thrombus is unequivocally identified, the study is terminated. The only exception would be a patient who is considered a candidate for surgical thrombectomy or thrombolytic therapy, where precise knowledge of the laterality, location, and extent of the thrombus is required. The overall complication rate of pulmonary angiography is 2% to 5% and can be divided into those related to contrast administration and those secondary to cardiac catheterization and injection of intrapulmonary arterial contrast. Mortality from pulmonary angiography is less than 0.5% and is usually related to sudden RV failure from transient elevation of pulmonary artery pressure secondary to contrast injection. Death from pulmonary angiography
1048
is seen almost exclusively in critically ill patients and those with preexisting severe PAH (pulmonary artery systolic pressure >70 mm Hg) or RV dysfunction (RV end diastolic pressure >20 mm Hg). However, there is no significant increase in the incidence of major, nonfatal reactions in patients with PAH. In addition, the majority of patients with severe RV dysfunction have uneventful studies. When one considers the added safety of selective contrast injections using nonionic contrast agents and the high mortality of untreated PE in this population, pulmonary angiography should be performed in these patients
when
indicated.
Noninvasive Imaging for Deep Venous Thrombosis. The use of noninvasive techniques for the diagnosis of DVT has altered the conventional approach to the evaluation of pulmonary thromboembolic disease (see Chapter 40). Because 90% of PEs arise from the lower extremities, and because the treatment for proximal (i.e., above-the-knee) P.428 DVT is identical to that for proven PE, a confident diagnosis of proximal DVT can provide an endpoint in patient evaluation for thromboembolic disease.
FIGURE
14.12. Subsegmental
Pulmonary
1049
Embolism
on
Isotropic Multidetector CT (MDCT). Axial (A) and coronal reconstructed CT (B) on CT angiography from a 40-slice MDCT scanner show an isolated subsegmental embolus to a branch of the right lower lobe lateral basal segmental artery (arrows) .
FIGURE
14.13. Pulmonary Embolism on Pulmonary
Arteriogram. A frontal radiograph from a left pulmonary arteriogram reveals a large intraluminal filling defect within the main and left interlobar pulmonary artery, diagnostic of pulmonary embolism. Note the typical meniscus of contrast outlining the trailing edge of the thrombus (curved black arrows) and a rim of contrast around the body of the thrombus (small
1050
white
arrows) .
When performed by skilled personnel, compression US has a sensitivity of 90% to 95% and a specificity of 95% to 98% for the diagnosis of acute DVT when compared to contrast venography. False-negative studies occur when DVT is limited to the calf or pelvis, or in patients with duplicated deep venous systems. False-positive studies are seen most often in patients with prior DVT. In addition to providing an accurate diagnosis of the presence of DVT, US offers the advantage of imaging the nonvenous structures in the leg, allowing the radiologist to diagnose conditions that may simulate DVT clinically, such as Baker cysts, enlarged lymph nodes, pseudoaneurysms, and pelvic masses compressing the iliac vein. Although accurate for the diagnosis of proximal DVT, a negative compression US study does not exclude PE. Thus, patients with a negative US study should undergo evaluation of the pulmonary arteries with CT or conventional angiography. Indirect CT venography (CTV), typically performed after contrast injection has been administered for CTPA, has been utilized to allow detection of thigh and pelvic DVT. Axial or helical scans performed from the popliteal fossa to the diaphragm obtained approximately 3 minutes after the initiation of contrast injection for CTPA have been shown in preliminary studies to have a high accuracy in the detection of proximal lower-extremity and pelvic DVT. The addition of CTV to CTPA can provide incremental information for the diagnosis of venous thromboembolic disease, particularly when a proximal DVT is detected in a patient with a poor quality, equivocal, or negative CTPA study. P.429 MR venography and radionuclide scintigraphy can be used to detect DVT but are not used routinely in clinical practice for this purpose. Nonthrombotic pulmonary embolism can occur rarely. The most commonly described conditions are (1) air embolism, usually as a result of air within a venous catheter or air injected during contrast-
1051
enhanced CT; (2) macroscopic fat embolism following long bone fracture, with pulmonary embolization of marrow elements; (3) methylmethacrylate embolization complicating vertebroplasty; and (4) radioactive seed implant embolization from prostate brachytherapy.
FIGURE
14.14. Pulmonary
Arterial
Hypertension.
Posteroanterior (A) and lateral (B) chest radiographs in a 32year-old woman with pulmonary hypertension from chronic pulmonary thromboembolic disease show enlarged main (M), right (R), and left (L) pulmonary arteries with diminutive peripheral vessels.
1052
FIGURE 14.15. Eisenmenger Syndrome. Pulmonary arterial hypertension from left-to-right shunt resulting from atrial septal defect. Enhanced CT scan in a 49-year-old woman with prior repair of an atrial septal defect (A) shows massive dilatation of the main, right, and left pulmonary arteries, reflecting pulmonary arterial hypertension. Scan at the level of the heart (B) shows right ventricular hypertrophy and dilatation as a result of pulmonary
hypertension.
Pulmonary tumor emboli can develop in a small percentage of patients with malignancies such as bronchoalveolar cell carcinoma, breast cancer, hepatoma, and GI malignancies. These tumor emboli may lead to significant respiratory symptoms because of occlusion of small vessels. Imaging features are uncommon but include central pulmonary arterial dilation and enlarged, nodular peripheral pulmonary artery branches on thin-section CT. P.430 In patients suspected of this disorder, aspiration cytology from a wedged pulmonary arterial occlusion (Swan-Ganz) catheter can be useful for diagnosis.
PULMONARY
ARTERIAL
HYPERTENSION
PAH is defined as a systolic pressure in the pulmonary artery exceeding 30 mm Hg, measured either directly, by catheterization of
1053
the pulmonary artery, or estimated by echocardiography. The diagnosis of PAH is usually evident from the clinical history, physical findings, and appearance on chest radiographs. The typical radiographic findings of PAH are enlarged main and hilar pulmonary arteries that taper rapidly toward the lung periphery (Fig. 14.14) . Associated enlargement of the RV, seen on lateral radiographs as prominence of the anterosuperior cardiac margin with obliteration of the retrosternal airspace, is an additional clue to the diagnosis. Occasionally, hypertension-induced atherosclerotic lesions in the large elastic arteries can produce mural calcifications on radiographs or CT, a rare finding that is specific for PAH. A useful measurement for enlargement of the central pulmonary arteries, usually indicating PAH in the absence of a left-to-right shunt, is a transverse diameter of the proximal interlobar pulmonary artery on PA chest radiograph that exceeds 16 mm. Another specific indicator of PAH is a transverse measurement of the main pulmonary artery on CT or MR that exceeds 28.6 mm. However, a normal measurement of the main or right interlobar pulmonary artery does not exclude PAH, as patients with mild or even moderate elevation of pulmonary artery pressure may have normal-size arteries. Those patients with longstanding PAH will develop RV hypertrophy, with eventual RV dilatation and failure (“cor pulmonale―). In addition, MR may also demonstrate intraluminal signal during the early diastolic phase of the cardiac cycle, a finding indicative of turbulent flow caused by the increased vascular resistance that is sometimes seen with marked elevation of pulmonary artery pressure. In addition to PAH, enlargement of the central pulmonary arteries may be seen in conditions associated with increased flow through the pulmonary circulation. This occurs in patients with a high cardiac output, such as anemia, thyrotoxicosis, or those with left-to-right shunts. The latter includes atrial and ventricular septal defects, patent ductus arteriosus, and partial anomalous pulmonary venous return. Early in the course of left-to-right shunts, the pulmonary artery pressure is normal or slightly elevated, because pulmonary vascular resistance drops to compensate for the increased flow. In these patients, there is enlargement of both central and peripheral
1054
pulmonary arteries, producing “shunt vascularity― on chest radiographs. Later, usually in young adulthood, the muscular pulmonary arterioles develop medial hyperplasia and intimal fibrosis, with resultant increased pulmonary vascular resistance (Fig. 14.15) . When this occurs, the chest radiograph demonstrates findings typical of PAH that are indistinguishable from PAH caused by other etiologies.
TABLE 14.4 Causes of Pulmonary Arterial Hypertension
Pulmonary venous Left heart disease
hypertension
L V failure Mitral valve disease Obstruction of pulmonary venous return Fibrosing mediastinitis Cor triatriatum L A myxoma Pulmonary
venoocclusive
hemangiomatosis) Lung disease/chronic
disease
(pulmonary
hypoxemia
Emphysema/chronic bronchitis Cystic lung disease Langerhans cell histiocytosis Lymphangioleiomyomatosis Cystic fibrosis Interstitial fibrosis Idiopathic pulmonary fibrosis Sarcoidosis Radiation fibrosis (rare) Small airways disease Constrictive bronchiolitis Hypoventilation Obesity Chest wall deformity
(kyphoscoliosis)
1055
capillary
Pulmonary arterial disease Left-to-right shunt (Eisenmenger ASD VSD PDA Partial
anomalous
pulmonary
Primary pulmonary hypertension arteriopathy) Pulmonary vasculitis
syndrome)
venous
return
(plexogenic
pulmonary
Connective tissue disease HIV infection Drugs (fenfluramine, dexfenfluramine, “fen-phen―) Chronic pulmonary thromboembolic disease
ASD = atrial septal defect. VSD = ventricular septal defect. PDA = patent ductus arterious.
An increase in resistance to pulmonary blood flow is the most common cause of PAH (Table 14.4). The disorders producing increased pulmonary vascular resistance are pulmonary venous hypertension, parenchymal lung disease, chest wall deformity, diffuse pleural fibrosis, pulmonary arterial disease, and idiopathic pulmonary vascular disease. The most common cause of chronic elevation of pulmonary venous pressure is mitral stenosis, although any impedance to pulmonary venous return to the left heart can produce venous hypertension. Less common entities in this group include chronic LV failure, atrial P.431 myxoma, cor triatriatum, and pulmonary vein stenosis or occlusion. In addition to the characteristic pulmonary arterial changes of PAH, patients may show LV dilatation in LV failure or LA enlargement in mitral stenosis or cor triatriatum. The radiographic signs of pulmonary venous hypertension and pulmonary edema may be seen early in the course of these disorders but are often absent by the
1056
time PAH has developed.
FIGURE 14.16. Chronic Thromboembolic Pulmonary Hypertension. Enhanced CT scan at the level of the main pulmonary artery (A) shows dilated main and left pulmonary arteries, with thrombosis of the truncus anterior branch of the right pulmonary artery (arrow). At the level of the hila (B), there is an eccentric filling defect (arrow) in the right interlobar artery and a weblike filling defect (arrowhead) containing calcification in the left interlobar artery. These findings are characteristic of chronic unresolved emboli.
Parenchymal
lung
disease,
particularly
centrilobular
emphysema
and
diffuse interstitial fibrosis, are common causes of PAH. The mechanisms by which these disorders produce increased vascular resistance include chronic hypoxemia and reflex vasoconstriction and the development of irreversible changes in pulmonary arteriolar caliber, with widespread obliteration of the pulmonary vascular bed. The radiographic findings of emphysema and interstitial fibrosis are usually evident on plain radiographs by the time PAH has developed. Chronic hypoxemia from alveolar hypoventilation is the likely mechanism for PAH that complicates pleural fibrosis, kyphoscoliosis, and the obesity-hypoventilation syndrome. Pleural thickening and kyphoscoliosis are readily evident radiographically. The obesity-
1057
hypoventilation (obstructive sleep apnea) syndrome is usually associated with marked truncal obesity and lungs that are diminished in volume (mostly owing to diaphragmatic elevation) but are normal in
appearance.
Disorders of the pulmonary arteries that produce PAH include chronic PEs, vasculitis, and pulmonary arteriopathy resulting from longstanding increased pulmonary blood flow from left-to-right shunt. Occlusion of lobar and segmental vessels producing PAH can result from failure of pulmonary thromboemboli to lyse or completely recanalize (Fig. 14.16). Rarely, pulmonary vasculitis resulting from diseases such as rheumatoid lung disease or Takayasu arteritis can produce obliteration of the pulmonary vasculature and lead to PAH. The diagnosis of large vessel thromboembolic pulmonary hypertension is usually made by echocardiography, which provides an indirect estimate of pulmonary artery pressure. CT angiographic findings of chronic thromboembolic pulmonary hypertension (CTPH) correlate with conventional angiographic findings and include focal stenoses, bandlike or weblike filling defects, and eccentric wall thickening. Lung windows in patients with CTPH classically demonstrate a pattern of mosaic attenuation, with the hyperlucent regions
demonstrating
attenuated
vascular
markings
(mosaic
oligemia) as compared to areas of increased attenuation that result from hyperemia from intact pulmonary artery branches. Idiopathic or primary
pulmonary
hypertension
encompasses
diseases
of the pulmonary arterioles and venules that are not attributable to other etiologies and have characteristic histologic findings. Plexogenic pulmonary arteriopathy, recurrent microscopic PE, and pulmonary venoocclusive disease (PVOD) are the three diseases that comprise this category. Plexogenic pulmonary arteriopathy is a disease of young women in whom medial hypertrophy and intimal fibrosis obliterate the muscular arteries. Dilated vascular channels within the periphery of the obliterated vessel produce the plexogenic lesions seen on biopsy in virtually all patients with this disease. Progressive dyspnea and fatigue develop with characteristic physical findings of PAH and cor pulmonale. In plexogenic pulmonary
1058
P.432 arteriopathy, pulmonary perfusion scans typically show normal perfusion or small, nonsegmental peripheral perfusion defects, allowing distinction from large-vessel thromboembolic disease. Microembolic disease is clinically and radiographically indistinguishable from plexogenic arteriopathy. In this entity, plexogenic lesions within arterioles are absent. Perfusion scans are more likely to show small perfusion defects in this disorder. The presence of small microemboli histologically is not a distinguishing feature, because in situ thrombosis within diseased arterioles can have a similar appearance. In PVOD, the obliteration of small intrapulmonary venules results in interstitial pulmonary edema. A condition related to PVOD is pulmonary capillary hemangiomatosis (PCH), which is characterized by the proliferation of capillaries throughout the pulmonary interstitium, resulting in venular obstruction. The transmission of increased pressure to the arterial side leads to medial hypertrophy and obliteration of vessel lumina with resultant arterial hypertension. Chest radiographs often show interstitial or airspace pulmonary edema with a normal heart size. Perfusion lung scanning is usually normal or shows small peripheral nonsegmental defects. The combination of pulmonary edema with a normal heart size, absent findings for pulmonary venous hypertension, normal PCWP, and the insidious onset of dyspnea should suggest this diagnosis rather than left heart failure, mitral valve disease, or large-vessel pulmonary venous occlusion. Thinsection CT features of PVOD and PCH are those of pulmonary venous hypertension and include interlobular septal thickening, centrilobular nodular ground-glass opacities, and pleural effusions (9). A definitive diagnosis can only be made by characteristic findings on open lung biopsy. The prognosis is universally poor, with most patients succumbing to their disease within 2 years of diagnosis.
REFERENCES 1. Pistolesi M, Miniati M, Milne ENC, et al. The chest roentgenogram in pulmonary edema. Clin Chest Med
1059
1985;6:315–344. 2. Ketai L, Godwin D. A new view of pulmonary edema and acute respiratory distress syndrome. J Thorac Imaging 1998;13:147–171. 3. Milne ENC, Pistolesi M, Miniati M, et al. The radiologic distinction of cardiogenic and noncardiogenic edema. AJR Am J Roentgenol
1985;144:879–894.
4. Albelda SM, Gefter WB, Epstein DM, et al. Diffuse pulmonary hemorrhage: a review and classification. Radiology 1985;154:289–297. 5. The PIOPED investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism: results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). JAMA 1990;263:2753–2759. 6. Buckner CB, Walker CW, Purnell GL. Pulmonary embolism: chest radiographic abnormalities. J Thorac Imaging 1989;4:23–27. 7. Schoepf UJ, Costello P. CT angiography for diagnosis of pulmonary embolism: state of the art. Radiology 2004;230:329–337. 8. Stein PD, Athanasoulis C, Alavi A, et al. Complications and validity of pulmonary angiography in acute pulmonary embolism. Circulation 1992;85:462–468. 9. Hansell DM. Small-vessel diseases of the Lung: CT-Pathologic correlates.
Radiology
2002;225:639–653.
1060
1061
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section III - Pulmonary > Chapter 15 Pulmonary Neoplasms
Chapter
15
Pulmonary
Neoplasms
Jeffrey S. Klein
THE
SOLITARY
PULMONARY
NODULE
The radiologic evaluation of a solitary pulmonary nodule (SPN) is one of the most common and most difficult diagnostic dilemmas in thoracic radiology (1 ). The prevalence of SPNs has increased recently as a result of the growing use of chest CT—specifically its use to screen for occult lung cancer. Before embarking on a detailed diagnostic evaluation of an SPN, one must determine whether the nodular opacity seen on the chest radiograph is real or artifactual. If the opacity is judged to represent a real lesion, the radiologist must then determine whether it is an SPN. First, is the lesion truly solitary? Not uncommonly, a dominant pulmonary nodule on chest radiographs is associated with smaller nodules or nodules that are obscured by the heart or the hemidiaphragms. A careful search on posteroanterior (PA) and lateral chest radiographs will identify most of these patients, although CT may be necessary to identify additional nodules not seen on chest radiographs. Multiple pulmonary nodules of similar size and appearance are almost always metastases or granulomas and require an evaluation that differs from that for a solitary lesion. Second, is the lesion intrapulmonary? Intrapulmonary lesions are discrete opacities that are completely circumscribed by aerated lung on both frontal and lateral radiographs. A pleural or mediastinal
1062
lesion may be outlined by lung when it projects inward. However, the base of the lesion, which forms obtuse angles with the lung, is not outlined by lung when it arises from the pleura or mediastinum. Skin lesions and chest wall lesions, including bone islands and healing rib fractures, can also mimic intrapulmonary nodules. Some skin lesions are circumscribed by air, but a careful physical examination of the patient, usually following the chest radiograph, will reveal the surface lesion responsible for the “nodule― seen on chest film. One of the most troublesome skin “lesions― to distinguish from a true pulmonary nodule is the nipple shadow. This is usually readily identified by performing chest films with nipple markers. Occasionally, chest fluoroscopy or CT will be necessary for confident localization of a nodular opacity seen on conventional
radiographs.
Finally, is the lesion a nodule? A nodule is a discrete round or oval opacity 4 to 30 mm in diameter; linear or angular opacities are not nodules and represent scars or areas of linear atelectasis. Threedimensional analysis of the shape of an SPN on volumetric CT can help distinguish flat scars from true SPNs. When a focal opacity is seen in the lung apices, an apical lordotic film or CT scan may be necessary to distinguish a linear from a nodular opacity. A round opacity greater than 3 cm in diameter is termed a mass. Because the majority of lung masses in patients over the age of 35 represent bronchogenic carcinoma, these lesions are not considered SPNs. Once an SPN has been identified, the radiologist should initiate a series of investigations to determine whether the nodule is definitely benign or suspicious for malignancy (i.e., indeterminate). This stepwise approach is summarized in Fig. 15.1 .
1063
FIGURE 15.1. Imaging Algorithm for the Solitary Pulmonary Nodule. SPN, solitary pulmonary nodule; Ca+ + , calcification; TNB, transthoracic needle biopsy; CXR, chest x-ray; GGO, ground-glass opacity; Bx, biopsy; F/U, follow-up.
P.434
Clinical
Factors
Before considering the radiologic characteristics used to distinguish benign from indeterminate nodules, several important clinical factors may be helpful in making this distinction. In a patient under the age of 35, particularly a nonsmoker without a history of malignancy, an SPN is invariably a granuloma, hamartoma, or inflammatory lesion. These nodules can be followed with plain radiographs to confirm their benign nature. Patients over the age of 35, particularly those who are current or recent cigarette smokers, have a significant incidence of malignant SPNs: approximately 50% of noncalcified SPNs in patients over 50 years of age are malignant at thoracotomy. Therefore, as a rule, an SPN in a patient over 35 years of age should never be followed radiographically without tissue confirmation unless a benign pattern of calcification or the presence of intralesional fat is identified on radiographs or HRCT, or there has been radiographically documented lack of growth over a minimum of 2 years. There are exceptions to this rule: a history of smoking, asbestos exposure, or both raises the level malignancy in a patient with an SPN. Alternatively, if from an area where histoplasmosis or tuberculosis is
1064
cigarette of concern for the patient is endemic, the
likelihood of a granuloma is greater; in such patients, a conservative approach may be warranted. Finally, the finding of an SPN in a patient with an extrathoracic malignancy raises the possibility of a solitary pulmonary metastasis. If the lung is the sole site of metastatic disease, distinguishing between a primary bronchogenic carcinoma and a pulmonary metastasis is usually not important, as many surgeons will resect a solitary pulmonary metastasis. An SPN that arises more than 2 years after the diagnosis of an extrathoracic malignancy is almost always a primary lung tumor rather than a metastasis; breast carcinoma and melanoma are notable exceptions to this rule.
Growth
Pattern
Pulmonary malignancies grow at a relatively predictable rate. The growth rate of an SPN is usually expressed as the doubling time , or the time it takes for a nodule to double its volume. For a sphere, this corresponds to a 25% increase in diameter. Although some benign lesions (mostly hamartomas and histoplasmomas) may exhibit a growth rate similar to that of malignant lesions, the absence of growth or an extraordinarily slow or rapid rate of growth is reliable evidence that an SPN is benign. Studies have shown that bronchogenic carcinoma has a doubling time of between 1 month and 2 years. Therefore, a doubling time of less than 1 month or greater than 2 years reliably characterizes a lesion as benign. Infectious lesions and rapidly growing metastases from choriocarcinoma, seminoma, or osteogenic sarcoma comprise the majority of rapidly growing solitary nodules, while lack of growth or a doubling time exceeding 2 years is seen in hamartomas and histoplasmomas. However, there are exceptions to this rule. Giant cell carcinoma, a subtype of large cell carcinoma, and pulmonary carcinosarcomas and blastomas may have a doubling time of less than 1 month. Even more common pulmonary malignancies, such as the occasional adenocarcinoma or carcinoid tumors, may have a doubling time of greater than 2 years. Any malignancy that hemorrhages into its substance will appear to enlarge rapidly. There are two important caveats to using the growth rate of an SPN to
1065
determine benignity. The first is that the growth rate of an SPN that is not visible on prior radiographs cannot be estimated, since noncalcified nodules less than 1 cm in diameter are not usually visible on conventional radiographs. Most important, with few exceptions, a patient over the age of 35 with a noncalcified SPN should not be evaluated prospectively to determine benignity by following the growth rate on serial chest radiographs. The assessment of growth rate to determine benignity should only be used retrospectively in comparing the size of an SPN with prior radiographs from at least 2 years previously. In patients with clinical and imaging characteristics suggesting a probably benign SPN, particularly for lesions 15 H) enhancement of a solid nodule 6 to 30 mm in diameter after intravenous iodinated malignancy (sensitivity = 98%).
contrast
effectively
excludes
PET using fluorine-18-labeled fluorodeoxyglucose (FDG) has shown a
1070
high accuracy in the distinction between benign and malignant SPNs (Fig. 15.6 ) (3 ). For lesions >10 mm diameter, the sensitivity and specificity of FDG-PET is 97%, with a specificity of 78%, mostly as a result of inflammatory lesions such as active granulomas that are FDG-avid. False-negative PET studies are seen in patients with lesions 2 cm from the tracheal carina T3 Any tumor with localized chest wall, diaphragmatic, mediastinal pleural, or pericardial invasion; the tumor may be 3 cm of contact between tumor and pleura, pleural thickening adjacent to the mass, and infiltration of extrapleural fat. Extrathoracic extension of the mass or rib destruction are specific but insensitive CT findings for chest wall invasion (Fig. 15.16 ). Additional techniques that have been described to assess parietal pleural invasion by tumor include assessment of respiratory movement on dynamic expiratory CT and the use of diagnostic pneumothorax.
1102
FIGURE 15.16. T3 Tumor With Localized Chest Wall Invasion. A CT scan through a large left lower lobe adenocarcinoma shows invasion of the posterior chest wall. A portion of the chest wall was removed en bloc with the tumor.
P.448 MR is equal to CT in its ability to diagnose chest wall invasion. T2WIs show excellent contrast between tumor and chest wall muscle and fat and are used in selected cases to detect chest wall invasion. MR also detects early obliteration of the high-signal extrapleural fat that may be an early finding in chest wall invasion. Coronal MR images are useful in superior sulcus tumors to determine chest wall, brachial plexus, or subclavian artery involvement.
Mediastinal
Invasion
1103
Tumor invasion of the mediastinum with involvement of the heart, great vessels, trachea, or esophagus (T4 tumor) precludes resection. Localized invasion of the mediastinal pleura or pericardium (T3 tumor) does not prevent resection, although extensive invasion with replacement of mediastinal fat does. On conventional radiographs, a mediastinal mass, mediastinal widening, or diaphragmatic elevation (from phrenic nerve involvement) suggests invasion. As with the diagnosis of chest wall invasion, CT demonstration of tumor mass in contiguity with the mediastinal pleura or thickening of the mediastinal pleura does not necessarily indicate mediastinal extension or unresectability. However, a significant mediastinal mass that is contiguous with a lung tumor, compresses mediastinal vessels or esophagus, or replaces mediastinal fat strongly suggests this diagnosis. Other findings that may suggest mediastinal invasion include: (1 ) obliteration of the fat plane adjacent to the descending aorta or other mediastinal vessels, (2 ) tumor contacting more than one fourth of the circumference of the aortic wall, or (3 ) tumor contacting more than 3 cm of the mediastinum. If none of these findings are present, the tumor is potentially resectable, even though 29% of resectable lesions lacking any of these findings are found to invade the mediastinum locally (Fig. 15.17 ) (8 ).
1104
FIGURE
15.17. CT to Assess Mediastinal Invasion of Lung
Cancer. In a patient with a medial segment, middle lobe, non–small cell carcinoma, CT shows a fat plane (arrows ) between the lesion and the heart. Surgical resection confirmed the absence of pericardial
or
mediastinal
invasion.
As with CT, MR is incapable of accurately demonstrating mediastinal pleural invasion or minimal invasion of mediastinal fat. Mediastinal invasion can be diagnosed with a reasonable degree of accuracy when there is significant obliteration of fat planes or compression or displacement of mediastinal vessels. In one study, MR was found to be significantly more accurate than CT in diagnosing mediastinal invasion, but this result was based on a small number of patients who had invasion, and the study predated the advent of isotropic MDCT (8 ). Other studies have shown no significant advantage of MR over CT for this purpose. MR is occasionally performed when vascular invasion is suspected and is likely more accurate than CT in this
regard.
Central
Airway
Involvement
Tumors that extend into a main bronchus within 2 cm of the tracheal
1105
carina (T3 tumors) are resectable. Although tracheal carinal involvement (T4 tumor) (Fig. 15.18 ) can be treated by carinal resection with end-to-side anastomosis of the remaining bronchus to the tracheal stump (“sleeve pneumonectomy―), most surgeons would consider this an unresectable tumor. Although plain films can occasionally demonstrate P.449 a mass within the main bronchus or trachea, CT is more accurate in assessing the relationship of the mass to the tracheal carina (Fig. 15.18B ). However, CT is known to underestimate the mucosal or submucosal extent of tumor as seen bronchoscopically. Therefore, any patient with a central lesion should undergo bronchoscopy to determine the proximal extent of the tumor, unless CT shows obvious carinal or tracheal invasion.
FIGURE
15.18. Tracheal Carinal Involvement in Squamous Cell
Carcinoma. A. A frontal chest radiograph in a middle-aged woman with hemoptysis shows a mass in the lower right hilum (large arrow ) with right lower lobe atelectasis (small arrows ). B . A CT scan demonstrates a mass surrounding the main bronchi, with irregular narrowing of the right main bronchial lumen and infiltration of the tracheal carina (arrow ). An enlarged precarinal lymph node (open arrow ) and small bilateral pleural effusions are also seen. Bronchoscopy revealed invasion of the tracheal carina by squamous cell carcinoma.
1106
Multiple Tumor Nodules in the Same Lobe The recent update to the International Staging System for non-small cell lung cancer classifies cases of satellite tumor nodules in the same lobe as the primary tumor as T4 disease, based on prognosis. Despite their IIIb categorization, most patients with multiple nodules in the same lobe and absent nodal or distant metastases will undergo attempts at curative resection.
Pleural
Effusion
Malignant pleural effusion (T4 tumor) precludes curative resection of a tumor. In a patient with bronchogenic carcinoma, pleural effusion can occur for a variety of reasons, including pleural invasion, obstructive pneumonia, and lymphatic or pulmonary venous obstruction by tumor. Although the presence of effusion associated with lung cancer indicates a poor prognosis, only those patients with tumor cells in the pleural fluid or on pleural biopsy are considered unresectable. Other patients with effusion are considered to have “resectable― lesions, despite their poor prognosis. Usually plain radiographs, including decubitus films, are sufficient to diagnosis a pleural effusion. Thoracentesis with cytologic examination and/or pleural biopsy is necessary for definitive diagnosis of malignant pleural involvement. Pleural thickening >1 cm, lobulated pleural thickening, or circumferential pleural thickening (i.e., involvement of the mediastinal pleura) on CT or MR strongly suggests pleural invasion (Fig. 15.19 ). While PET can be useful in characterizing pleural effusions in patients with lung cancer as malignant, caution is advised in patients who have undergone prior pleurodesis, as focal plaques from intrapleural talc administration can be FDG avid on PET.
Lymph
Node
Metastases
(N)
Selected patients with ipsilateral mediastinal or subcarinal node metastases are classified as N2 and are considered potentially resectable. However, patients with N2 nodal disease (i.e., stage IIIa
1107
lung cancer) have a significantly worse prognosis as compared to patients classified with stage IIIa disease because of a T3 lesion. Those patients with N2 disease from nonbulky intracapsular nodal metastases limited to one mediastinal nodal station have the best 5year survival rates following extensive mediastinal nodal dissection. Contralateral hilar or mediastinal, supraclavicular or infraclavicular nodal metastases represent N3 disease and are unresectable (Fig. 15.20 ). The detection of a large mediastinal mass on chest radiograph in a patient with lung cancer requires mediastinoscopic or transthoracic biopsy confirmation of tumor invasion before deeming the patient unresectable. A P.450 normal chest radiograph or the suggestion of hilar or mediastinal adenopathy should prompt a chest CT to assess the status of the lymph nodes. No single measurement allows completely accurate distinction of normal from malignant nodes. This is because malignant involvement does not always enlarge the lymph node (producing
false-negative
findings
and
reducing
sensitivity),
while
enlarged nodes in patients with lung cancer may represent reactive hyperplasia rather than tumor replacement (producing false-positive findings and reducing specificity). If a small nodal diameter (5 mm) is used as the dividing point between benign and malignant, sensitivity will be excellent but specificity will be low. However, choosing a large nodal diameter (2 cm) increases specificity but decreases sensitivity. Most radiologists use a short-axis nodal diameter of 1 cm because this value achieves the best compromise of sensitivity and specificity.
1108
FIGURE 15.19. Pleural Metastases (T4) in Bronchogenic Carcinoma. CT images through the lower chest in a patient with a large right lower lobe tumor show a large pleural effusion with discrete pleural nodules (arrows ) representing pleural metastases.
FIGURE
15.20. N3 Nodal Metastases in Bronchogenic
Carcinoma. A. CT scan through the aortic arch in a 57-year-old man shows a right hilar mass (white arrow ) invading the distal trachea and mediastinal fat (T4 tumor). There are enlarged prevascular lymph nodes (N2) (black arrow ). B . Scan at the supraclavicular level shows marked bilateral supraclavicular
1109
lymphadenopathy
(arrows ) representing N3 disease.
Recent studies have shown that CT is relatively inaccurate in determining the nodal status of the patient with lung cancer. Both sensitivity and specificity for nodal metastases, when a short-axis diameter of 1 cm or greater is used as abnormal, are approximately 60% to 65% on a patient-by-patient basis and may be even lower when looking at individual nodal stations (9 ). Although CT cannot be considered accurate enough to determine with certainty whether or not mediastinal lymph nodes are involved by tumor, it can provide information of value in guiding invasive staging procedures such as mediastinoscopy, transcarinal Wang biopsy, endoscopic US–guided biopsy, and transthoracic or open biopsy. As discussed earlier, PET—particularly integrated PET-CT—provides superior accuracy in nodal staging of lung cancer (1 0 ). P.451 In select institutions, mediastinoscopy complements CT in the nodal staging of lung cancer. Most patients with enlarged mediastinal nodes on CT that are accessible to transcervical mediastinoscopy (pretracheal, anterior subcarinal, and right tracheobronchial nodes) should have mediastinoscopy and biopsy. The decision of whether patients with negative CT studies for nodal enlargement should undergo mediastinoscopy depends on the local surgical practice. In patients with small, peripheral lung nodules, mediastinal metastases are uncommon, and thoracotomy may be warranted without prior CT or mediastinoscopy, but this remains controversial. Patients with borderline pulmonary function benefit most from mediastinoscopy, because a positive mediastinoscopic biopsy almost certainly precludes any attempt at resection. The accuracy of MR is equal to that of CT in the diagnosis of mediastinal lymph node metastases, although it is rarely used for this purpose alone. There are specific advantages and disadvantages of MR in characterizing mediastinal lymph nodes. Clusters of normalsize nodes may be mistaken for a single enlarged nodal mass because of the limited spatial resolution of MR. MR is incapable of
1110
demonstrating calcification within nodes, which is diagnostic of benign lymph node disease. Aortopulmonary window and subcarinal nodes are nicely demonstrated on coronal MR images, although coronal reconstructions from MDCT can demonstrate enlarged nodes in these regions with equal efficacy.
Metastatic
Disease
(M)
Each patient with proven lung cancer should be carefully evaluated for the presence of distant metastases (M1). Unequivocal evidence of metastases can obviate an unnecessary thoracotomy. Common sites of extrathoracic spread in patients with lung cancer include lymph nodes, liver, adrenal gland, bone, and brain. Metastases to lobes outside the primary lobe or to the other lung, although intrathoracic, are also considered M1 disease. Involvement of these sites probably represents hematogenous spread of tumor from the lung. CT of the chest and upper abdomen is part of the initial evaluation in virtually all patients evaluated for bronchogenic carcinoma. This is adequate for assessing the liver, spleen, adrenal glands, and upper abdominal lymph nodes for evidence of metastases. US or MR may be used to distinguish soft tissue hepatic masses from incidental cysts. Technetium-99m-methylene diphosphonate radionuclide bone scanning or whole-body FDG-PET imaging is used to detect bone metastases, with preliminary studies suggesting that PET is as sensitive but more specific than bone scanning (7 ). Plain films are obtained to assess specific foci of abnormally increased bone tracer uptake or to evaluate localized bone pain. Imaging of the brain is routinely performed in patients with symptoms or signs suggesting intracranial metastases. This usually involves MR or contrast-enhanced head CT. Head scanning in patients without clinical evidence of CNS involvement is somewhat more controversial. Because virtually all patients with isolated or asymptomatic brain metastases are found to have adenocarcinoma or large cell carcinoma, patients with these subtypes of bronchogenic carcinoma should have head CT scans, regardless of
1111
the clinical findings, to identify silent metastases. Patients with positive findings can be spared an unnecessary thoracotomy. Approximately 60% to 65% of patients with small cell carcinoma have metastatic disease at the time of diagnosis. Because it is likely that all patients with small cell carcinoma have gross or microscopic metastatic foci at presentation, these patients are generally not candidates for curative surgical resection. However, accurate staging of these patients for extrathoracic involvement determines prognosis and allows for proper assessment of response to chemotherapy. An additional reason for extrathoracic staging of small cell carcinoma is the ability to manage localized bone or soft tissue involvement with radiation or resection. Adrenal masses are seen in approximately 10% of patients undergoing staging CT examinations for bronchogenic carcinoma. However, approximately 5% of normal individuals are known to have benign adrenal cortical adenomas. In fact, isolated adrenal masses in patients with non–small cell bronchogenic carcinoma are twice as likely to be adenomas than metastases. In many patients, the adrenal mass is the only extrathoracic site of abnormality, making accurate diagnosis of the adrenal mass crucial in determining management. Methods used to distinguish adenomas from malignant (primary or metastatic) adrenal lesions include CT, chemical-shift MR, FDG-PET, and fine-needle aspiration biopsy (see Chapter 33 ). The combined ability of unenhanced CT to detect lipid-rich adenomas (≤10 H) and delayed enhanced CT to detect lipid-poor adenomas (≥ 60% relative washout at 15 minutes) has been utilized with high accuracy to distinguish between adenomas and malignant adrenal lesions (Fig. 15.21 ). Chemical-shift MR is rarely used nowadays to characterize adrenal lesions. PET has a sensitivity approaching 100% for detecting adrenal metastases, such that a negative study effectively excludes this possibility. However, adenomas can be FDG avid and produce false-positive studies; therefore, isolated, FDG-positive adrenal lesions may require biopsy for definitive characterization.
1112
TRACHEAL
AND
BRONCHIAL
Tracheal
Neoplasms.
MASSES
Intratracheal masses may be divided into neoplastic (1 1 ) and nonneoplastic masses. Primary tracheal tumors are rare; however, 90% of all primary tracheal tumors in adults are malignant. The majority of primary tracheal malignancies arise from P.452 tracheal epithelium or mucous glands (90%); the remainder arise from the mesenchymal elements of the tracheal wall (10%). Squamous cell carcinoma is the most common primary tracheal malignancy, accounting for at least 50% of all malignant tracheal neoplasms (Fig. 15.22 ). These tumors affect middle-aged male smokers and are associated with laryngeal, bronchogenic, or esophageal malignancies in up to 25% of cases. The majority arise in the distal trachea within 3 to 4 cm of the tracheal carina, with the cervical trachea the next most common site. Cough, hemoptysis, dyspnea, and wheezing are common presenting symptoms. Patients may be mistakenly treated for asthma before the correct diagnosis is made. Adenoid cystic carcinoma (formerly called cylindroma) is a malignant neoplasm that arises from the tracheal salivary glands and accounts for 40% of primary tracheal malignancies. This neoplasm tends to involve the posterolateral wall of the distal two thirds of the trachea or main or lobar bronchi.
1113
FIGURE
15.21. Adrenal Mass Characterization by Delayed
Enhanced CT. In a patient with a cavitary right upper lobe cancer (A) , dynamic enhanced scan (B) , and 10-minute delayed enhanced (C) scans show a washout percentage of 90% ([74–71/74 × 100), which is indicative of an adenoma.
The diagnosis of a primary tracheal malignancy is rarely made prospectively on chest radiographs, although well-penetrated radiographs can demonstrate distortion of the tracheal air column by a mass. CT typically shows a lobulated or irregular soft tissue mass that eccentrically narrows the tracheal lumen and has a variable extraluminal component (Fig. 15.22B ). Masses >2 cm in diameter are likely to be malignant, while those Table of Contents > Section III - Pulmonary > Chapter 16 Pulmonary Infection
Chapter
16
Pulmonary
Infection
Timothy J. Higgins Jeffrey S. Klein
INFECTION IN THE NORMAL HOST The bronchopulmonary system is open to the atmosphere and therefore is relatively accessible to airborne microorganisms. Multiple host defense mechanisms exist at the level of the pharynx, trachea, and central bronchi. When these mechanisms fail, pathogenic organisms can penetrate to the small distal bronchi and the pulmonary parenchyma. Once the invading organisms penetrate the parenchyma, there is activation of both the cellular and humoral immune systems. This response may manifest clinically and radiographically as pneumonia, and in a normal host will often lead to eradication or at least suppression of the infecting organisms. If the immune response is impaired, a lower respiratory tract infection may lead to a very severe illness and often death, despite appropriate antibiotic therapy.
Mechanisms of Disease Radiographic Patterns
and
Microorganisms responsible for producing pneumonia enter the lung and cause infection by three potential routes: via the tracheobronchial tree, via the pulmonary vasculature, or via direct
1133
spread from infection in the mediastinum, chest wall, or upper abdomen. Infection via the tracheobronchial tree is generally secondary to inhalation or aspiration of infectious microorganisms and can be divided into three subtypes based on gross pathologic appearance and radiographic patterns: lobar pneumonia, lobular or bronchopneumonia, and interstitial pneumonia. As will be discussed in later sections, certain organisms will typically produce one of these three patterns, although there may be considerable overlap. Lobar pneumonia is typical of pneumococcal pulmonary infection. In this pattern of disease, the inflammatory exudate begins within the distal airspaces. The inflammatory process spreads via the pores of Kohn and canals of Lambert to produce nonsegmental consolidation. If untreated, the inflammation may eventually involve an entire lobe (Fig. 16.1). Because the airways are usually spared, air bronchograms are common and significant volume loss is unusual (see Table 12.7) . Bronchopneumonia is the most common pattern of disease and is most typical of staphylococcal pneumonia. In the early stages of bronchopneumonia, the inflammation is centered primarily in and around lobular bronchi. As the inflammation progresses, exudative fluid extends peripherally along the bronchus to involve the entire pulmonary lobule. Radiographically, multifocal opacities that are roughly lobular in configuration produce a “patchwork quilt― appearance because of the interspersion of normal and diseased lobules (Fig. 16.2). While bronchopneumonia is the most common cause of multifocal patchy airspace opacities, there is a broad list of differential diagnostic considerations (see Table 12.9). Exudate within the bronchi accounts for the absence of air bronchograms in bronchopneumonia. With coalescence of affected areas, the pattern may resemble lobar pneumonia. In interstitial pneumonia, seen in viral and mycoplasma infection, there is inflammatory thickening of bronchial and bronchiolar walls and the pulmonary interstitium. This results in a radiographic pattern of airways thickening and reticulonodular opacities (see Table
1134
12.12). Air bronchograms are absent because the alveolar spaces P.461 remain aerated. Segmental and subsegmental atelectasis from small airways obstruction is common.
FIGURE 16.1. Pneumococcal Pneumonia. Posteroanterior (A) and lateral (B) radiographs in a 57-year-old man with fever, chills, and productive cough demonstrate airspace opacification in the right upper lobe with air bronchograms. Sputum culture was positive for Streptococcus pneumonia. CT scan in another patient with pneumococcal pneumonia (C) shows dense multifocal segmental airspace opacification in the upper lobes. Note the lobular pattern of consolidation in the right upper lobe and superior segment of the right lower lobe (arrows),
1135
reflecting
bronchopneumonia.
The spread of infection to the lung via the pulmonary vasculature usually occurs in the setting of systemic sepsis. The pattern of parenchymal involvement is patchy and bilateral. The lung bases are most severely involved, because blood flow is greatest in the dependent portions of the lungs. Pulmonary infection from direct spread usually results in a localized parenchymal process adjacent to an extrapulmonary source of infection. If an organism causes extensive parenchymal necrosis, abscess formation may result.
Bacterial
Pneumonia
Gram-Positive
Bacteria
Streptococcus
pneumoniae
(pneumococcus)
S pneumoniae is a gram-positive organism that may cause infection in healthy individuals but is much more commonly seen in the elderly, alcoholics, and other compromised hosts. Patients with sickle cell disease or who have undergone splenectomy are at particular risk for severe pneumococcal pneumonia. Pneumococcal pneumonia tends to begin in the lower lobes or the posterior segments of the upper lobes. Initially there is involvement of the terminal airways, but rather than remaining localized to this site, there is rapid development of an airspace inflammatory exudate. The spread of infection to contiguous airspaces via interalveolar connections accounts for the nonsegmental and homogeneity of the resultant consolidation.
distribution
The typical radiographic appearance of acute pneumococcal pneumonia is lobar consolidation (Fig. 16.1). Air bronchograms are usually evident. Cavitation in pneumococcal pneumonia is rare, with the exception of infections caused by serotype 3. Uncomplicated parapneumonic effusion or empyema may be seen in up to 50% of patients. With appropriate therapy, complete clearing may be seen in
1136
10 to 14 days. In older patients or those P.462 with underlying disease, complete resolution may take 8 to 10 weeks.
FIGURE
16.2.
Pseudomonas
aeruginosa Pneumonia.
A.
Frontal radiograph in an HIV-positive man with fever and progressive respiratory symptoms shows multifocal airspace opacities with dense apical opacification with cavitation (arrows) . B . A CT scan through the apices shows airspace opacification with left apical cavitation. C . A scan at the level of the tracheal carina shows airspace disease in the anterior segments of right and left upper lobes with sparing of the dependent portions of lung. Bronchoscopy revealed Pseudomonas.
Patients with pneumococcal pneumonia occasionally present with atypical radiographic patterns of disease (1). Patchy lobular opacities
1137
similar to those seen with bronchopneumonia (Fig. 16.1C) or rarely, a reticulonodular pattern may be seen. In some patients, the atypical appearance may relate to the presence of preexisting lung disease (e.g., emphysema), partial treatment, or an impaired immune response (e.g., AIDS). In children, pneumococcal pneumonia may present as a spherical opacity (“round pneumonia―) simulating a parenchymal mass. Staphylococcus aureus pneumonia is most common in hospitalized and debilitated patients. It may also develop following hematogenous spread to the lung in patients with endocarditis or indwelling catheters and intravenous drug users. Community-acquired infection may complicate influenza or other viral pneumonias. S aureus typically produces a bronchopneumonia and appears radiographically as patchy opacities (Fig. 16.3). In severe cases, the opacities may become confluent to produce lobar opacification. Because the inflammatory exudate fills the airways, air bronchograms are rarely seen. In adults, the process is often bilateral and may be complicated by abscess formation in 25% to 75% of patients. In patients who develop pulmonary infection from hematogenous seeding, one sees multiple bilateral poorly defined nodular opacities that eventually become more sharply defined and cavitate. Parapneumonic effusion and empyema are common. Pneumatocele formation is common in children and may lead to pneumothorax. Pneumatoceles may be distinguished from abscesses by their thin walls, rapid change in size, and tendency to develop during the late phase of infection.
Streptococcus
pyogenes
Acute streptococcal pneumonia is rarely seen today, though it can occasionally complicate viral infection or streptococcal pharyngitis. Its radiographic appearance is similar to that of staphylococcal pneumonia, with lobular or segmental lower lobe opacities. The process may be complicated by abscess formation and cavitation; empyema is relatively common.
1138
Bacillus
anthracis
Anthrax is caused by a sporulating gram-positive bacillus that is distributed worldwide. Naturally occurring inhalational anthrax is rare; however, anthrax has been used as an agent of bioterrorism in the United States. The primary radiographic manifestations of inhalational anthrax are related to the underlying pathology of hemorrhagic lymphadenitis and mediastinitis accompanied by hemorrhagic pleural effusions. Conventional radiographs demonstrate mediastinal widening, hilar enlargement, and often pleural effusion. Frank areas of consolidation are not usually present, but peribronchial opacities may be seen. CT scans of recent bioterrorism victims, performed in 2001 without intravenous contrast, demonstrated high-attenuation lymphadenopathy and pleural effusions secondary to hemorrhage. CT scans may show extensive adenopathy in the setting of P.463 normal radiographs and should be obtained if the suspicion of anthrax is high (2) .
FIGURE
16.3.
Staphylococcus
aureus Pneumonia. CT scans
at the top of the aortic arch (A) and central pulmonary arteries (B) show a combination of abscess and cavity formation (arrowheads) and lobular consolidation (arrows). Sputum cultures showed S aureus pneumonia.
1139
Gram-Negative
Bacteria
Gram-negative bacteria are increasingly important causes of pneumonia in hospitalized patients, accounting for over 50% of nosocomial pulmonary infections. While gram-negative organisms may be isolated from only a small percentage of healthy individuals, the isolation rate in hospitalized and severely ill patients ranges from 40% to 75%. The organisms most often responsible for pneumonia include members of the Enterobacteriaceae family (Klebsiella, Escherichia coli, Proteus); Pseudomonas aeruginosa; Haemophilus influenzae; and Legionella pneumophila (3) . The radiographic appearance of gram-negative bacterial pneumonia varies from small ill-defined nodules to patchy areas of opacification that may become confluent and resemble lobar pneumonia. Involvement is usually bilateral and multifocal, and the lower lobes are most frequently affected. Abscess formation and cavitation are relatively common. Parapneumonic effusion is common and is often complicated
by
empyema
Klebsiella
formation.
pneumoniae
Klebsiella pneumonia occurs predominantly in older alcoholic men and debilitated hospitalized patients. Radiographically it appears as a homogeneous lobar opacification containing air bronchograms. Three features help distinguish it radiographically from pneumococcal pneumonia: (1) the volume of the involved lobe may be increased by the exuberant inflammatory exudate, producing a bulging interlobar fissure; (2) an abscess may develop, with cavity formation, which is uncommon in pneumococcal pneumonia; and (3) the incidence of pleural effusion and empyema is higher. Pulmonary gangrene may be seen but is uncommon.
Haemophilus
influenzae
In adults, H influenzae infection is most common in patients with chronic obstructive pulmonary disease (COPD), alcoholism, diabetes
1140
mellitus, and those with an anatomic or functional splenectomy. It most often causes bronchitis, although it may extend to produce bilateral lower lobe bronchopneumonia. Pseudomonas aeruginosa pneumonia most often affects debilitated patients, particularly those requiring mechanical ventilation. There is a high mortality rate associated with the disease. The radiographic pattern of parenchymal involvement depends upon the method by which the organisms reach the lung. Patchy opacities with abscess formation, which mimic staphylococcal pneumonia, are common when the infection reaches the lung via the tracheobronchial tree (Fig. 16.2). Diffuse, bilateral, ill-defined nodular opacities usually reflect hematogenous dissemination. Pleural effusions are common and are usually
small.
Legionella
pneumophila
Legionnaires disease is caused by infection with L pneumophila, a gram-negative bacillus commonly found in air conditioning and humidifier systems. This infection tends to affect older men. Community-acquired infection is seen in patients with COPD or malignancy, while nosocomial infection primarily affects immunocompromised patients or those with renal failure or malignancy. The
characteristic
radiographic
pattern
is
airspace
opacification,
which is initially peripheral and sublobar. In some patients, the airspace opacities appear as a round pneumonia. The infection progresses to lobar or multilobar involvement despite the initiation of antibiotic therapy. At the peak of disease, the parenchymal involvement is usually bilateral. Pleural effusions are seen in approximately 30% of patients. Cavitation is not seen except in the immunocompromised patient (Fig. 16.4). The radiographic resolution of pneumonia is often prolonged and may lag behind symptomatic improvement.
1141
FIGURE 16.4. Legionella Pneumonia in an Immunocompromised Patient. Frontal chest radiograph in a 35-year-old man with AIDS demonstrates a middle lobe airspace opacification with areas of cavitation. Bronchoscopy showed L pneumophila pneumonia.
P.464
Anaerobic
Bacterial
Infection
The majority of anaerobic lung infections arise from aspiration of infected oropharyngeal contents (4). Approximately 25% of patients give a history of impaired consciousness, and many are alcoholic. The most common organisms responsible are the gram-negative bacilli Bacteroides and Fusobacterium, although the majority of pulmonary infections are polymicrobial. All anaerobic pulmonary infections produce a similar radiographic appearance. The distribution of parenchymal opacities reflects the gravitational flow of aspirated material. When aspiration occurs in the supine position, it
1142
is the posterior segments of the upper lobes and superior segments of the lower lobes that are predominantly involved, whereas aspiration in the erect position leads to involvement of basal segments of the lower lobes. The typical radiographic appearance is peripheral lobular and segmental airspace opacities. Cavitation within areas of consolidation is relatively common, and discrete lung abscesses may be seen in up to 50% of patients. Hilar and/or mediastinal lymph node enlargement may be seen in those with lung abscesses. Empyema, with or without bronchopleural fistula formation, is a common complication and is seen in up to 50% of patients.
Atypical
Bacterial
Infections
Actinomycosis Actinomyces israelii is an anaerobic gram-positive filamentous bacterium that is a normal inhabitant of the human oropharynx. It causes disease when it gains access to devitalized or infected tissues that facilitate its growth. Actinomycosis most commonly follows dental extractions, manifesting as mandibular osteomyelitis or a soft tissue abscess. The lungs may be infected by aspiration of infectious oral debris or, less commonly, by direct extension from the primary site of disease. The radiographic pattern of actinomycosis is often indistinguishable from that of nocardiosis. Findings consist of nonsegmental airspace opacities in the periphery of the lower lobes. In some cases, the infection manifests as a localized masslike opacity that mimics bronchogenic carcinoma. If therapy is not instituted, a lung abscess may develop. Thoracic actinomycosis is characterized by its ability to spread to contiguous tissues without regard for normal anatomic barriers. Extension into the pleura will cause empyema, while chest wall involvement is characterized by osteomyelitis of the ribs and chest wall abscess. Involvement of the ribs is seen as wavy periosteal reaction or lytic rib destruction (5). If the pleuropulmonary disease becomes chronic, extensive fibrosis may be seen. Rarely, the
1143
disease is disseminated and a miliary pattern is seen. Mycoplasma display both bacterial and viral characteristics and are considered as a separate group. They are probably the most common atypical pneumonia and account for 10% to 30% of all communityacquired pneumonia. Affected patients usually have a subacute illness of 2 to 3 weeks’ duration. Symptoms include fever, nonproductive cough, headache, and malaise. Unusual physical findings include bullous myringitis and rash. In the early stages of infection, interstitial inflammation leads to a fine reticular pattern on the chest radiograph. This may progress to patchy segmental airspace opacities (Fig. 16.5), which may coalesce to produce lobar consolidation. CT of mycoplasma pneumonia usually appears as patchy airspace opacities with a tree-in-bud appearance that reflects infectious bronchiolitis (Fig. 16.6). The process is often unilateral and tends to involve the lower lobes. Pleural effusion may be seen in the consolidative form of disease and occurs most commonly in children. Lymph node enlargement is uncommon but may be seen in children. Radiographic resolution may require 4 to 6 weeks.
Mycobacterial
Infections
Mycobacterium tuberculosis is an aerobic acid-fast bacillus. Two principal forms of tuberculous pulmonary disease are recognized clinically and radiographically: primary tuberculosis (TB) and “reactivation― or postprimary disease. The inflammatory response to M tuberculosis differs from the normal response to bacterial organisms in that it involves cell-mediated immunity (delayed hypersensitivity). Initially, droplet nuclei laden with bacilli are inhaled and implant in a subpleural location. In most P.465 patients, the bacilli are phagocytized and killed by alveolar macrophages. If the bacilli overcome the immune response of the host, an inflammatory focus is established. The macrophages are then transformed into epithelioid cells, which aggregate to form granulomas. The granulomas are usually well-formed by 1 to 3
1144
weeks, coinciding with the development of delayed hypersensitivity. The granulomas typically demonstrate central caseous necrosis, thereby distinguishing them from the granulomas seen in sarcoidosis. Inflammation and enlargement of draining hilar and mediastinal lymph nodes is common in primary disease, particularly in children and immunocompromised patients.
FIGURE
16.5.
Mycoplasma Pneumonia.
Posteroanterior (A)
and lateral (B) radiographs in a 21-year-old woman demonstrate mixed diffuse interstitial and bibasilar airspace opacities. Immunofluorescent staining of induced sputum samples revealed M
pneumoniae.
1145
FIGURE 16.6. CT of Mycoplasma Pneumonia. Thin-section CT scans through the upper (A) and lower (B) lungs in a patient with mycoplasma pneumonia show patchy ground-glass opacities (arrows) with scattered tree-in-bud opacities (arrowheads) .
In primary infection, the parenchymal disease and adenopathy may completely resolve, or there may be a residual focus of scarring or calcification. In some situations, usually in infants under the age of 1 year, local parenchymal disease progresses and is termed progressive primary TB. More commonly, the disease will be contained by the granulomatous response and recur years later (reactivation or postprimary TB) in the setting of weakened host defenses from aging, alcoholism, diabetes, cancer, or HIV infection. Postprimary TB develops under the influence of hypersensitivity, with caseous necrosis seen histologically. Primary tuberculosis has classically been a disease of childhood, although the incidence of primary disease has P.466 increased because of the HIV epidemic. Most patients with primary TB are asymptomatic and have no radiographic sequelae of infection. In some patients a Ranke complex, consisting of a calcified parenchymal focus (the Ghon lesion) and nodal calcification, is seen. If the patient is symptomatic, a nonspecific focal pneumonitis occurs and is seen as small, ill-defined areas of segmental or lobar
1146
opacification (Fig. 16.7). The parenchymal consolidation may mimic a bacterial pneumonia, but the clinical and radiographic course is much more indolent. Cavitation is relatively uncommon in the immunocompetent patient (6). The pulmonary focus may resolve completely or persist as a Ghon lesion or a Ranke complex. Tuberculomas are discrete nodular opacities that may develop in primary TB but are much more common in postprimary disease. Unilateral pleural effusions are seen in 25% of cases and are usually associated with parenchymal disease. If a tuberculous empyema develops, it may break through the parietal pleura to form an extrapleural collection (empyema necessitatis). Unilateral hilar or mediastinal lymph node enlargement is common, particularly in children, and may be the sole radiographic manifestation of infection. Bilateral hilar or mediastinal lymph node enlargement may be seen, but this is uncommon and is almost invariably asymmetric in distinction to lymph node enlargement in sarcoidosis. During the primary
tuberculous
infection,
there
is
hematogenous
dissemination
of the organism to regions with a high partial pressure of oxygen; these include the lung apices, renal medullae, and bone marrow. These microscopic foci are clinically silent and serve as a source of reactivation disease.
1147
FIGURE
16.7. Primary
Tuberculosis. A posteroanterior chest
radiograph in a 32-year-old homeless man shows airspace disease within the anterior segment of the right upper lobe, with right hilar (solid arrow) and paratracheal (open arrow) lymph node enlargement. Sputum stains and cultures revealed Mycobacterium
tuberculosis.
Postprimary TB patients often present with cough and constitutional symptoms, including chills, night sweats, and weight loss. Reactivation tends to occur in the apical and posterior segments of the upper lobes and the superior segments of the lower lobes. Illdefined patchy and nodular opacities are commonly seen. Cavitation is an important radiographic feature of postprimary infection and usually indicates active and transmissible disease (Fig. 16.8). The cavitary focus may lead to transbronchial spread of organisms and result in a multifocal bronchopneumonia. Erosion of a cavitary focus into a branch of the pulmonary artery can produce an aneurysm (Rasmussen aneurysm) and cause hemoptysis. With appropriate antimicrobial treatment, the disease is usually controlled by a
1148
granulomatous response. Parenchymal healing is associated with fibrosis, bronchiectasis, and volume loss (cicatrizing atelectasis) in the upper lobes. There are several late complications of pulmonary TB. Interstitial fibrosis can cause pulmonary insufficiency and secondary pulmonary arterial hypertension. Hemoptysis may be secondary to bronchiectasis, mycetoma formation in an old tuberculous cavity, or erosion of a calcified peribronchial lymph node (broncholith) into a bronchus. Bronchostenosis is a result of healed endobronchial TB. Miliary TB may complicate either primary or reactivation disease. It results from hematogenous dissemination of tubercle bacilli and produces diffuse bilateral 2- to 3-mm pulmonary nodules (Fig. 16.9) . Miliary disease is associated with a high mortality and requires prompt therapy.
Atypical
Mycobacterial
Infection
There are several nontuberculous mycobacteria that may cause pulmonary disease (7). The most common organisms responsible for pulmonary disease are Mycobacterium avium-intracellulare (MAI) or M kansasii. Disease in nonimmunocompromised patients typically affects patients with underlying COPD. The radiographic features are often indistinguishable from those of reactivation TB, with chronic fibrocavitary opacities involving the upper lobes. While cavitation is common, pleural effusion, lymph node enlargement, and miliary spread are distinctly unusual. A second pattern of disease with MAI has recently been described in middle-aged and elderly women, with small peribronchial nodules and bronchiectasis seen in a middle lobe and lingular distribution (Fig. 16.10). Although the disease caused by nontuberculous mycobacteria tends to be more indolent than that seen with M tuberculosis, it is often difficult to treat effectively.
Viral
Pneumonia
Viruses are a major cause of upper respiratory tract and airways infection, although pneumonia is relatively uncommon. The diagnosis
1149
of viral pneumonia is often one P.467 of exclusion. Chest radiographic features are nonspecific and usually demonstrate a pattern of bronchopneumonia or interstitial opacities (8). Resolution is usually complete, but permanent sequelae may be seen, including bronchiectasis, bronchiolitis obliterans (which may produce a unilateral hyperlucent lung or Swyer-James syndrome), and interstitial fibrosis.
FIGURE
16.8. Postprimary
(Reactivation)
Tuberculosis.
A.
Frontal chest film in a 69-year-old Asian immigrant with a cough and severe wasting reveals hyperinflation with marked fibrotic and cavitary disease in the upper lobes with severe volume loss. B . A CT scan through the lung apices demonstrates consolidative and cavitary changes with air–fluid levels and pleural and parenchymal calcifications. Sputum cultures were positive for Mycobacterium tuberculosis.
Influenza In adults, the most common cause of viral pneumonia is influenza.
1150
Outbreaks of influenza can occur in pandemics, epidemics, or sporadically. In most patients the disease is confined to the upper respiratory tract, but in elderly persons, those with underlying cardiopulmonary disease or immunocompromise, and pregnant women, a severe hemorrhagic pneumonia may develop. In adults with influenzal pneumonia, there is often bilateral lower lobe patchy airspace opacification. In children, a diffuse interstitial reticulonodular pattern is more commonly seen. Bacterial superinfection with Streptococcus or Staphylococcus organisms contributes to a fulminating course that may result in death. The development of lobar consolidation, pleural effusion, or cavitation suggests bacterial superinfection. Respiratory syncytial virus and parainfluenza virus are common causes of epidemic viral pneumonia in children. When seen in adults, the disease is usually in the setting of a debilitated or immunocompromised patient (Fig. 16.11). Findings are similar to other viral pneumonias: patchy airspace opacities, bronchial wall thickening,
and
tree-in-bud
opacities.
Varicella-zoster, which causes chickenpox and shingles, may produce a severe pneumonia in adults. Patients on immunosuppressive therapy or with lymphoma are at greatest risk. Chest radiographs characteristically show diffuse bilateral ill-defined nodular opacities 5 to 10 mm in diameter. These opacities usually resolve completely, although in some patients they involute and calcify to produce innumerable small (2 to 3 mm) calcified nodules (Fig. 16.12) . Adenovirus is a frequent cause of upper and, occasionally, lower respiratory tract infection. Overinflation and bronchopneumonia accompanied by lobar atelectasis are the most frequent radiographic manifestations of adenovirus pneumonia; however, adenovirus in children may present as lobar or segmental consolidation. SARS-Associated
Coronavirus
(SARS-CoV). Severe Acute
Respiratory Syndrome (SARS) is a recently described respiratory illness caused by a new coronavirus not previously seen in humans, SARS-CoV. The disease appears to have originated in southern China
1151
and rapidly spread to other areas of the world, causing over 8,000 reported cases in late 2002 and early 2003. The clinical symptoms and signs as well as the radiographic manifestations of SARS are nonspecific. Unilateral or bilateral areas of airspace opacity are seen on initial radiographs in the majority of affected patients. The opacities are typically peripheral and lower zone in location, progressively involving the central lungs. Occasionally, initial radiographs are negative; CT demonstrates areas of ground-glass opacity P.468 and/or not
consolidation
(9). Lymphadenopathy and pleural effusions are
characteristic.
1152
FIGURE 16.9. Miliary Tuberculosis. A cone-down view of a frontal radiograph demonstrates innumerable micronodular opacities characteristic of micronodular (miliary) interstitial disease. Transbronchial biopsy demonstrated caseating granulomas containing acid-fast bacilli.
Fungal
Pneumonia
Fungal infections are now seen with increased frequency because of an increase in the incidence of disease caused by pathogenic fungi in healthy hosts and the emergence of opportunistic species in immunocompromised hosts. Fungi can cause pulmonary disease by several mechanisms. Some fungi, including Histoplasma capsulatum, Coccidioides immitis, and Blastomyces dermatitidis, are primary pathogens and most commonly infect normal hosts (1 0). Other fungi, most notably Aspergillus, Candida, and Cryptococcus, are opportunistic pathogens in immunocompromised individuals. In all cases, the fungi elicit a necrotizing granulomatous reaction. The high mortality of untreated invasive infection and the availability of effective antifungal therapy with intravenous amphotericin B and the oral azoles (e.g., fluconazole, itraconazole) has made the early and accurate diagnosis of fungal infection imperative. A number of serologic assays (complement fixation, immunodiffusion) and histologic methods are available for the accurate diagnosis of fungal infection.
Histoplasmosis H capsulatum is endemic to certain areas of North America, most notably the Ohio, Mississippi, and St. Lawrence River valleys and Mexico. The overwhelming majority (95% to 99%) of infections by H capsulatum are asymptomatic. A routine chest film demonstrating multiple well-defined calcified nodules less than 1 cm in size, with or without calcified hilar or mediastinal lymph nodes, may be the only indication of prior infection.
1153
FIGURE
16.10.
Mycobacterium
avium-intracellulare (MAI)
Infection. Thin-section CT scans through mid-lungs (A) and lower (B) lungs in a patient with MAI pulmonary infection show bronchiectasis (arrowheads), scattered nodules, and tree-in-bud (arrows) .
FIGURE 16.11. Parainfluenza Virus Pneumonia. CT through upper lungs (A) and mid-lungs (B) in a patient with acute myelogenous leukemia (AML) shows striking bronchopneumonia and bronchiolitis (arrowheads). Parainfluenza virus was isolated
1154
from
bronchoalveolar
lavage
fluid.
P.469 Acute histoplasma infection most often presents with the abrupt onset of flulike symptoms. The chest radiograph in such patients may be normal or may show nonspecific changes, including subsegmental airspace opacities with or without associated hilar lymph enlargement. If the patient inhales a large inoculum of organisms, widespread, fairly discrete nodular opacities 3 to 4 mm in diameter are seen with hilar adenopathy. Alternatively, acute histoplasmosis may result in a solitary, sharply defined nodular opacity, 1 month) and may mimic bronchogenic carcinoma. A third pattern of disease is diffuse reticulonodular opacities. Pleural effusion and lymph node enlargement are uncommon. A disseminated miliary form may be seen in immunocompromised hosts. Aspergillus species are responsible for a spectrum of pulmonary diseases in humans. These include aspergilloma or mycetoma formation
within
preexisting
cavities,
semi-invasive
(chronic
necrotizing) aspergillosis in patients with mildly impaired immunity, invasive pulmonary aspergillosis in the neutropenic lymphoma or leukemia patient, and allergic bronchopulmonary aspergillosis in the hyperimmune
patient.
An aspergilloma (mycetoma, fungus ball) is a ball of hyphae, mucus, and cellular debris that colonizes a preexisting bulla or a parenchymal cavity created by some other pathogen or destructive process. Invasion into adjacent lung parenchyma does not occur unless host defense mechanisms are compromised. The mycetoma is usually asymptomatic, but may cause hemoptysis, which may be massive (>350 mL/24 hours). An aspergilloma is seen as a solid round mass within an upper lobe cavity, with an “air crescent― separating the mycetoma from the cavity wall (Fig. 16.15). The mycetoma is usually free within the cavity and can be seen to roll dependently on decubitus radiographs or CT. Progressive apical pleural thickening adjacent to a cavity is a common radiographic
1158
finding and should prompt a search for a complicating mycetoma. Semi-invasive and invasive aspergillosis are discussed later in this chapter, while allergic bronchopulmonary aspergillosis is reviewed in Chapter 18.
FIGURE 16.14. Disseminated (Miliary) Coccidioidomycosis. A . Frontal radiograph in a 42-year-old patient with AIDS shows miliary nodulation with enlarged hilar and mediastinal nodes. B . CT confirms the presence of diffuse nodular and reticular opacities. Transbronchial biopsy showed Coccidioides infection.
P.471
Parasitic
Infection
Parasitic infections of the lung are relatively uncommon in the United States. However, increases in travel to countries where parasites are endemic, the immigration of people from these regions to the United States, and growing numbers of immunocompromised patients require a familiarity with these infections. In general, parasitic diseases of the thorax are manifested by either a direct invasion of lungs and pleura or, less commonly, a hypersensitivity reaction (1 1) .
1159
FIGURE
16.15. Aspergilloma.
A.
Posteroanterior
chest
radiograph in a 32-year-old woman with prior tuberculosis and recent hemoptysis shows a mass (M) in the right upper lung capped by a crescent of air with adjacent apical pleural thickening (asterisk). B . CT scan shows a mass representing a mycetoma within a preexisting cavity. Sputum showed heavy Aspergillus.
Amebiasis Symptomatic infection with Entamoeba histolytica is usually confined to the GI tract and liver. If the infection remains confined to the subphrenic space, a right pleural effusion and basilar atelectasis may result from local diaphragmatic inflammation. The most common method of pleuropulmonary involvement by amebiasis is by the direct intrathoracic
extension
of
infection P.472
from a hepatic abscess. This transdiaphragmatic spread of organisms may extend into the right pleural space to produce an empyema or may involve the right lower lobe to produce an amebic pneumonia or lung abscess.
1160
Hydatid
Disease
(Echinococcosis)
of
the
Lung Echinococcus granulosus is the cause of most cases of human hydatid disease. The disease is endemic in sheep-raising areas and is relatively uncommon in the United States. Dogs are the usual definitive hosts, with sheep acting as intermediate hosts. When a human becomes an accidental intermediate host, disease may result. The larval organisms travel to the liver and lungs and, if they survive host defenses, encyst and gradually enlarge. Pulmonary echinococcal cysts are composed of three layers: an exocyst (chitinous layer), which is a protective membrane; an inner endocyst, which produces the “daughter cysts―; and a surrounding capsule of compressed, fibrotic lung known as the pericyst. Pulmonary echinococcal cysts characteristically present as wellcircumscribed, spherical soft tissue masses. In distinction to hepatic cysts, lung cysts do not have calcified walls. The cysts range in size from 1 to 20 cm, with a predilection for the lower lobes and the right side. While most cysts remain asymptomatic, patients may present when the cyst develops a communication with the bronchial tree. If the pericyst ruptures, a thin crescent of air will be seen around the periphery of the cyst, producing the “meniscus― or “crescent― sign. If the cyst itself ruptures, the contents of the cyst are expelled into the airways, producing an air–fluid level. On occasion, the cyst wall may be seen crumpled and floating within an uncollapsed pericyst, producing the pathognomonic “sign of the camalote― or “water lily― sign. Rarely, a cyst will rupture into the pleural space, producing a large pleural effusion. Paragonimiasis results from infection with the lung fluke Paragonimus westermani. The organism is found predominantly in eastern Asia and is usually acquired by eating raw crabs or snails. Infestation of the lung may be asymptomatic, or a patient may present with cough, hemoptysis, dyspnea, and fever. In 20% of affected patients, the chest radiograph is normal. The most common radiographic finding is multiple cysts with variable wall thickness. These cystic opacities may become confluent and are often
1161
associated with focal atelectasis and subsegmental consolidation. Dense linear opacities representing the burrows of the organisms may be identified. Because the flukes penetrate the pleura, effusions are common and may be massive.
Schistosomiasis Human schistosomiasis is caused by three blood flukes: Schistosoma mansoni, S japonicum, and S haematobium. It is one of the most important parasitic infestations of humans worldwide, although it is rarely acquired in the United States. The life cycle of the fluke is complex, with human infestation acquired through contact with infested water. The larvae penetrate the skin or oropharyngeal mucosa and travel via the venous circulation to the pulmonary capillaries. As the larvae pass through the lungs, an allergic response may develop, presenting radiographically as transient airspace opacities (eosinophilic pneumonia) that resolve spontaneously. The larvae then pass through the pulmonary capillaries into the systemic circulation. S japonicum and S mansoni eventually migrate to the mesenteric venules, while S haematobium migrates to bladder venules. The mature flukes produce ova, which may embolize to the lungs, where they implant in and around small pulmonary arterioles. The organism induces granulomatous inflammation and fibrosis, which leads to an obliterative arteriolitis, resulting in pulmonary hypertension and cor pulmonale. Radiographically, a diffuse fine reticular pattern is most commonly seen in association with dilatation of the central pulmonary arteries. Small nodular opacities resembling miliary TB may be seen as granulomata forming around ova.
INFECTION IN THE IMMUNOCOMPROMISED AIDS
HOST
AND
IN
Immunocompromise is defined as a decrease in the normal host defense mechanisms that fight infection. Immunocompromised patients include those with HIV infection, underlying hematologic
1162
malignancy, and individuals receiving chemotherapeutic and immunosuppressive therapy. The types of pulmonary infection seen in the immunocompromised patient depend on the specific defect(s) in host defense mechanisms. While the majority of pulmonary complications in immunocompromised patients are infectious in nature, noninfectious complications of disease can account for up to 25% of lung disease in this population. The accurate identification of the predominant radiographic pattern of abnormality in the immunocompromised patient helps limit the differential diagnostic considerations (Tables 16.1, 16.2) (1 2). With the advent of highly active antiretroviral therapy (HAART) and effective prophylaxis, the incidence of opportunistic infection in HIV/AIDS has decreased dramatically. Bacterial respiratory infections now account for most pulmonary infections in individuals living with HIV in the developed world (1 3,1 4) .
Bacterial
Pneumonia
Bacteria are the most common cause of pneumonia in immunocompromised hosts. In HIV-infected patients, bacterial pneumonia may occur early in the course of infection and has an incidence six times that seen in the normal population. The occurrence of two or more episodes of bacterial pneumonia within 1 year is categorized as an AIDS-defining illness for patients with HIV infection. The most common organisms causing pneumonia in HIV patients are S pneumoniae, P.473 H influenzae, S aureus, E coli, and P aeruginosa. Uncommon causes of bacterial pneumonia in the AIDS population include Nocardia asteroides, Rhodococcus equi, Bartonella henselae, and B quintana (bacillary angiomatosis). In the non-HIV immunocompromised patient, S aureus and gram-negative aerobes including Klebsiella, Proteus, E coli, Pseudomonas, Enterobacter, and Serratia are the most common bacterial pathogens. Bacterial pneumonia is characterized by focal segmental or lobar airspace opacities. Cavitation is more frequent in the immunocompromised population than in normal individuals and may occur as multiple microabscesses.
1163
Multilobar involvement and diffuse pneumonia may occur and are distinctly unusual in normal individuals. Pleural effusions and empyema are uncommon (1 3) .
TABLE 16.1 Radiographic Patterns of Abnormality in NonHIV
Immunocompromised
Pattern Lobar/segmental consolidation
Patients
Potential Gram-negative Gram-positive
Etiology bacteria bacteria
Legionella
Nodules ± cavitation
Fungi Aspergillus species Coccidioides Cryptococcus
immitis neoformans
Mucor species Nocardia asteroides Legionella micdadei Neoplasm Other
Diffuse lung disease
Pneumocystis jiroveci Viral pneumonia Fungi Toxoplasma gondii Strongyloides Drug reaction Hemorrhage
stercoralis
Radiation pneumonitis Nonspecific interstitial pneumonia (NSIP) Lymphangitic carcinomatosis
1164
Modified from McLoud and Naidich (1 2); material used with permission.
Renal transplant recipients and patients on high-dose corticosteroids are at increased risk of pneumonia caused by Legionella pneumophila and L micdadei (Pittsburgh agent). L pneumophila causes multilobar focal areas of consolidation, sometimes with cavitation and pleural effusion. The Pittsburgh agent causes a characteristic appearance of multiple, well-circumscribed, centrally cavitating nodules. Nocardia is a gram-positive, branching, filamentous bacillus that is weakly acid fast. N asteroides is the most important cause of pulmonary disease. It is usually an opportunistic infection in patients on immunosuppressive therapy, those with lymphoma or leukemia, and patients with alveolar proteinosis. The most frequent radiographic presentation is a homogeneous, nonsegmental airspace opacity or a mass. Cavitation is frequent (Fig. 16.16). Infection may extend into the pleural space and chest wall to produce empyema and osteomyelitis, respectively. Hilar lymph nodes may be enlarged. Treatment is with sulfur antibiotics.
TABLE 16.2 Radiographic Patterns of Abnormality in AIDS Patients
Pattern
Potential
Etiology
Normal
Pneumocystis jiroveci pneumonia (PCP) Tuberculosis or fungal infection Nonspecific interstitial pneumonia (NSIP)
Focal lung disease
Bacterial pneumonia PCP Mycobacterial/fungal infection Non-Hodgkin lymphoma
1165
Diffuse lung disease
PCP PCP + other infection (cytomegalovirus, Mycobacterium avium-intracellulare, miliary tuberculosis, fungus) Mycobacterium tuberculosis Fungal infection NSIP Lymphocytic interstitial Kaposi sarcoma
Nodules
Non-Hodgkin
Mycobacterial
(LIP)
lymphoma
Kaposi sarcoma Septic emboli Mycobacterial/fungal
Adenopathy
pneumonia
or
infection
fungal
infection
Kaposi sarcoma Non-Hodgkin lymphoma PCP
(uncommon)
Pleural
Kaposi
sarcoma
effusion
Mycobacterial/fungal
infection
Non-Hodgkin lymphoma Pyogenic empyema PCP (uncommon)
Modified from McLoud and Naidich (1 2); material used with permission.
Tuberculosis The incidence of TB has increased considerably since the onset of the AIDS epidemic. Most cases are caused by reactivation of previously
1166
acquired disease. The diagnosis of TB in immunocompromised hosts is complicated because skin reactivity and sputum analysis are less sensitive in immunocompromised hosts and the yield of bronchoalveolar lavage is decreased in this patient population. The chest radiographic findings depend on the stage of HIV infection and the degree of immune dysfunction, which can be estimated by the CD4 count. In P.474 the early stages of AIDS (CD4 >200 cells/mm3 ), a postprimary pattern of upper lobe fibrocavitary disease indistinguishable from that seen in the immunocompetent patient is most common. Later in the course of AIDS (CD4 50 to 200 cells/mm3 ), the radiographic features most often associated with primary disease are seen and include lobar consolidation, mediastinal and hilar lymphadenopathy, and pleural effusion (6). Rim-enhancing nodes with central necrosis on CT scans are a characteristic finding and should strongly suggest TB in a patient with AIDS. In advanced AIDS (CD4 1 mm) and more numerous in patients with diseases primarily affecting the interlobular interstitium, such as interstitial pulmonary edema, idiopathic pulmonary fibrosis, and lymphangitic carcinomatosis (Fig. 17.4). Interlobular lines on HRCT are the equivalent of Kerley B lines seen in the inferolateral portions of the lungs on frontal radiographs. Within the central regions of the lung, long (2 to 6 cm) linear opacities representing obliquely oriented connective tissue septa can be seen, which are the equivalent of radiographic Kerley A lines.
TABLE 17.3 Differential Diagnostic HRCT Features in Interstitial Lung Disease
HRCT Finding Interlobular lines
(septal)
Differential Interstitial
Diagnosis
edema
Lymphangitic
carcinomatosis
Sarcoidosis
Idiopathic pulmonary fibrosis (and other forms of usual
1188
(IPF)
interstitial
Intralobular
lines
pneumonia
[UIP])
IPF (UIP)
Asbestosis
Alveolar
proteinosis
Hypersensitivity (chronic)
“Thickened― fissures
Pulmonary
pneumonitis
edema
Sarcoidosis
Lymphangitic
Peribronchovascular interstitial
Pulmonary
carcinomatosis
edema
(smooth)
thickening
Sarcoidosis
(nodular)
Lymphangitic carcinomatosis (smooth or nodular)
Centrilobular
nodules
Hypersensitivity
pneumonitis
Bronchiolitis obliterans organizing pneumonia (BOOP)/cryptogenic pneumonia (COP)
1189
with
organizing
Respiratory bronchiolitis–associated interstitial
Subpleural
lines
lung
disease
(RB-ILD)
Asbestosis
IPF (UIP)
Parenchymal
bands
Asbestosis
IPF (UIP)
Sarcoidosis
Honeycombing
IPF (UIP)
Asbestosis
Hypersensitivity
pneumonitis
(chronic)
Sarcoidosis
Thin-walled
cysts
Eosinophilic
granuloma
(EG)
Lymphangioleiomyomatosis
Tuberous
sclerosis
Neurofibromatosis (emphysema)
1190
(pneumatocele)
Irregular lung interfaces
Pulmonary
edema
IPF (UIP)
Sarcoidosis
Micronodules,
random
distribution
Miliary
tuberculosis
or
histoplasmosis
Hematogenous
metastases
Silicosis/coal worker's pneumoconiosis (CWP)
EG
Micronodules, perilymphatic distribution
Sarcoidosis
Lymphangitic
carcinomatosis
Silicosis/CWP
Ground-glass
opacities
UIP
Desquamative pneumonia
Acute
interstitial
1191
interstitial
pneumonia
(AIP)
Hypersensitivity
pneumonitis
BOOP/COP
RB-ILD
Hemorrhage
Pneumocystis
jiroveci
Cytomegalovirus
Alveolar
IPF/UIP
Traction bronchiectasis
Sarcoidosis
Silicosis/CWP
Conglomerate
mass
Sarcoidosis
Silicosis
CWP
Radiation
Consolidation
BOOP/COP
1192
pneumonia
proteinosis
Architectural distortion
fibrosis
pneumonia
Sarcoidosis
AIP
UIP
Intralobular
Lines
In some patients, a lattice of fine lines is seen within the central portion of the pulmonary lobule radiating out toward the thickened lobular borders to produce a “spoke-and-wheel― or “spiderweb― appearance. These lines are not normally visible on HRCT and represent thickening of the intralobular or parenchymal interstitium. Intralobular lines usually represent fibrosis and are most commonly seen in idiopathic pulmonary fibrosis (IPF) and other forms of usual interstitial pneumonia (UIP). However, intralobular lines can also be seen in other infiltrative diseases such as pulmonary
alveolar
proteinosis
“Thickened―
(PAP).
Fissures
The apparent thickening of interlobar fissures in patients with interstitial lung disease is usually a direct extension of the thickening of interlobular septa to involve the subpleural interstitium of the lung. While such a process normally involves all pleural surfaces equally, the “thickening― is usually best appreciated on the fissures, where two layers of visceral pleura—and therefore two layers of subpleural P.483 interstitium—are seen outlined on either side by aerated lung. The fissural thickening can be smooth or nodular. Smooth fissural thickening is virtually indistinguishable from a small amount of pleural fluid within the fissure and is most commonly seen with pulmonary edema. Nodular fissural thickening is commonly seen in sarcoidosis
and
lymphangitic
carcinomatosis
1193
(Fig. 17.4), where the
nodules lie within the subpleural lymphatics.
FIGURE 17.4. Interlobular Septal Lines in Lymphangitic Carcinomatosis. An HRCT scan through the upper lobes in a patient with lymphangitic carcinomatosis shows thickened interlobular septa (small arrow). Note the presence of nodular fissural thickening (large arrows), another common finding in this entity.
Thickened bronchovascular structures of the lung result from thickening of the peribronchovascular interstitium. This produces apparent enlargement of perihilar vascular structures and thickening of bronchial walls, which is the HRCT equivalent of peribronchial cuffing and tram tracking seen radiographically. While pulmonary edema causes smooth thickening of the peribronchovascular interstitium, nodular or irregular thickening can be seen in sarcoidosis or UIP (Fig. 17.5). Lymphangitic carcinomatosis can result in either smooth or irregular peribronchovascular thickening, although the former is more common (Fig. 17.6) .
Centrilobular
(Lobular
Core)
1194
Abnormalities
Thickening of the axial interstitium within the lobular core produces an abnormal prominence of the “dotlike― or branching centrilobular arteriole. Diseases that commonly produce this appearance include pulmonary edema, lymphangitic carcinomatosis, and UIP. The centrilobular bronchiole is not normally seen on HRCT but may be rendered visible as a result of luminal dilatation or thickening of the centrilobular interstitium. Small airways disease can produce centrilobular bronchiolar abnormalities, which are seen on HRCT as fluid-filled dilated branching Y-shaped structures that produce a “tree-in-bud― appearance. Ill-defined centrilobular nodules represent disease of the bronchiole and adjacent parenchyma and can be seen in subacute hypersensitivity pneumonitis (Fig. 17.7), cryptogenic organizing pneumonia (COP), and other disorders.
FIGURE 17.5. Intralobular Lines in Idiopathic Pulmonary Fibrosis (IPF). A targeted HRCT through the right lower lobe in a patient with IPF shows thickening of intralobular (long arrows) and interlobular (arrowheads) lines associated with ground-glass opacity.
1195
Subpleural
Lines
These 5- to 10-cm-long curvilinear opacities are found within 1 cm of the pleura and parallel the chest wall. They are most frequent in the posterior portions of the lower lobes and remain unchanged on prone scans. This finding, which probably represents an early phase of lung fibrosis, should be distinguished from a similar line that is seen as a result of atelectasis in the P.484 dependent portion of the lungs in normal individuals. Subpleural lines are most often seen in patients with asbestosis and, less commonly, IPF.
FIGURE 17.6. Thickened Bronchovascular Structures in Lymphangitic Carcinomatosis. In a patient with lymphangitic carcinomatosis, an HRCT shows both smooth and nodular thickening of the bronchovascular structures (arrows) that represents lymphatic tumor surrounding the axial interstitium.
1196
FIGURE
17.7. Centrilobular
Ground-Glass
Nodules
in
Subacute Hypersensitivity Pneumonitis. HRCT shows the typical poorly defined centrilobular nodules (arrows) of subacute hypersensitivity
Parenchymal
pneumonitis
(“bird-fancier's
lung―).
bands are nontapering linear opacities, 2 to 5 cm in
length, that extend from the lung to contact the pleural surface. These fibrotic bands can be distinguished from vessels and thickened septa by their length, thickness, course, absence of branching, and their association with regional parenchymal distortion. Parenchymal bands are usually seen in asbestosis, IPF, and sarcoidosis. Honeycombing, seen as small (6 to 10 mm) cystic spaces with thick (1 to 3 mm) walls, most often in the posterior subpleural regions of the lower lobes, represents end-stage pulmonary fibrosis of various etiologies. Pathologically, the cysts are lined by bronchiolar epithelium and are the result of bronchiolectasis. Most patients show additional signs of interstitial disease, including thickened interlobular and intralobular lines, parenchymal bands, irregularity of lung interfaces, and areas of ground-glass opacification.
1197
Honeycombing is frequently seen in IPF (and other forms of UIP) (Fig. 17.8), chronic hypersensitivity pneumonitis, and occasionally sarcoidosis. Thin-walled cysts are a common manifestation of late stages of Langerhans cell histiocytosis of lung (LCH) and lymphangioleiomyomatosis (LAM). These cysts are slightly larger in diameter (10 mm) than honeycomb cysts, are more uniform in size, and have thinner walls. Honeycomb cysts usually have shared walls, while the cysts of LCH and LAM do not. The cysts of LCH and LAM are usually evenly distributed from central to peripheral portions of the upper lobes (Fig. 17.9), with or without lower lobe involvement, while honeycombing tends to occur in the subpleural regions of the lower lobes. While normal lung may be found in the intervening spaces between the cysts of LCH and LAM, honeycombing uniformly destroys lung and produces distortion of lung interfaces and traction bronchiectasis, features not found in eosinophilic granulomatosis (EG) and LAM.
FIGURE 17.8. Honeycomb Lung in Idiopathic Pulmonary Fibrosis. HRCT in a patient with IPF shows peripheral honeycombing (arrows) indicative of end-stage pulmonary
1198
fibrosis.
Irregularity
of
Lung
Interfaces
A common HRCT sign of interstitial disease, irregularity of the normally smooth interface between the bronchovascular bundles and the surrounding lung reflects edema or fibrosis of the axial interstitium or infiltration by granulomas (Fig. 17.6) or tumor. Similarly, irregularity of the interface between fissures or pleural surfaces and adjacent lung indicates P.485 peripheral interstitial disease. Pulmonary edema, IPF, and sarcoidosis are the most common causes of irregular lung interfaces.
FIGURE 17.9. Thin-Walled Cysts in Lymphangioleiomyomatosis (LAM). An HRCT of a patient with LAM shows multiple, variably sized, round, thin-walled cysts.
1199
FIGURE 17.10. Nodules and a Conglomerate Mass in Silicosis. A. Posteroanterior radiograph of a 79-year-old patient with silicosis shows diffuse nodules as well as a conglomerate mass in the right upper lobe (arrow). B . HRCT scan through the upper lobes shows peribronchovascular and subpleural micronodules
(small
arrows), larger nodules (curved
arrow), and
a conglomerate mass representing progressive massive fibrosis in the right upper lobe (large arrow). The pleural effusions are caused by concomitant congestive heart failure.
Micronodules These 1- to 3-mm, sharply marginated, round opacities seen on HRCT represent conglomerates of granulomas or tumor cells within the interstitium. These are most often seen in sarcoidosis, EG, silicosis (Fig. 17.10), miliary tuberculosis (TB) or histoplasmosis, metastatic adenocarcinoma, and lymphangitic carcinomatosis. They may be seen along the central bronchovascular structures (sarcoidosis, EG); within interlobular septa or subpleural interstitium (sarcoidosis, lymphangitic carcinomatosis, silicosis); or within the substance of the pulmonary lobules (metastatic adenocarcinoma, miliary granulomatous infection). Nodules predominating in the
1200
peribronchovascular, interlobular, and subpleural regions—those portions of the interstitium where the lymphatics lie—are said to have a “perilymphatic― distribution. Because it may be difficult to distinguish vertically oriented small upper and lower lobe vessels from interstitial nodules on HRCT, contiguous, thick (10 mm) scans are often helpful.
Ground-Glass
or
Hazy
Increased
Density
Multifocal areas of increased density can sometimes be identified in patients with diffuse interstitial lung disease. These regions, which often respect lobular borders, are distinguished from typical airspace opacification by their granular appearance with maintained visibility of pulmonary vessels and the absence of air bronchograms. These opacities are most often produced by thickening of the alveolar septa, with or without lining of the alveolar spaces by inflammatory exudate or fluid. Diseases commonly associated with this appearance include desquamative interstitial pneumonia (DIP), Pneumocystis jiroveci (formerly P carinii) pneumonia, acute hypersensitivity pneumonitis (Fig. 17.11), nonspecific interstitial pneumonia (NSIP), and interstitial pulmonary edema. The ground-glass densities are occasionally confined to the immediate centrilobular regions of the pulmonary lobules, where they appear as fuzzy nodular densities that outline the normally invisible centrilobular bronchiole (Fig. 17.7) . This P.486 reflects involvement of the peribronchovascular interstitium and surrounding alveoli by an inflammatory process and is seen in hypersensitivity pneumonitis, COP, and panbronchiolitis. The presence of ground-glass opacities is important because it often implies an active inflammatory process or edema that is reversible and warrants aggressive treatment. However, ground-glass abnormality associated with a predominant pattern of honeycombing can represent microscopic pulmonary fibrosis.
1201
FIGURE
17.11. Ground-Glass Opacity in Acute
Hypersensitivity
Pneumonitis. An HRCT through the upper
lobes shows confluent ground-glass opacity in a patient with hypersensitivity pneumonitis. Note that the pulmonary vessels are still visible within the areas of abnormality.
FIGURE 17.12. Architectural Distortion and Traction Bronchiectasis in Idiopathic Pulmonary Fibrosis. HRCT through the lower lobes shows peripheral honeycombing, traction bronchiectasis (arrow), and resultant architectural distortion.
1202
Architectural
Distortion
and
Traction
Bronchiectasis Processes that result in extensive parenchymal fibrosis can distort the normal architecture of the lung, creating irregularities of the lung–mediastinal, lung–pleural, and lung–vascular interfaces. Parenchymal distortion is often better appreciated on HRCT than on plain radiographs. Sarcoidosis and UIP (Fig. 17.12) are the diseases most commonly associated with architectural distortion.
FIGURE 17.13. Consolidation in Cryptogenic Organizing Pneumonia (COP). A. Posteroanterior radiograph in a 53-yearold patient with fever, dyspnea, and a dry cough shows patchy consolidation and diminished lung volumes. B . HRCT scan shows multifocal areas of consolidation in a peribronchial distribution. Note air bronchograms with mild bronchial dilatation within the consolidated areas. An open lung biopsy showed COP.
A finding commonly associated with architectural distortion is traction bronchiectasis, in which fibrosis causes traction on the walls of
1203
bronchi, resulting in irregular dilatation. While this usually involves segmental and subsegmental bronchi, it also can be seen at the intralobular level, where traction bronchiolectasis contributes to honeycombing. Traction bronchiectasis is most commonly seen in IPF (Fig. 17.12) and other forms of UIP but is also common in longstanding sarcoidosis.
Conglomerate
Masses
In some patients with extensive pulmonary fibrosis, masses of fibrotic tissue develop in the parahilar regions of the upper lobes, often associated with peripheral bullae. On CT and HRCT, these masses are seen to contain crowded vessels and dilated bronchi. These conglomerate masses are most often seen in patients with end-stage sarcoidosis but can occur in complicated silicosis with progressive massive fibrosis (PMF) (Fig. 17.10) or radiation fibrosis following treatment of Hodgkin lymphoma or lung cancer. A similar finding is seen rarely in intravenous drug users when a granulomatous fibrosis results as a response to intravenous talc or starch mixed with narcotics. Consolidation refers to increased lung density that obscures underlying blood vessels; air bronchograms are commonly present. This finding can be seen with any P.487 airspace-filling process (Fig. 17.13) but occasionally occurs in interstitial diseases such as UIP and sarcoidosis.
TABLE 17.4 Differential Diagnostic Features in Chronic Interstitial
Finding Upper zone distribution
Lung
Disease
Differential Tuberculosis
1204
Diagnosis
(postprimary)
Chronic
fungal
infection
Histoplasmosis
Coccidioidomycosis
Sarcoidosis
Eosinophilic
granuloma
Silicosis
Ankylosing
spondylitis
Hypersensitivity (chronic)
pneumonitis
Radiation fibrosis from treatment of head and neck malignancy
Lower
zone
Idiopathic
pulmonary
distribution
Asbestosis
Rheumatoid
lung
Scleroderma
Neurofibromatosis
1205
fibrosis
Dermatomyositis/polymyositis
Chronic
Normal or increased lung volumes
aspiration
Sarcoidosis
Eosinophilic
granuloma
Lymphangioleiomyomatosis
Tuberous
sclerosis
Interstitial
disease
superimposed
emphysema
Honeycombing
Idiopathic
pulmonary
fibrosis
Sarcoidosis
Eosinophilic
granuloma
Rheumatoid
lung
Scleroderma
Pneumoconiosis
Hypersensitivity
Chronic
1206
pneumonitis
aspiration
on
Radiation
Miliary
nodules
fibrosis
Tuberculosis
Fungi
Histoplasmosis
Coccidioidomycosis
Cryptococcosis
Silicosis
Metastases
Thyroid
Renal
carcinoma
cell
carcinoma
Bronchogenic
carcinoma
Melanoma
Choriocarcinoma
Sarcoidosis
Eosinophilic
Hilar/mediastinal
Sarcoidosis
1207
granuloma
lymph node enlargement
Lymphangitic
carcinomatosis
Lymphoma
Hematogenous
metastases
Tuberculosis
Fungal
infection
Silicosis
Pleural
disease
Asbestosis
(plaques)
Lymphangitic (effusion)
carcinomatosis
Rheumatoid lung disease (effusion/thickening)
Lymphangioleiomyomatosis effusion)
Abnormalities of soft
Skin
nodules
tissues and bony thorax
Neurofibromatosis
1208
(chylous
Subcutaneous
calcifications
Dermatomyositis
Scleroderma
Erosion of distal clavicles
Rheumatoid
lung
Scleroderma
Rib
lesions
Ribbon ribs/erosion of inferior rib margins
Neurofibromatosis
Erosion of superior margins
Rheumatoid
lung
Scleroderma
Kyphoscoliosis
Neurofibromatosis
Lytic bone lesions
1209
Metastases
Eosinophilic
CHRONIC
INTERSTITIAL
granuloma
LUNG
DISEASE
Chronic interstitial lung disease usually results from diffuse inflammatory processes that primarily affect the axial and parenchymal interstitium of the lung. A wide variety of disease processes can result in diffuse damage to the pulmonary interstitium (2). Careful evaluation of all available radiologic studies and correlation with clinical findings and laboratory data are essential to the accurate diagnosis of chronic interstitial lung disease (Table 17.4). However, the majority of patients with interstitial lung disease will require histologic examination of lung tissue for definitive diagnosis.
Chronic
Interstitial
Pulmonary
Edema
Chronic elevation of pulmonary venous pressure may lead to increased interstitial markings on plain radiographs. The interstitial thickening is caused by distention of pulmonary lymphatics and chronic
interstitial
edema,
which P.488
may lead to fibrosis. This is seen most commonly in patients with long-standing mitral stenosis or LV failure. Radiographically, peribronchial cuffing, tram tracking, poor definition of vascular markings, and linear or reticular opacities may be seen. Redistribution of blood flow to the upper lobes, a manifestation of pulmonary venous hypertension, and prominence of the fissures caused by subpleural edema and fibrosis are concomitant findings. Honeycombing is not a feature of chronic pulmonary venous hypertension; its presence in a patient with cardiac disease should suggest another cause of pulmonary fibrosis (e.g., amiodarone lung toxicity).
1210
Connective
Tissue
Disease
These disorders are associated with immunologically mediated inflammation and damage to connective tissues throughout the body. The most common thoracic manifestations of this group of heterogeneous disorders are vasculitis and interstitial fibrosis, although the pleura, chest wall, diaphragm, and heart may also be affected (3) . Rheumatoid
Lung
Disease (Table 17.5).
Rheumatoid
arthritis
produces a chronic arthritis of peripheral joints. Extra-articular manifestations are seen in up to 75% of patients. In contrast to the disease as a whole, which is more common in women, pulmonary involvement is more common in men. The pleuropulmonary manifestations of rheumatoid disease typically follow the onset of joint disease and tend to be seen in patients with high serum rheumatoid factor titers and eosinophilia. However, in up to 15% of patients, pleuropulmonary involvement precedes the joint disease. The most common radiographic manifestation of parenchymal lung involvement is an interstitial pneumonitis and fibrosis, which histologically is a form of UIP. This begins as an alveolitis (inflammation of the alveolar interstitium) that is seen radiographically as fine reticular or ground-glass opacities with a lower zone predominance. There is gradual progression to end-stage pulmonary fibrosis with the development of a bibasilar medium or coarse reticular or reticulonodular pattern (honeycombing) (Fig. 17.14). HRCT is more sensitive in detecting the earliest parenchymal changes than conventional radiographs and is also more sensitive in depicting the development of interstitial fibrosis (Fig. 17.15) . Predominant upper lobe fibrosis and cavity or bulla formation are rare. This less common pattern of lung involvement is indistinguishable from that seen with ankylosing spondylitis and must be distinguished from postprimary fibrocavitary TB by acid-fast staining of sputum.
1211
TABLE 17.5 Manifestations of Rheumatoid Lung Disease
Manifestation
Radiographic
Findings
Serositis
Pleuritis
Pleural
Pericarditis
Pericardial
Interstitial pneumonitis
Necrobiotic
Caplan
nodules
syndrome
Bronchiolitis
effusion,
thickening
effusion
Pulmonary fibrosis predominance)
(basilar
Multiple
peripheral
cavitating
nodules
Multiple
peripheral
cavitating
nodules
Hyperinflation
obliterans
Pulmonary
arteritis
Pulmonary arterial hypertension right heart enlargement
Pulmonary
1212
hemorrhage
and
FIGURE
17.14. Honeycombing
in
Rheumatoid
Lung.
Posteroanterior radiograph in a patient with end-stage rheumatoid lung disease demonstrates a medium reticular process representing honeycomb lung. Note the predominant peripheral distribution of disease. Bilateral pleural effusions and cardiac enlargement caused by pericardial effusion are also evident.
1213
FIGURE
17.15. Rheumatoid
Lung
Disease. An HRCT through
the lung bases in a patient with rheumatoid arthritis shows a focal area of honeycombing at the periphery of the right lung base (curved arrow). Note also the left pleural effusion (straight arrow), a frequent finding in rheumatoid disease.
P.489 Less common parenchymal manifestations of rheumatoid disease are lung nodules and changes attributable to COP. Necrobiotic (rheumatoid) nodules in the lung can produce peripheral well-defined nodular opacities on chest radiographs that are indistinguishable from the subcutaneous rheumatoid nodules seen on the extensor surfaces of the elbows and knees in these patients. The lung nodules commonly evolve into thick-walled cavities, which tend to wax and wane in parallel with the flares of arthritis. Similar nodules may develop in the lungs of coal miners and silica or asbestos workers with rheumatoid arthritis as a hypersensitivity response to inhaled dust particles (Caplan syndrome). Caplan syndrome is usually indistinguishable radiographically from the necrobiotic nodules of simple rheumatoid disease, although the presence of the associated characteristic small nodular or irregular parenchymal opacities of simple pneumoconiosis helps make this distinction. COP and bronchiolitis obliterans (constrictive bronchiolitis) are associated rheumatoid disease. The clinical, functional, and radiographic
1214
with
findings are similar to those of COP or bronchiolitis obliterans associated with systemic lupus erythematosus (SLE), drugs, or viral infection. Pleuritis is the most common thoracic manifestation of rheumatoid disease and is found in 20% of patients. As with pulmonary involvement, there is a male predilection for pleural disease. Unilateral or bilateral pleural effusions may be seen that are exudative and have a characteristically low glucose concentration. Enlargement of the central pulmonary arteries and RV dilatation may be seen on chest radiographs in patients with pulmonary arterial hypertension. This is an uncommon manifestation of rheumatoid disease that usually develops secondary to diffuse interstitial fibrosis. Rarely, the pulmonary arteries are involved as a part of the systemic vasculitis seen in extra-articular rheumatoid disease. There are no parenchymal abnormalities associated with rheumatoid pulmonary arteritis. Abnormalities that may be seen in the chest wall of individuals with rheumatoid arthritis include tapered erosion of the distal clavicles, rotator cuff atrophy with a high-riding humeral head, bilateral symmetric glenohumeral joint space narrowing with or without superimposed degenerative joint disease, and superior rib notching or
erosion.
Systemic
Lupus
Erythematosus
This disease of young and middle-aged women typically involves inflammation of multiple organs mediated by autoantibodies and circulating immune complexes. The thorax is commonly affected and may be the initial site of involvement. The thoracic disease is often limited to the pleura and pericardium, although the lung, heart, diaphragm, and intercostal muscles are involved in as many as one third of patients. In the pleura and pericardium, a fibrinous serositis produces painful pleural and pericardial effusions that are exudative in nature. Radiographically, the pleural effusions are small or moderate in size and can be unilateral or bilateral. The effusions usually resolve with corticosteroid therapy. Pleural fibrosis, seen in a
1215
majority of patients with long-standing disease, results in diffuse pleural thickening. Pulmonary involvement may take the form of acute lupus pneumonitis or chronic interstitial disease. Acute lupus pneumonitis is characterized by rapid onset of fever, dyspnea, and hypoxemia, which occasionally requires mechanical ventilation. These patients have pathologic changes that are indistinguishable from those seen in ARDS, with diffuse alveolar damage producing an exudative intraalveolar edema with hyaline membrane formation. Radiographically, rapidly coalescent bilateral airspace opacities are seen, while the typical HRCT finding is one of ground-glass opacity. These findings are difficult to distinguish from those seen in diffuse alveolar hemorrhage associated with pulmonary vasculitis, severe pneumonia related to immunosuppressive therapy, or pulmonary edema secondary to renal failure. The diagnosis of acute lupus pneumonitis is made by excluding pneumonia and pulmonary edema and by noting an improvement following the initiation of immunosuppressive therapy. Radiographic evidence of IPF is distinctly uncommon in SLE, but fibrosis is said to be present pathologically in one third of patients. When seen radiographically, the pattern is one of bibasilar reticular opacities that are indistinguishable from those seen in rheumatoid lung disease or scleroderma. Therefore, the presence of severe interstitial fibrosis in a patient with clinical features of SLE should prompt consideration of the diagnosis of an overlap syndrome (mixed connective tissue disease). As with rheumatoid lung disease and scleroderma, HRCT is the most sensitive technique for demonstrating early interstitial disease. Additional chest radiographic findings in SLE include elevation of the hemidiaphragms with decreased lung volumes and resultant bibasilar areas of linear atelectasis. Diaphragmatic elevation is present in as many as 20% of patients and is the result of diaphragmatic weakness from a primary myopathy unrelated to corticosteroid therapy. Rarely, the central pulmonary arteries are enlarged from pulmonary arterial hypertension secondary to pulmonary vasculitis. Pulmonary embolism
1216
with or without infarction may produce peripheral parenchymal opacities and results from deep venous thrombosis that develops in the presence of a circulating lupus anticoagulant. COP has been described in patients with SLE but is indistinguishable clinically and radiographically from lupus pneumonitis, because both conditions produce parenchymal opacities that are responsive to steroids. Superior rib erosions may be seen that are indistinguishable from similar findings in rheumatoid arthritis or scleroderma. Scleroderma produces inflammation and fibrosis of the skin, esophagus,
musculoskeletal
system,
heart,
lungs, P.490
and kidneys in young and middle-aged women. The etiology and pathogenesis are unknown. The lungs are involved pathologically in nearly 90% of patients, although only 25% of patients have respiratory symptoms or radiographic evidence of pulmonary involvement. Pulmonary function testing is more sensitive than conventional radiography in the diagnosis of lung disease and shows the typical diminished lung volumes, preserved flow rates, and low diffusing
capacity
of
interstitial
pulmonary
fibrosis.
Pathologically,
the sequence of parenchymal and radiographic changes is indistinguishable from rheumatoid lung disease, IPF, and other forms of UIP. Severe pulmonary involvement is reflected radiographically as a coarse reticular or reticulonodular pattern involving the subpleural regions of the lower lobes. The most common HRCT findings are interlobular
septal
thickening,
ground-glass
opacities,
and
honeycombing (Fig. 17.16). HRCT is more sensitive than the chest radiograph in detecting early interstitial disease. Progressive loss of lung volume is seen with advancing pulmonary fibrosis. The development of large (1 to 5 cm) subpleural lower lobe cysts may lead
to
spontaneous
pneumothorax.
1217
FIGURE 17.16. Scleroderma. An HRCT of a patient with scleroderma demonstrates minimal interstitial disease in the periphery of the middle lobe and lingula (small arrows), with honeycombing in the posterior subpleural portions of the lower lobes (large arrows) .
Pulmonary arterial hypertension with enlarged central pulmonary arteries and RV dilatation is seen in up to 50% of patients with scleroderma and may be seen in the absence of interstitial fibrosis. In these patients, thickening and obliteration of small muscular pulmonary arteries and arterioles are responsible for the development of pulmonary arterial hypertension. Pleural effusions are significantly less common in scleroderma than in rheumatoid disease or SLE and may be a helpful distinguishing feature radiographically. Pleural thickening is more often attributable to extension of pulmonary interstitial fibrosis into the interstitial layer of the pleura than to pleuritis. Several additional chest radiographic findings may be seen in patients with scleroderma. Eggshell calcification of mediastinal lymph nodes has been reported, although it is more common in silicosis and
1218
sarcoidosis. A dilated air-filled esophagus may be identified on the upright chest radiograph and is a manifestation of esophageal dysmotility from smooth muscle atrophy and fibrosis. An air–fluid level within a dilated esophagus suggests secondary distal esophageal stricture formation from chronic reflux esophagitis. The functional or anatomic esophageal obstruction may result in aspiration with the development of lower lobe pneumonia. Because patients with scleroderma are at a greater risk for developing lung cancer, particularly bronchioloalveolar cell carcinoma, the appearance of a mass or persistent possibility. Patients with the CREST calcification, Raynaud phenomenon, sclerodactyly, and telangiectasia), a have
radiographically
visible
airspace opacity should raise this syndrome (subcutaneous esophageal dysmotility, variant of scleroderma, may
calcifications
within
the
subcutaneous
tissues of the chest wall. Superior rib notching or erosion may be seen. Dermatomyositis
and
polymyositis involve autoimmune
inflammation and destruction of skeletal muscle, producing proximal muscle pain and weakness (polymyositis) and occasionally associated with a skin rash (dermatomyositis). The thoracic manifestations of these diseases include respiratory and pharyngeal muscle weakness and an associated interstitial pneumonitis. Interstitial pneumonitis, indistinguishable from that associated with rheumatoid lung disease, SLE, scleroderma, or IPF, is seen in 5% to 10% of patients. A fine reticular interstitial pattern in acute disease leads to a chronic, coarse reticular or reticulonodular process that is predominantly basilar in distribution. Most patients with polymyositis and interstitial lung disease have clinical manifestations of rheumatoid arthritis or scleroderma, and these patients tend to respond favorably to corticosteroids. As with scleroderma, the early parenchymal changes may be subradiographic but can be demonstrated on HRCT studies through the lower lobes (Fig. 17.17). Additional chest radiographic findings in polymyositis reflect the involvement of skeletal muscle. Small lung volumes with diaphragmatic elevation and basilar linear atelectasis are secondary to diaphragmatic and intercostal muscle involvement. Pharyngeal and upper esophageal muscle weakness
1219
predispose to aspiration pneumonia. The chest radiograph should be examined carefully for lung masses because bronchogenic carcinoma accounts for a significant percentage of the malignancies seen with a higher-than-normal frequency in patients with dermatomyositis or polymyositis.
Sjögren
Syndrome
This autoimmune disorder of middle-aged women is characterized by the sicca syndrome of dry eyes (keratoconjunctivitis sicca), dry mouth P.491 (xerostomia), and dry nose (xerorhinia), which result from lymphocytic infiltration of the lacrimal, salivary, and mucous glands, respectively. Most patients with the sicca syndrome have associated manifestations of other collagen vascular diseases, such as rheumatoid
arthritis,
scleroderma,
or
SLE.
FIGURE 17.17. Polymyositis. An HRCT through the lung bases shows reticulation and nodules, reflecting interstitial pneumonitis in a patient with polymyositis.
1220
The chest is involved in approximately one third of patients with Sjögren syndrome with or without associated collagen vascular disease. The most common manifestation is interstitial fibrosis, which is indistinguishable from that seen with other collagen vascular disorders. Involvement of tracheobronchial mucous glands leads to thickened sputum with mucus plugging and recurrent bronchitis, bronchiectasis, atelectasis, and pneumonia. HRCT demonstrates both interstitial opacities and the presence of small airways involvement with bronchiolectasis and a “tree-in-bud― and pleural effusion are less common.
appearance.
Pleuritis
Patients with Sjögren syndrome are at increased risk for developing lymphocytic interstitial pneumonitis (LIP) and non-Hodgkin pulmonary lymphoma. The radiographic appearance of LIP is lower lobe coarse reticular or reticulonodular opacities that are indistinguishable from interstitial fibrosis. HRCT shows ground-glass opacity with scattered, thin-walled cysts. The development of lymphoma in these patients should be suspected when nodular or alveolar opacities develop in the lung in association with mediastinal lymph
node
enlargement.
Ankylosing
Spondylitis
Approximately 1% to 2% of individuals with ankylosing spondylitis develop pulmonary disease in the form of upper lobe pulmonary fibrosis. The fibrotic changes are commonly associated with the development of bullae and cavities, which are prone to mycetoma formation with Aspergillus. The diagnosis should be suspected in a young to middle-aged man with characteristic spine changes (kyphosis, spinal ankylosis) seen in association with abnormally increased lung volumes and upper lobe fibrobullous disease, the latter of which simulates postprimary fibrocavitary TB.
Overlap Syndromes Tissue Disease
and
Mixed
Connective
Some patients with collagen vascular disease have features of more
1221
than one of the recognized syndromes discussed above. These patients are classified as having an overlap syndrome with thoracic manifestations characteristic of the other disorders. Patients with a distinct form of overlap syndrome, called mixed connective tissue disease, have clinical features of SLE, scleroderma, and polymyositis and have serum antibodies to extractable nuclear antigen. The thoracic manifestations of mixed connective tissue disease include IPF, pulmonary arterial hypertension caused by plexogenic pulmonary arteriopathy, and pleural effusion and thickening from a fibrinous pleuritis typical of SLE.
Idiopathic Chronic Pneumonias
Interstitial
The idiopathic interstitial pneumonias are characterized by an inflammatory process in the lung that can result in pulmonary fibrosis. These histologic terms provide the most precise method of classifying these disorders and include UIP, acute interstitial pneumonia (AIP), COP, respiratory bronchiolitis–associated interstitial lung disease (RB-ILD), DIP, and nonspecific interstitial pneumonia (NSIP) (2). Unfortunately, confusion arises when clinical terms are used interchangeably with the aforementioned histologic terms in describing these disorders. When possible (when the histology is known), it is most accurate to use the histologic term to describe a particular disorder, while reserving clinical terms such as IPF or rheumatoid lung for interstitial disease associated with specific clinical diseases for which histology is unavailable.
Usual
Interstitial
Pneumonia
UIP is the most common of the idiopathic interstitial pneumonias. It is likely the result of repetitive injury to the lung. The initial response in the lung is inflammation, which is followed by repair and eventually fibrosis. The pathologic abnormalities seen in UIP represent a spectrum of findings, characterized in the early stage of disease by marked proliferation of macrophages in the alveolar airspaces associated with a mild and uniform thickening of the
1222
interstitium by mononuclear cells. Late in the course of disease, the pathologic findings are characterized by thickening of the alveolar interstitium by mononuclear inflammatory cells and fibrous tissue. A distinguishing histologic feature of UIP is that different stages of the disease are seen simultaneously within different portions of the lung. Patients with UIP typically present in the fifth to seventh decades, with a slight male preponderance. Presenting P.492 symptoms include progressive dyspnea or a nonproductive cough. Pulmonary function tests show restrictive disease and a decreased diffusing capacity for carbon monoxide (DLCO). Most cases of UIP are sporadic, but up to 30% of patients with UIP have an associated collagen vascular or immunologic disorder. This is most often rheumatoid arthritis, but it can also be SLE, scleroderma, or dermatomyositis/polymyositis.
FIGURE
17.18. Usual Interstitial Pneumonia (UIP). A.
Posteroanterior radiograph in a patient with UIP demonstrates bilateral coarse reticular opacities and diminished lung volumes. B . An HRCT through the mid-lungs shows honeycombing in a peripheral, subpleural distribution. Traction bronchiectasis is evident (arrow) .
The radiographic manifestations of UIP parallel the pathologic
1223
changes. In the early phase of disease, the chest radiograph may appear normal despite the presence of clinical symptoms and abnormalities on pulmonary function testing. The earliest radiographic changes are bibasilar fine to medium reticular opacities or ground-glass density (Fig. 17.18). As the disease progresses, a coarse reticular or reticulonodular pattern is seen, which almost invariably leads to the formation of honeycomb cysts (3 to 10 mm in diameter) and progressive loss of lung volume. Extensive pulmonary fibrosis may be associated with findings of pulmonary arterial hypertension. Upper lobe bullae may be seen and predispose to the development of spontaneous pneumothorax. Hilar lymph node enlargement and pleural effusions have been described but are rare and should suggest an alternative diagnosis. HRCT findings in UIP differ with the stage of the disease and vary from one lung region to another. Patients with active inflammatory areas of disease, as demonstrated histologically by interstitial and intra-alveolar inflammatory changes, show areas of ground-glass density on HRCT. As fibrosis develops, findings include irregular septal or subpleural thickening (in contrast to the smooth septal thickening seen with edema or lymphangitic spread of carcinoma), intralobular lines, irregular interfaces, honeycombing, and traction bronchiectasis (Fig. 17.12). The changes are typically most severe in the peripheral and basal portions of the lungs, which can be helpful in differential diagnosis (Fig. 17.18). Mildly enlarged mediastinal lymph nodes are often seen. In most patients, the disease progresses inexorably, with an overall mean survival of 80 H) may be seen; associated rib fractures or subcutaneous emphysema should suggest the diagnosis. An acute hemothorax is treated with thoracostomy tube drainage, while thoracotomy is generally reserved for persistent bleeding or hypotension.
TABLE 19.2 Empyema Versus Lung Abscess on CT
Feature Shape
Empyema Oval,
oriented
Abscess Round
longitudinally
Margin
Thin,
smooth
(“split
Thick,
irregular
pleura―
sign)
Angle with
Obtuse
Acute
Effect on lung
Compression
Consumption
Treatment
External
Antibiotics, postural drainage
chest wall
drainage
Esophageal perforation from prolonged vomiting (Boerhaave syndrome) or as a complication of esophageal dilatation may produce a pleural effusion, most commonly on the left side. Extravascular placement of a central line can result in a hydrothorax when intravenous solution is inadvertently infused into the pleural or extrapleural space.
1317
Collagen
Vascular
and
Autoimmune
Disease
Systemic lupus erythematosus has a reported incidence of pleural effusions ranging from 33% to 74%. These exudative effusions are a result of pleural inflammation; patients present with pleuritic chest pain. In some cases, the nephrotic syndrome associated with systemic
lupus
erythematosus
may
produce
transudative
effusions. P.530
Cardiomegaly is a common chest radiographic finding and may be caused by pericardial effusion, hypertension, renal failure, or lupusassociated endocarditis or myocarditis.
FIGURE 19.2. Malignant Pleural Effusion: CT Diagnosis. HRCT in a patient with breast carcinoma and a right pleural effusion on chest radiographs demonstrates multiple bilateral pleural nodules and masses (arrows) representing pleural metastases. The diagnosis was confirmed by US-guided right pleural
biopsy.
Pleural effusion is the most common intrathoracic manifestation of rheumatoid arthritis and is most frequently seen in male patients following the onset of joint disease (Fig. 19.3). The effusions occur
1318
independent of pulmonary parenchymal involvement but may develop following intrapleural rupture of peripheral rheumatoid nodules. The effusions of rheumatoid arthritis are exudative, with lymphocytosis, low glucose concentration, and low pH (1 cm) and lobulated diffuse pleural thickening (1 0). Calcification or, rarely, ossification is seen in 20% of tumors, although calcified pleural plaques may be seen in uninvolved areas of the pleura. A pleural effusion is often present, which, if large, may obscure the pleural tumor. Malignant involvement of the mediastinal pleural surface may prevent contralateral mediastinal shift despite extensive pleural tumor volume and effusion, a finding that may help distinguish mesothelioma from metastatic disease. CT is the imaging modality of choice in the evaluation of malignant mesothelioma and depicts the extent of pleural involvement and invasion of the chest wall and mediastinum (Fig. 19.14) . Diaphragmatic invasion by tumor, best assessed by coronal MR or reformatted multidetector CT (MDCT) scans, is important only in those patients who are considered for resection (Fig. 19.15) . Adenopathy is seen in the ipsilateral hilum and mediastinum in approximately 50% of patients. While the radiologic findings may be highly suggestive of mesothelioma, metastatic pleural malignancy can have a similar appearance, so histologic confirmation is necessary. The diagnosis of malignant mesothelioma is made histologically and often requires the use of special stains. The epithelial type of malignant mesothelioma may be indistinguishable from adenocarcinoma on light microscopy. While surgical resection by pleurectomy or extrapleural pneumonectomy may benefit selected patients with limited disease and good pulmonary reserve, the median survival from the time of diagnosis is only 6 to 12 months.
CHEST
WALL
Disorders of the soft tissues or bony structures of the chest wall may come to attention because of local symptoms P.540 or physical findings, during evaluation of pulmonary or pleural
1345
disease, or as an incidental finding on radiographic studies (Table 19.6) .
FIGURE
19.14. Malignant
Mesothelioma.
A. A posteroanterior
chest radiograph in a 34-year-old man evaluated for a positive purified protein derivative tuberculin skin test (PPD) reveals lobulated right pleural thickening encompassing the right lung. B . A CT scan through the lung bases demonstrates a circumferential pleural soft tissue process traversing the major fissure (arrows) . US-guided pleural biopsy revealed malignant mesothelioma.
1346
FIGURE 19.15. Diaphragmatic Invasion of Malignant Mesothelioma on MR. Sagittal T1WI through the dome of the right hemidiaphragm in a 53-year-old man with known pleural mesothelioma shows a basilar pleural mass (straight arrows) invading through the right diaphragm (curved arrows) into the liver. Laparoscopy confirmed the MR findings, which rendered the patient
Soft
unresectable.
Tissues
Congenital absence of the pectoralis muscle results in hyperlucency of the affected hemithorax on frontal P.541 radiographs. Poland syndrome is an autosomal recessive disorder characterized by unilateral absence of the sternocostal head of the pectoralis major, ipsilateral syndactyly, and rib anomalies. There may be associated aplasia of the ipsilateral breast (Fig. 19.16) . Patients who have had a mastectomy will also show unilateral
1347
hyperlucency. In those who have undergone a modified radical mastectomy, the horizontally oriented inferior edge of the hypertrophied pectoralis minor muscle may be identified on frontal radiographs.
TABLE 19.6 Chest Wall Lesions
Tumors
Benign Mole Nevus Wart Neurofibroma Lipoma Hemangioma Desmoid Malignant Fibrosarcoma Liposarcoma Metastases Melanoma Bronchogenic
carcinoma
Askin tumor (primitive neuroectodermal tumor)
Infection (abscess)
Staphylococcus Tuberculosis
Trauma
Hematoma
1348
FIGURE
19.16. Poland
Syndrome. Contrast-enhanced CT in a
62-year-old woman with Poland syndrome shows hypoplasia of right anterior ribs with absence of right pectoral muscles.
A variety of skin lesions such as moles, nevi, warts, neurofibromas, and accessory nipples may produce a nodular opacity on frontal radiographs that mimics a solitary pulmonary nodule. Examination of the skin surface should be performed in any patient with a new nodular opacity seen on chest radiographs, and repeat radiographs obtained with a radiopaque marker over the skin lesion will confirm the nature of the opacity and avoid unnecessary follow-up radiographs and chest CT. Chest wall abscesses may present as localized, painful, fluctuant subcutaneous masses. Staphylococcus and Mycobacterium tuberculosis are the most common organisms responsible. The diagnosis is usually obvious clinically. Chest radiographs demonstrate a poorly defined opacity on the frontal radiograph when the abscess involves the anterior or posterior chest wall. CT shows a localized fluid collection with an enhancing wall and is used to determine the location and extent of the collection prior to open drainage.
1349
Soft tissue neoplasms of the chest wall are rare (1 1). They are most often detected clinically as a mass protruding from the chest wall and appear as nonspecific extrathoracic soft tissue masses on chest radiographs. The most common benign neoplasm of the chest wall is a lipoma. Lipomas may be intrathoracic or extrathoracic, or they may project partially within and outside the thorax (dumbbell lipoma). CT shows a sharply circumscribed mass of fatty density (Fig. 19.17) , while MR shows characteristic high and intermediate signal intensity on T1WIs and T2WIs, respectively. A desmoid tumor is a rare fibroblastic tumor arising within striated muscle that is histologically benign but has a tendency for local invasion. Desmoids are most common in the abdominal wall musculature of multiparous women but may arise in the chest wall musculature following local trauma. Hemangiomas are uncommon chest wall tumors. While they are often indistinguishable from other soft tissue tumors radiographically, the recognition of phleboliths, hypertrophy of involved bones, or the identification of vascular channels on contrast-enhanced CT or MR studies should suggest the diagnosis.
1350
FIGURE 19.17. Chest Wall Lipomas. Unenhanced CT scan shows sharply circumscribed homogeneous fatty masses in the left pectoral (straight arrows) and rhomboid major (curved arrows) muscles.
Fibrosarcomas and liposarcomas are the most common malignant soft tissue neoplasms of the chest wall in adults. Malignant tumors often present with symptoms of localized chest wall pain and a visible, palpable mass. Patients who have received chest wall radiation are at particular risk for developing sarcomas. Radiographically, these soft tissue masses are often associated with bony destruction. CT best depicts the bone destruction and intrathoracic component of tumor, while MR shows the extent of tumor and delineates tumor from surrounding muscle and subcutaneous fat (8). A rare malignant neoplasm arising from the chest wall of children and young adults is an Askin tumor, which arises from primitive neuroectodermal P.542 rests in the chest wall (Fig. 19.18). These lesions are very aggressive and associated with a high mortality rate.
FIGURE 19.18. Askin Tumor (Primitive Neuroectodermal Tumor) of Chest Wall. A. Contrast-enhanced CT in a 32-yearold man demonstrates a left pleural mass with adjacent
1351
involvement of the rib and associated pleural effusion. B . Repeat CT obtained 1 month later shows enlargement of the mass with progressive rib involvement and a large pleural effusion with contralateral mediastinal shift. Surgical resection revealed an Askin tumor.
The Bony Thorax Congenital Anomalies (Table 19.7). The most common congenital anomalies of the ribs are bony fusion and bifid ribs, neither of which have clinical significance. Intrathoracic ribs are extremely rare congenital anomalies where an accessory rib arises from a vertebral body or the posterior surface of a rib and extends inferolaterally into the thorax, usually on the right side. Osteogenesis imperfecta and neurofibromatosis may be associated with thin, wavy, “ribbon― ribs. A relatively common congenital anomaly is the cervical rib, which arises from the seventh cervical vertebral body. Cervical ribs are usually asymptomatic, although in a minority of individuals with the thoracic outlet syndrome, the rib or associated fibrous bands can compress the subclavian artery, producing secondary ischemic symptoms, or compress the subclavian vein and brachial plexus, producing pain, weakness, and swelling of the upper extremity. Surgical resection of the cervical rib can relieve the symptoms in selected patients. Rib
notching is seen in a variety of pathologic conditions. Inferior
rib notching is much more common than superior rib notching and is caused by enlargement of one or more of the structures that lie in the subcostal grooves (intercostal nerve, artery, or vein). The notching predominantly affects the posterior aspects of the ribs bilaterally and may be narrow, wide, deep, or shallow. The most common cause of bilateral inferior rib notching is coarctation of the aorta distal to the origin of the left subclavian artery. In this condition, blood circumvents the aortic obstruction and reaches the descending aorta via the subclavian, internal mammary, and intercostal arteries. The increased blood flow in the
1352
intercostal arteries produces tortuosity and dilatation of these vessels, which erodes the inferior margins of the adjacent ribs. Other causes of aortic obstruction that can lead to inferior rib notching include aortic thrombosis and Takayasu aortitis. Congenital heart diseases associated with decreased pulmonary blood flow may be associated with rib notching as the intercostal arteries enlarge in an attempt to supply collateral blood flow to the oligemic lungs. Superior vena cava obstruction can cause increased flow through intercostal veins and produce rib notching. Patients with aortic coarctation develop rib notching gradually; it is most common in adolescents and is rare in children under age 7. The first two ribs are uninvolved because the first and second intercostal arteries arise from the superior intercostal branch of the costocervical trunk of the subclavian artery and therefore do not communicate with the descending thoracic aorta. Coarctation may produce unilateral left rib notching when the aortic narrowing occurs proximal to an aberrant right subclavian artery. Unilateral right-sided notching occurs when the coarctation is proximal to the left subclavian artery. Additional causes of unilateral inferior rib notching include subclavian artery obstruction and surgical anastomosis of the proximal subclavian artery to the ipsilateral pulmonary artery (Blalock-Taussig procedure) Multiple intercostal neurofibromas in neurofibromatosis type 1 are the most common nonvascular cause of inferior rib notching. The neurofibromas
appear
as
multiple P.543
extrapleural soft tissue masses, most often seen in the upper paravertebral regions. Other thoracic bony manifestations of neurofibromatosis include ribbon ribs, thoracic kyphoscoliosis, and scalloping of the posterior aspect of the vertebral bodies caused by dural ectasia.
TABLE 19.7 Rib Lesions
1353
Congenital
Fusion anomalies Cervical rib Ribbon ribs Rib notching Inferior Coarctation of the aorta Tetralogy of Fallot Superior vena cava obstruction Blalock-Taussig shunt (unilateral right) Neurofibromatosis Superior Paralysis Collagen
vascular
disease
Rheumatoid arthritis Systemic lupus erythematosus
Trauma
Healing rib fracture
Nonneoplastic tumors
Fibrous dysplasia Eosinophilic granuloma Brown tumor
Neoplasms
Benign Osteochondroma Enchondroma Osteoblastoma Malignant Primary Chondrosarcoma Osteogenic sarcoma Fibrosarcoma Metastatic Multiple myeloma Metastases Breast carcinoma
1354
Bronchogenic carcinoma Renal cell carcinoma Prostate carcinoma
Osteomyelitis
Staphylococcus
aureus
Tuberculosis Actinomycosis Nocardiosis
Superior rib notching is much less common than inferior rib notching. The pathogenesis of superior rib notching is unknown, although a disturbance of osteoblastic and osteoclastic activity and the stress effect of the intercostal muscles are proposed mechanisms. Paralysis is the most common condition associated with superior rib notching. Other etiologies include rheumatoid arthritis, systemic lupus erythematosus, and rarely, marked tortuosity of the intercostal arteries
from
severe,
long-standing
aortic
obstruction.
Trauma Rib and costal cartilage fractures may result from blunt or penetrating trauma to a normal ribcage or from minimal trauma to abnormal ribs, such as those affected by metastases. An acute rib fracture is seen as a thin vertical lucency; malalignment of the superior and inferior cortices of the rib may occasionally be the only radiographic finding. The tendency to affect the posterolateral aspects of the ribs explains the utility of obtaining ipsilateral posterior oblique radiographs for suspected fracture, because this projection best displays the fracture line. In any patient with an acute rib fracture, a careful search should be made for associated pneumothorax, hemothorax, and pulmonary contusion or laceration. Since the first three ribs are well protected by the clavicles, scapulae, and shoulder girdles, fracture of these ribs indicates severe trauma and should prompt a careful evaluation for associated great vessel and visceral injuries. Fracture of the tenth, eleventh, or twelfth ribs may be associated with injury to the liver or spleen.
1355
Severe blunt trauma to the ribcage, in which multiple contiguous ribs are fractured in more than one place, is termed a “flail chest.― This results in a free segment of the chest wall that moves paradoxically inward on inspiration and outward on expiration. Healing rib fractures will demonstrate callus formation, which may be exuberant in patients receiving corticosteroids. Multiple contiguous healed rib fractures, particularly if bilateral, should suggest chronic alcoholism or a prior motor vehicle accident.
Nonneoplastic
Lesions
The ribs are the most common site of involvement by monostotic fibrous dysplasia. The typical appearance is an expansile lesion in the posterior aspect of the rib with a lucent or ground-glass density; rarely, the lesion is sclerotic. Multiple rib involvement from polyostotic fibrous dysplasia can result in severe restrictive pulmonary disease. Eosinophilic granuloma can cause lytic lesions in patients under age 30. These are usually solitary lytic lesions, which can be expansile but do not have sclerotic margins; this latter feature helps distinguish these lesions from fibrous dysplasia (Fig. 19.19). Brown tumors from hyperparathyroidism can also produce lytic rib lesions.
Neoplasms Primary osteochondral neoplasms or metastatic disease can involve the ribs. Osteochondromas are the most common benign neoplasm of ribs, followed in relative frequency by enchondromas and osteoblastomas (Fig. 19.20). Primary malignant neoplasms of the ribs in adults are uncommon. Chondrosarcoma is the most common primary rib malignancy, with osteogenic sarcoma and fibrosarcoma less common. Rib involvement from multiple myeloma or metastatic carcinoma can produce solitary or multiple lytic lesions and is much more common than primary tumors. Myeloma can also cause permeative bone destruction that is indistinguishable from severe osteoporosis. The diagnosis of myeloma is made by identification of a monoclonal spike on serum protein
1356
P.544 electrophoresis and typical findings of abnormal aggregates of plasma cells on bone marrow biopsy. The most common metastatic lesions to ribs are from bronchogenic and breast carcinoma, which produce multiple lytic lesions when dissemination is hematogenous or localized rib destruction when invasion is by direct contiguous spread. Expansile lytic rib metastases are seen most commonly from renal cell and thyroid carcinoma. Sclerotic rib metastases are most commonly seen in breast and prostate carcinoma.
1357
FIGURE 19.19. Eosinophilic Granuloma of Rib. Chest radiograph (A) in a 24-year-old with pleuritic chest pain shows a lytic lesion of the posterolateral right seventh rib (arrowhead) . Thin-section CT (B) confirms the presence of an expansile lytic lesion (arrowhead), which was positive (arrowhead) on bone scan (C). Surgical resection showed eosinophilic granuloma of bone.
Infection Chest wall infection and osteomyelitis of the ribs usually develop from contiguous spread from the lung, pleural space, and vertebral column. Less commonly, infection complicates penetrating chest trauma or spreads to the ribs hematogenously. Pleuropulmonary infections that may traverse the pleural space and produce a chest wall infection include TB, actinomycosis, and nocardiosis. Radiographs may demonstrate bone destruction, periostitis, and subcutaneous emphysema; bone scans can detect subradiographic bone involvement. CT can demonstrate bone destruction, soft tissue swelling, and abscesses within the chest wall. Additionally, CT may show involvement of the adjacent pleural space, lung, sternum, or vertebral column.
Costal
Cartilages
Ossification of the costal cartilages is a normal finding on frontal chest radiographs in adults. Female costal cartilage ossification involves the central portion of the cartilage, extending from the rib toward the sternum in the shape of a solitary finger, while male costal cartilage ossification involves the peripheral portion of the cartilage and has the appearance of two fingers (“peace― sign). These typical patterns of male and female costal cartilage ossification are seen in 70% of patients (Fig. 19.21) and do not apply to the first rib.
Scapula
1358
Scapular abnormalities that are visible on frontal radiographs include congenital, posttraumatic, and neoplastic lesions. Sprengel deformity is a congenital anomaly in which the scapula is hypoplastic and elevated. The association of Sprengel deformity with an P.545 omovertebral bone, fused cervical vertebrae, hemivertebrae, kyphoscoliosis, and rib anomalies is termed the Klippel-Feil syndrome. Scapular fractures may result from direct trauma to the upper back and shoulder or from impaction of the humeral head into the glenoid. A winged scapula is identified when the scapula is superiorly displaced from its normal position and the inferior portion is posteriorly displaced from the chest wall, thereby foreshortening its appearance on the frontal radiograph. This deformity results from disruption in the innervation of P.546 the serratus anterior muscle that maintains the scapula against the chest wall. Metastatic disease to the scapula is recognized by the presence of lytic destructive lesions; bronchogenic and breast carcinomas are the most common primary malignancies.
FIGURE 19.20. Chondrosarcoma of Rib. A. Posteroanterior chest radiograph in a 37-year-old man with a 3-month history of
1359
right shoulder pain demonstrates a right apical extrapulmonary mass. B . A CT scan reveals a bone-forming mass arising from the right third costotransverse junction, with erosion of the adjacent vertebral body. This chondrosarcoma was successfully resected by a combined thoracic and neurosurgical approach.
FIGURE 19.21. Normal Ossification Patterns in Men and Women. Shaded-surface three-dimensional reconstructions of the anterior chest wall show typical ossification patterns of costal cartilages in a woman (A) and a man (B).
Clavicle A variety of diseases can affect the clavicle. The clavicle is involved in cleidocranial dysostosis, in which there is partial or complete aplasia of the clavicle. The distal third of the clavicle is commonly fractured in blunt trauma. Rheumatoid arthritis and hyperparathyroidism are associated with erosion of the distal clavicles. The distal clavicle is sharply defined in rheumatoid arthritis and tapers to a point, whereas in hyperparathyroidism it is often
1360
widened and irregular. Additional findings in rheumatoid arthritis include narrowing of the glenohumeral joint and a high riding humeral head caused by rotator cuff atrophy. Primary malignant neoplasms of the clavicle include Ewing or osteogenic sarcoma. Metastases to the clavicle are usually associated with lesions in other portions of the bony thorax. Osteomyelitis of the clavicle is uncommon and is most often seen in intravenous drug users. Paget disease can involve the clavicle, but there is often concomitant pelvic bone and calvarial involvement.
Thoracic
Spine
Numerous thoracic spine abnormalities are visible on chest radiographs. Congenital anomalies, including hemivertebrae, butterfly vertebra, spina bifida, and scoliosis, can be seen on wellpenetrated frontal radiographs. Vertebral compression fractures caused by trauma, osteoporosis, or metastases are best seen on lateral radiographs and may produce an exaggerated kyphosis. Large bridging osteophytes may mimic a paraspinal mass on frontal radiographs or a pulmonary nodule on lateral films. Vertebral osteomyelitis is seen as destruction of vertebral bodies and intervertebral discs, often associated with a paraspinal abscess (Fig. 19.22). Chronic anemia in patients with thalassemia major or sickle cell disease may result in prevertebral or paravertebral masses of extramedullary
hematopoiesis,
which
represent
herniated
hyperplastic bone marrow. Sickle cell anemia produces a characteristic appearance of H-shaped or “Lincoln log― vertebrae on lateral chest radiographs that is pathognomonic of this disease. Similarly, a “rugger jersey― appearance to the thoracic spine on lateral chest films suggests renal osteosclerosis.
Sternum Developmental
sternal
deformities
include
pectus
excavatum
(funnel
chest), pectus carinatum (pigeon breast), and abnormal segmentation. In pectus excavatum, the sternum is inwardly depressed and the ribs protrude anterior to the sternum. It often has
1361
an autosomal dominant pattern of inheritance but may occur sporadically. Pectus excavatum is commonly associated with congenital connective tissue disorders, such as Marfan syndrome, Poland syndrome, osteogenesis imperfecta, and congenital scoliosis. Most patients are asymptomatic. A clinically insignificant systolic murmur can result from compression of the right ventricular outflow tract, although some patients with pectus deformities and systolic murmurs have mitral valve prolapse. Pectus excavatum has a characteristic appearance on frontal chest radiograph. The heart is displaced to the left, and the combination of the depressed soft tissues of the anterior chest wall and the vertically oriented anterior ribs results in loss of the right heart border. The findings on frontal radiographs may be mistakenly attributed to middle lobe opacification from pneumonia or atelectasis. The typical inward depression of the midsternum and lower sternum is seen on lateral chest radiographs (Fig. 19.23). CT helps define the deformity and its effect upon the heart and mediastinal structures.
FIGURE
19.22. Vertebral
Osteomyelitis. Sagittal maximum
1362
intensity projection reconstruction in a 72-year-old woman with back pain and staphylococcal sepsis shows an expansile lesion (arrow) of a midthoracic vertebral body with prevertebral soft tissue mass. CT-guided aspiration of the paravertebral mass revealed Staphylococcus.
Pectus carinatum is an outward bowing of the sternum that may be congenital or acquired. The congenital form is seen more commonly in boys and in families with a history of chest wall deformities or scoliosis. Congenital atrial or ventricular septal defects and severe childhood asthma account for the majority of the acquired cases of pectus carinatum. Affected patients are asymptomatic. The characteristic outward bowing of the sternum with deepening of the retrosternal airspace is seen on lateral radiographs. Severe blunt trauma to the chest, most often associated with deceleration injury from a motor vehicle accident, can result in fracture or dislocation of the sternum. Sternal body fracture and sternomanubrial dislocation are associated with a 25% to 45% mortality rate from concomitant P.547 injuries to the aorta, diaphragm, heart, tracheobronchial tree, and lung. Sternal films or lateral radiographs will show the fracture and often demonstrate a retrosternal hematoma; CT may be useful in those patients with normal plain films and a high suspicion of sternal injury.
1363
FIGURE
19.23. Pectus
Excavatum.
Posteroanterior (A) and
lateral (B) chest radiographs show changes of pectus excavatum. Note the apparent middle lobe opacity that is typical of this condition.
A prior median sternotomy is the most common sternal abnormality seen on conventional radiographs and chest CT. Circular wires encompassing the sternum are seen spaced along its length within the interspaces between costal cartilages. The vertical lucency representing the sternotomy may heal, but in many patients bony union does not occur. In the early postoperative period, a retrosternal hematoma may be seen, which normally resolves within the first several weeks. The radiologist plays a key role in the evaluation of possible sternal wound infection. Plain film evidence of bony destruction and air in the sternal incision appearing days to weeks after sternotomy are specific but insensitive findings for osteomyelitis. Bone scans are not particularly useful, as there will be increased radionuclide uptake for months following sternotomy. CT is
1364
the modality of choice in the evaluation of sternal wound infection. The CT findings of sternal osteomyelitis include bone destruction, peristernal soft tissue mass, enhancing fluid collection, and gas. The extent of infection, specifically associated mediastinitis, can also be determined.
DIAPHRAGM Unilateral
Diaphragmatic
Elevation
The differential diagnosis of unilateral diaphragmatic elevation is listed in Table 19.8. Eventration of the diaphragm is a result of congenital absence or underdevelopment of diaphragmatic musculature. This produces a localized elevation of the anteromedial portion of the hemidiaphragm on frontal radiographs in older individuals (Fig. 19.24), which is indistinguishable on the right from the rare foramen of Morgagni hernia. Complete diaphragmatic eventration is usually left sided and is indistinguishable radiographically
from
diaphragmatic
paralysis.
Unilateral diaphragmatic paralysis is usually caused by surgical injury or neoplastic involvement of the phrenic nerve, which affects the right and left hemidiaphragms with equal frequency. Idiopathic phrenic nerve dysfunction resulting from a viral neuritis is a common cause of diaphragmatic paralysis in male patients and is usually right sided. A positive fluoroscopic or ultrasonographic sniff test (paradoxical superior movement of the diaphragm with sniffing, a result of the effects of negative intrathoracic pressure on a flaccid diaphragm during inspiration) is diagnostic. Chronic loss of lung volume, particularly from collapse or resection of the lower lobe, results in diaphragmatic P.548 elevation. This is also a common sequela of chronic cicatrizing atelectasis of the upper lobe from TB.
1365
TABLE
19.8
Unilateral
Diaphragmatic
Elevation
Eventration
Diminished volume
lung
Congenital Hypoplastic Acquired Lobar/lung Pulmonary
Paralysis
lung atelectasis resection
Idiopathic Iatrogenic phrenic nerve injury Phrenic crush (tuberculosis) Intraoperative Malignant invasion of phrenic nerve Bronchogenic carcinoma Inflammation of diaphragmatic muscle Pleuritis Lower
lobe
Subphrenic
Upper abdominal mass
*Apparent
diaphragmatic
pneumonia abscess
Hepatomegaly or liver mass Splenomegaly Gastric/colonic distention Ascites (usually bilateral) Diaphragmatic hernia* Subpulmonic pleural effusion*
elevation.
1366
FIGURE
19.24. Eventration of the Diaphragm.
Posteroanterior (A) and lateral (B) chest radiographs in an asymptomatic 61-year-old woman reveal marked elevation of the left hemidiaphragm representing diaphragmatic eventration.
An enlarged liver or hepatic mass can produce right hemidiaphragmatic elevation by direct pressure on the undersurface of the hemidiaphragm. Similarly, an enlarged spleen, gas-distended stomach, or enlarged splenic flexure can produce an elevated left hemidiaphragm. Irritation of the superior surface of the hemidiaphragm by a pleural or pleura-based parenchymal process (e.g., infarct, a subphrenic diaphragm to effusion may Bilateral
pneumonia) or of the undersurface of the diaphragm by abscess, hepatitis, or cholecystitis may cause the become flaccid, leading to elevation. A subpulmonic simulate an elevated hemidiaphragm.
Diaphragmatic
Elevation that is not effort related may
1367
be caused by a neuromuscular disturbance or intrathoracic or intraabdominal disease. Radiographically, the diaphragms are elevated on both frontal and lateral views. Bibasilar linear atelectasis or passive lobar or segmental lower lobe atelectasis may be seen. Bilateral phrenic nerve disruption or intrinsic diaphragmatic muscular disease will produce bilateral diaphragmatic paralysis and elevation. Common disorders include cervical cord injury, multiple sclerosis, and the myopathy P.549 associated with systemic lupus erythematosus. In these patients, fluoroscopic or real-time US imaging of the diaphragms demonstrates a positive sniff test. Lung restriction caused by interstitial fibrosis, bilateral pleural fibrosis, or chest wall disease (most commonly from obesity) can produce bilateral diaphragmatic elevation. An increase in intraabdominal volume, most often from ascites, hepatosplenomegaly, or pregnancy, can restrict diaphragmatic motion. These conditions may be distinguished from bilateral paralysis by observation of normal but diminished inferior excursion of the diaphragms on fluoroscopy, US, or inspiratory/expiratory radiographs.
Diaphragmatic
Depression
Depression and flattening of one hemidiaphragm is seen with unilateral overinflation of a lung, usually as a compensatory mechanism when the contralateral lung is small or as a result of a large ipsilateral pneumothorax. Distinction between these two entities is usually possible by the clinical history and by characteristic findings in those with pneumothorax. A tension pneumothorax may cause inversion of the hemidiaphragm. Bilateral diaphragmatic depression is either a permanent finding—a result of abnormally increased lung compliance in patients with emphysema—or a transient finding in those with asthma and expiratory air trapping.
1368
FIGURE
19.25. Hiatal
Hernia. Axial (A) CT scan with coronal
reconstruction (B) in a 73-year-old man shows a sliding hiatal hernia in the posterior mediastinum.
Diaphragmatic
Hernias
There are three types of nontraumatic diaphragmatic hernias. The most common is the esophageal hiatal hernia, which represents herniation of a portion of the stomach through the esophageal hiatus. These are usually seen as incidental asymptomatic masses on chest radiographs, although some patients may have symptoms of gastroesophageal reflux or, rarely, severe pain from strangulation of the herniated stomach. Hiatal hernias are seen projecting behind the heart on frontal chest radiographs in the immediate supradiaphragmatic region of the posterior mediastinum. An
1369
air–fluid level may be seen in the hernia. An esophagram is confirmatory. CT shows widening of the esophageal hiatus and depicts the contents of the hernia sac, which often include stomach, omental fat, and, rarely, ascitic fluid (Fig. 19.25) .
Bochdalek
Hernia
The foramen of Bochdalek is a defect site of the embryonic pleuroperitoneal the Bochdalek foramen present in the hypoplasia of the ipsilateral lung and
in the hemidiaphragm at the canal. Large hernias through neonatal period with respiratory distress. In adults,
small hernias through this foramen are common and are predominantly seen on the left side, presumably because of the protective effect of the liver, which prevents herniation of right infradiaphragmatic fat through the right foramen of Bochdalek. The hernia typically appears as a P.550 posterolateral mass above the left hemidiaphragm, although it can occur anywhere along the posterior diaphragmatic surface (Fig. 19.26). CT shows the diaphragmatic defect with herniation of retroperitoneal fat, omentum, spleen, or kidney.
1370
FIGURE 19.26. Foramen of Bochdalek Hernia. Posteroanterior (A) and lateral (B) chest radiographs in an asymptomatic 82year-old man show a mass arising from the posterolateral aspect of the left hemidiaphragm. C . A CT scan through the diaphragm shows fat herniating through bilateral Bochdalek hernias. Note the top of the right kidney within the herniated fat (arrow) .
Morgagni
Hernia
A defect in the parasternal portion of the diaphragm, the foramen of Morgagni, is the least common type of diaphragmatic hernia. A
1371
Morgagni hernia is invariably right sided and appears as an asymptomatic cardiophrenic angle mass. The diagnosis is made by noting herniation of omental fat, liver, or transverse colon through the paracardiac portion of the right hemidiaphragm on CT scans through the lung bases. The presence of omental vessels within a fatty paracardiac mass is diagnostic (Fig. 19.27). Coronal MR or US can demonstrate the diaphragmatic defect, distinguishing this entity from partial eventration of the hemidiaphragm. Traumatic
herniation of abdominal contents through a tear or rupture
of the central or posterior aspect of the hemidiaphragm may follow blunt thoracoabdominal trauma or penetrating injury (1 2). The left side is affected in more than 90% of cases because the liver dissipates the traumatic forces and protects the right hemidiaphragm from injury (Fig. 19.28). Radiographically, the diagnosis should be suspected when the left hemidiaphragmatic contour is indistinct or elevated or when gas-filled loops of bowel or stomach are seen in the left lower thorax following severe trauma. Early diagnosis is often difficult because associated thoracic and abdominal injuries may obscure the clinical and radiographic findings. The diagnosis P.551 is often made after the traumatic episode, with symptoms caused by intestinal obstruction with strangulation (pain, vomiting, fever) or compression of the left lung (cough, dyspnea, chest pain). In addition to the stomach, the small intestine, colon, omentum, spleen, kidney, and the left lobe of the liver can also herniate through the defect. The diagnosis is usually made by upper or lower GI contrast studies demonstrating bowel herniating into the thorax through a constricting diaphragmatic defect. The resultant narrowing or “waist― of the herniated intestine as it traverses the diaphragmatic defect differentiates a hernia from simple diaphragmatic elevation. Large diaphragmatic defects may be demonstrated on MDCT scans with coronal and sagittal reconstructions, which also characterize the P.552 herniated tissues and detect associated visceral injuries. In addition to the detection of intrathoracic herniation of abdominal contents,
1372
MDCT can directly depict the diaphragmatic defect, even in the absence of visceral herniation. Other CT findings suggestive of traumatic diaphragmatic injury include thickening or retraction of the diaphragm away from the traumatic injury, a narrowing or waist of the diaphragm on the herniated viscus (“collar― or “waist― sign) and contact between the posterior ribs and the liver (right-sided injury) or stomach (left-sided injury), termed the “dependent viscera― sign. US or MR are difficult to obtain in the acute trauma setting but are occasionally useful (1 2) .
1373
FIGURE 19.27. Foramen of Morgagni Hernia. Frontal (A) and lateral (B) chest radiographs in a 60-year-old woman reveal a large mass in the right cardiophrenic angle. C . CT scan at the level of the diaphragm shows a fatty pericardiac mass containing omental vessels. D . A more inferior scan demonstrates an abnormally high transverse colon (arrow), which is characteristic
1374
of this entity.
FIGURE
19.28. Traumatic
Diaphragmatic
Hernia. Frontal
chest radiograph (A) in a 37-year-old man who sustained blunt chest trauma shows a left pneumothorax (arrow) and an apparently elevated left diaphragm. CT through the lower chest (B) shows colon (arrowhead) in the lower left thorax with surrounding atelectasis and effusion. Note the posterior rib fracture (asterisk). Surgery confirmed left diaphragmatic injury.
Diaphragmatic
Tumors
Primary diaphragmatic tumors are rare, with an equal incidence of benign and malignant lesions. Benign lesions include lipomas, fibromas, schwannomas, neurofibromas, and leiomyomas. Echinococcal cysts and extralobar sequestrations may be found within the diaphragm. Fibrosarcomas are the most common primary malignant diaphragmatic lesion. Radiographically, they appear as focal extrapulmonary masses obscuring all or part of the hemidiaphragm and are indistinguishable from masses arising within
1375
the diaphragmatic pleura. CT may show the origin of the mass, although the relationship of the mass to the diaphragm is best appreciated on coronal MR images or transabdominal US. Direct invasion of the diaphragm by lower lobe bronchogenic carcinoma, mesothelioma, or a subphrenic neoplasm is much more common than primary diaphragmatic malignancy.
CONGENITAL
LUNG
DISEASE
Bronchogenic cysts represent anomalous outpouchings of the primitive foregut that no longer communicate with the tracheobronchial tree. They are commonly present as asymptomatic mediastinal masses and are discussed in detail in Chapter 13. Cystic
adenomatoid
malformation (CAM) is a lesion usually seen
in newborn infants, although it occasionally presents in childhood or early adulthood. Three pathologic subtypes of CAM have been described. The most common subtype is comprised of one or several large cysts that are lined by respiratory epithelium with scattered mucous glands, smooth muscle, and elastic tissue in their walls. Multiple smaller cystic structures are present in the intervening lung between the larger cysts. Radiographically, these lesions often appear as round, air-filled masses, which exert mass effect on the adjacent lung and mediastinum (Fig. 19.29). A CAM in the left lower lobe may be difficult to distinguish from a congenital diaphragmatic hernia. Delayed clearance of fetal fluid in the newborn may give the radiographic appearance of an intrapulmonary soft tissue mass. These lesions may be identified on prenatal US examination.
1376
FIGURE 19.29. Congenital Cystic Adenomatoid Malformation (CCAM). A. Frontal chest radiograph in a newborn shows a multicystic mass in the right mid-lung and lower lung. B . CT scan demonstrates a complex mass occupying the middle and right lower lobes with air-filled cysts and a solid component posteriorly. Surgery revealed a CCAM of the middle lobe.
P.553 Bronchial
atresia, a developmental stenosis or atresia of a lobar or
segmental bronchus, produces bronchial obstruction with resultant distal bronchiectasis. Most patients are asymptomatic and are first recognized by typical findings on frontal chest radiographs, namely a rounded, oval, or branching central lung opacity representing the obstructed, mucus-filled, dilated bronchus (mucocele) with hyperlucency in that portion of lung supplied by the atretic bronchus. The overinflated lobe or segment results from air trapping in the obstructed lung as air enters by collateral air drift on inspiration but cannot empty through the proximal tracheobronchial tree on expiration. The most common site of involvement is the apicoposterior segment of the left upper lobe, followed by the segmental bronchi of the right upper and middle lobes. The combination of a central mucocele with peripheral hyperlucency in a
1377
young, asymptomatic patient is virtually diagnostic of this disorder (1 3) . Neonatal
lobar
hyperinflation
(congenital
lobar
emphysema)
may develop from a variety of disorders that produce a check-valve bronchial obstruction. These include extrinsic compression by mediastinal bronchogenic cysts, anomalous left pulmonary artery, congenital deficiency of bronchial cartilage, and congenital or acquired bronchial stenosis. The bronchial obstruction leads to air trapping on expiration, with resultant overinflation of the distal lung. In order of decreasing frequency, the left upper lobe, right middle lobe, and right upper lobe are the most common sites of involvement. Respiratory difficulties are usually evident within the first month of life, with a minority presenting later. Radiographically, hyperlucency of the affected lobe is seen with compression of adjacent lung, diaphragmatic depression, and contralateral mediastinal shift (Fig. 19.30). These findings are accentuated on expiratory films or on decubitus films obtained with the affected side down. CT, particularly when performed in expiration or with the affected side down, shows a hyperlucent, overexpanded lobe with attenuated blood vessels. Because many of these cases are not truly congenital but rather arise in the neonatal period from acquired abnormalities and because overinflation of normal alveoli without destruction of alveolar walls is seen pathologically, the term neonatal lobar hyperinflation has been used to more appropriately describe this syndrome. Treatment is surgical for symptomatic patients, whereas relatively asymptomatic patients are observed for spontaneous resolution. The findings in bronchial atresia and congenital lobar emphysema are reviewed in Table 19.9. Bronchopulmonary
sequestration is a congenital abnormality
resulting from the independent development of a portion of the tracheobronchial tree that is isolated from the normal lung and maintains its fetal systemic arterial supply. Grossly, the sequestered lung is cystic and bronchiectatic. These patients most often present with recurrent pneumonia from recurrent infection in the sequestered lung, although some (mostly extralobar sequestrations) are discovered as asymptomatic posterior mediastinal masses on routine
1378
radiographs. Pulmonary sequestration is divided into intralobar and extralobar forms (Table 19.10). Intralobar sequestration is contained within the visceral pleura of the normal lung. Extralobar sequestration is enclosed by its own visceral pleural envelope and may be found adjacent to the normal lung or within or below the diaphragm. Most patients P.554 with intralobar sequestration present with pneumonia. Extralobar sequestration is usually asymptomatic and is seen as an incidental finding in a neonate with other severe congenital anomalies. Intralobar by a ratio extralobar one third
sequestration is more common than the extralobar type, of 3 to 1. Both forms are found in the lower lobes, but sequestration is predominantly left sided (90%), whereas of intralobar sequestrations are right sided. A major
differentiating feature between the two types is the arterial supply to and venous drainage from the sequestered lung. An intralobar sequestration is supplied by a single large artery that arises from the infradiaphragmatic aorta and enters the sequestered lung via the pulmonary ligament. The venous drainage is via the pulmonary veins. In contrast, an extralobar sequestration receives several small branches from systemic and occasionally pulmonary arteries, with venous drainage into the systemic venous system (inferior vena cava, azygos, or hemiazygos veins).
1379
FIGURE
19.30. Neonatal
Lobar
Hyperinflation.
A. Frontal
chest radiograph in a 1-year-old boy shows a hyperlucent left upper lobe producing contralateral mediastinal shift. B . CT scan confirms the presence of an overexpanded and hyperlucent left upper lobe, representing congenital lobar emphysema.
TABLE 19.9 Bronchial Atresia Versus Neonatal Lobar Hyperinflation
Diagnostic Variable
Bronchial
Atresia
Neonatal Lobar Hyperinflation
Age at presentation
Teens/young adults
Neonatal
Symptoms
Asymptomatic
Respiratory distress
Location
LUL >RUL >RML
LUL >RML >RUL
1380
period
Radiographic/C T findings
Hyperlucent segment with mucocele
Hyperlucent lobe Diaphragmatic depression Mediastinal displacement
Treatment
None
Resection
LUL, left upper lobe; RUL, right upper lobe; RML, right middle lobe.
Sequestration appears as a solid posterior mediastinal mass or as a solitary or multicystic air collection (1 3). Air–fluid levels are seen when infection has produced communication of the sequestered lung with the normal tracheobronchial tree. The definitive diagnosis is made by the demonstration of abnormal systemic arterial supply to the abnormal lung, which is usually accomplished by thoracic aortography, contrast-enhanced MDCT (Fig. 19.31), US, or coronal MR and MR angiography. Arteriography is usually reserved for preoperative patients in whom precise demonstration of the origin and number of the systemic feeders is necessary. Hypoplastic
lung is a developmental anomaly resulting in a small
lung. It occurs secondary to congenital pulmonary arterial deficiency or following compression of the developing lung in utero from a variety of causes. Grossly, the lung is small, with a decrease in the number and size of airways, alveoli, and pulmonary arteries. Radiographically, the small lung and hemithorax are associated with ipsilateral diaphragmatic elevation and mediastinal shift, with herniation of the hyperinflated contralateral lung anteriorly toward the affected side. Hypoplastic lung can simulate total lung collapse radiographically but can usually be distinguished on clinical grounds and review of prior radiographic studies that show a small lung without evidence of pleural or parenchymal scarring.
1381
TABLE
19.10
Pulmonary
Sequestration
Diagnostic
Intralobar
Extralobar
Variable
Sequestration
Sequestration
Frequency
(of
Common
(75%)
Uncommon
(25%)
all sequestrations)
Age at presentation
Young adult
Neonate/infant
Mode of presentation
Recurrent pneumonia
Asymptomatic
Location
Left lower lobe: 60%
Left lower lobe: 90%
Right lower lobe: 40%
Right lower lobe: 10%
Pleural
Within
Separate
covering
pleura
layer
Associated
Rare
Common
visceral
congenital anomalies
pleural
(diaphragmatic eventration/hernia)
Radiographic appearance
Cystic lung mass with or without air–fluid levels
Solid peridiaphragmatic mass
Arterial
Single vessel from peridiaphragmatic
Multiple small systemic/pulmonary
supply
1382
Venous drainage
aorta
arteries
Pulmonary (leftto-left shunt)
Systemic (left-toright shunt)
P.555 Hypogenetic lung-scimitar syndrome, a variant of the hypoplastic lung, is characterized by an underdeveloped right lung with abnormal venous drainage of the lung to the inferior vena cava just above or below the right hemidiaphragm. The systemic venous drainage of the lung produces an extracardiac left-to-right shunt. The anomalous vein, which drains all or most of the right lung, may be seen as a vertically oriented curvilinear density shaped like a scimitar in the medial right lower lung, thereby giving this syndrome its common name of scimitar syndrome. The anomalies of venous drainage and lobar
bronchial
anatomy
(usually
bilateral
left-sided
[hyparterial]
bronchial branching) have given rise to the term congenital pulmonary venolobar syndrome. The right pulmonary artery is invariably hypoplastic, with supply to all or part of the lung (usually the lower lobe) from the systemic circulation. Associated anomalies include eventration of the right hemidiaphragm, horseshoe lung (congenital fusion of the right and left lungs posteroinferiorly), and cardiac anomalies such as atrial septal defect (most common), coarctation of the aorta, patent ductus arteriosus, and tetralogy of Fallot. The frontal chest radiographic P.556 findings are diagnostic and include a small right hemithorax with diaphragmatic elevation or eventration, dextroposition of the heart, and herniation of left lung anteriorly into the right hemithorax (Fig. 19.32). The classic appearance of a solitary scimitar vein is seen in only one third of cases, with the remainder having multiple small draining veins. Although plain film findings are usually diagnostic, CT or MR shows the abnormal draining vein and associated abnormalities. Most patients are asymptomatic, but some may present with recurrent infection or symptoms related to a left-to-
1383
right shunt or the associated cardiac anomalies.
FIGURE 19.31. Intralobar Pulmonary Sequestration. Contrast-enhanced CT scan (A) in a 34-year-old pregnant woman with recurrent right lower lobe pneumonia shows focal consolidation with a feeding artery visible medially (arrow) . Shaded-surface reconstruction (B) from CT angiogram shows feeding artery arising from celiac axis (arrowhead) to supply the right lower lobe.
1384
FIGURE
19.32. Congenital
Pulmonary
Venolobar
(Scimitar)
Syndrome. Frontal (A) and lateral (B) chest radiographs in a patient with Scimitar syndrome show a small right lung with rightward cardiomediastinal shift and a characteristic draining vein (arrow in A). The lateral film shows the interface of the hypoplastic right lung with the anteriorly situated heart and mediastinal fat (arrowheads) that have shifted as a result of the hypoplasia.
Arteriovenous Pulmonary
arteriovenous
Malformation malformations
(AVMs)
are
abnormal
vascular masses in which a focal collection of congenitally weakened capillaries dilates to become a tortuous complex of vessels fed by a single pulmonary artery and drained by a single pulmonary vein. Most pulmonary AVMs do not come to attention until early adulthood. They are detected either incidentally, as part of a screening evaluation in patients with hereditary hemorrhagic telangiectasia (a condition that is present in approximately 80% of all patients with pulmonary AVMs), or because of a variety of symptoms. The most common pulmonary symptoms are hemoptysis and dyspnea, the
1385
latter attributable to hypoxia caused by the intrapulmonary right-toleft shunt. Nonpulmonary symptoms most often relate to CNS disease. Stroke may occur from paradoxical right-to-left cerebral emboli or from thrombosis resulting from secondary polycythemia caused by chronic hypoxemia. Brain abscess may develop from paradoxical septic emboli. The chest radiograph of a pulmonary AVM usually shows a solitary pulmonary nodule, most often located in the subpleural portions of the lower lobes. Approximately one third of patients have multiple lesions. The lesion is often lobulated and has feeding and draining vessels emanating from the mass and extending toward the hilum. The morphology of the lesions is best demonstrated on MDCT with reconstructions. The feeding and draining vessels can be demonstrated by CT or MR. Angiography is reserved for preoperative evaluation and for patients undergoing therapeutic transcatheter embolization with spring coils or detachable occlusion balloons, which is the treatment of choice for patients with multiple AVMs.
TRAUMATIC Pulmonary
LUNG
DISEASE
contusion usually follows blunt chest trauma and
typically develops adjacent to the site of impact. Blood and edema fluid fill the alveoli of the lung within the first 12 hours after trauma, producing scattered areas of airspace opacification that may rapidly become confluent and may be difficult to distinguish from aspiration pneumonia (Fig. 19.33). Patients may have shortness of breath and hemoptysis; blood can usually be suctioned from the P.557 endotracheal tube. The typical radiographic course is stabilization of opacities by 24 hours and improvement within 2 to 7 days. Progressive opacities seen more than 48 hours after trauma should raise the suspicion of aspiration pneumonia or developing ARDS.
1386
FIGURE 19.33. Pulmonary Contusion/Traumatic Lung Cysts. CT in a 26-year-old man who sustained severe blunt left chest trauma shows extensive contusion in the left lung, with small traumatic lung cysts (arrowheads) and minimal superior segment right lower lobe involvement.
Pulmonary Cyst, and
Laceration, Traumatic Pulmonary Hematoma
Lung
Pulmonary laceration is a common sequela of penetrating or blunt chest trauma. In the latter situation, it represents a shearing injury to the substance of the lung. The elastic properties of the lung quickly transform the linear laceration into a rounded air cyst. These cysts may be filled with varied amounts of blood as a result of laceration of pulmonary capillaries; those that are completely filled with blood are more appropriately termed pulmonary hematomas. On radiographs and CT, these cysts appear as rounded lucencies that may contain air or an air–fluid level (Fig. 19.33) (1 4). Initially, these cysts are often obscured by the adjacent contused lung, only to be recognized after resorption of the blood. The cysts tend to shrink gradually over a period of weeks to months. The term traumatic
1387
air
cysts rather than pneumatoceles should be used for these lesions; the latter term is reserved for air cysts that result from a checkvalve overdistention of the distal lung, as seen in staphylococcal pneumonia.
ASPIRATION Aspiration pneumonia and pneumonitis are terms used to describe the different pulmonary inflammatory responses to aspirated material. As was discussed in the chapter on infection, aspiration pneumonia describes a mixed anaerobic infection resulting from the aspiration of infected oropharyngeal contents. The aspiration of oropharyngeal or gastric secretions may also occur in a “pure― form uncomplicated by anaerobic infection, producing aspiration pneumonitis. Aspiration of oropharyngeal or gastric secretions, with or without food particles, is not an uncommon event. It is seen in debilitated patients with chronic diseases, in patients with tracheal or gastric tubes, in unconscious patients, and in those who have suffered strokes, seizures, or trauma. More chronic and less easily recognizable forms of aspiration may occur in patients with anatomic abnormalities of the upper GI tract (Zenker diverticulum, esophageal stricture) or functional disorders (gastroesophageal reflux, neuromuscular
dysfunction).
Gastric fluid is highly irritating to the lungs and often stimulates explosive coughing and associated deep inspirations, leading to widespread distribution of the fluid throughout both lungs and into the peripheral airspaces. The hydrochloric acid contained in gastric fluid causes direct damage to both the bronchiolar lining and the alveolar wall. The severity of the resultant pneumonitis depends upon several factors: it is increased with a pH of the aspirated fluid 5 MHz), handheld linear array transducer is most commonly used. A targeted evaluation of the mammographically visible abnormality is performed. Simple cysts are easily distinguishable from complex or solid masses. This differentiation is extremely important, because simple cysts are always benign and require no further workup, whereas noncystic masses may represent cancers.
ANALYZING
THE
MAMMOGRAM
Masses Complete assessment of a mammographically visible, potentially malignant mass requires several steps. First, the radiologist must decide whether the mass is real. The left and right breasts must be compared in each view. Most women have reasonably symmetric parenchyma; however, at least 3% of women have areas of asymmetric but histologically normal breast tissue. When attempting to distinguish asymmetric normal breast tissue from a true abnormality, the radiologist must look for the mammographic features of a mass. Masses have convex borders P.573 and become denser toward the center. They distort the normal breast architecture. True masses are seen in multiple projections and can still be visualized when focal compression is utilized (Fig. 20.5) . Asymmetric breast parenchyma has an amorphous quality. On spot compression, the tissue spreads apart and fat can be seen interspersed with the denser breast structures in a pattern of normal architecture (Fig. 20.6). The appearance of asymmetric tissue varies significantly from one mammographic projection to another. When evaluating the breast for a possible mass, it is important to correlate the mammographic findings with the physical examination. When a suspicious palpable abnormality corresponds to an area of asymmetry seen on mammography, a biopsy should be undertaken.
1423
In a study of 221 patients with mammographically visible asymmetries, only three patients had malignancies, and all three had suspicious, palpable abnormalities that corresponded to the visualized
asymmetries
(2 1) .
Summation shadows that resemble masses on mammography can be produced by overlapping breast tissue. They are visible in only one view and usually disappear when focal compression spreads the tissues apart. Once the radiologist has concluded that a mass is present, its margins, density, location, and size should be assessed. The number of mammographically visible masses and their similarities or differences should be analyzed. Previous films should be compared with the current study to look for new masses or an increase in the size of a mass. It is impossible to evaluate one characteristic independent of the others.
FIGURE 20.5. Infiltrating Duct Carcinoma. A. CC views of both breasts, showing an asymmetric area of increased density in
1424
the outer aspect of the right breast (arrows). B . Magnification compression view shows this to be a true mass with defined, convex borders and increasing density toward its center.
Margins The margins of a mass are probably the most important characteristics to be assessed. Overlying breast parenchyma often obscures margin analysis, but liberal use of magnification compression views, in multiple projections, will aid the radiologist.
Spiculated Breast
Margins
carcinoma classically appears as a spiculated mass on
mammography (Fig. 20.7); however, fewer than 20% of nonpalpable cancers present as such (1 8). Most spiculated-appearing breast cancers will be infiltrating ductal carcinoma; however, tubular and lobular carcinomas can present as such. Tubular carcinomas are more well-differentiated histologically and carry a better prognosis. Lobular carcinomas comprise about 10% of all P.574 invasive
carcinomas.
They
are
not
mammographically
distinguishable
from invasive ductal carcinomas, although they are frequently more subtle. Single rows of lobular cancer cells can infiltrate surrounding tissues, so they generally cause less tissue distortion.
1425
FIGURE
20.6. Asymmetric
Breast
Parenchyma.
A. CC views
of both breasts in an asymptomatic woman. An area of asymmetric density is seen in the outer aspect of the right breast (arrows). B . Compression magnification view demonstrates normal breast architecture in the area of increased density. These findings are consistent with histologically normal but asymmetric mammary parenchyma.
1426
FIGURE
20.7. Classic
Breast
Carcinoma. This spiculated
breast mass is an infiltrating duct carcinoma.
A very limited differential exists for a spiculated mass. Fat necrosis from a previous surgical biopsy can appear spiculated (Fig. 20.8) . Scars from previous breast surgery should be marked carefully with radiopaque wires. Comparison should be made with previous films, both to determine the location of the abnormality that underwent biopsy and to assess for any increase in size of the presumed scar. Many scars will regress with time, but others will be stable in appearance and size. Any increase in size should be viewed with suspicion and biopsy should be undertaken. Radial scars or complex sclerosing lesions can also present as spiculated lesions. These are spontaneous lesions that are benign and consist histologically of central sclerosis and varying degrees of epithelial proliferation, represented by strands of fibrous connective tissue. Histologic differentiation of these lesions from carcinoma is mandatory.
1427
Indistinct
(Ill-Defined)
Margins
Breast carcinoma can also present as a round mass with indistinct or ill-defined borders (Fig. 20.9). Benign lesions that can present as such include abscess, hematoma, and focal fibrosis.
FIGURE
20.8. Postsurgical
Fat
Necrosis. This spiculated mass
had been stable for 7 years. The radiopaque wire indicates the scar on the patient's skin from the previous lumpectomy.
P.575
1428
Breast abscesses are most commonly seen in a subareolar location in lactating women (Fig. 20.10). Clinically, there is associated pain, swelling, and erythema. Spontaneous hematomas are seen in women on anticoagulant therapy or in those with blood dyscrasias. They can, of course, also be secondary to trauma, needle aspiration, or surgery. Correlation with the patient's history and physical examination will be helpful in discerning whether a lesion represents a hematoma. If doubt persists as to the nature of a possible hematoma, short-interval follow-up mammograms (4 to 6 weeks later) to demonstrate resolution will be helpful (Fig. 20.11) .
Circumscribed
(Well-Defined)
Margins
Circumscribed masses are almost always benign; however, up to 5% of masses that appear well circumscribed on conventional mammograms may represent carcinomas (2 2). The “halo sign,― which is a partial or complete radiolucent ring surrounding a mass, is not helpful in determining benignity. Sonography should be used to assess circumscribed masses prior to any additional mammographic views; if a simple cyst is diagnosed by US, no further imaging workup is required. Magnification compression views will be of great assistance in clarifying the nature of borders of an apparently well-circumscribed solid mass. Masses that appear well circumscribed on conventional views may have indistinct or microlobulated margins on compression magnification views (2 3); such masses should undergo biopsy. If a solid mass appears circumscribed on magnification views and no previous mammograms are available for comparison, the mass can generally be characterized as one that has a high probability of being benign. Such masses are frequently subjected to a course of follow-up mammography. The first of these surveillance mammograms should be performed 6 months following the original study.
1429
FIGURE 20.9. Infiltrating Duct Carcinoma. Lesion presenting as a round mass with indistinct, microlobulated borders.
Cysts are the most common well-circumscribed masses seen in women between the ages of 35 and 50 (Fig. 20.12). They are rare after menopause unless hormone replacement therapy has been instituted. Cysts can be accurately diagnosed by US and are virtually never malignant. A high-frequency (generally 5 to 10 MHz) US transducer is utilized in a targeted examination of the mass in question. On sonography, cysts are round or oval, smooth-walled, anechoic, and produce enhanced through transmission of sound. They can frequently be deformed with gentle
1430
P.576 pressure from the transducer. It is essential that the focal zone and gain of the US unit be optimally adjusted for the lesion so that cysts can be accurately diagnosed sonographically. The cyst must be thoroughly examined in two projections to rule out any irregularities or masses emanating from the walls.
FIGURE 20.10. Large Subareolar Abscess. The indistinct borders of the mass are the result of surrounding inflammation.
1431
Fibrosis is another manifestation of fibrocystic change that can be seen mammographically. It can be quite focal, giving it the appearance of a well-defined mass on the films. Such areas of focal fibrosis may also present with ill-defined borders, making them difficult to differentiate from carcinomas. Fibroadenomas are the most common well-defined solid masses seen on mammography (Fig. 20.13). They are homogenous but frequently show large, coarse calcifications. They may have a lobulated contour, but there are usually only a few large lobulations. If a fibroadenoma is not calcified, it cannot be distinguished from a cyst by mammography. Sonography will allow characterization of fibroadenomas as solid hypoechoic masses. The peak age of patients with clinically detected fibroadenomas is 20 to 30 years; however, fibroadenomas are seen into the eighth decade. They rarely appear or grow after menopause.
FIGURE
20.11. Infiltrating
Duct
1432
Carcinoma.
Magnification
view of a palpable abnormality in the upper outer quadrant. The patient had undergone a negative fine-needle aspiration biopsy the previous day; the mammographic differential diagnosis included hematoma and carcinoma. Follow-up mammogram 6 weeks later demonstrated no resolution. Surgical biopsy showed infiltrating duct carcinoma.
Primary
breast
malignancies to be considered when a well-defined
density is visualized on mammography are infiltrating duct carcinoma, papillary carcinoma, mucinous carcinoma, and medullary carcinoma. Lymphoma, either primary or metastatic, may also present as a well-circumscribed mass. Metastatic
disease to the breast from other sources may present as
a well-circumscribed nodule. The most common primary cancer to produce breast metastases is melanoma, but a large variety of other primary sites have also been reported to metastasize to the breast. When these malignancies are encountered, magnification compression views of the abnormality often demonstrate some irregularity to the contour of the mass (Fig. 20.14) .
Density Density is relevant to analysis of mammographically detected masses when these masses contain lucent areas indicative of fat. Breast masses that clearly contain fat are benign. The assessment of density in homogeneous nonfatty masses is not, however, useful in the prediction of benignity or malignancy.
1433
FIGURE
20.12. Simple Breast Cyst. A. CC mammogram
demonstrates a 1.5-cm mass in a 50-year-old-woman (arrows) . The mass is at least partially well circumscribed. B . US of the mass demonstrates a round, anechoic structure with well-defined margins that was enhanced through transmission of sound. These features are diagnostic of a simple cyst.
P.577
Fat
Density
Benign breast lesions that are purely fat density include oil cysts from fat necrosis, lipomas, and sometimes galactoceles. Oil cysts are generally the result of trauma (Fig. 20.15). They are round lucent lesions surrounded by a thin capsule; often they are multiple and can
1434
demonstrate rim calcifications. Lipomas are similar to oil cysts in appearance; they are also lucent with a surrounding capsule. The surrounding breast architecture may be distorted because of the mass effect of the lipoma. Galactoceles usually occur in lactating or recently lactating women and are probably the result of an obstructed duct. If the inspissated milk is of sufficient fat quantity, these lesions will appear lucent; however, they can also be of mixed or water density.
FIGURE 20.13. Fibroadenoma. A. MLO view of a 1.8-cm, partially well-defined mass (arrow). B . US demonstrates a solid hypoechoic mass with a macrolobulated, well-defined margin.
1435
1436
FIGURE 20.14. Infiltrating Duct Carcinoma. A. A wellcircumscribed, 8-mm mass that had enlarged compared with a study done 1 year previously. B . Magnification view shows a spiculation anteriorly proven at biopsy.
(arrow). Infiltrating duct carcinoma was
P.578
Mixed Fat and Water Density Other benign masses that are mixed fat and water density are hamartomas, which are rare benign tumors, and intramammary lymph nodes. The latter are frequently seen on mammograms. They are generally located in the upper outer quadrant in the posterior three fourths of the breast parenchyma. They normally contain a fatty center or a lucent notch, representing fat in the hilus of the node (Fig. 20.16). Fat–fluid levels can occasionally be seen on MLO mammograms in galactoceles and postsurgical hematomas.
1437
FIGURE 20.15. Oil Cysts. Multiple lucent masses with thin capsules (arrows) are characteristic of oil cysts. The patient had suffered trauma to the breast.
Location Breast cancers can occur in any location within the breast. As such, the location of a lesion is helpful in mammographic diagnosis in only two situations. The first occurs when the mammographer is considering an intramammary lymph node in the differential. The second occurs when a lesion can be localized to the skin.
1438
Intramammary nodes visualized on mammograms are almost always located in the upper outer quadrant of the breast. They have been noted in other locations in autopsy series, and there are rare case reports of visualization of such nodes by mammography in other locations in the breast.
Skin
Lesions
If a lesion is located only on the skin, it does not represent a breast carcinoma. Frequently, however, skin lesions project over the parenchyma and can P.579 appear to be within the breast. Such lesions are usually recognizable by air trapping around the edges or in the interstices. This air trapping can produce a dark halo around one edge (Fig. 20.17). Air trapping will not, however, be evident with flat, pigmented skin lesions or sebaceous cysts.
FIGURE
20.16. Intramammary
Lymph
1439
Node.
Intramammary
lymph node displays a characteristic lucent center (arrow) and well-circumscribed margins. The node was located in the upper outer quadrant.
It is helpful to examine the patient and place a radiopaque marker on any skin lesions or possible sebaceous cysts. The technologist can then perform a repeat film in the projection that the lesion was visualized. If necessary, this view can be followed by a tangential view to demonstrate that the lesion is located in the skin.
FIGURE 20.17. Skin Nevus. The dark halo produced around one edge is the result of air trapping (arrows) .
1440
Size By itself, the size of a mammographically discovered mass is not particularly helpful in determining its etiology. A spiculated or illdefined mass should undergo biopsy no matter what its size. However, when the mammographer is dealing with a circumscribed mass that has a much lower chance of being malignant, size may play a role in determining the next step in the workup. US is not usually helpful when lesions are less than about 3 to 5 mm in size, particularly in fatty breasts. Frequently, patients with such lesions will be asked to return in 6 months for a follow-up study to assess for interval growth. If the lesion increases in size, further investigation with US and possible biopsy can be performed. After the first 6-month follow-up, stable lesions should be followed at yearly intervals for a minimum of 3 years. Larger, clinically occult masses require both US—to prove they are solid—and magnification views to prove they are circumscribed before surveillance mammography is suggested. Some experts advocate an upper size limit of 1 to 1.5 cm for masses that are to undergo follow-up, but recent research has shown that nonpalpable, circumscribed breast masses can be managed by periodic mammographic surveillance regardless of size (2 4). Generally, a 6month follow-up of the affected breast is advocated; this is followed by a bilateral mammogram 6 months later and then annual mammography for at least 3 years to document stability. P.580
Number Multiple
Of
Masses
Masses
In many cases, multiple well-defined round masses will be seen on mammography. When evident, such masses are also frequently bilateral. Multiple, bilateral round masses are usually benign. They most often represent cysts or fibroadenomas, although multiple papillomas can also present in this way (Fig. 20.18). In patients with a history of previous malignancy, however, metastasis may also be
1441
considered, although metastatic disease is much more commonly unifocal. All lesions should be evaluated carefully. Benign and malignant lesions can coexist in the same breast. A lesion with a different, suspicious morphology should prompt a biopsy. When evaluating the patient with similar appearing multiple, bilateral, rounded breast masses, it is not generally advisable to utilize US. US is confusing and frequently demonstrates hypoechoic areas that, although disconcerting to the radiologist, do not prove to be malignant. Multifocal primary breast cancers generally present as obvious ill-defined or stellate lesions that are suspicious in appearance (Fig. 20.19) .
Calcifications Clustered pleomorphic microcalcifications, with or without an associated soft tissue mass, are a primary mammographic sign of breast cancer. Such calcifications are seen in more than half of all mammographically discovered cancers; about one third of all nonpalpable cancers are manifest by calcifications alone, without an associated mass (1 9) .
1442
FIGURE 20.18. Multiple Benign Masses. Bilateral CC views show multiple large round masses in both breasts. The patient was asymptomatic. Differential diagnosis was cysts or fibroadenomas.
The calcifications associated with malignancy are dystrophic; they are the result of abnormalities in the tissues. Some malignant calcifications occur in necrotic tumor debris; others are the result of calcification of stagnant secretions that are trapped in the cancer (2 5) . Calcifications are a frequent finding on mammographic examinations. In the majority of cases, such calcifications will be benign and their origin, as such, will be easily identifiable. There is, however, a significant overlap in the appearance of benign and malignant calcifications. Only 25% to 35% of all calcifications that undergo biopsy will be malignant. The importance of technically optimal mammography cannot be overstated when calcifications are being studied. The film exposure must be appropriate; an underexposed film can hide calcifications in
1443
a background of white breast tissue. Slight overpenetration of films is optimal for detection of calcifications. Magnification views are extremely helpful for assessing the malignant potential of a group of calcifications. Careful analysis of the form, size, distribution, and number of calcifications, as well as any association with other soft tissue structures, will allow the radiologist to determine which calcifications are unequivocally benign and which require biopsy or follow-up studies.
Form Benign
Calcifications
Some shapes of calcifications can be easily identified as benign. Any calcification with a lucent center should not cause concern. Calcifications with lucent centers are often located in the skin. A skin marker can be placed over the calcifications and a subsequent tangential view taken to confirm their location in the skin (Fig. 20.20). Calcifications with lucent centers are also seen as a result of fat necrosis. Such calcifications can be smooth and round, or they can be eggshell-type calcifications in the walls of an oil cyst (Fig. 20.21) .
1444
FIGURE
20.19. Multifocal
Carcinoma. CC view. The largest
mass was palpable. The others were discovered by mammography (straight arrows). The more well-defined nodule (curved arrow) probably represented an intramammary lymph node.
P.581 Calcifications that layer into a curvilinear or linear shape on 90° lateral films, yet appear as smudged clusters on CC views, are also representative of a benign process (Fig. 20.22). Such calcifications represent sedimented calcium (“milk of calcium―) within the fluid of tiny breast cysts. Similar benign calcifications can also be seen within larger cysts and oil cysts. Sedimented calcium is a common finding in approximately 5% of women presenting for mammography.
1445
Other benign calcifications that are easily recognizable by their form include arterial calcifications, the calcifications in a degenerating fibroadenoma, and calcifications associated with secretory disease. Arterial calcifications generally present as tubular parallel lines of calcium (Fig. 20.23). Occasionally, early arterial calcification can present a diagnostic problem, but this can usually be resolved by looking for soft tissue of the vessel in association with the calcification. Magnification in multiple projections can be helpful (Fig. 20.24) . Fibroadenomas can calcify in various patterns. Sometimes the calcifications are indeterminate, but the classic calcifications associated with an atrophic fibroadenoma are large, coarse, and irregular in shape (Fig. 20.25) .
Secretory
Disease
The calcifications associated with secretory disease are smooth, long, thick linear calcifications that radiate toward the nipple in a generally orderly pattern (Fig. 20.26). These calcifications are located in ectatic ducts. When periductal inflammation has occurred, these calcifications may appear more lucent centrally because calcium is deposited in the tissues adjacent to the ducts.
1446
FIGURE 20.20. Skin Calcifications. Tangential view showing calcifications in the skin. A radiopaque marker had been placed on the skin at the site of the calcifications. This was done to facilitate positioning for the tangential view.
Malignant calcifications vary in shape and size (Fig. 20.27). The margins of the calcifications are jagged and irregular. Malignant calcifications are often branching. Ductal carcinoma in situ (DCIS), or noninvasive breast cancer, is most often detected mammographically as a result of such calcifications. Groups of pleomorphic calcifications that are more linear or “dot-dash― in appearance are more
1447
commonly associated with high-nuclear-grade intraductal carcinomas that have luminal necrosis (comedocarcinomas) (Fig. 20.28). The lower-grade (cribriform and micropapillary) types are often manifest by more punctate or granular appearing calcifications. The morphology of the calcification cannot, however, be used to predict the subtype of DCIS because there is P.582 considerable overlap in the forms of the calcification associated with each subtype; frequently, multiple DCIS subtypes exist together in the same lesion. In the high-grade (comedo) subtype, the calcifications can be an approximate indication of the size of the tumor, although the extent of disease is often greater than mammographically predicted. In the lower-grade varieties, correlation is even poorer. The biologic behavior of these subtypes also differs; high-grade types are the most likely to recur (2 6) .
FIGURE 20.21. Eggshell Calcifications in Oil Cysts. These are large calcifications with lucent centers and are benign.
Pleomorphic microcalcifications in association with a malignant soft tissue mass can also indicate areas of extensive intraductal
1448
component within or adjacent to the invasive tumor. It is especially important to recognize malignant calcifications that occur in tissues surrounding invasive cancers so they can be excised with the invasive tumor. Such extensive intraductal component–positive cancers also have a greater tendency to recur.
FIGURE
20.22. Milk of Calcium in Breast Cysts. A.
Magnification of a 90° lateral mammogram showing diffuse linear calcifications (arrows). B . CC magnification view of the same area showing smudged, rounded calcifications (arrows) . This change in configuration between views is typical of sedimented calcium. The calcium is layering in the bottom of microcysts, so it appears as a line or meniscus when viewed from the side in the lateral projection. When viewed from the top, these calcifications simply appear smudged and rounded.
Indeterminate Morphologically
Calcifications
indeterminate
calcifications
account
for
the
majority
of mammographically generated biopsies of calcifications (Fig. 20.29). Such calcifications are most often associated with fibrocystic change. Diagnoses included under the general category of fibrocystic disease are fibrosis, adenosis, sclerosing adenosis, epithelial
1449
hyperplasia, cysts, apocrine metaplasia, and atypical hyperplasia. Occasionally, biopsy of indeterminate calcification will yield a diagnosis of lobular carcinoma in situ (LCIS), also called lobular neoplasia. Although it is not an invasive cancer, LCIS places a woman at higher risk for development of invasive breast cancer. Mammographically, LCIS has no distinct features. If it is clinically occult, it is most often found serendipitously, P.583 adjacent to a focus of mammographically indeterminate, but histologically benign, calcifications.
FIGURE
20.23. Arterial
Calcifications. Arterial calcifications in
the breast are identified by their location in the wall of a tortuous vessel.
Distribution Calcifications that are widely scattered and seen bilaterally are usually indicative of a benign process, such as sclerosing adenosis or adenosis. Multiple, bilateral clusters of calcifications that appear
1450
morphologically similar are also generally benign. Careful analysis with a magnifying lens is essential in these cases so that a morphologically dissimilar cluster is not overlooked. Such calcifications should be thoroughly examined with magnification views.
FIGURE
20.24. Early
Arterial
Calcification.
Magnification
view. The calcification can be seen clearly in the walls of an artery (arrows). The soft tissue of the artery was difficult to appreciate on the conventional views.
1451
FIGURE 20.25. Fibroadenoma. Typical large, coarse, irregular calcifications are seen in a fibroadenoma.
Malignant calcifications usually occur in tight clusters within a small volume of tissue, but DCIS can produce calcifications that encompass large areas of the breast. Calcifications that are morphologically suspicious or indeterminate and occupy a segment of the breast should
undergo
biopsy.
Size Malignant calcifications are generally smaller than 0.5 mm. Because the calcifications associated with carcinoma are so small, they are frequently referred to as microcalcifications. Within a cluster, there will be a variety of sizes. Benign calcifications are often larger. When benign P.584 disease produces clusters of calcifications, the size of these calcifications is usually similar.
1452
FIGURE
20.26. Secretory
Calcifications. Craniocaudal view
demonstrates long and thick calcifications in ectatic ducts that radiate toward the nipple.
Number Calcifications associated with malignancy are generally quite numerous. The greater the number of calcifications, the more likely they are associated with malignant disease. Establishing the lower limit of the number of calcifications in a cluster that would require biopsy is extremely difficult. Assessment of the morphology of these calcifications by magnification views will influence this decision more than the actual number of calcifications.
1453
FIGURE
20.27. Malignant
Calcifications. Magnification view of
infiltrating ductal carcinoma. Note the irregular forms as well as the variety of sizes and shapes.
1454
FIGURE 20.28. Malignant Calcifications. Dot-dash or “casting’’ calcifications of the comedo subtype of ductal carcinoma in situ. Note the pleomorphism in the size and shape of the calcifications. (From Kline TS, Kline IK. Breast. New York: Igaku-Shoin, 1989:201, Guides to Clinical Aspiration Biopsy Series; used with permission.)
Architectural
Distortion
Breast cancer is occasionally heralded by distortion in the normal architecture of the breast (Fig. 20.30). Differential diagnosis includes fat necrosis related to scarring from previous surgery and a complex sclerosing lesion, also known as radial scar. On close inspection, fat may be seen interspersed with fibrous elements in the center of fat necrosis or complex sclerosing lesions, but this appearance is not specific for benignity. Similar findings can be seen in malignant lesions. Biopsy is necessary for differentiation.
Increased Hormone
Density
of
Breast
Tissue
Therapy
Increasing parenchymal density of breast tissue can be bilateral or unilateral. Bilateral increased density is usually the result of estrogen replacement therapy in postmenopausal women. Such hormone therapy can give the breasts a more glandular, premenopausal appearance. Intrinsic hormonal fluctuations in premenopausal, pregnant, or lactating women may cause similar changes in the density of the breasts. Hormonally related changes in breast density are not associated with skin thickening.
Inflammatory
Carcinoma
A unilateral increase in breast density with associated skin thickening may be caused by several processes. The most ominous of these P.585
1455
is inflammatory carcinoma of the breast (Fig. 20.31). Clinically, this disease is manifest by a warm, erythematous, firm, tender breast. Histologically, the dermal lymphatics are diffusely involved. Mammographically, a focal mass may be seen within the dense tissue, but often the breast appears homogeneously dense. Inflammatory carcinoma of the breast is a locally advanced disease that carries a poor prognosis.
FIGURE 20.29. Indeterminate Calcifications. Magnification view of cluster of calcifications. There is some irregularity in shape and variation in size, but these calcifications were benign. They were associated with fibrocystic change.
Radiation
Therapy
A unilateral increase in parenchymal density with skin thickening can also be seen in patients who have undergone radiation therapy to the breast. Radiation changes are most pronounced during the first 6 months following therapy. They usually resolve gradually over a period of years.
1456
Diffuse mastitis can produce a generalized skin thickening and increase in breast density. Clinical differentiation from inflammatory carcinoma is usually possible.
FIGURE
20.30. Architectural
Distortion
Representing
Breast Carcinoma. Note how the cancer pulls the surrounding parenchyma toward it (arrows) .
Obstruction to the lymphatic or venous drainage from metastatic disease, surgical removal, or thrombosis can produce a unilateral increase in breast density, with skin thickening caused by edema. The anasarca associated with congestive heart failure, renal failure, cirrhosis, or hypoalbuminemia most often presents as bilateral increased breast density with skin thickening; however, asymmetric involvement of the breasts can occur. Correlation of physical examination findings and history will usually allow differentiation of the various causes of an increase in breast density.
Axillary
Adenopathy
1457
Axillary lymph nodes are frequently visualized on the MLO mammogram. Normally they are less than 2 cm in size and have lucent centers or notches resulting from fat in the P.586 hilum. Fatty infiltration of the nodes themselves can cause lucent enlargement
and
replacement.
FIGURE 20.31. Inflammatory Carcinoma. CC view demonstrates a diffuse increase in parenchymal density, along with skin thickening laterally (arrows) .
Pathologic axillary nodes are homogeneously dense and enlarged. A
1458
variety of processes can result in replacement of normal nodal architecture. Malignant involvement of axillary nodes can be the result of primary breast cancer, metastatic disease, lymphoma, or leukemia (Fig. 20.32). Axillary nodes can also become pathologically enlarged because of inflammation. Patients with rheumatoid arthritis, systemic lupus erythematosus, scleroderma, and psoriasis may also have
axillary
adenopathy.
Coarse calcifications in axillary nodes may reflect granulomatous disease. Microcalcifications are occasionally seen in nodes involved with metastatic breast cancer. Gold deposits, seen in patients being treated for rheumatoid arthritis, are occasionally seen in axillary nodes and may be confused with calcifications.
The
Augmented
Breast
More than 1.5 million women in the United States have undergone augmentation mammoplasty. Imaging of the augmented breast poses unique challenges. Special techniques must be employed both to screen for breast cancer and to evaluate the patient for possible complications related to the implant.
1459
FIGURE
20.32. Lymphoma. Hodgkin disease involves the
axillary lymph nodes. The nodes are homogeneous, dense, and enlarged (arrows) .
Various types of implants have been used in augmentation procedures. They include silicone envelopes filled with saline or with viscous silicone gel, as well as double-lumen implants containing an inner core of silicone gel surrounded by an outer envelope filled with saline. Silicone is more radiopaque than saline, although neither
1460
allows adequate visualization of immediately surrounding tissue. Implants can be placed either anterior (prepectoral) or posterior (subpectoral) to the pectoralis muscle. A fibrous capsule develops around the implant. Patients with prepectoral implants are subject to a greater risk of fibrous and calcific contractures around the implant. Such contractures are not only painful and deforming, but they also make mammography more difficult. Screening mammography in the woman with implants requires the use of at least two extra views of each breast. P.587 Standard MLO and CC views are performed with moderate compression. Then the implants are displaced posteriorly against the chest wall while the breast tissue is pulled anteriorly and compressed more vigorously (Fig. 20.33). The compression paddle keeps the implant from migrating into the field of view. Greater compression of anterior tissues allows more optimal imaging (Fig. 20.34). Both MLO and CC views are repeated using this technique. These modified views are called implant displacement views (2 7) . Implant displacement views are more difficult to accomplish in patients with prepectoral implants with associated capsular contractures around the implant. The implants are not easily displaced, so that less of the anterior breast tissue is depicted on the modified views. In such cases, a 90° lateral view may also be helpful in screening. Although some breast tissue may be obscured in patients with implants, these women, when in the appropriate age groups, deserve the same careful screening examinations at the same intervals as patients without implants. The indeterminate mammogram in an implant patient should be evaluated in a manner similar to that in a patient without implants.
1461
FIGURE
20.33. Breast Implants. A. Standard MLO view of a
patient with a subpectoral silicone implant. Note the pectoralis muscle (arrows) anterior to the implant. B . MLO implant displacement view on the same patient. The implant has been displaced posteriorly, out of view, while compression has been applied anteriorly.
1462
FIGURE
20.34. Infiltrating Duct Carcinoma. A. Standard MLO
view in a patient with prepectoral silicone implants. Note the pectoralis muscle (white arrows) extending posterior to the implant. A poorly defined 1-cm mass (black arrows) was noted in the subareolar tissues. B . MLO implant displacement view in the same patient. The subareolar mass (black arrow) is more clearly defined because of greater compression of the tissues anterior to the implant. Histologic examination of the mass showed infiltrating duct carcinoma.
Women who have undergone augmentation mammoplasty may also present with abnormalities related to their implants. These include
1463
capsular contractures, herniations of the implant through rents in the capsules, implant rupture with free (extracapsular rupture) or contained (intracapsular rupture) silicone, and deflation of saline implants. Many patients will present for breast imaging subsequent to noticing a change in implant contour or size (Fig. 20.35) . Mammography is generally the first examination performed if the woman is over the age of 30; however, mammography is not useful in the detection of intracapsular silicone implant ruptures because the silicone is contained within the fibrous capsule that has developed around the implant. Extracapsular silicone implant ruptures can sometimes be detected by mammography, but often the free silicone is obscured by the overlying implant P.588 or is in an area of the breast or chest wall not imaged on the mammogram (2 8) .
1464
FIGURE 20.35. Ruptured Implant. Standard MLO view of a patient with prepectoral silicone implants. The patient had noted a new mass superolaterally in her breast. The mammogram shows an extracapsular rupture, with silicone outside the implant capsule (arrows) that corresponded to the palpable abnormality.
Other imaging modalities can be used for the assessment of implant complications. MR is the most accurate in identifying silicone implant rupture and in localizing free silicone (2 9). The protocol for breast
1465
implant evaluation consists of axial, sagittal, and/or coronal T2W sequences with and without water suppression and inversion recovery (IR) sequences with water suppression. It is essential to use several projections in implant evaluation. The most effective sequence is the IR sequence, which suppresses the fat signal. The addition of water saturation results in a silicone-only image. T2W fast spin-echo sequences, without and with water suppression, are useful for saline implants. In intracapsular silicone implant rupture, the implant shell has ruptured but the silicone remains within the fibrous capsule. Signs of intracapsular rupture on MR can be subtle. A linguine sign indicating intracapsular rupture occurs when the collapsed implant shell floats within the silicone gel contained in the fibrous capsule (Fig. 20.36) . The noose, teardrop, or keyhole signs of intracapsular rupture indicate small amounts of silicone collected in a radial fold (Fig. 20.37). Over time, microscopic silicone can leak through the intact implant shell and collect at the implant shell surface, giving a subcapsular line sign. This can be difficult to differentiate from a small intracapsular rupture. In extracapsular rupture, the envelope and fibrous capsule lose integrity, resulting in the extrusion of free silicone gel into breast tissue (Fig. 20.38). US is also used to detect implant rupture, but it has a lower sensitivity (70%) than MR (94%) (3 0). The specificity of both US and MR is similar (92% to 97%). The success of US in the assessment of implant integrity is highly dependent on the operator; an experienced radiologist must scan the breasts in a methodical manner.
The Male Breast The most common indication for breast imaging in men is a palpable asymmetric thickening or mass. Gynecomastia is usually the cause. Breast cancer is rare, but can occur.
1466
FIGURE
20.36. MR of Intracapsular Silicone Implant
Rupture. Sagittal inversion-recovery T2WI with water suppression shows multiple low-intensity curvilinear lines (arrows) contained within the fibrous capsule, representing the collapsed implant shell (“linguine sign―). There is no extracapsular silicone.
1467
FIGURE 20.37. MR of Subtle Intracapsular Silicone Implant Rupture. Sagittal fast spin-echo T2WI shows a focus of silicone gel trapped within a fold of the implant shell (arrows), known as “noose sign,― “inverted teardrop sign,― or “keyhole sign.―
P.589 Normal male breast appears on mammography as a mound of subcutaneous fat without glandular tissue (Fig. 20.39). The nipple is small. Gynecomastia generally appears as a triangular or flame-shaped area of subareolar glandular tissue that points toward the nipple. Fat is interspersed with parenchymal elements. A gradual merging of the more glandular elements with the fat occurs at the deep margin (Fig. 20.40). Gynecomastia can be unilateral or bilateral. When bilateral, it is most frequently asymmetric. Many causes have been reported,
1468
including ingestion of a variety of drugs, such as reserpine, cardiac glycosides, cimetidine, and thiazides, as well as marijuana. Testicular, adrenal, and pituitary tumors are associated with gynecomastia. Chronic hepatic disease, by virtue of the body's reduced ability to clear endogenous estrogens, can also cause male breast enlargement.
FIGURE
20.38. MR of an Extracapsular Silicone Implant
Rupture. Sagittal (A) and coronal (B)
inversion-recovery
T2WIs
with water suppression show extracapsular silicone (white arrows) in the superior and lateral left breast. The partially collapsed implant shell (black arrow) is seen within the silicone gel contained within the fibrous capsule that surrounds the implant.
Male breast cancer is mammographically similar to that found in women. It can have a variety of appearances, including an illdefined, spiculated, or circumscribed mass (Fig. 20.41) . Microcalcifications
can
occur.
1469
Comparison
With
Previous
Films
The importance of comparing current mammograms with previous films cannot be overstated. In one series, developing densities accounted for 6% of nonpalpable breast carcinomas (1 9) . Comparison with previous films will allow detection of subtle changes, in turn suggesting the need for further evaluation of such areas at an earlier time than might be possible if no comparison had been made (Fig. 20.42). It must, of course, be remembered that benign masses may appear or enlarge over time. In fact, in the majority of cases, interval change will be benign, but such changes should be fully evaluated by correlation with the history and physical examination, as well as the use of ancillary testing methods such as US, aspiration, and biopsy.
1470
FIGURE 20.39. Male Breast. Relatively normal male breast, which is a mound of subcutaneous fat. Note the lack of glandular tissue.
1471
P.590 Malignant masses that were stable in size for up to 4.5 years have been reported. Although such a long period of stability is unusual, these reports emphasize the need for suspicious lesions to undergo biopsy regardless of their apparent lack of change in size on serial films. Such lesions may have been overlooked or misinterpreted on a previous study. Any new microcalcifications or increase in number of such calcifications deserve special consideration. Appropriate workup with magnification views will allow analysis of the morphology of such calcifications. Any calcifications that are not clearly benign deserve biopsy.
1472
FIGURE
20.40. Gynecomastia. MLO view of a man with breast
enlargement. Glandular tissue is seen in the subareolar area. This tissue gradually intersperses with the fat and does not appear as a mass.
1473
MAGNETIC
RESONANCE
IMAGING
Indications There is growing consensus on the indications for MR of the breast. As described in the section on imaging of implants, MR is the most accurate modality for evaluating breast implant integrity. In cases of newly diagnosed invasive breast cancer, MR is superior to mammography and clinical breast examination in defining the local extent of disease (3 1). MR has been reported to be more sensitive than mammography in detection of multifocal or multicentric invasive tumors and in the evaluation of P.591 residual
disease
postlumpectomy
(3 2,3 3,3 4). However, there is
controversy about whether the use of MR offers meaningful clinical benefit. It is possible that women who opt for mastectomy based on MR findings may have disease that could have been controlled with radiation therapy. Further clinical trials are needed to assess the clinical utility of breast MR for determining the local extent of disease.
1474
FIGURE
20.41. Male Breast Cancer. MLO view of the breast in
a male. The mass has a defined interface with the surrounding
1475
fat (arrows). (Courtesy of Patricia Bell, M.D., Auburn, California.)
In patients who have undergone breast conservation therapy, differentiating lumpectomy and radiation changes from local recurrence can be difficult. MR can be helpful in these cases (3 5). MR is also useful in diagnosing breast cancer when the patient presents with axillary metastases and no known primary tumor (3 6) . The role of breast MR in screening remains controversial. MR is more costly than mammography and has high false-positive rates, making it unacceptable for the general population; however, MR does appear to have a role in screening women at high risk for development of breast
cancer.
Technique Breast MR is usually performed in a standard 1.0- to 1.5-tesla magnet. A dedicated bilateral breast surface coil should be used. The patients are imaged prone to minimize respiratory motion. Ideally, imaging should be done between days 6 and 17 of the menstrual cycle. Bilateral studies should be performed. The breast should be imaged in axial or sagittal planes or a combination of the two. Core pulse sequences when evaluating the breast for cancer include a three-plane localizer, T1WIs, T2WIs, and two- or three-dimensional fat-suppressed gradient echo series performed before contrast administration, immediately after, and delayed. The number of postcontrast series can vary, but at least three are needed to perform kinetic enhancement curves. The T1WIs allow clear differentiation of adipose tissue from glandular tissue. T2W fat-suppressed images allow identification of fluid-filled structures such as cysts. Dynamic images obtained prior to and after IV gadolinium enhancement help to identify potential malignancies based on morphology and enhancement kinetics. The intravenous gadolinium DTPA dose ranges from 0.1 to 0.2 mmol/kg body weight. Fat suppression can be accomplished before gadolinium administration using chemical selective fat saturation or water-only
1476
excitation techniques. After IV contrast administration, passive fat suppression can be accomplished with postprocessing image subtraction, but patient movement between precontrast and postcontrast enhanced images can degrade the images because of misregistration. Kinetic curves can be performed on enhancing lesions.
Interpretation Each lesion should be evaluated for its shape, margin, internal architecture, precontrast T1 and T2 signal characteristic, enhancement characteristics, and change from prior studies. Predictors of benignity include smooth margins, nonenhancing internal septations, minimal or no enhancement, and diffuse patchy enhancement.
Features
suggestive
of
malignancy
include
spiculated
or irregular borders, peripheral or rim enhancement, regional enhancement, and ductal enhancement. On the precontrast T1WI, bright T1 signal intensity is suggestive of benign etiologies such as a complicated or hemorrhagic cyst, fresh fat necrosis, or the fatty hilum of an intramammary lymph node. Simple cysts have high T2 signal intensity, whereas most invasive carcinomas have low T2 signal intensity. Medullary or mucinous carcinoma can have high T2 signal intensity and look similar to cysts on MR.
1477
FIGURE 20.42. Infiltrating Duct Carcinoma. A. CC mammogram shows dense mammary parenchyma but no evidence of malignancy. B . Mammogram 1 year later shows development of a subtle new mass (arrows). C . US of the mass shows an irregular solid mass with indistinct margins. Biopsy
1478
demonstrated
infiltrating
duct
carcinoma.
P.592 Kinetic curves improve the specificity of breast MR. These curves can be evaluated qualitatively according to the curve shape and classified as a persistent pattern of enhancement, a plateau of enhancement, or washout of signal intensity (3 7). Most invasive carcinomas demonstrate rapid initial enhancement with a plateau or washout (Figs. 20.43, 20.44). Some malignant lesions, such as DCIS, invasive lobular carcinoma, tubular carcinoma, and mucinous carcinoma, may demonstrate slow enhancement. A curve showing progressive increase in signal intensity after the first 2 minutes is more suggestive of a benign etiology. Enhancement curves are helpful for lesions that are indeterminate or benign in morphology and may influence the decision to biopsy. Any morphologically suspicious lesion, however, requires biopsy regardless of its enhancement kinetics.
FIGURE
20.43. Breast MR Kinetic Curves. Schematic drawing
of kinetic curves showing hypothetical signal intensities of a lesion after contrast injection. The shape of the curve aids in
1479
differentiating benign from malignant lesions. Rapid enhancement in the early postcontrast phase is more often associated with malignant lesions. Washout of contrast in the immediate and late postcontrast phases also has a higher likelihood of malignancy. Photo courtesy of Hologic, Inc, Bedford, Massachusetts.
P.593
THE RADIOLOGIC REPORT AND PLAN The radiologic report should be clear and concise. The American College of Radiology has developed a standardized format and terminology called the Breast Imaging Reporting and Data System (BI-RADS) (3 8) for mammograms, breast US, and breast MR. All reports should begin with description of the overall breast composition. With mammography, this description of breast density will allow the clinician to gauge the sensitivity of the exam. The breast should be characterized as: (1) composed almost entirely of fat (75% glandular), which lowers the sensitivity of mammography. A description of the significant findings on the mammogram, US or MR should follow, and there should be comparison to any previous available examinations. The most important part of the breast imaging report is the assessment category, which should fall into one of the following seven categories: BI-RADS
Categoryw
(0): Need
additional
imaging
evaluation
and/or prior mammograms for comparison. This category is reserved for screening exams that require further imaging workup or comparison films to fully characterize a potential abnormality. The suggested additional studies, such as US or additional mammographic views, should be specified in the report. Prior mammograms are always helpful in the interpretation of a screening study. Category 0
1480
should, however, be used for film comparison only in cases where the radiologist feels that such films are essential to the final assessment for the patient. BI-RADS Category (1): Negative. No significant findings are present on a negative mammogram. The patient should return for routine screening. BI-RADS Categoryw (2): Benign finding. There is a benign finding such as a lipoma, oil cyst, galactocele, intramammary lymph node, hamartoma, fibroadenoma, cyst, scattered round calcifications of adenosis, arterial calcifications, sedimented calcium within microcysts, secretory calcifications, duct ectasia, skin calcifications, or multiple bilateral well-circumscribed masses representing cysts or fibroadenomas. These patients should return for routine screening. BI-RADS Category (3): Probably benign; initial short-interval follow-up suggested. The findings that should be included in this category are circumscribed masses, asymmetric parenchymal densities that are not associated with palpable masses and, occasionally, clusters of smooth, round, similar-appearing microcalcifications. The probability that such abnormalities represent cancer is less than 2% (3 9); therefore, most mammographers recommend a plan of careful follow-up. The first follow-up mammogram of the affected breast should be performed 6 months following discovery of the abnormality. If the abnormality is stable, a bilateral study should be performed 6 months later, and follow-ups should occur at yearly intervals for a period of at least 3 years. Progression of a cancerous lesion depends on tumor biology and doubling time; hence the necessity of a lengthy follow-up. Some cancers may grow slowly and others may change rapidly. BI-RADS Category (4): Suspicious abnormality; biopsy should be considered. Included in this category are lesions that are not classically malignant but are suspicious enough to warrant biopsy. The probability that such a lesion will represent malignancy is approximately 25% to 35% in most practices in the United States. Category 4 lesions can be divided into three subdivisions (4A, 4B, and 4C, with 4A the lowest suspicion for malignancy and 4C the
1481
highest); this division is optional, but it may allow more meaningful correlation with biopsy results. BI-RADS Category (5): Highly
suggestive
of
malignancy;
appropriate action should be taken. These are lesions that have a very high probability of being malignant and should undergo biopsy. Spiculated masses and pleomorphic clusters of calcifications are included in this category. BI-RADS Category (6): Known biopsy; proven malignancy; appropriate action should be taken. These are lesions that are already known to be malignant but have not undergone definitive therapy. For example, this category should be used for proven cancers that are being P.594 imaged to assess their response to neoadjuvant chemotherapy prior to definitive surgery. Clinicians must be cautioned that mammography has a false-negative rate of 9% to 16%; therefore, a negative mammogram should not preclude biopsy of a clinically suspicious
mass.
1482
FIGURE 20.44. MR of Infiltrating Carcinoma. Precontrast (A), early postcontrast (B), and late postcontrast (C) enhanced fat-suppressed T1W fast spoiled gradient-echo sagittal MR images of the left breast show an oval, spiculated 16-mm enhancing mass at 12:00. The mass demonstrates rapid initial enhancement with washout, as shown in the kinetic curve (D).
INTERVENTIONAL PROCEDURES NONPALPABLE LESIONS Mammographically
suspicious
abnormalities
1483
require
FOR histologic
or
cytologic examination for definitive diagnosis. Percutaneous, imagedirected core biopsy or aspiration performed in the radiology department is the standard of care. Needle localization followed by surgical excision is reserved for cases in which the percutaneous biopsy is inconclusive or for definitive surgery after percutaneous biopsy yields a malignant diagnosis.
Percutaneous
Biopsy
Increasing use of mammographic screening has led to the discovery of greater numbers of potentially malignant but clinically occult breast lesions. In the past, excisional biopsy was the only method for definitive diagnosis, but the development of imaging-guided percutaneous core biopsy of the breast now provides a more costeffective
alternative
(4 0). Nearly all suspicious lesions are amenable
to core biopsy with stereotactic, US, or MR guidance. Core biopsy is superior to fine-needle aspiration biopsy for the following
reasons: P.595
Histologic evaluation of core biopsy specimens can be performed by all pathologists, whereas cytologic diagnosis of fine-needle aspirates requires that the pathologist have special expertise and training. The amount of tissue obtained from core biopsies is usually sufficient for diagnosis; insufficient material for diagnosis is a frequent problem with fine-needle aspiration. Differentiation of invasive from noninvasive carcinomas is usually possible with core biopsy, whereas it is not with fine-needle aspiration cytology. Indications for core biopsy are similar to those for surgical biopsy. A full breast imaging workup must be completed before core biopsy is recommended. Core biopsy should not be substituted for shortinterval follow-up of probable benign lesions, because this approach is not cost-effective and may induce increased anxiety in some
1484
women. Technical difficulties such as inadequate visualization of the lesion may occasionally preclude the use of a core biopsy. Core biopsies can be guided by stereotactic images, US, or MR (4 1) . Currently two types of stereotactic units are available. One can be added onto a standard mammography machine but has limited working space and requires a seated patient. The other is a prone dedicated unit that is more costly but offers the advantages of having the patient in a prone position so as to minimize movement and vasovagal reactions (Fig. 20.45). A stereotactic unit allows the x-ray tube to move independent of the compressed breast. The lesion is centered in the aperture within the compression plate, and images at negative and positive 15° are obtained. Calculation of the amount of deviation of the lesion in these two views allows the exact determination of the depth of the lesion. The needle guide is adjusted for exact positioning of the needle in three dimensions to the center of the lesion. Following the injection of local anesthetic, a small skin incision is made to permit needle entry into the breast. Positioning of the needle is verified with stereotactic views, and biopsies are taken (Fig. 20.46) . When US is used, the needle can be observed in real time as the biopsy is performed (Fig. 20.47).
Adequate
sonographic
visualization
of the lesion is essential if core biopsy is to be performed with US guidance. Most microcalcifications and some masses, particularly those in fatty-replaced breasts, cannot be visualized and hence cannot be biopsied using US. Aspiration of fluid cannot be performed through a core biopsy needle. Many lesions chosen for US-guided biopsy will be atypical cysts; in such cases it is prudent to attempt aspiration with a 22-gauge needle. If fluid is not obtained, a core biopsy can be performed.
1485
FIGURE 20.45. Dedicated Stereotactic Biopsy Unit. The xray tube (arrowhead) moves independent of the compressed breast so that stereotactic images can be obtained. The needle guide (arrow) is adjusted so that the needle will be centered in the lesion. (Reproduced with permission from Fischer Imaging Corporation, Denver, Colorado.)
Lesions that are seen only on MR can be biopsied in the magnet using a grid system specifically designed to fit on the breast coil. There are several MR-compatible biopsy devices that allow vacuumassisted biopsies under MR guidance. In general, contrast enhancement is required to ensure appropriate targeting. A marking clip can be placed following an MR-guided biopsy. The clip can then be used as the target for a mammographic needle localization procedure, should that be necessary. Either a 14-gauge automated biopsy gun or a 9- to 11-gauge vacuum-assisted needle can be used for a core biopsy. The standard 14-gauge gun works by a spring-action mechanism that fires the needle through the lesion. The inner cannula containing the tissue notch is projected through the lesion first, and the cutting cannula is then fired over it so that a small core of tissue is retained within the specimen notch. With the vacuum-assisted devices, suction is used to bring the tissue into the specimen notch of the needle, which is then cut by an inner rotating cannula. Vacuum-assisted devices generally
1486
require only a single needle pass to obtain multiple specimens, whereas standard core biopsy requires multiple passes, one for each specimen. The vacuum-assisted needle offers improved ability to adequately sample microcalcifications in comparison to the standard biopsy gun (4 2) . The accuracy of core biopsy in diagnosing breast carcinoma approaches that of surgical biopsy, with reported sensitivities of 85% to 100% and specificities of 96% to 100% (4 3). To achieve such high sensitivities and specificities, it is essential that the mammographic, sonographic, and MR appearance of the lesion be correlated with the pathologic diagnosis. If there is discordance, repeat core biopsy or excisional biopsy should be performed. In cases where atypical ductal or lobular hyperplasia is diagnosed by core biopsy, excisional biopsy should be performed, P.596 because 10% to 48% of these lesions will ultimately prove to be carcinoma (4 1). Post–core biopsy management of papillary lesions, mucin-containing lesions, LCIS, and radial scars is controversial.
1487
FIGURE
20.46. Stereotactic Core Biopsy. A. On the scout
view, the lesion is centered in the aperture of the compression paddle. B . Stereotactic views at -15° and +15° are obtained, and the center of the lesion is marked in both views with the square target mark. C . After injection of local anesthetic, an 11gauge vacuum-assisted core biopsy needle is inserted and prefire stereo images are obtained to verify appropriate positioning of the needle; the needle should be inserted to a depth that is 5 mm short of the targeted center of the lesion. The vacuumassisted device is then fired into the lesion and multiple biopsy samples are obtained. D . After the biopsy is performed, a marking clip (arrows) is inserted and stereotactic images are obtained to verify appropriate positioning of the clip. Note air within the lesion where the biopsy specimens were obtained. In this case, the histologic diagnosis was invasive lobular carcinoma.
1488
Localization
of
Occult
Breast
Lesions
If surgical excision of a nonpalpable abnormality is to be performed, a localization will be required so that the surgeon is accurately directed to the lesion. Many methods for localization have been described. Localizations can be performed using needle-wire systems or dye injection. There are several commercially available needle-wire systems. All allow placement of a wire through an introducing needle that has been positioned in the breast at the site of the abnormality. The wires differ mainly in the configuration of the anchoring end. Injection of blue dye is a less frequently used method of localization. A needle is placed at the site of the abnormality and dye is injected. However, if there is a delay between the time of injection and surgery, diffusion of the dye through the tumors can occur, resulting in a biopsy specimen that may be larger than necessary. Methylene blue, formerly in common use for localizations, also interferes with estrogen
receptor
analysis.
Most mammographic units are equipped with a compression paddle that contains either one large hole marked on the edge with a grid, or a series of smaller holes marked P.597 with letters or numbers. The seated patient is placed in the mammographic unit so that the lesion to be localized is located under a hole in the compression plate. The skin surface closest to the lesion should be used for needle placement. For example, if the lesion is located at 12:00, a CC approach should be used. The breast is then filmed to determine the exact location of the abnormality. A needle is inserted parallel to the x-ray beam and through the abnormal area. The position of the needle with respect to the lesion is then checked by obtaining another film. If the needle position is satisfactory, the patient, with needle in place, is carefully removed from the mammography unit so that the tube can be rotated 90°. The patient is then positioned in the unit and the affected breast
1489
compressed along an axis parallel to the needle. A film is taken to assess the depth of the needle tip with respect to the lesion. The needle must be beyond the lesion before the technician proceeds. This ensures a fixed relationship between the localizer and the lesion. Optimally, the tip of the needle for a wire localization should be 1 to 2 cm beyond the lesion. Once the depth of the needle tip is satisfactory, the wire can be inserted through the needle and the needle withdrawn, leaving the wire in place (Fig. 20.48) . Alternatively, dye can be injected through the needle, which for this technique should be within the lesion or a few millimeters beyond it, and then the needle can be withdrawn. The patient is then sent to the operating room for surgical biopsy (4 4) .
FIGURE
20.47. US-Guided
Core
Biopsy. Prefire (A)
longitudinal US showing a 14-gauge core biopsy needle (black arrows) at the edge of a solid hypoechoic mass (white arrows) . The postfire image (B) shows the lesion (white arrows) pierced by the needle (black arrows) .
Bracketed localization is advocated for nonpalpable lesions over 2 cm in size. More than one localization wire is placed to demarcate the extent of the lesion. This technique is particularly helpful for areas of microcalcifications over 2 cm in diameter because it promotes complete removal of such lesions.
1490
After the biopsy has been performed, the excised tissue should be sent for x-ray. This ensures that the mammographic abnormality has been removed. In a small number of cases (1% to 5%), localization will fail and the lesion will not be removed. In most of these cases, the localization will have to be repeated. Most localizations are performed under mammographic guidance, but US and MR can also be used to guide such procedures. The technique used in US is similar to that for US-guided percutaneous biopsy. A high-frequency transducer is placed over the lesion, and the needle is introduced obliquely under real-time monitoring. When the tip is seen beyond the lesion, the wire can be inserted or the dye injected. Wire position should be confirmed by mammography. US is most useful in guiding a localization when the abnormality is seen well in one projection but is obscured by dense tissue in the second. It may also be useful when lesions are located in areas of the breast that are difficult to position within the hole in the localized compression paddle. US can only be used when the lesion can be visualized. Microcalcifications, in general, cannot be imaged, and not all soft tissue masses are well delineated by US. Lesions seen only on MR can be localized using the grid system that is used for MR-guided core biopsy. Freehand localizations can also be performed. Contrast enhancement is generally required to confirm the location of the lesion prior to needle placement. X-ray specimen radiography may not identify the lesion, because the contrast is no longer in the tissues. MR and pathologic correlation is, thus, extremely important. Discordant cases require postoperative MR to confirm removal of the lesion.
Other
Interventional
Procedures
Aspiration of sonographically atypical cysts can be performed for confirmation of the diagnosis using either US or mammographic guidance. The majority of such lesions will be smooth-walled masses that are atypical either because they lack through-transmission or because the fluid within them is not anechoic. In such cases, a 22gauge needle can be inserted using a technique similar to that used
1491
P.598 for core biopsy. If fluid is withdrawn, the lesion should be completely aspirated. If fluid cannot be withdrawn, the lesion is presumably solid, and core biopsy can be performed.
FIGURE 20.48. Needle Localization. CC (A) and MLO (B) mammograms show a highly suspicious spiculated mass in the upper outer quadrant (arrows). C . Localization was performed by placing the fenestrated compression plate over the lesion (arrow) and then placing a needle through the lesion, parallel to the xray beam. D . The hub of the needle (open arrow) is superimposed on the lesion; the tip (solid arrow) is at the posterior edge. A film is then taken in the 90° orthogonal
1492
projection and, after the depth is adjusted, the hook wire is passed through the needle. E . A film in the same projection demonstrates the final depth of the wire. F . The excised tissue is sent for specimen x-ray to confirm that the abnormality (arrows) has been removed. Histologic examination in this case revealed invasive lobular carcinoma.
If there is irregularity or nodularity of the cyst wall, as detected by sonography, devices are one needle surrounding
core biopsy should be undertaken. Vacuum-assisted preferable for biopsy of these types of lesions since only pass is required for sampling. It is likely that the fluid such lesions will leak into the surrounding tissues at
biopsy, thus rendering the lesion difficult to visualize for multiple passes. Cytologic evaluation of fluid surrounding an intracystic lesion is unreliable for diagnosis. Ductography can be used to investigate the cause of a spontaneous nipple discharge. The procedure involves the injection of contrast material into a duct, after which films are taken to look for intraductal tumors. These are most frequently papillomas and less commonly
carcinomas.
The P.599
utility of this study is controversial. If the patient has a bloody discharge, some surgeons prefer to inject the discharging duct with methylene blue in the operating room before dissecting along it. Others utilize preoperative ductography to evaluate bloody discharge, reasoning that if the ductogram is negative, the patient can be observed. The use of ductography in the evaluation of a unilateral, spontaneous serous discharge is similarly controversial, since both bloody and serous fluid can be associated with small cancers that may not be visible mammographically.
CONCLUSION Breast cancer represents a significant public health problem. More than 180,000 new cases are diagnosed and nearly 45,000 women die
1493
of the disease each year in the United States. Early detection with screening mammography is the only proven way to lower mortality from breast cancer. Diagnostic accuracy can be increased with the use of special mammographic views, US, and percutaneous biopsy techniques. Other modalities, such as MR and PET, are under study to determine their potential utility in detection and diagnosis of breast diseases. Utilization of breast imaging has increased over the last two decades, and mortality from breast cancer is declining. Our challenge, as radiologists, is to maintain the highest standards of quality in performance and interpretation of breast imaging studies; it is also to encourage all women to take regular advantage of these lifesaving techniques.
REFERENCES 1. Shapiro S. Evidence on screening for breast cancer from a randomized
trial.
Cancer
1977;39:2772–2782.
2. Nystrîm L, Rutqvist LE, Wall S, et al. Breast cancer screening with mammography: overview of Swedish randomized trials. The Lancet 1993;341:973–978. 3. Hendrick RE, Smith RA, Rutledge JH, et al. Benefit of screening mammography in women ages 40–49: a meta-analysis of new randomized controlled trial results. In: NIH Consensus Development Conference: Breast Cancer Screening for Women Ages 40–49, Program and Abstracts. Bethesda, MD: National Institutes of Health, 1997. 4. Paci E, Duffy SW, Giorgi D, et al. Quantification of the effect of mammographic screening on fatal breast cancers: the Florence Programme: 1990–1996. Br J Cancer 2002;87:65–69. 5. Tabar L, Vitak B, Chen HHT, Yen MF, Duffy SW, Smith RA. Beyond randomized controlled trials. Organized mammographic
1494
screening substantially reduces breast Cancer 2001;91:1724–1731.
carcinoma
mortality.
6. Smith RA, Saslow D, Sawyer KA, et al. American Cancer Society guidelines for breast cancer screening: update 2003. CA Cancer J Clin 2003;53:141–169. 7. von Eschenbach AC. NCI remains committed to current mammography
guidelines.
The
Oncologist
2002;7:170–171.
8. White E, Miglioretti DL, Yankaskas BC, et al. Biennial versus annual mammography and the risk of late-stage breast cancer. J Natl Cancer Inst 2004;96(24):1832–1839. 9. Tabar L, Larsson LG, Andersson I, et al. Breast-cancer screening with mammography in women aged 40–49 years. Int J Cancer 1996;68:693–699. 10. Tabar L, Fagerberg G, Day NE, Holmberg L. What is the optimum interval between mammographic screening examinations? An analysis based on the latest results of the Swedish two-county breast cancer screening trial. Br J Cancer 1987;55:547–551. 11. Kerlikowske K, Grady D, Barclay J, et al. Effect of age, breast density, and family history on the sensitivity of first screening mammography. JAMA 1996;276:33–38. 12. Curpen BN, Sickles EA, Sollitto RA, et al. The comparative value of mammographic screening for women 40–49 years old versus women 50–64 years old. AJR Am J Roentgenol 1995;164:1099–1103. 13. Linver MN. Mammography outcomes in a practice setting by
1495
age: prognostic factors, sensitivity, and positive biopsy rate. In: National Institutes of Health Consensus Development Conference Syllabus, Breast Cancer Screening for Women Ages 40–49, January 1997, Bethesda, Maryland. Bethesda, MD: National Institutes of Health, 1997. 14. Feig SA, Ehrlich SM. Estimation of radiation risk from screening mammography: recent trends and comparison with expected benefits. Radiology 1990;174:638–647. 15. Kolb TM, Lichy J, Newhouse JH. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology 2002;225:165–175. 16. Kriege M, Brekelmans CTM, Boetes C, et al. Efficacy of MRI and mammography for breast-cancer screening in women with a familial or genetic predisposition. N Engl J Med 2004;351:427–437. 17. de Paredes ES, Marsteller LP, Eden BV. Breast cancers in women 35 years of age and younger: mammographic findings. Radiology
1990;177:117–119.
18. Lewin JM, D’Orsi CJ, Hendrick RE. Digital mammography. Radiol Clin North Am 2004;42:871–884. 19. Sickles EA. Mammographic features of 300 consecutive nonpalpable breast cancers. AJR Am J Roentgenol 1986;146:661–663. 20. Bassett LW, U.S. Agency for Health Care Policy and Research, et al. Clinical Practice Guideline no. 13: Quality Determinants of Mammography. AHCPR Publication No. 95-0632. Rockville, MD:
1496
U.S. Department of Health and Human Services, 1994:25–31. 21. Kopans DB, Swann CA, White G, et al. Asymmetric breast tissue. Radiology 1989;171:639–643. 22. Marsteller LP, de Paredes ES. Well defined masses in the breast. Radiographics 1989;9:13–37. 23. Sickles EA. Breast masses: mammographic evaluation. Radiology 1989;173:297–303. 24. Sickles EA. Nonpalpable, circumscribed, noncalcified solid breast masses: likelihood of malignancy based on lesion size and age
of
patient.
Radiology
1994;192:439–442.
25. Bassett LW. Mammographic analysis of calcifications. Radiol Clin North Am 1992;30:93–105. 26. Harris JR, Lippman ME, Veronesi U, Willet W. Breast cancer (second of three parts). N Engl J Med 1992;327:390–398. 27. Eklund GW, Busby RC, Miller SH, Job JS. Improved imaging of the augmented breast. AJR Am J Roentgenol 1988;151:469–473. 28. Destouet JM, Monsees BS, Oser RF, et al. Screening mammography in 350 women with breast implants: prevalence and findings of implant complications. AJR Am J Roentgenol 1992;159:973–978. 29. Gorczyca DP, Schneider E, DeBruhl ND, et al. Silicone breast implant rupture: comparison between three-point Dixon and fast spin-echo MR imaging. AJR Am J Roentgenol 1994;162:305–310.
1497
30. DeBruhl ND, Gorczyca DP, Ahn CY, et al. Silicone breast implants:
US
evaluation.
Radiology
1993;189:95–98.
31. Morris E. Breast cancer imaging with MRI. Radiol Clin North Am
2002
May;40(3):443–466.
32. Lee SG, Orel SG, Woo IJ, et al. MR Imaging screening of the contralateral breast in patients with newly diagnosed breast cancer: preliminary results. Radiology 2003;226:773–778. 33. Liberman L, Morris EA, Kim CM, et al. MR imaging findings in the contralateral breast of women with recently diagnosed breast cancer. AJR Am J Roentgenol 2003;180:333–341. 34. Lee JM, Orel SG, Czerniecki BJ, Solin LJ, Schnall MD. MRI before reexcision surgery in patients with breast cancer. Am J Roentgenol
2004;182:473–480. P.600
35. Lee CH, Smith RC, Levine JA, Troiano RN, Tocino I. Clinical usefulness of MR imaging of the breast in the evaluation of the problematic mammogram. AJR Am J Roentgenol 1999;173:1323–1329. 36. Orel SG, Weinstein SP, Schnall MD, et al. Breast MR imaging in patients with axillary node metastases and unknown primary malignancy. Radiology 1999;212:543–549. 37. Kuhl CK, Mielcareck P, Klaschik S, et al. Dynamic breast MR imaging: are signal intensity time course data useful for differential diagnosis of 1999;211:101–110.
enhancing
lesions?
Radiology
38. American College of Radiology (ACR), BI-RADS Committee.
1498
ACR Breast Imaging Reporting and Data System: Breast Imaging Atlas. Reston, VA. American College of Radiology, 2003. 39. Sickles EA. Periodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases. Radiology 1991;179:463–468. 40. Lindfors KK, Rosenquist CJ. Needle core biopsy guided with mammography: a study 1994;190: 217–222.
of
cost-effectiveness.
Radiology
41. Berg WA. Image-guided breast biopsy and management of high-risk lesions. Radiol Clin North Am 2004;24:935–946. 42. Meyer JE, Smith DN, DiPiro PJ, et al. Stereotactic breast biopsy of clustered microcalcifications with a directional, vacuumassisted device. Radiology 1997;204:575–576. 43. Bassett L, Winchester DP, Caplan RB, et al. Stereotactic coreneedle biopsy of the breast: a report of the joint task force of the American College of Radiology, American College of Surgeons, and College of American Pathologists. Cancer 1997;47:171–190. 44. Kopans DB, Lindfors K, McCarthy KA, Meyer JE. Spring hookwire breast lesion localizer: use with rigid-compression mammographic systems. Radiology 1985;157:537–538.
1499
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section V - Cardiac Radiology > Chapter 21 Cardiac Anatomy, Physiology, and Imaging Modalities
Chapter
21
Cardiac Anatomy, Physiology, Imaging Modalities
and
David K. Shelton
Imaging
Methods
Thorough knowledge of cardiac anatomy and physiology is important as a basis for cardiac imaging. Comprehensive knowledge of cardiac imaging also requires consideration of virtually all the available imaging modalities. Chest radiography provides the initial evaluation of most cardiac patients. A barium esophagram can provide additional information because of the close relationship of the esophagus to the cardiac structures. Fluoroscopy increases the detectability of coronary and valvular calcification and provides dynamic
and
positional
information.
Transthoracic
echocardiography,
including pulse-wave and color-flow Doppler, and transesophageal echocardiography provide additional detailed imaging of internal cardiac anatomy and function. Nuclear cardiology, PET, and pharmacologic testing provide key functional, perfusion, and physiologic information. Cardiac and coronary angiography, although invasive, provide detailed anatomic information that can lead directly to interventional or surgical therapy. CT, multidetector CT (MDCT), CT angiography (CTA), and ultrafast CT with the use of IV iodinated contrast material are capable of providing critical information, particularly for pericardial or intracardiac disease. Recent technologic advances in the latter also allow detection of premature coronary
1500
calcification, which may have prognostic implications. MR adds threedimensional (3D) tomographic and motion studies of the myocardium, valves, and chambers without using ionizing radiation or intravascular contrast. Cardiac imaging requires familiarity with all imaging techniques and their associated physics, 3D cardiac anatomy, cardiac physiology, and cardiac disease processes.
ANATOMY The four-chambered heart lies primarily in the anterior left hemithorax, with the LV lying on the left hemidiaphragm (Figs. 21.1, 21.2). The RA extends to the right of midline as it receives systemic blood from the superior vena cava (SVC), inferior vena cava (IVC), and coronary sinus. The RA and RV lie primarily anterior to the planes of the LA and LV. The RV is the most anterior chamber and abuts P.604 the sternum (Fig. 21.3). The LA is subcarinal and midline in the thorax, being supplied by the right and left superior and inferior pulmonary
veins.
1501
FIGURE 21.1. Normal Posteroanterior Chest Radiograph. Frontal view of the chest demonstrates normal heart size, contours, and chamber size. The hila and pulmonary vascularity are normal. The LV (arrowheads) is border-forming on the left. The RA (curved arrow) is border-forming on the right. The aortic knob (arrow) is of normal contour, and the pulmonary artery (open arrow) is concave.
Frontal
Projection
The right border of the cardiac silhouette is formed primarily by the RA, with the SVC entering superiorly and the IVC often seen at its lower margin (Figs. 21.1, 21.3). The left border of the heart is created primarily by the LV and LA appendage. The pulmonary artery, aortopulmonary window, and aortic knob extend superiorly.
Lateral
Projection
The RV is border-forming anteriorly adjacent to the sternum, with its outflow tract extending superiorly and posteriorly (Fig. 21.2). The LA is border-forming in the high posterior, subcarinal region. The LV is border-forming
Right
inferiorly
and
posteriorly.
Atrium
The RA is divided into two portions. The smooth posterior wall develops from the sinus venosus, with the attached SVC and IVC in continuity posteriorly (Fig. 21.4). The trabeculated anterior wall is derived from the embryonic RA. The RA appendage extends superiorly and medially from the SVC opening. The crista terminalis is a muscular ridge that runs from the mouth of the SVC and fades inferiorly to the mouth of the IVC. It divides the two portions of the atrium and corresponds to an external sulcus terminalis. The medial or posterior wall of the RA is the interatrial septum, which contains a smooth, central dimpled area called the fossa ovalis. Inflow from the
1502
SVC, IVC, and coronary sinus enters the smooth posterior portion of the RA. The SVC has a free opening, whereas the IVC is partially guarded by a thin eustachian valve, which is occasionally absent or perforated (network of Chiari). The large draining coronary vein or coronary sinus enters the RA anterior and medial to the IVC. Its opening is guarded by the thebesian valve between the orifice of the IVC and the tricuspid valve.
FIGURE 21.2. Normal Lateral Chest Radiograph. This wellpositioned left lateral chest radiograph demonstrates the right ribs projected posterior to the left ribs because of divergence of the x-ray beam. The right and left bronchi are overlapped, and the sternum is seen in the lateral view. The true lateral projection allows evaluation of the inferior vena cava intersection (arrow) with the LV. There is no evidence of posterior displacement of the left bronchus (curved arrow) to indicate left
1503
atrial enlargement. There is no evidence of right ventricular encroachment into the retrosternal clear space.
Right
Ventricle
The RV (Figs. 21.4, 21.5) lies anterior to the left ventricular outflow tract and wraps around it and to the left. The right ventricular outflow is directed superiorly, posteriorly, and to the left. The RV is divided into a posterior or inferior portion (inflow or sinus portion), which is heavily trabeculated, and a less trabeculated anterior or superior portion (outflow tract or pulmonary conus). The two portions of the RV are divided by the crista supraventricularis, which is a muscular ridge with a septal band called the moderator band. This band is present in more than 40% of patients, connects the interventricular septum to the anterior papillary muscle, and contains the right bundle branch. The infundibulum (conus arteriosus) P.605 is the smooth cephalic portion of the RV that leads to the pulmonary trunk.
1504
FIGURE 21.3. Cardiothoracic Anatomy: Frontal View of the Heart After Cutaway of the Chest Wall, Pleural Surfaces, and Pericardial Surface. Note the relationship of the RA, RV, left atrial appendage, and LV to the great vessels. (Reproduced with permission. Drawing by Frank H. Netter MD, from Atlas of Human Anatomy. The CIBA Collection of Medical Illustrations, Clinical Symposia. West Caldwell, NJ: CIBA-Geigy Corp, 1989.)
1505
Pulmonary
Arteries
The muscular pulmonary conus extends to the semilunar tricuspid pulmonary valve, with the pulmonary trunk extending superiorly and to the left. The left pulmonary artery extends posteriorly as a continuation of the main pulmonary artery, coursing over the top of the left main stem bronchus and then descending posteriorly. The right pulmonary artery extends horizontally to the right, bifurcates within the pericardial sac, and exits the right hilum as the truncus anterior and interlobar arteries. The left mainstem bronchus is hyparterial, meaning that it lies below the pulmonary artery. The right bronchus is eparterial, meaning that it lies next to the right pulmonary
artery.
The ligamentum arteriosum arises from the superior proximal left pulmonary artery and crosses through the aorticopulmonary window to the floor of the aorta. The ligamentum arteriosum is the remnant of the ductus arteriosus, which closes functionally in the first 24 hours and closes anatomically by 10 days following birth. Desaturated blood from the right heart circulates through the lungs and returns as oxygenated blood through the right and left superior and inferior pulmonary veins into the LA.
Left
Atrium
The LA is the highest and most posterior chamber (Fig. 21.6). Its smooth walls are nestled between the right and left bronchi, and its posterior wall abuts the anterior wall of the esophagus. The left atrial appendage is a small pouch that projects superiorly and to the left and is smoother and longer than the right atrial appendage. The left atrial appendage extends anterior to the left superior pulmonary veins and is readily seen on MR and CT scans. The foramen ovale within the interatrial septum remains nominally patent in up to 25% of adults. Its inferior margin is a remnant of the septum primum and may be somewhat scalloped. The mitral valve is located anterior and inferior to the body of the LA, with the mitral valve leaflets extending into the LV.
1506
Left
Ventricle
The mitral valve is the conduit for blood flow from the LA to the LV and is in the high posterior P.606 “valve plane― of the LV (Figs. 21.5, 21.6). The anterior or septal leaflet of the mitral valve lies near the interventricular septum and extends to the posterior (noncoronary) cusp of the aortic valve. The smaller posterior mitral leaflet lies posteriorly and to the left. The chordae tendineae are strong fibrous cords that extend from the mitral leaflets to the papillary muscles of the LV. The inflow portion of the LV is posterior to the anterior leaflet of the mitral valve. The outflow portion of the LV is anterior and superior to the anterior mitral leaflet. The interventricular septum has a high membranous portion that is contiguous with the aortic root. The more muscular inferior portion of the septum extends to the left ventricular apex. The esophagus passes immediately posterior and is in contact with the muscular wall of the LV.
1507
FIGURE 21.4. Cutaway Views of the Right Atrium and Right Ventricle. (Reproduced with permission. Drawing by Frank H. Netter, MD, from Atlas of Human Anatomy. The CIBA Collection of Medical Illustrations, Clinical Symposia. West Caldwell, NJ: CIBA-Geigy Corp, 1989.)
1508
Aorta The outflow tract of the LV leads into the aortic root through the aortic valve, which is composed of right, left, and posterior (noncoronary) cusps. The sinuses of Valsalva are the reservoirs created by the closure of the aortic valve and from which the right and left coronary arteries arise. The posterior wall of the aorta is continuous with the anterior leaflet of the mitral valve and more superiorly abuts the anterior wall of the LA. The anterior wall of the aorta is continuous with the interventricular septum. After coursing superiorly and then to the left, the aorta gives off the right innominate artery, left common carotid artery, and left subclavian artery. The aortic arch is the transverse portion of the aorta that abuts the left wall of the trachea, causing a characteristic indentation.
Conduction
System
The sinoatrial node consists of specialized neuromuscular tissue that measures approximately 5 to 20 mm and is located on the anterior endocardial surface of the RA, just above the SVC and right atrial appendage junction, near the crista terminalis. Electrical propagation spreads to both atria via Purkinje-like fibers and is recorded as the P wave on an electrocardiogram. The atrioventricular node is a 2- × 5-mm region of neuromuscular tissue on the endocardial surface, along the right side of the interatrial septum, just inferior to the ostium of the coronary sinus. The impulse is collected and P.607 delayed approximately 0.7 seconds in the atrioventricular node before passing into the bundle of His. The bundle of His is a 20-mmlong tract that extends down the right side of the membranous interventricular septum. The bundle of His bifurcates into right and left bundles before arborizing through the two ventricles via the Purkinje system. The interventricular septum activates from superior to inferior, with the anterior or septal RV being the first to activate and the posterior or basal LV being the last to activate. This information is particularly useful when evaluating phase analysis or
1509
phase propagation in gated cardiac scintigraphy.
FIGURE 21.5. Bisection Through the Heart Simulating a Four-Chamber View. (Reproduced with permission. Drawing by Frank H. Netter, MD, from Atlas of Human Anatomy, The CIBA Collection of Medical Illustrations, Clinical Symposia. West Caldwell, NJ: CIBA-Geigy Corp, 1989.)
1510
CARDIAC Left-sided
CATHETERIZATION
catheterization
is
normally
accomplished
via
arterial
puncture in the femoral or brachial artery (Fig. 21.7). It is typically used for aortography, coronary and coronary bypass graft angiography, ventriculography, and evaluation for patent ductus arteriosus. Right-sided catheterization is typically accomplished by venous puncture in the femoral or brachiocephalic vein (Fig. 21.8). It is used for pulmonary angiography, catheterization of the RA and RV, or evaluation of shunt lesions, such as an atrial septal defect. Important considerations include determination of the catheter course to help diagnose atrial septal defects, ventricular septal defects, patent ductus arteriosus, or persistent left SVC. During catheterization,
oxygen
saturation
percentages
are
commonly
determined, along with pressure measurements and pressure gradients (Table 21.1). Contrast is injected to demonstrate additional details of anatomy, as well as to evaluate for valvular lesions, chamber size, ventricular function, and wall motion. Right atrial pressures are normally 2 to 5 mm Hg and oxygen saturation is 65% to 75%. Elevated right atrial pressures are seen with right heart failure, decreased compliance, and tricuspid valve disease. A 7% or greater increase in saturation from the IVC to the RA is considered evidence of a left-to-right shunt atrial septal defect (ASD). P.608 Right ventricular pressures are typically 25 mm Hg systolic and 0 to 5 mm Hg diastolic. Elevated systolic pressures are seen with pulmonary hypertension, pulmonic valve stenosis, and congenital heart lesions such as transposition and truncus arteriosus. Diastolic pressures increase with right heart failure. Saturations should be nearly the same as right atrial saturations. A 5% increase in saturation from RA to RV suggests a ventricular septal defect.
1511
FIGURE 21.6. Cutaway Views of the LV and LA. (Reproduced with permission. Drawing by Frank H. Netter, MD, from Atlas of Human Anatomy, The CIBA Collection of Medical Illustrations, Clinical Symposia, 1989.)
Pulmonary
arterial
pressures are normally 25 mm Hg systolic and
1512
10 mm Hg diastolic, with a mean pulmonary artery pressure of 15 mm Hg. A significant pressure gradient (>10 mm Hg) across the valve implies pulmonic valve stenosis. Increased pressures are seen with shunt lesions, pulmonary vascular disease, and pulmonary venous obstruction. Pulmonary arterial saturation should be approximately the same as right ventricular saturation, with a 3% difference considered significant for a shunt lesion. Pulmonary capillary wedge pressure is typically 2 to 8 mm Hg and approximates the left atrial pressure unless there is evidence of pulmonary venous obstruction. Elevations in the left atrial or wedge pressure are usually seen with mitral stenosis and left-sided congestive heart failure (CHF). Normal left atrial saturation is approximately 94%, and a decrease greater than 5% implies a rightto-left shunt. Left ventricular pressures are normally, approximately 120 mm Hg systolic and 0 to 5 mm Hg diastolic. Decreased systolic pressures are seen with shock and CHF. Elevated systolic pressures imply systemic hypertension or outlet obstruction. Increased diastolic pressure is seen with CHF. Decreased saturation at the left ventricular level would imply a right-to-left shunt. Normal aortic pressure is approximately 120 mm Hg systolic and 80 mm Hg diastolic, with a mean pressure of 70 to 100 mm Hg. With each systolic contraction, the average stroke volume of each ventricle is 70 mL of blood (Table 21.2). End-diastolic volume is normally 125 to 150 mL for the LV and 165 mL for the RV. A normal cardiac output is 4 to 5 L/min, with a normal cardiac index of 2.8 to 4.0
L/min/m2 P.609
of body surface area. The normal ejection fraction is 50% to 75% for the LV and 45% to 55% for the RV. Typical end-diastolic volumes are 57 mL for the RA and 50 mL for the LA. Coronary blood flow averages approximately 224 mL/min and increases by up to sixfold during
exercise.
1513
FIGURE 21.7. Aortogram Via Transfemoral Approach. The catheter is placed in the mildly dilated ascending aorta (straight arrow). Notice the reflux of contrast from the aortic valve into the LV (curved
arrow) in this patient with aortic insufficiency.
1514
FIGURE
21.8. Right Heart Catheterization Via the Right
Subclavian Vein. The catheter is positioned in the pulmonary conus. Contrast fills the main, right, and left pulmonary arteries. Note the arteriovenous malformation with a large feeding artery (arrow) .
1515
TABLE 21.1 Normal Values for Cardiac Catheterization
Site
Pressure (mm Hg)
Saturation
Vena cava
5
60–65
Right
atrium
2–5
65–75
Right
ventricle
25/0
70
25/10
73
Pulmonary
artery
Left
atrium
2–8
94–98
Left
ventricle
120/0–5
94–98
120/80
94–98
Aorta
Aortic
(%)
Valve
The normal aortic valve orifice is 3 cm2 . Symptoms result from aortic stenosis usually when either the orifice is less than 0.7 cm,2 or if aortic stenosis and insufficiency are present, when it is less than 1.5 c m2 . Mild stenosis is indicated by a pressure gradient across the aortic valve greater than 25 mm Hg, moderate stenosis by a gradient greater than 40 to 50 mm Hg, and severe stenosis by a gradient exceeding 80 mm Hg.
Mitral
Valve
The mitral valve orifice usually measures 4 to 6 cm2 . Mild mitral stenosis occurs with an orifice smaller than 1.5 cm2 , moderate mitral
1516
stenosis at smaller than 1.0 cm2 , and severe mitral stenosis at smaller than 0.5 cm2 . Pulmonic
stenosis is considered significant if the right ventricular
systolic pressure exceeds 70 mm Hg. Pulmonary artery hypertension is defined as a mean pulmonary artery pressure of more than 25 mm Hg.
CHEST
RADIOGRAPHY
The chest radiograph remains the mainstay for imaging of the heart and lungs. There are many approaches to P.610 reading the radiograph. Although most radiologists initiate the process with “global perception,― it is important to develop a checklist scan technique. This discussion concerns adult posteroanterior and lateral radiographs.
TABLE 21.2 Average Physiologic Data for Cardiac Chambers
Parameter
Left
Chambers
Right Chambers
Atrial end-diastolic volume
50 mL
Ventricular enddiastolic volume
125–150
Ejection
50%–75%
45%–55%
70 mL
70 mL
Stroke
fraction
volume
1517
57 mL
mL
165 mL
Cardiac
output
4–5
Cardiac
index
2.8–4 L/min/m2
Cardiac
L/min
4–5
L/min
2.8–4 L/min/m2
Silhouette
Size The cardiothoracic ratio should not exceed 0.5 on a 72-inch erect posteroanterior (PA) radiograph or 0.6 on a portable or anteroposterior (AP) examination. Other factors should be considered, such as fat pads and pectus deformity.
Shape Various contour effects can offer clues to underlying disease. A “water bottle― configuration occurs with pericardial effusion or generalized cardiomyopathy. Left ventricular or “Shmoo― configuration (after Al Capp's Shmoo) describes lengthening and rounding of the left heart border, with a downward extension of the apex resulting from left ventricular enlargement. “Hypertrophy― configuration describes increased convexity of the left heart border and apex. Right ventricular hypertrophy and enlargement tend to lift the apex and create a more horizontal vector to the cardiac axis. Hypertrophy of either ventricle usually causes little enlargement of the silhouette unless dilatation is also present. Hypertrophy typically results from increased afterload, whereas dilation occurs with failure or diastolic overload. “Straightening― of the left heart border is seen with rheumatic heart disease and mitral stenosis.
“Moguls
of
the
Heart.―
“Skiing the moguls of the heart― refers to the left mediastinal
1518
outline beginning at the aortic knob. A prominent knob is a clue to ectasia, aneurysm, or hypertension. Notching or a “figure 3― sign of the aorta suggests coarctation (Fig. 21.9). The second mogul is the main pulmonary artery segment. Excessive convexity is seen with poststenotic dilatation, chronic obstructive pulmonary disease, pulmonary artery hypertension, left-to-right shunts, and pericardial defects. Severe concavity suggests right-to-left shunts. The third mogul is a prominent left atrial appendage that in 90% of cases indicates prior rheumatic carditis (Fig. 21.10). It is not usually seen with other bulge just ventricular caused by
causes of above the aneurysm. pericardial
Chamber Left
atrial
left atrial enlargement. The fourth mogul is a cardiophrenic angle, seen with infarction or A fifth bulge at the cardiophrenic angle is cysts, prominent fat pads, or adenopathy.
Enlargement
enlargement is best confirmed by measuring the
distance from the midinferior border of the left mainstem bronchus to the right lateral border of the left atrial density (see Fig. 21.10). This distance is less than 7 cm in 90% of normal patients and greater than 7 cm in 90% of patients with left atrial enlargement, as P.611 proven by echocardiography. This measurement can be approximated by placing one's right fifth finger under the left bronchus and, while keeping the fingers closed, determining whether the LA is seen beyond one's four fingertips; if so, the LA is enlarged. Less-sensitive signs of left atrial enlargement include splaying of the carinal angle, uplifting of the left mainstem bronchus, and prominence of the left atrial appendage. On occasion, the enlarged LA will displace the descending aorta to the left. Massive left atrial enlargement can result in the LA becoming border-forming on the right side, so-called “atrial escape.― On lateral views, an enlarged LA will displace the left bronchus posteriorly, with the bronchi creating right and left legs for the “walking man sign.― An enlarged LA also impresses against the esophagus.
1519
FIGURE
21.9. Aortic
Coarctation. Notice the “figure 3
sign― or notching of the aorta near the aortic knob (straight arrow). The ascending aorta (curved arrow) is prominent, and the LV is excessively rounded (arrowheads). Rib notching is noted along the right fifth rib margin inferiorly (curved open arrow) .
1520
FIGURE
21.10. Rheumatic
Heart
Disease. The left atrial
appendage is strikingly prominent (curved
arrow). Splaying of
the carina and a double density along the right heart border indicates left atrial enlargement (arrowheads). When the distance from the lateral margin of the LA to the midpoint on the undersurface of the left bronchus exceeds 7 cm, left atrial enlargement is likely (black arrows) .
Right atrial enlargement is more difficult to define on chest radiographs than left atrial enlargement, but fortunately, it is less common. Clues include a prominent atrial bulge too far to the right of the spine (more than 5.5 cm from the midline on a well-positioned PA radiograph). Another sign is elongation of the right atrial convexity to exceed 50% of the mediastinal or cardiovascular shadow. Right atrial enlargement usually accompanies right ventricular
Left
enlargement.
Ventricular
Enlargement
1521
On the PA view, an enlarged LV creates an elongated left heart border with the apex pointing downward. Prominent rounding of the inferior left heart border is also seen (Fig. 21.11). The lateral view shows an enlarged LV extending behind the esophagus. The HoffmanRigler sign for left ventricular enlargement exists when the LV extends more than 1.8 cm posterior to the posterior border of the IVC at a level 2 cm cephalad to the intersection of the LV and the IVC (Fig. 21.12). This sign requires a true lateral radiograph and can be a false-positive result if the lateral view is slightly oblique or if there is volume loss in either lower lobe. This sign can be quickly applied by using one of the “2-cm fingertips― for a quick check without a ruler.
FIGURE 21.11. Left Ventricular Enlargement on Posteroanterior Chest Radiograph. Prominence of the LV with rounding along the inferior heart border and an apex that is pointing downward (arrowheads) are indicative of “left ventricular configuration.― The ascending aorta (arrow) is dilated because of aortic stenosis and insufficiency.
1522
FIGURE
21.12. Left Ventricular Enlargement on Lateral
Chest Radiograph. The posterior margin of the LV (arrowheads) projects prominently behind the inferior vena cava (open arrow) and overlaps the thoracic spine. The Hoffman-Rigler sign is positive.
Right ventricular enlargement is not as easily detected as leftsided enlargement. If the heart is enlarged and the Rigler sign does not show left ventricular enlargement, then consider right-sided enlargement. If the RV fills too much of the retrosternal clear space or “climbs― more than one third of the sternal length, then right ventricular enlargement is likely. Indirect signs such as enlargement of the pulmonary outflow tract or hilar arteries add confidence.
Abnormal
Mediastinal
1523
Contours
Aorta Dilatation of the ascending aorta as a result of poststenotic dilatation is seen in approximately 80% of patients with aortic stenosis (Fig. 21.11). It can also be seen in patients older than 50 years when there is tortuosity of the entire aorta or systemic hypertension. Ascending aortic aneurysm (calcific with P.612 syphilis, not calcified with Marfan syndrome) is another possibility (Fig. 21.13). A ductus bump adjacent to the aortic knob can be an indication of patent ductus arteriosus.
FIGURE 21.13. Calcified Aortic Aneurysm on Lateral Chest Radiograph. The ascending aorta is enlarged in this patient with a syphilitic calcified aortic aneurysm. The anterior margin is identified by soft tissue prominence (straight arrow) overlapping the retrosternal clear space. The posterior margin is identified by calcification in the wall (curved
arrow) .
1524
Azygos vein dilatation (>6 mm on upright PA radiograph or >1 cm on supine radiograph) is seen with intravascular volume expansion, elevated central venous pressure, and right heart failure (see Fig. 23.17). Additional causes include the Valsalva maneuver, pregnancy, renal failure, vena cava obstruction, or azygos continuation of the IVC. Dilatation of the SVC often accompanies volume expansion or elevated central venous pressure but is more difficult to detect with certainty.
Cardiac
Calcifications
Coronary
Calcification
Radiographs commonly demonstrate coronary artery calcification in a 3-cm triangle along the upper left heart border, called the CAC (coronary artery calcification) triangle (Figs. 21.14, 22.1). If chest pain and coronary calcification are present, there is a 94% chance the patient will show occlusive coronary artery disease at angiography. Fluoroscopic detection of coronary calcification actually has higher sensitivity and specificity in screening asymptomatic individuals than does exercise-tolerance testing. In symptomatic patients, the detection of coronary calcification approaches exercisetolerance testing in sensitivity and exceeds exercise-tolerance testing in specificity. More than 82% of the patients with fluoroscopically demonstrated coronary artery calcification and positive exercisetolerance testing show significant coronary artery disease at angiography. Calcifications have more significance when seen in patients under 60 years of age. Heavier and more extensive calcification correlates with more severe coronary disease. Detection of coronary calcification helps to differentiate patients with ischemic cardiomyopathy from those with nonischemic cardiomyopathy.
1525
FIGURE 21.14. Coronary Artery Calcification on Posteroanterior Chest Radiograph. The calcification (arrow) is most commonly detected in the coronary artery calcification triangle along the upper left heart margin (arrowheads). The presence of coronary artery calcification may be indicative of coronary stenosis and ischemic heart disease.
Valvular
calcification is seen in 85% of patients with acquired
valvular disease but is rarely detected in patients under 20 years of age. Aortic valve calcification is highly suggestive of valve disease. Calcific aortic stenosis is most often degenerative or atherosclerotic in origin and is usually seen in older men. Extensive aortic annulus calcification is atherosclerotic in nature and has been associated with conduction blocks.
1526
Mitral valve calcification is highly suggestive of rheumatic valvular disease and is seen on the chest radiographs of approximately 40% of patients with mitral stenosis. It is even more common in patients with stenosis and regurgitation. Atherosclerotic calcification of the mitral annulus occurs in approximately 10% of the elderly population (Fig. 21.15). It appears as circular, ovoid, C-shaped, or J-shaped calcification in the mitral annulus and can lead to mitral valve incompetence. P.613 Sinus of Valsalva aneurysm calcification is seen as a curvilinear density anterior and lateral to the ascending aorta.
FIGURE 21.15. Mitral Annulus Calcification on Lateral Chest Radiograph. Ovoid calcification of the mitral annulus (arrow) is secondary to atherosclerosis and is commonly associated with mitral insufficiency. Mitral calcification is best seen on a lateral radiograph.
Calcified
ligamentum
arteriosum is seen as a linear calcification in
1527
the aortopulmonary window connecting the top of the left pulmonary artery to the floor of the aortic arch.
Calcified
LA
Thin curvilinear calcification in the wall of the LA is usually associated with mitral stenosis, left atrial enlargement, atrial fibrillation, and left atrial thrombus. Calcified pericardium is typically anterior and inferior in location. It can be single or double layered and is associated with a high incidence of constrictive pericardial hemodynamics. Causes include viral, hemorrhagic, and tuberculous pericarditis as well as postsurgical scarring.
Calcified
Infarct
Dystrophic calcification may occur in the myocardial wall from prior myocardial
infarction.
Calcified
ventricular
aneurysm
Thin curvilinear calcification anterolaterally near the apex is most often seen with true aneurysms. Posterior curvilinear calcification is usually seen in pseudoaneurysms (Fig. 21.16) . Calcified
thrombus is seen as clumpy calcification in the LA or, less
commonly, in the LV.
Calcified
Pulmonary
Arteries
Thin, eggshell-like calcification in the walls of the pulmonary arteries is virtually diagnostic of long-standing pulmonary arterial hypertension (see Figs. 22.15, 22.16) .
Tumors Rounded or stippled calcifications are seen occasionally in atrial myxomas and rarely in other cardiac neoplasms (see Fig. 22.29) .
1528
Pulmonary
Vascularity
The lungs have dual blood supply, with the pulmonary arteries and the systemic bronchial arteries.
FIGURE Lateral
21.16. Calcified Chest
Ventricular
Pseudoaneurysm
on
Radiograph. Thin, curvilinear calcification along
the posterior wall of the LV (arrowhead) is indicative of a ventricular pseudoaneurysm.
Pulmonary
Arteries
Increased circulation from left-to-right shunts results in enlargement of the main and hilar pulmonary arteries with increased blood flow to the upper and lower lobes. Asymmetric blood flow can be seen with pulmonary hypoplasia, Swyer-James syndrome, and congenital lesions, such as pulmonary stenosis (increased to the left lung) or tetralogy of Fallot (increased to the right lung) (Fig. 21.17) . Bronchial
arteries arise from the aorta and penetrate into the
1529
lungs, traveling with the bronchi. Tetralogy of Fallot and pseudotruncus arteriosus result in a shift to bronchial circulation. Bronchial arteries are also important in Rasmussen aneurysms from tuberculosis and systemic hypervascularity of any chronic infection. Pulmonary arterial hypertension (Fig. 21.18) results in (1) dilated main pulmonary artery, (2) right-sided cardiac enlargement, (3) central enlargement of left and right pulmonary arteries, (4) rapid pruning of the peripheral pulmonary arteries, (5) decreased peripheral pulmonary circulation, (6) calcification of the central pulmonary arteries, and (7) secondary enlargement of the azygos vein. Pulmonary
aneurysms and peripheral pulmonic stenosis can also
cause unusual enlargements of the pulmonary arteries and may be seen in Williams syndrome, Marfan syndrome, and collagen disorders. Pulmonary
venous
hypertension (Fig. 21.19) results from mitral
stenosis, mitral regurgitation, or elevated left P.614 ventricular pressure (aortic stenosis or CHF). The normal vessel caliber in the lower lobes is greater than that in the upper lobes by a 3:2 ratio because of hydrostatic pressure and the high compliance of the venous system. Elevated venous pressure causes progressive, edematous perivascular cuffing, which occurs first in the lower vessels, which have higher hydrostatic pressures. Perivascular edema in the lower lobes results in decreased compliance and progressive cephalization of blood flow. The chest radiograph will show decreased caliber of the lower lobe vessels and increased caliber of the upper lobe vessels. Cephalization of blood flow is the earliest radiographic sign of CHF and pulmonary venous hypertension. Cephalization begins at 10 to 13 mm Hg wedge pressure. Equalization of upper to lower pulmonary blood flow occurs at 14 to 16 mm Hg. Reversal of the normal distribution, with the upper lobe vessels distended and the lower lobe vessels constricted, occurs at 17 to 20 mm Hg. Hilar fullness, the “Viking helmet sign― in the hila, and filling out of the right hilar angle commonly accompany reversed flow distribution.
1530
FIGURE
21.17. Chest Radiograph on Patient With Tetralogy
of Fallot. Asymmetric blood is evident with increased flow to the right. Note also right ventricular hypertrophy configuration and concave
pulmonary
artery
segment.
1531
FIGURE
21.18. Idiopathic
Pulmonary
Hypertension. The
main (curved arrow), right, and left (arrowheads) pulmonary arteries are dilated. The pulmonary arteries taper rapidly and peripheral pulmonary vascularity is decreased.
1532
FIGURE 21.19. Pulmonary Venous Hypertension. Cephalization of blood flow is evident in this patient with mitral stenosis and enlarged left atrial appendage (curved arrow). The lower lobe vessels are constricted, and the upper lobe vessels are distended. Fullness in the hilar angle (straight arrow) is caused by enlargement of the superior pulmonary veins crossing between the interlobar artery and the upper lobe artery.
Pulmonary
Edema
Interstitial edema with Kerley A, B, and C lines and thickened pulmonary fissures occurs at 20 to 25 mm Hg wedge pressure (Fig. 21.20). Kerley lines represent thickened interlobular septa: A lines are long, straight lines radiating toward the hila; B lines are horizontal lines connecting to the pleural surface near the costophrenic angle; and C lines are random reticular lines seen throughout the lungs. Alveolar edema begins at 25 to 30 mm Hg wedge pressure (Fig. 21.21). Chronic failure “toughens― the
1533
interstitium (often resulting in hemosiderosis and pulmonary ossification) and can add an additional protective zone of 5 mm Hg prior to the development of interstitial or alveolar edema. These P.615 progressive signs of failure have been classified as stages 1 to 4 (Table 21.3) .
FIGURE
21.20. Interstitial
Edema. The edema is indicated by
prominent Kerley lines. Thickening of the fissures (arrow) is also present, along with prominence of the LV and LA and cephalization of blood flow.
Congestive
Heart
Failure
Radiographic findings include (1) cardiomegaly; (2) left ventricular and left atrial enlargement; (3) cephalization of blood flow; (4) azygos vein and SVC distention; (5) perivascular cuffing with haziness and unsharpness of the pulmonary vessels; (6) peribronchial cuffing with thickening of the bronchial walls, seen as small “Cheerios― when viewed end on; (7) Kerley lines; (8)
1534
thickening of the pulmonary fissures; (9) subpleural edema; (10) pleural effusions, usually larger in the right hemithorax; and (11) alveolar edema in a “bat's wing― or “butterfly― distribution, also often more pronounced on the right.
FIGURE 21.21. Alveolar Pulmonary Edema. Classic bat's wing or butterfly perihilar alveolar infiltrates are present in a symmetric cloudlike pattern.
1535
TABLE 21.3 Signs of Progressive Cardiac Failure
Stage
Sign
Wedge Pressure (mm Hg)
1
Progressive
2
Interstitial edema and septal lines
20–25
3
Alveolar edema, often in bat's wing
>25–30
perihilar
4
cephalization
10–20
distribution
Chronic or severe pulmonary venous hypertension resulting in
>30–35
hemosiderosis, pulmonary ossification, and chronic interstitial disease such as from long-standing mitral
Right
Heart
stenosis
Failure
The most common cause of right heart failure is left heart failure. Elevated left-sided pressures manifest in the pulmonary circuit and then in the right side of the heart. Long-standing venous hypertension leads to pulmonary arterial hypertension. Elevated right-sided pressures cause right ventricular hypertrophy and dilatation, as well as systemic venous dilatation involving the azygos vein, the SVC, and the jugular veins. Dilatation of the right heart can also cause tricuspid valve incompetence. Right heart failure protects the pulmonary circuit by accumulating edema and fluid outside the lungs, similar to the old therapeutic maneuver of rotating tourniquets.
1536
Right heart failure may also occur with the dilated cardiomyopathies, including viral and alcoholic cardiomyopathy. When right heart failure is the result of pulmonary disease, such as chronic obstructive pulmonary disease, destructive lung disease, or primary pulmonary hypertension, the term cor pulmonale is used.
The
Pericardium
The pericardium is composed of one continuous fibrous membrane that is folded back on itself, creating two layers. The inner layer of visceral pericardium, or epicardium, is closely attached to the myocardium and subepicardial fat. The outer layer or parietal pericardium is thicker and is often referred to simply as the pericardium.
Pericardial
Effusion
Between the visceral and parietal layers is the pericardial space, which usually contains 20 mL of serous fluid. More than 50 mL of fluid is clearly abnormal, but a volume of about 200 mL is required for detection by plain film radiography. Mediastinal and epicardial fat enable the pericardium to be visualized as a thin arcuate line paralleling the anterior heart border in the retrosternal region. A pericardial stripe exceeding 2 to 3 mm is indicative of pericardial thickening or effusion. Unfortunately, the thickened pericardial stripe can be seen on the P.616 lateral radiograph in only about 15% of patients with pericardial effusion. The “differential density sign― refers to a lucent margin along the left heart border on the PA radiograph or along the posterior cardiac border on the lateral radiograph. It is seen in up to 63% of patients with pericardial effusion but is less specific than the thickened pericardial stripe. Large pericardial effusions cause the heart to appear on frontal radiographs in the shape of a sac of water sitting on a tabletop (Fig. 21.22) .
1537
FIGURE
21.22. Pericardial
Effusion.
“Water-bottle
configuration― of the cardiac silhouette is indicative of pericardial effusion or dilated cardiomyopathy. This patient with systemic lupus erythematosus has an enlarged azygos vein (arrow), decreased pulmonary vasculature, and clear lung parenchyma.
Pneumopericardium appears on plain films as radiolucency surrounding the heart and separated from the lung by a thin white line of pericardium (Fig. 21.23). Air may also be seen outlining the pulmonary arteries or the undersurface of the heart. Pneumopericardium can be caused by trauma, infection, or pneumomediastinum. Firm attachment of the pericardium to the ascending aorta just above the main pulmonary artery acts to contain the pneumopericardium.
Other Signs of Cardiac Disease
1538
Situs
Anomalies
Careful attention should be directed at the location of the aortic arch, gastric fundus, heart, pulmonary fissures, and the branching pattern of the bronchi. Normal anatomic positioning is termed situs solitus. Situs inversus means that the patient's entire anatomic arrangement is reversed in a right-to-left direction as a “mirror image.― Situs inversus is associated with a 5% to 10% incidence of congenital heart disease, compared with less than 1% incidence for situs solitus. Dextrocardia indicates that the heart is in the right hemithorax. The apex of the heart lies to the right, with the long axis of the heart directed from left to right. Kartagener syndrome is a combination of situs inversus with dextrocardia, bronchiectasis, and sinusitis (Fig. 21.24). The latter findings are caused by the abnormal mucosal
cilia.
FIGURE 21.23. Pneumopericardium. Air within the pericardial sac enables visualization of the pericardium (arrowheads), seen as a thin white line paralleling the left heart border.
Dextroposition means the heart is shifted toward the right
1539
hemithorax. It is associated with hypoplastic right lung and an increased incidence of congenital heart P.617 disease, particularly left-to-right shunts. Dextroversion means the cardiac apex is to the right, but the stomach and aortic knob remain on the left. The LV remains on the left but lies anterior to the RV.
FIGURE
21.24. Kartagener
Syndrome. Situs inversus is
evident with dextrocardia and the gastric air bubble (black arrow) on the patient's right. Evidence of bronchiectasis is present behind the heart and in the left lower lobe (arrowheads) .
Dextrocardia with situs ambiguous and polysplenia is also called “bilateral left-sidedness.― Each lung contains only two lobes and hyparterial bronchi. Bilateral SVCs are also common. The incidence of congenital heart disease is increased, most commonly that of atrial septal defect or anomalous pulmonary venous return.
1540
Dextrocardia with asplenia is referred to as “bilateral rightsidedness― because of bilateral minor fissures and three lobes in each lung. The cardiac anomalies are usually more complex and severe than in polysplenia.
Bony
Abnormalities
Postoperative changes of sternotomy suggest and the presence of cardiac disease. Sternal vehicle accidents are associated with a 50% contusion. Hypersegmentation of the sternum
prior cardiac surgery fractures from motor incidence of cardiac (more than four to five
segments) is present in 90% of patients with Down syndrome and offers a clue to the presence of endocardial cushion defect or complete atrioventricular canal. Wavy retrosternal linear opacities suggest dilated internal mammary arteries associated with coarctation of the aorta. Pectus excavatum is associated with an increased incidence of mitral valve prolapse and Marfan syndrome. A barrel-shaped chest with pectus carinatum is associated with ventricular septal defects and complete atrioventricular canal. Scoliosis with a “shield chest― is seen with Marfan syndrome, aortic valve disease, coarctation, and aortic dissection. The presence of 11 or fewer ribs is highly associated with Down syndrome and atrioventricular canal. “Ribbon ribs― or bifurcated ribs and an overcirculation pattern suggest truncus arteriosus, whereas their association with a pattern of undercirculation suggests tetralogy of Fallot. Rib notching and inferior rib sclerosis indicate collateral circulation through intercostal arteries and occur with coarctation of the aorta and Blalock-Taussig operations. The third through the eighth ribs are most commonly involved. Fractures of the first and second ribs indicate that highvelocity blunt trauma has occurred, and there is an increased risk of aortic injury. The spine offers clues to the presence of aortic valve disease when changes of ankylosing spondylitis, neurofibromatosis, or rheumatoid arthritis are present. Scoliosis is associated with an increased incidence of congenital heart disease.
1541
NUCLEAR
CARDIOLOGY
Cardiac nuclear medicine is a central modality in cardiac imaging and is covered in detail in Chapter 57. Perfusion scans with thallium or new technetium agents are useful for diagnosing coronary ischemia and myocardial infarcts. Normal perfusion scans appear in the shape of a horseshoe in the vertical and long axes and in the shape of a doughnut in the short axis (see Fig. 57.2). The scans are accomplished during rest, with controlled exercise, or with pharmacologic stress with IV dipyridamole. The stress and redistribution or rest images appear identical in normal patients. Hypoperfused segments on stress images that fill in on rest are indicative of ischemia. Hypoperfused segments on both rest and stress images are usually infarcts or scars. Myocardial infarction scanning can be accomplished using rest perfusion agents for “cold spot― imaging or technetium pyrophosphate for “hot spot― imaging (see Figs. 22.8, 22.9). Antimyosin antibody scans have also been utilized for diagnosing and sizing myocardial infarction. Electrocardiogram (ECG)-gated myocardial blood pool studies examine wall motion and allow calculations of left ventricular ejection fraction
(see Figs.
57.10, 57.11).
Ventricular
function,
aneurysms,
and valvular disease may be studied with volume curves and functional images. Right ventricular ejection fraction calculations require first-pass examinations because of anatomic overlap of the RV with the atria in the left anterior oblique projection. First-pass cardiac studies can also diagnose SVC obstruction and left-to-right cardiac shunts. Right-to-left cardiac shunts can be evaluated and quantified with technetium macroaggregated albumin or microspheres. SPECT imaging has greatly improved the diagnostic capabilities of myocardial perfusion imaging and infarct scans. ECG-gated SPECT is readily accomplished and adds wall motion evaluation, ventricular volumes, and ejection fraction information to the study as well. PET is a newer technology with increased resolution compared to SPECT imaging. PET can assess cardiac metabolism as well as perfusion,
1542
enhancing its ability to evaluate cardiomyopathies, ischemia, infarction, and “hibernating― or viable myocardium.
ECHOCARDIOGRAPHY Echocardiography includes M-mode, real-time two-dimensional US, range-gated and color-flow Doppler, and transesophageal US. Transesophageal echocardiography uses a nasogastric probe with a steerable ultrasonic beam that views the heart and aorta from the close posterior position provided by the esophagus (Fig. 21.25). Mmode echocardiograms are produced by a narrow ultrasonic beam that is directed at cardiac structures and observed over time or is swept across an area of anatomy (see Figs. 21.26, 21.27, 21.28) . The returning echoes produce a time–motion study of cardiac structures. With a transthoracic technique, anterior structures are usually displayed at the top of the image. The thickness and motion of the myocardium can be evaluated throughout the cardiac cycle. Pericardial effusions are shown as echo-free spaces adjacent to the myocardium
(Fig. 21.26). Large pleural P.618 P.619
effusions create an echolucent space posterior to the LV and pericardium.
1543
FIGURE
21.25. Transesophageal
Echocardiogram. A five-
chamber view of the heart is provided by an US probe within the esophagus. The probe is behind the LA and is depicted at the top of the image. All four chambers and the aortic valve are seen in one plane, the “five-chamber view.― The LA and the RA are separated by the interatrial septum. The aortic valve (A) is readily identified in the midplane. The RV and the LV are separated by the interventricular septum. The tricuspid valve (curved arrow) is seen between the RA and the RV, and the mitral valve (straight arrow) is seen between the LA and the LV.
1544
FIGURE 21.26. Pericardial Effusion. An M-mode echocardiogram with the ultrasonic probe at the top of the image demonstrates the RV, interventricular septum (curved arrow) , and LV. Note the normal myocardial contractility with the interventricular septum contracting toward the posterior left ventricular wall during systole. A pericardial effusion (PEff) is seen as an echolucent space posterior to the left ventricular wall.
1545
FIGURE
21.27. Aortic
Root. An M-mode echocardiogram
demonstrates anterior movement of the anterior (curved arrow) and posterior (straight arrow) walls of the aortic root during systole. The RV is seen anterior to the aortic root and the LA is seen posterior to the aortic root. Aortic valve motion can be seen within the aortic root.
The interventricular septum appears as a band of echoes near the midplane. It normally thickens and moves toward the posterior wall of the LV during systole (Fig. 21.26). Paradoxic septal motion may be seen in pericardial effusion, with cardiac tamponade, chronic obstructive pulmonary disease, asthma, atrial septal defects pulmonary hypertension, left bundle-branch block, and septal ischemia. The interventricular septum measures less than 10 to 11 mm at end-diastole and is compared with the thickness of the posterior wall of the LV for asymmetric or concentric hypertrophy.
1546
FIGURE 21.28. Normal Mitral Valve. An M-mode echocardiogram demonstrates the right ventricular cavity and left ventricular cavity separated by a band of echoes representing the interventricular septum (arrow). The moving mitral valve can be seen within the left ventricular cavity. Because of the plane of section, the full systolic motion of the myocardium is not well visualized. The points of the mitral waveform are labeled with letters.
The aortic root lies immediately posterior to the RV and measures 8 to 12 mm in neonates and 20 to 40 mm in adults (Fig. 21.27). The thin parallel aortic walls move anteriorly during systole. The aortic root is dilated in aortic stenosis, aortic insufficiency, tetralogy of Fallot, and aortic aneurysm. The thin aortic cusps seen within the aortic root should open widely during systole and should not reverberate. The LA is seen posterior to the aortic root (Fig. 21.27). The normal
1547
size is no larger than 40 mm during diastole in adults. The LA is free of internal echoes and has a thin posterior wall that merges with the thicker left ventricular wall. The LV lies inferior and lateral to the LA and is an echo-free space except for the thin chordae tendineae and the echogenic projections of the papillary muscles. The left ventricular posterior wall thickens during systole and contracts anteriorly. The transverse diameter of the LV does not normally exceed 5.7 cm during diastole. The wall measures approximately the same as the ventricular septum (10 to 11 mm). The mitral valve produces a saw-toothed or M-shaped pattern posterior to the interventricular septum (Fig. 21.28). The anterior leaflet is the dominant echo and is continuous with the posterior wall of the aortic root. Immediately posterior to the anterior leaflet is the W-shaped pattern of the posterior leaflet. The two leaflets close during systole. The echo pattern of the anterior leaflet should be carefully scrutinized for evidence of thickening, delay in closure (seen with mitral stenosis), vegetations, prolapse, myxoma, or highfrequency vibration secondary to aortic regurgitation (Austin Flint phenomenon). The specific points of the mitral waveform are (Fig. 21.28) the following: A point: Atrial contraction with peak anterior opening motion B point: notch Between the A and C points representing elevated left
ventricular
end-diastolic
pressure
C point: Closure of the mitral valve occurs with contraction of the LV during systole D point: early Diastole when mitral valve begins to open E point: maximal Excursion of the valve opening. This is the peak of early diastolic opening and the most anterior position of the valve during diastole. F point: most posterior point of early diastolic Filling prior to atrial contraction
1548
The E–F slope is a function of the left atrial emptying rate and should be steep. With mitral stenosis, the slope will be flattened and look more squared off than M-shaped. With valve thickening and calcification,
the
squared-off
part
appears
thickened.
The tricuspid valve is identified by locating the mitral valve and rotating the transducer medially. It has an M-shaped echo pattern similar to that of the mitral valve. The E–F slope is decreased with tricuspid stenosis and is increased with Ebstein anomaly, tricuspid regurgitation, and atrial septal defect. The pulmonic valve is rather difficult to evaluate by M-mode echocardiography. The diameter of the pulmonary trunk is similar to that of the aortic root. Pulmonary valve motion is similar to aortic valve motion, except that only the posterior leaflet is well seen and there may be a small “A wave― because of atrial contraction.
CORONARY
ANGIOGRAPHY
While CT coronary angiography (CTCA) is playing an increasingly important role, true coronary angiography will remain vitally important, especially in preparation for coronary intervention. Selective catheterization of the coronary arteries was first accomplished in 1959 by Sones with the use of a flexible, tapered-tip catheter using a cut-down procedure on the brachial artery. In 1966, Amplatz used J-shaped, preformed catheters with better torque control from a transfemoral approach. In 1968, Judkins used separate preformed catheters for the right and left coronary arteries. After selective catheterization of the coronary artery, hand injections of contrast verify the size and flow of the artery. The left coronary artery generally requires 7 to 9 mL of contrast at 4 to 6 mL/s, whereas 6 to 8 mL at 3 to 5 mL/s is sufficient for the smaller right coronary artery. Pressure limits for power injectors should be set at less than 150 psi. The catheter tip should not be left wedged in the coronary ostium, as this might occlude blood flow. Complications of coronary angiography include hematoma, pseudoaneurysm, and fistula formation at the puncture site; arrhythmias,
including
premature
ventricular
1549
contractions;
heart
block and asystole; myocardial infarction; stroke; emboli; and death. Indications for coronary arteriography include (1) confirmation of an anatomic cause for angina, (2) identification of high-risk lesions, (3) evaluation of asymptomatic patients with abnormal exercise tolerance test or occupational risk, P.620 (4) preoperative evaluation for cardiac surgery, (5) evaluation of patients with coronary artery bypass grafts for stenosis or occlusion, and (6) evaluation of interventional therapy after myocardial infarction.
FIGURE 21.29. Left Coronary Artery (LCA) in the Left Anterior Oblique Projection. The LCA divides into the circumflex artery, which makes up the left side of the circle, and the left anterior descending artery, which makes up the anterior portion of the loop. Obtuse marginal branches extend from the circumflex artery; diagonal and septal branches extend from the left anterior descending artery. (Reproduced with permission from Kubicka RA, Smith C. How to interpret coronary arteriograms. Radiographics 1986;6:661–701.)
1550
Coronary
Anatomy
The right coronary artery (RCA) arises from the right coronary cusp, and the left coronary artery (LCA) arises from the left coronary cusp. Approximately 85% of patients are right dominant, meaning that the RCA supplies the posterior descending artery and the posterior and inferior surface of the myocardium. In 10% to 12% of patients, the LCA is dominant and supplies the inferior and posterior surface. Approximately 4% to 5% of patients are codominant. The LCA measures 0.5 to 1.5 cm in length before it divides beneath the left atrial appendage (Figs. 21.29, 21.30). artery (LAD) extends anteriorly in the circumflex artery extends laterally and atrial appendage to the atrioventricular
The left anterior descending interventricular groove. The posteriorly under the left groove. An occasional third
branch is the ramus intermedius, which extends as a first diagonal branch (d1) or a first marginal branch (m1). The LAD gives off several septal branches that penetrate into the septum. One or more diagonal branches extend toward the anterolateral wall. Occasionally, a conus branch comes off after the first septal branch and extends to the right ventricular infundibulum. The circumflex artery gives off one or more obtuse marginal branches that supply the lateral wall of the LV. The RCA passes anterior and to the right between the pulmonary artery and the RA (Figs. 21.31, 21.32). Its first branch is a conus branch to the pulmonary outflow tract. The second branch is the sinus node branch, with a smaller branch to the RA. Muscular branches extend into the right ventricular myocardium. At the posterior turn, a large acute marginal branch is often given off anteriorly toward the diaphragmatic surface of the RV. The RCA then extends posteriorly in the atrioventricular sulcus and makes a 90° turn toward the apex in right-dominant systems. As the posterior descending artery, it supplies branches to the diaphragmatic myocardium and the posterior one third of the interventricular septum. The distal RCA may also give off a variable number of posterolateral
ventricular
branches.
1551
The coronary arteries can be visualized as a circle and loop, with the atrioventricular groove being the circle and the interventricular septum being the attached loop (Fig. 21.29, Fig. 21.30, Fig. 21.31, Fig. 21.32). In the right anterior oblique projection, the circle is superimposed on itself and the loop is in profile. In the left anterior projection, the circle is more open and the loop is foreshortened. In the left anterior craniad view, there is a better, elongated view of the left main coronary artery, LAD, and ramus intermedius. P.621
Coronary Fixed
Pathology
Coronary
Stenosis
A 75% reduction in cross-sectional area is required to cause a significant reduction in blood flow (see Fig. 22.4). A 50% reduction in diameter corresponds to a 75% reduction in cross-sectional area. Other significant findings include coronary calcification, ulcerative plaques, and aneurysm formation. Collateral flow typically develops when there is greater than 85% stenosis.
FIGURE Oblique
21.30. Left Coronary Artery in the Right Anterior Projection. The loop is more open in this projection,
1552
whereas the circle is superimposed. The left anterior descending artery makes up the anterior portion of the loop. The circumflex artery and its obtuse marginal branches make up the left side of the circle. (Reproduced with permission from Kubicka RA, Smith C. How to interpret coronary arteriograms. Radiographics 1986;6:661–701.)
FIGURE
21.31. Right Coronary Artery (RCA) in the Left
Anterior Oblique Projection. The right portion of the circle represents the RCA, and the posterior portion of the loop represents the posterior descending artery. S-A, sinoatrial; A-V, atrioventricular. (Reproduced with permission from Kubicka RA, Smith C. How to interpret coronary arteriograms. Radiographics 1986;6:661–701.)
Catheter
spasm is most often seen in the RCA as a smooth
transient narrowing 1 to 2 mm distal to P.622 the catheter tip. The patient usually remains asymptomatic.
1553
FIGURE
21.32. Right Coronary Artery (RCA) in the Right
Anterior Oblique Projection. The RCA forms the atrioventricular circle. The loop is more opened in this projection with the posterior descending artery making up its inferior margin. S-A, sinoatrial. (Reproduced with permission from Kubicka RA, Smith C. How to interpret coronary arteriograms. Radiographics 1986;6:661–701.)
Prinzmetal
variant
angina is angina secondary to prolonged
coronary spasm. IV ergonovine may be used in a provocative test to incite coronary spasm, typical symptoms, and electrocardiographic changes. Prinzmetal angina is usually treated medically. Kawasaki
syndrome is an inflammatory condition of the coronary
arteries, probably attributable to a prior viral syndrome, that results in coronary stenosis and coronary aneurysms. It occasionally persists into adulthood. Myocardial bridging describes a normal variant in which the coronary arteries penetrate and then emerge from the myocardium rather than running along the surface of the epicardium. This causes arterial constriction during systole, which reverts to normal flow during
diastole.
1554
Anomalies of the coronary arteries include multiple coronary ostia with more than one coronary artery arising directly from one coronary cusp, a single coronary artery, and origination of the LCA from the PA (Fig. 21.33) .
Therapeutic
Considerations
The primary modes of therapy for coronary artery disease include many efficacious medical regimens, percutaneous coronary angioplasty and stenting, and coronary artery bypass graft surgery. Coronary artery bypass grafting usually uses saphenous vein grafts or native internal mammary arteries. Surgical bypass has been shown to prolong life in left main coronary artery disease and threevessel disease. Percutaneous coronary angioplasty (see Fig. 22.5) is considered useful for both single-vessel and multivessel disease and has an 85% to 90% initial success rate. Restenosis remains a significant problem in up to 50% of cases, P.623 typically occurring within the first 6 months. Restenosis is less frequent with newer stents. Angioplasty is typically accomplished by balloon dilatation of the stenotic lesion over a guidewire. Angioplasty is considered successful when the stenosis is reduced to less than 50% of diameter narrowing, although long-term prognosis is better when there is less than a 30% residual stenosis. Directional and rotational atherectomy and atherectomy with the transluminal extraction catheter and laser angioplasty are additional percutaneous techniques that are currently used in specific situations.
1555
FIGURE 21.33. Aberrant Left Coronary Artery (LCA). The catheter in the ascending aorta (Ao) opacifies a dilated right coronary artery (RCA) (straight arrow). The LCA (curved arrow) arises from the pulmonary artery (PA) and is filled in a retrograde fashion via collateral flow from the RCA.
CARDIAC
ANGIOGRAPHY
Angiography of the heart in adults most often involves left-sided catheterization via arterial puncture with retrograde examination of the aorta, LV, and LA. Selective catheterization of the coronary arteries is also accomplished from the arterial side. Right heart angiography uses puncture of a neck or femoral vein with catheter placement in the RA, RV, pulmonary outflow tract, or pulmonary artery. Additionally, the LA or LV may be seen on delayed or “levo-phase― views from a right-sided injection. It is also possible to access the left side during right heart catheterization by
1556
puncturing the atrial septum. End-hole catheters are used for pressure measurements, and pigtail or multiple–side-hole catheters are used for intracardiac injections to avoid contrast injection into the myocardium itself. Blood flow is estimated with standard oximetry, thermodilution, and indicator dilution techniques.
FIGURE
21.34. Left
Ventricular
Aneurysm. Diastole (A) and
end-systole (B). The left ventriculogram is accomplished with the pigtail catheter entering the LV from the aortic root (Ao). A paradoxic bulge near the apex (arrowheads) indicates a left ventricular
Wall
aneurysm.
motion is evaluated globally and regionally. Hypokinesia
describes diminished contractility or less systolic motion than normal. Akinesia denotes no systolic wall motion. Dyskinesia means there is paradoxic wall motion during systole. Tardikinesia refers to delayed contractility. Asynchrony refers to cardiac motion that is out of phase with the remainder of the myocardium. Ventricular
aneurysms appear as a bulge in the wall that moves
paradoxically compared with other areas of the LV (Fig. 21.34). True aneurysms are lined by thinned, scarred myocardium and are
1557
typically located near the apex or anterolateral wall. Pseudoaneurysms are focal, contained ruptures that are often larger but have narrower ostia, and they are most commonly located at the inferior and posterior aspect of the LV. Intramural thrombi may be seen in up to 50% of ventricular aneurysms.
CARDIAC
CT
MDCT is useful in evaluating aortic aneurysms, aortic dissections (Fig. 21.35), aortic injuries, vascular anomalies (Fig. 21.36), central pulmonary emboli (Fig. 21.37), intracardiac masses and thrombi (Fig. 21.38), pericardial thickening, fluid collections, and pericardial calcifications. Optimal contrast enhancement, ECG gating, and breath-hold technique are required for optimal studies. Ultrafast or electron beam CT (EBCT) offers the advantage of high-speed scanning to better stop action and eliminate motion artifact. Angled couch views supplement standard axial imaging. With cardiac gating, cine-CT
can P.624
provide wall motion studies, ejection fraction, and valve evaluation.
1558
FIGURE 21.35. Type B Aortic Dissection. Contrast-enhanced multidetector CT scan demonstrates descending aortic aneurysm with intimal flap. The ascending aorta is normal.
Coronary
Artery
Calcium
Screening
with
CT As previously described, coronary artery calcification has been studied extensively with chest radiography and fluoroscopy. Radiography has a sensitivity of 42% and fluoroscopy has sensitivity of 40% to 79% and specificity of 52% to 95% for detecting coronary calcification as an indicator of coronary stenosis. Coronary calcification is a significant marker for underlying atherosclerosis. EBCT has P.625 been studied thoroughly since the early 1990s as a coronary calcification screening modality and has a sensitivity of 70% to 74%, specificity of 70% to 91%, and negative predictive value of 97% when compared to coronary angiography (see Fig. 22.2). Now multidetector CT (MDCT) has been shown to be equivalent to EBCT for coronary calcification detection and scoring.
1559
FIGURE 21.36. Aberrant Left Pulmonary Artery. Contrastenhanced multidetector CT demonstrates anomalous origin of left pulmonary artery (arrow) from the right pulmonary artery, crossing posterior to the trachea, consistent with a pulmonary sling.
1560
FIGURE 21.37. Pulmonary Artery Embolus. Contrastenhanced CT examination shows a filling defect within the right pulmonary artery (arrowheads). Disappearance with time confirms an embolus. Primary neoplasms or metastatic emboli could cause a similar filling defect.
1561
FIGURE
21.38. Ultrafast CT (Electron Beam CT [EBCT]).
Contrast-enhanced EBCT shows intraventricular clot (black arrow), thinned myocardium (white arrow), and akinesis, secondary to anteroapical infarct. (Courtesy of William Stanford, MD.)
EBCT allows 40 to 60 sections, each 1.5 to 3.0 mm, with an exposure time of 100 ms, with single breath-hold acquisition and ECG gated to end-diastole. New 16-slice, and now 64-slice MDCT equipment has taken rotation speeds down to 330 ms (0.33 sec) and resolution to 0.33 mm. MDCT coronary calcium screening is also done with ECG gating, single breath-hold, and arms up. One method of scoring utilizes the Agatston method, in which coronary calcification is defined as an area with greater than 130 Hounsfield units (HU) and larger than 2 mm2 . A score of 1 is given for 130 to 200 HU, 2 for 201 to 299 HU, 3 for 300 to 399 HU, and 4 for 400 HU or greater. This factor is assigned and multiplied by the area of the lesion for each coronary artery territory. This score is then summed for a total coronary calcification score or Agatston
1562
score (Fig. 21.39). A score of 0 to 10 is very low to low risk, 11 to 100 is moderate risk, 101 to 400 is moderately high risk, and greater than 400 is high risk for underlying stenosis and future cardiac events. However, the specific calcified area or artery may not correlate with specific stenoses.
FIGURE 21.39. Coronary Calcification Scoring from Multidetector CT. The report shows the score for each coronary artery and location. The summed score is over 1,100, placing the patient in the very-high-risk category.
The utility of coronary calcium screening lies in (1) early detection of calcium in asymptomatic patients for risk stratification and risk factor modification, (2) evaluation of progression or even regression of calcification as a indicator of atherosclerotic coronary disease, (3) its ability to demonstrate the absence of calcification, thereby essentially ruling out significant underlying coronary stenosis.
1563
CT
Coronary
Angiography
EBCT and now MDCT have been also shown to be efficacious for noninvasive CTCA. Many laboratories have utilized 16-slice MDCT and more recently 64-slice MDCT for CTCA. In development are 128-slice, 256-slice, and area-detector technology. Resolution is now down to 0.33 mm with rotation speeds to 300 ms. Because faster heart rates can lead to motion artifact, slowing the heart rate to 60 or 70 bpm with oral and IV beta blockers is necessary. Contrast is delivered utilizing a peripheral or jugular vein, an 18- to 20-gauge needle, and 100 to 150 mL of isosmolar contrast at 4 mL/sec. The study is acquired with arms up, single breath-hold (10 to 30 sec) and ECG gating (prospective or retrospective). The contrast bolus is immediately followed by a saline flush of 25 to 40 mL. The scan timing can be judged with a test P.626 bolus or begin at the end of contrast injection. Optimal image quality has peak opacification in the LV and coronary arteries, with less dense concentration in the RV and pulmonary arteries.
1564
FIGURE 21.40. Three-Dimensional Volume-Rendered CT Coronary Angiogram. The left anterior descending (LAD), branching diagonals, and circumflex coronary arteries are well visualized in this left anterior oblique projection from 16-slice multidetector CT. The left main coronary artery is partially seen under the left atrial appendage.
ECG “pulsing― can reduce tube current during systole and increase it during diastole where the target images are usually constructed. This can reduce the radiation dose by up to 50%. Reconstruction is done to 1-mm slice thickness and a medium smooth reconstruction algorithm. Past processing is very important and is often done by the radiologist, especially for 3D reconstruction. The coronaries can be evaluated for congenital abnormalities,
1565
presurgical anatomy, coronary calcifications, and coronary plaque or stenosis utilizing volume-rendered 3D views (Fig. 21.40), 2D views, multiplanar views (Fig. 21.41), maximal intensity projections, and coronary “straightening― views (Fig. 21.42). Stenoses greater than 50% are considered hemodynamically significant and stenoses greater than 75% are considered high grade. Problems occur in grading stenoses with heavy coronary calcification and with stents. Patency, however, can be determined by evaluating coronary enhancement downstream. CTCA has also been shown to be useful and accurate for the follow-up of coronary artery bypass graft patency.
FIGURE 21.41. Maximum Intensity Projection (MIP) CT Coronary Angiogram. The aortic valve, right coronary artery
1566
(RCA), and posterior descending artery (PDA) are well seen in this left anterior oblique MIP from 16-slice multidetector CT.
FIGURE
21.42. Right Coronary Artery in
“Straightened― Maximum Intensity Projection View. This computer-reconstructed view effectively takes out the curves and makes it easier to see that, while there are atherosclerotic
CARDIAC
irregularities,
there
is
no
significant
stenosis.
MR
Cardiac MR combines many of the capabilities of the other imaging modalities into one examination. These include excellent static anatomic images and dynamic motion studies for function. Cardiac MR applications include congenital heart disease, aortic and pulmonary artery disease, pericardial disease, ventricular function, valvular function, cardiomyopathies, and cardiac masses. Cardiac pacemakers are considered contraindications but most prosthetic valves can be safely studied.
1567
FIGURE
21.43. Spin-Echo
MR. A tomographic slice in the
short-axis projection demonstrates the RV, the interventricular septum, and the LV. The anterior (straight arrow) and posterior (curved arrow) papillary muscles are seen within the left ventricular cavity. The spin-echo technique creates a “black blood― appearance because of the signal void of moving blood.
P.627 The best anatomic depiction is accomplished on spin-echo T1WI in which the moving blood produces a signal void or “black blood― appearance (Fig. 21.43). Gradient-echo or fast-field echo images impart bright signal to coherently flowing blood, creating a “white blood― appearance that is similar to contrast studies (Fig. 21.44). Electrocardiographic gating can be used similar to gated cardiac SPECT and gated cardiac blood pool scintigraphy. Slicespecific information is acquired with reference to specific phases within the cardiac cycle. With gradient-recalled echo technique, motion studies can show flowing blood as well as myocardial
1568
contractility.
FIGURE
21.44. Fast-Field Echo MR. The fast-field echo
technique creates a “white blood― depiction that shows flowing blood and turbulence during motion studies. The enddiastole image (straight
arrow) has the largest ventricular size.
The end-systole image (curved arrow) has the smallest ventricular cavity and the thickest wall.
MR images are acquired as tomographic slices through any selected plane. The planes may be angled to match cardiac (e.g., short axis, four chamber) or vascular anatomy (e.g., left anterior oblique [LAO] aorta). Tissue characterization of the myocardium is accomplished using T1WI and T2WI, contrast enhancement, and spectroscopy. This may be useful for neoplastic, infiltrative, or inflammatory conditions of the myocardium.
1569
Cardiac MR motion studies provide functional information, including wall motion analysis, systolic wall thickening, chamber volumes, stroke volumes, right and left ventricular ejection fractions, and valve evaluation (Fig. 21.45). Flowing blood becomes turbulent and loses its coherence when it passes through stenotic or regurgitant valves. The high-velocity stenotic jet or regurgitant flow is displayed as a wedge-shaped puff of dark turbulent flow, readily identified on the white blood background with the gradient-echo technique (see Figs. 22.21, 22.23). Visual and region-of-interest grading can be accomplished for stenotic or regurgitant flow based on distance, area, or regurgitant volume (see Chapter 22). The regurgitant fraction is calculated by comparing the right and left stroke volumes. Velocity-encoded cine-MR techniques using phase analysis can calculate flow velocities and flow volumes in addition to the regurgitant volumes. These techniques can be used in lieu of angiography for many cases. An understanding of MR signal characteristics and the details of 3D cardiac anatomy displayed in different tomographic planes is critical to the accurate utilization of cardiac MR.
FIGURE 21.45. MR Ejection Fraction Technique. Regions of interest are drawn on the diastolic image (straight arrow) and the end-systolic image (curved
arrow) of each slice. An area
1570
ejection fraction (EF) is then calculated for each slice. Volume ejection fraction calculations are calculated using sequential slices that include the entire ventricular volume. EDV, enddiastolic volume; ESV, end-systolic volume; SV, stroke volume; CO, cardiac output; ED, end-diastole; ES, end-systole.
P.628
Suggested
Readings
Becker CR, Jakobs TF, Aydemir S, et al. Helical and single-slice conventional CT versus electron beam CT for the quantification of coronary artery calcification. AJR Am J Roentgenol 2000;174:543–547. Bogaert J, Dymarkowski S, Taylor AM, eds. Clinical Cardiac MRI. Berlin, New York: Springer-Verlag, 2005. Budoff MJ, Achenbach S, Duerinckx A. Clinical utility of computed tomography and magnetic resonance techniques for noninvasive coronary angiography. J Am Coll Cardiol 2003;42(11):1867–1878. Feigenbaum H, Armstrong WF, Ryan T. Echocardiography. 6th ed. Philadelphia: Lippincott Williams & Wilkins, 2004. Gedgaudas E, Moffer JH, Castaneda-Zuniga WR, Amplatz K. Cardiovascular
Radiology.
Philadelphia:
WB
Saunders,
1985.
Higgins CB. Essentials of Cardiac Radiology and Imaging. Philadelphia: JB Lippincott, 1992. Kelley MJ, ed. Symposium on Chest Radiography for the Cardiologist. Philadelphia: WB Saunders, 1983. Cardiol Clin
1571
1(4):543–750. Kubicka RA, Smith C. How to interpret coronary arteriograms. Radiographics 1986;6:661–701. Lardo A, Chronos NA, Fayad ZA. Cardiovascular Magnetic Resonance: Established and Emerging Applications. London: Martin Dunitz; Independence, KY: Taylor and Francis, 2003. Lawler LP, Pannu HK, Fishman EK. MDCT evaluation of the coronary arteries, 2004: how we do it—data acquisition, postprocessing, display and interpretation. AJR Am J Roentgenol 2004;184:1402–1412. Leschka S, Alkadhi H, Plass A, et al. Accuracy of MSCT coronary angiography with 64-slice technology: first experience. Eur Heart J 2005;26(15):1482–1487. Lipton MJ, Boxt LM, eds. Cardiac imaging. Radiol Clin North Am 2004;42:487–697. Marcus ML, Schelbert HR, Skorton DJ, Wolf GL. Cardiac Imaging—A Companion to Braunwald's Heart Disease. Philadelphia: WB Saunders, 1991. Miller SW. Cardiac Imaging: The Requisites. 2nd ed. Philadelphia: Elsevier Mosby, 2005. Netter FH. Atlas of Human Anatomy. The CIBA Collection of Medical Illustrations. West Caldwell, NJ: CIBA-Geigy, 1989. Oudkerk M. Coronary Radiology. New York: Springer-Verlag, 2004.
1572
Reddy GP, Higgins CB, Chao KH, Tung PP. Cardiac MR Imaging. CD-ROM. Philadelphia: Lippincott Williams & Wilkins, 2001. Rensing BJ, Surruys PW, de Feyter PM. CT-based coronary angiography. J Invasive Cardiol 2000;12(1):23–24. Schoepf UJ, Becker CR, Ohnesorge BM, Yucel EK. CT of coronary artery disease. Radiology 2004;232:18–37. Schoepf UJ, Schoepf UJ. CT of the Heart: Principles and Applications. Totowa, NJ: Human Press, 2004. Schoenhagen P, Halliburton SS, Stillman AE, et al. Noninvasive imaging of coronary arteries: current and future role of multidetector
row
CT.
Radiology
2004;232:7–17.
Stanford W, Thompson BH. Imaging of coronary artery calcification. Its importance in assessing atherosclerotic Radiol Clin North Am 1999;37:257–272.
disease.
Stanford W, Thompson BH, Burns, TL, et al. Coronary artery calcium quantification at multi-detector row helical CT versus electron-beam CT. Radiology 2004;230:397–402.
1573
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section V - Cardiac Radiology > Chapter 22 Cardiac Imaging in Acquired Diseases
Chapter
22
Cardiac Imaging in Acquired Diseases David K. Shelton Cardiac disease remains among the most common problems affecting patient morbidity and mortality today, despite many important dietary, pharmaceutical, interventional, and surgical advances. Most acquired cardiac diseases can be classified under six general categories: ischemic heart disease, cardiomyopathies, pulmonary vascular disease, acquired valvular disease, cardiac masses, and pericardial disease. Use of plain film radiography, fluoroscopy, US, CT, MR, nuclear imaging, and angiocardiography must be integrated with knowledge of specific disease processes.
ISCHEMIC Coronary
HEART Artery
DISEASE Disease
Coronary artery disease is the most common cause of mortality in the United States, with approximately one American dying every minute. Six million to 7 million Americans have active symptoms related to ischemic heart disease. Approximately 300,000 coronary artery bypass graft (CABG) surgeries are performed per year in the United States, with a similar number of percutaneous transluminal coronary angioplasties (PTCA). There were 1.83 million cardiac catheterizations in the US in 1999 and it has been estimated to
1574
reach 3 million by 2010. Clinical presentations include (1) stable angina, (2) unstable angina (often preinfarction), (3) acute myocardial infarction, (4) congestive heart failure secondary to chronic ischemia or prior infarction sequelae, (5) arrhythmias, and (6) sudden death. Clinical symptoms are caused by luminal abnormalities of the coronary arteries, including (1) atheromatous disease, (2) coronary thrombosis, (3) intraluminal ulceration and hemorrhage, (4) vasoconstriction, and (5) coronary ectasia and aneurysm. Vulnerable plaque is initiated by lipoprotein deposition into susceptible areas of the coronary walls and other arteries. Chronic inflammation elsewhere in the body, as well as in the developing plaque, is associated with cytokine and macrophage activity. A thin fibrous cap develops over the lipid core, and mechanical stress can lead to exposure to the blood products, which can then trigger the thrombotic cascade. A sequence of vulnerable plaque development, sudden rupture, and thrombosis is now known to be the leading cause of myocardial infarction. Risk factors for development of atherosclerotic coronary artery disease include elevated serum cholesterol and C-reactive protein, tobacco
smoking,
diabetes,
hypertension,
sedentary
lifestyle,
obesity, age, male gender, chronic inflammation, and heredity. Aggravating conditions include aortic stenosis, ventricular hypertrophy, cardiomyopathy, coronary embolism, congenital anomalies, Kawasaki syndrome, and anemia. Noninvasive imaging is often used as a screening test. Selective coronary angiography with ventriculography and now CT coronary angiography (CTCA) can be utilized to determine coronary anatomy and to direct the specific therapy. A typical imaging workup includes chest radiography, nuclear medicine myocardial perfusion scans, and consideration for coronary angiography. Indications for coronary P.630 angiography include angina refractory to medical therapy, unstable angina, high-risk occupation (e.g., pilot), and abnormal
1575
electrocardiograms or stress perfusion tests. Coronary angiography is considered following myocardial infarction when PTCA or intracoronary thrombolysis are being deliberated. Additional indications include development of mechanical dysfunction, progressive congestive failure, refractory ventricular arrhythmias, and follow-up after administration of IV thrombolytic agents.
FIGURE
22.1. Coronary
radiograph
demonstrates
Coronary
artery
Artery coronary
Calcification. Lateral chest artery
calcification
(arrow ).
calcification occurs in the intima and is directly
related to advanced atheromatous disease and coronary narrowing (Fig. 22.1 ; see Fig. 21.14 ). Coronary calcification is detected at angiography in 75% of patients with 50% diameter stenosis. Only 11% of men without significant coronary artery disease have coronary calcification. In the asymptomatic population, the detection of coronary calcification has a predictive accuracy of 86%. In symptomatic patients, coronary calcification is seen in 50% of patients with single-vessel disease, 77% of those with two-vessel disease, and 86% of those with three-vessel disease. Fluoroscopically detected coronary calcification in the presence of
1576
angina-like chest pain is associated with coronary stenosis 94% of the time. Overall, fluoroscopic detection of coronary artery calcification has a 73% sensitivity and 84% specificity for symptomatic patients. Exercise-tolerance testing has a sensitivity of 76% to 88% and a specificity of 43% to 77%. Exercise testing with planar thallium imaging has a sensitivity and a specificity of approximately
85%.
Use of electron-beam CT (EBCT) and multidetector CT (MDCT) has improved the sensitivity for detecting coronary artery calcification to approximately 95% (Fig. 22.2 ). Importantly, CT also allows the grading of the severity of coronary calcification and thus can establish risk scores, which can help determine patient risk and allow follow-up after medical intervention. The absence of coronary calcification is associated with a very low risk of significant coronary disease. On the other hand, patient youth and higher calcification scores are associated with a higher risk of underlying coronary artery disease and future cardiac events. The negative predictive value of a zero calcification score is 94% to 100%. With scores greater than 400 there is a sensitivity of 82% and specificity of 62% for predicting an abnormal myocardial perfusion SPECT scan (Table 22.1 ).
1577
FIGURE
22.2. CT
Coronary
Calcification. Calcification is seen in
the left anterior descending artery, with regions of interest (white boxes ) placed for quantification. Electron beam CT, 100 ms, 3-mm slice, without contrast. (Courtesy of William Stanford, MD.)
Myocardial perfusion scanning , which uses thallium, technetium (Tc) 99m-sestamibi, Tc-99m-tetrofosmin, or Tc-99m-teboroxime, is one of the primary imaging modalities for detecting myocardial ischemia. Stress images are obtained with exercise or pharmacologic agents such as adenosine dipyridamole. SPECT has increased the sensitivity to 90% to 94% and the specificity to 90% to 95%. The hallmark for segmental ischemia is a perfusion defect on stress testing that fills in during rest (Fig. 22.3 ). A defect that appears stable during both stress and rest examinations is usually an infarction. “Hibernating― regions of viable myocardium associated with tight coronary stenosis may appear as fixed defects on sestamibi or tetrofosmin images or on redistribution thallium images obtained 4 hours after stress. Stress echocardiography using either exercise or pharmacologic stress modalities has also become a widely accepted method to
1578
detect significant (>50% to 70%) P.631 coronary artery stenosis. With the advent of digital image acquisition and cine-loop playback, prestress echocardiographic views can be simultaneously compared with views taken either immediately postexercise or at peak pharmacologic doses. Development of new segmental wall motion abnormalities or worsening of resting abnormalities suggests stress-induced ischemia. One advantage of these techniques is that they also allow prior assessment of resting wall motion abnormalities that are consistent with either profoundly ischemic, stunned, hibernating, or infarcted
myocardium.
0 No identifiable atherosclerotic plaque; very low cardiovascular disease; less than 5% chance of presence of coronary artery disease A negative examination 1–10 Minimal plaque burden Significant coronary artery disease very unlikely 11–100 Mild plaque burden Likely mild or minimal coronary stenosis 101–400 Moderate plaque burden Moderate nonobstructive coronary artery disease highly likely Over 400 Extensive plaque burden High likelihood of at least one significant coronary stenosis (>50% diameter) The summed coronary calcification score can be assigned a percentile ranking for sex and age as well as a risk statement. Appropriate clinical response depends on other risk factors as well.
1579
Calcium
Score
Interpretation
TABLE 22.1 Coronary Calcium Scoring
FIGURE
22.3. Myocardial
Perfusion
Scan. SPECT images of the
LV in short axis projection demonstrate a defect (arrows ) in the inferior wall of the LV during stress, which is well perfused on the rest images. This is strong evidence of ischemic heart disease when using
technetium-99m-sestamibi
dipyridamole
for
pharmacologic
as
the
stress
radionuclide
and
testing.
The overall sensitivity of exercise echocardiography is 76% to 97% using pharmacologic stress agents; the sensitivity is 72% to 96% with dobutamine, approximately 85% with adenosine, and 52% to 56% with a standard dose of dipyridamole. The sensitivities for these tests are lowest for single-vessel disease and improve incrementally for two- and three-vessel disease. Stress echocardiography has a specificity of 66% to 100%. Gated blood pool scintigraphy will demonstrate exercise-induced wall motion abnormalities in 63% of patients with significant coronary artery disease. With exercise, the ejection fractions
1580
normally increase by at least 5%. Failure of ejection fraction to increase with exercise is an indication of myocardial dysfunction. Using these two findings, exercise gated blood pool scintigraphy has a sensitivity of 87% to 95% and a specificity of 92% for detection of coronary artery disease. Coronary angiograms and CTCA (Fig. 22.4 ) should be evaluated for the percent of stenosis, the number of vessels P.632 involved, focal versus diffuse disease, coronary anatomy, ectasia or aneurysm, coronary calcification, and collateral flow (Fig. 22.5 ). Collaterals may include epicardial, intramyocardial, atrioventricular, intercoronary, or intracoronary vessels (i.e., “bridging collateral―). The angiographer must count the number of major epicardial vessels with greater than 50% diameter narrowing. Patients are divided into one-vessel, two-vessel, or three-vessel disease on the basis of involvement of the right or left main coronary artery, left anterior descending artery and left circumflex artery. A 50% diameter narrowing roughly predicts a 75% crosssectional area reduction, which is the physiologic point at which flow is restricted enough to result in ischemia under stress conditions. Reliability for estimating the percent diameter narrowing depends on the observer, projection, resolution, and presence of coronary calcification or ectasia. The degree of coronary disease may be assessed using percent stenosis of each individual coronary artery or of 5-mm segments of the coronary arteries. The right coronary artery is 10 cm long, the left main coronary artery is 1 cm long, the left anterior descending (LAD) is 10 cm long, and the left circumflex is 6 cm long, for a total of 27 cm. These may be divided into fiftyfour 5-mm segments. This scoring system allows the interpreter to quantify the number of 5-mm segments with stenoses in the 0% to 25%, 25% to 50%, 50% to 75%, and 75% to 100% ranges. The significance of 30% to 70% lesions is often clarified by correlation with stress-induced myocardial perfusion scintigraphy.
1581
FIGURE 22.4. CT Coronary Angiogram of Left Anterior Descending Artery (LAD). Left anterior oblique view of maximum intensity projection from 16-slice multidetector CT, CT coronary angiogram demonstrates focal soft plaque in the proximal LAD with 70% stenosis. Percutaneous transluminal coronary angioplasty was subsequently performed.
1582
FIGURE
22.5. Coronary Stenosis. A. An 80% stenotic lesion
(arrow ) is identified in the left anterior descending artery (LAD) on coronary angiography. This patient was experiencing classic angina. B . Marked improvement in the LAD lesion (arrow ) is evident following percutaneous transluminal coronary angioplasty. The angina
symptoms
Percutaneous
resolved.
transluminal
angioplasty
has
traditionally
been
reserved for localized lesions in one- or two-vessel disease (Fig. 22.5 ), but recent published series comparing PTCA with CABG in multivessel disease revealed no difference in the endpoints of death and myocardial infarction. The PTCA group, however, required a significantly higher number of repeat procedures during follow-up, although this has improved with more frequent use of stents. CABG with the use of saphenous vein grafts or internal mammary arteries is usually reserved for more complex or longer-segment disease. CABG markers are usually placed at the anastomotic site to help the angiographer during future selective angiography. Use of the internal mammary artery has better long-term results than saphenous vein grafts and has been correlated with increased survival. Recurrence of symptoms after CABG may be because of occlusion, graft stenosis, or progression of native vessel disease. Graft stenoses and acute occlusions may be amenable to percutaneous interventional techniques. Grafts and stents can be readily evaluated with CTCA (Fig. 22.6 ), although the metallic stent can cause imaging problems.
1583
Echocardiography is useful in detecting some of the long-term complications of ischemic disease, including ventricular aneurysm, thinning of myocardium, akinesia, or dyskinesia. Aneurysms are best seen at the apex and septum. Mural thrombi may also be diagnosed but are difficult to visualize at the apex. Stress echocardiography with either exercise or pharmacologic stress techniques is increasingly used to evaluate for ischemia. CTCA is capable of establishing the patency of CABGs. Ultrafast CT (EBCT) and now MDCT have 93% sensitivity, 89% specificity, and 92% accuracy for establishing patency of CABGs. EBCT and MCDT have also shown to be extremely sensitive for detecting coronary calcification. EBCT and MDCT with contrast can also evaluate wall motion, thrombi, old infarcts, aneurysm, and pericardial abnormalities. P.633 M R can be used (1) to define the location and size of previous myocardial infarctions, (2) to demonstrate complications of previous infarctions, (3) to establish the presence of viable myocardium for possible revascularization, (4) to differentiate acute versus chronic myocardial infarction, (5) to evaluate regional myocardial wall motion and systolic wall thickening (Fig. 22.7 ), (6) to demonstrate global myocardial function with right ventricular and left ventricular ejection fractions, (7) to demonstrate regional myocardial perfusion, and (8) to evaluate papillary muscle and valvular abnormalities. Gadolinium-enhanced T1WIs demonstrate areas of ischemia and reperfusion after myocardial infarction. MR with spectroscopy targeting myocardial phosphate metabolism can distinguish acute from chronic ischemia and reperfused infarcted myocardium from reperfused viable myocardium. With spin-echo imaging, MR has a 78% accuracy for establishing patency of CABGs. Cine MR with gradient echo has sensitivity of 88% to 93%, specificity of 86% to 100%, and overall accuracy of 89% to 91% for patency of CABGs.
1584
FIGURE 22.6. CT Coronary Angiogram (CTCA) of Left Circumflex. Left anterior oblique view of 16-slice multidetector CT, CTCA maximum intensity projection shows patent coronary stent with good flow and no evidence of obstruction. Coronary calcification is also evident downstream.
Myocardial
Infarction
After acute infarction, the chest radiograph will initially show a normal heart size in 90% of cases. Cardiomegaly and congestive failure will eventually develop in 60% to 70% of cases, more frequently with anterior wall infarction, multivessel disease, or left ventricular aneurysm. Increasing stages of pulmonary venous hypertension, particularly alveolar edema, are associated with worsened
prognosis.
1585
FIGURE
22.7. Wall Motion MR Evaluation.
Short-axis
tomographic views of the LV are used for evaluation of systolic wall thickening. Regions of interest are drawn around the myocardium in diastole (left ) and systole (right ). The inferior wall (black
arrow )
demonstrates hypokinesia and poor systolic wall thickening. The functional graph (below ) confirms the findings (white arrow ). The patient had a previous inferior wall myocardial infarction.
Complications of myocardial infarction include the following. Cardiogenic shock implies that systolic pressure is less than 90 mm Hg and is typically associated with acute pulmonary edema and worsened prognosis. Atrioventricular block is common, especially after inferior wall infarcts resulting from either ischemia or injury to the atrioventricular nodal branch of the right coronary artery or increased vagal tone. Complete heart block occurs with larger infarcts and has a worse prognosis. Right ventricular infarction occurs in approximately 33% of inferior wall infarctions. Symptoms are caused by the reduction in right ventricular ejection fraction, which returns to normal within 10
1586
days in approximately 50% of cases. The diagnosis may be established using technetium pyrophosphate (PYP) radionuclide scans. Complications include cardiogenic shock, elevated right atrial pressure, and decreased pulmonary artery pressure. Right precordial electrocardiographic leads can also assist in making the diagnosis. Myocardial
rupture (3.3% of infarcts) may occur 3 to 14 days after
infarction. The mortality rate approaches P.634 100% and accounts for 13% of myocardial infarction deaths. The chest radiograph shows acute cardiac enlargement secondary to leakage of blood into the pericardium. Rupture of the interventricular septum (1%) typically occurs between days 4 and 21, usually as a complication of anterior myocardial infarction and LAD disease. Mortality is 24% within 24 hours and 90% within 1 year. Swan-Ganz catheter measurements show an acute increase in saturation in the RV, although the wedge pressures may be normal. Chest radiographs show acute pulmonary vascular engorgement and right-sided cardiac enlargement because of left-to-right shunt. Pulmonary edema is not a typical feature. Echocardiography readily demonstrates the septal defect. Papillary muscle rupture (1%) is suggested by abrupt onset of mitral regurgitation, with acute pulmonary edema on the radiograph. Typically, the LV is only minimally enlarged, whereas the LA enlarges quickly. Inferior infarcts are associated with posteromedial papillary rupture. Anterior infarcts less commonly affect the anterolateral papillary muscle. Mortality is 70% within 24 hours and 90% within 1 year. Echocardiography confirms the diagnosis. Ventricular aneurysm develops in approximately 12% of survivors from myocardial infarction. Ventricular aneurysms may also be caused by Chagas disease or trauma and are rarely congenital—usually seen in young black males. Aneurysms present with congestive failure, arrhythmias, and systemic emboli. True aneurysms are broad-mouthed, localized outpouchings that do not contract during systole (see Fig. 22.33 ). They are typically anterior or apical and result from LAD disease. The chest radiograph shows a
1587
localized bulge along the left cardiac border and may show rimlike calcification in the wall (Fig. 22.8 ). Fluoroscopy detects up to 50% of cases, whereas 96% are detected by radionuclide ventriculography or myocardial perfusion scan. Echocardiography, contrast-enhanced CT, and MR are also accurate at detecting true aneurysms. Pseudoaneurysms are contained myocardial ruptures consisting of a localized hematoma surrounded by adherent pericardium. Causes include infarction and trauma. Patients are at high risk for delayed rupture. Pseudoaneurysms are typically posterolateral or retrocardiac in location and have smaller mouths than true aneurysms. MR is the most accurate at detecting pseudoaneurysms, but they can also be seen with echocardiography. Dressler
syndrome (4%) is also known as the postmyocardial
infarction syndrome and is similar to the postpericardiotomy syndrome complicating cardiac surgery. Onset is typically 1 week to 3 months postinjury (peak at 2 to 3 weeks), but relapses occur up to 2 years later. Presentation includes fever, chest pain, pericarditis, pericardial effusion, and pleuritis, with pleural effusion usually more prominent on the left. Dressler syndrome responds well to antiinflammatory
medications.
1588
FIGURE
22.8. Left
Ventricular
Aneurysm. A localized calcified
bulge (arrow ) is seen along the left heart border, secondary to prior myocardial infarction complicated by left ventricular aneurysm.
Infarct
Imaging
The indications for myocardial infarct imaging include late admission, equivocal enzymes, equivocal electrocardiogram, recent cardiac surgery or trauma, and suspicion of right ventricular infarction.
Radionuclide
Imaging
“Cold spot― imaging is accomplished with thallium or technetium perfusion agents (Fig. 22.9 ). Sensitivity is 96% within 6 to 12 hours but is P.635 only 59% for remote infarction. Acute infarction cannot be distinguished from remote infarction. “Hot spot― infarct imaging uses Tc-PYP (Fig. 22.10 ), Tc-tetracycline, Tcglucoheptonate, indium-111 antimyosin antibodies, or F-18 sodium
1589
fluorine. Pyrophosphate uptake occurs in myocardial necrosis as a result of PYP complexing with calcium deposits. The PYP scans turn positive at 12 hours, have peak sensitivity at 48 to 72 hours, and revert to normal by 14 days. Persistent abnormal uptake implies a poor prognosis or developing aneurysm. Cardiomyopathies and diffuse myocarditis show diffuse increased uptake. Contusions and radiation myocarditis show increased regional uptake of Tc-PYP.
FIGURE
22.9. Myocardial
Infarction. Resting, planar thallium
image in the left anterior oblique projection demonstrates a defect in the inferoposterior wall (arrow ), consistent with a myocardial infarction. Cold spot imaging can be accomplished almost immediately after the acute event.
1590
FIGURE 22.10. Myocardial Infarct Scan. Hot spot imaging was accomplished using pyrophosphate. Notice the uptake in the anterolateral wall of the myocardium (arrows ), which is almost as “hot― as the sternum (curved arrow ). Images were obtained in right anterior oblique (A) , left lateral (B) , and left anterior oblique (C) projections.
EBCT and MDCT with contrast demonstrate poor perfusion of the infarcted segment immediately after administration of contrast. After a delay of 10 to 15 minutes, the normal myocardium washes out, leaving a contrast-enhanced periphery of the infarcted zone. M R demonstrates prolongation of T1 and T2 times secondary to edema of the acutely infarcted segment. Edema occurs within 1 hour after infarct and may be associated with myocardial hemorrhage. MR has 93% sensitivity, 80% specificity, and 87% accuracy for acute myocardial infarction. The infarcted region is best delineated by high signal on T2WIs; however, surrounding edema tends to overestimate the size of the infarct. T1WIs with gadolinium demonstrate the acutely ischemic region and will help to differentiate reperfusion from occlusive myocardial infarction. Regional wall thinning and lack of systolic thickening are the best evidence of the size of the infarcted segment (Figs. 22.9 , 22.11 ). Scar tissue will not contract, whereas viable myocardium will contract and thicken by at least 2 mm. Very high-grade stenotic lesions may result in chronically ischemic myocardium with altered metabolism. This “hibernating myocardium― may act like postinfarction scar, but it remains viable and may improve in function with revascularization. Unfortunately, it also remains at risk for acute infarction. “Stunned myocardium― describes postischemic, dysfunctional myocardium without complete necrosis, which is potentially salvageable. Echocardiography demonstrates hypokinesis, akinesis, or dyskinesis in previously infarcted myocardial segments; however, this cannot be distinguished from stunned or hibernating myocardium. Global hypokinesis can also be seen with
1591
cardiomyopathic processes. Thinned, hyperechoic walls with wall motion abnormalities suggest transmural scar. Use of echocardiographic
resting
P.636 microbubble contrast can enhance the infarcted region by highlighting perfused areas, resulting in a negative contrast effect at the site of the infarct.
FIGURE 22.11. Old Septal Infarction. Spin-echo MR demonstrates fixed thinning of the myocardial wall (arrow ) attributable to prior myocardial infarction.
CARDIOMYOPATHIES The prevalence of cardiomyopathies is approximately 8 cases per 100,000 people in developed countries. One percent of cardiac deaths in the United States are attributable to cardiomyopathy. The mortality rate in men is twice that in women, and in Blacks it is twice that of Whites. In developing countries and in the tropics, the prevalence and mortality rates are much higher, probably because of nutritional deficiency, genetic factors, physical stress, untreated hypertension,
and
infection,
especially
1592
Chagas
disease.
The cardiomyopathies are a group of anomalies with three basic features: (1) failure of the heart to maintain its architecture, (2) failure of the heart to maintain normal electrical activity, and (3) failure of the heart to maintain cardiac output. General features of cardiomyopathies include cardiomegaly; congestive heart failure, often with relatively clear lungs; dilated LV and RV with elevated end-diastolic pressures and decreased contractility; and decreased ejection fractions. These findings are seen only in the later stages of hypertrophic and restrictive cardiomyopathies. Causes of congestive heart failure are listed in Table 22.2 . The cardiomyopathies may also be divided into dilated, hypertrophic, restrictive, and right ventricular forms (Table 22.3 ).
Dilated
Cardiomyopathy
In the Western world, dilated cardiomyopathy accounts for 90% of all cardiomyopathies (Fig. 22.12 ). The term “congestive cardiomyopathy― should be reserved for a subgroup of the dilated cardiomyopathies for which the etiology is unknown. Specific causes for dilated cardiomyopathies should be pursued because the specific therapy may vary: (1) ischemic cardiomyopathy (the most common cause) because of chronic ischemia, prior infarction, or anomalous coronary arteries; (2) long-term sequelae of myocarditis (Coxsackie virus, most commonly); (3) toxins (ethanol and Adriamycin [doxorubicin]); (4) metabolic conditions (mucolipidosis, mucopolysaccharidosis, glycogen storage disease); (5) nutritional deficiencies (thiamin and selenium); (6) infants of diabetic mothers; and (7) muscular dystrophies. Myocardial Cardiomyopathy Myocarditis Postpartum Coronary
(dilated,
restrictive,
cardiomyopathy
Transient ischemia Chronic ischemic cardiomyopathy Prior infarct or aneurysm
1593
hypertrophic)
Endocardial Fibrosis Löffler syndrome Valvular Stenosis Regurgitation Pericardial Effusion Constrictive Vascular Hypertension Pulmonary emboli Arteriovenous fistula Vasculitis Extracardiac Endocrinopathy
(thyroid,
adrenal)
Toxic Anemic Metabolic TABLE 22.2 Causes of Congestive Heart Failure
Dilated L V thin L V dilated Decreased Normal to decreased Hypertrophic L V thick L V normal to decreased Increased Decreased Restrictive Normal Normal Normal to decreased
1594
Severely decreased Uhl anomaly R V thin R V dilated Decreased Normal to decreased Type
Ventricular Wall
Ventricular Cavity
Contractility
Compliance
TABLE 22.3 Types of Cardiomyopathies Clinical presentation is related to congestive heart failure, although the initial presentation may include cardiac arrhythmias, conduction disturbances, thromboembolic phenomena, or sudden death. Presentation may also differ, depending on left-sided dominance, right-sided
dominance,
or
biventricular
involvement.
The chest radiograph commonly demonstrates global cardiomegaly. Larger heart sizes are associated with worse prognosis. Coronary artery calcification may be a clue to ischemic cardiomyopathy. Gated myocardial scintigraphy shows decreased left ventricular ejection fraction, prolonged pre-ejection period, shortened left ventricular ejection time, and a decreased rate of ejection. Echocardiography shows a dilated LV with global hypokinesia, thinning of the left ventricular wall and interventricular septum, decreased myocardial thickening,
left
atrial P.637
enlargement, and often right ventricular hypokinesia. MR shows dilatation of the specific chambers, decreased thickness of the myocardium with nonuniformity seen in prior infarctions, pericardial effusions, right and left ventricular ejection fractions, stroke volumes, wall-stress physiology, and quality of systolic wall thickening.
1595
FIGURE 22.12. Dilated Cardiomyopathy. The typical appearance of a dilated cardiomyopathy is demonstrated with a water-bottle configuration and dilatation of the azygos vein (arrow ). Pulmonary infiltrates are the result of pulmonary edema and capillary leak in this patient with viral myocarditis.
Hypertrophic cardiomyopathy may be familial (60%), autosomal dominant with variable penetrance, associated with neurofibromatosis and Noonan syndrome, or secondary to pressure overload. The hypertrophic cardiomyopathies are divided into two basic types: (1) concentric hypertrophy , which may be diffuse, midventricular, or apical in distribution; and (2) asymmetrical septal hypertrophy (ASH), also known as idiopathic hypertrophic subaortic stenosis (IHSS) (Fig. 22.13 ). Either form may cause some degree of muscular outflow obstruction with a systolic pressure gradient. Systemic hypertension may cause left ventricular hypertrophy followed by dilation, pulmonary venous hypertension, and increased risk of coronary artery disease.
1596
FIGURE 22.13. Hypertrophic Cardiomyopathy. Gradient-echo MR demonstrates marked left ventricular hypertrophy on these shortaxis views of the LV obtained during diastole (A) and systole (B). Note the asymmetric thickening of the septum (arrow ) compared with the remainder of the left ventricular myocardium. Ejection fraction is 92%.
The clinical presentation includes angina, syncope, arrhythmias, and congestive heart failure. Sudden death occurs in up to 50% of patients. The overall mortality rate is 2% to 3% per year. On chest radiography, 50% of patients with hypertrophic cardiomyopathy will have a normal chest radiograph and 30% will have left atrial enlargement, commonly because of mitral regurgitation. Echocardiographic features of ASH include (1) hypertrophy of the interventricular septum (>12 to 13 mm), (2) abnormal ratio of thickness of the interventricular septum to the left ventricular posterior wall (>1.3:1), (3) systolic anterior motion of the mitral valve with mitral regurgitation, (4) narrowing of the left ventricular outflow tract during systole, (5) high velocity across the left ventricular outflow tract with delayed systolic peaks on Doppler examination, (6) midsystolic closure of the aortic valve, and (7) normal or hyperkinetic left ventricular function. Restrictive cardiomyopathy is the least frequent form of cardiomyopathy. Etiologies include infiltrative disorders such amyloid,
glycogen
storage
disease,
1597
mucopolysaccharidosis,
as
hemochromatosis, sarcoidosis, and myocardial tumor infiltration. In the tropics, endomyocardial fibrosis is highly prevalent. A rare form of endomyocardial fibrosis associated with eosinophilia is called Löffler endocardial fibrosis. Restrictive cardiomyopathy should be considered when patients present with symptoms of congestive failure without radiographic evidence of cardiomegaly or ventricular hypertrophy (Fig. 22.14 ). The primary differential diagnosis is constrictive pericardial disease that can be differentiated by CT or MR. Signs and symptoms are related to congestive failure, arrhythmias, and heart block. In late stages, the electrocardiogram shows low voltage. Pathophysiology includes impaired diastolic function with decreased ventricular compliance, poor diastolic filling, and elevation of right and P.638 left ventricular filling pressures. Early in the progression of the disease, ventricular systolic function is normal or near normal. There may be a significant decline in later stages.
FIGURE 22.14. Restrictive Cardiomyopathy. Spin-echo MR demonstrates a variable high-density signal within the myocardium, a dilated RA (closed curved arrow ), and an enlarged inferior vena cava (open curved arrow ). The interventricular septum has an
1598
abnormal contour (straight arrow ) because of high right ventricular pressures in this biopsy-proven case of amyloid cardiomyopathy.
The chest radiograph often shows a normal-sized heart with pulmonary congestion. Left atrial enlargement and pulmonary venous hypertension may be present. The PYP nuclear scans demonstrate hot spots in abnormal areas of myocardium in 50% to 90% of patients. Echocardiography may show decreased systolic and diastolic function, with normal to decreased ejection fractions. Mild left ventricular wall hypertrophy is often present, with a granular or “snowstorm― appearance to the myocardium, especially noted in the case of cardiac amyloidosis. MR shows high signal in the myocardium on T2WIs in patients with amyloidosis and sarcoidosis. The atria are enlarged because of elevated diastolic pressures, but ventricular volumes are often normal. Mitral regurgitation and tricuspid regurgitation are readily depicted with gradient-echo cine MR and Doppler echocardiography. The inferior vena cava and superior vena cava may be greatly dilated.
Right Cor
Ventricular
Cardiomyopathies
pulmonale is defined as right ventricular failure secondary to
pulmonary parenchymal or pulmonary arterial disease. It may be considered a secondary form of right ventricular cardiomyopathy. Etiologies include (1) destructive pulmonary disease, such as pulmonary fibrosis and chronic obstructive pulmonary disease; (2) hypoxic pulmonary vasoconstriction resulting from chronic bronchitis, asthma, CNS hypoxia, or upper airway obstruction; (3) acute and chronic pulmonary embolism; (4) idiopathic pulmonary hypertension; and (5) extrapulmonary diseases affecting pulmonary mechanics such as chest deformities, morbid obesity (pickwickian syndrome), and neuromuscular diseases.
1599
FIGURE 22.15. Cor Pulmonale. A posteroanterior chest radiograph demonstrates marked hyperinflation caused by chronic obstructive pulmonary disease. The anterior junction line (arrow ) is herniated to the left of the aortic knob because of marked emphysema in the anterior segment of the right upper lobe.
The end result is alveolar hypoxia leading to hypoxemia, pulmonary hypertension, elevated right ventricular pressures, right ventricular hypertrophy, right ventricular dilation, and right ventricular failure. Symptoms include marked dyspnea and decreased exercise endurance out of proportion to pulmonary function tests. Blood gases demonstrate hypoxemia and hypercapnia. The chest radiograph shows a normal-sized heart or mild cardiomegaly (Fig. 22.15 ). Right ventricular and right atrial enlargement may be present. The main and central pulmonary arteries are prominent, and the periphery is oligemic. The interlobar artery typically measures more than 16 mm. The lungs show signs of chronic obstructive pulmonary disease, emphysema, or pulmonary fibrosis. Nuclear scintigraphy shows right ventricular enlargement with decrease in the right ventricular ejection fraction on first-pass examination. Echocardiography, CT, and MR show right ventricular
1600
and right atrial enlargement with thickening of the anterior right ventricular wall. M-mode echocardiography of the tricuspid valve shows a diminished A wave and a flat E–F slope. Therapy is aimed at the underlying pulmonary disorder. Uhl anomaly was initially described as a congenital disorder with “parchment-like thinning― of the RV. More recently it has been described as an acquired disorder in infants or adults and is called “arrhythmogenic right ventricular dysplasia.― This rare form of cardiomyopathy P.639 is limited to dilation of the RV, with marked thinning of the anterior right ventricular wall. MR may also show fatty infiltration of the anterior RV free wall, which is diagnostic. Clinical presentation includes syncope, recurrent ventricular tachycardia, and premature death from early congestive failure or arrhythmias. Familial occurrence has been reported, and males outnumber females by 3 to 1. Right ventricular ejection fractions are commonly reduced to less than half of normal, with mild reductions in the left ventricular ejection
fraction.
PULMONARY
VASCULAR
DISEASE
Enlargement of the pulmonary outflow tract is seen in congenital heart disease with left-to-right shunts. Outflow tract prominence without evidence of a shunt lesion is usually the result of poststenotic dilatation secondary to pulmonary stenosis, pulmonary arterial hypertension, Marfan syndrome, Takayasu arteritis, or idiopathic dilatation of the pulmonary artery. Idiopathic dilatation of the pulmonary artery demonstrates a dilated main pulmonary artery, normal peripheral pulmonary arteries, and normal, balanced circulation. This entity is much more common in women and is often associated with a mild systolic ejection murmur, but without evidence of pulmonary stenosis. Pulmonary arterial hypertension should be considered whenever the main pulmonary artery and left and right pulmonary arteries are enlarged (Fig. 22.16 ). Signs of right atrial and ventricular
1601
enlargement or hypertrophy are often present. Systolic right ventricular and pulmonary artery pressures exceed 30 mm Hg. Other findings include rapid tapering and tortuosity of the pulmonary arteries. The peripheral lung zones appear clear. Calcification within the pulmonary arterial walls is virtually diagnostic of pulmonary arterial hypertension (Fig. 22.17 ).
FIGURE 22.16. Pulmonary Arterial Hypertension. The main pulmonary artery (curved arrow ), left pulmonary artery (open arrow ), and right pulmonary artery (straight arrow ) are extremely enlarged. Faint calcification is seen in the right pulmonary artery. The patient had schistosomiasis with resultant vasculitis and pulmonary
arterial
hypertension.
1602
FIGURE
22.17. Pulmonary
Arterial
Hypertension.
Noncontrast
CT demonstrates calcification in the wall of the right pulmonary artery (arrow ).
The differential diagnosis for pulmonary arterial hypertension includes long-standing pulmonary venous hypertension (e.g., mitral stenosis), Eisenmenger physiology (from long-standing left-to-right shunts), pulmonary emboli, vasculitides (such as rheumatoid arthritis or polyarteritis nodosa), and primary pulmonary hypertension. Polyarteritis nodosa is a necrotizing vasculitis involving the medium-sized pulmonary arteries. Radiographic findings include small pulmonary arterial aneurysms, focal stenoses, small infarctions, and signs of pulmonary hypertension. Primary pulmonary hypertension is most common in women in their third and fourth decades. Histologic examination reveals plexiform and angiomatoid lesions with no evidence of emboli or venous abnormalities. Symptoms include dyspnea, fatigue, hyperventilation, chest pain, and hemoptysis. Increased pulmonary blood flow is caused by high output states and left-to-right shunts. High output states include volume loading, pregnancy, peripheral shunt lesions (arteriovenous malformations), hyperthyroidism, anemia, and leukemia (Fig. 22.18 ). The main and
1603
central pulmonary arteries are enlarged, with increased circulation to the lower lobes, upper lobes, and peripheral lung zones. Bronchovascular pairs show enlargement of the vascular component. The most common shunts in the adult are the acyanotic lesions, including atrial septal defect, ventricular septal defect, patent ductus arteriosus, and partial anomalous pulmonary venous return. Cyanotic
lesions
with P.640
increased blood flow to the lungs include transposition of the great vessels, truncus arteriosus, total anomalous pulmonary venous return, and endocardial cushion defects. Ventricular septal defects with left-to-right shunting may occur acutely following myocardial infarction.
FIGURE 22.18. High Output Failure. Chest radiograph demonstrates cardiomegaly, vascular engorgement, and distension of the azygos vein in this pregnant patient with severe anemia. The azygos vein (arrow ) is a good marker of intravascular volume expansion or elevated central venous pressures.
Decreased pulmonary blood flow with a small heart is caused by chronic obstructive pulmonary disease, hypovolemia,
1604
malnourishment, and Addison disease. When the cardiac silhouette is enlarged, the differential diagnosis includes cardiomyopathy, pericardial tamponade, Ebstein anomaly, and right-to-left shunts from congenital heart disease. Asymmetric pulmonary blood flow may be evident on chest radiography, angiography, or nuclear medicine pulmonary perfusion scans (Fig. 22.19 ; see Figs. 22.15 , 22.26 ). This may result from either decreased or increased blood flow to one lung. Pulmonary valvular stenosis often results in increased blood flow to the left lung. With resultant left pulmonary artery dilatation, tetralogy of Fallot may cause increased blood flow to the right lung (Fig. 22.15 ). Surgical shunts, such as the Blalock-Taussig procedure, also increase blood flow to one lung. Decreased blood flow to one lung occurs in peripheral pulmonic stenosis (see Fig. 22.26 ), interruption of the pulmonary artery, scimitar syndrome, pulmonary hypoplasia, Swyer-James syndrome, pulmonary emphysema, pulmonary embolism, fibrosing mediastinitis, or carcinoma affecting one artery (Fig. 22.19 ). When examining a chest radiograph, one must be careful to exclude technical artifacts such as lateral decentering and soft tissue asymmetry such as mastectomy. The balance of circulation and size of the central pulmonary arteries should be compared, along with the size of the bronchovascular pairs.
FIGURE
22.19. Asymmetric
Pulmonary
1605
Blood
Flow.
Technetium-
99m macroaggregated albumin pulmonary perfusion lung scan demonstrates marked reduction in the pulmonary blood flow to the left lung (arrows ) in comparison with the right lung. A subtle left hilar mass was causing compression of the left pulmonary artery. POST, posterior; RPO, right posterior oblique; RLAT, right lateral; ANT, anterior; LLAT, left lateral; LPO, left posterior oblique.
Pulmonary venous hypertension may be identified on radiographs, pulmonary angiograms, or nuclear medicine
perfusion
scans (Fig. 22.20 ; see Fig. 21.19 ). Pulmonary venous hypertension is considered mild with wedge pressures of 10 to 13 mm Hg, moderate with equalization of upper and lower lobe blood flow and wedge pressures of 14 to 16 mm Hg, or severe with the upper lobe vessels being distended more than the lower lobe vessels and wedge pressure 17 to 20 mm Hg. Progressive cephalization is accompanied by progressive secondary enlargement of the pulmonary arteries and filling out of the hilar angles. The most common cause of pulmonary venous hypertension is elevation of left atrial pressures secondary to left ventricular failure (Table 22.4 ).
ACQUIRED
VALVULAR
HEART
DISEASE
Mitral stenosis in the adult is usually caused by rheumatic heart disease, with 50% of patients giving a history of P.641 rheumatic fever. Rarely, an atrial myxoma may mimic mitral stenosis on chest radiography. The incidence of mitral stenosis is higher in women by a ratio of 8:1. Lutembacher syndrome is a combination of mitral stenosis with a preexisting atrial septal defect, which results in marked right-sided enlargement.
1606
FIGURE
22.20. Moderate
Mitral
Stenosis. A chest radiograph
demonstrates mild cardiomegaly with straightening of the left heart border, prominence of the left atrial appendage (open arrow ), and evidence of left atrial enlargement (arrows ). Cephalization of blood flow and enlargement of the pulmonary arteries indicate pulmonary venous and pulmonary arterial hypertension. The normal mitral valve area is 4 to 6 cm2 . With mild mitral stenosis (mitral valve area Table of Contents > Section V - Cardiac Radiology > Chapter 23 Cardiac MRI
Chapter
23
Cardiac
MRI
Christopher M. Kramer Because of advances in both technology and acquisition techniques, cardiovascular MR (CMR) has evolved rapidly, especially over the last decade. With the rising prevalence of coronary artery disease (CAD) and heart failure, the mounting enthusiasm for CMR-based cardiac assessment is readily appreciated. With one examination, LV structure, function at rest and under stress, perfusion, and viability can be evaluated with a high degree of precision while avoiding the potentially harmful effects of ionizing radiation and nephrotoxic contrast agents. Further developments are needed to bring CMR coronary arteriography into the clinical routine. Atherosclerotic plaque imaging is a growing research application that will become clinically useful with time. Other common indications for CMR will be reviewed, including imaging of cardiomyopathies, pericardial disease, valvular heart disease, congenital heart disease, cardiac masses, and pulmonary veins.
ISCHEMIC
HEART
DISEASE
Global Structure and Function at Rest. In patients with congestive heart failure, abnormalities in ventricular structure and function, whether characterized by visual inspection or measurement of ejection fraction, mass, or volumes, are powerful predictors of prognosis and have important therapeutic implications. Careful visual
1636
assessment of LV function by an experienced clinician is adequate in most clinical scenarios, with the caveat that great care must be taken to analyze each segment. Although a number of strategies have been devised for partitioning the myocardium, the American Heart Association currently endorses the use of a 17-sector model divided on the basis of a normal coronary distribution. For assessment of LV function with CMR, images are obtained in shortaxis views covering the LV from apex to base as well as twochamber, three-chamber, and four-chamber orientations (Fig. 23.1) to ensure visualization of all segments. Steady-state free precession (SSFP) cine imaging is currently the optimal technique for visually assessing LV function. SSFP, in contrast to older gradient echo cine techniques, does not rely on the inflow of unsaturated spins to create contrast between the LV cavity and the endocardium. Because SSFP images are flow independent, endocardial border detection is significantly enhanced. Furthermore, SSFP provides excellent spatial and temporal resolution and a high signal-to-noise ratio, and it requires relatively short breath-holding times. Because of its high spatial resolution, reproducibility, and 3D data that require the operator to make no geometric assumptions, CMR has evolved into the reference standard for measuring mass, chamber volumes, and ejection fraction. Because of the exceptional contrast generated between the myocardium and blood pool (Fig. 23.1), CMR enables the operator to precisely delineate both the endocardial and epicardial borders. SSFP is the favored technique for assessing ventricular dimensions. The flow-independent contrast facilitates the accurate demarcation of border contours around the papillary muscles and ventricular trabeculations, where blood pooling typically P.653 occurs, particularly in patients with low-flow states such as congestive heart failure.
1637
FIGURE 23.1. LV Function. Steady-state free precession images in two standard planes at different phases in the cardiac cycle in a patient with a prior anterior myocardial infarction (MI). The upper panels demonstrate a four-chamber long-axis image at end-diastole on the left and end-systole on the right. Note the thinned, distal anterior wall and apex (arrow) with reduced wall thickening during systole, which are suggestive of prior MI in that
region.
To calculate LV ejection fraction, mass, and volume, short-axis slices stacked from apex to base of the LV are acquired throughout the cardiac cycle, and the epicardial and endocardial contours are carefully outlined at end-diastole and end-systole. Most CMR systems have specialized software with automated border detection to assist
1638
the operator, but the technology remains limited for patients with non-geometric configurations and in those with prominent papillary muscles or heavy trabeculations. Determination of the slice area, distance between slices, slice thickness, and both end-systolic and end-diastolic volumes can then be performed using Simpson's Rule. Once these values are obtained, stroke volume (end-diastolic volume – end-systolic volume) and ejection fraction (systolic volume/enddiastolic volume × 100%) are readily measured. Myocardial mass can then be calculated by dividing the density of myocardium (1.05 g/cm3 ) by the measured LV volume. The same techniques can be applied to the RV, but because of the complex 3D structure of this chamber and heavy trabeculations, the value of currently available automated border detection software is limited. MR flow velocity mapping, a technique that can quantify flow velocity and direction on a per-pixel basis, can also be used to quantify global cardiac function. After an imaging plane is selected perpendicular to flow, velocities are acquired in the through-plane direction at multiple points in the cardiac cycle. Through summation of velocities in a selected portion of the proximal aorta, stroke volume and cardiac output can then be determined.
Regional
Myocardial
Conventional techniques primarily on evaluating methods are insensitive well as translation and
Function
for assessing ventricular motion of the endocardial to the deformation within torsion during contraction
motion rely service. These the myocardium as and relaxation.
Early efforts to characterize this dynamic 3D geometry provided insight into the complexities of cardiac motion but required implantation of radiopaque markers within the myocardium. Myocardial tagging (Fig. 23.2) places virtual markers within the heart through the manipulation of the magnetic field to facilitate visualization and quantification of regional function, including the rotational and translational motion that has been previously difficult to analyze. CMR is currently the only noninvasive technique with this capability. To generate a tagged sequence, a grid consisting of nulled orthogonal lines is applied to the heart at end-diastole by altering
1639
the local magnetization with narrow and radiofrequency pulses. Because saturated rows of protons comprise this grid structure, the tagging is “embedded― in the tissue and its motion can be reliably tracked throughout the cardiac cycle. Intramural deformation can be visualized and the strain quantified at various sites within the myocardium. Strain analysis is more accurate than planar wall thickening for detecting regional myocardial dysfunction, as this technique takes into account the motion of a selected segment in all directions simultaneously. Quantification of strain with tagged CMR can be performed with a high degree of precision, allowing for separation of the subendocardial, midmyocardial, and subepicardial layers. Although a precise assessment of 3D LV function can be achieved with this technique, the data analysis remains cumbersome and time consuming. New methods of image acquisition and postprocessing analysis are currently under investigation, such as HARP (harmonic phase) and DENSE (displacement encoding with stimulated echoes), both of which allow more rapid analysis. Although it is not ready for routine clinical application, CMR strain imaging may become the diagnostic reference standard of the future, one that will enhance our ability to identify subtle abnormalities in function during stress testing and allow for earlier detection of disease states. Myocardial tagging techniques have already enabled researchers to achieve a better understanding of cardiac function in both normal and diseased states. Tagging has characterized regional myocardial dysfunction in acute and chronic myocardial infarction (MI), hypertrophic P.654 cardiomyopathy, valvular heart disease, and pulmonary hypertension. Furthermore, tagging has facilitated the detailed analysis of the local functional response of the myocardium to a number of therapies for congestive heart failure, including pharmacologic agents and surgical reduction treatments.
1640
FIGURE
23.2. Myocardial
Tagging. This technique is shown in
a patient on day 3 after anterior myocardial infarction. The left panel represents a short-axis midventricular tagged image at end-diastole, while the right panel depicts the end-systolic frame at the same location. Because the tags (dark
stripes) remain
embedded within the tissue throughout the cardiac cycle, deformation can be tracked and the strain quantified. Note the normal deformation in the posterolateral wall (3:00 to 6:00 in the image) (arrows) and the reduced deformation in the anterior wall.
The role of tagging in ischemic heart disease is being established. In a recent study of 211 consecutive patients with chest pain, dobutamine CMR (DCMR) with tagging was shown to be superior to nontagged sequences for detection of significant coronary lesions with a 96% sensitivity. Among the patients with normal dobutamine CMR and no resting wall motion abnormalities, the event free survival after 17 months was 98%. Dobutamine CMR tagging has also been demonstrated to be equivalent to dobutamine stress echo in predicting functional improvement following MI with reperfusion. More recently, tagged CMR images were employed to assess timing
1641
of cardiac contraction and asynchrony in candidates for biventricular pacing.
Global
Function:
Stress
Dobutamine is the drug most commonly administered when performing stress CMR because of its proven efficacy in stress echocardiography (DSE). Dobutamine increases myocardial oxygen demand through its potent chronotropic and ionotropic effects. In a territory supplied by a stenotic coronary artery, inadequate blood flow is available to compensate for increased demand, resulting in ischemia. This manifests visually as a segmental wall motion abnormality, which is a reliable sign of ischemia that precedes both angina and electrocardiographic changes. Side effects of dobutamine infusion are common but typically mild, and the test is generally well tolerated, even by elderly patients. Although a number of studies have confirmed the safety and efficacy of DSE for diagnosing CAD, suboptimal image quality remains a frequently encountered drawback, occurring in up to 15% of patients studied, although problems are less common when contrast is used for LV cavity opacification (LVO). DCMR compares favorably to DSE. DCMR requires no imaging window, which results in high-quality reproducible images, irrespective of the operator or particular patient (Fig. 23.3) . Furthermore, slice position can be better reproduced during the study, ensuring that the same regions are compared both at rest and during stress. An older study using a breath-hold gradient-echo cine technique compared high-dose DCMR and DSE in a group of 208 consecutive patients. The accuracy of DCMR in detecting ischemia was superior to that seen with second-harmonic DSE without LVO (86% versus 73%), primarily because of superior image quality. Another study enrolled 153 patients with inadequate windows by second-harmonic DSE (without LVO) and performed DCMR with both sensitivity and specificity >80%. DCMR was able to accurately diagnose ischemia in patients who were poor candidates for DSE. Subsequent follow-up
1642
demonstrated that the results could be used to predict subsequent MI and cardiovascular death. Recently, DCMR, adenosine stress-function CMR, and adenosine firstpass perfusion were compared in 79 consecutive patients with suspected or known CAD who were scheduled for coronary angiography. Using quantitative coronary angiography as the gold standard, the sensitivity and specificity of DCMR, adenosine stressfunction CMR, and adenosine magnetic resonance first-pass perfusion (MRFP) was 89% and 80%, 40% and 96%, and 91% and 62%, respectively. This study underscores the superior accuracy of dobutamine over adenosine for functional stress testing and relative advantages in specificity compared to stress-perfusion imaging. Despite the apparent superiority of DCMR, widespread application has yet to be realized. In regard to safety, as long as online assessment of ventricular function during dobutamine infusion is used, the inability to reliably interpret electrocardiogram (ECG) recordings during testing is not a problem, since wall motion abnormalities precede ECG changes in the ischemic cascade. With respect P.655 to patient monitoring, all precautions observed with other modalities are mandatory with DCMR, including presence of trained clinicians and MR-compatible monitoring equipment. In case of a medical emergency, staff can provide critical life support while the patient is being moved out of the magnetic environment, a procedure that can be executed in less than 30 seconds. Significant complications from dobutamine infusion can be expected to occur in up to 0.6% of patients and include MI, sustained ventricular tachycardia, and even death. In a recent review of 1,035 consecutive patients who underwent DCMR, the only major complication consisted of one patient developing sustained ventricular tachycardia, which was successfully cardioverted.
1643
FIGURE 23.3. Dobutamine Cardiac MR Stress Test. Two endsystolic steady-state free precession four-chamber long-axis image frames from a dobutamine cardiac MR stress test in a patient with chest pain 10 years following left internal mammary bypass graft to the left anterior descending artery. The left frame is at a low dose of dobutamine (10 µg/kg/min) and demonstrated cavity obliteration at end-systole, implying normal systolic function. The right frame is during peak dobutamine dose (40 µg/kg/min) at a heart rate of 150 beats per minute. At peak dobutamine, note the lack of wall thickening in the apical septum consistent with disease in the bypass graft that was proven at subsequent catheterization.
Myocardial
Perfusion
Stress-perfusion nuclear imaging is commonly performed in patients with suspected or known CAD. SPECT is the noninvasive technique most frequently employed to analyze myocardial perfusion, and on a more limited scale, PET. Although both modalities are well validated, they are hampered by limited spatial resolution, the potential risks of radiation exposure, and an inability to reliably detect subendocardial perfusion defects. First described in humans in 1992, stress magnetic resonance first-pass perfusion (MRFP) has generated a substantial
1644
amount of interest because of its favorable spatial resolution in comparison to SPECT and PET. MRFP is performed using a T1W sequence to visualize a gadoliniumbased contrast agent in transit through the heart. Following peripheral injection, the contrast is detected against the background of nulled myocardium, first in the right-sided chambers, then opacifying the left-sided cavities. The myocardium begins to enhance when the contrast reaches the ascending aorta and coronary ostia, with a peak in enhancement approximately 10 heartbeats after the LV opacifies (Fig. 23.4). Signal intensity correlates with contrast concentration. Because passage of this extracellular agent can occur in as little as 5 to 10 seconds during hyperemic conditions, rapid acquisition of images is essential, and numerous techniques to improve speed have P.656 been employed. With correct patient breath-holding and proper ECG gating, a high degree of spatial resolution is preserved.
1645
FIGURE
23.4. MR
First-Pass
Perfusion.
First-pass
gadolinium-
enhanced hybrid gradient-echo/echo planar perfusion images in a basal short-axis plane following adenosine stress (upper panel) and at rest (lower panel) demonstrating a reversible perfusion defect in the inferior wall (5:00 to 7:00 in the image) (arrows) . The patient was later shown to have a 99% distal right coronary artery stenosis at cardiac catheterization.
MRFP analysis relies on time-intensity curves and can be performed in a quantitative, semiquantitative, or qualitative fashion. Because the contrast agents rapidly redistribute into the extracellular space, analysis is limited to the initial upslope in the time-intensity curves, which has been shown to correlate well with measures of microsphere blood flow. In clinical practice, qualitative analysis by an experienced clinician is generally performed and relies on observer detection of regional differences in signal intensity over time. Studies evaluating the sensitivity and specificity of visual detection of perfusion defects have been encouraging. Occlusive CAD results in a reduction in flow reserve, a characteristic feature that can be detected before either clinical symptoms or LV dysfunction are generally evident. In a viable myocardial territory supplied by a 70% stenotic coronary artery, vasodilated flow is reduced by approximately 50% when compared to flow in segments subtended by a normal coronary artery. Hence, any technique employed to detect lesions of 70% or greater must be able to differentiate at least a twofold difference in perfusion. Validation of MRFP in humans has been performed in a series of small clinical studies employing a variety of contrast agents, analysis techniques, and reference standards. One recent study examined signal-intensity time curves in both patients and controls following dipyridamole infusion and bolus injection of gadolinium diethylenetriaminepentaacetate (Gd-DTPA). Using a linear fit to determine the upslope, the researchers defined a cutoff value between normal and ischemic myocardial segments. Diagnostic
1646
accuracy was 87% and interobserver agreement was high (0.96). In another study of 48 patients, MRFP, (13)N-ammonia PET, and quantitative coronary angiography were performed. During first pass of contrast, signal-intensity up-slopes were calculated in 32 myocardial sectors. Analysis of the subendocardial upslope data showed sensitivity and specificity of 91% and 94%, respectively, when compared to PET imaging. When compared to quantitative angiography, sensitivity and specificity were lower but were still >85%. In another study of 84 patients referred for diagnostic coronary angiography, first-pass contrast-enhanced CMR during adenosine vasodilatation was performed. Myocardial perfusion reserve index was determined from the upslope of signal intensity within the myocardium. The best results were achieved when the innermost slices were evaluated; diagnostic accuracy using this approach was 89%. One of the more novel aspects of MRFP is that its superior spatial resolution allows for the discrimination of perfusion across layers in the same myocardial segment. Investigators have found transmural perfusion analysis useful in a number of clinical scenarios, including transplant arteriopathy, syndrome X, and transmyocardial laser revascularization. MRFP is already a powerful tool, with its capacity to detect perfusion abnormalities isolated to the endocardium, discern transmural flow gradients, and identify reductions in flow reserve that are undetectable with radionuclide imaging. With optimization of imaging sequences, development of new intravascular contrast agents, and imaging at higher field strengths, MRFP has tremendous potential for the detection of CAD.
Myocardial
Viability:
Contractile
Reserve
Viable cardiac myocytes are alive and capable of recovery of function, while nonviability denotes cell death or irreversible myocyte damage. The identification of viable myocardium in patients with chronic LV dysfunction is of significant clinical relevance. With restoration of adequate blood flow, viable segments have the potential for improvement in contractile function. The prognosis of patients with viability is substantially improved with
1647
revascularization. Alternatively, patients without viability who are revascularized have higher perioperative morbidity and worse longterm outcomes. Because routine histopathologic examination of cellular elements is impractical, a number of techniques have been developed and proven useful for measuring indirect parameters of viability. Because of its ability to evaluate a number of markers proven to predict viability, CMR is rapidly becoming established as the test of choice. The demonstration of contractile reserve is an established predictor of myocardial viability. As with echocardiography, CMR assessment of contractile reserve is performed using infusion of low dose dobutamine. Low-dose DCMR has been compared to fluorodeoxyglucose PET (FDG-PET) in 35 patients with chronic MI and segmental LV dysfunction. With end-diastolic wall thickness and improvement in systolic thickening used as markers of viability, CMR had a sensitivity, specificity, and diagnostic accuracy of 88%, 87%, and 92%, respectively, when compared to FDG-PET. When this technique was used to predict the improvement in LV function following revascularization, systolic thickening by CMR had a sensitivity and specificity of 89% and 92% for predicting recovery. Furthermore, segments that failed to improve following revascularization had a significantly lower diastolic wall thickness (6 mm versus 9.8 mm) than those that showed functional improvement. Echo-based studies have corroborated the value of end-diastolic wall thickness in predicting viability. The utility of incorporating tagged sequences into the assessment of contractile reserve has also been demonstrated. Both DSE and lowdose DCMR with tagging were performed in a study of 22 patients with acute MI following reperfusion therapy (day 3 ± 1). At 8 weeks, both echocardiography and CMR with tagging were performed at rest to evaluate for functional recovery. Using echo as the P.657 gold standard, the overall accuracy of DSE and low-dose DCMR with tagging was similar. Recent studies of hibernating myocardium prior to and after multivessel revascularization using tagged low-dose dobutamine have demonstrated that half of dysfunctional segments
1648
recover resting function, and of the remainder, half demonstrate rest dysfunction but contractile reserve. Half of these latter segments were found to have recovered rest function when examined 3 years after
revascularization.
Myocardial
Viability:
Infarct
Imaging
Late contrast-enhanced CMR (ce-CMR) using Gd-DTPA, an extracellular/interstitial contrast agent, has proven to be a valuable tool in identifying MI and predicting the potential of infarcted myocardium to recover following revascularization (Fig. 23.5). In regions of infarction, T1 relaxation is enhanced as a consequence of increased regional uptake and retention of contrast. This phenomenon is thought to occur within infarcted zones because of increased volume of distribution as well as delayed washout of contrast. Animal studies have demonstrated that the spatial extent of late contrast enhancement on CMR very closely mirrors the distribution of myocyte necrosis early post-MI and that of collagenous scar seen at 8 weeks. Furthermore, studies have shown that in regions of the heart subjected to reversible injury, the retention of contrast does not occur. In reality, zones of ventricular dysfunction typically consist of a combination of reversibly injured (hibernating or stunned) and irreversibly injured (infarcted) myocardium. The power of ce-CMR is its ability to distinguish these two states within the same segment of myocardium, a feat that cannot be replicated with other techniques.
1649
FIGURE 23.5. Myocardial Infarction. Contrast-enhanced inversion-recovery gradient-echo image in a four-chamber longaxis plane 10 minutes following gadolinium infusion at 0.2 mM/kg in a patient with a prior lateral wall infarction. Note the area of bright enhancement (arrow) in the lateral wall that subtends the inner 50% of the wall.
The ability of ce-CMR to identify and characterize myocardial scar has been directly compared to both SPECT and PET. In a study of 91 patients with suspected or known CAD, both ce-CMR and SPECT imaging were performed to determine the extent, location, and size of infarct scarring. Although SPECT correctly identified all patients with transmural or near-transmural scar seen on ce-CMR, SPECT failed to correctly identify nearly half of those with subendocardial infarction. Another study compared ce-CMR to PET in 31 patients with ischemic cardiomyopathy. Infarct mass correlated well between the two modalities, but ce-CMR more frequently identified scar than PET, again reflecting its superior spatial resolution.
1650
In 2000, a landmark study was published demonstrating the utility of ce-CMR for identifying myocardial viability. Fifty patients scheduled to undergo revascularization procedures were recruited, 40 of whom had some region of hyperenhancement. Of the 2,093 myocardial segments analyzed, 804 demonstrated abnormal contractility. The likelihood of functional recovery strongly correlated with the transmural extent of scar. In patients with no hyperenhancement, 78% of segments demonstrated recovery of function. In contrast, only 1 of 58 segments with >75% transmural extent showed any improvement following revascularization, demonstrating the powerful negative predictive value of this finding. Within zones that exhibited 51% to 75% hyperenhancement, only 10% improved following revascularization. Even in regions with akinesis or dyskinesis prior to revascularization, lack
of hyperenhancement predicted functional
recovery in 100% of cases. Several subsequent studies have validated this approach for predicting recovery of function after revascularization. Some controversy persists as to which test is best for identifying viability among CMR-based techniques. A recently completed study compared the performance of low-dose DCMR with ce-CMR in predicting viability. Although no difference was appreciated in identifying viability in segments without hyperenhancement or those with scar≥75%, DCMR was superior in predicting recovery in zones demonstrating between 1% and 74% transmural scar by late enhancement. The performance of late enhancement varied greatly depending on the cutoff values used to define viability. Ultimately, complementary use of DCMR and ce-CMR may prove to be the optimal strategy for defining viability. Although some evidence suggests that revascularization confers clinical benefits irrespective of systolic functional recovery, this hypothesis remains untested at present. As our understanding of the nature and importance of late contrast enhancement has evolved, investigators have found that more than just the size and extent of enhanced regions have significance. Many large acute P.658 infarcts demonstrate hypoenhanced regions at the core of
1651
hyperenhanced zones, which represent areas of microvascular obstruction (Fig. 23.6). The presence of microvascular obstruction correlates strongly with infarct size and identifies regions that are truly nonviable. Moreover, microvascular obstruction has been shown to be a powerful predictor of lack of functional recovery in the infarct region and poor cardiovascular outcome in the patient post-MI. In a recent study of 110 patients post-MI, microvascular obstruction was a predictor of LV remodeling, as defined by an increase in enddiastolic volume, and was a more powerful predictor of survival than either ejection fraction or infarct size.
FIGURE 23.6. Microvascular Obstruction. Short-axis late contrast-enhanced image using an inversion-recovery gradientecho sequence 10 minutes following gadolinium infusion at 0.2 mM/kg shows infarct scar in the septum with a small hypoenhanced zone, which is consistent with microvascular obstruction (arrow). This region would be deemed nonviable based on the transmural extent of hyperenhancement and presence of microvascular obstruction.
1652
Coronary
Arteriography
Because about 35% of patients referred for their first invasive x-ray angiogram have normal epicardial coronary arteries, an appealing role for CMR in ischemic heart disease would be the noninvasive assessment of the coronary arteries with high temporal and spatial resolution, during relatively short acquisition times. Coronary magnetic resonance angiography (CMRA) has not matured to that point yet, but substantial improvements have been achieved over the past few years. Given the small size, tortuous course, and motion of the coronary arteries, several technical challenges must be overcome to obtain images of diagnostic quality. Best in-plane resolution for CMRA is about 600 to 900 µm, which is still about twice the pixel size available in conventional angiography. Compensation for cardiac and coronary arterial motion is achieved by using short acquisition times and optimizing the timing of acquisition in middiastole, when cardiac motion is least. Respiratory motion correction can be achieved by several different techniques. The advantages of conventional
breath-holding
techniques
are
shorter
acquisition
times
and the freedom to repeat the acquisition if the images are suboptimal, but the shorter acquisition time results in lower signalto-noise ratio. The signal-to-noise ratio can be greatly improved upon by longer acquisition times, but this requires respiratory compensation to avoid blurring of the images. The most commonly used techniques rely on diaphragmatic navigators, in which the lungdiaphragm interface is tracked and is used to predict the motion and position of the coronary arteries. Using this method, each acquisition takes about 5 to 10 minutes, with the current navigator efficiency of 30% to 50% during free breathing. CMRA has evolved since the early 1990s when two-dimensional fatsuppressed breath-hold techniques held early promise, with sensitivity and specificity of 90% and 92%, respectively, but fell prey to less optimal results in later studies. Several new approaches are in development in an attempt to optimize accuracy. Threedimensional approaches are most promising; the one used most
1653
frequently is a navigator echo technique that tracks diaphragmatic motion and acquires images when the diaphragm is in a specified position. A recent multicenter study of a T2 -preparatory pulse with a gradient-echo 3D sequence and navigator echoes showed some promise, with an overall accuracy of 72%. Results in patients with left main or three-vessel CAD were quite good, with 100% sensitivity, 85% specificity, and 87% accuracy. However, the specificity on an individual vessel basis remained inadequate for screening purposes, ranging from 52% to 72% in the three major coronary
vessels.
The advent of high-field (3T) coronary imaging offers enhanced image quality and resolution (Fig. 23.7) that may allow improved accuracy for detection of CAD, although large patient studies have not been performed to date using these higher field strengths. In clinical practice, CMRA is used to assess anomalous coronary arteries, where it has been shown to be equivalent or superior to conventional coronary angiography. Other clinical applications include the evaluation of patients with Kawasaki disease. Detection of aneurysm and thrombus within aneurysms has been demonstrated with CMRA, and serial MRA avoids the risk of radiation exposure in children with Kawasaki disease. Another clinical application of CMRA is the imaging of bypass grafts, both arterial and venous. Navigator-gated 3D CMRA has been shown to be quite accurate for identifying occlusion or stenosis. In one study of 56 vein grafts, the area under the curve by receiveroperator characteristic P.659 analysis was 0.89 and 0.89 for occlusion and 0.82 and 0.79 for stenosis >70%, with excellent interobserver agreement. A subsequent study of bypass graft patency with SSFP angiography in 25 patients showed lower specificity. Measuring flow in bypass grafts has been shown to be helpful as well. In one study of 69 patients who underwent velocity encoded flow mapping by MR at baseline and with vasodilator stress, receiver operating characteristic (ROC) analysis demonstrated sensitivity of 96% and specificity of 92% for
1654
identification of stenosis >70%, suggesting improved performance over angiography alone. However, flow scans could not be obtained in 80% of grafts.
FIGURE
23.7. Coronary
MR
Angiography. Shown are curved
multiplanar reformats of a three-dimensional, navigator-gated, T2-prepared gradient-echo coronary MR angiogram performed at 3.0 Tesla in a healthy volunteer. The image on the left demonstrates normal left main artery (LM, arrow) and left anterior descending artery (LAD, arrow) at high spatial resolution (0.6 × 0.6 × 3 mm voxel size) that allows visualization of diagonal and septal branches (broken arrows). The image on the right demonstrates the right coronary artery (RCA). AO, aorta; RV, right ventricle; LV, left ventricle; PA, pulmonary artery; LCX, left circumflex. (From Flamm SD, Muthupillai R. Coronary artery magnetic resonance angiography. J Magn Reson Imaging 2004;19:686–709; reprinted with permission.)
Recent efforts have been aimed at imaging not the coronary lumen but the coronary wall itself. Black-blood MR images that null the signal from the blood are essential to visualize the coronary wall. A
1655
high-resolution black-blood CMR method without motion or bloodflow artifacts and with 0.46-mm in-plane resolution has been used to visualize the wall of major epicardial coronary arteries in a group of normal subjects and five patients with CAD. The average coronary wall thickness for each cross-sectional image in normal subjects was 0.7 mm, with a range from 0.5 to 1.0 mm. The patients with coronary disease with Table of Contents > Section VI - Vascular and Interventional Radiology > Chapter 24 - Thoracic Aorta, Pulmonary Arteries, and Peripheral Vascular Disorders
Chapter
24
Thoracic Aorta, Pulmonary Arteries, and Peripheral Vascular Disorders Michael J. Miller Jr. Tony P. Smith
INTRODUCTION
TO
VASCULAR
RADIOLOGY
The walls of both arteries and veins are made up of three layers; from the inside out, they are the intima , the media , and the adventitia . The intima is a single cell layer thick and has the primary function of interacting with the flowing blood—in particular, preventing thrombosis. It is by far the most chemically active layer. The media is composed for the most part of smooth muscle—more, of course, in large muscular arteries and less in smaller arteries and veins. Smooth muscle cells can contract to augment normal hemodynamic function and to respond to stress, such as with vasoconstriction (vasospasm). The adventitia is a layer of supportive connective tissue of varying thickness that surrounds and supports the media. The anatomic structure of the vessel wall is actually simple, but its physiologic and pathologic function is very complex and for the most part poorly understood. A wide variety of disease processes can affect the vessel wall, particularly the arterial wall. These diseases include inflammation (vasculitis), fibromuscular disease, connective tissue disease, trauma, and of course degeneration (atherosclerosis), to name
1685
a few. Although we often do not understand the exact pathophysiology underlying a vessel's reaction to a particular disease process in a particular individual, the arterial wall and vessel itself have only a limited number of radiographic manifestations. When the vessel wall is “attacked,― it can weaken and dilate, producing an aneurysm, or it can even rupture, causing extravasation, pseudoaneurysm formation, or arteriovenous fistulas. It can thicken either by growth of the vessel layers (intimal hyperplasia for example) or by the deposition of material, such as an atherosclerotic plaque, causing the vessel to narrow and producing stenosis or even occlusion. The vessel may lose its ability to prevent coagulation, resulting in thrombosis. As yet unknown genetic factors may induce the proliferation of vessels, resulting in arteriovenous malformations, or vessels may be induced to grow and proliferate by “acquired― factors such as within a tumor. If one keeps in mind the vessel wall, in particular the arterial wall, much of what is seen angiographically is more easily understood.
Angiographic
Suite
Most angiographic suites have two major types of equipment: patient monitoring devices and radiographic equipment. Patient monitoring devices are essential to patient care during angiographic procedures, especially for conscious sedation, and there are usually one or more channels for pressure measurements. Radiographic equipment today is based on a C-arm design that allows complex angulation and is equipped for digital acquisition only. C-arm configuration allows one to set the angle for imaging; therefore the technologist no longer has to place the patient based on landmarks into the “named― positions. One is also able to acquire images at a rapid rate for long periods of time, allowing for special imaging P.672 such as bolus chasing for leg angiography and rotational angiography, which
provides
three-dimensional
image
viewing.
Tools There are a number of catheters used in interventional radiology that
1686
can be loosely divided into five types: diagnostic angiographic catheters, microcatheters, drainage catheters, balloon catheters, central venous catheters.
and
There are a host of properties considered in manufacturing, buying, and using diagnostic angiographic catheters ; these include size (the smaller the better for access but size limits the lumen), shape, radiopacity, torque capacity, and softness of the distal tip. Larger-lumen diagnostic catheters are available (guiding catheters) for placement of microcatheters and angioplasty balloons in a coaxial fashion. Microcatheters are 3 French or smaller and are designed for very distal catheterization. These catheters are placed over 0.010- to 0.018-inch guide wires. Most of their use is in the neurointerventional arena, but they are very helpful for peripheral intervention to select small vessels for embolization or infusion (e.g., chemotherapy). These catheters have a distal platinum marker but are otherwise not very radiopaque. Drainage catheters are used frequently in interventional radiology for drainage of fluid collections, including nephrostomy, abscess, biliary gallbladder, pleural fluid, ascites, and lymphoceles. The same basic catheters are used for drainage in all such sites. Characteristics of drainage catheters include size, biocompatibility, radiopacity, and softness shape/retention property. Catheter shape is usually based on the size of the fluid collection for drainage and the retention property. The most common retention device for pigtail catheters is the retention suture. Straight catheter retention devices include the mushroom tip or inflatable
balloon.
Balloon catheters can either be very soft and pliable, such as occlusion balloons or Fogarty balloons to clear thrombosis, or can be more rigid and used for dilations (angioplasty). Balloons for dilation can be divided into two main categories regarding the size of guide wire over which they are placed: 0.018 inch (or even smaller, including 0.014 inch) and 0.035 inch. The smaller guide wire lumen obviously allows the balloon to be smaller. These are the balloons used for coronary angioplasty, but they have become popular recently in peripheral and neuroradiologic interventional procedures. The smaller systems do not have the guide wire support of the larger systems, and the balloons cannot be
1687
constructed in very large diameters. Most peripheral intervention is performed with 0.035-inch wire balloon systems. There are an array of balloons in a variety of sizes. Large balloons require large access sites (introducer sheaths), especially for removal after inflation, as they do not “re-wrap― very well. An important concept to understand for the angioplasty balloon is compliance. Once the balloon reaches the manufacturer's stated size, it can be very firm (noncompliant) or it can “grow― a little with inflation (compliant). There are advantages to each. With the former, very difficult, hard lesions can be dilated. With the latter, one can “size― the balloon a little larger than its stated size by increasing the atmospheric pressure during dilation. This allows better fine-tuning of the angioplasty. Central venous lines can be differentiated into those that have an implanted portion and those that are placed by direct puncture without implantation. The former are implanted subcutaneously and are designed for extended usage. The latter are the traditional central lines (central venous lines, Swan-Ganz catheters) that are placed for temporary care and monitoring. A special one of these is the peripherally inserted central catheter, which is placed via a peripheral arm vein coursing into the central veins for up to 6 weeks. Implanted devices can be divided into two main types: tunneled external catheters , placed using subcutaneous tunnels, and implanted ports , placed in a subcutaneous pocket. The type of device chosen depends on its clinical use and is based on a variety of factors, including number of lumens needed, frequency of use, type of use, length of use, device location, and individual patient factors. Tunneled external catheters are tunneled from the venous access site and have a retaining fabric (usually Dacron polyester) cuff to prevent dislodgment by the ingrowth of connective tissue. They are less expensive and less invasive to place than completely implanted ports but tend to have higher infection rates and are less cosmetically desirable. There are two main designs based on usage: cuffed central venous designs (e.g., Hickman) and those for hemodialysis access (e.g., PermCath). Implanted ports contain catheters that are tunneled a short distance from the venous access, where the port device is implanted in a subcutaneous pocket. Most guide
wires for standard angiography and interventional
1688
procedures fall into two categories based on their construction: spring guide wires , which are constructed of stainless steel wire tightly wound on itself to form a spring, and nitinol guide wires , which are constructed of a nickel-titanium alloy and an organic coating to which is bound a hydrophilic coating. This coating absorbs water and becomes very slippery. Guide wires range in size from 0.010 to 0.038 inches. There are two types of stents based on mechanism of delivery: selfexpanding and balloon-expandable . Self-expanding stents are composed of various metal combinations, most often either stainless steel or nitinol alloys. Balloon-expandable stents are composed for the most part of stainless steel. Self-expanding stents have the advantage of being quite flexible and can be constructed in quite large diameters but are often more difficult to place precisely because of foreshortening upon stent opening. This problem has been lessened somewhat with the nitinol varieties. Balloon-expandable stents are less flexible and are limited by balloon size but can be more precisely placed. There is much work being performed regarding P.673 drug delivery (coated stents) and covering of stents (stent grafts). Drug-coated stents are available for coronary use, but will soon be available in sizes for peripheral work as well. A number of stents are available that are covered by fabric, producing in effect a stent graft . The fabric is most often polytetrafluoroethylene. Smaller varieties are used for the peripheral circulation and larger varieties are used for the aorta. In their current form, peripheral stent grafts require relatively large introducer sheaths (minimum of 7 French, up to 12 French). Those for the aorta require very large introducers (up to 30 French) and thus necessitate surgical access to the arterial entry site (usually common femoral artery). Embolic agents can be conveniently divided into three categories, based on (1) location within the vascular bed where they occlude, (2) permanence of occlusion, and (3) radiopacity. Agents that occlude large vessels are considered to be proximal agents; chief among these are coils. Distal agents are smaller and flow, for example, into the nidus of an arteriovenous malformation or into a tumor bed for occlusion of
1689
small vessels. Chief among these agents are particles and liquids. A temporary agent is absorbed by the body and is principally represented by gelatin sponge. However, recanalization around an agent should also be considered temporary, as can occur with polyvinyl alcohol sponge. The major available embolic agents are outlined in Table 24.1 .
Medications A number of medications are used in the interventional suite for conscious sedation. Probably the two most common are fentanyl (Sublimaze: 25 to 100 µg bolus, 25 to 75 µg maintenance IV) and midazolam (Versed: 0.5 to 2 mg bolus, 1 mg maintenance). If these medications are administered, it is essential that the reversal agents are understood. Fentanyl is reversed with naloxone (Narcan: 0.4 to 2 mg IV). Midazolam is reversed with flumazenil (Remazicon: 0.2 mg IV over 15 seconds, with additional doses as required). Macrocoils (0.035–0.038 Stainless steel
inch)
Platinum Proximal Excellent Permanent Microcoils* Platinum
(0.010–0.018
inch)
Proximal Excellent Permanent Polyvinyl alcohol sponge Denatured ethanol particles Distal, based on size of particles, to arteriolar level Mixed with contrast material Temporary Detachable balloons* * Silicone balloon Balloon inflated to size Inflated with isosmolar contrast
1690
Permanent Glue Polymerization
of
cyanoacrylate
Distal based on rate of polymerization Mixed with Ethiodol (ethiodized oil) and tantalum powder Permanent Alcohol Sclerosing agent Distal, to capillary level Mixed with contrast powder Permanent Gelatin sponge Derivative of purified pork skin Proximal, based on size of pieces None, except when injected with contrast Temporary Microspheres Acrylic polymer Distal, based on size of spheres, to arteriolar level Mixed with contrast material Unknown *Coils are placed in the desired location by being pushed through a diagnostic
angiography
catheter
or
microcatheter.
Although
usually
effective, the coil is not easily controlled (i.e., cannot be easily retrieved after exiting from the catheter). There are ways to control coil delivery electronically or mechanically, which are most often used in neurointerventional procedures and are not given in this table. **Detachable balloons are currently not available in the United States. Embolization Agent
Constitution
Site of Occlusion
Radiopacity
Period of Occlusion
TABLE 24.1 Summary of Embolization Agents Antibiotic somewhat
prophylaxis for vascular and interventional radiology is controversial. However, most clinicians suggest
administration at least for contaminated (the presence of inflammation
1691
consistent with infection but no gross pus) or dirty (infected purulent site or infected GI or genitourinary (GU) site) procedures. Intra-arterial
pharmacoangiography
consists
of
either vasodilators or
vasoconstrictors . Vasodilators are used to treat vasospasm whose etiology is either iatrogenic (catheter-induced) or from other causes (trauma, medications, etc.). Two vasodilators are commonly used in vascular radiology: nitroglycerin (given intra-arterially in 100-µg doses) and papaverine (given intra-arterially in 25- to 100-mg doses). The only vasoconstrictor used with any frequency in vascular radiology is pitressin (Vasopressin), which was historically given intra-arterially for transcatheter therapy of lower GI bleeding, which has been mostly supplanted by embolization. Antithrombotic agents fall into two broad categories: anticoagulants and antiplatelet agents . Anticoagulants are pharmacologic agents that inhibit thrombin generation in vivo and are usually heparin intravenously and warfarin (Coumadin) orally. These two anticoagulants have been the main stays of antithrombotic therapy for years. Newer anticoagulants such as direct thrombin inhibitors (bivalirudin, hirudin, argatroban) might offer significant advantages over heparin, but further studies are needed to demonstrate their safety and effectiveness. Antiplatelet
agents P.674
are either oral or IV. Oral antiplatelet medications include acetylsalicylic acid (either 81 mg or 325 mg) and the thienopyridines, which include ticlopidine (Ticlid: 250 mg) and clopidogrel (Plavix: 75 mg). Intravenous antiplatelet agents consist of the glycoprotein IIb/IIIa antagonists, which are the “big gun― antiplatelet agents. The best known of these is abciximab (Reopro). Antiplatelet agents have demonstrated clinical benefits in coronary interventions, particularly following stent placement. Very little data exist regarding use of these agents for peripheral interventions. However, they are used in selected situations to decrease the likelihood of thrombus formation. Streptokinase Streptase Indirect activation
of
plasminogen
conversion
1692
10,000-unit bolus 5,000 units/hour 18 Urokinase Abbokinase Tissue plasminogen
activator
100,000- to 250,000-unit bolus 100,000–200,000 units/hour 15 Alteplase (rt-PA ) Activase Tissue plasminogen activator ≤2 mg/hr infusion 50% have cardiovascular complications. There is weakening of the aortic root, producing aortic ectasia and aortic insufficiency, making the patient prone to aortic dissection. Dissection or LV failure causes death in one third of patients by 32 years of
1708
P.681 age and in two thirds by age 50 if untreated. The classic aortogram appearance is that of a very large aneurysmal aortic root with sinotubular ectasia (the “tulip bulb― appearance) (Fig. 24.8A ). When present, aortic dissection involves the ascending aorta with or without extension into the descending (Fig. 24.8 ).
FIGURE 24.8. Aortic Dissection. A. Aortogram of a 47-year-old man with Marfan syndrome. Oblique arch view shows classic tulip bulb appearance of the aortic root (arrow ). Dissection flap (arrowheads ) is
1709
noted from the aortic root across the arch and into the descending aorta. B . Upper abdominal angiogram shows spiraling course of the dissection. This lumen, the true lumen, gives rise to the superior mesenteric artery (arrow ) and the left renal artery (arrowhead ). C . Later of image from abdominal aortogram shows intimal flap (arrow ) as well as filling of the right renal artery (arrowhead ) from the false lumen. (Courtesy Dr Joseph M. Stavas, Durham, NC.)
Ehlers-Danlos
syndrome is a genetically heterogeneous group of
heritable connective tissue disorders characterized by hyperextensible joints and tissue fragility. Multiple types have been described and categorized. Type IV, the type of interest here, has a defect in type III collagen presenting the characteristic vascular features, although other types have been reported to have vascular problems. Classically, type IV does not have the hyperextensibility of the large joints, although smaller joints may be minimally hypermobile. Angiographically, it tends to involve the ascending aorta, resulting in aneurysms, which are prone to dissection and rupture. Angiography should be carried out carefully as vessels are very thin and can even be perforated during catheterization. Aortic
dissection represents a laceration of the aortic intima and the
inner layer of the media, resulting in a cleavage of the aortic media. Blood penetrates the aortic wall via the primary entry site and dissects the medial layers for a variable distance both upstream and particularly downstream, creating a false lumen. Blood flow may occur in both the true and false lumens, but either may be thrombosed. When blood flow exists in both lumens, there P.682 are one or more re-entry points of the false lumen into the true lumen. Aortic rupture characteristically occurs at the site of the primary entry and is the most common cause of death, with early mortality as high as 1% per hour if left untreated. Approximately three fourths of all cases have involvement of the ascending aorta, arch, or both at autopsy. Less than 25% begin beyond the arch and 25% to 45% of dissections originate in the ascending aorta
1710
and reach the abdominal aorta. The dissection plane usually spirals as it courses downstream and may take any course. However, the typical course of an extensive dissection is usually described as the false aortic channel, expanding on the right in the arch and disrupting the right coronary artery. It then courses along the superior aspect of the arch, often involving the great vessels. If it extends distally, the false lumen most often courses to the left, involving the left renal artery. Still more distally, it tends to continue on the left side of the abdominal aorta and into the left pelvic system. Although dissection into the great vessels is quite a common finding, neurologic symptoms occur in only 20% of patients dying from dissections. There are two basic classification systems for aortic dissection, which are based on extent of involvement. In the DeBakey classification , a type 1 dissection begins in the proximal aorta and courses into the descending thoracic aorta, a type 2 dissection is limited to the ascending aorta, and a type 3 dissection is limited to the descending aorta.
The Stanford
classification is based upon whether or not the
ascending aorta is involved: a type A dissection involves the ascending aorta, and a type B dissection does not involve the ascending aorta. This classification scheme is based upon the need for surgical treatment of ascending aortic dissection. There are a number of etiologic factors associated with aortic dissection. Cystic medial degeneration may be the chief predisposing factor in aortic dissection. Hypertension is present in 80% of surgical patients treated for aortic dissection and appears to be the most important predisposing factor. Atherosclerosis is present in up to two thirds of patients with an aortic dissection, although it may be coincidental rather than causative of aortic dissection. Other etiologic factors include inflammatory diseases (aortitis), blunt trauma, and iatrogenic trauma, including patients receiving catheterization, particularly patients with intra-aortic balloon counterpulsation devices. Congenital anomalies and inheritable disorders of elastic tissue (Marfan, Turner, Ehlers-Danlos syndromes) and congenitally abnormal aortic valves (particularly bicuspid aortic valve) are also associated with aortic dissection. There are four modalities for imaging the thoracic aorta for the
1711
diagnosis of dissection: US, CT, MR, and catheter angiography. Excellent results, including sensitivities and specificities of greater than 90%, for all three noninvasive modalities have been reported. CT and MR have become the diagnostic imaging studies of choice, although transesophageal echocardiography is also applicable. However, because of its 24-hour availability and lack of invasiveness, CT is the most often employed
imaging
study.
Angiography was long considered the diagnostic standard for the evaluation of aortic dissection. However, prospective studies have found that for the diagnosis of aortic dissection, although the overall sensitivity of aortography is about 90%, it falls to only 77% when the definition of aortic dissection included intramural hematoma with a noncommunicating dissection. It does have advantages in that it is able to delineate the extent of the aortic dissection, including branch vessel involvement, the presence of aortic regurgitation, and patency of the coronary arteries. It is most often performed when stent grafting is being considered in the thoracic aorta or percutaneous fenestration for the abdominal aorta. Complete diagnosis of aortic dissection requires visualization of both a true and a false lumen (Fig. 24.9 ). A supportive but incomplete finding for aortic dissection by angiography is compression of the true lumen by the unopacified false lumen. A number of important factors should be analyzed using angiography, including the extent of dissection, identification of the primary intimal tear (entry site), re-entry site(s), status of the aortic valve, and assessment of brachiocephalic and visceral vessels. Although a major point of aortography has been identification of the coronary vessels in relation to the exact site of intimal tear, most surgeons can visually inspect for these structures during surgery. The classic angiographic finding of a “double barrel― aorta with an interposed intimal flap is seen in 87% of cases. The intimal flap usually begins in the right anterolateral ascending aorta and spirals to the left posterolateral aspect of the descending aorta into the abdomen (Fig. 24.8 ). Thus the left renal artery is frequently supplied by the false lumen, and the left iliac artery is more commonly involved when the
1712
dissection extends distally. Flow within the false lumen is slow, leading to late filling of branch vessels with their origin from this lumen. Thrombus in the false channel (25% of patients) appears as thickening of the aortic wall up to 1 cm. The true lumen is compressed and narrowed by the false channel in 85% of cases, deviating the course of a catheter. Most diagnostic modalities for aortic dissection are aimed at the acute proximal dissection, as such a dissection requires emergency surgical intervention to prevent rupture into the pericardium. Surgery for more distal dissections (arch and beyond) in the acute setting continues to be controversial. Certainly those with perforated descending or abdominal aortas would require emergency surgery, as would those with mesenteric ischemia. However, if clinically stable, most patients are managed medically, which results in the chronic dissection often seen in imaging. P.683 The strength of angiography may be in its possibility for endovascular therapy. Although still early in the experience, promising results have been obtained and, of course, endovascular surgery avoids the risks of major thoracic surgery. However, stent grafting is limited to the descending thoracic aorta only; an ascending injury still requires open repair. Endovascular fenestration of the aorta is a method of creating an opening in the intimal flap to allow blood flow into both lumens, preserving side branch patency. Although often performed in the thoracic aorta, it is indicated for acute abdominal and distal limb ischemia and is discussed along with the abdominal aorta (see Chapter 2 5 ). Although the roles of stent grafting and fenestration are yet to be proven, they will have an impact on future therapeutic strategies.
1713
FIGURE
24.9. Thoracic Aortic Dissection. A. This 68-year-old man
presented with chest pain. CT of thorax demonstrates the typical intimal flap of an aortic dissection (arrow ). B . Lateral aortogram shows filling of the true lumen (arrows ), which is compressed by the false lumen. The false lumen fills on later images.
Within the differential diagnosis of aortic dissection are the intramural hematoma and penetrating
aortic
ulcer . These three entities share
much in common and in fact together constitute the acute aortic syndrome . An intramural hematoma represents a localized hematoma within the aortic wall. This usually occurs in the elder, hypertensive patient and may represent a controlled dissection, although not all experts agree on this point. It is thought to represent a localized dissection without an identifiable entry/re-entry point. It may, however, progress to dissection. Angiography plays little role in the diagnosis. CT, MR, or US are the diagnostic tests of choice and demonstrate the characteristic intramural hematoma. An atherosclerotic plaque may ulcerate into the media, resulting in a penetrating aortic ulcer . The presentation is usually an elderly hypertensive patient with marked atherosclerotic disease. The diagnosis is best made by CT, which
1714
demonstrates aortic ulcer, frequently associated with an intramural hematoma (Fig. 24.10A ). Penetrating ulcer of the thoracic aorta is defined as an atherosclerotic lesion of the descending thoracic aorta with ulceration that penetrates the internal elastic lamina, allowing hematoma formation in the media (Fig. 24.10B ). There is controversy regarding whether this lesion differs from classic acute type III aortic dissection. The plaque may precipitate localized intramedial dissection associated with a variable amount of hematoma within the aortic wall, may break through into the adventitia to form a pseudoaneurysm, or may rupture completely into the right or left hemithorax. The diagnosis is made at CT with demonstration of a contrast material–filled outpouching in the aorta in the absence of a dissection flap or false lumen and often in the presence of extensive aortic calcification. Although aortography once was the standard for the diagnosis of many aortic diseases, it has largely been replaced by CT and MR. However, angiography still plays a role when endovascular therapy with a stent graft is employed. Penetrating ulcers appear to have a greater propensity to rupture in the acute setting during conservative treatment. Thus, aggressive management is P.684 recommended for penetrating ulcers, and a high index of suspicion must be maintained for rupture.
1715
FIGURE
24.10. Penetrating Aortic Ulcer. A. CT of a 73-year-old man
with chest pain shows an area of contrast filling with surrounding area of hemorrhage (arrow ). Areas of calcification in the aorta signify underlying
atherosclerotic
disease. B . Oblique descending thoracic
aortogram shows the ulcer crater filling with contrast material (arrow ). Traditionally, treatment was effected by open surgery, but today this condition can be effectively treated with stent grafting.
PULMONARY
ANGIOGRAPHY
Pulmonary angiography is usually performed from the common femoral vein, but it can be performed from the internal jugular or brachial/axillary veins. Specially shaped pulmonary catheters in a 5 to 7 French multi-sidehole pigtail design with a near right-angle curve (such as the Grollman catheter) can be easily placed into the right or left pulmonary artery. Traditionally, pressures are obtained in the pulmonary artery as well as in the RA and RV. Such pressures have diagnostic value (pressures reflective of right heart function) and there has been controversy regarding perceived complications of pulmonary angiography at higher arterial pressures (discussed subsequently). Nonionic, low-osmolar contrast material is used, resulting in fewer
1716
complications and a decrease in cough reflex. Imaging is performed digitally today, with acquisition of at least 6 frames/second. The anatomy of the pulmonary arteries is variable but for the most part follows the bronchi. Complications of pulmonary angiography have mostly been reported for the diagnosis of pulmonary embolism and are felt to be increased in the presence of pulmonary artery hypertension (usually defined as pulmonary artery systolic pressure >30 mmHg). It has been reported that pulmonary angiography is contraindicated in patients with high pulmonary artery pressures and left bundle branch block. The presence of pulmonary artery hypertension can result in right heart strain, which is exacerbated by contrast injection. However, it is not problematic with low-osmolar agents injected into the right or left pulmonary artery rather than the main pulmonary artery. The presence of an existing left bundle branch block is problematic because of its possibly inducing a right bundle block during catheterization of the right heart, resulting in total heart block. Transvenous or external pacing is recommended in this group of patients. There may be a host of indications for pulmonary angiography in a particular patient. Such indications may include trauma, congenital anomalies (particularly with congenital heart disease), tumor encasement of vessels, pulmonary hypertension (primary), vasculitis, and stenosis. However, two overriding indications in the typical interventional radiology practice are pulmonary embolism (PE) and pulmonary
arteriovenous
malformations (PAVMs). P.685
Pulmonary artery aneurysms deserve a brief discussion, even though they are rare, and even more rarely require pulmonary angiography, which is most often performed in anticipation of endovascular therapy. There are multiple etiologies for pulmonary artery aneurysms. The most striking is associated with tuberculous infection, forming the Rasmussen aneurysm. Antibiotic therapy has all but eradicated this in the United States, where the most common cause of a pulmonary artery aneurysm (pseudoaneurysm) is from iatrogenic trauma related to placement of a pulmonary artery catheter (mostly of the Swan-Ganz type).
1717
FIGURE 24.11. Pulmonary Embolism (PE). A. CT of a 78-year-old man with hypoxia demonstrates filling defect in the right and left pulmonary arteries, diagnostic of PE (arrows ). B . Left pulmonary artery injection of a 72-year-old man who fainted demonstrates intraluminal filling defects (arrows ) and areas of occlusion (arrowheads ), diagnostic of acute PE. C . Right pulmonary angiogram of a 35-year-old woman with severe shortness of breath shows enlarged main and right proximal pulmonary arteries, with pruning of vessels distally and areas of narrowing (webbing) (arrows ); this is diagnostic of chronic PE.
1718
Pulmonary
Embolism
It has been estimated that PE occurs in approximately 650,000 patients annually in the United States and contributes to up to 50,000 deaths. It is said to be responsible for up to 15% of all in hospital deaths. PE disease can be divided into chronic and acute forms based on history and angiographic appearance. Pulmonary angiography for acute PE has been all but replaced by multislice CT and is now often relegated to difficult diagnostic situations (Fig. 24.11A ). There are still some advantages of pulmonary angiography. Although it is certainly less than perfect, pulmonary angiography is P.686 the imaging “gold standard.― It allows visualization of the pelvic veins and inferior vena cava (IVC), and it provides hemodynamic parameters (pulmonary and right heart pressures) as well as an opportunity for therapy in the same sitting (filter, thrombolysis). The main disadvantage of pulmonary angiography is its invasive nature, which is not only uncomfortable for the patient but has a finite complication rate, including arrhythmias, cardiac injury (perforation), cardiac
arrest,
respiratory
hematoma/thrombosis,
and
insufficiency, even
contrast
reactions,
access
death.
The diagnosis of acute PE by pulmonary angiography is only reliable when intraluminal filling defects or an occluded pulmonary artery with or without a trailing edge of clot are identified (Fig. 24.11B ). Less reliable findings include area(s) of decreased flow, abnormal parenchymal stain, presence of collateral vessels, and delayed venous return. Pulmonary angiography should be performed as soon as possible, as the body tends to dissolve thrombus at a variable rate. Approximately 0.1% to 0.2% of patients with acute PE develop chronic pulmonary hypertension. Pulmonary angiography for chronic pulmonary embolic disease is usually performed to confirm the diagnosis and for surgical planning for pulmonary endarterectomy. Chronic pulmonary embolic disease can be suggested from CT or MR angiography findings, but the diagnosis is confirmed by pulmonary angiography. Diagnosis of chronic PE by pulmonary angiography is based on the identification of
1719
webs, luminal irregularities, areas of abrupt vessel narrowing and/or obstruction, and dilated central pulmonary arteries consistent with arterial hypertension (Fig. 24.11C ). These findings are usually bilateral. Pulmonary angiographic techniques are the same as for acute PE. Thrombolytic therapy for acute PE has the goal of rapid clot dissolution, resulting in greater pulmonary perfusion, thus providing improved hemodynamic (right heart) status and better gas exchange. Complete clot resolution should also serve to decrease chronic vascular obstruction, hopefully preventing chronic pulmonary hypertension. All of these should reduce the morbidity and mortality of PE. Unfortunately, most of this is currently unproven. In theory, catheter-directed thrombolytic therapy should be superior to IV administration because the agent is concentrated to the region of concern and continued until thrombus has been significantly reduced. Unfortunately, the results are not clear on these points, and intra-arterial administration of thrombolytic therapy for acute PE is limited to patients who are severely ill and in need of rapid thrombus dissolution. Further studies are needed, but early data do not support the use of local thrombolytic agents over IV administration except in highly selected cases. There are currently available a number of mechanical thrombectomy devices that serve to debulk (break up) the thrombus. Small series have been published in which these instruments were applied to PE. Although the theory is very attractive, the data are scarce, very early, and completely
uncontrolled.
Pulmonary arteriovenous malformations represent direct, lowpressure, artery-to-vein connections (fistulas) of the lung. Although they are associated with hereditary hemorrhagic telangiectasia (HHT) (also called Rendu-Osler-Weber syndrome) in 60% to 90% of reported cases, PAVMs may occur spontaneously (without HHT), or they may be associated with other causes such as trauma or erosion of a vessel by aneurysm, infection, or tumor. The clinical presentation of PAVMs may be difficult to discern, because only 72% of patients have symptoms referable to the PAVM or underlying HHT. The presence of symptoms correlates best with lesion size. A single AVM Table of Contents > Section VI - Vascular and Interventional Radiology > Chapter 25 - Abdominal Arteries, Venous System, and Nonvascular Intervention
Chapter
25
Abdominal Arteries, Venous System, and Nonvascular Intervention Michael J. Miller Jr. Tony P. Smith
ABDOMINAL
AORTA
AND
Abdominal Aortography Intervention
ITS
BRANCHES
and
Although individualized to a particular patient and their clinical situation, angiography of the abdominal aorta is most often performed for atherosclerotic disease, both aneurismal and occlusive. Angiography is also performed for aortic dissection as well as trauma. Rarely, involvement with the vasculitides, including Takayasu arteritis, midaortic syndrome, and other etiologies, necessitates angiography.
Aneurysms As with the thoracic aorta, there are multiple possible etiologies for abdominal aortic aneurysms, but two are of primary importance: atherosclerosis and infection. The most common etiology of
1751
abdominal aortic aneurysms (AAAs) is degenerative, which for the most part is synonymous with atherosclerotic. An atherosclerotic AAA is defined as enlargement of the aorta at least 1.5 times greater than the normal vessel diameter. Atherosclerotic AAAs are for the most part fusiform and often lined with mural thrombus. Although US does demonstrate the aneurysm, CT has become the diagnostic study of choice. Angiography is usually only indicated for a particular patient and their special situation. Angiography only demonstrates the true lumen, not the portion of the aneurysm, which is thrombus filled, and, of course, has a finite complication rate, as does any invasive procedure. Angiographically, an AAA is seen as an irregular, often calcified, fusiform aneurysm (Fig. 25.1A). Angiography does demonstrate the patency of the other major vessels (renals, visceral, iliacs), as well as their relationship to the AAA. Of greatest importance is the relationship of the renal arteries to the AAA, as it influences the surgical or endovascular repair. The indication for the elective repair of an asymptomatic AAA is when the diameter exceeds 5.0 cm. It is at this diameter that the chance of rupture increases dramatically. It is common for an AAA to extend into the iliac arteries, and 99% of atherosclerotic iliac artery aneurysms are associated with an AAA. Treatment of AAAs has been traditionally by open surgical repair. However, stent graft placement has become widely used because of its minimal invasiveness (Fig. 25.1). A number of stent grafts are available; all require large (up to 30 French) access sites via the common femoral artery for placement and therefore are most often placed using surgical access to one or both groins. The grafts differ by design including segments without covering, which can be anchored above the renal arteries as well as bifurcated sections for the iliac arteries. Stent grafts can be successfully placed into AAA in over 90% of cases. However, approximately 25% will require additional endovascular procedures. To that end, one of the major concerns with the placement of grafts is the continued P.701 filling (opacification) of the AAA following stent graft placement, termed endoleaks. Such leaks are best studied with CT (Fig. 25.1C)
1752
as well as by angiography, particularly as endovascular techniques are utilized to repair such leaks. Endoleaks are categorized into four types: type 1 is a leak at the superior or inferior attachment site, type 2 represents AAA filling via a patient arterial side branch such as a lumbar or the inferior mesenteric artery, type 3 is loss of integrity of the stent graft, and type 4 is a leak through the porous graft material. Isolated common iliac artery atherosclerotic aneurysms can be handled much like AAAs using smaller stent grafts. Internal iliac artery atherosclerotic aneurysms are probably best managed by embolizing the internal iliac aneurysm, primarily with coils.
FIGURE 25.1. Abdominal Aortic Aneurysm (AAA). A. Aortic arteriogram in a 68-year-old man with an infrarenal demonstrates an AAA (arrows). B . The aneurysm is completely excluded by a covered stent (stent graft) (arrows), which was placed below the renal arteries extending into both external iliac arteries (arrowheads). (Courtesy Andrew H. Cragg, MD, Minneapolis, MN.) C . CT scan of a different patient showing a stent graft in place but with persistent filling of the aneurysm, as noted by the contrast material (arrow). This is a type 1 endoleak occurring at the proximal attachment site of the graft. This was
1753
repaired by placing another shorter graft (cuff) over the site.
Mycotic aneurysms or pseudoaneurysms of the abdominal aorta are a rare but life-threatening condition and are often surgical emergencies. Radiographically, they often appear saccular and very irregular. Their rate of growth and the patient's constitutional symptoms will suggest the infectious nature. The commonest pathogen is Salmonella species, accounting for an incidence of up to 74%. Aortoiliac occlusive disease is most commonly caused by atherosclerosis. Patients with aortoiliac disease usually present with claudication. Bilateral buttock claudication, impotence, and absent femoral pulses are seen in the Leriche syndrome, although this title is often used to signify aortic occlusion by imaging rather than based on clinical symptoms. The aorta and iliac segments can be the source of distal emboli to the legs, such as the blue toe syndrome discussed later. Radiographically, atherosclerotic involvement of the distal aorta and iliac arteries is identical to that elsewhere, with plaque formation, calcification, and vessel narrowing. Visualization of the aorta and iliac arteries as well as the entire lower extremity can be carried out with CT or MR as well as angiography. CT and MR work well for screening, but angiography remains the gold standard, and endovascular therapy can be carried out in the same sitting. Aortoiliac occlusive disease is probably best initially approached with endovascular techniques, including angioplasty with or without stent placement (Fig. 25.2). Following angiographic assessment and prior to treatment of aortoiliac lesions, pressure measurements should be obtained; a gradient of 10 mm Hg systolic is considered significant. If no gradient is noted, pressures should be augmented, with vasodilators injected intra-arterially down the leg to simulate increased blood flow to chemically mimic exercise. Another pressure gradient is then obtained, again looking for a difference of 10 mm Hg P.702 or higher. Indications for angioplasty should take into account the
1754
patient's symptoms, appearance of the lesion, and pressure gradients.
FIGURE
25.2. Aortoiliac
Stenting.
A. Pelvic angiogram in a
55-year-old woman with severe claudication demonstrates severe aortic and bilateral proximal common femoral artery atherosclerotic disease (arrows). B . Angioplasty with stents placed “kissing― in the aorta provides an excellent anatomic
result.
Isolated aortic lesions are quite rare but can be effectively treated by endovascular techniques. Disease affecting both the aorta and proximal iliac arteries is usually treated with “kissing― balloons at the aortic bifurcation placed from each common femoral artery. Success can readily be achieved, even with complete occlusion of one or both iliac arteries. Stent placement for aortic or iliac disease is based on the success of angioplasty, as determined radiographically, by follow-up pressure gradients, or by intravascular sonography. Any residual gradient, vessel irregularity, or intimal flap is an indication for stent placement, although many clinicians stent these vessels primarily instead of attempting angioplasty alone. Complications of angioplasty are the same as angioplasty at other sites and include
1755
acute thrombosis, distal embolization, and vessel perforation, which occur in fewer than 5% of cases. Aortoiliac occlusive disease can be caused by inflammatory
diseases,
in particular Takayasu arteritis, which produces a long segment, smooth narrowing of the abdominal aorta that may extend into branch vessels. Hypoplastic aortic syndrome is a congenital process of unknown etiology producing long segment narrowing of the aorta and is usually seen in young females. Neurofibromatosis may also involve the aorta, and the iliac arteries are the third most common location
for fibromuscular
disease.
Abdominal aortic dissection is almost always associated with, or is in fact an extension of, thoracic dissection. Abdominal aortic dissections are best imaged by CT, although MR does provide excellent diagnostic images. CT demonstrates both lumens, how well they fill with contrast material, the proximal and distal extent of the dissection, as well as the patency and relationship of branch vessels to the dissection. In addition to rupture, of greatest importance is how the dissection plane affects abdominal aortic branch vessels, including visceral, renal, and aortic bifurcation. Angiography is usually reserved for symptomatic patients prior to intervention. As with other imaging, it is essential that angiography demonstrate the extent of the dissection and the patency of branch vessels. Recently, a number of newer interventional radiologic or minimally invasive techniques appear to have significantly improved the management of patients with aortic dissection. These include stent grafting for entry site closure to prevent aneurysmal widening of the false lumen, as well as percutaneous techniques, such as balloon fenestration of the intimal flap and aortic true lumen stenting to alleviate branch vessel ischemia. Aortic fenestration is a technique whereby a long needle is introduced via the groin access. Using endovascular sonography, a puncture is made from one lumen to the other. Following placement of a guide wire, a balloon is used to create an opening between the two aortic lumens to equalize the pressure between the two, which should hopefully allow reasonable flow into branch vessels from either the true or the false lumen.
1756
Aortic fenestration is a procedure reserved for emergency situations. False lumen thrombosis following entry closure with stent grafts has been observed in 86% to 100% of patients, whereas percutaneous interventions are able to effectively relieve organ ischemia in approximately 90% P.703 of cases. In the years to come, it is to be expected that endoluminal techniques will become the method of choice for treating most type B dissections.
FIGURE
25.3. Traumatic Hemorrhage in the Pelvis. A. Pelvic
angiogram in a 22-year-old man following a motor vehicle crash reveals avulsion of the superior gluteal artery with extravasation (arrow). B . Embolization of the superior gluteal was performed with coils (arrow), and the hemorrhage was well controlled. This unsubtracted image shows pelvic fracture (arrowhead) and deviation of the bladder (B) by a large pelvic hematoma.
Trauma Patients rarely survive abdominal aortic trauma, but pelvic trauma resulting in significant bleeding is more commonly seen. Most pelvic bleeding associated with pelvic fractures can be controlled by fixation
1757
devices, but those that cannot routinely turn to angiography for diagnosis and embolotherapy. Arterial bleeding is most often associated with injury to one or more branches of the internal iliac arteries. Embolization is most often performed with proximal agents—either gelatin sponge if a temporary agent is desired, or coils for more permanent occlusion. Bleeding from pelvic injury can be effectively controlled with embolotherapy, and angiographic embolization is highly effective in controlling arterial bleeding associated with pelvic fractures (Fig. 25.3). However, repeat angiography should be performed in patients with pelvic fractures and ongoing evidence of hemorrhage, demonstrated by persistent base deficit and hypotension once other potential sources of bleeding have been excluded.
Renal
Angiography
and
Intervention
A single renal artery occurs approximately 60% of the time, with either multiple arteries or an early renal artery division in the other 40%. The most common reason to study the renal arteries in the United States is for occlusive disease and trauma; it is done less often for other abnormalities such as neoplasms. Imaging of the renal arteries includes US, CT, and MR, with nuclear scintigraphy also playing a role in the relative perfusion of the kidney. US allows visualization of the proximal renal artery and evaluates flow, including resistance to distal flow. CT and MR provide excellent images and are useful screening tools in many centers for renal artery stenosis. However, renal angiography remains the gold standard once only provides applied in the angiographic
screening suggests an abnormality. Angiography not diagnostic images, but transcatheter therapy can be same sitting and is often the primary goal for the procedure.
Renal arteries are often studied in patients with decreased renal function. In such patients, alternative imaging that does not require iodinated contrast is recommended, but when angiography is required there are several useful options. Alternative contrast agents such as carbon dioxide and gadolinium have been successfully used,
1758
particularly in conjunction with small doses of iodinated contrast. Alternatively, traditional low-molar or isosmolar iodinated contrast material may be used and kept to a minimum, especially if the patient has received preventative therapy. Chief among these is hydration, which is currently the only well proven renal protective maneuver, although agents such as fenoldopam and antioxidants (chiefly acetylcysteine) have been used. There are a number of etiologies for renal artery occlusive disease, including dissection, vasospasm, vasculitis, coarctation syndromes, and neurofibromatosis, but the two most common causes by far are atherosclerosis and fibromuscular disease (FMD). These two account for 99% of the stenoses encountered in the United States, with atherosclerosis representing 65% of stenoses overall (Fig. 25.4) . Fibromuscular disease is the most common cause P.704 of renovascular hypertension in patients younger than 40 years and represents an assortment of histologic patterns, which are usually classified into three main types based on the morphologic appearance and the layer of the involved arterial wall—intimal (7% to 8%), medial (85%), and periarterial (7% to 8%)—although there are also schemes that include subclassifications of each category. Medial fibroplasia accounts for the majority of cases and has the classic “string of beads― appearance on angiography, which represents alternating weblike stenoses and aneurysms (Fig. 25.5) . The middle and distal portions of the main renal artery are most frequently involved. The proximal renal artery is rarely involved alone. FMD is the most common cause of hypertension in children. It has been well shown that atherosclerotic renal disease is progressive, but it is also well documented that FMD is also progressive in nature. FMD classically responds well to angioplasty, with success rates approaching 98%. Stenting is rarely required, usually for dissections from angioplasty.
1759
FIGURE 25.4. Atherosclerotic Renal Artery Stenosis. A. Aortogram in a 68-year-old woman with hypertension shows bilateral renal artery stenosis that is more pronounced on the right (arrow) than the left. B . Following angioplasty, a significant intimal dissection (flap) (arrow) is noted. C . Aortogram following placement of a balloon-expandable stent shows an excellent radiographic result (black arrow). Left renal artery stenosis (white arrowhead) is better seen here.
Atherosclerotic
renal
occlusive
disease clinically presents with
1760
hypertension, renal failure, or both. Patients with atherosclerotic renal disease are usually over the age of 60 years, and the atherosclerotic lesions look like atherosclerosis elsewhere (Fig. 25.4). Atherosclerotic renal artery stenosis is amenable to angioplasty but is not as straightforward as simply dilating the lesion (Fig. 25.4). Anatomically, the atherosclerotic renal artery lesion is in both the proximal renal artery and the aorta, with the latter poorly responsive to angioplasty alone. Stenting of renal atherosclerotic lesions P.705 overcomes this problem for the most part and is now the standard, but even in light of an anatomically successful procedure, the patient may not respond very well. In fact, the vast majority of patients with hypertension (more than 95%) have essential hypertension, and etiology of renal insufficiency is often multifactorial. Whether treatment of a renal artery stenosis will actually produce a desired clinical result is difficult to predict. Renal vein renins can be measured, and elevated renin values do tend to correlate with a positive clinical result but have little predictive value when not elevated. Even pressure gradients across the stenosis are not very predictive. Therefore one can see that the results are never going to be extremely successful, as one never really knows whether the disease process (hypertension and/or renal failure) is actually caused by the renal artery stenosis until treatment is undertaken. To that end, the results of renal angioplasty are often reported all antihypertensive medications), improved (requiring less medication), or failed (no change or worsened by Based on meta-analysis data, with a mean of almost 2 follow-up, approximately 20% of patients are cured of
as cured (off significantly treatment). years of hypertension
with stenting, compared to only 10% following percutaneous angioplasty alone. Improvement is about the same with angioplasty alone (53%) versus stent placement (49%). Improvement in renal function is greater with angioplasty alone (38%) versus stent placement (30%). However, restenosis rates are better with stent placement (17%) when compared to angioplasty alone (26%). Angioplasty alone is usually performed for renal artery FMD, with cure rates usually in the range of 25% and improvement in the vast
1761
majority of the others.
FIGURE
25.5. Fibromuscular Disease of the Right Renal
Artery. Arteriogram in a 38-year-old with hypertension shows the typical beaded appearance of the right renal artery, diagnostic of medial fibroplasia. This is the classic “string of beads― appearance, which stenoses and aneurysms.
represents
alternating
weblike
Finally, as with any invasive procedure, there can be complications. Unfortunately, complications from renal angioplasty are not rare, with an overall complication rate of approximately 5% to 10%, most commonly worsening renal function and injury to the renal artery. In spite of the difficulties with results and the possible complications from angioplasty, endovascular therapy remains the best available treatment following medical failure. The reason to image the renal arteries in a patient with hypertension and/or renal failure is to look for a treatable cause, and a large portion of that treatment is angioplasty
and
stenting.
1762
Neurofibromatosis causes renal artery stenosis by extrinsic compression of the renal artery by neurofibromata or from disorganized intimal and medial proliferation at the renal artery orifice or in the proximal renal artery. Angiography demonstrates smooth or nodular stenoses with or without associated aneurysms. Hypertension secondary to neurofibromatosis is seen mainly in children. Renal transplant artery narrowing is most often caused by surgical technical factors (acutely), intimal hyperplasia at the anastomotic site (late), or even atherosclerosis (much later), all of which may be amenable to balloon angioplasty with or without stenting. Renal
artery
aneurysms, exclusive of trauma, are rare (less than
0.1%). Aneurysms are basically of two types: extrarenal, caused by atherosclerosis and FMD; and small multiple ones within the kidney, mostly indicative of polyarteritis nodosa (PAN). For unruptured hilar aneurysms, treatment is said to be indicated when the aneurysm exceeds 2 cm in size; such lesions are often amenable to endovascular therapy. PAN is a rare necrotizing vasculitis that affects the small and medium-sized arteries of multiple organs, most commonly the renal (85%) and hepatic (65%) arteries. Characteristic subcutaneous nodules are seen in 15%. The major angiographic findings are multiple, small, saccular microaneurysms; occlusions; and irregular stenoses throughout the abdominal viscera (Fig. 25.6) . Microaneurysms are seen in 50% of patients, range in size from 1 to 12 mm, and are typically located at branch points. The differential diagnosis of microaneurysms includes PAN, Wegener granulomatosis, systemic lupus erythematosus, rheumatoid vasculitis, and drug abuse. Angiography for renal neoplasms is rarely performed, most often in our
experience
for
preoperative
embolization P.706
prior to surgical resection or palliative for hematuria. Most hypervascular renal lesions are renal cell carcinoma, but definitive diagnosis cannot be made angiographically.
1763
FIGURE 25.6. Polyarteritis Nodosa. Right renal angiogram of a 37-year-old man with a history of hypertension and IV drug abuse demonstrates multiple small renal arterial aneurysms (arrows). Although these findings have been reported with IV drug abuse, the most likely radiographic diagnosis is polyarteritis nodosa, which was confirmed in this patient.
1764
FIGURE
25.7. Traumatic
Arteriovenous
Fistula.
A. Right
renal angiogram in a 50-year-old man who underwent percutaneous nephrolithotomy shows a lower pole arteriovenous fistula (arrow) at the catheter insertion site (arrowheads). B . The vessel was occluded with microcoils (arrows), and the follow-up angiogram shows a good radiographic result.
Trauma Injury to kidney is not rare in either penetrating or major blunt trauma. A full range of vascular injuries can be seen from hematuria without visible injury (grade 1) to a shattered kidney or renal hilum avulsion (grade 5). Angiography is most often performed in these patients to determine the extent of injury and therapeutic planning, which may be surgical or via endovascular means. A special type of trauma to the kidney, of course, is iatrogenic, including biopsy or catheter placement (Fig. 25.7). This also includes the renal transplant, which might be undergoing repeated biopsies. Angiography is performed to determine the site and nature of the injury, which could be pseudoaneurysm, arteriovenous fistula, or even extravasation. Most of these injury sites are amenable to
1765
endovascular and coiling.
Splenic
treatment,
most
often
Angiography
by
subselective
and
catheterization
Intervention
The splenic artery arises from the celiac artery at the hepatosplenogastric trunk. In 25% of the population, the left gastric, splenic, and hepatic arteries arise as a tripod celiac. Infrequently, the splenic may arise from the aorta or the superior mesenteric artery. The splenic artery tends to become increasingly tortuous with age. Common branches off the splenic artery include the dorsal pancreatic, arteria pancreatica magna, caudal pancreatic, left gastroepiploic, short gastric, and splenic polar branches. There are few indications for splenic arteriogram and intervention. Prior to intervention, the patient should be informed and consent obtained. Potential complications should be discussed with the patient and the patient's family. Complications of any organ embolization also apply to the spleen, including dissection or vascular injury. Specific to the spleen are pancreatitis and splenic infarction with abscess formation. Patients should be given pneumococcal vaccine prior to the procedure, when time allows. Trauma is a common indication for arteriographic imaging of the spleen to confirm vascular injury. The risk of delayed rupture increases with severity of splenic injury. Embolization of the splenic artery is performed as close to the injury as possible to preserve splenic tissue through collateral supply from pancreatic and short gastric collaterals (Fig. 25.8). When acute extravasation or a vascular abnormality such as pseudoaneurysm or fistula is discovered, subselective arteriography and focal occlusion of the branch of the splenic artery are performed. Coils are preferred for embolization of the main splenic artery or its branches. Delayed complications include rebleeding and abscess, which mandate splenectomy. Hypersplenism is caused by hemolytic anemia, splenic vein thrombosis, portal venous hypertension, tumor, infiltrative diseases, myelofibrosis,
and
polycythemia
vera.
1766
Anemia,
thrombocytopenia,
splenomegaly causing discomfort, and gastric varices caused by splenic vein thrombosis are indications for partial splenic infarction. P.707 Embolization is performed with a distal agent, such as Gelfoam (absorbable gelatin-derived sponge), 355- to 500-µm polyvinyl alcohol particles, or 500- to 700-µm calibrated gelatin microspheres, until 60% to 70% of the splenic tissue is ablated. The catheter should be distal to the pancreatic branches prior to infusion of the embolic agent to prevent pancreatitis. Splenic abscess is a complication of splenic embolization.
FIGURE 25.8. Splenic Trauma. A. Splenic arteriogram demonstrates diffuse injury of the spleen, with multiple areas of extravasation of contrast (arrowheads). B . Completion arteriogram shows occlusion of the splenic artery following coil embolization.
Splenic artery aneurysm is the most common aneurysm outside of the aorta and iliac arteries. They are seen between the third and sixth decade of life and are more common in women. Congenital aneurysm has a higher rate of rupture. The diagnosis is made with CT or MR. Aneurysms and pseudoaneurysms are treated with coil obliteration, trapping of the arterial segment with coils, or occlusion
1767
balloons (Fig. 25.9). Blood flow to the spleen is usually preserved through short gastric collaterals. Multiple aneurysms of the main and branch vessels of the splenic artery can be seen in the setting of cirrhosis.
FIGURE
25.9. Splenic
Artery
Pseudoaneurysm.
A. Celiac
arteriogram demonstrates a large pseudoaneurysm (arrowheads) from the splenic artery in a patient with pancreatitis. B . Arteriogram following coil embolization shows a small residual neck (arrowhead) .
Hepatic
Angiography
and
Intervention
The common hepatic artery arises from the celiac trunk. The gastroduodenal artery (GDA) is its first major branch, with the artery continuing as the proper hepatic artery. Branch arteries parallel the portal veins and supply the liver segments. In approximately 55% of people, the right, left, and middle hepatic arteries arise from the common hepatic artery. The cystic artery often arises from the right posterior branch. The middle hepatic artery may arise from either the left or right hepatic artery. It supplies liver segments IVa and IVb. P.708 In approximately 2.5% of people the common hepatic artery arises from the superior mesenteric artery. An aberrant right hepatic artery
1768
exists in up to 26% of people. The most common variations are either a replaced right hepatic artery (Fig. 25.10) or an accessory artery arising from the superior mesenteric artery (SMA). An aberrant left hepatic artery exists in up to 25% of people. The left hepatic artery may arise from the left gastric artery (15%) or from the gastroduodenol artery, splenic artery, or aorta in 4% of cases. Hepatic arteriography is most commonly performed today for trauma and neoplastic disease, usually as a precursor to transcatheter intervention.
FIGURE
25.10. Replaced Right Hepatic Artery. Superior
mesenteric arteriogram demonstrates a replaced right hepatic artery (arrowhead) arising from the superior mesenteric artery.
Trauma Hepatic arteriography is used to increase the sensitivity for vascular injury, intervene upon a vascular injury demonstrated on screening CT examination, or treat a delayed complication from conservative management of a liver laceration (Fig. 25.11). Penetrating trauma is often iatrogenic, resulting from biopsies or percutaneous cholangiography with or without drainage. Indications for
1769
embolization in the acute phase include continued hemodynamic instability, arterioportal fistula, and hemobilia. In the delayed setting, pseudoaneurysm discovered on CT or US carries a 44% risk of rupture and may require emergent management.
FIGURE 25.11. Active Hemorrhage: Liver Laceration. A. CT shows active hemorrhage with contrast extravasation (arrows) from liver lacerations caused by blunt trauma. B . Hepatic arteriogram demonstrates multiple arterial injuries, indicated by contrast extravasation (arrowheads). C . Arteriogram through a microcatheter (black arrow) shows extravasation (black arrowhead). Previously placed embolization coils (white arrows) are evident.
Neoplasms are diagnosed by CT, MR, and US. Angiography is performed to determine resectability, provide an arterial roadmap, or
1770
to
deliver
transcatheter
therapy.
Capillary hemangiomas demonstrate uniform dense staining in the late arterial phase, which persists beyond the venous phase. They usually have well-defined (but irregular) borders with a feeding artery, which is near normal in size. Cavernous hemangiomas have the classic appearance of contrast puddles near the periphery in wellmarginated vascular spaces, while the stain persists beyond the venous phase. The lesions may be up to 15 cm in size. The feeding artery is usually normal in size. Hemangioendotheliomas present in infancy, either with mass effect or hepatomegaly. Most (90%) are associated with extrahepatic hemangiomas (cutaneous lesions). Lesions usually involute within 1 to 2 years. Treatment may be required if the lesions cause symptoms. Angiography shows dilated irregular vascular lakes, staining beyond the venous phase, and dilated
feeding
vessels.
Arteriovenous P.709
shunting and early opacification of hepatic veins are also described. Angiography of hepatoma demonstrates a hypervascular mass with large, distorted feeding arteries. Neovascularity, intratumoral puddling of contrast, and portal vein invasion with arterioportal shunting may be demonstrated. Up to 25% of tumors are hypovascular. The combination of portal venous invasion and arterioportal
shunting
is
pathognomonic
for
hepatoma.
Angiography
of cholangiocarcinoma demonstrates a hypovascular or avascular tumor without neovascularity. The most common malignant liver lesions are metastases. As shown on physiologic studies, the degree of vascularity and staining on angiography has little relation to tumor blood flow. Even with hypovascular metastases, the blood flow is increased relative to normal liver parenchyma. Angiography may show displacement of adjacent vessels and compressed or occluded portal veins. Arterial encasement or shunting is rare. Embolization of metastases has mixed results. Hypervascular metastases include neuroendocrine tumor, renal cell carcinoma, thyroid carcinoma, and choriocarcinoma. Hypovascular metastases include lung, esophagus, and pancreas carcinoma. Mixed vascularity is seen in breast carcinoma, ocular carcinoma, cholangiocarcinoma, and sarcoma.
1771
Embolization The decision to embolize a lesion involves a number of important points. First, is the portal vein patent, and in what direction is it flowing? Typically 70% of hepatic parenchymal supply is provided by the portal vein, while metastases receive blood flow primarily from the hepatic artery. Late-phase imaging of the portal vein following selective injection of the celiac, splenic, or superior mesenteric arteries can be used to confirm portal vein patency. Portal venous flow is needed to preserve functional hepatocytes in the field of planned embolization. Embolization of the hepatic artery is usually well tolerated if portal venous flow is available. Portal venous thrombosis or hepatofugal flow increases the risk of hepatic infarction. Embolization can be done in the setting of portal vein occlusion if a modified, low-dose, super-selective technique is used. Second, are you embolizing a tumor? Tumors, which are responsive to
embolization,
include
hepatocellular
carcinoma,
neuroendocrine
tumors, melanoma, sarcoma, and colorectal metastases (Fig. 25.12) . Tumor replacement of greater than 50% to 75% of normal liver is a contraindication to embolization. Tumors are treated with distal agents such as polyvinyl alcohol particles (100 to 300 µm or 300 to 500 µm) or calibrated gelatin microspheres (300 to 500 µm). Cytotoxic agents may improve outcomes from embolization. Most chemical embolizations are performed with a mix of Isovue (iopamidol), ethiodized oil, and cytotoxic agent in addition to the particles. Lipiodol may stay within hepatomas for up to a year while cleared from normal or cirrhotic liver within 4 weeks. Doxorubicin is used for neuroendocrine tumors, while colorectal metastases may be treated with fluorouracil and mitomycin. Attention to the cystic and gastroduodenal arteries must be taken to prevent nontarget embolization. Microcatheters facilitate subselective embolization. Patients are followed with CT or MR to detect contrast-enhancing viable tumor. Finally, is it trauma you are treating? Proximal agents are the theme in the treatment of hepatic arterial injury. Coil embolization of the vascular injury should be performed as selectively as possible to
1772
avoid complications, including infection, ischemia, and biliary stricture. Additional complications include pseudoaneurysm, fistula, and P.710 hemobilia. These can be treated with microcatheter subselective embolization of the source vessel. Cyanoacrylate (glue), large particles, and gelatin sponge may be used for vessels that cannot be selected primarily. Technical success is between 85% and 95 %.
FIGURE
25.12. Tumor
Embolization.
A. Celiac arteriogram
demonstrates a hypervascular mass (arrowheads) in the right hepatic lobe in a patient with metastatic carcinoid. B . Completion arteriogram following embolization with particles demonstrates absent enhancement of the mass (arrowheads) .
1773
FIGURE 25.13. Liver Transplant. A. Celiac arteriogram reveals a high-grade anastomotic stenosis (arrowhead) in a patient post–liver transplant. B . Balloon is in place, stenting the stenosis. C . Completion arteriogram demonstrates resolution of the stenosis.
Liver
transplantation is now a well-accepted surgical procedure,
and angiography plays a role in its planning and in the care for the patient posttransplantation. The initial planning of the transplant may be impacted by variant anatomy. Posttransplant care is usually indicated in the setting of hepatic failure or biliary strictures. Arterial anastomotic stenosis can be treated with angioplasty and stent placement (Fig. 25.13). The technique for angioplasty and stenting is covered in the aortoiliac section of this chapter. In some cases, the arteriogram precedes surgical revision. In the case of splenic arterial steal, management includes proximal embolization of the splenic
1774
artery or partial distal embolization with particles in the setting of hypersplenism. Polyarteritis
nodosa, as mentioned earlier, is a rare necrotizing
vasculitis that affects the small and medium-sized arteries of multiple organs, most commonly the renal (85%) and hepatic (65%) arteries. The major angiographic findings are multiple, small saccular aneurysms, occlusions, and irregular stenoses throughout the abdominal viscera.
Mesenteric
Angiography
and
Intervention The celiac artery, SMA, and inferior mesenteric artery (IMA) are the main arterial supply to the gastrointestinal tract. The celiac axis originates at the T12 level, giving rise to the splenic, common hepatic, and left gastric arteries. The common hepatic artery becomes the proper hepatic artery after giving off the gastroduodenal artery, which then branches into the superior pancreaticoduodenal (anterior and posterior) and right gastroepiploic arteries. The left gastroepiploic artery and short gastric arteries are distal branches of the splenic artery. The right gastric artery is a small artery with variable origin, usually from the proper or left hepatic artery. The left gastric artery supplies the distal esophagus and the majority of the stomach while running along the lesser curvature. The gastroepiploic arteries form an anastomosing arc along the greater curvature of the stomach, supplying the bulk of the remainder of gastric flow. The SMA originates at the T12/L1 level and supplies the entire small intestine and the proximal two thirds of the colon. The first branch is the inferior pancreaticoduodenal artery, which freely anastomoses with the superior pancreaticoduodenal artery to supply the duodenum. The remaining branches, in order of origin, are the jejunal, ileal, middle colic, right colic, and terminal ileocolic arteries. The middle colic divides into the left and right branches, which freely anastomose with the respective right and left colic (IMA) arteries. The ileocolic supplies the terminal ileum and cecum; the right colic
1775
supplies the ascending colon and hepatic flexure, and the middle colic supplies the transverse colon. The IMA originates at the L3 level and gives rise to the left colic, sigmoid, and superior hemorrhoidal (rectal) arteries. The superior hemorrhoidal branches freely anastomose with the hemorrhoidal branches of the internal iliac system. P.711 There are three collateral communications of the mesenteric vessels. (1) The marginal artery of Drummond provides anastomosis between the right colic, right and left branches of the middle colic, and the left colic arteries. It is found along the mesenteric border of the colon and is an important collateral supply in IMA occlusions. (2) The arc of Riolan is a variable communication between the SMA and IMA located more centrally in the mesentery than the marginal artery. (3) The arc of Buehler is a short, ventral artery between the main celiac and SMA and represents a persistent fetal communication. GI hemorrhage is the most common reason to perform angiography of the mesenteric vessels. The evaluation of GI bleeding includes nasogastric tube aspirate, esophagogastric duodenoscopy, colonoscopy, radionuclide imaging (tagged red blood cells and sulfur colloid), and angiography. The application of these modalities is dependent on the likely source of bleeding and the clinical status of the patient. Hemodynamically unstable patients may require emergency angiography and/or surgery, whereas a stable patient is able to undergo a more controlled, systematic evaluation and treatment. For suspected upper GI bleeding (proximal to ligament of Treitz), the evaluation begins with gastric aspiration and endoscopy. For lower GI bleeding, evaluation by colonoscopy or radionuclide imaging localizes the site of bleeding and guides the angiographic examination to the most likely vascular territory. The most reliable angiographic sign of GI bleeding is contrast extravasation, which is seen as an amorphous contrast collection that persists through the venous phase. If the bleeding rate is rapid enough, the extravasated contrast may outline mucosal folds. The pseudovein sign is a linear collection of contrast between mucosal folds that simulates an
1776
enlarged vein. Bleeding must occur at a rate of at least 0.5 mL/min to be identified by angiography. Upper GI hemorrhage is typically suspected in the setting of positive gastric aspirate and melanotic stool. First-line therapy usually consists of endoscopic evaluation and intervention. This can direct endovascular intervention when there is no angiographic evidence of bleeding. Etiologies for upper GI bleeding include Mallory-Weiss tear, hemorrhagic gastritis, gastric or duodenal ulceration, recent GI surgery, and tumor. The distribution for embolization is determined by the site of the lesion. Bleeding from the duodenum is managed with embolization of the gastroduodenal artery. Attention must be made to the retrograde filling from the pancreaticoduodenal artery. This is remedied by trapping the bleeding vessel or site. Gastric ulceration is treated with right, or more typically left, gastric arterial embolization. The technical success of embolization for upper GI bleeding is greater than 90%, while the clinical success rate ranges between 75% and 90%. Proximal agents such as coils and Gelfoam are preferred. In the setting of tumor, distal agents such as particles may be required for clinical success. In the setting of clinical failure, repeat embolization or surgical intervention may be used. Patients with prior surgical alteration require close attention if embolization is the chosen treatment, given that collateral supply to a region may be compromised, thus increasing the risk of bowel infarction. Tumor is the most common cause for bleeding from the small bowel; it is responsible for 20% to 50% of cases. Angiography depicts tumor neovascularity (enlarged, bizarre, irregular vessels with arteriovenous shunting) with or without contrast extravasation. Aortoenteric fistula accounts for 10% of small bowel bleeding and is usually a complication of AAA surgery as soon as 3 weeks postoperatively. The duodenum, where it crosses over the aorta, is the source in 80% of cases. Angiography demonstrates an anterior nipplelike projection from the aortic graft anastomosis or, rarely, contrast extravasation at the fistula site. Angiography is performed using an aortic injection and treatment is urgent surgery. Diverticula of the small bowel are an uncommon cause of small bowel bleeding.
1777
They are located along the mesenteric border of the bowel. The jejunum is a more common bleeding source than the ileum. Bleeding is typically slow and difficult to diagnose angiographically. Meckel diverticulum, the omphalomesenteric duct remnant, is found along the antimesenteric border in the distal ileum. Patients present with painless bleeding caused by an ileal ulcer adjacent to the heterotopic gastric mucosa contained in the diverticulum. A radionuclide Meckel scan is more sensitive than angiography, as this demonstrates the gastric mucosa. Inflammatory bowel disease is identified angiographically as diffuse hyperemia, arteriovenous shunting, and oozing. Vascular malformations are responsible for 20% of small bowel bleeding. They may be solitary or multiple, as seen in RenduOsler-Weber syndrome, and usually present as chronic, recurrent bleeds. Lower GI hemorrhage is caused most commonly by colonic diverticula. Although diverticula are much more common in the left colon, a bleeding diverticulum is three times more likely to be found in the right colon. Angiodysplasia shows the classic angiographic features of early opacification of an enlarged draining vein, persistent dense opacification of the vein, and vascular tufts along the antimesenteric border of the cecum or ascending colon. Treatment can be embolic or surgical. Tumors usually cause slow bleeding and anemia and are only infrequently the source of massive lower GI bleeding. When massive bleeding occurs, selective or superselective embolization with coils or particles may be performed if the patient is not a candidate for surgical resection. Treatment of colonic bleeding has shifted toward superselective embolization with coils or particles. Despite advances in microcatheter technology, the technical success has remained constant over the years. The use of superselective embolization has increased the technical demands of the procedure (Fig. 25.14). If the catheter can P.712 be advanced into the arcuate branches of the colon, collateral flow to the segment of bowel is preserved. This is thought to decrease the risk of infarction. The technical success is 70% to 100%, with a
1778
clinical success of 60% to 100%. The rate of recurrence is 19%. Coagulopathy, multiorgan failure, shock, and corticosteroids may cause clinical failure. The rate of minor complications is between 15% and 20%, while the major complication rate is 1% to 11%. In the modern series there has been no bowel infarction. This is likely related to the targeting of the vessels beyond the marginal artery. Treatment with vasopressin has been effective in lower GI bleeding but has fallen out of favor. Vasopressin administration involves placing a catheter within the source vessel and starting an initial 20minute infusion of 0.1 IU/min, increasing up to 0.4 IU/min if necessary. Also, infusion is quite time consuming relative to embolotherapy. Complication rates have been reported as high as 20% for major and 40% for minor. Complications include myocardial infarction, bowel infarction, groin hematoma, and catheter malfunction.
FIGURE 25.14. Lower GI Hemorrhage. A. Superior mesenteric artery arteriogram demonstrates extravasation (arrowhead) from a branch of the right colic artery. B . Completion arteriogram demonstrates superselective embolization of the segment (arrows) of the right colic artery, with preserved collateral flow (arrowhead) to the segment.
1779
Mesenteric
ischemia comprises a group of disorders that have a
common endpoint: bowel necrosis. The mortality rate approaches 70%. Mesenteric ischemia can be divided into acute and chronic varieties. Arterial embolism and thrombosis, nonocclusive ischemia, and mesenteric venous thrombosis are causes of acute mesenteric ischemia. Embolism and thrombosis account for 75% of acute ischemic episodes. Acute mesenteric ischemia may be precipitated by an embolic source, which is most commonly cardiac in origin. The angiographic appearance is of an abrupt cutoff of the SMA at the site of its most proximal branches—typically 4 to 6 cm from its origin. The abrupt cutoff has a reverse meniscus appearance because of contrast partially enveloping the embolus, which has lodged at the branch point of the SMA. Arterial thrombosis occurs on a background of preexisting severe atherosclerotic occlusive disease of the celiac artery and SMA. Symptoms of postprandial abdominal pain, weight loss, and altered bowel habits are typical. Intra-arterial infusion with thrombolytics such as tissue plasminogen activator allows for the dissolution of the clot and subsequent intervention on vascular lesions such as an ostial stenosis. Surgical intervention may be needed if bowel infarction is present. Nonocclusive ischemia accounts for 10% of acute ischemia and is caused by conditions that produce low flow states, such as hypotension, dehydration, and low cardiac output. The bowel responds with disproportionate vasoconstriction leading to ischemia. Angiography confirms diffuse vasoconstriction without underlying structural abnormality. The classic appearance of alternating areas of vasospasm has been termed sausage link narrowings. Vasodilators such as papaverine can be used to improve bowel perfusion and maximize recovery. Mesenteric venous occlusion generally affects the medium-sized veins of the middle small bowel and accounts for about 10% of cases. This can be treated with catheter-directed P.713 thrombolysis from the arterial access or via portal access from a percutaneous or transjugular route. Both mesenteric venous
1780
thrombosis and nonocclusive ischemia can present with GI bleeding.
FIGURE 25.15. Mesenteric Ischemia. A. Superior mesenteric arteriogram demonstrates a high-grade proximal stenosis (arrows). B . A balloon-mounted stent is used to treat the stenosis. C . Completion angiogram demonstrates the stent (arrowheads) and complete resolution of the stenosis.
Chronic
mesenteric
ischemia occurs with occlusion or high-grade
1781
stenosis of at least two of the three mesenteric arteries. Etiologies include atherosclerosis, fibromuscular dysplasia, and various vasculitides. Patients typically present with postprandial pain. The treatment is to relieve the flow-limiting lesion or occlusion with angioplasty and stenting or surgical bypass (Fig. 25.15). The vasculitides that involve the mesentery can be divided into their vascular distribution but are not very predictable for diagnosis. Treatment is specific to the etiology of the vasculitis. Drug-induced vasospasm or vasculitis must also be excluded. Stenosis of the celiac artery may be caused by extrinsic compression by the median arcuate ligament of the diaphragm. This condition has been implicated in chronic mesenteric ischemia, although this is very controversial. Angiography in the lateral projection shows a superior impression upon the proximal celiac axis that is more pronounced with
expiration.
DIAGNOSIS AND INTERVENTION VENOUS SYSTEM
OF
THE
The superior vena cava (SVC) is formed by the junction of the short, vertically oriented right and the longer obliquely oriented left brachiocephalic veins at the level of T1. The third tributary to the SVC is the azygos vein, which enters the dorsal aspect of the SVC at its midpoint. The SVC contains no valves and is usually less than 2 cm in diameter. The inferior vena cava (IVC) is formed by the junction of the right and left common iliac veins at L5. It ascends on the right of the abdominal aorta and anterior to the spine to enter the RA at about T8. A rudimentary valve (eustachian valve) is present just prior to its entrance into the RA. The main tributaries of the IVC are the hepatic (T10), renal (L2), right adrenal, right gonadal, and lumbar veins. The azygos venous system is an asymmetrically paired paravertebral venous complex, which provides an important collateral communication between the SVC and IVC. This system is divided into the azygos and hemiazygos veins, which lie to the right and the left of the spine,
1782
P.714 respectively. Both are continuations of the ascending lumbar and subcostal veins and begin at the L1 level. The azygos follows the aorta through the diaphragm to the T6 level, where it arches anteriorly over the right mainstem bronchus to join the SVC. The hemiazygos ascends into the chest and traverses the midline to join the azygos vein at approximately T8. The azygos system serves as the most important collateral pathway when the IVC is occluded.
FIGURE 25.16. Left Superior Vena Cava (SVC). A chest radiograph demonstrates abnormal position of a central venous catheter (arrowheads). Blood gas analysis and contrast injection confirmed catheter position within a left-sided SVC.
1783
FIGURE
25.17. Duplication Inferior Vena Cava. A. Inferior
vena cavagram demonstrates the right component of the inferior vena cava (IVC) with a large inflow from the left renal vein (arrowheads). The left iliac vein did not fill on this injection. B . Contrast injection into the left femoral vein opacifies the left IVC (LIVC), which joins the right IVC by draining through the left renal vein (LRV, arrowheads). RRV, right renal vein.
Central venous variants are relatively common, particularly for the IVC, and are quite important during venous intervention. A left SVC occurs in 0.3% of the population and descends through the left mediastinum anteriorly to join the coronary sinus, which drains into the RA (Fig. 25.16). A double SVC (left SVC with a normal right SVC) is the most common variation (85%). A single left SVC is rare and is associated with congenital heart disease. Azygos continuation of the IVC is caused by the absence of the intrahepatic portion of the IVC, with failure of the right subcardinal vein to anastomose with the hepatic veins. The hepatic veins drain into the RA. The renal and iliac veins drain via the azygos and hemiazygos veins into the SVC. Findings include dilatation of the azygos vein, the azygos arch, and
1784
the SVC. Duplicated IVC is present in 3% of the population and is a persistence of both right and left supracardinal veins (Fig. 25.17) . The left IVC is a continuation of the left iliac vein and ascends to the left of the aorta before crossing over to join the right IVC, usually via the left renal vein. Left IVC without a right IVC occurs in 0.2% of the population and crosses the midline at the level of the renal vein. The retroaortic left renal vein (2%) crosses behind the aorta instead of its usual path anterior to the aorta. The presence of both retroaortic and preaortic renal veins forms the circumaortic left renal vein (8%), which encircles the aorta to join the IVC. P.715 The deep venous drainage of the upper extremity consists of the brachial vein, which travels with its like-named artery and becomes duplicated peripheral to the elbow. The superficial venous system consists of the basilic vein, which drains into the proximal brachial vein, and the cephalic vein, which drains into the axillary vein but may drain as centrally as the subclavian or jugular veins. The axillary vein becomes the subclavian vein at the margin of the first rib. The predominant venous system of the lower extremity is the deep system. In the calf, the deep trunks follow the named arteries. The confluence of calf veins forms the popliteal vein. The femoral vein is a continuation of the popliteal vein at the adductor hiatus. The femoral vein joins the deep femoral vein (profunda femoris) below the inguinal ligament to form the common femoral vein, which continues as the external iliac vein. The confluence of the external and internal iliac veins at the pelvic brim forms the common iliac vein. The superficial system consists of the greater and lesser saphenous veins. The greater saphenous vein courses along the medial aspect of the lower limb and enters the femoral vein below the inguinal ligament. The lesser saphenous vein commences posterior to the lateral malleolus and enters the popliteal vein above the knee joint. Both the deep and superficial venous systems contain valves. Diagnostic evaluation of the extremity veins typically begins with US (see Chapter 40). MR, CT, and venography are useful for evaluation
1785
of both the central and peripheral venous system. Conventional venography has fallen out of favor, given the ability to evaluate for venous thrombosis on a noninvasive basis with US, CT, and MR.
Venous
Access
The need for central venous access continues to increase in both hospitalized and ambulatory patients. Decisions need to be made as to which access device is most appropriate for the clinical situation. Central venous access catheters can be categorized as temporary or “permanent,― tunneled or non-tunneled. The indications for placement of these catheters include access for antibiotic therapy, chemotherapy, parenteral nutrition, pain management, and hemodialysis. Non-tunnelled, temporary catheters include triple-lumen catheters and peripherally inserted central catheters (PICCs). PICCs are inserted through a peripheral upper extremity vein. Access into the brachial, basilic, or cephalic veins allows for central placement of the catheter tip at the cavoatrial junction to reduce the risk of subclavian venous thrombosis. They can be used for access for up to 4 weeks. Tunneled
access is used to provide access for intervals of greater
than 4 weeks; with proper care, these can last for longer than a year. Access via the external jugular vein is favored because its size and position lends itself for easy access and avoids some of the complications associated with subclavian access, such as pinch-off syndrome (compression between the clavicle and first rib) and subclavian stenosis. Potential complications include venous thrombosis, catheter obstruction by impingement against the vein wall, stenosis of cava, or occlusion of the catheter. A single-lumen Hickman catheter is adequate for antibiotic therapy, while hyperalimentation may require multiple-lumen access. Flow rates are the critical issue in the setting of dialysis access, so larger sized catheters are used, ranging in size from 12 French in the pediatric patient to 14 or 16 French in the adult. The catheters come premeasured and are chosen for length that will allow the catheter to bridge the cavoatrial junction into the RA. The distal catheter lumens
1786
are offset to prevent blood mixing. This is essential to prevent recirculation of the dialysate, which will prolong pump time. Subcutaneous
ports are the most cost-effective devices for venous
access for more than 6 months. These are ideal for cancer patients and for patients with sickle cell disease who require pain management. The catheter is completely under the skin, which prolongs the life of the catheter and prevents infection. Heparin solution is instilled into catheters and ports to prevent thrombosis. The access site should be kept clean, and antibiotic ointments are used at the access site to prevent catheter seeding from the skin flora. Additional complications include air embolism, access site bleeding, vessel injury, and pneumothorax. Catheter retrieval may be required for removal of a fragment of catheter that was lost because of pinch-off syndrome, during catheter exchange, or because of placement misadventure. Retrieval is performed with a gooseneck snare, which has a snare loop at right angles to its cable. The snare comes in a variety of sizes, which are chosen to match the diameter of the vessel containing the catheter fragment (Fig. 25.18). Other foreign bodies, such as coils lost during embolotherapy, can be retrieved in a similar manner.
Inferior
Vena
Cava
Filters
Pulmonary embolism is a major cause of morbidity and mortality, with up to 90% of pulmonary emboli originating from venous thrombosis in the lower extremity or pelvis. IVC filters are placed in patients with deep venous thrombosis (DVT) to prevent fatal pulmonary emboli. Indications for placement of IVC filters include: contraindication to anticoagulation, decreased cardiopulmonary reserve, patient noncompliance, and free-floating thrombus within the IVC. Filters may be placed prophylactically in patients with spinal injury or multiple traumatic injuries and in those undergoing pulmonary embolectomy and venous thrombolysis. The ideal filter should be efficient at trapping emboli, allow maintenance of the access site and caval patency, be easy to insert, have an indefinite life span, be potentially removable, and be MR compatible. Currently
1787
there are nine IVC filters available in the United States. These differ in P.716 deployment, with a trend toward smaller delivery systems and the option of removal. At the time of this writing, there are three filters that have the option of removal (Günther Tulip, Cook Inc; OptEase, Cordis Corp; Recovery, C.R. Bard Inc) (see Fig. 25.21) .
FIGURE
25.18. Catheter
Retrieval.
A. Fluoroscopic image
demonstrates a peripherally inserted central catheter (PICC) line (arrows), which has migrated centrally through the heart and into the pulmonary artery. A gooseneck snare (arrowhead) has been placed in the right pulmonary artery, accessed through the IVC. The snare looped is then tightened around the wayward catheter to accomplish retrieval. B . The PICC line has been
1788
captured by the snare and pulled back into the IVC.
When placing IVC filters, access via the right femoral or right jugular vein is preferred, although the left femoral or left jugular vein may be used if other sites are occluded or otherwise unsuitable. An IVCgram is performed to determine presence of caval thrombus, to assess the anatomy and diameter of the IVC, and to check for venous anomalies (Fig. 25.19). Reflux of contrast into, or flow of unopacified blood from, the iliac and renal veins is used to locate these vessels. Either a bony landmark or a radiopaque ruler is used as a reference marker for the renal vein level. This aids in accurate positioning of the filter just inferior to the renal veins. The filter is then deployed under fluoroscopic guidance. In patients with duplicated IVC, placement of a filter in each cava or a single suprarenal filter may be performed. Thrombus within the infrarenal cava or clot extension from a renal vein may necessitate placement in a suprarenal location. In the gravid female, placement in a suprarenal location has been recommended to avoid compression of the filter by the uterus. In addition, placement in this location should prevent embolization occurring through an enlarged ovarian vein from pelvic vein thrombosis. Recurrence of pulmonary embolism following filter placement is in the 2.7% to 4% range. Filters vary the most in incidence of caval thrombosis, with rates in the 3% to 9% range (Fig. 25.20) . Additional
complications
include
filter
migration,
caval
perforation,
tilting of the filter, and filter fracture. Concerns about the long-term safety of IVC filters have resulted in a trend toward retrievable filters, especially in patients with a long life expectancy or a short-term risk of thromboembolism. Indications for filter removal include migration of the filter (occasionally into the right heart) and filter infection. Retrievable filters should be removed within 10 to 14 days of placement (Fig. 25.21). Endothelialization lead to difficulty, with explantation as soon as 12 days after placement.
1789
can
FIGURE
25.19. Inferior Vena Cava (IVC) Filter Placement.
A . Venogram demonstrates clot (arrows) extending into the IVC from the left iliac vein. B . A filter (arrowhead) is demonstrated above the clot (arrow) within the IVC.
1790
FIGURE
25.20. Thrombus of Inferior Vena Cava (IVC)
Filter. CT of the abdomen demonstrating thrombus (arrowheads) surrounding the top (arrow) of an IVC filter. The patient had abdominal and lower extremity swelling 1 month after filter placement.
P.717
Venous
Thrombolysis
Anticoagulation therapy is the standard of care for the prevention of pulmonary embolism and recurrent DVT, but it does not protect the patient from the long-term effects of DVT. Chronic venous outflow obstruction and injury to valves produce P.718 postthrombotic syndrome in 40% to 80% of patients with DVT. Postthrombotic syndrome refers to the chronic pain, swelling, and development of cutaneous ulcers that may follow DVT. The goal of venous thrombolysis is to remove the obstructing thrombus and to preserve venous valve function. Thrombolysis is most effective in improving vein patency and relieving symptoms when the clot extends centrally.
1791
FIGURE
25.21. Inferior Vena Cava Filter Retrieval. A.
Cavagram demonstrates a Günther Tulip filter (arrow) without a clot present. B . A gooseneck snare is used to engage the hook at the top of the filter. C . The filter is pulled back into and constrained by the sheath (arrows) and then is withdrawn through the jugular vein.
Patients with upper or lower extremity symptoms and documented clot may be considered for thrombolysis if they are without a contraindication for lytic therapy. Contraindications include: internal bleeding, stroke within the past 6 months, cranial or spinal surgery within the past 2 months, intracranial neoplasm, bleeding diathesis, uncontrolled hypertension, and contraindication to anticoagulation. US documents DVT in the extremity. Central extent of the clot can be evaluated with MR and CT (Fig. 25.22) . Patients with upper extremity involvement are treated with removal of the offending catheter and anticoagulation. If the patient is more symptomatic and has a documented central clot, catheter-directed
1792
thrombolysis with venous intervention may be undertaken (Fig. 25.23). For the upper extremity, the preferred access is via the brachial or basilic vein, with catheter placement into the area of thrombosis. Once lysis is complete, the offending lesion may be treated endovascularly with venoplasty or stenting. Compression syndromes may require additional surgical correction.
FIGURE 25.22. Deep Venous Thrombosis: Iliac Vein. A. Postcontrast CT shows low-attenuation clot within the right iliac vein (arrowhead) in comparison to the higher attenuation within the normal right iliac artery (arrow). B . Gradient MR of the pelvis shows absent flow signal (white arrowhead) caused by thrombus within the right iliac vein compared with the normal right iliac artery (white arrow) and contralateral left iliac vein (gray arrowhead). C . T2WI through the region demonstrates extensive
1793
perivascular edema (white arrows) and intermediate signal clot (black arrowhead) within the expanded right iliac vein.
In the lower extremity, the popliteal vein is the preferred access to iliofemoral DVT. The popliteal vein is large enough to accommodate sheath sizes for most iliac and caval interventions (Fig. 25.24). The administration of thrombolytic agents into thrombi in the venous system is similar to the process performed in the arterial system, in that a multiple-sidehole infusion catheter is imbedded throughout the thrombus. The thrombolytic agent of choice is infused into the clot. The patient usually returns to the angiographic suite for imaging at 12 to 24 hours to assess progression of thrombolysis. Once dissolution of the thrombus has resulted in the restoration of antegrade flow, any underlying lesion can be treated with venoplasty and/or stenting. Stenting is avoided below the inguinal ligament (superficial femoral vein) because of frequent stent failure and thrombosis. For iliofemoral DVT in which symptoms have been present for less than P.719 4 weeks, 80% to 85% of patients have completely or greatly improved partial thrombolysis. In patients in whom there is no malignancy, the 2-year iliac vein patency is as high as 75%. In wellselected patients, thrombolysis is effective and has been shown to improve quality of life with a good safety profile.
1794
FIGURE
25.23. Deep Venous Thrombosis: Subclavian Vein.
A . Left arm venogram in a patient with pacemaker shows lack of contrast opacification of the subclavian vein (white arrowhead) indicative of thrombosis along the pacemaker lead. The brachiocephalic vein (black arrows) fills via collaterals. B . Imaging following administration of thrombolytics and balloon angioplasty shows residual clots (arrows), which were treated with repeat angioplasty and stent placement.
1795
FIGURE
25.24. Deep Venous Thrombosis: Femoral Vein. A.
Prone left leg venogram demonstrates clot (arrows) within the femoral vein. B . Venogram following 24-hour infusion of thrombolytics demonstrates complete resolution of the clot.
Complications of venous thrombolysis include: bleeding at the access site, hemorrhagic stroke, GI bleeding, retroperitoneal hematoma, and pulmonary embolus. Bleeding is increasingly common with higher doses of tissue plasminogen activator and anticoagulation during thrombolysis. The rate of pulmonary embolism is low. A filter may be placed if there is free-floating clot within the IVC or if clot extends above an indwelling thrombosed filter. Phlegmasia cerulea dolens is arterial compromise of a limb caused by elevated venous pressures from massive acute venous thrombosis. In the majority of patients, DVT spares the collateral pathways. In patients with phlegmasia cerulea dolens, thrombosis involves both
1796
main and collateral venous drainage, causing swelling and severe elevations in vascular resistance and resulting in ischemia. These patients require acute treatment. Paget-von Schrötter syndrome is compression of the subclavian vein by a cervical rib, soft tissue anomaly, or scar tissue after clavicle fracture resulting in thrombosis and arm swelling. Treatment of the offending subclavian venous stenosis becomes multidisciplinary. Stenting of these lesions should be avoided prior to surgery. Stenting often fails because of persistent extrinsic compression with stent fracture. Angioplasty may be performed to improve flow. SVC syndrome is caused by extrinsic compression of the SVC with thrombosis. Causes include bronchogenic carcinoma (most common: up to 82%); granulomas (histoplasmosis and tuberculosis); lymphoma; intravascular foreign bodies (pacemaker leads, central venous catheters); and venous stenoses caused by chronic dialysis and venous hypertension (Fig. 25.25). In the setting of carcinoma, treatment with external beam radiation may be used initially. If symptoms persist or thrombosis has occurred, endovascular intervention can be undertaken with venoplasty and stenting.
1797
FIGURE 25.25. Superior Vena Cava (SVC) Obstruction. A. Dilated and tortuous collateral veins (arrows) carry venous return from the upper extremities. B . Coronal MR shows bright signal from thrombus (arrow) within the SVC. C . Contrast injection into the right subclavian vein reveals nearly complete obstruction of the SVC caused by tumor invasion (arrows) by bronchogenic carcinoma.
May-Thurner syndrome involves the compression of the left iliac vein by the right iliac artery crossing over it. This is normal anatomy; however, in some patients arterial pressure on the vein results in thickening of the vein wall and narrowing of the lumen, with resulting thrombosis. In some patients, exercise alone may cause symptoms
1798
without thrombosis being present. Angioplasty and stenting may offer improvement in symptoms and long-term venous patency (Fig. 25.26) .
Venous
Access
for
Dialysis
More than 200,000 people in this country require hemodialysis. Access for hemodialysis is generally achieved through the use of arteriovenous P.720 fistula (AVF) (Brescia-Cimino), bridge graft fistula, or a central venous catheter or dialysis port.
1799
FIGURE 25.26. May Thurner Syndrome. A. Pelvic venogram demonstrates occlusion of the left common iliac vein. B . Following thrombolysis and angioplasty, a self-expanding stent (arrows) has been placed in the left iliac vein. C . Completion venogram demonstrating resolution of the occlusion and a patent
1800
left iliac vein.
AVFs provide convenient access for hemodialysis and can last for years. In patients in whom an AVF cannot be fashioned because of inadequate native veins, a bridge graft (loop or straight) can be created between an artery and a vein using either a PTFE (polytetrafluoroethylene) material or a bovine vein. Use of the AVF is complicated by hematomas, pseudoaneurysms, stenoses, and thrombosis. Monthly monitoring of graft function is recommended. Intervention prior to graft thrombosis is preferred. High venous pressure is evidence of venous anastomotic stenosis. Both primary fistulas and bridge grafts can be managed with percutaneous techniques in interventional radiology (Fig. 25.27) . Grafts that are not amenable to percutaneous management are those that are infected, have been revised or placed within the past week, or are in patients in whom repeated graft failure has occurred within a short time of percutaneous management. Techniques for thrombectomizing dialysis grafts include mechanical thrombectomy devices, balloon thrombectomy, and thrombolytic agents. Access is gained with the puncture directed toward the P.721 venous anastomosis. The arterial end of the graft should be compressed to prevent displacement of thrombus into the parent arterial supply. The venous portion of the graft may be cleared with a balloon or mechanical device. The entire then be pulse-sprayed with a thrombolytic advanced through the venous anastomosis, outflow and central veins is obtained. Any
length of the graft may agent. A catheter is and venography of the central or outflow vein
stenoses and the venous anastomosis are treated with venoplasty. The graft is then accessed toward the arterial anastomosis, and a balloon thrombectomy of the arterial plug may be performed. Once function is restored, hemostasis may be achieved with manual compression or with the use of a purse-string suture. Inability to cross the venous anastomosis is a cause of technical failure. Minor complications occur in fewer than 10% of patients. Major
1801
complications occur in 1% of patients. Most of these are distal arterial emboli, which can usually be managed with catheter-directed therapy. Budd-Chiari syndrome refers to occlusion of the hepatic veins, which occurs as result of hepatic venous or IVC outflow obstruction. Most patients have an underlying condition that predisposes to blood clotting. In up to 30% of patients, no predisposing factors are identified. Post-sinusoidal portal hypertension occurs. Patients may present with acute hepatic failure, portal hypertension, or chronic hepatic dysfunction. Angiography of the celiac and superior mesenteric artery is nonspecific. Venous studies classically show a “steeple― or “pencil point― configuration of the intrahepatic IVC (from compression by swollen liver and enlarged caudate lobe). Thrombus may be present. Hepatic vein studies show a “spiderweb― pattern of collateral veins and lymphatics (Fig. 25.28). Webs may be present but can be missed unless the catheter is directly adjacent to the web. There is usually no normal hepatic sinusoidal filling. Venoplasty of caval webs or of a segmental IVC occlusion may be used to manage these patients. The 1-year patency rate is 80% to 100%. Frequent repeat venography is needed to follow up these patients. More aggressive maneuvers such as IVC or hepatic vein stenting or placement of a transjugular intrahepatic portosystemic shunt (TIPS) may be required. P.722
Transjugular Shunt
Intrahepatic
Portosystemic
The pressure in the portal veins normally varies between 5 and 10 mm Hg. The portosystemic pressure gradient (i.e., the pressure gradient between the RA and the portal vein) normally ranges from 3 to 6 mm Hg. The definition of portal hypertension is an absolute portal pressure greater than 11 mm Hg or a portosystemic gradient above 6 mm Hg. Formation of varices (portosystemic collaterals) and subsequent bleeding occur when the portosystemic gradient is greater than 11 to 12 mm Hg. Variceal hemorrhage, ascites,
1802
spontaneous bacterial peritonitis, coagulopathy, hepatic encephalopathy, hepatorenal syndrome, and hepatopulmonary syndrome may all occur in patients with portal hypertension. Common collaterals are the coronary veins, which anastomose with the azygos system in the submucosa of the distal esophagus and gastric cardia. These abnormal vascular structures thin the overlying mucosa, project into the esophageal lumen, and are prone to erode and bleed. These are termed “uphill― varices, as opposed to the “downhill― varices seen with SVC obstruction.
FIGURE
25.27. Hemodialysis
Fistula.
A.
Fistulogram
demonstrates venous anastomotic stricture (arrow) and a pseudoaneurysm (arrowhead) in a polytetrafluoroethylene graft. B . Balloon dilatation of a wall graft that was placed to exclude the pseudoaneurysm. C . Completion fistulogram demonstrates exclusion of the pseudoaneurysm (white arrowhead) and stenting of the venous anastomotic stricture (black arrow), which was resistant
to
balloon
angioplasty.
1803
FIGURE
25.28. Budd-Chiari
Syndrome.
A. Injection of the
hepatic vein shows the classic spiderweb appearance that is characteristic of Budd-Chiari Syndrome. B . Venogram of the intrahepatic inferior vena cava shows the mass effect (arrows) from the swollen liver and clot extending from the hepatic veins.
The modified Child-Pugh score is a classification scheme used to assess the overall severity and prognosis of liver disease. Child class A is the mildest and Child class C P.723 is the most severe. In symptomatic patients, medical management is the first line of therapy. Endoscopic management may be needed for the control of variceal bleeding from the esophagus. When medical and endoscopic management of the patient are no longer effective or fail, then decompression of the portal system should be undertaken.
1804
FIGURE
25.29. Transjugular
Intrahepatic
Portosystemic
Shunt: Step by Step. A. The needle is shown passing from the proximal right hepatic vein into the right portal vein 2 to 3 cm from the portal vein bifurcation. Note the corkscrew appearance of the dilated coronary vein. B . After balloon angioplasty of the tract, a flexible wall stent is being deployed to bridge the parenchymal tract. C . The stent is completely deployed, creating the shunt. The coronary vein is embolized with coils in patients with ongoing or recent active bleeding. (Adapted from Zemel G, Becker GJ, Bancroft JW, et al. Technical advances in transjugular intrahepatic portosystemic shunts. Radiographics 1992;12:615–622. Used with permission.)
1805
TIPS is an endovascular procedure that creates a portosystemic shunt. A TIPS most closely resembles a side-to-side portocaval shunt physiologically. The shunt is typically formed between the right hepatic vein and the right portal vein (Fig. 25.29). Technical success rates for creating a TIPS are greater than 95%. Doppler US or a CT study of the liver should be obtained to confirm patency of the portal vein. Access is gained into the right internal jugular vein and a TIPS set is utilized. A number of TIPS sets are available for use by the interventionist. The right hepatic vein is accessed, and curved needle passes are made in an anterior direction to access the right portal vein. If the middle hepatic vein is used, passes are made in a posterior direction. Once the portal vein is encountered, the pressure gradient between the portal and systemic (right atrial) venous systems is measured, and portal venography is performed to determine the access point into the portal vein, length of tract, and the area to be covered by the stent. Care must be taken not to enter the extrahepatic portions of the portal vein. Following balloon dilation of the stent tract, an appropriately sized stent is deployed (Fig. 25.30). Typically a 10-mm-diameter stent will be used. The current trend is to use covered stents, as these improve patency by excluding the parenchymal tract, which promotes thrombosis and instent stenosis. However, covered stents are associated with trapping of portal vein branches or the IVC, so careful measurement and planning of stent placement are needed. The portosystemic pressure gradient and venography are repeated. The goal is to decrease the portosystemic gradient to less than 12 mm Hg and to see no significant filling of varices. When TIPS have been created to treat ascites, portosystemic gradients lower than 15 mm Hg are adequate. If varices opacify following TIPS placement, embolization of the varices with coils is performed (Fig. 25.31). Proximal embolization prevents filling of the varices by its usual retrograde fashion. The coils are selected based upon the size of the vessel and should be slightly oversized to reduce the risk of distal embolization. TIPS must be followed closely, as there is a high incidence of shunt stenosis and occlusion. Doppler US of the TIPS is required
1806
periodically for surveillance (see Chapter 40). The primary patency of TIPS at 1 year is 20% to 66%. The primary assisted patency at 1 year is 72% to 83%. Shunt dysfunction may occur early or late. Bile duct transection has been described as a potential etiology for acute shunt thrombosis. The incidence of acute thrombosis is diminished with the use of covered stents. US scans are obtained at 1, 3, and 6 months and then every 6 months thereafter. Absolute contraindications to TIPS include severe hepatic failure and severe right heart failure. Relative contraindications include polycystic liver disease, hepatic neoplasm, hepatic encephalopathy, and portal vein thrombosis. Procedure-related complications include intraperitoneal hemorrhage (1% to 6%), hemobilia (1% to 4%), sepsis, and transient renal failure. New or worsening encephalopathy may occur in 25% of patients. In all but 4% to 7% of patients, the encephalopathy can be controlled with medical therapy. Recurrent bleeding is seen in 15% to 30% of patients. TIPS stenosis or thrombosis may be treated with balloon angioplasty with or without placement of another stent. The population of patients receiving TIPS is often very ill. The 30-day survival of patients in whom TIPS has been P.724 P.725 created is 85% to 97%. A patient's status prior to the TIPS is directly related to survival post-TIPS. TIPS is best suited for Child class B and C patients, particularly those who are candidates for a liver transplant. Severe compromise of liver function and elevation of the Child score decrease survival and increase the risk of complication.
1807
FIGURE
25.30. Transjugular
Intrahepatic
Portosystemic
Shunt (TIPS) Procedure. A. Portal venogram with injection of the sheath demonstrates the hepatic vein (white arrowhead) , right portal vein access (white arrow), and the parenchymal tract (black
arrowheads). B . After TIPS creation, the shunt
(arrowheads) bridges the parenchymal tract between the hepatic and portal veins.
1808
FIGURE
25.31. Persistent Filling of Varices. A. Portal
venogram demonstrates filling of a varix (arrows) with a splenorenal shunt (arrowheads). B . Portal venogram shows proximal occlusion of the varix using coils (arrowhead) .
NONVASCULAR
INTERVENTION
Percutaneous nephrostomy is indicated for ureteral obstruction or treatment of stones within the renal collecting system. It can also be used to access ureteral strictures for dilatation, to perform a brush biopsy of suspected tumor, or to place a ureteral stent. Urinary diversion with a nephrostomy tube is used in the setting of urine leakage including vesical-vaginal or vesical-cutaneous fistulae. Diversion of urine flow allows for closure of the fistulous tract. The procedure may be guided by fluoroscopy using IV contrast, CT, or US. The patient is placed in a prone position. A 18- or 22-gauge needle is passed into a posterior calyx, usually at the middle or lower pole (Fig. 25.32). Ideally, the nephrostomy catheter is positioned below the 12th rib to avoid pneumothorax. However, it is often necessary to obtain access between the 11th and 12th ribs. For
1809
intercostals placement, the puncture should be made close to the top of the 12th rib to avoid the parietal pleura and the intercostal vessels. Punctures close to the neurovascular bundle may lead to increased
pain.
The needle is passed through the least vascular area of the kidney, Brödel plane, by positioning the needle 30° to 45° from the vertical (Fig. 25.32). The needle is passed down the barrel of the posterior calyx, avoiding the large vascular structures in the renal pelvis. When urine is aspirated from the needle, a guidewire is advanced through the needle into the renal collecting system. The tract is dilated and the nephrostomy catheter is placed. An antegrade nephrostogram is performed to confirm catheter placement, detect urine leakage, and determine the etiology of the hydronephrosis. Patients are instructed in proper tube care. Nephrostomy catheters are typically exchanged every 3 months to prevent encrustation and occlusion. If pyonephrosis is encountered, a urine sample is sent to microbiology, and overdistension of the collecting system is avoided to prevent bacteremia. The patient is observed carefully for septic shock. Complications of nephrostomy include hemorrhage; sepsis; pneumothorax (with an intercostal approach); urine leakage (urinoma); visceral injury (colon, liver, spleen); and catheter malfunction. Ureteral
stenting may be used for identification of the ureter during
surgery or to bypass a ureteral obstruction caused by tumor, inflammation, or benign stricture. A ureteral stent catheter, usually composed of polyurethane, can be placed at the time of the initial percutaneous nephrostomy if there is no significant bleeding or edema, the ureter shows very little tortuosity, and the ureteral obstruction can be easily crossed (Fig. 25.33). If not, then a few days of drainage are recommended before stenting is undertaken. A catheter of appropriate length for the ureter is placed with the distal pigtail in the bladder and proximal pigtail in the renal pelvis. A double-J ureteral stent is a completely internal catheter that is used only if the patient can have this exchanged cystoscopically every few
1810
months. Occasionally, a ureteral stricture will need balloon dilatation prior to stent placement. Strictures result from iatrogenic trauma, tumors of the ureter, transitional cell cancer of the bladder affecting the trigone region, and cervical cancer. Balloon dilatation with a highpressure 5- or 6-mm balloon will suffice in most cases and allow passage of a stent. Dilatation with a cutting balloon may be necessary for difficult lesions. Complications of ureteral stent placement include perforation of the ureter, bladder irritation with urinary frequency, catheter dislodgement, catheter encrustation resulting in occlusion, and incorrect positioning, in addition to complications of nephrostomy. Several situations require special attention when placing nephrostomy catheters. When a nephrostomy catheter is being placed for percutaneous stone extraction (percutaneous nephrolithotomy), a calyx should be chosen that provides the best access to the stone. Two access sites may be needed to retrieve fragments from large staghorn calculi. Sometimes, collecting distension is needed to make room for a wire to pass into the already crowded collecting system. This can by achieved by instillation of contrast via a retrograde ureteral stent or antegrade nephrostomy. Wire access into the bladder with a second “safety wire― is recommended for the eventual dilation of the tract up to 30 French. Special attention also needs to be paid for percutaneous nephrostomy into a horseshoe kidney. Intervention is more complex because of altered vascular anatomy and the more anterior position of the kidney. Longer access systems are necessary, especially for stone removal cases. CT correlation is a helpful guide to anatomy and measurement. Renal transplant nephrostomy is usually performed with US guidance for ureteral pelvic or ureteral vesicle level obstructions. It is best in this situation to temporize with a percutaneous nephrostomy initially and then determine the best manner of treatment after discussion with transplant surgical team. Percutaneous biliary drainage is performed most often for obstruction and less often for leakage. Biliary obstruction may be the
1811
result of extraluminal compression of the biliary tree or an intraluminal mass. Benign causes include calculi, sclerosing cholangitis, previous surgery or invasive procedure, and ischemia. Malignant
obstruction P.726
includes cholangiocarcinoma, pancreatic carcinoma, ampullary carcinoma, metastases, and lymphadenopathy. US, CT, and MR may show causes of biliary obstruction. Indications for percutaneous biliary drainage include relief of obstructive jaundice associated with pruritus, cholangitis with sepsis, brachytherapy access for malignant lesions, failed endoscopic biliary drainage, or surgically altered anatomy
(Billroth
II).
1812
FIGURE 25.32. Percutaneous Nephrostomy. A. Cross section of kidney demonstrating needle path for percutaneous
1813
nephrostomy placement. Needle entry into a posterior calyx poses the least risk of hemorrhage, since it courses in the avascular plane of Brodel. (Redrawn from Kandarpa K, Aruny JE. Handbook of Interventional Radiologic Procedures. 2nd ed. Boston: Little, Brown, and Company, 1996.) B . Nephrostogram in a patient with obstructing stone (white arrow) within the proximal ureter. Note the middle-pole access of the nephrostomy with self-retaining pigtail (black arrows). The middle-pole access offers mechanical advantage if nephroureteral stenting is needed.
Patients are typically given broad-spectrum antibiotics prior to the procedure because of the high risk of bacteremia. More aggressive antibiotic therapy is needed if sepsis occurs after the procedure. Previous imaging studies are reviewed prior to the procedure for access planning. The presence of ascites is a relative contraindication. US or fluoroscopic guidance is used. From the right, the skin entry site may be up to 2 cm posterior to the midaxillary line, at about the 11th intercostal space level. Traversal of the pleural space should be avoided. From the left, access into the biliary tree may be obtained using a subxiphoid approach. Contrast injection confirms entry P.727 of the needle into the biliary tree (Fig. 25.34). When access into a suitable peripheral bile duct is obtained, a guidewire is inserted, and dilation of the tract is performed. Excessive manipulation within the biliary tree should be avoided. If the obstruction in the biliary tree is not traversed, a locking loop catheter may be placed above the obstruction. The biliary system should be drained for at least a day to allow for resolution of local edema and distension before making another attempt. When the obstruction is traversed, a biliary drainage catheter is inserted and the locking loop is formed within the duodenum. The catheter is placed to external gravity drainage. The catheter should be flushed with 10 mL of sterile saline two to three times daily. Routine catheter exchange is performed every three
months.
1814
FIGURE
25.33. Ureteral Stent Placement. A.
Nephrostogram
in prone position demonstrates a stricture (arrowheads) at the anastomosis between the ureter and the ileal loop diversion. B . Postprocedure radiograph shows the nephroureteral stent (arrows) bridging the stenosis.
FIGURE 25.34. Percutaneous Biliary Stent. A. A percutaneous cholangiogram demonstrates an anastomotic stricture (arrow) in a patient post–biliary diversion for a
1815
laparoscopic cholecystectomy injury. B . A biliary drain has been placed across the anastomosis (arrow) .
Complications of biliary drainage include sepsis, hemorrhage, bile leak, cholangitis, catheter dislodgement or malfunction, fluid and electrolyte imbalance, and pneumothorax. Sepsis is caused by overdistension of an infected biliary system. Pneumothorax is a complication of P.728 high access within the liver resulting in traversing of the diaphragm and lung. In the case of malignant disease, self-expanding metallic stents can be placed within the biliary system so that patients do not have to spend the remaining few months of their lives with an external biliary drainage catheter (Fig. 25.35). Balloon dilatation of the stricture with a high-pressure or cutting balloon is performed before stent placement. Appropriate sedation must be attained, as this is very painful to the patient. The stent is placed across the obstructing biliary stricture with adequate coverage. Extension into the duodenum may be required for ampullary lesions. Tumor ingrowth through the interstices of the stent or overgrowth at the margins of the stent may lead to recurrent biliary obstruction.
1816
FIGURE 25.35. Self-Expanding Biliary Stent. A. Percutaneous cholangiogram shows common bile duct obstruction (arrows) in a patient with pancreatic carcinoma. B . Postprocedure cholangiogram shows placement of a self expanding stent (arrows) across the obstruction.
Laparoscopic cholecystectomy injury to the biliary ducts deserves special attention. Biliary drainage may act as a temporizing or definitive therapy. External drainage is used to divert bile from the site of injury or from a leak from a cystic duct remnant or accessory cystic duct. When the common bile duct has been transected, access into the biliary tree may be difficult, because the biliary tree decompresses into the peritoneal cavity. An abdominal drainage catheter should be inserted into an associated biloma. Percutaneous biliary drainage may be required when endoscopic intervention fails. If possible, the site of injury is traversed with a biliary drainage catheter. This catheter may have to stay in place for weeks. If the lacerated area does not heal, then surgical repair will be needed. In cases of complete transection of the biliary tree, surgical repair is required. Percutaneous cholecystostomy may be required in cases of calculus or acalculous cholecystitis in unstable patients. Usually, placement of a cholecystostomy tube is a temporary measure for managing a very sick patient until a cholecystectomy can be performed. The same steps apply to preparation of the patient, as described above for percutaneous biliary drainage. Antibiotic coverage is started before the procedure. US or fluoroscopic guidance is used. US guidance allows for bedside placement of the tube in very ill patients. A transhepatic approach to the gallbladder is preferred so that any leakage around the needle or catheter will be extraperitoneal. If the patient does not undergo surgery, the catheter should be left in place for approximately 6 weeks. This allows maturation of the tract and reduces the possibility of bile leakage and peritonitis. Before removal, a tube cholangiogram is performed to ensure that the cystic duct and common bile duct are
1817
patent. The catheter should not be removed if retained stones are demonstrated within the gallbladder.
SUGGESTED
READINGS
Boyer H, Haskal Z. American Association for the Study of Liver Disease Practice Guidelines: the role of transjugular intrahepatic portosystemic shunt creation in the management of portal hypertension. J Vasc Interv Radiol 2005;16:615–629. Burke D, Lewis C, Cardella J, et al. Quality improvement guidelines for percutaneous transhepatic cholangiography
and
biliary drainage. J Vasc Interv Radiol 2003;14:243S–246S. Comerota AJ. Quality of life improvement using thrombolytic therapy for iliofemoral deep venous thrombosis. Rev Cardiovasc Med 2002;3(suppl 2):S61–67. Darcy M. Treatment of lower GI bleeding: vasopressin infusion versus embolization. J Vasc Interv Radiol 2003;14:535–543. Ferrel H, Patel N. Selection criteria for patients undergoing transjugular intrahepatic portosystemic shunt procedures: status. J Vasc Interv Radiol 2005;16:449–455.
current
P.729 Goffette PP, Laterre PF. Traumatic injuries: imaging and intervention in post traumatic complications (delayed intervention).
Eur
Radiol
2002;12:994–1021.
Kinney TB. Update on inferior vena cava filters. J Vasc Interv Radiol 2003;14:425–440. Leertouwer TC, Gussenhoven EJ, van Jaarsveld BC, et al. Stent
1818
placement for renal artery stenosis: where do we stand? A metaanalysis. Radiology 2000;216:78–85. Mewissin MW, Seabrook GR, Meissner MH, et al. Catheter-directed thrombolysis for lower extremity deep venous thrombosis: report of a national multicenter registry. Radiology 1999;211:39–49. Ramchandani P, Cardella J, Grassi CJ, et al. Quality improvement guidelines for percutaneous nephrostomy. J Vasc Interv Radiol 2003;14:277S–281S. Shapiro M, McDonald AA, Knight D, Johannigman JA, Cuschieri J. The role of repeat angiography in the management of pelvic fractures.
J
Trauma
2005;58:227–231.
1819
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section VII - Gastrointestinal Tract > Chapter 26 - Abdomen and Pelvis
Chapter
26
Abdomen and Pelvis William E. Brant
IMAGING
METHODS
Plain film radiographs of the abdomen are important for the assessment of the acute abdomen and to serve as “scout― films prior to contrast studies. CT, US, and MR provide comprehensive evaluation of the abdomen, including the peritoneal cavity,
retroperitoneal
compartments,
abdominal
and
pelvic
organs,
blood vessels, and lymph nodes.
COMPARTMENTAL ANATOMY ABDOMEN AND PELVIS
OF
THE
Knowledge of the complex compartmental anatomy of the abdomen and pelvis is fundamental to understanding the effects of pathologic processes and to correctly interpret imaging studies. An understanding of the shape and extent of anatomic compartments and their normal variations may clarify imaging findings that would otherwise be incomprehensible or lead to misdiagnosis (1) . Fundamental considerations include constant anatomic landmarks, ligaments and fascia that define compartments, and normal variations in size and appearance of the various compartments and recesses. Identifying the precise compartment in which an abnormality is located determines to a great extent the origin of the
1820
abnormality. The peritoneal cavity is divided into the greater peritoneal cavity and the lesser peritoneal cavity (the lesser sac) (Fig. 26.1). Within both portions of the peritoneal cavity are numerous recesses in which pathologic processes tend to loculate. The right subphrenic space communicates around the liver with the anterior subhepatic and posterior subhepatic space (Morison pouch). The Morison pouch (the right hepatorenal fossa) is the most dependent portion of the abdominal cavity in a supine patient, and it collects ascites, hemoperitoneum, metastases, and abscesses. The right subphrenic and subhepatic spaces communicate freely with the pelvic peritoneal cavity via the right paracolic gutter. The left subphrenic space communicates freely with the left subhepatic space, but it is separated from the right subphrenic space by the falciform ligament and from the left paracolic gutter by the phrenicocolic ligament. The P.734 left subphrenic (perisplenic) space distends with fluid from ascites and with blood from splenic trauma. It is a common location for abscesses and for disease processes of the tail of the pancreas. The left subhepatic space (gastrohepatic recess) is affected by diseases of the duodenal bulb, lesser curve of the stomach, gallbladder, and left lobe of the liver.
1821
FIGURE
26.1. Anatomy of the Peritoneal Cavity. A. Diagram
of an axial cross section of the abdomen illustrates the recesses of the greater peritoneal cavity and the lesser sac. B . CT scan of a patient with a large amount of ascites nicely demonstrates the recesses of the greater peritoneal cavity and the lesser sac. The lesser sac is bounded by the stomach (St) anteriorly, the pancreas (P) posteriorly, and the gastrosplenic ligament (curved arrow) laterally. The falciform ligament (arrowhead) separates
1822
the right and left subphrenic spaces. Fluid from the greater peritoneal cavity extends into the Morison pouch (arrow) between the liver and the right kidney. Fluid in the gastrohepatic recess (asterisk) separates the stomach from the liver (L). S, spleen; GB, gallbladder; RK, right kidney; IVC, inferior vena cava; Ao, aorta; LK, left kidney.
Free fluid, blood, infection, and peritoneal metastases commonly settle in the pelvis because the pelvis is the most dependent portion of the peritoneal cavity (in the upright patient) and communicates with both sides of the abdomen. The falciform ligament consists of two closely applied layers of peritoneum that extend from the umbilicus to the diaphragm in a parasagittal plane. The caudal free end of the falciform ligament contains the ligamentum teres, which is the remnant of the obliterated umbilical vein. Paraumbilical veins (portosystemic collateral vessels) that form in the falciform ligament are a specific sign of portal hypertension. The reflections of the falciform ligament separate over the posterior dome of the liver to form the coronary ligaments, which define the “bare area― of the liver not covered by peritoneum. The coronary ligaments reflect between the liver and diaphragm and prevent ascites and other intraperitoneal processes from covering the bare area of the liver. The lesser omentum, composed of the gastrohepatic and hepatoduodenal ligaments, suspends the stomach and duodenal bulb from the inferior surface of the liver. The lesser omentum separates the gastrohepatic recess of the left subphrenic space from the lesser sac (Figs. 26.1, 26.2). The lesser omentum transmits the coronary veins (which dilate as varices) and contains lymph nodes (which enlarge with involvement by gastric carcinoma and lymphoma). The lesser sac is the isolated peritoneal compartment between the stomach and the pancreas. It communicates with the rest of the peritoneal cavity (the greater sac) only through the small foramen of Winslow. Pathologic processes in the lesser sac usually occur because of disease in adjacent organs (pancreas, stomach) rather than
1823
spread from elsewhere in the abdominal cavity. The lesser sac is normally collapsed but can become huge when filled with fluid. The greater
omentum is a double layer of peritoneum that hangs
from the greater curvature of the stomach and descends in front of the abdominal viscera, separating bowel from the anterior abdominal wall (Fig. 26.2). The greater omentum encloses fat and a few blood vessels. It serves as fertile ground for implantation of peritoneal metastases and assists in loculation of inflammatory processes of the peritoneal cavity (abscesses, tuberculosis). The retroperitoneal space between the diaphragm and the pelvic brim is divided into anterior pararenal, perirenal, and posterior pararenal compartments by the anterior and posterior renal fascia (Fig. 26.3) . The anterior pararenal space extends between the posterior parietal peritoneum and the anterior renal fascia. It is bounded laterally by the lateral conal fascia. The pancreas, duodenal loop, and ascending and descending portions of the colon are within the anterior pararenal space. Disease in the anterior pararenal space usually originates from these organs (pancreatitis, perforating/penetrating
ulcer,
diverticulitis).
1824
FIGURE
26.2. The Lesser Sac. Sagittal plane diagrams of the
1825
medial (A) and lateral (B) aspects of the lesser sac illustrate its position posterior to the stomach and anterior to the posterior parietal peritoneum covering the pancreas. Note that projections of the lesser sac extend to the diaphragm, resulting in the potential for disease processes in the lesser sac to cause pleural effusions. The coronary ligaments reflect between the liver and the diaphragm producing a bare area of liver not covered by peritoneum.
FIGURE
26.3. Retroperitoneal
Compartmental
1826
Anatomy.
Diagrams illustrate two normal variations of the reflections of the posterior parietal peritoneum around the descending colon. In (A) the colon is entirely retroperitoneal, and in (B) the peritoneum forms a deep pocket lateral to the colon, allowing intraperitoneal fluid to extend far posteriorly. Fluid or disease processes in the anterior pararenal space from the pancreas or colon may also extend posteriorly to the kidney by separating the two layers of the posterior renal fascia.
P.735 The anterior and posterior renal fascia encompass the kidney, adrenal gland, and perirenal fat within the perirenal space. The anterior renal fascia is thin and consists of one layer of connective tissue. The posterior renal fascia is thicker, consisting of two layers of connective tissue (Fig. 26.3). The anterior layer of the posterior renal fascia is continuous with the anterior renal fascia. The posterior layer of the renal fascia is continuous with the lateroconal fascia, forming the lateral boundary of the anterior pararenal space. The anterior and posterior layers of the posterior renal fascia may be separated by inflammatory processes, such as pancreatitis, extending from the anterior pararenal space. The renal fascia is bound to the fascia surrounding the aorta and vena cava; this usually prevents spread of disease to the contralateral perirenal space. However, disease processes arising from the perivascular space may extend into the perirenal space (hemorrhage from aortic aneurysm rupture, lymphoma). Fluid collections in the perirenal space are usually renal in origin (infection, urinoma, hemorrhage). Bridging septa extending between the renal fascia and the renal capsule tend to cause loculations of fluid processes in the perirenal space. The right perirenal P.736 space is open superiorly to the bare area of the liver, allowing spread of disease processes (infection, tumor) between the kidney and liver.
1827
FIGURE 26.4. Compartmental Anatomy of the Pelvis. Diagram in the coronal plane illustrates the major anatomic compartments of the pelvis.
The posterior
pararenal
space is a potential space, usually filled
only with fat, extending from the posterior renal fascia to the transversalis fascia. The posterior pararenal fat continues into the flank as the properitoneal fat “stripe― seen on plain films of the abdomen. The compartment is limited medially by the lateral edges of the psoas and quadratus lumborum muscles. Isolated fluid collections are rare and most commonly caused by spontaneous hemorrhage into the psoas muscle as a result of anticoagulation therapy. The pelvis is divided into three major anatomic compartments (Fig. 26.4). The peritoneal cavity extends to the level of the vagina, forming the pouch of Douglas (cul-de-sac) in females (Fig. 26.5), or to the level of the seminal vesicles, forming the rectovesical pouch in males. The broad ligaments reflect over the uterus, fallopian tubes, and parametrial uterine vessels and serve as the anterior boundary of the rectouterine pouch of Douglas. The cul-de-sac is the most dependent portion of the peritoneal cavity and collects fluid, blood, abscesses, and intraperitoneal drop metastases. The extraperitoneal space of the pelvis is continuous with the retroperitoneal space of the abdomen, extends to the pelvic
1828
diaphragm, and includes the retropubic space (of Retzius). Pathologic processes from the pelvis spread preferentially into the retroperitoneal compartments of the abdomen. The perineum lies below the pelvic diaphragm. The ischiorectal fossa serves as its anatomic landmark (Fig. 26.6) .
FIGURE 26.5. Pouch of Douglas. A CT of the pelvis in a woman with abundant ascites demonstrates fluid distension of the pouch of Douglas (PD) (cul-de-sac) posterior to the uterus (U) and anterior to the rectum (curved arrow). The broad ligament (long arrows) is outlined by fluid anteriorly and posteriorly.
FLUID IN THE PERITONEAL CAVITY Fluid in the peritoneal cavity originates from many different sources
1829
and varies greatly in composition. Ascites is serous fluid in the peritoneal cavity and is most commonly caused by cirrhosis, hypoproteinemia, or congestive heart failure. Exudative ascites results from inflammatory processes such as abscess, pancreatitis, peritonitis, or bowel perforation. Hemoperitoneum results from trauma, surgery, or spontaneous hemorrhage. Neoplastic ascites is associated with intraperitoneal tumors. Urine, bile, and chyle may also spread freely within the peritoneal cavity. Plain film diagnosis of ascites requires that at least 500 mL of fluid be present. Findings are: (1) diffuse increase in density of the abdomen (gray abdomen); (2) indistinct margins of the liver, spleen, and psoas muscles; (3) medial displacement of gas-filled colon, liver, and spleen away from the properitoneal flank stripe; (4) bulging of the flanks; (5) increased separation of gas-filled small bowel loops; and (6) “dog's ears― appearance of symmetric densities in the pelvis caused by fluid spilling out of the cul-de-sac on either side of the bladder. CT demonstrates fluid density in the recesses of the peritoneal cavity (Figs. 26.1B, 26.5). The CT density of the fluid gives a clue as to its composition. Serous ascites has attenuation values near water (-10 to +10 Hounsfield units [H]). Exudative ascites is usually above +15 H, but acute bleeding into the peritoneal cavity averages +45 H. US is sensitive to small amounts of fluid in the peritoneal recesses. Care must be taken to examine the most gravity-dependent portions of the peritoneal cavity (the Morison pouch and the pelvis). Simple ascites is anechoic, and exudative, hemorrhagic, or neoplastic ascites often contains floating debris. Septations in ascites are associated with an inflammatory or P.737 malignant process. MR shows limited specificity for defining the type of fluid present. Serous fluid is low intensity on T1WIs and markedly increased in intensity on T2WIs. Hemorrhagic fluid shows high signal intensity on both T1WIs and T2WIs. Serous ascites is commonly bright on gradient-echo images because of fluid motion. Pseudomyxoma peritonei refers to gelatinous ascites that occurs as a result of intraperitoneal spread of mucin-producing cells caused by rupture of appendiceal mucocele or intraperitoneal spread of
1830
mucinous adenocarcinoma of the ovary, colon, or rectum. Plain films may demonstrate punctate or ringlike calcifications scattered through the peritoneal cavity. CT demonstrates mottled densities, septations, and calcifications within the fluid. The mucinous fluid is typically loculated and causes mass effect on the liver and bowel (Fig. 26.7) . US demonstrates intraperitoneal nodules that range from hypoechoic to
strongly
echogenic.
FIGURE 26.7. Pseudomyxoma Peritonei. CT scan of a 60year-old man with intraperitoneal spread of mucinous adenocarcinoma of the colon shows loculations (arrowheads) of fluid indenting the surface of the liver (L), giving evidence of mass effect. The attenuation of the fluid measured 32 H, indicating exudative ascites.
PNEUMOPERITONEUM Free air within the peritoneal cavity is a valuable sign of bowel perforation, most commonly caused by duodenal or gastric ulcer
1831
perforation. However, additional causes of pneumoperitoneum include trauma, recent surgery or laparoscopy, and infection of the peritoneal cavity with gas-producing organisms. Postoperative pneumoperitoneum usually resolves in 3 to 4 days. Serial films demonstrate a progressive decrease in the amount of air present. Failure of progressive resolution or an increase in the amount of air present suggests a leak of bowel anastomosis or sepsis. Pneumoperitoneum in the absence of a ruptured viscus may occur with air introduced through the female genital tract by orogenital insufflation or in association with pulmonary emphysema, alveolar rupture, and dissection of air into the peritoneal cavity.
FIGURE 26.6. Perineal Tumor. A CT scan of a 12-year-old girl with a history of a rhabdomyosarcoma of the right leg demonstrates a tumor metastasis (T) in the right ischiorectal fossa. The left ischiorectal fossa (IRF) shows its normal
1832
appearance as a triangle of fat bordered by the rectum (R), obturator internus muscle (OI), and the gluteus muscles (GM). The ischiorectal fossa is entirely below the levator ani and is part of the perineum. c, tip of the coccyx; IT, ischial tuberosities.
Plain film evidence of pneumoperitoneum is best seen on radiographs obtained with the patient in the standing or sitting position. Upright chest radiographs are the most sensitive for free air. Small amounts of air are clearly demonstrated beneath the domes of the diaphragm. Left lateral decubitus and cross-table lateral views may be used with very ill patients to demonstrate air outlining the liver. Signs of pneumoperitoneum on supine radiographs (Fig. 26.8) include the following: (1) gas on both sides of the bowel wall (Rigler sign), (2) gas outlining the falciform ligament, (3) gas outlining the peritoneal cavity (the “football sign―), and (4) triangular or linear localized extraluminal gas in the right upper quadrant. On CT, small amounts of extraluminal gas may be confused with gas within the bowel and be surprisingly difficult to recognize. Images should be examined at lung windows (window level –600 H, window width 1,000 H) to detect free P.738 intraperitoneal air. The peritoneal recess between the liver and diaphragm (Fig. 26.9) is a good place to look for pneumoperitoneum on CT.
1833
FIGURE
26.8. Pneumoperitoneum:
Radiograph. Supine
radiograph in a patient with a perforated gastric ulcer demonstrates visualization of both sides of the bowel wall (arrowheads), free air outlining the falciform ligament (arrow) , free air outlining the edge of the liver (curved arrow), and free air outlining the pericolic gutters (asterisk) .
ABDOMINAL
CALCIFICATIONS
Intra-abdominal calcifications may be an important sign of intraabdominal disease and should be searched for on every imaging study of the abdomen. CT and US are more sensitive to detection of calcifications than are plain radiographs. However, the high spatial
1834
resolution of plain film radiography commonly provides characteristic findings that allow a specific diagnosis of the nature of the calcification (2) . Vascular calcifications are common in the aorta and iliac vessels (see Fig. 26.13) of older individuals. Plaquelike vascular calcifications overlie the lumbar spine and sacrum and commonly require detailed inspection to detect. Aneurysms of the aorta are manifest by luminal diameter exceeding 3 cm, as measured between calcifications in the aortic wall (Fig. 26.10). Ringlike calcified aneurysms most commonly involve the splenic or renal arteries. Phleboliths are calcified thrombi in veins—most commonly visualized in the lateral aspects of the pelvis. They are round or oval calcifications up to 5 mm that commonly contain a central lucency. They may be mistaken for urinary tract calculi.
FIGURE 26.9. Pneumoperitoneum: CT. A collection of air (arrow) is seen within the peritoneal space between the liver (L) and the diaphragm (arrowhead). This is a prime area to search to
1835
detect small amounts of free intraperitoneal air on CT. This patient had a torn jejunum as a result of trauma from a motor vehicle collision.
Calcified lymph nodes result most commonly from granulomatous diseases such as tuberculosis or histoplasmosis. The calcification is usually mottled and 10 to 15 mm in size. Mesenteric nodes are the most commonly calcified.
Gallstones
and
Gallbladder
Only about 15% of gallstones contain sufficient calcium to be identified on plain film. Most calcified gallstones contain calcium bilirubinate and have a laminated appearance, with a dense outer rim and more radiolucent center. When multiple gallstones are present, they are commonly faceted. Calcifications in the gallbladder wall (porcelain
gallbladder)
(Fig. 26.11) are plaquelike and oval in
configuration, conforming to the size and shape of the gallbladder. Milk of calcium bile is a suspension of radiopaque crystals within gallbladder bile. Layering of the suspension can be demonstrated on erect films.
Urinary
Calculi
About 85% of urinary calculi are visible on plain film. They range in size from punctate up to several centimeters. Most characteristic are the staghorn calculi, which assume the shape of the renal collecting system (Fig. 26.12). Renal calculi are differentiated from gallstones on radiographs by oblique P.739 projections that confirm their posterior position, as opposed to the more anterior positions of gallstones. Ureteral calculi may be seen anywhere along the course of the ureter, but they are most common at the areas of narrowing: the ureteropelvic junction, the pelvic brim, and the vesicoureteral junction. Bladder calculi (Fig. 26.13) are single or multiple and commonly laminated, may be any size, and
1836
usually lie near the midline of the pelvis. Calculi within bladder diverticula may be eccentric to the bladder.
FIGURE
26.10. Abdominal
Aortic
Aneurysm. Plain radiograph
demonstrates an aneurysm of the abdominal aorta, evidenced by wide separation of calcifications in the aortic wall (arrowheads) . Calcifications in the wall overlying the spine may be difficult to visualize. A film taken with the patient in the left posterior oblique position will project the aorta away from the spine and make visualization of aortic wall calcifications easier.
1837
FIGURE 26.11. Bladder Calculi. Numerous calculi (arrows) in the bladder are evident on this plain radiograph of the pelvis. The large prostate (P, between closed arrows), responsible for urinary stasis leading to stone formation, makes a mass impression on the layering stones. Also evident are atherosclerotic calcifications in the iliac arteries (curved
1838
arrows) .
FIGURE
26.12. Porcelain
Gallbladder.
Cone-down
radiograph
of the right upper quadrant of the abdomen demonstrates calcification in the wall of the gallbladder (arrow). This finding is indicative of chronic obstruction of the cystic duct, chronic gallbladder inflammation, and an increased risk of gallbladder carcinoma.
Liver and spleen granulomas are usually multiple, small, and dense. They are healed foci of tuberculosis, histoplasmosis, or other granulomatous disease. Appendicoliths
and
enteroliths are concretions within the lumen
of the bowel. Most are round or oval and have concentric laminations. Appendicoliths are strongly indicative of acute appendicitis in patients with acute abdominal pain. Enteroliths are most common in the colon and often caused by calcium deposition on an undigestible material such as a fruit pit.
1839
Calcified adrenal glands are associated with adrenal hemorrhage in the newborn, tuberculosis, and Addison disease. The calcification is mottled and in the location of the adrenal glands on either side of the first lumbar vertebra (Fig. 26.14) . Pancreatic pancreatitis
calcification is associated with chronic alcohol-induced and hereditary pancreatitis. P.740
The calcifications are caused by pancreatic calculi and are usually coarse and of varying size (Fig. 26.15) .
FIGURE
26.13. Staghorn
Calculus. A plain radiograph reveals
a large calculus occupying the collecting system of the left kidney and assuming its shape. Staghorn calculi (S) are usually composed of struvite and form in the presence of chronic urinary infection.
Calcified cysts may be found in the kidneys, spleen, liver, appendix, and peritoneal cavity. Calcifications in the wall of a cyst
1840
are curvilinear or ring shaped (Fig. 26.16). Echinococcus cysts commonly calcify and may be found in any intra-abdominal organ as well as within the peritoneal cavity.
FIGURE 26.14. Adrenal Calcifications. Plain radiograph of the abdomen in a 4-year-old demonstrates calcification of both adrenal glands (arrows) resulting from bilateral adrenal hemorrhage as an infant.
1841
FIGURE 26.15. Pancreatic Calcifications. Coarse and punctate calcifications (arrow) extend upward across the left upper quadrant in this patient with chronic alcoholic pancreatitis. Calcifications in the pancreatic head (arrowhead) are obscured by the spine.
Tumor
calcification
A wide variety of different tumors of abdominal organs may contain calcifications. The coarse “popcorn― calcifications of uterine leiomyomas are most characteristic. Benign cystic teratomas may form teeth or bone. Calcified peritoneal metastases of ovarian or colon mucinous cystadenocarcinoma may outline the peritoneal cavity (Fig. 26.17). Renal cell carcinoma calcifies in up to 25% of cases. Soft tissue calcifications may be seen with hypercalcemic states, idiopathic calcinosis, and old hematomas. P.741 Calcified injection granuloma from quinine, bismuth, and calcium
1842
salts of penicillin are commonly evident in the buttocks. Cysticercosis causes characteristic “rice-grain― calcifications in muscles.
FIGURE
26.16. Calcified Renal Cyst. Scout radiograph for an
excretory urogram shows the rim calcification (arrow) characteristic of wall calcification in a renal cyst.
1843
FIGURE 26.17. Tumoral Calcifications. Radiograph of the abdomen demonstrates cloudlike calcifications in the distribution of peritoneal recesses. These calcifications were caused by intraperitoneal spread of a papillary serous cystadenocarcinoma of the ovary.
Bowel contents may include bone, pits, seeds, birdshot, or medications containing iron or other heavy metals that result in abdominal
opacities.
Peritoneal calcifications may be nodular or sheetlike and result most commonly from peritoneal dialysis, previous peritonitis, or peritoneal
carcinomatosis
(Fig. 26.17) .
1844
ACUTE
ABDOMEN
The differential diagnosis of patients presenting with acute abdominal pain is extremely broad (Table 26.1). Accurate and efficient diagnosis requires cooperation between the referring physician and the radiologist to select the imaging method most likely to provide the correct diagnosis. Routine assessment of the acute abdomen commonly includes the “acute abdomen series,― which consists of an erect posteroanterior chest radiograph and supine and erect or decubitus radiographs of the abdomen. The chest radiograph provides optimal detection of pneumoperitoneum and intrathoracic diseases that may present with abdominal complaints. The supine abdominal film permits diagnosis of many acute abdominal conditions, and a horizontal-beam abdominal film adds confidence to the diagnosis. CT or US are routinely obtained to provide a definitive diagnosis (3,4,5) .
TABLE 26.1 Common Causes of Acute Abdomen
Appendicitis
Peritonitis
Acute
cholecystitis
Intraperitoneal
abscess
Acute
pancreatitis
Retroperitoneal
abscess
Acute
diverticulitis
Bowel
Acute
ulcerative
colitis
Pseudomembranous
Amebiasis
colitis
obstruction
Urinary
tract
infection
Urinary
tract
obstruction
Pelvic
1845
inflammatory
diseases
Acute
intestinal
Normal
ischemia
Abdominal
Gas
Pattern
Interpretation of plain abdominal radiographs routinely includes assessment of gas, fluid, soft tissue, fat, and calcium densities (Fig. 26.18). Normal gas in the abdomen is predominantly swallowed air. Air–fluid levels are seen in normal patients, commonly in the stomach, often in the small bowel, but never in the colon distal to the hepatic flexure. Normal air–fluid levels in the small bowel should not exceed 2.5 cm in length. Small bowel gas usually appears as multiple small, random gas collections scattered throughout the abdomen. Small bowel gas is increased in patients who chronically swallow air or drink carbonated beverages. A normal intestinal gas pattern varies from no intestinal gas to gas within three to four variably shaped small intestinal loops measuring less than 2.5 to 3 cm in diameter. The normal colon contains some gas and fecal material and varies in diameter from 3 to 8 cm, with the cecum having the largest diameter. The term “nonspecific abdominal gas pattern― has no precise meaning and should not be used.
Dilated
Bowel
The small bowel is dilated when it exceeds 2.5 to 3.0 cm in diameter. The colon is dilated when it exceeds 5 cm in diameter, and the cecum is dilated when it exceeds 8 cm in diameter. In adults, dilated small bowel can usually be differentiated from dilated large bowel by assessment of location and anatomic features. The small bowel is more central in the abdomen and is characterized by valvulae conniventes, which cross the entire diameter of the lumen. Dilated small bowel rarely exceeds 5 cm in diameter, although the large bowel is not considered dilated until it exceeds 5 cm diameter. The large bowel is more peripheral in the abdomen and is characterized by haustra that extend only part way across the lumen. The large bowel contains fecal material that has a characteristic mottled
1846
appearance. The cecum, which has the largest P.742 normal diameter of the large bowel, always dilates to the greatest extent irrespective of the site of obstruction.
FIGURE 26.18. Normal Bowel Gas Pattern. Supine radiograph shows the normal distribution of gas in the stomach (large arrow) and duodenum (small arrow). The normal mottled pattern of stool is seen in the distribution of the right colon (arrowhead) . A few gas collections within the small bowel (curved arrow) are
1847
seen in the pelvis.
Adynamic
Ileus
The word ileus means “stasis― and does not differentiate mechanical obstruction from nonmechanical stasis. The terms adynamic ileus, paralytic ileus, and nonobstructive ileus are used interchangeably and refer to stasis of bowel contents because of decreased or absent peristalsis. Common causes of adynamic ileus are listed in Table 26.2. Adynamic ileus typically demonstrates diffuse symmetric, predominantly gaseous, distension of bowel. The small bowel, stomach, and colon are proportionally dilated without an abrupt termination. More loops are dilated than with obstruction. Occasionally, adynamic ileus may result in a gasless abdomen with dilated loops of bowel that are filled only with fluid. US is useful in confirming decreased or absent peristalsis, although examination may be difficult if large amounts of gas are present.
TABLE 26.2 Common Causes of Adynamic Ileus
Drugs Atropine, glucagon, morphine, barbiturates, phenothiazines Metabolic causes Diabetes mellitus, hypothyroidism, hypokalemia, hypercalcemia Inflammation Intestinal: gastroenteritis Extra-intestinal: peritonitis, cholecystitis, abscess
pancreatitis,
Postoperative: resolves in 4 to 7 days Posttraumatic Post–spinal
injury
1848
appendicitis,
Sentinel loop refers to a segment of intestine that becomes paralyzed and dilated as it lies next to an inflamed intraabdominal organ. In essence, it is a short segment of adynamic ileus that appears as an isolated loop of distended intestine that remains in the same general position on serial films (Fig. 26.19). A sentinel loop should alert the clinician to the presence of an adjacent inflammatory process. A sentinel loop in the right upper quadrant suggests acute cholecystitis, hepatitis, or pyelonephritis. In the left upper quadrant, pancreatitis, pyelonephritis, or splenic injury may be suspected. In the lower quadrants, diverticulitis, appendicitis, or Crohn disease are causes of a sentinel loop. Toxic
salpingitis,
cystitis,
megacolon is a manifestation of fulminant colitis
characterized by extreme dilation of all or a portion of the colon. In this state, peristalsis is absent and the large bowel loses all tone and contractility. The patient has progressive abdominal distension and is toxic, febrile, and P.743 obtunded. Bowel sounds and bowel movements are absent. The bowel wall becomes like “wet blotting paper,― and the risk of perforation is extreme. Mortality approaches 20% in toxic megacolon. Acute ulcerative colitis is the most common cause of toxic megacolon (Table 26.3). Plain films demonstrate distension of the colon with absent haustra. Dilation of the transverse colon up to 15 cm diameter is often the most striking observation. The diagnosis is suggested when the diameter of the colon exceeds 5 cm and the mucosa appears abnormal. Pseudopolyps caused by islands of edematous mucosa surrounded by extensive ulceration appear as soft tissue nodules within the air-distended colon. CT demonstrates a distended colon filled with air and fluid. The wall of the colon is thin but has an irregular nodular contour; air may be seen within the colon wall. Barium enema is absolutely contraindicated because of risk of perforation.
1849
FIGURE
26.19. Sentinel
Loop. Daily serial radiographs on this
patient demonstrated a persistent loop of dilated small bowel (arrow) in the same location. This sentinel loop was caused by acute pancreatitis. Normal gas pattern is present in the right colon (arrowhead). The abdomen is otherwise devoid of intestinal gas.
TABLE 26.3 Causes of Toxic Megacolon
Ulcerative colitis: 75% of cases
Amebic
Pseudomembranous
Ischemic
colitis
Bacterial
colitis:
Crohn
colitis
colitis
typhoid
1850
colitis
cholera,
Mechanical bowel obstruction refers to stasis of bowel contents above a focal lesion. The obstruction may be caused by obturation (occlusion by a mass in the lumen), stenosis owing to intrinsic bowel disease, or compression of the lumen by extrinsic disease. The goal of imaging is to confirm the presence of an obstruction, identify its level, and demonstrate its cause. Radiographs can confirm the presence of bowel obstruction 6 to 12 hours before the diagnosis can usually be made clinically. When bowel obstruction occurs, the lumen of the bowel proximal to the obstruction progressively dilates because of continued secretions, swallowed fluid, air, and food, and eventual cessation of absorption. Stasis results in the overgrowth of bacteria and production of toxins that may injure the mucosa. Compromise of blood supply may occur because of distension of the bowel wall and increased intraluminal pressure. A variety of terms used clinically must be understood. Complete obstruction means the lumen is totally occluded, but partial obstruction means some bowel contents pass through. Simple obstruction refers to blockage of the luminal contents without interference of blood supply. Strangulation obstruction means that the blood supply to the bowel wall is impaired. Most strangulation obstructions are closedloop obstructions— i.e., the bowel loop segment is blocked at both ends. This occurs with incarcerated hernias and volvulus.
SMALL
BOWEL
OBSTRUCTION
Small bowel obstruction accounts for 20% of surgical admissions for acute abdominal pain and 80% of all intestinal tract obstruction. The causes of small bowel obstruction are listed in Table 26.4. In the Western world, postsurgical adhesions account for 75% of small bowel obstructions, whereas in developing nations, 80% of small bowel obstructions are caused by incarcerated hernia but only 10% are caused by adhesions. Patients present clinically with crampy abdominal pain, abdominal distention, and vomiting. Plain films are diagnostic in only 50% to 60% of cases. Findings of small bowel obstruction are as follows: (1) dilated loops of small bowel (>3 cm) disproportionate to more distal small bowel or colon, (2) small bowel air–fluid levels that exceed 2.5 cm in length, (3) air–fluid levels
1851
at differing heights within the same loop (strong evidence of obstruction) (Fig. 26.20), and (4) small bubbles of gas trapped P.744 between folds in dilated, fluid-filled loops producing the “string of pearls― sign, a row of small gas bubbles oriented horizontally or obliquely across the abdomen. The level of obstruction is determined by dilated loops above the obstruction and normal or empty loops below the obstruction. Stepladder or hairpin loops of small bowel are most characteristic. Inguinal hernias, easily overlooked clinically in the obese, may be evident on radiographs. CT has become the imaging method of choice when the diagnosis is equivocal (6,7). CT offers the advantage of revealing the cause of obstruction in 70% to 90% of cases. CT diagnosis is based upon demonstration of a transition site between small bowel loops dilated with fluid or air and collapsed bowel loops distal to the obstruction (Fig. 26.21). A potential pitfall is the common finding of a collapsed descending colon, even in patients with adynamic ileus. Bowel obstruction should not be diagnosed in this setting unless an obstructing lesion is visualized at the splenic flexure. The “small-bowel feces― sign is uncommon but definitive CT evidence of bowel obstruction. Particulate feculent matter mixed with gas bubbles is seen within dilated small bowel. Abrupt beaklike narrowing, without other evidence of lesions, is indicative of adhesions as the cause of obstruction. Other causes, including tumor, abscess, inflammation, hernia,
intussusception,
etc.,
have
characteristic
1852
findings.
TABLE 26.4 Causes of Small Bowel Obstruction
Adhesions
Volvulus
Postsurgical
Gallstone
Postinflammatory
Parasites
Incarcerated
Malignancy:
hernia
usually
Bolus
metastatic
Intussusception
1853
Foreign
ileus
of Ascaris
body
FIGURE
26.20. Small
Bowel
Obstruction. Erect radiograph of
the abdomen reveals dilated air-filled loops of small bowel containing air–fluid levels at different heights within the same loop (arrows). Note the valvulae conniventes (arrowhead) that extend across the entire diameter of the bowel lumen. The small bowel obstruction was caused by adhesions.
1854
FIGURE
26.21. Small
Bowel
Obstruction. CT demonstrates
dilated fluid- and air-filled loops of small intestine (wide
arrows) .
A transition to nondilated bowel is evident in the distal ileum (arrowhead), indicating an obstructing adhesion at that point. The more distal small bowel (thin colon (curved
arrows) and the descending
arrow) are collapsed.
Strangulation obstruction is associated with changes in the bowel wall and mesentery caused by impairment of blood supply. CT findings are: (1) circumferential wall thickening (>3 mm), (2) edema of the bowel wall (target or halo appearance of lucency in the bowel wall), (3) lack of enhancement of the bowel wall (most specific sign), (4) haziness or obliteration of the mesenteric vessels, and (5) infiltration of the mesentery with fluid or hemorrhage. Because most cases are caused by closed-loop obstruction, findings of that condition are commonly present as well. Closed-loop obstruction is indicated by the following CT signs: (1) radial distribution of dilated small bowel with mesenteric vessels converging toward a focus of torsion, (2) U-shaped or C-shaped dilated small bowel loop, (3) “beak― sign at the site of torsion seen as fusiform tapering of a dilated bowel loop, (4) “whorl―
1855
sign of tightly twisted mesentery seen with volvulus. Intussusception is a major cause of small bowel obstruction in children but is less common in adults (8). In adults, intussusception is often chronic, intermittent, or subacute, and it is usually caused by a polypoid tumor, such as lipoma. Additional causes are malignant tumor, Meckel diverticulum, lymphoma, mesenteric nodes, and foreign bodies. Enteroenteric intussusception occurs with small bowel tumors and sprue. Ileocolic intussusception is usually idiopathic in children but is caused by a mass in adults. Colocolic intussusception is common in adults but rare in children. Plain films in intussusception demonstrate small bowel obstruction and a soft tissue mass. Barium studies demonstrate barium trapped between the intussusceptum and the receiving bowel, forming a coiled-spring appearance. CT is usually diagnostic, demonstrating a characteristic target-like intestinal mass (Fig. 26.22). On transverse section, the inner central density is the invaginating loop, surrounded by fatdensity mesentery that is enveloped by the receiving loop. US exhibits a similar “donut― configuration of alternating hyperechoic and hypoechoic rings representing alternating mucosa, muscular wall, and mesenteric fat tissues in cross section. Asymptomatic,
incidental,
transient
intussusception
without P.745
associated small bowel infection is an increasingly common finding on CT.
1856
FIGURE
26.22. Enteroenteric
Intussusception. CT performed
without intravenous contrast demonstrates intussusception of small bowel, which was caused by a renal cell carcinoma metastasis to small bowel. The invaginating bowel (arrowhead) and its accompanying fat-containing mesentery (thin arrow) are seen inside of the receiving bowel (wide arrow) .
Gallstone ileus is a misnomer, since it is a cause of obstruction that should be suspected in any elderly woman with small bowel obstruction. It is the cause of 24% of small bowel obstructions in patients over age 70. Because it is a disease of the elderly, insidious in onset, and difficult to diagnose, mortality is increased fivefold over mortality for small bowel obstruction caused by adhesions. Bowel obstruction is caused by a large gallstone that erodes through the gallbladder wall and passes into the intestine, usually creating a cholecystoduodenal fistula. The gallstone most commonly lodges in the distal ileum. Specific radiographic signs are present in only about half of patients. Rigler's triad consists of the following: (1) dilated small bowel loops (80% of cases), (2) air in the biliary tree or gallbladder (67%), and (3) calcified gallstone in an ectopic location (50%). Barium studies should include instillation of contrast into the duodenum to demonstrate passage of barium into the biliary tree. Nonopaque obstructing gallstones are demonstrated as an
1857
intraluminal
mass.
LARGE
BOWEL
OBSTRUCTION
Large bowel obstruction is predominantly a condition of older adults and accounts for about 20% of all bowel obstructions. The cecum dilates to the greatest extent, irrespective of the site of large bowel obstruction. When the cecum exceeds 10 cm in diameter, it is at high risk for perforation, with attendant risks of peritonitis and septic shock. The common causes of large bowel obstruction are listed in Table 26.5. Most colonic obstructions occur in the sigmoid colon, where the bowel lumen is narrower and stool is more formed. Plain films are commonly diagnostic in large bowel obstruction, demonstrating dilation of the colon from the cecum to the point of obstruction. The colon distal to the obstruction is devoid of gas. When the ileocecal valve is competent, the small bowel usually contains little gas; the colon is unable to decompress into the small bowel and gaseous distension of the cecum is progressive. When the ileocecal valve is incompetent, gaseous distension of the small bowel is often present; the colon can decompress into the ileum and jejunum, and the risk of perforation of the cecum is reduced. Air–fluid levels distal to the hepatic flexure are strong evidence of obstruction unless the patient has had an enema.
TABLE 26.5 Causes of Large Bowel Obstruction
1858
Colon
carcinoma
Metastatic disease, malignancies
(50%–60%)
especially
pelvic
Diverticulitis
Volvulus
Fecal
impaction
Amebiasis
Ischemia
Adhesions
Sigmoid
volvulus is most common in the elderly and in individuals
on high-residue diets. The sigmoid colon twists around its mesentery, resulting in a closed-loop obstruction. The proximal colon dilates while the rectum empties. Plain radiographs are usually diagnostic. The sigmoid colon appears markings, arising from the abdomen and often to the by the lateral walls of the
as a large gas-filled loop without haustral pelvis and extending high into the diaphragm. The three white lines formed loop and the summation of the two
opposed medial walls of the loop converge inferiorly into the left iliac fossa (Fig. 26.23). Barium enema demonstrates obstruction that tapers to a beak at the point of the twist, usually approximately 15 cm above the anal verge. Mucosal folds spiral into the beak at the point of obstruction. CT is rarely utilized. Sigmoid volvulus causes 3% to 8% of large bowel obstructions in adults and has a reported mortality of 20% to 25%.
1859
FIGURE 26.23. Sigmoid Volvulus. Radiograph of the abdomen demonstrates the characteristic massive dilation of the sigmoid colon (S) arising from the pelvis and extending to the left diaphragm. Three lines representing the walls of the twisted loop converging to the left lower quadrant are evident (1, 2, 3) .
P.746 Cecal volvulus causes 1% to 3% of large bowel obstructions in adults and occurs most frequently in adults 30 to 60 years of age. Twisting of the cecum occurs in the ascending colon above the ileocecal valve. Two types of twisting may occur. Twisting along the long axis of the ascending colon accounts for two thirds of cases and
1860
is usually referred to as true cecal volvulus. The cecum may twist and remain in the right lower quadrant of the abdomen, or it may twist and invert to occupy the left upper quadrant. Cecal bascule refers to a folding of the cecum into a position anteromedial to the ascending colon, rather like folding the toe of a sock back on itself. Bascule accounts for about one third of cases. Radiographs are usually diagnostic, showing a markedly distended loop of bowel extending from the right lower quadrant upward to the left upper quadrant or epigastrium (Fig. 26.24). A single air–fluid level is often present within the whereas the distal colon the cecum as it rotates. loop and the site of the
dilated loop. The small bowel is distended, is decompressed. The distal ileum encircles CT is rarely needed but shows the dilated twist or fold. Signs of small bowel
obstruction may be present. Displacement of the appendix to the upper abdomen is a definitive sign. Images should be examined carefully for evidence of ischemia. A contrast enema demonstrates a beaklike or foldlike termination at the point of obstruction in the ascending colon. Mortality rates of 20% to 40% are reported because of delays in diagnosis.
1861
FIGURE 26.24. Cecal Volvulus. Supine abdominal radiograph demonstrates displacement of the dilated cecum (arrowheads) to the epigastrium. The more distal colon is collapsed. Previously administered IV contrast is seen in the bladder (B). The bladder base is markedly elevated by an enlarged prostrate (P).
Fecal Impaction is the most common cause of large bowel obstruction in elderly and bedridden patients. Radiographs demonstrate a large mass of stool with a characteristic mottled appearance in the distal colon. Following disimpaction, colonoscopy or barium enema should be performed to search for an obstructing carcinoma that may have caused the fecal impaction.
1862
BOWEL
ISCHEMIA
AND
INFARCTION
Bowel ischemia, potentially leading to infarction, is a true emergency, with high associated morbidity and mortality. Insufficient blood supply to small or large bowel may be transient and reversible or lethal (9). Causes include arterial occlusion of the mesenteric arteries by thrombus, embolus, volvulus, vasculitis, or external compression; hypotension related to congestive heart failure, sepsis, or blood loss; vasoconstrictive medications such as ergotamine, digitalis, or norepinephrine; and impaired venous drainage caused by venous thrombosis, tumor, adhesions, or volvulus. Ischemic injury starts at the mucosa and extends progressively through the bowel wall to the serosa. Contrast-enhanced multidetector CT is the imaging method of choice. Findings of bowel ischemia include: (1) circumferential or nodular thickening (>5 mm) of the bowel wall with infiltration of low-density edema or high-density blood, resulting from mucosal injury; (2) “thumbprinting― (see Fig. 32.20) resulting from this nodular infiltration of the bowel wall; (3) dilatation of the bowel lumen (>3 cm for small bowel; >5 cm for colon; >8 cm for cecum); (4) pneumatosis intestinalis (see following paragraph); (5) edema or hemorrhage into the mesentery; (6) engorged mesenteric vessels; (7) thrombosis of mesenteric arteries or veins; (8) poor enhancement of the bowel wall along its mesenteric border, which is evidence of ischemia; (9) poor or absent mucosal enhancement with thinning of the bowel wall, which is evidence of bowel infarction; (1 0) ascites, which is commonly present (1 0) . Pneumatosis intestinalis refers to the presence of gas within the bowel wall (1 1). It may occur as a benign entity without clinical significance, or it may be an important finding of bowel ischemia. Causes of pneumatosis intestinalis may be lumped into four categories: (1) bowel necrosis, usually associated with other radiographic and clinical signs of bowel ischemia; (2) mucosal disruption caused by ulcers, mucosal biopsies, trauma, enteric tubes, or inflammatory bowel disease; (3) increased mucosal permeability related to immunosuppression in AIDS, organ transplantation, or chemotherapy; and (4) pulmonary disease resulting in alveolar
1863
disruption and dissection of air P.747 along interstitial pathways to the bowel wall. Causes of the latter include chronic obstructive pulmonary disease, asthma, cystic fibrosis, mechanical ventilation, and chest trauma. Interpretation of the imaging finding of pneumatosis must be correlated with the clinical condition of the patient. Pneumatosis in asymptomatic patients is very likely benign and incidental. Pneumatosis in seriously ill patients with abdominal pain or distension is more likely to be a sign of bowel ischemia. Pneumatosis appears on radiographs or CT as cystic air bubbles (up to several centimeters) or linear streaks of air within the bowel wall, especially in its most gravity-dependent aspect (Fig. 26.25). On CT, air bubbles within the lumen may mimic pneumatosis but should always be seen adjacent to the nondependent bowel wall. Turning the patient and rescanning may clarify the diagnosis. Air may also be evident within mesenteric vessels or within portal veins in the liver.
FIGURE 26.25. Pneumatosis Intestinalis. A. Digital radiograph scout scan from CT reveals pneumatosis of the colon as dark linear streaks of air (arrowheads) in the colon wall. Both small and large bowel are markedly dilated. B . CT image of the same patient viewed with lung windows confirms the presence of
1864
air in the colon wall (arrowheads). The small bowel (SB) is dilated. At surgery, both the small and large bowels were infarcted. The patient expired.
ABDOMINAL
TRAUMA
CT of the abdomen and pelvis has become an integral part of the emergency evaluation of victims of blunt abdominal trauma (1 2,1 3) . CT characterizes the precise nature of traumatic injury and is used to direct therapy, especially in patients with coexisting injuries, head trauma, or impaired consciousness caused by injury, drugs, or alcohol. Candidates for CT are patients with a history of significant blunt trauma who are hemodynamically stable. Focused abdominal sonograms for trauma (FAST scans) are often used to detect the presence of intraperitoneal fluid and triage trauma patients for CT (1 4). CT findings of traumatic injury include: (1)
hemoperitoneum:
acute blood within the peritoneal cavity measuring 30 to 45 H (Fig. 26.26); (2) sentinel clot: a focal collection of clotted blood (>60 H) that may be seen in the peritoneal cavity adjacent to an injured organ (Fig. 26.26); (3) active bleeding, as evidenced by hyperdense fluid (85 to 370 H) (Fig. 26.27) seen during the arterial phase of scanning with multidetector CT; (4) free air within the peritoneal cavity (Fig. 26.9), which is an insensitive sign of bowel injury provided that diagnostic peritoneal lavage has not been performed; (5) free contrast within the peritoneal cavity, which may result from oral contrast leaking from injured bowel or IV contrast leaking from a ruptured bladder; (6) subcapsular hematomas, which appear as crescent-shaped collections confined by the capsule of the injured organ; (7) intraparenchymal hematomas, which appear as irregularly shaped low-density areas within a contrast-enhanced solid organ; (8) lacerations, defined by organ; (9) the organ's
which appear as jagged linear defects (Fig. 26.27) lower-density blood within a contrast-enhanced injured absence of organ enhancement, which reflects damage to arterial supply; and (1 0) infarctions, which are seen as
zones of decreased contrast enhancement that extend to the capsule
1865
of a solid organ (Fig. 26.28) (1 5,1 6,1 7,1 8) .
FIGURE
26.26. Hemoperitoneum
and
Sentinel
Clot. CT scan
shows high-attenuation fluid in the peritoneal recesses, indicating hemoperitoneum (H). A sentinel clot (arrow) stands out as a high-attenuation collection within the lower-attenuation liquid blood. The location of the clot suggests injury to the liver (L). A laceration of the left lobe of the liver was found at surgery.
1866
FIGURE
26.27. Active
Hemorrhage:
Liver
Laceration. CT
shows a jagged laceration (arrowheads) of the liver (L) filled with blood. A focus of continuing active hemorrhage (arrow) is seen as an ill-defined collection of high-attenuation contrast agent. Hemoperitoneum (H) is evident in the peritoneal recesses. Sp; spleen. St; stomach.
1867
FIGURE
26.28. Renal
Infarction. Postcontrast CT reveals a
lack of enhancement (arrow) of the posterior portion of the left kidney (LK), which occurred as a result of an intimal tear and thrombosis of a branch renal artery occurring during a motor vehicle collision. Note that the defect in enhancement extends to the capsule of the kidney, indicating acute renal vascular injury.
P.748
LYMPHADENOPATHY The abdomen and pelvis contain more than 230 lymph nodes that may be involved in a wide variety of neoplastic and inflammatory diseases (1 9). CT, US, and MR can evaluate the entire abdominopelvic lymphatic system and have replaced the older technique of lymphangiography. Unfortunately, none of the crosssectional imaging methods can demonstrate tumor involvement of a lymph node by alteration of internal architecture. The criteria for pathologic involvement are based primarily on alterations in node
1868
size (Table 26.6). Short-axis measurements of lymph node size are preferred to determine abnormal enlargement. Morphologic patterns of pathologic lymphadenopathy include single enlarged nodes, multiple separate lobulated enlarged nodes, or bulky conglomerate masses of lymph nodes (Fig. 26.29). Calcification in enlarged nodes may be seen with inflammatory adenopathy, mucinous carcinomas, sarcomas, and treated lymphoma. CT to detect adenopathy requires optimal contrast opacification of blood vessels and the GI tract. Normal nodes are oblong in shape, homogeneous in configuration, and have short-axis diameters below the limits listed in Table 26.6. Most pathologically enlarged nodes have CT densities that are slightly lower than that of skeletal muscle. Low-density nodal metastases are commonly seen with nonseminomatous testicular carcinoma, tuberculosis, and occasionally lymphoma. US is P.749 almost equal to CT in accuracy for detection of lymphadenopathy; however, a skillful dedicated examination is required. Lymphoma typically produces hypoechoic or even anechoic lymphadenopathy. Masses of retroperitoneal nodes may silhouette segments of the normally echogenic wall of the aorta (the “sonographic silhouette sign―). The “sandwich sign― refers to entrapment of mesenteric vessels by masses of enlarged lymph nodes in the mesentery. MR usually provides excellent differentiation of lymph nodes from blood vessels because of flow void within vessels. However, because of the current lack of an effective GI contrast agent, loops of bowel are commonly confused with masses of nodes. On T1WIs, lymph nodes show low signal intensity compared to surrounding fat. On T2WIs, lymph nodes show high signal intensity compared to muscle. The fat-saturation technique highlights pathologic adenopathy.
TABLE 26.6 Abdominal and Pelvis Lymphadenopathy: Upper Limits of Normal Node Size by Location
1869
Node
Location
Maximum Dimension
Comments
(mm) Retrocrural
6
May enlarge from disease above or below the diaphragm
Retroperitoneal
10
Multiple nodes 8 to 10 mm in size are usually
Gastrohepatic ligament
8
abnormal
Must differentiate lymphadenopathy from coronary varices
Porta
hepatis
6
May cause biliary obstruction
Celiac and superior mesenteric arteries
10
Also called nodes
preaortic
Pancreaticoduodenal
10
Commonly involved by lymphoma and G I carcinoma
Perisplenic
10
Involved
by
lymphoma and G I carcinoma
Mesenteric
10
In the small bowel mesentery
1870
Pelvic
15
Most
commonly
involved by pelvic tumors
FIGURE
26.29. Hodgkin
adenopathy
Lymphoma. CT shows bulky confluent
(arrows) in the retroperitoneum surrounding the
aorta (Ao) and displacing the inferior vena cava (IVC) anteriorly. Masses of lymphoma (arrowhead) are also present in the spleen.
Hodgkin lymphoma is responsible for 20% to 40% of all lymphoma and is characterized histologically by the presence of the ReedSternberg cell. Hodgkin lymphoma has a bimodal age distribution; it most commonly affects patients aged 25 to 30 years and over 70 years. At presentation, abdominal adenopathy is present in about 25% of cases. The spleen is involved in about 40% of cases and the liver in about 8%. Involvement of the GI and urinary tracts is much less common with Hodgkin than with non-Hodgkin lymphoma. Non-Hodgkin
lymphoma is responsible for 60% to 80% of
lymphomas. Non-Hodgkin lymphoma is a heterogeneous group of
1871
disorders with a confusing array of changing names and classifications. Disease severity ranges from indolent to very aggressive. Non-Hodgkin lymphomas are particularly common in patients with AIDS and other immunocompromising conditions. The non-Hodgkin lymphomas commonly involve extranodal sites, including the GI and urinary tracts. At presentation, abdominal adenopathy is present in about 50% of cases. The spleen is involved in about 40% of cases and the liver in about 14%. Posttransplantation
lymphoproliferative
disorder is a spectrum
of lymphoid hyperplasias and neoplasias in patients who have received solid organ transplants and immunosuppressive therapy. Up to 20% of transplant recipients may be affected. The disorder is thought to result from an Epstein-Barr virus–induced proliferation of B lymphocytes that is usually opposed by functioning T cells. However, T-cell function is limited by immunosuppressive therapy. The proliferation ranges from polyclonal, P.750 benign, and reversible to aggressive and difficult-to-treat monoclonal lymphoma. Extranodal involvement with discrete solitary, multiple, or infiltrative masses within organs is most common. Lymph node enlargement occurs near the transplanted organ but may also occur at remote sites, i.e., in the abdomen but associated with a heart or lung transplant. CT may reveal lymphadenopathy before the patient becomes symptomatic. Treatment is reduction of immunosuppressive therapy.
ABDOMINOPELVIC
TUMORS
AND
MASSES
Peritoneal mesothelioma is an uncommon primary tumor of the peritoneal membrane. Approximately 20% to 30% of mesotheliomas arise from the peritoneum, but most of the remainder arise from the pleura. All are closely associated with asbestos exposure. CT demonstrates nodular, irregular peritoneal and omental thickening and masses, which merge to form large plaques and cakelike thickening of the omentum (“omental cake).― Adjacent bowel may be invaded and become fixed. US demonstrates the sheetlike
1872
superficial masses. Rare multilocular cystic forms of the tumor also occur. Prognosis is poor, with most patients dying within 1 year of diagnosis. Peritoneal metastases are most commonly associated with ovarian, colon, stomach, or pancreas carcinoma. The preferential sites for tumor implantation are the pelvic cul-de-sac, right paracolic gutter, and the greater omentum. CT demonstrates tumor nodules on peritoneal surfaces (“omental cake;― Fig. 26.30), which displace bowel away from the anterior abdominal wall, tumor nodules in the mesentery, thickening and nodularity of the bowel wall owing to serosal implants, and ascites that is commonly loculated. US may directly visualize the peritoneal tumors and demonstrates secondary signs of malignant ascites, including echogenic debris in the fluid, septation, and matted bowel loops.
FIGURE
26.30. Peritoneal
Metastases. A CT scan
demonstrates intraperitoneal spread of ovarian carcinoma. The tumor is implanted on the omentum (arrows), causing the appearance of “omental cake― as the thickened omentum floats in ascites (A) between bowel loops and the abdominal wall. Nodules of tumor (arrowhead) are implanted on the peritoneal surface.
1873
Lymphangiomas are benign cystic lesions that arise from lymphatic vascular channels. The cystic mass contains septations and multiple loculations containing chylous, serous, hemorrhagic, or mixed fluid. Lesions occur in the omentum, mesentery, mesocolon, and retroperitoneum. CT shows a fluid-density mass with enhancing wall and septa. US shows better the multilocular nature of the mass. The fluid contains echogenic debris. MR shows low signal on T1WIs and high signal on T2WIs for serous lymphangiomas. Those complicated by infection or hemorrhage are high signal on T1WIs. Primary retroperitoneal neoplasms arise in the retroperitoneal tissues outside of the retroperitoneal organs. Many tumors grow quite large before discovery. Tumors displace and compress abdominal and pelvic organs. Well-defined tumors with homogeneous fat are lipomas. Heterogeneous tumors that contain areas of distinct fat density may be liposarcomas (Fig. 26.31), the most common sarcoma of the retroperitoneum, or teratomas. Other fat-containing mass lesions include adrenal myelolipoma, angiomyolipoma, omental infarction, and mesenteric panniculitis (2 0). Cystic tumors that enhance minimally are likely lymphangiomas. Other considerations include neurogenic tumors, such as schwannomas, neurofibromas, and ganglioneuromas; lymphoma; desmoid tumors; and malignant mesenchymomas. Retroperitoneal fibrosis is a rare condition manifested by formation of a fibrous plaque in the lower retroperitoneum that encases and compresses the aorta, inferior P.751 vena cava, and ureters. Two thirds of cases are idiopathic. Methysergide, an ergot prescribed for migraine headache, causes 12% of cases. Small foci of metastatic malignancy that elicit a fibrotic reaction in the retroperitoneum account for another 8% to 10% of cases. Inflammatory aneurysms, which induce a rind of perianeurysmal fibrosis, are responsible for 5% to 10% of cases. Other possible causes include tuberculosis, syphilis, actinomycosis, and fungi. About 15% of patients have additional fibrosing processes, including mediastinal fibrosis, Riedel fibrosing thyroiditis, sclerosing
1874
cholangitis, and fibrotic orbital pseudotumors. The fibrotic plaque is usually located over the anterior surfaces of the L4 and L5 vertebrae. In the early stages, the plaque is highly cellular and edematous; when mature, it consists of dense hyalinized collagen with few cells. Cases induced by malignancy have a few malignant cells scattered within the collagen.
FIGURE
26.31. Liposarcoma. CT shows a large liposarcoma
(arrows) that arose in the retroperitoneum as a mottled fatdensity mass that distorts the inferior vena cava (IVC), surrounds the aorta (Ao), and displaces small and large bowel (B) laterally.
The hallmark of retroperitoneal fibrosis on excretory urography is smooth extrinsic narrowing of one or both ureters in the region of L4–L5. Proximal hydronephrosis results from impairment of ureteral peristalsis. The process may extend into the pelvis and cause a teardrop configuration to the bladder and narrowing of the sigmoid colon. CT demonstrates a fibrous plaque that envelops the inferior vena cava, aorta, and often the ureters. The plaque may be midline or asymmetric, well-defined or poorly defined, localized or
1875
expansive. On MR the plaque is typically of low signal intensity on both T1WIs and T2WIs. Plaque that shows high signal intensity on T2WIs should be considered suspicious for malignancy as a cause, although early edematous plaques may have the same appearance. On US, retroperitoneal fibrosis is easily confused with lymphoma in the retroperitoneum. Both appear as confluent hypoechoic masses encasing the IVC and aorta. Typically, lymphoma extends behind the vessels and displaces them anteriorly, but retroperitoneal fibrosis does not. Foreign bodies may be ingested or inserted, enter the abdomen or pelvis as a result of penetrating trauma, or be left behind at surgery. Recognition is important to avoid complications, which include hemorrhage, abscess formation, septicemia, bowel perforation or obstruction, or embolization. Many orally ingested foreign bodies are radiopaque, such as coins, pins, parts of toys, etc. Most will pass through the GI tract, causing only minimal mucosal damage. Large or elongated pointed objects may impinge at flexures or narrowed areas of the GI tract, such as the pylorus, duodenojejunal junction, ileocecal valve, or appendix. Button-sized batteries, such as those used in watches and hearing aids, contain highly toxic substances that can erode or perforate the bowel and can cause heavy metal poisoning if the battery ruptures. These should be followed to ensure they pass entirely through the bowel. Endoscopic or surgical removal should be considered if they fail to progress. Objects inserted into the vagina, rectum, or urethra can be removed manually or endoscopically. Retained bullets and shotgun pellets may lead to abscess formation or lead intoxication. CT is utilized to determine their exact position, complications, and the difficulty of removal. Wooden foreign bodies are usually not visualized on radiographs. CT shows high attenuation of the wooden object. US demonstrates high echogenicity with acoustic shadowing. MR shows wood to have variable intensity—usually less than that of skeletal muscle on T1WIs and T2WIs. Retained surgical sponges are a rare but dreaded complication of surgery. Retained sponges may be asymptomatic, cause an abscess, or generate a granulomatous response, inducing fibrosis and calcification. Sponges are usually detectable because of
1876
an incorporated tapelike or stringlike radiopaque marker (Fig. 26.32). CT shows a mass of soft tissue density, frequently containing air bubbles. Radiologists should be familiar with an ever-expanding number of medical devices that appear in images of the abdomen and pelvis, including intestinal tubes, postoperative apparatus, genitourinary devices, and monitoring instruments and attachments (2 1) . Abscesses occur within the peritoneal cavity because of spillage of contaminated material from perforated bowel or as a complication of surgery, trauma, pancreatitis, sepsis, or AIDS. Development of an abscess is commonly insidious, and the clinical presentation is often nonspecific and confusing. The pelvis is the most common site for abscess formation. Radiograph findings include soft tissue mass, collection of extraluminal gas, viscus displacement, localized or generalized ileus, elevation of the diaphragm, pleural effusion, and pulmonary basilar changes. A focal collection of extraluminal gas is the most specific sign of abscess but is unfortunately uncommon. CT shows a loculated fluid collection, often with internal debris and fluid–fluid levels. The walls of the fluid collection are often thick and irregular. Gas within the fluid collection is strong evidence of abscess (Fig. 26.33). The fascia adjacent to the abscess is thickened, and fat surrounding the abscess may be increased in density and contain soft tissue strands because of inflammation. US demonstrates a focal fluid collection that often contains echogenic fluid, floating debris, and septations. However, completely anechoic fluid collections may also be infected. A thickened wall is usually evident. Gas within the fluid collection is evidenced by echogenic foci that produce comet-tail or reverberation artifacts. CT-directed or USdirected aspiration confirms the diagnosis, provides material for culture, and offers the opportunity for percutaneous catheter drainage.
AIDS IN THE ABDOMEN Although advances in antiretroviral treatment have allowed many HIV-infected individuals to live productive lives, AIDS remains a
1877
worldwide
epidemic,
with P.752
20 million dead and 40 million infected. Each year almost 5 million people become newly infected with HIV, and nearly 4 million die of AIDS. In the United States, approximately 40,000 people are newly infected with HIV each year and approximately 1 million people are living with HIV.
FIGURE
26.32. Retained Surgical Sponge. A. Digital
radiograph of the abdomen taken at bedside reveals the characteristic radiopaque tape (arrow) that marks a surgical sponge inadvertently left within the abdominal cavity. Metallic cutaneous staples (arrowheads) identify the patient as having had recent surgery. B . CT reveals the difficulty of identifying the surgical sponge if the radiopaque marker (straight arrow) had not been present. The sponge (arrowheads) contains fluid, blood, and air bubbles, producing a pattern very similar to stool in the colon. The descending colon (curved
1878
arrow) is displaced medially.
FIGURE 26.33. Abscess. CT reveals an abscess (arrows) in the retroperitoneum. The abscess contains fluid and gas (arrowhead). Note the discrete enhancing wall of the abscess. Duodenum (D) containing intraluminal gas is displaced anteriorly and is draped over the collection.
AIDS in the abdomen is characterized by multiple coexisting diseases with multicentric involvement. Up to 90% of patients with AIDS develop complaints related to the GI or hepatobiliary systems. Genitourinary tract disease affects 38% to 68% of AIDS patients. Manifestations of infectious and neoplastic processes in AIDS patients are effectively demonstrated by abdominal imaging techniques (Table 26.7) . AIDS is a disease of impaired cellular immunity caused by the retrovirus designated human immunodeficiency virus (HIV). The disease is characterized by multiple opportunistic infections and aggressive malignancies, most commonly Kaposi sarcoma and AIDSrelated lymphoma. Infection by multiple organisms at multiple sites
1879
is the rule. Primary infection with HIV causes only minor symptoms, which may resemble infectious mononucleosis or other viral syndromes, with fevers, myalgias, transient adenopathy and skin rash. This is the stage of active viral replication and dissemination. With development of the immune response, usually within 3 months, virus P.753 levels decrease dramatically and the patient enters a clinically “silent― period, which often lasts many years. However, the CD4 receptor–coated T lymphocytes, which are primarily responsible for cell-mediated immunity, gradually but progressively decrease in number in the peripheral blood. A CD4 count below 200 cells/mm3 (normal is 800 to 1,000 cells/mm3 ) is diagnostic of AIDS.
TABLE 26.7 Abdominal Imaging Findings in AIDS
Adenopathy
Persistent generalized lymphadenopathy (reactive lymphoid hyperplasia)—mild retroperitoneal adenopathy (nodes 1.5 cm; suggests ARL, KS, MTB, MAI
Liver
Hepatitis/cirrhosis caused by HBV and HCV—especially in intravenous
drug
abusers
Hepatomegaly without focal lesions caused by HCV, MAI, histoplasmosis
1880
Hepatomegaly with focal lesions caused by bacillary angiomatosis or ARL
Masses >5 cm caused by ARL, KS, or amebic abscess
Masses (2–4 cm) caused by ARL, hepatocellular carcinoma, metastatic disease
Microabscesses (2 cm caused by PJ, MTB, ARL
Focal lesions 1.5 cm) in retroperitoneum and mesentery
G I tract wall thickening, nodules, plaques, polypoid lesions, thickened folds
Focal lesions in liver and spleen
1883
AIDS-related
lymphoma
Bulky adenopathy (>1.5 cm): mesentery, para-aortic, pelvic
Hepatosplenomegaly
Focal lesions in liver, spleen, kidney
Focal masses/wall thickening in G I tract, especially rectum and perianal area
Mycobacterium
avium-intracellulare infection
Bulky adenopathy (>1.5 cm): retroperitoneal + mesenteric
Hepatosplenomegaly
Rare focal lesions in liver + spleen
Pneumocystis
jiroveci infection
Focal lesions in liver + spleen
Diffuse or punctate calcification in liver, spleen, kidney, adrenal glands, lymph nodes
ARL, AIDS-related lymphoma; CMV, cytomegalovirus; HBV, hepatitis B virus; HCV, hepatitis C virus; HSV, herpes simplex virus; KS, Kaposi sarcoma; MAI, Mycobacterium aviumintracellulare; MTB, Mycobacterium tuberculosis; OI, opportunistic infections; PJ, Pneumocystis jiroveci.
1884
Clinical immunodeficiency presents with signs and symptoms of impaired cellular immunity. Cellular immunity is primarily responsible for the body's defense against mycobacteria, fungi, parasites, certain viruses, and tumors. Because drug therapy has become more effective, AIDS patients live longer, develop a broader spectrum of opportunistic infections, and have a greater risk of developing AIDSrelated tumors. Malignancy, rather than infection, is now a major cause of death. Although AIDS-related diseases may affect all body systems and involve every region of the body, abdominal disease is increasingly common. Abdominal imaging studies are used to document the presence and severity of complications and, in some cases, to suggest the specific cause. Patients with abdominal disease and AIDS may present with dysphagia, abdominal pain, diarrhea, fever, or progressive weight loss with muscle wasting. Barium studies are being used with P.754 declining frequency because of the lack of specificity of most findings and the expanding clinical practice of treating AIDS symptomatically without identification of a specific GI pathogen. US and CT are the most useful modalities for evaluating the solid visceral organs, adenopathy, and the peritoneal cavity. MR has made no significant contribution to evaluating the abdomen in AIDS patients. CD4 cell counts are correlated with a high risk of specific pathogens in AIDS patients. Thrush and tuberculosis are seen most commonly in patients with CD4 counts between 250 and 500 cells/mm3 . Pneumocystis jiroveci (formerly P carinii) pneumonia usually presents in patients with CD4 counts lower than 200 cells/mm3 . Kaposi sarcoma and AIDS-related lymphoma appear with counts in the range of 150 to 200 cells/mm3 . Counts of 75 to 125 cells/mm3 are associated with esophageal candidiasis and infection with herpes simplex virus, toxoplasmosis, and cryptococcosis. Patients with CD4 counts below 50 cells/mm3 have a median survival time of less than 1 year. Opportunistic
infections are caused by organisms that are usually
effectively controlled by normal cellular immunity. P jiroveci causes pneumonia in nearly 80% of AIDS patients. In patients treated with
1885
prophylactic aerosolized pentamidine, extrapulmonary P jiroveci infection is common, affecting the liver, spleen, kidney, pancreas, and lymph nodes. Mycobacterium avium-intracellulare and M tuberculosis are also frequent infections. M avium-intracellulare is a cause of bulky abdominal adenopathy, hepatosplenomegaly, and focal lesions in the liver and spleen. Candida albicans and cytomegalovirus are common causes of esophagitis as well as gastric antritis and duodenitis. Cryptosporidium and Isospora belli are protozoans, previously found only in animals, that infect the GI tract and cause severe diarrhea. Cryptosporidium and cytomegalovirus are implicated as causes of AIDS-related cholangitis. Herpesvirus, Toxoplasma gondii, Entamoeba histolytica, Giardia lamblia, and Cryptococcus neoformans are additional pathogens in AIDS patients. Kaposi
sarcoma (KS) is the most common malignancy associated
with AIDS and is second only to Pneumocystis pneumonia as the most common AIDS-defining illness. The tumor is always multicentric and arises from lymphatic epithelium found in all organs and tissues. The typical lesion is a vascular nodule on the skin or mucous membranes, in the GI tract, or in any solid visceral organ. AIDSassociated KS is divided into two clinical subtypes. Classic KS is a limited form of KS, with lesions mostly confined to the face, extremities, and oral mucosa; it may convert into epidemic KS at any time. Epidemic KS is disseminated and aggressive and requires therapy. Lesions are found in lymph nodes, visceral organs, GI tract, and bone marrow. Most patients with internal involvement have multiple lesions on the skin. KS affects the GI tract in 40% to 50% of cases, causing nodules, plaques, polypoid lesions, and thickened folds. Bulky adenopathy is sometimes present in the retroperitoneum and mesentery. Brightly enhancing lymphadenopathy is particularly suggestive of KS (79% positive predictive value). In the liver, KS lesions appear as hyperechoic nodules on US and as uniformenhancing or ring-enhancing lesions on CT. AIDS-related KS is now believed to be caused by a herpes-type virus that is transmitted primarily by anal intercourse. KS is most common (90% to 95% of cases) in homosexual and bisexual men and is uncommon in women and heterosexual men. The diagnosis should be confirmed with
1886
biopsy to avoid misdiagnosis. AIDS-related lymphomas are extremely aggressive neoplasms that respond poorly to therapy and commonly involve extranodal sites. Median survival is only 5 to 6 months. Extranodal involvement is found at presentation in 73% to 86% of patients, with the most common locations being the CNS (27%), bone marrow (22%), GI tract (17% to 54%), liver (12% to 29%), kidney (11%), and spleen (7%). Focal hepatic lesions are hypodense on contrast CT and vary from innumerable small lesions ( Table of Contents > Section VII - Gastrointestinal Tract > Chapter 27 - Liver, Biliary Tree, and Gallbladder
Chapter
27
Liver, Biliary Tree, and Gallbladder William E. Brant
LIVER Imaging
Methods
CT, MR, and US all produce high-quality images of the liver parenchyma. Dynamic bolus contrast–enhanced multidetector CT (MDCT) is the current method of choice for most hepatic imaging. Fast imaging techniques that control motion have increased the role of MR as a problem solver and often as the primary hepatic imaging modality. MR is preferred whenever iodinated contrast cannot be used. US is used as a screening method for patients with abdominal symptoms and suspected diffuse or focal liver disease. Color flow and spectral Doppler are used to assess hepatic vessels and tumor vascularity. Radionuclide imaging is used in the characterization of cavernous hemangiomas and focal nodular hyperplasia. MDCT of the liver is performed using a three- or four-phase protocol of multiple scans of the entire liver. Initial noncontrast images are followed by rapid bolus IV contrast injection by a mechanical injector. Immediate images are optimally obtained during the peak arterial enhancement phase to detect hypervascular tumors and other lesions supplied primarily by the hepatic artery. Arterial-phase-
1890
enhancing lesions, like hepatocellular carcinoma (HCC), are high attenuation on a background of lower-attenuation, minimally enhanced parenchyma. Maximum enhancement of the liver is attained during the portal venous phase to demonstrate hypovascular lesions as low-attenuating masses on a background of brightly enhanced parenchyma. Delayed images are obtained several minutes after contrast injection to document late contrast fill-in of hemangioma and delayed enhancement of cholangiocarcinoma. Hepatic MR is performed with a broad array of fast spin-echo, breath-hold gradient recall, short-time inversion-recovery, fatsuppressed, or in-phase/out-of-phase pulse sequences. The goal is to maximize lesion detection by the striking contrast resolution of MR while minimizing motion artifact by rapid-scan breath-hold sequences. Dynamic contrast enhancement is achieved with MR by repeating full liver scans multiple times in the first minutes following gadolinium injection. US is used as a rapid screening modality to detect diseases of the gallbladder, biliary tree, and liver. Hepatic US imaging is reviewed in Chapter 36. Radionuclide imaging of the liver is inferior to CT and MR for lesion detection but offers functional information in characterizing lesions, such as focal nodular hyperplasia. Radionuclide blood pool imaging is very useful for definitive diagnosis of cavernous hemangioma. Hepatic radionuclide imaging is reviewed in Section XII. Fine-needle aspiration for cytology and core needle biopsy for histology, guided by US or CT, are popular and safe methods to obtain tissue diagnoses.
Anatomy Couinaud
Segments
The vascular anatomy that defines the surgical approach to lesion resection is the anatomy most relevant to liver imaging. A numbering system
1891
P.757 developed by Couinaud (pronounced “kwee-NO―) is commonly used internationally and provides standardized identification of hepatic segments (Fig. 27.1) (Table 27.1). The eight Couinaud segments have separate vascular inflow, outflow, and biliary drainage and can each be resected without damaging the remaining segments. Division of the liver into eight segments is based on a concept of three longitudinal planes and two transverse planes. A longitudinal plane through the middle hepatic vein, inferior vena cava (IVC), and gallbladder fossa divides the liver into right and left lobes. A longitudinal plane through the right hepatic vein divides the right lobe into anterior (VIII and V) and posterior (VII and VI) segments. A longitudinal plane through the left hepatic vein divides the left lobe into medial (IVa and IVb) and lateral (II and III) segments. A transverse plane through the left portal vein divides the left lobe into superior (IVa and II) and inferior (IVb and III) segments. An oblique transverse plane through the right portal vein divides the right lobe into superior (VIII and VII) and inferior (V and VI) segments. Segment I is the caudate lobe, which extends between the fissure of the ligamentum venosum and the IVC. Hepatic venous drainage from the caudate lobe is directly into the IVC via small veins.
1892
FIGURE 27.1. Couinaud Liver Segments. A. Superior portion of the liver. B . Inferior portion of the liver. CT scans illustrate the Couinaud classification of numbering of liver segments. The longitudinal plane of the right hepatic vein divides VIII from VII in the superior portion of the liver and V from VI in the inferior portion of the liver. The longitudinal plane of the middle hepatic vein through the gallbladder fossa separates IVa from VIII in the superior liver and IVb from V in the inferior liver. The longitudinal plane of the left hepatic vein and fissure of the ligamentum teres separates IVa from II in the superior liver and IVb from III in the inferior liver. The axial plane of the left portal vein separates IVa superiorly from IVb inferiorly and II superiorly from III inferiorly in the left lobe. The axial plane of the right portal vein separates VIII and VII superiorly from V and VI inferiorly in the right lobe. The caudate lobe (segment I) extends between the fissure of the ligamentum venosum anteriorly and the inferior vena cava posteriorly.
Blood
supply to the liver is approximately two thirds via the portal
vein and one third via the hepatic artery. P.758 When IV contrast is administered as a bolus during rapid CT scanning, the maximum liver parenchymal enhancement will be delayed 1 to 2 minutes following initiation of injection. This delay reflects the transit time of contrast agent through the GI tract and spleen before accessing the liver through the portal vein. Tumors, which are supplied primarily by the hepatic artery, will enhance maximally during the early hepatic arterial phase, while the liver parenchyma enhances maximally during the portal venous phase.
TABLE 27.1 American and International Nomenclature for Anatomic Segments of the Liver
1893
American Caudate
lobe
International Caudate
lobe
Number I
Left lobe
Lateral segment
Medial
segment
Left lateral superior subsegment
II
Left
inferior
III
subsegment
IVa
lateral
subsegment
Left
medial
IVb
Right lobe
Anterior segment
Posterior segment
Right
anterior
inferior
V
superior
VIII
subsegment
Right
anterior
subsegment
Right posterior subsegment
inferior
VI
Right posterior subsegment
superior
VII
Adapted from Dodd GD. An American's guide to Couinaud's numbering system. AJR Am J Roentgenol 1993;161:574-575.
1894
Perfusion
abnormalities are seen on post–IV contrast CT and MR
because of variations in hepatic arterial and portal venous blood supply to various areas of the liver (1,2). This dual blood supply has a compensatory relationship: arterial flow increases when portal venous flow decreases. Transient enhancement differences are seen during either arterial phase imaging or portal venous phase imaging on MDCT and dynamic MR. Portal venous flow may be altered by: (1) portal blockade by tumor or thrombus; (2) extrinsic compression caused by ribs or diaphragmatic slips, or by tumors on the liver capsule; or (3) “third inflow― from systemic veins in the pericholecystic, parabiliary, and epigastric-paraumbilical venous systems (Fig. 27.2). Hepatic arterial flow may be increased by: (1) focal hypervascular lesions; (2) inflammation of adjacent organs (cholecystitis, pancreatitis); or (3) aberrant hepatic arterial supply. Regional differences in blood supply related to these factors explain patterns of enhancement abnormalities as well as altered patterns of diffuse liver disease, such as focal fatty deposition and focal fatty sparing in diffuse fatty infiltration.
1895
FIGURE 27.2. Perfusion Defect. A common perfusion defect (white arrow) is seen in segment IVb adjacent to the fissure of the ligamentum teres (black arrowhead). This perfusion defect is related to “third inflow― from paraumbilical systemic veins. Focal fatty infiltration is commonly seen in this location. Importantly, this normal variant must not be mistaken for a neoplasm.
On CT, the density of normal liver parenchyma is equal to or greater than the density of normal spleen parenchyma on unenhanced images. Following bolus IV contrast administration, the normal parenchymal enhancement is less than that of the spleen during the arterial phase and equal to or greater than that of the spleen during the portal venous phase. On T1WI MR, the normal liver is of slightly higher signal intensity than the spleen, and most focal lesions appear as lower-intensity defects. On T2WIs, the normal liver is less than or equal to the spleen in signal strength, and most lesions appear as high-intensity foci.
Diffuse
Liver
Disease
Hepatomegaly Enlargement of the liver is usually judged subjectively, based on imaging studies. Rounding of the inferior border of the liver and extension of the right lobe of the liver inferior to the lower pole of the right kidney are evidence of hepatomegaly. A liver length of greater than 15.5 cm, measured in the midclavicular line, is considered enlarged. A Reidel lobe is a normal variant of hepatic shape found most often in women. It refers to an elongated inferior tip of the right lobe of the liver. When a Reidel lobe is present, the left lobe of the liver is correspondingly smaller in size. The left lobe of the liver may, as a normal variant, be elongated and surround a portion of the spleen. Causes of hepatomegaly are listed in Table
1896
27.2. Fatty infiltration is a common and nonspecific response of hepatocytes to injury and toxins. Hepatocytes become filled with cholesterol and triglycerides. Causes include alcoholism, obesity, malnutrition, hyperalimentation, steroid therapy, diabetes mellitus, pancreatitis, glycogen storage disease, and chemotherapy. Imaging is the best diagnostic method to document the condition, since laboratory evaluation may be normal. On CT, fat infiltration lowers the attenuation of the hepatic parenchyma and makes the liver appear less dense than the spleen (Fig. 27.3). Differences in density between liver and spleen are most reliably judged on noncontrast images. On postcontrast images, the spleen enhances maximally 1 to 2 minutes before maximal liver enhancement and is thus transiently brighter than the normal liver. Fatty-infiltrated livers enhance less than normal livers. On US, the liver parenchyma is increased in echogenicity in areas of fat infiltration. The echogenicity of the fatty liver is significantly greater than the echogenicity of the normal kidney parenchyma. This density difference (fat is bright on US and dark on CT) has been called the “flip-flop sign.― Conventional spin-echo MR images show no significant
abnormalities
with
fat
infiltration.
Gradient-echo
imaging
with fat and water molecules in-phase and out-of-phase P.759 is the MR method most sensitive to fatty infiltration. On in-phase images, the signal from water and fat molecules are additive. On out-of-phase images, the signals from water and fat cancel out each other. A loss of signal intensity between in-phase and out-of-phase images is indicative of fatty infiltration (Fig. 27.4). This is the same technique used to characterize benign adrenal adenomas (see Chapter 33) .
1897
TABLE 27.2 Causes of Hepatomegaly
Vascular
Congestion
Congestive heart failure Hepatic vein thrombosis Metabolic/diffuse infiltration Fatty infiltration Alcohol Drugs/chemotherapy Hepatic toxins Gaucher disease Carbohydrate Glycogen
and
storage
lipidoses
diseases
Diabetes mellitus Iron Hemochromatosis Amyloid Amyloidosis Tumor/cellular Diffuse
infiltrate
metastases
Diffuse hepatocellular carcinoma Lymphoma Extramedullary hematopoiesis Systemic Cysts Polycystic
mastocytosis disease
Inflammation/infection Hepatitis Sarcoidosis Tuberculosis Malaria
1898
FIGURE
27.3. Diffuse Fatty Infiltration Seen on CT. CT
reveals the density of the enhanced liver parenchyma (L) to be significantly less than the density of the enhanced splenic parenchyma (S). Portal (p) and hepatic (h) veins run their normal courses without displacement or distortion. V, inferior vena cava; Ao, aorta.
Characteristic features of fatty infiltration include lack of mass effect (no bulging of the liver contour or displacement of intrahepatic blood vessels) and angulated geometric boundaries between involved and uninvolved parenchyma. Areas of fat infiltration may be multifocal, with interdigitating fingers of normal and abnormal parenchyma. Fatty changes can develop within 3 weeks of hepatocyte insult and may resolve within 6 days of P.760 removing the insult. Patterns of fatty infiltration are strongly related to hepatic blood flow.
1899
FIGURE
27.4. Diffuse Fatty Infiltration Seen on MR. A. In-
phase gradient-recall MR. B . Out-of-phase gradient recall MR. The out-of-phase image shows distinct loss of signal (darkening) of the entire liver parenchyma compared to the in-phase image. The out-of-phase MR image is easily recognized by the black line surrounding soft tissue structures at the interface with abdominal fat.
Diffuse
fatty
infiltration involving the entire liver is the most common
pattern. However, the degree of fat infiltration is commonly not uniform throughout the liver. Focal
fatty
infiltration involves a geographic or fan-shaped portion of
the liver with the same imaging features as diffuse infiltration. Vessels run their normal course through the area of involvement. Focal fatty infiltration may simulate a liver tumor; however, the area of involvement has a density characteristic of fat. Focal fatty infiltration is most common adjacent to the falciform ligament, gallbladder, and porta hepatis. Focal sparing in a diffusely fatty-infiltrated liver may be the most confusing pattern, because spared areas of normal parenchyma may simulate a liver tumor (Fig. 27.5). The fat-spared area commonly located in segment IV. The fat-spared area relative to the rest of the liver on US and is of higher the rest of the liver on CT. The remainder of the liver
1900
is most is hypoechoic density than demonstrates
features
characteristic
of
diffuse
fatty
infiltration.
Nonalcoholic steatohepatitis (NASH) describes fatty liver caused by an inflammatory response that is not caused by excessive alcohol intake. NASH is related to obesity, type 2 diabetes, hyperlipidemia, and anorexia nervosa. It is a rare cause of acute fulminant hepatic failure and may progress to cirrhosis. Histology shows steatosis with parenchymal inflammation and fibrosis. Imaging features are those of diffuse fatty liver.
FIGURE 27.5. Fatty Infiltration With Focal Sparing. A. US image demonstrates a focal hypoechoic area of normal liver (NL) near the portal vein (p) in a liver (FL) that is diffusely increased in echogenicity because of fatty infiltration. B . CT image obtained without contrast enhancement demonstrates the spared area of normal liver (arrow) to be high density compared to the lower density of the fatty replaced liver (FL). Note the characteristic “flip-flop― appearance of fat density between CT and US. S, spleen.
Acute hepatitis most commonly causes no abnormalities on hepatic imaging. In some patients, diffuse edema lowers the parenchyma
1901
echogenicity and causes the portal venules to appear unusually bright on US (3). In acute fulminant hepatitis, areas of necrosis show as ill-defined areas of low density on CT. Chronic hepatitis is characterized pathologically by portal and perilobular inflammation and fibrosis. Imaging studies are insensitive to early pathologic changes. Fatty changes are minimal and the liver is usually not enlarged. Perihepatic lymph nodes are commonly visualized. US may show a subtle coarse increase in hepatic echogenicity. The primary role of imaging patients with chronic hepatitis is to detect hepatocellular carcinoma. Cirrhosis is characterized pathologically by diffuse parenchymal destruction, fibrosis with alteration of hepatic architecture, and innumerable regenerative nodules that replace normal liver parenchyma. Causes of cirrhosis include hepatic toxins (alcohol, drugs); infection (viral hepatitis, especially types B and C); biliary obstruction; and heredity (Wilson disease). In the United States, 75% of patients with cirrhosis are chronic alcoholics. In Asia P.761 and Africa, most cases of cirrhosis are caused by chronic active hepatitis. A variety of morphologic alterations are seen on imaging studies (Fig. 27.6) (4,5). These include: (1) hepatomegaly (early), (2) hepatic atrophy (late), (3) coarsening of hepatic parenchymal texture, (4) irregularity (nodularity) of the liver surface, (5) hypertrophy of the caudate lobe with shrinkage of the right lobe, and (6) regenerating nodules. Extrahepatic signs of cirrhosis include evidence of portal hypertension, splenomegaly, and ascites. The pathologic changes of cirrhosis are irreversible, but disease progression can be limited or stopped by eliminating the causative agent (stop drinking alcohol). Transjugular intrahepatic portosystemic shunts (TIPS) are effective treatment for portal hypertension and long-term control of esophageal variceal bleeding. Liver transplantation is now established as effective treatment for end-stage liver disease.
1902
FIGURE
27.6. Cirrhosis
and
Portal
Hypertension. A CT scan
reveals atrophy of the liver (L) with diffuse nodularity of its surface (open arrows) and splenomegaly (S). Numerous enhancing portosystemic collateral vessels are evident, including perihepatic
(long black arrows), gastrohepatic (arrowheads), and
gastric varices (short arrow). A dilated periumbilical vein (white arrow) is seen coursing out of the fissure of the ligamentum teres into the falciform ligament. Ascites (a) is also evident. St, stomach; V, inferior vena cava; A, aorta.
US demonstrates heterogeneous parenchyma with coarsening of the echotexture and decreased visualization of small portal triad structures. High-frequency detailed scanning of the liver surface reveals fine nodules. CT may be normal in the early stages, or it may reveal parenchymal inhomogeneity with patchy areas of increased and decreased attenuation (6). Fine or coarse nodularity of the liver surface is characteristic. Areas of fatty replacement may be evident. MR shows heterogeneous parenchymal signal on T1WIs and T2WIs. High-signal fibrosis on T2WIs is the predominant cause of the heterogeneous
appearance.
1903
TABLE 27.3 Causes of Nodules in a Cirrhotic Liver
Regenerative nodules (nodules 10 mm) Hepatocellular carcinoma Confluent fibrosis Focal fat infiltration Focal fat sparing Metastases
Nodules
in
Cirrhosis
Nodules are a constant feature of cirrhosis (Table 27.3), and the challenge is to differentiate the ubiquitous benign nodule from HCC (7). The most common nodules are regenerative nodules (Fig. 27.7) , which are a routine pathologic feature of cirrhosis caused by the body's attempted repair of hepatocyte injury. Regenerative nodules are composed primarily of hepatocytes that are surrounded by coarse fibrous septations. They have the same imaging characteristics as hepatic parenchyma but stand out as nodules because of their surrounding fibrous bands. Regenerative nodules are 3 to 10 mm in size. Small regenerative nodules cause the micronodular pattern of alcoholic cirrhosis. On CT, most regenerative nodules are isointense with parenchyma and are not detected. Iron deposits in regenerating nodules (siderotic nodules) will cause a slight increase in CT attenuation, causing them to appear slightly hyperintense. Regenerative nodules may be hypointense, isointense, or hyperintense on T1WIs, but they do not show the early arterial enhancement post–gadolinium administration that is typical of HCC. On T2WIs, regenerative nodules are hypointense (siderotic nodules) or isointense but are not P.762 hyperintense,
providing
differentiation
1904
from
most
liver
metastases
and from dysplastic nodules. Dysplastic nodules (adenomatous hyperplastic nodules) are proliferative precancerous lesions found in 15% to 25% of cirrhotic livers. On CT and US, adenomatous hyperplastic nodules resemble regenerative nodules but are larger than 10 mm in diameter. On MR, non–iron-containing adenomatous hyperplastic nodules are hyperintense on T1WIs and hypointense to isointense on T2WIs. High signal on a T2WI is indicative of foci of cellular atypia or malignancy. Small HCCs are variable in signal on both T1WIs and T2WIs. The most characteristic MR appearance of small HCCs is hyperintense on T1WIs and moderately hypointense on T2WIs, with homogeneous early arterial phase enhancement with rapid washout postcontrast (see Fig. 27.17). Rapid growth is characteristic. Lesions >3 cm usually have a distinct capsule. On CT, small HCCs appear as low-density, encapsulated masses that enhance rapidly and quickly become hypodense with bolus contrast dynamic scanning. Ringlike enhancement of the tumor capsule is characteristic.
FIGURE 27.7. Regenerative Nodules in Cirrhosis. CT image filmed at a narrow window shows innumerable low-density small nodules evident throughout the liver in this patient with cirrhosis. Needle biopsy confirmed benign regenerative nodules.
1905
Confluent fibrosis describes the masslike areas of fibrosis found in livers with advanced cirrhosis. Extensive fibrosis produces wedgeshaped or radiating bandlike masses that are hypodense on noncontrast CT and may become hyperdense post–contrast administration. Volume loss of the affected portion of the liver is a key feature. A whole hepatic segment or lobe may be replaced by fibrosis. The areas of fibrosis are hypointense to liver parenchyma on T1WIs and hyperintense on T2WIs. As mentioned previously, focal fat infiltration or focal fat sparing may also produce nodules in the cirrhotic liver. Metastases, especially those from breast carcinoma, may mimic the appearance of cirrhosis, with regenerative nodules. Following chemotherapy, tissue retraction of necrosing metastases with surrounding scarring may produce a condition termed pseudocirrhosis (6) . Portal hypertension is a pathologic increase in portal venous pressure that results in the formation of portosystemic collateral vessels that divert blood flow away from the liver and into the systemic circulation. Causes of portal hypertension include progressive vascular fibrosis associated with chronic liver disease, portal vein thrombosis or compression, and parasitic infections (schistosomiasis). Portal hypertension carries the risk of hemorrhage from varices and hepatic encephalopathy. The signs of portal hypertension
include:
(1)
visualization
of
portosystemic
collaterals
(coronary, gastroesophageal, splenorenal, paraumbilical, hemorrhoidal, and retroperitoneal) (Fig. 27.6); (2) increased portal vein diameter (>13 mm); (3) increased superior mesenteric and splenic vein diameters (>10 mm); (4) portal vein thrombosis; (5) calcifications in the portal and mesenteric veins; (6) edema in the mesentery, omentum, and retroperitoneum; (7) splenomegaly owing to vascular congestion; and (8) ascites (8) .
1906
FIGURE
27.8. Portal
Hepatocellular
Vein
Carcinoma.
Thrombosis:
Multinodular
Contrast-enhanced
CT
demonstrates multiple hypodense nodules representing hepatocellular carcinoma that is replacing the right hepatic lobe. The portal vein (pv) is invaded by tumor (arrow), seen as a filling defect with the vein. The hepatic artery (arrowhead) is enlarged because of cirrhosis and portal hypertension.
Portal vein thrombosis may occur as a complication of cirrhosis, or it may be caused by portal vein invasion or compression by tumor (Fig. 27.8), hypercoagulable states, or inflammation (pancreatitis). The cause is unknown in 8% to 15% of patients. On CT and US, the thrombus is seen as a hypodense plug within the portal vein. On MR, the thrombus is hyperintense on T1WIs when acute and isointense when chronic. Signal in the thrombus is increased on T2WIs. Portal hypertension is exacerbated, or may be caused, by portal vein
1907
thrombosis. Cavernous transformation of the portal vein develops when small collateral veins adjacent to the portal vein expand and replace the obliterated portal vein. These collateral veins appear as a tangle of small vessels surrounding the thrombosed portal vein (8) . Budd-Chiari syndrome is caused by obstruction to hepatic venous outflow that may be a result of obstruction of the suprahepatic IVC by a congenital membranous web (primary type) or by thrombosis of the hepatic veins caused by tumor, hypercoagulable states, or trauma (secondary type). Blood flow to the right and left hepatic lobes is severely impaired, resulting in a characteristic “flipflop― pattern on contrast-enhanced CT. On early images, the central liver enhances prominently, whereas the peripheral liver enhances weakly (Fig. 27.9). On delayed images, the periphery of the liver is enhanced, whereas contrast has P.763 washed out of the central liver. The caudate lobe is spared because of its separate venous drainage to the IVC. The caudate lobe is characteristically enlarged and enhances normally. Thrombi may be seen in the hepatic veins, or they may be reduced in caliber and difficult to visualize. Comma-shaped intrahepatic collateral vessels may be seen on CT or MR (the “comma sign―). Multiple benign hepatic nodules up to 3 cm commonly develop. Most are detected by prominent contrast enhancement during the arterial phase or mild contrast enhancement during the portal venous phase (9) .
1908
FIGURE
27.9. Budd-Chiari
Syndrome. Early phase CT images
show the markedly heterogeneous liver with prominent central and weak peripheral enhancement that is characteristic of BuddChiari syndrome. Tumor invasion from a right adrenal carcinoma is seen as a filling defect (arrow) within the inferior vena cava.
Passive hepatic congestion is a common complication of congestive heart failure and constrictive pericarditis. Hepatic venous drainage is impaired, and the liver becomes engorged and swollen. Findings include distension of the hepatic veins and IVC, reflux of IV contrast into the hepatic veins and IVC, increased pulsatility of the portal vein, and inhomogeneous contrast enhancement of the liver. Secondary findings commonly present include hepatomegaly, cardiomegaly, pleural effusions, and ascites. Hemochromatosis may be primary (hereditary) or secondary owing to excessive iron intake from either parenteral or dietary sources. In severe cases, CT demonstrates a diffuse increase in liver density, to 75 to 130 H. MR is more sensitive to hepatic iron overload because of the superparamagnetic effect of iron. MR demonstrates marked diffuse low signal intensity on T2WIs (Fig. 27.10) and moderate loss of signal intensity on T1WIs (1 0). Long-standing hemochromatosis places the patient at risk for cirrhosis and hepatocellular carcinoma.
1909
Gas in the portal venous system is often an ominous sign associated with bowel ischemia in adults (Fig. 27.11) and necrotizing enterocolitis in infants. Additional causes include recent colonoscopy, enema administration, gastrostomy tube placement, abdominal trauma, inflammatory bowel disease, perforated gastric ulcer, necrotizing pancreatitis, diverticulitis, and abdominal abscess (8). CT reveals air in branching tubular structures that extends to the liver capsule. Air is commonly evident within P.764 the mesenteric and portal veins. Plain radiographs show streaks of low density in the periphery of the liver. In distinction, air in the biliary tree is more central, not extending to within 2 cm of the liver capsule.
FIGURE 27.10. Hemochromatosis. T2W MR image demonstrates marked low signal intensity in both liver and spleen. The low signal is caused by iron deposition in the reticuloendothelial system—in this case of secondary hemochromatosis caused by multiple blood transfusions.
1910
FIGURE
27.11. Portal Venous Gas. Noncontrast CT image
reveals gas in the portal vein, seen as air-density tubular structures extending to the periphery of the liver. Gas in the biliary tree is central and does not extend into the peripheral 2 cm of the liver. In this case, portal venous gas was associated with infarction of the small bowel.
Liver
Masses
A major challenge of liver imaging is to differentiate common and benign liver masses, such as cavernous hemangioma and simple hepatic cysts, from malignant masses such as metastases and hepatoma. US can definitively characterize hepatic cysts; however,
1911
benign and malignant solid masses overlap in sonographic appearance. CT can characterize most cysts and cavernous hemangiomas, but only with optimal technique and contrast administration. On MR, simple cysts and hemangiomas are hypointense on T1WIs and extremely hyperintense on T2WIs. These benign masses are typically homogeneous and have sharp outer margins. Malignant lesions on MR tend to be inhomogeneous with unsharp outer margins, peritumoral edema, and central necrosis. Most focal lesions are hypointense on T1WIs and hyperintense on T2WIs. Hyperintensity of focal lesions on T1WIs may be a result of the presence of fat, blood, proteinaceous material, or melanin in melanoma metastases (Table 27.4). Diffuse hypointensity of liver, caused by diffuse edema or iron overload, may make any lesion appear relatively hyperintense. Hypointensity on T2WIs is commonly a result of fibrosis (Table 27.5) .
TABLE 27.4 Causes of Hyperintensity in Focal Liver Lesions on MR T1WIs
Fat deposits Focal fat infiltration Fat deposition in tumor Hepatoma Lipoma Angiomyolipoma Hepatic adenoma Blood Hematoma Hemorrhage into tumor Proteinaceous material Proteinaceous fluid in cysts Necrosis/hemorrhage in tumor Abscess Hematoma Copper
1912
Intratumoral copper in Melanin Melanoma metastasis
hepatoma
Contrast enhancement Gadolinium administration Lipiodol administration Ghosting artifact Result of blood flow in adjacent vessels Hypointensity of liver parenchyma Edema caused by passive hepatic congestion Iron deposition in hepatocytes
TABLE 27.5 Causes of Hypointensity in Focal Liver Lesions on MR T2WIs
Fibrous capsule Hepatoma (24% to 42% of HCC) Hepatic adenoma Focal nodular hyperplasia Fibrous central scar Fibrolamellar Focal
HCC,
nodular
(rare)
HCC hyperplasia
hepatocellular
carcinoma.
Metastases are the most common malignant masses in the liver. Metastases are 20 times more common than primary liver malignancies. Of all patients who die of malignancy, 24% to 36% have liver involvement. Hepatic metastases most commonly originate from the GI tract, breast, and lung. A wide spectrum of appearance of metastatic disease is seen on all imaging studies (Fig. 27.12) (1 1) . Metastases may be uniformly solid, necrotic, cystic, or calcified; they
1913
may be avascular, hypovascular, or hypervascular; they are commonly irregular and poorly marginated but may be sharp and well defined. Most characteristic is bandlike peripheral enhancement on arterial phase images, with rapid washout on delayed CT and MR images. Metastatic disease must be considered in the differential P.765 of virtually all hepatic masses (Table 27.6). Multiplicity of lesions favors metastatic disease.
FIGURE 27.12. Metastases. Metastases from adenocarcinoma of the colon appear as numerous low-attenuation nodules of varying size on this portal venous phase, postcontrast CT. Note how the metastatic disease causes nodularity of the liver contour and the resemblance to regenerative nodules in cirrhosis (see Fig. 27.7) .
1914
TABLE 27.6 Causes of Multiple Small (10-mm) Lesions in the Liver
Regenerative nodules in cirrhosis Microabscesses (immunocompromised Multiple bacterial abscesses
patient)
Histoplasmosis Lymphoma Kaposi sarcoma (AIDS patient) Hepatocellular Sarcoidosis Gamna-Gandy
carcinoma bodies
(multinodular
(portal
form)
hypertension)
Metastases Breast
carcinoma
Lung carcinoma Ovarian carcinoma Gastric
carcinoma
Malignant melanoma Prostate carcinoma
Cavernous
hemangioma is second only to metastases as the most
common cause of a liver mass (1 2). It is the most common benign liver neoplasm, found in 7% to 20% of the population and more commonly in women. Up to 10% of patients have multiple lesions, which are easily mistaken for metastases. Many hemangiomas are discovered incidentally on hepatic imaging performed for other reasons. The tumor consists of large, thin-walled, blood-filled vascular spaces separated by fibrous septa. Blood flow through the maze of vascular spaces is extremely slow, resulting in characteristic imaging findings. Thrombosis within the vascular channels may result in central fibrosis and calcification. Most lesions are smaller than 5 cm, cause no symptoms, and are considered benign incidental findings. Larger lesions (>6 cm) occasionally cause symptoms by mass effect, hemorrhage, or arteriovenous shunting. The size of
1915
most cavernous hemangiomas is stable over time. Enlargement of a lesion is cause for reassessment. US demonstrates a well-defined, uniformly hyperechoic mass in 80% of patients. In a patient with no history of malignant disease and normal liver chemistries, follow-up only is generally recommended. No Doppler signal is obtained from most cavernous hemangiomas because the flow is too slow. CT generally shows a well-defined, hypodense mass on unenhanced scans. Because the lesion consists mostly of blood, attenuation of the hemangioma is similar to that of blood vessels within the liver. The characteristic pattern of enhancement with bolus IV contrast is nodular enhancement from the periphery of the lesion (Fig. 27.13) that gradually becomes isodense or hyperdense compared to the liver parenchyma. The degree of contrast enhancement parallels that of hepatic blood vessels during all postcontrast phases. The contrast enhancement persists for 20 to 30 minutes following injection because of slow flow within the lesion.
FIGURE 27.13. Cavernous Hemangioma. Images from a contrast-enhanced helical CT demonstrate the characteristic nodular pattern of enhancement from the periphery of the lesion.
1916
Radionuclide scanning using technetium-labeled red blood cells as a blood pool agent is extremely accurate in the diagnosis of cavernous hemangioma. Hemangiomas are characterized by prolonged, intense activity within the lesion on delayed images. MR demonstrates a well-defined homogeneous mass that is hypointense or isointense on T1WIs and brightens markedly with increasing amounts of T2 weighting. Areas of fibrosis remain dark on all image sequences. However, the MR appearance of cavernous hemangiomas overlaps that of cysts, abscesses, and hypervascular metastases. A specific diagnosis can be made by IV administration of gadolinium and observing early puddling peripheral enhancement in a pattern similar to that seen in contrast-enhanced CT. Biopsy may be required in atypical cases. Percutaneous biopsy can be safely performed using small needles (20-gauge and smaller). The characteristic finding is blood with normal epithelial cells and no malignant cells. Biopsy with large-bore needles has been associated with hemorrhage and death. Hepatocellular carcinoma is the most common primary malignancy of the liver. Risk factors include cirrhosis, chronic hepatitis, and a variety of carcinogens (sex hormones, aflatoxin, Thorotrast). In the United States, most HCCs are found in patients with cirrhosis (usually because of alcohol abuse). In Asia, most HCCs are found in patients with chronic active hepatitis. Hepatomas demonstrate three major growth patterns that affect their imaging appearance: diffuse infiltrative, solitary massive (Figs.
27.14, 27.15), and multinodular
(Fig. 27.8). Detection of the diffuse pattern of tumor is particularly difficult, P.766 especially when the liver parenchyma is already altered by diffuse hepatic disease. Approximately 24% of tumors are surrounded by a fibrous capsule. The encapsulated HCC, a variant of the solitary form, is found more frequently in Asian populations and has a better prognosis. Intratumoral hemorrhage and necrosis are common because of a lack of stroma within the tumor. Calcifications (punctate, stippled, or rimlike) occur in approximately 10% of cases.
1917
Most lesions are hypervascular and demonstrate contrast enhancement on arterial phase images, with diminishing enhancement on delayed phase images (see Fig. 27.17). Detection of hepatoma on a background of cirrhosis and regenerative nodules is a major imaging challenge. Elevation in serum α-fetoprotein is found in 90% of patients and is strongly suggestive of hepatoma in patients with cirrhosis. The tumor metastasizes to lungs, abdominal lymph nodes, adrenal glands, and bone (1 3) .
FIGURE
27.14. Hepatocellular Carcinoma on CT.
Three-phase
helical CT demonstrates the enhancement pattern of a larger hepatocellular carcinoma in the right lobe. The tumor is slightly hyperdense to parenchyma on the unenhanced scan (upper left) and shows intense enhancement on the early (arterial phase—upper right) scan and delayed (venous phase—lower) scan. The central low density is a result of necrosis. Note the satellite lesions (arrows) .
1918
FIGURE
27.15. Hepatocellular Carcinoma on MR.
Postcontrast T1WI shows the typical mosaic pattern of large hepatocellular carcinomas. Note the prominent enhancement in the periphery of the tumor.
TABLE 27.7 Fat-Containing Lesions in the Liver
Hepatic adenoma Hepatocellular carcinoma Focal fatty Lipoma Teratoma
deposition
Liposarcoma (primary or metastatic) Postoperative packing material (omentum) Focal
intrahepatic
extramedullary
hematopoiesis
Several imaging features are highly characteristic of HCC when
1919
present. Invasion of tumor into portal and hepatic veins occurs in up to 48% of cases (Fig. 27.8). Tumor thrombus is visualized as a lowdensity plug within an expanded vein. The intraluminal tumor enhances with contrast administration on CT and MR and may demonstrate arterial signal with Doppler US. Portal vein thrombus is more common than hepatic vein thrombus. A tumor capsule, when present, is visualized as a sharply marginated rim of tissue that enhances in 90% of cases. The capsule has low signal intensity on both T1WIs and T2WIs and is hypoechoic on US. Satellite nodules of tumor are common (Fig. 27.14). Fatty metamorphosis is a common histologic finding in HCC. CT may demonstrate a focal area of tumor with attenuation values of fat (Table 27.7) (1 4). MR confirmation of fat is performed with chemical shift imaging. A mosaic appearance of the tumor is considered to be characteristic but is found primarily in larger lesions (Fig. 27.15). The mosaic pattern appears as multiple nodular areas of differing CT attenuation. The pattern is more obvious with enhancement of septations on postcontrast scans. Arterioportal shunting is seen as early or prolonged enhancement of the portal vein or as a wedge-shaped area of parenchymal enhancement adjacent to the tumor. Abundant copper-binding protein in cancer cells may lead to excessive copper accumulation within the tumor. High copper concentration causes the tumor to appear hyperdense on noncontrast CT and hyperdense (because of T1 shortening effect) on T1WIs on MR. Focal
nodular
hyperplasia (FNH) forms a solid mass consisting of
abnormally arranged hepatocytes, bile ducts, and Kupffer cells. Most lesions are smaller than 5 cm in P.767 diameter and are hypervascular, with a central fibrous scar containing thick-walled blood vessels. Lesions are lobulated and well circumscribed but lack a capsule. These are benign lesions that do not require treatment but must be differentiated from hepatic adenoma and fibrolamellar carcinoma. In contrast to hepatic adenoma, hemorrhage, necrosis, and infarction are extremely rare in FNH. Similar to hepatic adenoma, FNH is found most commonly in women, but it is twice as common as hepatic
1920
adenoma and is not related to oral contraceptive use. Most tumors (80% to 95%) are solitary. Because of the presence of Kupffer cells, most (50% to 70%) FNH nodules will show normal or increased radionuclide activity on technetium sulfur colloid liver-spleen scans (1 5). This finding is highly indicative of FNH. On CT, MR, and US, most tumors appear homogeneous and solid. A central core scar with radiating septa is characteristic but present in only 60%. Because the cellular makeup of FNH is very similar to that of normal hepatic parenchyma, the lesion is usually isodense on noncontrast images (Fig. 27.16). The typical finding on contrastenhanced CT and MR is intense, brief (approximately 1 minute), uniform tumor enhancement during the arterial phase. FNH is inconspicuous on US, detectable only by mass effect (bulging contour) or slight alterations in parenchymal echotexture. Some lesions have a hypoechoic halo. The central scar, if present, is often poorly visualized. Color flow imaging may show the lesion's hypervascularity.
MDCT
shows
homogeneous
hyperenhancement
during the arterial phase. The lesion is often isodense during the portal phase, with enhancement of the central scar on delayed images. On MR, FNH appears homogeneous and isointense to slightly hypointense to normal parenchyma on T1WIs and isointense to slightly hyperintense on T2WIs (7). The central scar is hypointense on T1WIs and hyperintense on T2WIs. FNH shows a characteristic very intense homogeneous enhancement on arterial phase postcontrast images. The central scar and radiating septa enhance on delayed
postcontrast
images.
Hepatic adenomas are rare benign tumors that carry a risk of lifethreatening hemorrhage and potential for malignant degeneration. Surgical removal of the tumor is advocated. They are found most commonly in women on long-term oral contraceptives. Additional risk factors include androgen steroid intake and glycogen storage disease. The tumor consists of sheets and cords of benign hepatocytes without a distinct acinar architecture. The hepatocytes occasionally contain abundant fat, detectable by imaging studies. Kupffer cells are present in some tumors but are nonfunctional; thus, hepatic adenomas appear as cold defects on technetium sulfur colloid
1921
radionuclide scans. Tumors have poor connective tissue support, making them susceptible to hemorrhage. Most tumors are solitary, smooth, and encapsulated. They do not have central scars. Tumor size is commonly 8 to 15 cm but may be up to 30 cm. Areas of necrosis, hemorrhage, and fibrosis are common (1 6). Liver adenomatosis is characterized by the presence of multiple adenomas in an otherwise normal liver in patients without risk factors for hepatic adenomas. US is sensitive to high fat content or intratumoral hemorrhage, which makes the lesions appear hyperechoic. CT shows well-circumscribed tumors that are often low in attenuation because of internal fat, necrosis, or old hemorrhage. Calcifications in areas of old hemorrhage or necrosis are present in 15% of cases. Postcontrast scans show intense homogeneous enhancement during the arterial phase that becomes isodense with liver on portal venous and delayed phase scans. Tumors are hyperintense on T1WIs because of fat or hemorrhage. On T2WIs, most are hyperdense to liver and are commonly heterogeneous because of hemorrhage or necrosis. Dynamic postgadolinium scans are similar to postcontrast CT, showing intense arterial phase enhancement (Fig. 27.17), with isointensity on portal phase and delayed images. Fibrolamellar carcinoma is a hepatocellular malignancy with clinical and pathologic features that are distinct from HCC (1 7) . Tumors typically present as a large liver mass in a young adult (mean age, 23 years) with none of the risk factors for HCC and without elevation of α-fetoprotein levels. Cords of tumor are surrounded by prominent fibrous bands that emanate from a central fibrotic scar. The surrounding liver is usually normal, without features of cirrhosis or chronic liver disease. The characteristic appearance is a large, lobulated hepatic mass with central scar and calcifications. The central scar with radiating septa mimics the appearance of FNH. Satellite tumor nodules are occasionally present (10% to 15%). Hemorrhage and necrosis are uncommon (10%) but are occasionally massive, resulting in a multicystic appearance of the tumor. Although the tumor is less aggressive than HCC, the stage at presentation tends to be advanced, with malignant adenopathy
1922
present.
Aggressive
surgical
management
is
indicated.
US shows a large, lobulated, well-defined mass with mixed echogenicity. The central scar is echogenic, if visible. On precontrast CT, the tumor is low attenuation. Calcification may be evident within the fibrous scar. The tumor enhances prominently and heterogenously on both arterial and portal venous phases (Fig. 27.18). Enhancement of the scar is most evident on delayed scans. MR shows a usually homogeneous hypointense mass (86%) or an isointense mass (14%) on T1WIs. On T2WIs the mass is usually hyperintense and much more heterogeneous. The fibrous scar is hypointense on all image sequences. Gadolinium enhancement shows the same pattern as CT. Lymphoma involving the liver is usually diffusely infiltrative and undetectable by imaging methods. The multiple-nodule pattern found in 10% of cases resembles P.768 metastatic disease. Some cases present as a large, poorly defined, hypodense mass with or without satellite nodules. On MR, lesions are hypodense on T1WIs and hyperdense on T2WIs. Lesions enhance poorly or not at all (Fig. 27.19) .
1923
FIGURE
27.16. Focal
Nodular
Hyperplasia. The lesion
(arrows), consisting of liver elements, is isodense with the hepatic parenchyma on a T1WI (A) and a gradient-recall twodimensional time-of-flight image (B). The lesion is clearly depicted by intense enhancement during the arterial phase (C), post–gadolinium administration. This lesion lacks a central scar. Note the mottled early enhancement of the portion of the spleen (arrowhead) included on the image.
Benign hepatic cyst is the second most common benign hepatic
1924
mass, found in 5% of the population (1 8). Most are solitary, but they may be multiple, especially in patients with adult polycystic disease or tuberous sclerosis. Cysts range in size from microscopic to 20 cm. Hepatic cysts do not communicate with the biliary tree. Tiny cysts are responsible for many of the “hypoattenuating lesions too small to characterize― seen on MDCT. Larger cysts tend to occur in clusters with cysts of varying size, resulting in sharply defined, but lobulated, margins and septations. Hepatic polycystic disease is part of the spectrum of autosomal-dominant polycystic disease and occasionally occurs in the absence of polycystic kidneys.
FIGURE 27.17. Hepatic Adenoma. Postgadolinium, T1W, fatsuppressed MR image shows intense homogeneous enhancement during arterial phase of a biopsy-proven hepatic adenoma (arrow). The MR appearance is indistinguishable from that of a small hepatocellular carcinoma.
P.769 US is the best imaging modality to characterize hepatic cysts. Typical cysts are anechoic with thin walls and septa and posterior acoustic enhancement. Occasionally, hepatic cysts have internal debris, especially if they have been infected. CT shows low internal attenuation near water (Fig. 27.20) (1 8). MR shows low internal
1925
signal on T1WIs and high internal signal on T2WIs. Cysts do not enhance following contrast administration. Pyogenic
abscess is usually caused by Escherichia
coli,
Staphylococcus aureus, Streptococcus, or anaerobic bacteria (1 9) . Patients present with fever and pain. Destruction of liver results in a solitary cavity or a tight group of individual loculated abscesses (Fig. 27.21). A peripheral rim enhances with contrast. Gas is present within the lesion in 20% of cases (3). Diagnosis is confirmed by percutaneous aspiration. Catheter or surgical drainage is indicated.
FIGURE 27.18. Fibrolamellar Hepatocellular Carcinoma. CT scan demonstrates a large tumor (between white arrows) extending caudally from the right lobe of the liver. A characteristic stellate central scar (black arrow) is present.
1926
FIGURE
27.19. Hepatic
Lymphoma. CT shows a poorly
marginated, hypodense, nonenhancing nodule (arrow) that proved on biopsy to be non-Hodgkin lymphoma.
Amebic abscess is usually solitary, with thick nodular walls (3). The lesion may be indistinguishable from pyogenic abscess (Fig. 27.22); however, the patient is often P.770 more acutely ill and has a history of travel to endemic areas (India, Africa, East Asia, Central and South America). Amebic abscesses commonly occur in the right lobe of the liver, often cause elevation of the right hemidiaphragm, and may rupture through the diaphragm into the pleural space. In the United States, the diagnosis is typically confirmed by serology and the patient is treated with metronidazole. In endemic areas, the diagnosis is confirmed by aspiration of “anchovy paste― material, and the patient is treated by repeated aspiration or catheter drainage.
1927
FIGURE 27.20. Hepatic Cysts. Multiple hepatic cysts are an incidental finding on this postcontrast CT in a 78-year-old patient.
1928
FIGURE
27.21. Pyogenic
Abscess. CT scan shows multiple low-
density areas separated by enhancing septa and representing abscess locules. Several air bubbles (arrowhead) are evident within the lesion.
1929
FIGURE
27.22. Amebic
Abscess. Postcontrast CT image
reveals a thick-walled fluid collection in the right hepatic lobe. Differentiation of amebic from pyogenic liver abscess is made by history, serology, or image-guided aspiration.
Echinococcus cyst is caused by infestation with Echinococcus granulosus or E multilocularis tapeworm (1 9). The parasite is endemic in central and northern Europe, the Mediterranean, northern Asia, China, Japan, Turkey, and parts of North America. The liver is the most common organ affected (95%). Single or multiple cystic masses usually have well-defined walls that commonly calcify (50%). The cyst wall and septations usually enhance. Daughter cysts may be visualized within the parent cyst (75%). Diagnostic aspiration carries a risk of anaphylactic reaction. Treatment is mebendazole or surgical excision. Cystic/necrotic
tumor must always be considered for atypical
1930
cystic masses. Metastases may be necrotic or predominantly cystic. Biliary cystadenoma and cystadenocarcinomas are rare primary tumors that resemble mucinous cystic tumors of the pancreas. Undifferentiated embryonal sarcomas are seen in older children, adolescents, and young adults (1 8) . Small
hypoattenuating
lesions are detected with increased
frequency on MDCT because of thinner collimation, improved resolution, and rapid, multiphase, postcontrast scanning (Fig. 27.23) . Lesions smaller than 1 cm P.771 are difficult to characterize and often too small to biopsy. Differential diagnoses include cysts, hemangiomas, and metastases. Statistically, most of these tiny lesions are benign. In a patient with known malignancy, follow-up scans are needed to exclude metastatic disease. In a series of patients with known malignancy, lesions were metastases in 12% of cases (2 0) .
1931
FIGURE 27.23. Too Small to Characterize. Multidetector CT shows multiple, tiny, low-attenuation lesions (arrowhead) that are too small to definitively characterize. Even in patients with known malignancy, these lesions are usually benign. However, on follow-up in some patients, they will prove to be early metastatic lesions. They are usually identified only on high-quality postcontrast CT. Image-guided biopsy cannot usually be performed because the lesions cannot be identified on US or noncontrast CT.
BILIARY
TREE
Imaging
Methods
Imaging of the biliary tree uses assorted techniques that differ in degrees of invasiveness. US and CT are highly sensitive in the detection of dilation of the bile ducts, though they are somewhat less effective in identifying its cause. US is the preferred screening method for biliary obstruction because of its low cost and convenience. Unenhanced helical CT has a reported sensitivity of 88% in detection of stones in the common bile duct. MR can also demonstrate biliary dilation and may be more effective than CT or US in demonstrating associated tumors. MR cholangiopancreatography (MRCP), performed using extreme T2W sequences, offers a noninvasive method of high-resolution imaging of the biliary tree (2 1). MRCP takes advantage of the long T2 characteristic of bile. Extreme T2 weighting demonstrates bright bile ducts with dark surrounding soft tissues (Fig. 27.24). However, any static fluid will also be bright on MRCP images, so ascites, hepatic and renal cysts, and fluid in the bowel may obscure the biliary tree. Similar to contrast cholangiography, stones are seen as hypodense filling defects (Table 27.8) . Endoscopic retrograde cholangiography and percutaneous transhepatic cholangiography supplement cross-sectional imaging
1932
methods by providing access to the biliary tree for contrast injection and subsequent catheter drainage or biliary stent placement. Operative cholangiography is used to visualize nonpalpable bile duct stones at surgery, and T-tube cholangiography is used to visualize common duct stones following surgery. Radionuclide imaging utilizing technetium-99m-iminodiacetic acid is useful for showing the patency of biliary-enteric anastomoses and for demonstrating bile leaks and fistulae. Scintigraphy has the greatest sensitivity for early obstruction. IV cholangiography involves the use of highly toxic contrast agents and has been abandoned in favor of other techniques. Functional CT and MR cholangiography using oral cholangiographic agents, iopanoic acid (Telepaque) for CT, and mangafodipir trisodium for MR are under investigation.
FIGURE
27.24. Normal
MR
Cholangiopancreatography
(MRCP). Image from a MRCP in a patient who has had a cholecystectomy shows the cystic duct remnant (arrowhead) , normal common bile duct (straight arrow), normal pancreatic duct (curved arrow), and normal major intrahepatic bile ducts (squiggly
arrow) .
1933
TABLE 27.8 Causes of Filling Defects in the Bile Ducts
Biliary
stones
Air bubbles Blood clots Neoplasms Cholangiocarcinoma Ampullary carcinoma Granular cell myoblastoma Mesenchymal tumor Parasites Ascaris lumbricoides Liver
fluke
Anatomy of the Biliary Tract The bile ducts arise as bile capillaries between hepatocytes and join progressively larger branches until two main trunks are formed, draining the right and left lobes of the liver. The ducts of the left hepatic lobe are more anterior than those of the right hepatic lobe. This relationship must be kept in mind when contrast cholangiography is performed. Contrast agents flow to the most dependent portions of the biliary tree and may not opacify nondependent ducts. Failure to fill ducts before gravitational repositioning must not be interpreted as evidence of obstruction. The right and left hepatic ducts combine to form the common hepatic duct (CHD), which courses with the portal vein and hepatic artery in the porta hepatis. The cystic duct courses posteriorly and inferiorly from the gallbladder to join the CHD and form the common bile duct (CBD). The CBD runs ventral to the portal vein and to the right of the hepatic artery, descending from the porta hepatis along the free right margin of the hepatoduodenal ligament to the duodenal bulb. The distal third of the CBD turns caudally and descends in the groove
1934
between the descending duodenum and the head of the pancreas, just anterior to the IVC. The CBD tapers distally as it ends in the sphincter of Oddi, which protrudes into the duodenum P.772 as the ampulla of Vater. The CBD and the pancreatic duct share a common orifice in 60% of individuals and have separate orifices in the remainder. However, because of their close proximity, tumors of the ampulla region generally obstruct both ducts. The CHD and CBD are considered to be extrahepatic bile ducts (EHBDs). Normal intrahepatic bile ducts (IHBDs) are occasionally seen on US and on postcontrast helical CT with thin (5-mm) collimation. Normal IHBDs do not exceed 40% of the diameter of the adjacent portal vein, or 2 mm in diameter in the central liver or 1.8 mm in diameter in the peripheral liver. The extrahepatic CBD is routinely visualized and does not exceed 6 to 7 mm in internal diameter. Normal ducts appear larger on contrast cholangiography studies because of distention and magnification. Slightly larger common ducts are also normal in elderly patients because of elastic tissue degeneration with aging. Cholecystectomy is not proven to alter normal common duct size. Care must be taken to differentiate an enlarged common duct from an enlarged hepatic artery. Color Doppler is useful to make this differentiation on US. Contrast enhancement of blood vessels makes differentiation easy on CT. MRCP and cholangiographic studies demonstrate IHBD branches that parallel the portal veins and correspond to the Couinaud segments of the liver (2 2). The right hepatic duct drains segments V to VIII and is formed by the junction of the more horizontal coursing right posterior duct draining VI and VII and the more vertically coursing right anterior duct draining V and VIII. The left hepatic duct is formed by segmental ducts draining segments II, III, and IV. The duct of the caudate lobe (I) joins either the right or left hepatic duct. Variations include drainage of the right posterior duct into the left hepatic duct (13% to 19%); triple confluence, with the right posterior, right anterior, and left hepatic ducts uniting at a single position (11%); and anomalies of the cystic duct, including low insertion on the CBD, long parallel course with the CHD, and
1935
insertion on the medial rather than the lateral side of the CBD. These anomalies have significant importance to the biliary surgeon.
Biliary
Dilatation
CT, US, and MR are highly effective at demonstrating the anatomic finding of biliary dilatation, which is usually equated with biliary obstruction. However, biliary obstruction may be present intermittently or in the early stage without biliary dilation being present. Alternatively, biliary dilatation may be present without obstruction, such as after surgical decompression or bypass. Patients with clinical evidence of biliary obstruction (i.e., elevated alkaline phosphatase and direct hyperbilirubinemia) may not have biliary dilation. Hepatitis causes swelling of hepatocytes, which blocks biliary capillaries and causes intrahepatic cholestasis without surgical obstruction.
1936
FIGURE 27.25. Biliary Dilation. A CT scan demonstrates dilated intrahepatic ducts (black arrowheads), which are easily differentiated from portal veins (white arrowhead) and hepatic veins by contrast enhancement of the blood vessels. Note that the diameter of the bile ducts clearly exceeds 40% of the diameter of the adjacent portal vein. Biliary dilatation in this patient was caused by adenocarcinoma of the head of the pancreas.
Signs of biliary dilation include the following: (1) multiple branching tubular, round, or oval structures that course toward the porta hepatis; (2) diameter of intrahepatic bile ducts larger than 40% of the diameter of the adjacent portal vein (Fig. 27.25); (3) dilation of the common duct greater than 6 mm; and (4) gallbladder diameter greater than 5 cm, when obstruction is distal to the cystic duct. Benign disease is responsible for approximately 75% of cases of obstructive jaundice in the adult, while malignant disease causes the remainder. Gradual tapering of a dilated common duct suggests benign stricture. Gallstones may be identified in the bile duct surrounded by a crescent of bile. Abrupt termination of a dilated common duct is characteristic of a malignant process (2 3) . Infected bile is present in up to 10% of cases of complete biliary obstruction and 60% of cases of partial or intermittent biliary obstruction. IV antibiotic therapy is warranted prior to biliary interventional procedures in the obstructed patient. Causes of biliary dilation and obstruction (Table 27.9) include the following. Choledocholithiasis is responsible for approximately 20% of cases of obstructive jaundice in the adult (Fig. 27.26). Gallstones are present in the gallbladder in P.773 10% of the population, but the presence of stones in the gallbladder does not necessarily mean that stones are the cause of ductal obstruction. In addition, 1% to 3% of patients with
1937
choledocholithiasis will have no stones in the gallbladder.
TABLE 27.9 Causes of Biliary Tract Obstruction
BENIGN (75% of cases) Benign stricture Surgery/instrumentation Trauma Stone passage Pancreatitis Cholangitis Choledochal cyst Stone impacted in duct Parasite (Ascariasis) Liver cyst MALIGNANT (25% of cases) Pancreatic carcinoma Ampullary/duodenal carcinoma Cholangiocarcinoma Metastasis
The sensitivity of US for stones in the bile ducts ranges from 20% to 80%. Stone detection by US is much improved when the CBD is dilated and the pancreatic head is well visualized. CT sensitivity is 70% to 80%, with stones appearing as intraluminal masses of varying attenuation. The “target sign― or “crescent sign― describes the appearance of an intraluminal stone that is partially surrounded by a crescent of low-attenuation bile (2 4) . Contrast studies and MRCP have the highest sensitivity for stone detection (95% to 99%) and demonstrate stones as dark filling defects within bright bile (Fig. 27.27) .
1938
FIGURE 27.26. Obstructing Stone in Common Bile Duct. Serial CT images obtained from a jaundiced patient demonstrate dilatation of the common bile duct (small arrows) caused by an obstructing high-density gallstone (large arrow) impacted in the distal common bile duct. Note the course of the common bile duct in relationship to the head of the pancreas (p) and descending duodenum (d).
1939
FIGURE
27.27. Choledocholithiasis. MR
cholangiopancreatography image demonstrates two stones (arrow) seen as filling defects in the distal common bile duct. Ascites (a) outlines the liver. A normal gallbladder (gb) is evident.
Benign stricture is the cause of 40% to 45% of obstructive jaundice in the adult. Causes of benign stricture include trauma, surgery, prior biliary interventional procedures, recurrent cholangitis, previous passage of stones through the bile ducts, radiation therapy, and perforated duodenal ulcer. The wall of the involved CBD enhances minimally with benign strictures, while hyperenhancement of the CBD during the portal venous phase is evidence of malignant stricture
(2 5) .
Pancreatitis is responsible for approximately 8% of cases of biliary obstruction. Inflammation, fibrosis, and inflammatory masses narrow
1940
the bile ducts. Primary sclerosing cholangitis (PSC) is associated with a history of ulcerative colitis in 50% to 70% of cases (2 6). PSC is characterized by insidious onset of jaundice, with progressive disease affecting both IHBDs and EHBDs. Alternating dilation and stenosis (Fig. 27.28) produces a characteristic beaded pattern of the IHBDs. Small saccular outpouchings (duct diverticula), demonstrated on cholangiography, are considered to be pathognomonic. Complications include biliary cirrhosis (50%) and cholangiocarcinoma. AIDS-associated cholangitis is characterized by thickening of the walls of the bile ducts and the gallbladder P.774 because of inflammation and edema. Infection by opportunistic organisms,
most
commonly
cytomegalovirus
and Cryptosporidium, as
well as reaction to HIV itself, are implicated as the causes of observed disease. Bile ducts are commonly dilated in association with stenosis at the ampulla. Ulcers in the common duct, inflammatory changes in the duodenum, and additional evidence of infection with opportunistic organisms are commonly associated.
1941
FIGURE 27.28. Primary Sclerosing Cholangitis. Radiograph from an endoscopic retrograde cholangiography demonstrates the focal irregular strictures and focal mild dilatation of intrahepatic bile ducts that are typical of early stage sclerosing cholangitis.
Recurrent
pyogenic
cholangitis has been called Oriental
cholangiohepatitis because it is an endemic disease in Southeast Asia (2 7). It is characterized by recurrent attacks of jaundice, abdominal pain, fever, and chills. Intrahepatic and extrahepatic bile ducts are dilated and filled with soft pigmented stones and pus. The disease is associated with parasitic infestation (clonorchiasis, ascariasis) and nutritional deficiency. Findings include intraductal stones, severe extrahepatic biliary dilation, focal strictures, and straightening and rigidity of intrahepatic ducts. Complications include liver abscess, biloma, pancreatitis, and cholangiocarcinoma.
1942
Caroli
disease is an uncommon congenital anomaly of the biliary
tract characterized by saccular ectasia of the IHBDs without biliary obstruction (2 8). Only one hepatic lobe or segment may be affected. The EHBDs are spared in 50% of cases. Findings include: (1) saccular dilatation of IHBDs, giving the appearance on cross-sectional imaging of scattered intrahepatic cysts that communicate with the biliary tree; (2) enhancing fibrovascular bundles seen centrally within many of the dilated ducts, producing the characteristic “central dot sign;― (3) segmental distribution of the bile duct abnormality, with normal appearance of unaffected liver segments; (4) cholangiography showing a characteristic pattern of focal biliary narrowing and saccular dilatation; and (5) dilatation of the CBD (10 to 30 mm) in half of cases. The disease is associated with medullary sponge kidney and autosomal-recessive polycystic kidney disease. Complications include pyogenic cholangitis, liver abscess, and biliary stones. Cholangiocarcinoma develops in 7% of cases. Most cases present in childhood. Autosomal-recessive inheritance is evident in many cases. Choledochal cysts are uncommon congenital anomalies of the biliary tree characterized by cystic dilation of the bile ducts (2 9) . Many (60%) present in infancy or childhood, while others present in adulthood. Some are discovered by fetal US. The condition is much more common in females (70% to 84% of cases). Patients present with abdominal pain, mass, and jaundice. The Todani classification (1977) is typically used to describe choledochal cysts (Fig. 27.29) . Type 1 lesions are most common (80% to 90%) and appear as fusiform or saccular dilatations (Fig. 27.30) of the CHD, CBD, or segments of each. Type 2 lesions are diverticula of the CBD attached by a narrow P.775 stalk. Type 3 lesions are termed choledochoceles and are focal dilatations of the intraduodenal portion of the CBD, closely resembling ureteroceles. Type IV lesions are defined as multiple focal dilatations of the IHBDs and EHBDs usually with a focal large cystic dilatation of the CBD. Type 5 lesions are referred to as Caroli disease, which is more properly classified as a disease separate from
1943
choledochal
cyst.
FIGURE 27.29. Classification of Congenital Biliary Cysts. Type 1 choledochal cysts (80% to 90% of cases) are focal, saccular or fusiform, dilatations of the common bile duct. Type 2 cysts (2%) are true diverticula of the common bile duct. Type 3 cysts (1.4% to 5%) are termed choledochoceles and are dilatations of the terminal intraduodenal portion of the common bile duct. A type 4 classification (19%) refers to multiple intrahepatic and extrahepatic bile duct cysts. Caroli disease is classified as type 5.
1944
FIGURE 27.30. Type 1 Choledochal Cyst. Radiograph from endoscopic retrograde cholangiography demonstrates saccular dilation (arrow) of the common bile duct, which is typical of the most common form of choledochal cyst: type 1.
Pancreatic and ampullary carcinomas are the cause of 20% to 25% of cases of biliary obstruction in the adult. Metastatic disease from lung, breast, GI tumors, and lymphoma accounts for 2% of cases. Cholangiocarcinoma is the second most common malignant primary hepatic tumor (3 0). Tumors arise from the epithelium of bile ducts and are usually adenocarcinomas (90%). Growth patterns include
1945
mass forming, periductal infiltrating, and intraductal polypoid. Peripheral cholangiocarcinoma (10%) presents as an intrahepatic hypodense mass, with adjacent biliary dilatation present in only 25% of cases (Fig. 27.31). MDCT demonstrates delayed, mild, thin, incomplete, rimlike enhancement with low tumoral attenuation in most cases. Hilar cholangiocarcinoma (Klatskin tumor) (25%) occurs near the junction of the right and left bile ducts (Fig. 27.32). The tumor is usually small, poorly differentiated, and aggressive and causes obstruction of both ductal systems. Extrahepatic cholangiocarcinoma (65%) causes stenosis or obstruction of the CBD in most cases (95%) and presents as an intraductal polypoid mass in 5%. Infiltrating cholangiocarcinoma shows thickening of the wall of the involved bile duct, with P.776 hyperenhancement during the arterial phase (2 5).
Predisposing
conditions include choledochal cyst, ulcerative colitis, Caroli disease, Clonorchis sinensis infection, and PSC. The tumor may be infiltrative, desmoplastic, and small, making imaging detection as well as needle biopsy difficult. Abrupt stricture and thickening of the duct wall may be the only findings. Cross-sectional imaging is used to detect adenopathy and hepatic metastases. Prognosis is poor, with fewer than 20% of tumors being resectable.
1946
FIGURE 27.31. Cholangiocarcinoma: Peripheral. Postcontrast, T1W, fat-suppressed MR shows a heterogeneous mass (between arrows) within the liver. Biopsy confirmed cholangiocarcinoma. No dilated bile ducts were evident.
FIGURE 27.32. Cholangiocarcinoma: Hilar. Percutaneous transhepatic cholangiogram (PTC) demonstrates abrupt focal narrowing (short arrows) of the proximal common bile duct (cd) near the bifurcation. The intrahepatic bile ducts are diffusely dilated. The common bile duct shows normal narrowing at the ampulla of Vater (open arrow). The PTC needle (long arrow) is evident. D, duodenum.
1947
Gas in the Biliary Tract Gas in the biliary tract is most commonly encountered in the patient with a surgically created biliary-enteric anastomosis or who has received a sphincterotomy to facilitate stone passage (Table
27.10) .
Cholecystoduodenal fistula is most commonly caused by erosion of a gallstone through the gallbladder and into the duodenum. When the gallstone is large, it may cause small bowel obstruction, i.e., “gallstone ileus.― The gallstone may also erode into the colon and pass spontaneously in the feces. Cholecystoduodenal fistula is most common in women because of the higher incidence of gallstones. Choledochoduodenal
fistula is caused by a penetrating peptic
ulcer eroding into the common bile duct (Fig. 27.33) .
GALLBLADDER Imaging
Methods
US is the imaging method of choice for the gallbladder. It offers high anatomic detail, convenience, and cost efficiency. Gallbladder US is reviewed in detail in Chapter 36. Cholescintigraphy utilizing technetium-99m-iminodiacetic
acid
has
sensitivity
and
specificity
comparable to US for the diagnosis of acute cholecystitis. Oral cholecystograms have been abandoned in favor of other imaging methods. However, oral biliary contrast agents, previously used for oral cholecystograms, are currently utilized for CT cholangiography. Plain radiographs demonstrate calcified gallstones, porcelain gallbladder, and emphysematous cholecystitis. CT, as the imaging method of choice for the acute abdomen, frequently provides imaging diagnosis of gallbladder disease (3 1) .
1948
TABLE 27.10 Causes of Gas in the Biliary Tract
Postsurgical Sphincterotomy Choledochoduodenostomy Choledochojejunostomy Biliary-enteric fistula Cholecystoduodenal fistula (gallstone erodes into CBD) Choledochoduodenal fistula (ulcer penetrates CBD) Surgery/trauma Tumor erosion with fistula Infection Emphysematous Pyogenic
cholecystitis
cholangitis
CBD, common bile duct.
1949
FIGURE
27.33. Choledochoduodenal
Fistula. An upper GI
series demonstrates filling of the bile ducts because of a penetrating duodenal ulcer that created a fistula (large arrow) between the duodenum (d) and the bile ducts (small arrow) .
Anatomy The gallbladder lies on the underside of the liver, in the fossa formed by the junction of the left and right lobes. While the position of the fundus is inconsistent, the neck of the gallbladder is invariably positioned in the porta hepatis and major interlobar fissure. The gallbladder fundus frequently causes a mass impression on the top of the duodenal bulb. Kinking and folding of the gallbladder is common and generally easily recognized by careful image analysis. The socalled phrygian cap, which is descriptive of folding of the gallbladder fundus, is a common normal variant. Septa within the gallbladder may be partial or complete. The spiral valves of Heister are small folds in the cystic duct.
1950
The normal gallbladder is well distended with bile following a 4-hour fast and is easily visualized. A gallbladder larger than 5 cm in diameter is considered enlarged (hydropic), while a gallbladder smaller than 2 cm in diameter is considered contracted. The normal gallbladder wall does not exceed 3 mm in thickness—measured from gallbladder lumen to liver parenchyma—when the gallbladder is distended. The normal gallbladder lumen is free of particulate debris and is fluid density on imaging studies.
FIGURE 27.34. Cholelithiasis. A. CT reveals numerous subtle low-attenuation floating gallstones (arrow) within the gallbladder. The stones are nearly isodense with bile. Stones may be overlooked on CT because they are isodense with bile or because of their small size. B . Coronal T2WI shows a large gallstone
(arrow) as a filling defect within high-signal bile.
P.777 Gallstones are present in 8% of the general population and 15% of the population aged 40 to 60 years. Approximately 85% of gallstones are predominantly cholesterol, while 15% are predominantly bilirubin (pigment stones) related to hemolytic anemia. Approximately 10% of
1951
stones are sufficiently radiopaque to be detected by conventional radiographs as laminated or faceted calcifications. Fissures within gallstones may contain nitrogen gas that appears on plain film as branching linear lucencies resembling a “crow's foot.― Gallstones are most common in women (female:male = 4:1) and in patients with hemolytic anemia, diseases of the ileum, cirrhosis, and diabetes mellitus (3 2) . US detects 95% of all gallstones, whereas CT detects only 80% to 85%. Gallstones vary in CT attenuation, from fat density to calcium density (Fig. 27.34). Up to 20% of gallstones are isodense with bile and not detected by CT, while some gallstones are missed because of their small size. Care must be taken to avoid interpreting contrast in adjacent
bowel
as
cholelithiasis.
Contrast studies, MRCP, and T2WIs demonstrate gallstones as “filling defects―—rounded or faceted dark objects within high-density bile. Differential considerations for lesions in the gallbladder that may be mistaken for gallstones include the following. Sludge balls or tumefactive biliary sludge result from biliary stasis. The bile thickens and forms mobile masses that move with changes in patient position. Cholesterol polyps are common benign, polypoid masses that result from accumulation of triglycerides and cholesterol in macrophages in the gallbladder wall. They are of no clinical significance. All are 10 mm or smaller. Adenomyomatosis may be focal and present as a polypoid mass fixed to the gallbladder wall. Adenomatous polyps are small, usually flat masses fixed to the gallbladder wall. Gallbladder carcinoma may present as a polypoid mass. Most are 1 cm or larger. Gallstones are usually present.
Acute
Cholecystitis
1952
Acute inflammation of the gallbladder is caused by gallstones obstructing the cystic duct in 90% of cases. Acalculous cholecystitis occurs nearly always in patients with predisposing conditions (listed subsequently). Cholescintigraphy and US have comparable sensitivities and specificities in the diagnosis of acute cholecystitis. Scintigraphic diagnosis of acute cholecystitis is based on obstruction of the cystic duct with nonvisualization of the gallbladder. The normal gallbladder demonstrates progressive accumulation of radionuclide activity over 30 minutes to 1 hour following injection of technetium-99m-iminodiacetic acid. Delayed visualization of the gallbladder may be seen in patients with biliary stasis caused by fasting or hyperalimentation. Delayed images taken at 4 hours postradionuclide injection are needed to assess for this possibility. The test is considered positive if there is prompt tracer accumulation in the liver with excretion of tracer into the bowel and without gallbladder visualization at 4 hours. The test may be considered positive at 1 hour post-radionuclide injection if the gallbladder does not visualize within 20 minutes of IV injection of morphine.
FIGURE
27.35. Acute
Cholecystitis. Postcontrast CT
1953
demonstrates fluid (black arrow) around the enhancing mucosa (white arrowhead) of the gallbladder and a small, highattenuation gallstone (thin arrow) within the gallbladder lumen in a patient with acute, severe right upper quadrant pain. Surgery confirmed acute cholecystitis.
P.778 Confident US diagnosis of acute cholecystitis requires the presence of three findings: cholelithiasis; edema of the gallbladder wall, seen as a band of echolucency in the wall; and a positive sonographic Murphy sign. CT demonstrates the following (Fig. 27.35):
gallstones,
distended
gallbladder, thickened gallbladder wall, subserosal edema, highdensity bile, intraluminal sloughed membranes, inflammatory stranding in pericholecystic fat, pericholecystic fluid, blurring of the interface between gallbladder and liver, and prominent arterial phase enhancement of the liver adjacent to the gallbladder (3 1) . Acalculous cholecystitis causes special problems in diagnosis because the cystic duct is often not obstructed. Inflammation may be the result of gallbladder wall ischemia or direct bacterial infection. Patients at risk for acalculous cholecystitis include those with biliary stasis caused by lack of oral intake, posttrauma, postburn, postsurgery, or on total parenteral nutrition. Scintigraphy usually demonstrates lack of gallbladder visualization. Although this finding is 90% to 95% sensitive for acalculous cholecystitis, it is only 38% specific. False-positive conditions for nonvisualization include hyperalimentation and prolonged severe illness, which are predisposing conditions for acalculous cholecystitis. US demonstrates a distended tender gallbladder with thickened wall but without stones. Many patients are too ill to elicit a reliable sonographic Murphy sign. Sludge is a term used to describe the presence of thick particulate matter in highly concentrated bile. Calcium bilirubinate and cholesterol crystals precipitate in the bile when biliary stasis is
1954
prolonged because of a lack of oral intake, hyperalimentation, or biliary obstruction. Sludge appears as echodense bile on US and as high-attenuation bile on CT. Because sludge may be found in a fasting but otherwise normal patient, its presence is not definitive evidence of gallbladder disease. Pus, blood, and milk of calcium are additional causes of dense bile. Complications of acute cholecystitis include the following. Gangrenous cholecystitis indicates the presence of necrosis of the gallbladder wall. The patient is at risk for gallbladder perforation. Findings include mucosal irregularity and asymmetric thickening of the gallbladder wall with multiple lucent layers, indicating mucosal ulceration and reactive edema. Perforation of the gallbladder is a life-threatening complication seen in 5% to 10% of cases. Perforation may occur adjacent to the liver, resulting in pericholecystic abscess; into the peritoneal cavity, resulting in generalized peritonitis; or into adjacent bowel, resulting in biliary-enteric fistula. Overall mortality is as high as 24%. A focal pericholecystic
fluid
Emphysematous with gas-forming
collection
suggests
pericholecystic
abscess.
cholecystitis results from infection of the gallbladder organisms, usually E coli or Clostridium
perfringens. Approximately 40% of patients are diabetic. Gallstones may or may not be present. Gas is demonstrated within the wall or within the lumen of the gallbladder by plain film or CT. On US, intramural gas has an arc-like configuration that is difficult to differentiate from calcification and porcelain gallbladder. Mirizzi syndrome refers to the condition of biliary obstruction resulting from a gallstone in the cystic duct eroding into the adjacent common duct and causing an inflammatory mass that obstructs the common duct. Visualization of a stone at the junction of the cystic duct and the common hepatic duct in a patient with biliary obstruction and gallbladder inflammation suggests the diagnosis. Chronic
cholecystitis includes a spectrum of pathology that shares
the presence of gallstones and chronic gallbladder inflammation. Patients with chronic cholecystitis complain of recurrent attacks of
1955
right upper quadrant pain and biliary colic. Imaging findings include gallstones, thickening of the gallbladder wall, contraction of the gallbladder lumen, delayed visualization of the gallbladder on cholescintigraphy, and poor contractility. Variants of chronic cholecystitis include the following. Porcelain
gallbladder describes the presence of dystrophic
calcification in the wall of an obstructed and chronically inflamed gallbladder (Fig. 27.36). The condition is associated with gallstones in 90% of cases. Porcelain gallbladder carries a 10% to 20% risk of gallbladder
carcinoma.
Cholecystectomy
is
usually
indicated.
Milk of calcium bile, also called limy bile, is associated with an obstructed cystic duct, chronic cholecystitis, and gallstones. Particulate matter with a high concentration of P.779 calcium compounds is precipitated in the bile, making the bile radiopaque on plain films or CT. Dependent layering of bile can be demonstrated on plain film radiographs. The bile is extremely echogenic on US, and gallstones may be visualized within it.
FIGURE
27.36. Porcelain
Gallbladder.
1956
Conventional
radiograph of the right upper quadrant of the abdomen shows calcification (arrows) in the wall of the gallbladder (gb). This finding is indicative of chronic obstruction of the cystic duct with chronic cholecystitis. The risk of gallbladder carcinoma is increased.
FIGURE 27.37. Adenomyomatosis of the Gallbladder. An MR cholangiopancreatography image shows focal thickening of the wall of gallbladder fundus with a Rokitansky-Aschoff sinus (arrow) extending into the thickened wall.
Xanthogranulomatous cholecystitis is an uncommon variant of chronic cholecystitis characterized by nodular deposits of lipid-laden
1957
macrophages in the gallbladder wall and proliferative fibrosis. Imaging findings include marked wall thickening (2 cm), fat-density nodules in the wall, and narrowing of the lumen. Cholelithiasis is frequently present. The condition is difficult to differentiate from gallbladder carcinoma. Thickening of the gallbladder wall is present when the wall thickness at the hepatic aspect of the gallbladder exceeds 3 mm in patients who have fasted at least 8 hours. Conditions associated with wall thickening include the following.
Acute
and
Chronic
Cholecystitis
Wall thickening is a usual feature of acute cholecystitis and is present in 50% of cases of chronic cholecystitis. Hepatitis causes reduction in bile flow, which results in reduced gallbladder volume and thickening of the gallbladder wall in approximately half of patients. P.780 Portal venous hypertension and congestive heart wall thickening by passive venous congestion.
failure may cause
AIDS is associated with thickening of the gallbladder wall and the walls of the bile ducts. Opportunistic organisms are sometimes present. Hypoalbuminemia is associated with thickened gallbladder wall in 60% of patients. Gallbladder
carcinoma usually presents as a focal mass but may
cause only focal wall thickening. Adenomyomatosis is the most frequent benign condition of the gallbladder and is characterized by hyperplasia of the mucosa and smooth muscle. It is usually focal and in the fundus, but may be diffuse, involving the entire gallbladder. Outpouchings of mucosa into or through the muscularis form characteristic Rokitansky-Aschoff sinuses (Fig. 27.37). The condition has no malignant potential. Coexisting gallstones are commonly present (3 3) .
1958
Gallbladder
Carcinoma
Adenocarcinoma of the gallbladder is commonly overlooked or misdiagnosed preoperatively. The presence of gallstones in 70% to 80% of cases masks the findings of cancer, especially with US examination. Gallbladder carcinoma is most often a tumor of elderly women (>60 years, female:male = 4:1). Patients present with pain, anorexia, weight loss, and jaundice. Calcification of the gallbladder wall (porcelain gallbladder) is a risk factor. Imaging findings include the following: (1) intraluminal soft tissue mass (Fig. 27.38); (2) focal or diffuse thickening of the gallbladder wall; (3) soft tissue mass replacing the gallbladder; (4) gallstones; (5) extension of tumor into the liver, bile ducts, and adjacent bowel; (6) dilated bile ducts; and (7) metastases to periportal and peripancreatic lymph nodes and liver. Most tumors are unresectable at discovery (3 4) .
FIGURE 27.38. Gallbladder Carcinoma. Postcontrast CT shows an enhancing soft tissue mass (arrow) within the lumen of the gallbladder. Direct invasion of tumor into the adjacent liver parenchyma is evident (arrowhead) .
1959
REFERENCES 1. Colagrande S, Centi N, La Villa G, Villari N. Transient hepatic attenuation differences. AJR Am J Roentgenol 2004;183:459–464. 2. Yoshimitsu K, Honda H, Kuroiwa T, Irie H, et al. Unusual hemodynamics and pseudolesions of the noncirrhotic liver at CT. Radiographics 2001;21:S81-S96. 3. Mortele KJ, Segatto E, Ros PR. The infected liver: radiologicpathologic correlation. Radiographics 2004;24:937–955. 4. Dodd GDI, Baron RL, Oliver JHI, Federle MP. Spectrum of imaging findings of the liver in end-stage cirrhosis: part II, focal abnormalities.
AJR
Am
J
Roentgenol
1999;173:1185–1192.
5. Dodd GDI, Baron RL, Oliver JHI, Federle MP. Spectrum of imaging findings of the liver in end-stage cirrhosis: part I, gross morphology and diffuse abnormalities. AJR Am J Roentgenol 1999;173:1031–1036. 6. Gupta AA, Kim DC, Krinsky GA, Lee VS. CT and MRI of cirrhosis and its mimics. AJR Am J Roentgenol 2004;183:1595–1601. 7. Hussain SM, Zondervan PE, Ijzermans JNM, Schalm SW, et al. Benign versus malignant hepatic nodules: MR imaging findings with
pathologic
correlation.
Radiographics
2002;22:1023–1039.
8. Gallego C, Velasco M, Marcuello P, Tejedor D, De Campo L, Friera A. Congenital and acquired anomalies of the portal venous system. Radiographics 2003;22:141–159.
1960
9. Brancatelli G, Federle MP, Graziola L, Golfieri R, Lencioni R. Benign regenerative nodules in Budd-Chiari syndrome and other vascular disorders of the liver: radiologic-pathologic and clinical correlation.
Radiographics
2002;22:847–862.
10. Tani I, Kurihara Y, Kawaguchi A, et al. MR imaging of diffuse liver disease. AJR Am J Roentgenol 2000;174:965–971. 11. Sica GT, Ji H, Ros PR. CT and MR imaging of hepatic metastases. AJR Am J Roentgenol 2000;174:691–698. 12. Vilgrain M, Boulos L, Vullierme M-P, Denys A, Terris B, Menu Y. Imaging of atypical hemangiomas of the liver with pathologic correlation.
Radiographics
2000;20:379–397.
13. Lee KHY, O’Malley MEO, Kachura JR, Haider M, Hanbidge A. Hepatocellular carcinoma: imaging and imaging-guided intervention.
AJR
Am
J
Roentgenol
2003;180:1015–1022.
14. Prasad SR, Wang H, Rosas H, et al. Fat-containing lesions of the
liver:
radiologic-pathologic
correlation.
Radiographics
2005;25:321–331. 15. Hussain SM, Terkivatan T, Zondervan PE, et al. Focal nodular hyperplasia: findings at state-of-the-art MR imaging, US, CT, and pathologic analysis. Radiographics 2004;24:3–19. 16. Graziola L, Federle MP, Brancatelli G, et al. Hepatic adenomas: imaging and 2001;21:877–894.
pathologic
findings.
Radiographics
17. McLarney JK, Rucker PT, Bender GN, et al. Fibrolamellar carcinoma
of
the
liver:
radiologic-pathologic
1961
correlation.
Radiographics
1999;19:453–471.
18. Mortele KJ, Ros PR. Cystic focal liver lesions in the adult: differential CT and MR imaging features. Radiographics 2001;21:895–910. 19. Kawamoto S, Soyer PA, Fishman EK, Bluemke DA. Nonneoplastic liver disease: evaluation with CT and MR imaging. Radiographics
1998;18:827–848.
20. Schwartz LH, Gandras EJ, Colangelo SM, Ercolani MC, Panicek DM. Prevalence and importance of small hepatic lesions found at CT in patients with cancer. Radiology 1999;210:71–74. 21. Jara H, Barish MA. MR cholangiopancreatography techniques. Semin Ultrasound CT MRI 1999;20:281–293. 22. Mortele KJ, Ros PR. Anatomic variants of the biliary tree: MR cholangiographic findings and clinical applications. AJR Am J Roentgenol 2001;177:389–394. 23. Soto JA, Alvarez O, Lopera JE, et al. Biliary obstruction: findings at MR cholangiography and cross-sectional MR imaging. Radiographics 2000;20:353–366. 24. Miller FH, Hwang CM, Gabriel H, Goodhartz LA, Omar AJ, Parsons WG III. Contrast-enhanced helical CT of choledocholithiasis. AJR Am J Roentgenol 2003;181:125–130. P.781 25. Choi SH, Han JK, Lee JM, et al. Differentiating malignant from benign common bile duct stricture with multiphasic helical CT. Radiology 2005;236:178–183.
1962
26. Vitellas KM, Keogan MT, Freed KS, et al. Radiologic manifestations of sclerosing cholangitis with emphasis on MR cholangiopancreatography. Radiographics 2000;20:959–975. 27. Lim JH. Oriental cholangiohepatitis: pathologic, clinical, and radiologic features. AJR Am J Roentgenol 1991;157:1–8. 28. Levy AD, Rohrmann CA Jr, Murakata LA, Lonergan GJ. Caroli's disease: radiologic spectrum with pathologic correlation. AJR Am J Roentgenol 2002;179:1053–1057. 29. Kim OH, Chung HJ, Choi BG. Imaging of the choledochal cyst. Radiographics 1995;15:69–88. 30. Lim JH. Cholangiocarcinoma: morphologic classification according to growth pattern and imaging findings. AJR Am J Roentgenol 2003;181:819–827. 31. Grand D, Horton KM, Fishman EK. CT of the gallbladder: spectrum of disease. AJR Am J Roentgenol 2004;183:163–170. 32. Bortoff GA, Chen MYM, Ott DJ, Wolfman NT, Routh WD. Gallbladder stones: imaging 2000;20:751–766.
and
intervention.
Radiographics
33. Levy AD, Murakata LA, Abbott RM, Rohrmann CA Jr. Benign tumors and tumorlike lesions of the gallbladder and extrahepatic bile ducts: radiologic-pathologic correlation. Radiographics 2002;22:
387–413.
34. Levy AD, Murkata LA, Rohrmann CA Jr. Gallbladder carcinoma: radiologic-pathologic correlation. Radiographics 2001;21:295–314.
1963
1964
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section VII - Gastrointestinal Tract > Chapter 28 - Pancreas and Spleen
Chapter
28
Pancreas and Spleen William E. Brant
PANCREAS Imaging
Techniques
CT, US, and MR provide high-quality images of the pancreatic parenchyma and are used as the primary imaging modalities for the pancreas (Fig. 28.1). Multidetector CT (MDCT) optimizes contrast enhancement for detection of small tumors and provides the capability of CT angiography to detect vascular involvement by pancreatic tumor. Improved techniques and the use of gadolinium enhancement have increased the capability of MR to detect and characterize pancreatic lesions. Endoscopic retrograde cholangiopancreatography (ERCP) provides excellent visualization
of
the lumen of the pancreatic duct (Fig. 28.2), which is usually affected by any mass lesion of the pancreas. This procedure is performed by endoscopic cannulization of the bile and pancreatic ducts, followed by injection of a contrast agent and filming. MR cholangiopancreatography (MRCP) offers a noninvasive method of imaging the pancreatic duct as well as the biliary system (1) . Secretin administration during MRCP increases pancreatic secretions and improves visualization of the pancreatic duct. Arteriography is now routinely performed using CT and MR angiographic techniques (CTA, MRA). US- and CT-guided biopsy and drainage procedures play
1965
a major role in the diagnosis and treatment of pancreatic diseases.
Anatomy The pancreas is a tongue-shaped organ, approximately 12 to 15 cm in length, that lies within the anterior pararenal compartment of the retroperitoneum (Fig. 28.1). The pancreas is posterior to the left lobe of the liver, the stomach, and the lesser sac. It is anterior to the spine, the inferior vena cava, and the aorta. Pancreatic tissue is best recognized by identification of the vessels around it. The neck, body, and tail of the pancreas lie ventral to the splenic vein, with the tail extending into the hilum of the spleen. The splenic vein and pancreas are located anterior to the superior mesenteric artery. The head of the pancreas wraps around the junction of the superior mesenteric vein and the splenic vein, with the uncinate process of the pancreatic head extending under the superior mesenteric vein just anterior to the inferior vena cava. The splenic artery courses through the pancreatic bed in a tortuous course. Atherosclerotic splenic artery calcifications are easily mistaken for pancreatic calcifications. The lumen of the splenic artery may be mistaken for pancreatic cysts or a dilated pancreatic duct on CT without contrast or on US. Maximum dimensions for pancreatic size are a 3.0-cm diameter for the head, 2.5-cm diameter for the body, and 2.0-cm diameter for the tail. The gland is somewhat larger in young patients and progressively decreases in size with age. Because the gland is not encapsulated, fatty P.783 P.784 infiltration between the lobules in older patients gives the pancreas a delicate, feathery appearance on CT. The pancreatic duct is visualized with thin-slice CT and with US. It normally measures 3 to 4 mm in diameter in the head and tapers smoothly to the tail. Films from ERCP show that the normal duct is a bit larger owing to magnification effect and distension resulting from contrast injection (Fig. 28.2). The duodenum cradles the pancreatic head in the C-loop. Many pancreatic abnormalities show secondary effects on the
1966
duodenum, and occasionally on the stomach and colon.
FIGURE 28.1. Normal Pancreas as Seen on CT. A. Image through neck (n), body (b) and tail (t) of the pancreas. B . Image through head (h) and uncinate process (u) of the pancreas. The majority of the pancreas lies anterior to the splenic vein (s) and its junction with the superior mesenteric vein (v), which forms the portal vein (p). The head and uncinate process lie caudal to the majority of the pancreas. The superior mesenteric artery (a) arises from the aorta posterior to the splenic vein and courses caudally, just to the left of the superior mesenteric vein. The superior mesenteric artery is normally surrounded by a collar of clear fat.
1967
FIGURE 28.2. Normal Pancreatic Ducts. A. Radiograph from endoscopic retrograde cholangiopancreatogram demonstrates the main duct of Wirsung (DW, black arrows) and the accessory duct of Santorini (DS, open arrow). In this patient, the main duct drained separately into the major papilla (of Vater), with a different orifice for the common bile duct. The accessory duct drained into the minor papilla. Both ampullae were cannulated endoscopically and injected before this radiograph. A number of different variants of pancreatic duct anatomy exist. This variant is found in about 35% of individuals. Embryologically, the main duct is formed by the entire duct of the ventral pancreatic bud and the distal portion of the duct of the dorsal pancreatic bud. The main duct may join the common bile duct or it may have a separate orifice in the major papilla. The proximal portion of the duct of the dorsal pancreatic bud may be obliterated or persist as the accessory duct. E, endoscope. B . Image from an MR cholangiopancreatogram (MRCP) in a different patient. The pancreatic duct and the common bile duct are well visualized. This patient has had a cholecystectomy. MRCP offers the obvious advantage of being noninvasive. S, stomach. D, duodenal bulb.
On MR, the pancreas is best seen on fat-suppressed T1WIs (2,3) . High protein content in the exocrine pancreas results in high signal of the pancreatic parenchyma, which is difficult to differentiate from fat
1968
on non–fat-suppressed T1WIs. Tumors are typically of lower signal than pancreatic parenchyma on T1WIs. On T2WIs, pancreatic tissue is variable in signal intensity, from as low as the liver to as high as fat. Cystic lesions are bright and easily seen on T2WIs. Gadolinium will enhance the parenchyma, whereas adenocarcinoma enhances poorly and remains low signal on postcontrast T1WIs.
Pancreatitis Acute pancreatitis is generally diagnosed clinically. The role of imaging is to clarify the diagnosis when the clinical picture is confusing, to assess severity, to determine prognosis, and to detect complications (4). Inflammation of the pancreatic tissue leads to disruption of small pancreatic ducts, resulting in leakage of pancreatic secretions. Because the pancreas lacks a capsule, the pancreatic juices have ready access to surrounding tissues. Pancreatic enzymes digest fascial layers, spreading the inflammatory process to multiple anatomic compartments. Causes of acute pancreatitis are listed in Table 28.1. Imaging studies of acute pancreatitis may be normal in mild cases. Contrast-enhanced MDCT provides the most comprehensive initial assessment (5,6); however, US is useful for follow-up of specific abnormalities, such as fluid collections. Abnormalities that may be seen in the pancreas include: (1) focal or diffuse parenchymal enlargement, (2) changes in density because of edema, and (3) indistinctness of the margins of the gland owing to inflammation. Abnormalities in the peripancreatic tissues include stranding densities in the fat with indistinctness of the fat planes and thickening of affected fascial planes. Complications demonstrated by imaging are listed in Table 28.2, and a few are shown in Figs. 28.3, 28.4, and 28.5. US- or CT-directed aspiration biopsy may be needed to confirm the presence of pancreatic abscess. Image-directed catheter placement is an alternative to surgical drainage of pancreatic fluid collections (7). Contrast-enhanced MR is equivalent to CT in the assessment of pancreatitis (2) .
1969
TABLE 28.1 Causes of Acute Pancreatitis
Alcohol abuse—most common cause of chronic pancreatitis Gallstone passage/impaction—most common cause pancreatitis Metabolic disorders Hereditary pancreatitis—autosomal dominant Hypercalcemia Hyperlipidemia—types 1 and 5 Malnutrition Trauma Blunt abdominal
of
acute
trauma
Surgery Endoscopic retrograde Penetrating ulcer Malignancy
cholangiopancreatography
Pancreatic adenocarcinoma Lymphoma Drugs—steroids, tetracycline,
furosemide,
many
others
Infection Viral—mumps, hepatitis, infectious Parasites—Ascariasis, Clonorchis Structural
mononucleosis, AIDS
Choledochocele Pancreas divisum Idiopathic—20% of cases of acute pancreatitis
1970
TABLE 28.2 Complications of Acute Pancreatitis
Pancreatic
fluid
collections
(collections
of
enzyme-rich
pancreatic juice) Acute: resolve spontaneously in 50% of cases; may be intrapancreatic, anterior pararenal space, lesser sac, or extend anywhere in the abdomen, into solid organs, or even into the chest Pseudocyst: round or oval, encapsulated pancreatic fluid collection encased by a distinct fibrous capsule; require at least 4 weeks to develop; about 50% will spontaneously resolve, whereas the remainder will require catheter or surgical
drainage
Liquefactive necrosis of pancreatic parenchyma—seen as lack of parenchymal enhancement during bolus contrast administration on C T, often multifocal. Morbidity and mortality increase dramatically when necrosis is present. Infected necrosis—bacterial infection in necrotic tissue. Seen as an area of nonenhancing pancreatic tissue containing gas. Confirmed with needle aspiration. Infected necrosis generally requires surgical drainage. Abscess—circumscribed collection of pus in area with little or no necrosis tissues. Seen as a fluid collection with a thick wall. Effectively treated with catheter drainage. Hemorrhage—resulting from erosion of blood vessels and tissue necrosis. C T shows high-attenuation blood in the retroperitoneum. Pancreatic ascites—leakage peritoneal cavity.
of
pancreatic
secretions
into
Pseudoaneurysm—autodigestion of arterial walls by pancreatic enzymes results in pulsatile mass that is lined by fibrous tissue and maintains communication with parent artery.
1971
FIGURE
28.3. Acute
Necrotizing
Pancreatitis. CT scan
performed with rapid bolus administration of IV contrast demonstrates enhancement of only the distal body of the pancreas (p). The pancreatic head and neck did not enhance and are lost in the fluid (f) extending from the pancreatic bed. This CT finding is indicative of pancreatic necrosis. st, stomach; L, liver; ivc, inferior vena cava; ao, aorta; k, kidney.
P.785 Chronic
pancreatitis is caused by recurrent and prolonged bouts of
acute pancreatitis, which cause parenchymal atrophy and progressive fibrosis. Both the exocrine and endocrine functions of the pancreas may be impaired. The most common causes are alcohol abuse (70%) and biliary stone disease (20%). Many of the remaining patients may have autoimmune pancreatitis that responds to steroid therapy. The clinical diagnosis is often vague, so imaging is used both to confirm the diagnosis and to detect complications. The morphologic changes of chronic pancreatitis include (1) dilation of the pancreatic duct (70% to 90% of cases), usually in a beaded pattern of alternating areas of dilation and constriction (Fig. 28.6); (2) decrease in visible pancreatic tissue because of atrophy; (3) calcifications (40% to 50% of cases) in the pancreatic parenchyma that vary from finely stippled
1972
to coarse, usually associated with alcoholic pancreatitis (Fig. 28.7); (4) fluid collections that are both intrapancreatic and extrapancreatic; P.786 (5) focal enlargement of the pancreas owing to benign inflammation and fibrosis; (6) dilation of the biliary duct because of fibrosis or mass in the pancreatic head; and (7) fascial thickening and chronic inflammatory changes in surrounding tissues. Differentiation between an inflammatory mass resulting from chronic pancreatitis and that of pancreatic carcinoma frequently requires image-directed biopsy. MR reveals the fibrosis and parenchymal atrophy as a loss of the bright signal of pancreas parenchyma normally seen on T1W fat-suppressed images (3). Parenchymal enhancement on MR is heterogeneous early and increases on delayed images. MRCP and ERCP demonstrate the characteristic changes in the pancreatic duct (3). Calcifications are demonstrated by CT, US, and plain radiographs but are easily overlooked on MR. Autoimmune pancreatitis is characterized by diffuse narrowing of the pancreatic duct and a well-defined capsule that surrounds the pancreas and shows delayed contrast enhancement (8) .
1973
FIGURE 28.4. Pancreatic Fluid Collections. Three fluid collections (F) occurred as complications of acute pancreatitis. Pancreatic fluid dissected to subcapsular locations in the liver (L) and spleen (S), and one collection (arrow) developed within the peritoneal cavity.
FIGURE
28.5. Pancreatic
Abscess. Air (A) and fluid (f) extend
from the bed of the pancreas (p) on this CT scan performed without IV contrast. Air in the pancreatic bed is indicative of abscess and/or fistulous communication with bowel. st, stomach; l, liver; v, inferior vena cava; a, aorta; k, kidney.
1974
FIGURE 28.6. Chronic Pancreatitis. CT demonstrates marked beaded dilatation of the pancreatic duct (arrow) associated with atrophy (arrowhead) of the pancreatic parenchyma. These are characteristic findings of chronic pancreatitis.
FIGURE 28.7. Chronic Pancreatitis. CT in a patient with a history of chronic alcohol abuse reveals innumerable coarse calcifications (arrowhead) throughout the pancreas. This finding is most common in chronic pancreatitis caused by alcoholism.
1975
Solid Lesions of the Pancreas Pancreatic carcinoma (ductal adenocarcinoma) is a highly lethal tumor that is usually unresectable at presentation. The average survival time of a patient with this disease is only 5 to 8 months. It accounts for 3% of all cancers and is second only to colorectal cancer as the most common digestive tract malignancy. Radiographic assessment of resectability is critical, because surgical resection offers the only hope of cure, yet the surgery itself carries a high morbidity. Scanning by CT should include rapid bolus contrast injection and thin slices (9). Adenocarcinoma appears as a hypodense mass that distorts the contour of the gland. Associated findings include obstruction of the common bile duct and pancreatic duct and atrophy of pancreatic tissue beyond the tumor. Metastases commonly go to regional nodes, liver, and the peritoneal cavity. Signs of potential resectability (Fig. 28.8) include isolated pancreatic mass with or without dilation of the bile or pancreatic ducts, or combined dilation of both the bile and pancreatic ducts without an identifiable pancreatic head mass. Signs of unresectability include: (1) extension of the tumor beyond the margins of the pancreas, (2) tumor involvement of adjacent organs, (3) enlarged regional lymph nodes (>15 mm), (4) encasement or obstruction of peripancreatic arteries or veins (Fig. 28.9) (1 0) , P.787 (5) metastases in the liver, and (6) peritoneal carcinomatosis. Only 10% to 15% of patients have tumors that are potentially resectable using these criteria. Image-guided biopsy can confirm the diagnosis in patients whose tumors are deemed unresectable. Tumor recurrence following the Whipple procedure is best detected with MDCT (1 1). MR shows low-signal infiltrative tumor surrounded by high-signal enhanced parenchyma on a postcontrast T1WI. MRCP defines ductal anatomy with dilatation proximal to the stricturing tumor (Fig. 28.10). MRA and MR venography are excellent in identifying vascular involvement by tumor.
1976
FIGURE
28.8. Pancreatic
Carcinoma:
Resectable. This
adenocarcinoma (black arrow) of the pancreatic head proved to be surgically resectable. Central necrosis produced low density and air bubbles in the middle of the lesion. The superior mesenteric artery (white spared of involvement.
arrowhead) and vein (white
1977
arrow) are
FIGURE 28.9. Pancreatic Carcinoma: Nonresectable. Pancreas tumor (T) encases and narrows the celiac axis (arrowhead) and its branches, and partially envelopes the aorta (A). This cancer is not resectable by CT criteria.
FIGURE
28.10. Pancreatic
Carcinoma. Coronal plane image
from an MR cholangiopancreatogram demonstrates a dilated gallbladder (GB), diffuse dilatation of the intrahepatic biliary tree, and marked dilatation of the common bile duct (arrow) that ends abruptly at the tumor (not visualized on this image). The pancreatic duct (arrowhead) was not obstructed by the tumor and is normal in caliber.
Chronic pancreatitis may produce a mass that mimics pancreas carcinoma. Beaded dilatation of the pancreatic duct is characteristic
1978
of chronic pancreatitis, whereas smooth ductal dilatation is most frequent with carcinoma. Calcifications within the mass are common with chronic pancreatitis and are very rare with adenocarcinoma. Islet cell tumors more commonly contain calcifications. As many as 14% of patients with pancreas adenocarcinoma also have chronic pancreatitis. Image-guided biopsy is usually needed to provide a definitive diagnosis, but a negative biopsy is not always definitive because of sampling errors.
Islet
Cell
Tumors
Functioning islet cell tumors produce distinct clinical syndromes and usually present while the tumors are small (1 2). Insulinomas present with hypoglycemia, and gastrinomas present with peptic ulcers, diarrhea caused by gastric hypersecretion, or Zollinger-Ellison syndrome. Other islet cell tumors include glucagonoma (diabetes mellitus and painful glossitis), somatostatinoma (diabetes and steatorrhea),
and
VIPoma
(massive
watery
diarrhea).
Nonfunctioning
islet cell tumors are clinically silent until they present with symptoms of a growing, usually large, mass. Functioning tumors vary in malignant potential, from 10% for insulinoma to 60% for gastrinoma and 80% for glucagonoma. Up to 80% of nonfunctioning tumors are malignant. Functioning islet cell tumors vary in size from 0.4 to 4.0 cm and require strict attention to technique for accurate preoperative identification (1 3). Most small islet cell tumors cannot be identified on precontrast CT. Because the lesions tend to be hypervascular, bolus contrast administration during rapid, thin-slice, MDCT scanning through the pancreatic bed offers the best chance of lesion visualization. The tumor stands out as an enhancing nodule within the pancreas (Fig. 28.11). Sonography has proved extremely valuable for tumor localization during surgery. Islet cell tumors appear as hypoechoic masses within the pancreas. Octreotide is a somatostatin analogue that is used for scintigraphic detection of islet cell tumors. Nonfunctioning islet cell tumors tend to be much larger—6 to 20 cm diameter (Fig. 28.12). Imaging findings include coarse calcifications, cystic degeneration, necrosis, local and vascular invasion, and metastases. On MR, most islet cell tumors are
1979
hypointense on T1WIs and hyperintense on T2WIs and demonstrate bright arterial enhancement on dynamic postcontrast T1WIs (Fig. 28.13) .
FIGURE
28.11. Insulinoma. A small insulin-secreting islet cell
tumor (arrow) is identified by bright enhancement during arterial phase of contrast injection by multidetector CT.
1980
FIGURE 28.12. Nonfunctioning Malignant Islet Cell Tumor. A huge tumor mass (T) arises from the tail of the pancreas. This tumor grew to a large size before producing symptoms. Note the heterogeneous attenuation, which is characteristic of large islet cell malignancies.
P.788 Metastases to the pancreas are most frequent with renal cell carcinoma and bronchogenic carcinoma (1 4). Lesions may appear as a solitary, well-defined, heterogenously enhancing mass, as diffuse heterogeneous enlargement of the pancreas, or as multiple nodules. Tumors have no predilection for any particular portion of the pancreas. On MR, most lesions are low signal on T1WIs and high signal on T2WIs. Melanoma metastases are characteristically hyperintense on T1WIs because of the paramagnetic properties of melanin.
1981
FIGURE
28.13. Malignant Islet Cell Tumor.
Fat-suppressed
T1W early phase postcontrast MR demonstrates bright enhancement of the primary tumor (T) as well as its metastases (arrowheads) in the liver.
1982
FIGURE 28.14. Diffuse Fatty Infiltration of the Pancreas. CT shows diffuse fatty infiltration between the lobules of the pancreas (arrows) in a 70-year-old obese patient.
Lymphoma may involve the pancreas as a primary site (rare) or by direct extension from disease in the retroperitoneum (1 5). On CT, most lesions are homogeneous and of lower attenuation than muscle, and they show limited enhancement. Lesions may be a localized, well-defined mass, or they may be infiltrating diffusely enlarging or replacing the gland. Attenuation may be so low as to appear cystic. Fatty
lesions of the pancreas, similar to those of the liver, include
diffuse fatty infiltration, focal fatty infiltration, focal fatty sparing, and lipoma (1 6). Diffuse infiltration is associated with aging and obesity and is seen with pancreatic atrophy. Fat infiltrates between the lobules of pancreatic parenchyma (Fig. 28.14). In patients with cystic fibrosis, pancreatic parenchyma is eventually completely replaced by fat. Focal fatty sparing in diffuse infiltration may simulate a pancreatic mass, especially when it involves the head or uncinate process. Focal fatty infiltration may involve any portion of the pancreas. Lipomas are rare, usually solitary, fat-density masses that are usually incidental findings but may occasionally obstruct the pancreatic or bile ducts.
Cystic Lesions of the Pancreas Pseudocysts resulting from pancreatitis are the most common pancreatic cystic lesions (1 7). They are of fluid density and have a definable fibrous wall that may be calcified. Internal septations and multiple loculations are common (1 8) . Abscess must be considered in any patient with a cystic pancreatic lesion and a fever. Most abscesses have indistinct walls and contain fluid and debris. The presence of gas bubbles within the cystic mass is good evidence for abscess. Image-directed aspiration confirms the diagnosis and may be followed by percutaneous catheter placement P.789
1983
for treatment. Abscesses usually occur as a complication of pancreatitis. True pancreatic cysts, with epithelial lining, are found in 10% of patients with autosomal-dominant polycystic disease, 30% of patients with von Hippel-Lindau syndrome, and some patients with cystic fibrosis. They appear as well-defined, fluid-filled masses with walls of variable thickness. Patients with von Hippel-Lindau syndrome are prone to islet cell tumors and microcystic adenomas, in addition to multiple pancreatic cysts. Cystic tumors of the pancreas are uncommon (5% to 15% of pancreatic cysts) (1 8,1 9). Islet cell tumors may appear cystic because of extensive necrosis (1 3). Cystic teratomas rarely arise in the pancreas and usually have characteristic hair, fat, calcifications, and cystic and solid components. Microcystic
adenoma (serous cystadenoma) is a benign pancreatic
tumor composed of innumerable small cysts 1 mm to 2 cm in size. The lining epithelial cells are rich in glycogen, resulting in the alternate name of glycogen-rich cystadenoma. The cysts may be so small that the tumor appears as a solid lesion on imaging studies. A characteristic feature is a central stellate fibrous scar, which may be calcified. Approximately 80% of patients with this disease are age 60 or older. This tumor is common in patients with von Hippel-Lindau syndrome. Noncontrast CT shows a well-defined mass with low attenuation near water density. Contrast enhancement is usually marked, with demonstration of multiple internal septations in a honeycomb appearance (Fig. 28.15). US commonly shows an echogenic mass, with only a few of the larger cysts visible. Macrocystic serous cystadenoma, consisting of a unilocular or bilocular cyst larger than 2 cm, accounts for approximately 10% of serous cystadenomas (2 0). Most are smaller than 5 cm. These benign tumors are indistinguishable from potentially malignant mucinous cystic tumors (Fig. 28.16) (2 1) .
1984
FIGURE
28.15. Microcystic
Serous
Cystadenoma. Enhanced
CT shows a mass (arrow) in the pancreatic head consisting of innumerable cysts that are so small the low-attenuation mass appears almost solid.
1985
FIGURE
28.16. Macrocystic
Serous
Cystadenoma. MR
cholangiopancreatography image in sagittal plane shows a 3-cm cystic mass (straight arrow) in the head of the pancreas. A single fine septation is present within the cyst. The appearance suggests mucinous cystic neoplasm, but pathology confirmed a benign macrocystic serous cystadenoma. The common bile duct (arrowhead) and intrahepatic bile ducts are normal. The gallbladder (GB) shows a layer of gallstones (curved arrow). The patient is supine.
Mucinous cystic neoplasm has previously been known as macrocystic adenoma or mucinous cystadenoma/cystadenocarcinoma (Fig. 28.17) (2 0). Lesions are 2 to 30 cm in size and consist of a mucin-filled cyst, which is bounded by a thick fibrous capsule lined by mucin-producing cells. CT shows a multilocular or unilocular cyst in the body or tail of the pancreas. Individual cysts are generally
1986
larger than 2 cm and six or fewer in number. Thin sections and contrast enhancement best demonstrate the tumor detail. MR with contrast enhancement provides optimal detail of the structure inside the cyst. Because all mucinous tumors carry a risk of malignancy, surgical resection is the treatment of choice (1 9). Metastases to the liver tend to be cystic. Intraductal papillary mucinous neoplasm (IPMN) is the recently adopted preferred term for pancreatic tumors that produce an excessive amount of mucin, resulting in marked dilatation of the pancreatic duct and cystic enlargement of the branch ducts (2 2,2 3) . IPMN of the main P.790 duct type is associated with dilatation of the entire main pancreatic duct (Fig. 28.18). Papillary solid tumor excrescences may be seen within the dilated duct. Only a thin rim of atrophic pancreatic parenchyma is present. ERCP demonstrates a bulging papilla with mucin protruding from the orifice. IPMN of the branch duct–type appears as a focal group of small cysts (1 to 2 cm in diameter) that intercommunicate through dilated branch ducts (Fig. 28.19). These lesions are most common in the uncinate process. Some lesions consist of a single unilocular cyst. Pancreatic parenchyma adjacent to the lesions atrophies and becomes the capsule of the cystic mass (2 0) .
1987
FIGURE 28.17. Mucinous Cystic Neoplasm (Cystadenocarcinoma). CT demonstrates a 5-cm unilocular cystic tumor (arrow) (arrowheads) within the liver proved to be metastases.
Duodenal diverticula filled only with fluid may mimic a cystic pancreatic lesion (2 4) .
FIGURE
28.18. Intraductal
Papillary
1988
Mucinous
Neoplasm,
Main Duct Type. Axial T2WI shows massive dilatation of the main pancreatic duct (arrows). No discernible pancreatic parenchyma is evident.
FIGURE
28.19. Intraductal
Papillary
Mucinous
Neoplasm
(IPMN), Branch Duct Type. Coronal T2WI shows a multilobulated cystic mass (straight arrow) occupying the neck and head of the pancreas. Pathology after surgical removal confirmed IPMT. The common bile duct (arrowhead) is normal. A portion of the normal main pancreatic duct (curved arrow) is also evident on this image. A large gallstone (squiggly arrow) is present in the gallbladder.
SPLEEN Imaging
Techniques
CT and US remain the major techniques for imaging the splenic
1989
parenchyma (2 5). Technetium sulfur colloid radionuclide scanning images both the liver and the spleen and can be used to confirm the presence of functioning splenic tissue. MR is somewhat disappointing in its ability to demonstrate splenic abnormalities, because the signal intensity of the splenic parenchyma tends to parallel the signal intensity of pathologic lesions and shows insufficient contrast differentiation to identify the lesions. With the use of gadolinium enhancement, MR imaging of the spleen is improved.
Anatomy The spleen is the body's largest lymphoid organ. Although it serves as a site of blood formation in the fetus, there is no hematopoietic activity in the normal adult spleen. The spleen sequesters abnormal and aged red and white blood cells and platelets and serves as a reservoir for red blood cells. The spleen occupies the left upper quadrant of the abdomen, just below the diaphragm and posterior and lateral to the stomach. Its diaphragmatic surface is smooth and convex, conforming to the shape of the diaphragm, whereas its visceral surface has concavities for the stomach, kidney, colon, and pancreas. Spleen size varies with age, nutrition, and hydration. The spleen is relatively large P.791 in children, reaching adult size by age 15. The average spleen dimensions in adults are 12 cm in length, 7 cm in width, and 3 to 4 cm in thickness. The spleen progressively decreases in size with age. The splenic artery and vein course through the pancreas to the splenic hilum, where they divide into multiple branches. Splenic arteries are end arteries without anastomoses or collateral supply. Occlusion of the splenic artery or its branches produces infarction. On all imaging studies, the spleen has a homogeneous appearance. On both CT and MR, lesions are best demonstrated on contrastenhanced images. On noncontrast CT, the normal spleen density is less than or equal to the density of normal liver. On MR, the spleen signal intensity is lower than hepatic parenchyma on T1WIs and higher than liver parenchyma on T2WIs (2 6). US demonstrates a midlevel, even echo pattern for the splenic parenchyma.
1990
Transient pseudomasses are formed during rapid IV bolus administration of both CT and MR contrast agents because of variable rates of blood flow through the splenic parenchyma (Fig. 28.20) . Images obtained during the arterial phase demonstrate irregular defects in parenchymal enhancement (2 7). One or 2 minutes later, the entire spleen is homogeneously enhanced on both CT and MR. Diffuse liver disease is associated with more prominent splenic pseudomasses during early enhancement. Lobulations and clefts in the splenic contour are common and must not be mistaken for masses or splenic fractures.
FIGURE 28.20. Transient Pseudomasses in Spleen. Multidetector CT image obtained during the arterial enhancement phase of IV contrast injection shows normal early flow enhancement defects (arrows) in the spleen (S).
1991
FIGURE
28.21. Accessory
Spleen. An accessory spleen (arrow)
is seen in the splenic hilum. Accessory spleens have the same imaging characteristics as the parent spleen (S).
Accessory spleens are found in 10% to 16% of normal individuals (2 8). These appear as round masses, 1 to 3 cm in size, and of the same texture as normal splenic parenchyma (Fig. 28.21). They may be single or multiple and are usually located near the splenic hilum. Technetium sulfur colloid radionuclide scans can be used to confirm suspected accessory spleens as functioning splenic tissue. Wandering
spleen is the term applied to a normal spleen that is
positioned out of its normal location in the left upper quadrant.
1992
Laxity of the splenic ligaments, commonly found in association with abnormalities of intestinal rotation, allows the spleen to be positioned anywhere in the abdominal cavity. A wandering spleen may present as a palpable abdominal mass, although most cause no symptoms. The diagnosis is made by recognizing the normal shape and tissue texture of the spleen and the absence of normal spleen in the left upper abdomen and by identifying the blood supply from splenic vessels. Radionuclide scans can confirm functioning splenic tissue. Splenosis refers to multiple implants of ectopic splenic tissue that may occur after traumatic splenic rupture. Splenic tissue can implant anywhere in the abdominal cavity, or even in the thorax if the diaphragm has been ruptured. Splenosis complicates 40% to 60% of splenic P.792 injuries. The splenic implants are usually multiple and vary in size and shape. The tissue fragments enlarge over time and may simulate peritoneal metastases. Functioning splenic tissue is confirmed by radionuclide scanning.
1993
FIGURE 28.22. Splenic Regeneration. Hypertrophy of remnants of splenic tissue deposited on the diaphragm after traumatic splenic rupture has created a homogeneously enhancing mass of functioning splenic tissue (S). This patient has a history of splenectomy. LK, left kidney; St, stomach.
Splenic
Regeneration
After splenectomy, remaining accessory spleens or splenules resulting from traumatic peritoneal seeding of splenic tissue may enlarge and resume the function of the resected spleen. When the spleen is removed, bits of nuclear material, called Howell-Jolly bodies, are routinely seen in red cells on peripheral blood smears. Disappearance of these Howell-Jolly bodies from peripheral blood is a clinical sign of splenic regeneration. Imaging studies demonstrate single or multiple spleen-like masses (Fig. 28.22) in the abdominal cavity in patients with a history of splenectomy. Polysplenia is a rare congenital anomaly that features multiple small spleens, usually located in the right abdomen and associated with situs ambiguous. Both spleens are two-lobed. Most patients also have cardiovascular anomalies. Asplenia (Ivemark syndrome) is the congenital absence of the spleen, found in association with bilateral right-sidedness, midline liver, and bilateral three-lobed lungs. Major cardiac anomalies are present in 50% of cases. Most patients die before 1 year of age.
Splenomegaly The diagnosis of splenic enlargement on imaging studies is usually made subjectively. Although quantitative methods have been attempted, none have proved popular. Findings that suggest splenomegaly are (1) any spleen dimension greater than 14 cm, (2) projection of the spleen ventral to the anterior axillary line, (3) inferior spleen tip extending more caudally than the inferior liver tip, or (4) inferior spleen tip extending below the lower pole of the left
1994
P.793 kidney. Enlarged spleens frequently compress and displace adjacent organs, especially the left kidney (Fig. 28.23). The causes of splenomegaly are exhaustive (Table 28.3). Most do not produce a change in spleen density, so differentiation is based upon associated imaging findings or on clinical evaluation. MR offers no significant benefit in the differential diagnosis of splenomegaly. Mild to moderate splenomegaly is seen with portal hypertension, AIDS, storage diseases, collagen vascular disorders, and infection. More marked splenomegaly is usually associated with lymphoma, leukemia, infectious mononucleosis, hemolytic anemia, and myelofibrosis.
FIGURE 28.23. Splenomegaly. Radiograph from an excretory urogram demonstrates a massively enlarged spleen (S) in a patient with rheumatoid arthritis and Felty syndrome. The left kidney (LK) is compressed and rotated by the large spleen. The inferior margin of the spleen extends well below the inferior margin of the liver (L). RK, right kidney.
1995
TABLE 28.3 Causes of Splenomegaly
Congestive Portal hypertension (50% of cases) Portal vein thrombosis Myeloproliferative disorders Leukemia Lymphoma (30% of cases) Polycythemia vera Idiopathic thrombocytopenia purpura Sickle cell disease (in infants) Thalassemia major Hereditary
spherocytosis
Myelofibrosis Infection Malaria (universal in endemic areas) Schistosomiasis (endemic areas) Infectious mononucleosis Subacute bacterial endocarditis AIDS I V drug abuse Infiltrative Systemic lupus
erythematosus
Amyloidosis Gaucher disease
Cystic Lesions of the Spleen Posttraumatic cysts are false cysts that lack an epithelial lining (2 9). They generally have thick walls and septations that commonly become calcified (30% to 40%) (Fig. 28.24). The internal fluid may be complex owing to blood products, cholesterol crystals, or cellular debris. Posttraumatic cysts result from previous hemorrhage, infarction, or infection. They account for 80% of all splenic cysts.
1996
Epidermoid cysts are true epithelial-lined cysts that are probably developmental in origin. They have the same appearance as posttraumatic cysts but less frequently have calcification in their walls (5%). Pancreatic pseudocysts extend beneath the splenic capsule by tracking along the pancreatic tail to the splenic hilum. Splenic subcapsular pancreatic fluid collections develop in 1% to 5% of patients with pancreatitis (Fig. 28.4). Internal debris and hemorrhage are commonly present. Imaging studies demonstrate associated findings of pancreatitis.
FIGURE 28.24. Posttraumatic Splenic Cyst. The well-defined cyst with thick, densely calcified walls (arrow) seen in the spleen (S) on this CT scan is the result of an old intrasplenic hemorrhage.
Bacterial
abscesses occur most commonly in spleens that are
1997
already diseased. They present with vague symptoms but have a high mortality when left untreated. They result from hematogenous spread of infection (75%), trauma (15%), or infarction (10%). Abscesses appear as single or multiple low-density masses with illdefined thick walls. US commonly demonstrates internal echoes caused by inflammatory debris. Abscesses are low intensity on T1WIs and high intensity on T2WIs. They may contain gas or demonstrate air–fluid levels. Perisplenic fluid collections and left pleural effusions are common. Image-guided aspiration confirms the diagnosis. Treatment is by catheter drainage or splenectomy. Microabscesses are found in patients with immune systems compromised by AIDS, organ transplantation, lymphoma, or leukemia. The causes of microabscesses include fungi, tuberculosis, Pneumocystis jiroveci (Fig. 28.25) , P.794 histoplasmosis,
and
cytomegalovirus.
Imaging
studies
demonstrate
multiple small defects in the spleen, usually 5 to 10 mm but up to 20 mm, in size. The differential diagnosis of multiple small low-density splenic defects is listed in Table 28.4.
1998
FIGURE
28.25. Microabscesses in the Spleen. Multiple lucent
defects of varying size in the spleen (S) of this patient with AIDS are attributable to Pneumocystis jiroveci infection.
1999
TABLE 28.4 Causes of Multiple Small (10-mm) Lesions in the Spleen
Microabscesses (immunocompromised Multiple bacterial abscesses Histoplasmosis
patients)
Lymphoma Kaposi sarcoma (AIDS patients) Sarcoidosis Gamna-Gandy bodies Metastases Breast carcinoma Lung
(portal
hypertension)
carcinoma
Ovarian
carcinoma
Gastric carcinoma Malignant melanoma Prostate
carcinoma
Hydatid cysts in the spleen are found in only 2% of patients with hydatid disease. Hydatid cysts are usually also present in the liver or lung. The lesions consist of spherical mother cysts that contain smaller daughter cysts and have internal septations and debris representing hydatid sand. Ringlike calcifications in the wall are usually prominent in the chronic stage.
2000
FIGURE
28.26. Lymphoma. Contrast-enhanced
CT
demonstrates a lobulated low-attenuation mass (arrow) within the parenchyma of the spleen (S). Note the resemblance to the splenic flow defect illustrated in Fig. 28.20.
Solid Lesions of the Spleen Lymphoma is the most common malignant tumor involving the spleen. Commonly, a spleen involved with lymphoma appears normal on all imaging studies. CT is only 65% sensitive in demonstrating splenic involvement with lymphoma. Patterns of involvement that are visible on imaging studies include diffuse splenomegaly, multiple masses of varying size, miliary nodules resembling microabscesses, large solitary mass (Fig. 28.26), and direct invasion from adjacent lymphomatous nodes. Adenopathy is frequently evident elsewhere in the abdomen when the spleen is involved with lymphoma. Lymphoma is a common predisposing condition for splenic infarction.
2001
Metastases are found in the spleen on autopsy series in up to 7% of patients who die of cancer. Most splenic metastases are microscopic and are not detected by imaging studies. The most common tumors to metastasize to the spleen are malignant melanoma and lung, breast, ovary, prostate, and stomach carcinoma. Metastases appear as single or multiple low-density masses. On MR, metastases are low intensity on T1WIs and high intensity on T2WIs. The increased signal intensity of the lesions parallels the increased signal intensity of the normal splenic parenchyma on T2WIs, and the lesions may not be evident. Contrast enhancement is recommended for both CT and MR demonstration of metastases. Calcification is rare. Melanoma metastases commonly appear cystic.
FIGURE 28.27. Splenic Infarction. Coronal plane MR with T2 weighting demonstrates three splenic infarcts (arrows) as highsignal areas of parenchyma that extend characteristically to the splenic capsule.
2002
P.795 Infarction is produced by occlusion of the main or branch splenic arteries. Causes of infarction include emboli (owing to endocarditis, atherosclerotic plaques, or cardiac valve thrombi); sickle cell disease; pancreatitis; pancreatic tumors; and arteritis. Additional predisposing conditions include myeloproliferative disorders, hemolytic anemias, and sepsis. Infarcts classically appear as wedgeshaped defects in the splenic parenchyma. However, multiple infarcts may fuse, and the wedge shape may be lost. The key finding is extension of the abnormal parenchymal zone to an intact splenic capsule (Fig. 28.27). Splenomegaly, especially when caused by lymphoma, is a predisposing condition. Complications of splenic infarctions include subcapsular hematomas, rupture with hemoperitoneum. Gamna-Gandy
bodies (also called siderotic
infection,
and
splenic
nodules) are small
hemorrhages in the spleen caused by portal hypertension (2 6). They are seen best on MR as multiple small low-intensity nodules on T1WIs (Fig. 28.28) and T2*WIs. Signal intensity is low because of hemosiderin content. They do not enhance. Hemangioma is the most common primary neoplasm of the spleen, found in 14% of patients on autopsy P.796 series (3 0). The tumor consists of vascular channels of varying size lined by a single layer of endothelium. Imaging studies demonstrate an appearance similar to that of hemangiomas in the liver. US shows a well-defined hyperechoic mass. On CT, the lesion may appear solid and may have central punctate or peripheral curvilinear calcification. On MR, the lesion is low in signal intensity on T1WIs and high in signal intensity on T2WIs. The contrast enhancement pattern is variable (Fig. 28.29). The nodular enhancement from the periphery described for liver hemangiomas is not often seen with splenic hemangiomas.
2003
FIGURE
28.28. Gamna-Gandy
Bodies. Axial T1WI shows
numerous low-signal nodules (arrowhead) throughout the splenic parenchyma in a patient with splenomegaly and portal hypertension. These represent hemosiderin deposits from previous tiny intraparenchymal hemorrhages.
2004
FIGURE 28.29. Hemangioma Spleen. Postcontrast CT shows this splenic hemangioma (arrow) to be an inhomogeneous, minimally enhancing, lobulated, low-attenuation mass.
2005
FIGURE 28.30. Angiosarcoma Spleen. Axial T2WI shows nearly complete replacement of the parenchyma of the spleen (S) with numerous heterogeneous high-signal nodules of various sizes. Pathology confirmed nearly complete involvement of the spleen with angiosarcoma.
Angiosarcoma is very rare but is still the most common malignancy arising in the spleen (3 1). The tumor is aggressive, usually presenting with widespread metastases, especially to the liver. Imaging studies demonstrate multiple well-defined enhancing nodules or diffuse spleen abnormality (Fig. 28.30). Patients with Thorotrast exposure are at increased risk.
Aids Splenomegaly
associated
with
generalized
2006
lymphoid
hyperplasia
is
the most common finding in patients with AIDS. Focal lesions in the spleen are usually caused by opportunistic infections such as pneumocytes, Pneumocystis jiroveci, atypical mycobacterium, or Candida. Pneumocystis jiroveci infection may cause multiple splenic calcifications (Table 28.5). AIDS-associated lymphoma and Kaposi sarcoma may also cause single or multiple solid-appearing lesions in the
spleen.
TABLE 28.5 Causes of Multiple Splenic Calcifications
Histoplasmosis Tuberculosis Healed Pneumocystis Phleboliths Hemangiomas
jiroveci (AIDS patient)
REFERENCES 1. Reuther G, Kiefer B, Tuchmann A, Pesendorfer FX. Imaging findings of pancreaticobiliary duct diseases with single-shot MR cholangiopancreatography. 1997;168:453–459.
AJR
Am
J
Roentgenol
2. Miller FH, Keppke AL, Dalal K, et al. MRI of pancreatitis and its complications: part 1, acute pancreatitis. AJR Am J Roentgenol 2004;183:1637–1644. 3. Miller FH, Keppke AL, Wadhwa A, et al. MRI of pancreatitis and its complications: part 2, chronic pancreatitis. AJR Am J Roentgenol 2004;183:1645–1652.
2007
4. Balthazar EJ. Acute pancreatitis: assessment of severity with clinical and CT evaluation. Radiology 2002;223:603–613. 5. Mortele KJ, Wiesner W, Intriere L, et al. A modified CT severity index for evaluating acute pancreatitis: improved correlation with patient outcome. AJR Am J Roentgenol 2004;183:1261–1265. 6. Paulson EK, Vitellas KM, Keogan MT, et al. Acute pancreatitis complicated by gland necrosis: spectrum of findings on contrastenhanced CT. AJR Am J Roentgenol 1999;172:609–613. 7. Lee MJ, Wittich GR, Mueller PR. Percutaneous intervention in acute
pancreatitis.
Radiographics
1998;18:711–724.
8. Irie H, Honda H, Baba S, et al. Autoimmune pancreatitis: CT and MR characteristics. AJR Am J Roentgenol 1998;170:1323–1327. 9. Tamm EP, Silverman PM, Charnsangavej C, Evans DB. Diagnosis, staging, and surveillance of pancreatic cancer. AJR Am J Roentgenol 2003;180:1311–1323. 10. Novick SL, Fishman EK. Three-dimensional CT angiography of pancreatic carcinoma: role in staging extent of disease. AJR Am J Roentgenol 1998;170:139–143. 11. Bluemke DA, Abrams RA, Yeo CJ, Cameron JL, Fishman EK. Recurrent pancreatic adenocarcinoma: spiral CT following the Whipple procedure. Radiographics 1997;17:303–313.
evaluation
12. Buetow PC, Miller DL, Parrino TV, Buck JL. Islet cell tumors of the pancreas: clinical, radiologic, and pathologic correlation in
2008
diagnosis
and
localization.
Radiographics
1997;17:453–472.
13. Sheth S, Hruban RK, Fishman EJ. Helical CT of islet cell tumors of the pancreas: typical and atypical manifestations. AJR Am J Roentgenol 2002;179:725–730. 14. Scatarige JC, Horton KM, Sheth S, Fishman EK. Pancreatic parenchymal metastases: observations on helical CT. AJR Am J Roentgenol
2001;176:695–699.
15. Merkle EM, Bender GN, Brambs H-J. Imaging findings in pancreatic lymphoma: differential aspects. AJR Am J Roentgenol 2000;174:671–675. 16. Katz DS, Hines J, Math KR, et al. Using CT to reveal fatcontaining abnormalities of the pancreas. AJR Am J Roentgenol 1999;172:393–396. P.797 17. Demos TC, Posniak HV, Harmath C, Olson MC, Aranha G. Cystic lesions of the pancreas. AJR Am J Roentgenol 2002;179:1375–1388. 18. Kim YH, Saini S, Sahani D, Hahn PF, et al. Imaging diagnosis of cystic pancreatic lesions: pseudocyst versus nonpseudocyst. Radiographics 2005;25:671–685. 19. Spinelli KS, Fromwiller TE, Daniel RA, et al. Cystic pancreatic neoplasms: observe or operate. Ann Surg 2004;239:651–657. 20. Grogan JR, Saeian K, Taylor AJ, et al. Making sense of mucinproducing pancreatic tumors. AJR Am J Roentgenol 2001;176:921–929.
2009
21. Khurana B, Mortele KJ, Glickman J, Silverman SG, Ros PR. Macrocystic serous adenoma of the pancreas: radiologicpathologic correlation. AJR Am J Roentgenol 2003;181:119–123. 22. Lim JH, Lee G, Oh YL. Radiologic spectrum of intraductal papillary mucinous tumor of the pancreas. Radiographics 2001;21:323–340. 23. Sohn TA, Yeo CJ, Cameron JL, et al. Intraductal papillary mucinous neoplasms of the pancreas: an updated experience. Ann Surg
2004;239:788–797.
24. Stone EE, Brant WE, Smith GB. Computed tomography of duodenal diverticula. J Comput Assist Tomogr 1989;13:61–63. 25. Brant WE, Jain KA. Current imaging of the spleen. Radiologist 1996;3:185–192. 26. Ito K, Mitchell DG, Honjo K, et al. MR imaging of acquired abnormalities of the spleen. AJR Am J Roentgenol 1997;168:697–702. 27. Urban BA, Fishman EK. Helical CT of the spleen. AJR Am J Roentgenol 1998;170:997–1003. 28. Mortele KJ, Mortele B, Silverman SG. CT features of accessory spleen. AJR Am J Roentgenol 2004;183:1653–1657. 29. Urritia M, Mergo PJ, Ros LH, Torres GM, Ros PR. Cystic lesions of the spleen: radiologic-pathologic correlation. Radiographics 1996;16:107–129. 30. Abbott RM, Levy AD, Aguilera NS, Gorospe L, Thompson WM.
2010
Primary vascular neoplasms of the spleen: radiologic-pathologic correlation. Radiographics 2004;24:1137–1163. 31. Thompson WM, Levy AD, Aguilera NS, Gorospe L, Abbott RM. Angiosarcoma of the spleen: imaging characteristics in 12 patients. Radiology 2005;235:106–115.
2011
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section VII - Gastrointestinal Tract > Chapter 29 - Pharynx and Esophagus
Chapter
29
Pharynx
and
Esophagus
William E. Brant
IMAGING
METHODS
The upper GI series (UGI), also called a barium meal, is a barium examination of the alimentary tract from the pharynx to the ligament of Treitz. A barium swallow or esophagram is a study more dedicated to evaluation of swallowing disorders and suspected lesions of the pharynx and esophagus. Barium sulfate preparations are ingested orally, and filming is performed during fluoroscopy. The fluoroscopic examination is commonly videotaped or digitally stored to allow for more detailed review of swallowing dynamics and motility. Doublecontrast techniques, using mucosal coating with barium combined with luminal distension, are preferred for mucosal detail. Distension of the pharynx is provided by having the patient phonate. Distension of the esophagus is attained by having the patient ingest gasproducing crystals. The full-column, or single-contrast, technique uses barium suspension alone to fill and distend the esophagus. Mucosal relief views are collapsed views of the barium-coated esophagus. Cross-sectional imaging techniques are used to stage malignancies of the pharynx and esophagus and to clarify findings seen with other imaging methods. CT complements barium studies and endoscopy of the esophagus by demonstrating the esophageal wall and adjacent structures to determine the extent of disease (1). CT is poor at
2012
evaluating the mucosa and generally cannot differentiate inflammatory and neoplastic conditions. MR is preferred over CT for evaluation of the nasopharynx and is an alternative to CT for demonstrating the extent of esophageal disease. The clear depiction of blood vessels by MR is useful in confirming the presence of varices and in evaluating mediastinal vascular anatomy. Endoscopic sonography is useful for demonstration of tumor penetration of the esophageal wall (2) . This chapter reviews the pharynx, studied as part of a barium examination and for assessment of swallowing disorders. Crosssectional imaging of the neck and pharynx is reviewed in Chapter 9.
ANATOMY The pharynx extends from the nasal cavity to the larynx and is arbitrarily divided into three compartments (Fig. 29.1). The nasopharynx extends from the skull base to the soft palate. Its function is entirely respiratory, and the nasopharynx is not considered further in this chapter. The oropharynx is posterior to the oral cavity and extends from the soft palate to the hyoid bone. The hypopharynx (laryngopharynx) extends from the hyoid bone to the cricopharyngeus muscle. The base of the tongue forms the anterior boundary of the oropharynx. The outline of the surface of the tongue is nodular because of the presence of lymphoid tissue forming the lingual tonsils and the circumvallate papillae, which contain taste buds. The lingual tonsils may hypertrophy and mimic a neoplasm. The epiglottis and aryepiglottic folds separate the larynx P.799 from the oropharynx and hypopharynx. The valleculae are two symmetric pouches formed in the recess between the base of the tongue and the epiglottis. They are divided medially by the median glossoepiglottic fold and bounded laterally by the lateral glossoepiglottic folds. The piriform sinuses are deep, symmetric, lateral recesses formed by the protrusion of the larynx into the hypopharynx.
2013
FIGURE 29.1. Double-Contrast Pharyngogram. Three radiographs of the pharynx coated with barium demonstrate
2014
normal anatomic structures: (A) nondistended lateral view; (B) distended lateral view, obtained by having the patient phonate “eee…―; and (C) frontal (anteroposterior) view. The nasopharynx (NP) extends from the skull base to the soft palate. The oropharynx (OP) spans from the soft palate to the hyoid bone (HB). The hypopharynx (HP) extends from the hyoid bone to the cricopharyngeus muscle (C5–C6), which demarcates the pharynx and esophagus. The epiglottis (e) closes during swallowing to protect the larynx (L) from aspiration. The cricoid cartilage makes a prominent impression on the (long white arrows). The base of the tongue (T) lobulated appearance owing to nodular lymphoid valleculae (V) are recesses between the tongue
hypopharynx has a normal tissue. The and epiglottis,
bordered by the median glossoepiglottic fold (thick white arrow) and the lateral glossoepiglottic folds (black arrowheads). The pyriform recesses (P) extend laterally and posterior to the larynx. The pyriform recesses are commonly slightly asymmetric in size. The laryngeal ventricle (white arrowhead) is faintly visualized, outlined by air between the false vocal cords above and the true vocal cords below.
The esophagus extends from the cricopharyngeus muscle at the level of C5–C6 to the gastroesophageal junction (GEJ). The esophagus is a muscular tube formed by an outer longitudinal muscle layer and an inner circular muscle layer lined by stratified squamous epithelium. The esophagus lacks a serosal layer, which allows for rapid spread of tumor into adjacent tissues. The proximal third of the esophagus is predominantly striated muscle, whereas the distal two thirds, below the level of the aortic arch, is predominantly smooth muscle. Normal extrinsic impressions on the esophagus are made by the aortic arch, the left mainstem bronchus, and the LA. The normal esophageal mucosa is smooth and featureless when fully distended on aircontrast barium studies (3). With partial collapse, multiple longitudinal folds, 1 to 2 mm in thickness, become evident. Multiple regular, transverse folds, 1 mm thick, result from contraction of the longitudinal fibers in the muscularis mucosa (see Fig. 29.17). This
2015
pattern is called feline
esophagus because it is typical of a P.800
normal esophagus in cats. In humans, it may be an early sign of dysmotility or esophagitis.
FIGURE 29.2. Drug-Induced Esophagitis. Air-contrast esophagram demonstrates discrete shallow ulcers en face (black arrow) and in profile (arrowhead). The ulcers were caused by stasis of tetracycline capsules in the esophagus. Multiple regular, thin, transverse folds (white arrow) in the distal esophagus are typical of feline esophagus, a finding suggestive of esophagitis.
2016
On cross-sectional imaging, the esophagus appears as an oval of soft tissue density usually surrounded by fat. The esophagus may contain air or contrast located centrally within its lumen. Eccentric contrast or air should be considered abnormal. The wall of the distended esophagus should not exceed 3 mm in thickness. The anatomy of the esophagogastric region is complex (Fig. 29.2) . The length of the esophagus is tubular, and its termination is saccular. The saccular termination is called the esophageal vestibule. The tubulovestibular junction is formed by a symmetric muscular ring called the A ring. The B ring is an asymmetric mucosal ring or notch that occurs at the junction of esophageal squamous epithelium with gastric columnar epithelium. This squamocolumnar junction is also marked by the Z line, a thin ragged line of demarcation seen on double-contrast views of the lower esophagus. The B ring and the Z line are considered to be radiographic markers of the GEJ. The esophageal hiatus is an angled opening in the diaphragm formed by the edges of the diaphragmatic crura. On CT and MR, the crura appear, often prominently, as teardrop-shaped structures of muscle density. With normal breathing, the proximal vestibule and A ring lie in the thorax. The midvestibule is in the esophageal hiatus, and the distal vestibule and B ring are in the abdomen. With swallowing, the vestibule opens and moves upward and the B ring may be seen 1 cm above the diaphragm.
NORMAL
SWALLOWING
AND
MOTILITY
The normal process of swallowing can be divided into oral, pharyngeal, and esophageal phases. The oral stage involves the voluntary transport of a bolus from the oral cavity into the pharynx. The soft palate elevates and the tongue depresses to accommodate the bolus and channel it into the oropharynx. The oropharynx and hypopharynx receive the bolus and conduct it to the esophagus. Breathing is halted while the larynx elevates, the laryngeal vestibule closes, and the epiglottis and aryepiglottic folds close over the opening into the larynx and deflect the bolus through the lateral piriform
recesses.
2017
The functional upper esophageal sphincter (UES), formed by the cricopharyngeus and other pharyngeal muscles, opens to receive the bolus. Peristalsis conveys ingested material through the tubular esophagus to the stomach. Primary peristalsis consists of a rapid wave of inhibition that opens the sphincters, followed by a slow wave of contraction that moves the bolus. Normal peristalsis will clear the esophagus completely with peristalsis appears as a
each
swallow.
Radiographically,
primary P.801
stripping wave that traverses the entire esophagus from top to bottom. Secondary peristalsis is initiated by distension of the esophageal lumen. The peristaltic wave starts in the midesophagus and spreads simultaneously up and down the esophagus to clear reflux or any part of a bolus left behind. Secondary waves have the same radiographic appearance as primary waves, except that they start at the point of the retained barium bolus. Tertiary waves are nonproductive contractions associated with motility disorders. Irregular contractions follow one another at close intervals from the top to the bottom of the esophagus. These nonperistaltic contractions cause a corkscrew or beaded appearance of the esophageal barium column. The functional lower esophageal sphincter (LES) at the level of the esophageal vestibule relaxes and opens in response to swallowing, primary peristalsis, and proximal esophageal dilation.
2018
FIGURE
29.3. Anatomy of the Gastroesophageal Junction
(GEJ). Radiographs from a double-contrast barium study (A) and a single-contrast barium study (B) demonstrate the physiologic and anatomic landmarks of the GEJ. The Z line (Z, white arrowheads), seen best on the double-contrast study, marks the junction of the squamous epithelium of the esophagus (E) and the columnar epithelium of the stomach (S). The singlecontrast study demonstrates the esophageal vestibule (V) demarcated by the muscular A ring (A, white arrowheads) and the mucosal fold of the B ring (B, black arrowheads). The vestibule marks the location of the lower esophageal sphincter. The Z line and the B ring are markers of the GEJ. Their location
2019
relative to the esophageal hiatus in the diaphragm varies with swallowing and other physiologic motions. The double-contrast study shows the featureless mucosal pattern of the welldistended
normal
esophagus.
Oral and pharyngeal swallowing are evaluated fluoroscopically with the patient in an upright position simulating normal eating. The lateral projection is most useful. Studies are videotaped or digitally stored for subsequent detailed study. Esophageal motility is evaluated by observing fluoroscopically at least five separate swallows of barium with the patient in a prone oblique position. The patient must be instructed to swallow only once, as continuous swallowing distends the esophagus and makes impossible the evaluation
of
primary
MOTILITY
peristalsis.
DISORDERS
Difficulty with swallowing has an increasingly high prevalence with age. Symptoms of abnormal oral or pharyngeal swallowing include difficulty initiating swallowing, globus sensation (lump in throat), cervical dysphagia, nasal regurgitation, hoarseness, coughing, or choking. Symptoms suggesting esophageal dysfunction include heartburn,
dysphagia,
“indigestion,―
and
chest
pain. Dysphagia
is defined as the awareness of swallowing difficulty during the passage of solids or liquids from mouth to stomach. Patients complain of food “sticking in the throat― and of painful swallowing (odynophagia). These symptoms may be caused by anatomic abnormalities, tumors, or motility disorders. The patient's subjective assessment of the location of the abnormality is not reliable. Detailed dynamic barium studies of the entire oropharyngeal-esophageal pathway with videofluoroscopy are needed for complete evaluation. Motility disorders that may cause dysphagia or aspiration are reviewed in this section. Radiographic findings of functional abnormalities of the pharynx and esophagus increase in prevalence with age, may not correlate with specific symptoms, and must be interpreted with caution.
2020
Signs
of
Pharyngeal
Dysfunction
Pharyngeal stasis, indicative of impaired pharyngeal transport, is seen as increased residual volume of swallowed material filling the valleculae and piriform sinuses (4). Laryngeal penetration is defined as entry of barium into the laryngeal vestibule without passage below the vocal cords. Aspiration implies barium passage below the vocal cords (Fig. 29.3). Any of these findings may precipitate a cough. Laryngeal penetration and tracheobronchial aspiration are associated with an increased risk of pneumonia, especially in hospitalized patients. Nasal regurgitation occurs when the soft palate does not make a good seal against the posterior pharyngeal wall. Causes include neurologic impairment, muscular dystrophies, and structural defects in the palate. The major causes of pharyngeal dysfunction are listed in Table 29.1. Cricopharyngeal achalasia is attributable to failure of complete relaxation of the UES, commonly resulting in dysphagia and aspiration.
Barium
swallow
demonstrates P.802
a shelflike impression (cricopharyngeal bar) on the barium column at the pharyngoesophageal junction at the level of C5–C6. The pharynx is distended, and barium may overflow into the larynx and trachea. Because some normal individuals have a prominent cricopharyngeal impression, controversy exists as to how prominent the impression must be to be considered significant. Narrowing of the lumen by more than 50% of its usual diameter is generally accepted as a definite cause of dysphagia. Cricopharyngeal dysfunction is commonly associated with neuromuscular disorders of the pharynx.
2021
TABLE 29.1 Causes of Pharyngeal Swallowing Dysfunction
Aging
(primary
presbyphagia)
Neurologic disease Cerebrovascular accident Multiple sclerosis Movement disorders Neurodegenerative diseases CNS infections Muscle disease Muscular dystrophies Myasthenia gravis Structural
abnormalities
Pharyngeal webs Zenker diverticulum Tumors Medications Radiation Gastroesophageal Zenker
reflux
diverticulum
Trauma Postsurgical Malignancy
changes
Oral cavity Pharynx Larynx
Esophageal
achalasia is a disease of unknown etiology
characterized by (1) absence of peristalsis in the body of the esophagus, (2) marked increase in resting pressure of the LES, and (3) failure of the LES to relax with swallowing. The abnormal peristalsis and LES spasm result in a failure of the esophagus to empty. Pathologically, cases show a deficiency of ganglion cells in the myenteric plexus (Auerbach plexus) throughout the esophagus.
2022
The clinical presentation is insidious, usually at age 30 to 50 years, with dysphagia, regurgitation, foul breath, and aspiration. Radiographic signs include (1) uniform dilatation of the esophagus, usually with an air–fluid level present; (2) absence of peristalsis, with tertiary waves common in the early stages of the disease; (3) tapered “beak― deformity at the LES because of failure of relaxation (Fig. 29.4); and (4) increased incidence of epiphrenic diverticula and esophageal carcinoma. Achalasia is treated with balloon dilation or Heller myotomy. Diseases that may mimic esophageal
achalasia
include
the
following.
2023
FIGURE 29.4. Aspiration on a Barium Swallow. Frontal radiograph taken during a barium swallow examination demonstrates the appearance of aspiration. Barium coats the surface of the false cords (F), the intervening laryngeal ventricle (arrowhead), and the true vocal cords (T). Barium coating to this level would be diagnostic of laryngeal penetration. However, barium coating is seen in the proximal trachea (arrows) indicating that aspiration has occurred. Barium is also seen pooling in the piriform recesses (P). This is a normal finding.
Chagas disease is caused by the destruction of ganglion cells of the esophagus as the result of a neurotoxin released by the protozoa Trypanosoma cruzi, which is endemic to South America, especially eastern Brazil. The radiographic appearance of the esophagus is identical to that seen in achalasia. Associated abnormalities include cardiomyopathy, megaduodenum, megaureter, and megacolon. Carcinoma of the GEJ may mimic achalasia, but it tends to involve a longer (>3.5-cm) segment of the distal esophagus, is rigid, and tends to show more irregular tapering of the distal esophagus and mass
effect.
2024
FIGURE 29.5. Esophageal Achalasia. Radiograph from an upper GI series reveals uniform dilatation of the esophagus (E) to the level of the gastroesophageal junction, where a beak (arrow) is formed by the barium column. Repeated observation by fluoroscopy confirmed failure of relaxation of the lower esophageal sphincter and prolonged retention of barium in the esophagus, even in the upright position. The differential diagnosis for narrowing of the distal esophagus includes tumor, strictures (most often caused by gastroesophageal reflux disease), Chagas disease, and post-vagotomy effect. S, stomach.
2025
P.803 Peptic strictures are usually associated with normal primary peristalsis. Diffuse esophageal spasm is a syndrome of unknown cause characterized by multiple tertiary esophageal contractions (Fig. 29.5), thickened esophageal wall, and intermittent dysphagia and chest pain. Primary peristalsis is usually present, but the contractions are infrequent. Most patients are middle-aged. The LES is frequently dysfunctional, and the condition commonly improves with balloon dilatation of the LES. Neuromuscular
disorders are a common cause of abnormalities of
the oral, pharyngeal, or esophageal phases of swallowing. The most common causes of neurologic dysfunction are cerebrovascular disease and stroke. Additional causes include parkinsonism, Alzheimer disease, multiple sclerosis, neoplasms of the CNS, and posttraumatic CNS injury. Diseases of striated muscle, such as muscular dystrophy, myasthenia gravis, and dermatomyositis, predominantly affect the pharynx and proximal third (striated muscle portion) of the esophagus. Scleroderma is a systemic disease of unknown cause characterized by progressive atrophy of smooth muscle and progressive fibrosis of affected tissues. Women are most commonly affected and are usually age 20 to 40 years at the onset of disease. The esophagus is affected in 75% to 80% of patients. Radiographic findings (Fig. 29.6) include (1) weak to absent peristalsis in the distal two thirds (smooth muscle portion) of the esophagus, (2) delayed esophageal emptying, (3) a stiff dilated esophagus that does not collapse with emptying, and (4) wide gaping LES with free gastroesophageal reflux. Despite free reflux, tight strictures of the distal esophagus are uncommon. Postoperative states, including surgery for malignancy of the tongue, larynx, and pharynx, commonly impair swallowing function as well as alter the morphology. Surgical resection is aimed at providing at least a 1-cm
2026
P.804 margin free of tumor and often results in the removal of large blocks of tissue and functional alteration of the structures that remain.
FIGURE 29.6. Diffuse Esophageal Spasm. Image from a barium esophagram demonstrates numerous ineffective tertiary contractions throughout the esophagus. The lower esophageal sphincter was dysfunctional, not opening appropriately on fluoroscopic examination.
Esophagitis frequently results in abnormal esophageal motility and
2027
visualization
of
tertiary
esophageal
contractions.
Gastroesophageal reflux disease (GERD) occurs as a result of incompetence of the LES. The resting pressure of the LES is abnormally decreased and fails to increase with raised intraabdominal pressure. As a result, increases in intra-abdominal pressure exceed LES pressure, and gastric contents are allowed to reflux into the esophagus. Symptoms of GERD include substernal burning pain (“heartburn―), postural regurgitation (in the supine position), and development of reflux esophagitis, dysphagia, and odynophagia. Complications of GERD include reflux esophagitis (RE), stricture, and development of Barrett esophagus. The radiographic diagnosis of GERD may be difficult, because 20% of normal individuals show spontaneous reflux on UGI examination, and patients with pathologic GERD may not demonstrate reflux without provocative tests. Monitoring of esophageal pH for 24 hours in an ambulatory patient is the most sensitive means of diagnosing abnormal Hiatus
gastroesophageal
reflux.
hernia is often considered synonymous with GERD. There is,
however, a poor correlation between the presence of hiatus hernia and GERD or reflux esophagitis. One area of controversy is the definition of hiatus hernia and the criteria used for diagnosis. The simplest definition is protrusion of any portion of the stomach into the thorax. Three types of hiatal hernia are described (5). The most common (95%) is the sliding hiatus hernia, with the GEJ displaced more than 1 cm above the hiatus. The esophageal hiatus is often abnormally widened to 3 to 4 cm (Fig. 29.7). The upper limit of normal hiatal width is 15 mm, and this is most easily measured by CT. The gastric fundus may be displaced above the diaphragm and present as a retrocardiac mass on chest radiographs. The presence of an air–fluid level in the mass suggests the diagnosis. Small, sliding hiatus hernias commonly reduce in the upright position. The mere presence of a sliding hiatus hernia is of limited clinical significance in most cases. The function of the LES and the presence of pathologic gastroesophageal reflux are the crucial factors in producing symptoms and causing complications. Much less common is the paraesophageal hiatus hernia, in which the GEJ remains in its
2028
P.805 normal location while a portion of the stomach herniates above the diaphragm (Fig. 29.8). The mixed or compound hiatal hernia is the most common type of paraesophageal hernia (Fig. 29.9). The GEJ is displaced into the thorax with a large portion of the stomach, which is usually abnormally rotated. Paraesophageal hernias, especially when large with most of the stomach in the thorax, are at risk for volvulus, obstruction, and ischemia.
FIGURE
29.7. Scleroderma. Double-contrast esophagram in a
patient with scleroderma demonstrates a stiff esophagus with peristalsis. The gastroesophageal junction is gaping, and free gastroesophageal reflux was observed. Reflux esophagitis has resulted in mild stricturing (white arrows) of the esophagus and focal ulcers (black arrowhead) .
2029
FIGURE 29.8. Sliding Hiatus Hernia. CT demonstrates a 26mm gap between the crura (arrowheads) of the diaphragm. The normal esophageal hiatus should not exceed 15 mm. The stomach (S) extends through the hiatus and is positioned both above and below the diaphragm. The gastroesophageal junction was seen at a higher level in the thorax.
2030
FIGURE
29.9. Paraesophageal
Hiatal
Hernia.
Radiograph
from an upper GI series shows the characteristic findings of paraesophageal hiatal hernia. The gastroesophageal junction (arrow) and fundus (F) of the stomach are below the diaphragm, while a portion of the body (B) of the stomach herniates through the hiatus into the chest and then doubles back into the abdomen.
OUTPOUCHINGS Lateral pharyngeal diverticula are protrusions of pharyngeal mucosa through areas of weakness of the lateral pharyngeal wall and are most common in the region of the tonsillar fossa and thyrohyoid
2031
membrane. They reflect increased intrapharyngeal pressure and are seen most commonly in wind instrument players. Zenker
diverticulum arises in the hypopharynx just proximal to the
UES. It is located in the posterior midline at the cleavage plane (known as Killian dehiscence) between the circular and oblique fibers of the cricopharyngeus muscle. The diverticulum has a small neck that is higher than the sac, resulting in the trapping of food and liquid within the sac (Fig. 29.10). The distended sac may compress the cervical esophagus. Symptoms include dysphagia, halitosis, and regurgitation
of
food.
FIGURE 29.10. Compound Hiatus Hernia. Left posterior oblique view from an upper GI series demonstrates a large hiatus hernia. The fundus (F) of the stomach (S) extends well above the level of the left hemidiaphragm (open arrow). The widened (6cm) esophageal hiatus makes an impression (small black arrows) on the body of the stomach. The gastroesophageal junction (wide black arrow) is 5 cm above the left hemidiaphragm. The distal esophagus (wide white arrow) is bowed around the herniated
2032
stomach. The right hemidiaphragm (white arrow) projects well above the left hemidiaphragm on this view.
Midesophageal diverticula may be pulsion or traction diverticula. Pulsion diverticula occur as a result of disordered esophageal peristalsis (Fig. 29.11). Traction diverticula occur because of fibrous inflammatory reactions of adjacent lymph nodes. Most midesophageal diverticula have large mouths and empty well, so they are usually asymptomatic. Epiphrenic diverticula occur just above the LES, usually on the right side. They are rare and usually found in patients with esophageal motility disorders (Fig. 29.12). Because they have a small neck that is higher than the sac, they may trap food and liquids and cause symptoms. Sacculations are small outpouchings of the esophagus that usually occur as a sequela of severe esophagitis (Fig. 29.13). They are thought to result from the healing and scarring of ulcerations. Sacculations tend to change in size and shape during fluoroscopic observation. Smooth contours help to differentiate sacculations from ulcerations. Intramural pseudodiverticula are the dilated excretory ducts of deep mucous glands of the esophagus (6). They appear as flaskshaped barium collections that P.806 extend from the lumen or as lines and flecks of barium outside the esophageal wall. They tend to occur in clusters and in association with strictures. Liner tracks of barium (“intramural tracking―) commonly bridge adjacent pseudodiverticula.
2033
FIGURE 29.11. Zenker Diverticulum. Barium swallow examination demonstrates the characteristic barium-filled outpouching, indicating a Zenker diverticulum (ZD) at the junction of the hypopharynx (HP) and cervical esophagus (CE). Note that the neck of the diverticulum (arrowhead) is at a more cephalad location than its base, encouraging the trapping of food and liquid. TE, thoracic esophagus.
2034
ESOPHAGITIS Esophagitis is a common disease with many causes. Radiologic evaluation will detect most cases of moderate to severe esophagitis but will demonstrate fewer than half the cases of mild esophagitis. Attention to excellent technique and use of double-contrast studies are essential. Radiographic signs of esophagitis include (1) thickened esophageal folds (>3 mm), (2) limited esophageal distensibility (asymmetric flattening), (3) abnormal motility, (4) mucosal plaques and nodules, (5) erosions and ulcerations, (6) localized stricture, and (7) intramural pseudodiverticulosis (barium filling of dilated 1- to 3mm submucosal glands). Ulcers are a hallmark finding of esophagitis. Small ulcers (1 cm) are characteristic of cytomegalovirus, HIV, Barrett esophagus, and carcinoma. CT usually reveals nonspecific findings of thickening of the wall (>5 mm) and target sign with hypoattenuating thickened wall and high-attenuation enhancing mucosa (7) .
2035
FIGURE
29.12. Pulsion
examination
demonstrates
Diverticulum. A barium swallow a
persistent
mucosal
outpouching
(arrow) in the midesophagus. The patient was asymptomatic. Pulsion diverticula are formed when the mucosa and submucosa herniate
through
the
muscularis.
Reflux esophagitis is the result of esophageal mucosal injury caused by exposure to gastroduodenal secretions. The severity depends on the concentration of caustic agents, including acid, pepsin, bile salts, caffeine, alcohol, and aspirin, as well as the duration of contact with the esophageal mucosa. The findings of RE are always most prominent in the distal esophagus and GEJ (Fig. 29.13). Early changes of RE include mucosal edema, which is
2036
manifested as a granular or nodular pattern of the distal esophagus. In contrast to the distinct borders of Candida plaques and nodules, RE nodules have poorly defined borders. Inflammatory exudates and pseudomembrane esophagitis;
formation
may
mimic
fulminant Candida P.807
however, the patient has symptoms of reflux rather than severe odynophagia. RE is the most common cause of esophageal ulcerations. The ulcers appear as discrete linear, punctate, or irregular collections of barium, usually surrounded by a radiolucent mound of edema. Prominence of the ulcerations in the distal rather than proximal or midesophagus is the key to differentiating RE ulcers from those of herpes or drug-induced esophagitis. Complications of RE include ulceration, bleeding, stricture, and Barrett esophagus.
2037
FIGURE 29.13. Epiphrenic Diverticula. A stricture (long arrows) of the distal esophagus has resulted in the formation of pulsion diverticula (short arrows). The filling defects (curved arrow) in the barium column are caused by retained boluses of meat proximal to the stricture.
Barrett esophagus is an acquired condition of progressive columnar metaplasia of the distal esophagus caused by chronic GERD. Columnar rather than squamous epithelium lines the distal esophagus. The prevalence of Barrett esophagus in patients with reflux esophagitis is about 10%, but increases to 37% in patients with scleroderma. It is premalignant, with a 30- to 40-fold increased risk of developing adenocarcinoma, resulting in a 15% prevalence of adenocarcinoma in patients with Barrett esophagus. Clinical presentation is usually indistinguishable from that of RE. Adenocarcinoma may develop at any age. The characteristic radiographic appearance of Barrett esophagus is a high (midesophageal) stricture or deep ulcer in a patient with GERD. A reticular mucosal pattern of the esophageal mucosa, resembling areae gastricae of the stomach, is also suggestive. The diagnosis is confirmed by endoscopy and biopsy. Infectious
esophagitis is found most commonly in patients with
compromised immune systems. It is increasingly common because of the use of steroids and cytotoxic drugs and because of the increasing prevalence of AIDS. Candida albicans is by far the most common cause of infectious esophagitis and is highly prevalent in patients with AIDS. Additional risk factors include malignancy, radiation, chemotherapy, and steroid treatments. Candida of the oropharynx (thrush) is commonly present and is usually evident on physical examination. Odynophagia is a prominent symptom. Discrete plaquelike lesions demonstrated by double-contrast esophagrams are most characteristic (Fig. 29.14) . The plaques appear as longitudinally P.808
2038
oriented linear or irregular discrete filling defects etched in white with intervening normal-appearing mucosa. The lesions may be tiny and nodular, or they may be giant and coalescent with pseudomembranes. Ulcers tend to be small ( Table of Contents > Section VII - Gastrointestinal Tract > Chapter 31 - Mesenteric Small Bowel
Chapter
31
Mesenteric
Small
Bowel
William E. Brant
Imaging
Methods
Disease of the mesenteric small intestine is relatively rare (1) . Detailed radiographic study of the small bowel is justified only when clinical suspicion of small bowel disease is high. Small bowel disease is usually manifest by four major symptoms: colic, diarrhea, malabsorption, and bleeding. Colic is defined as recurrent and spasmodic abdominal pain, with periods of relief every 2 to 3 minutes. Diarrhea caused by small bowel disease is less urgent than that caused by colon disease. Malabsorption is manifest by steatorrhea, foul-smelling stools, and weight loss. Bleeding from small bowel disease is usually occult and manifest by anemia. Because the majority of the mesenteric small intestine is out of reach of the endoscopist, diagnostic radiology has the primary responsibility for its evaluation. The traditional method for radiographic examination of the small bowel is the small bowel follow-through examination (SBFT) (Fig. 31.1) tacked onto a standard upper GI series. The patient is asked to continue drinking barium while a series of supine abdominal films are obtained until the terminal ileum and cecum are filled with barium. Fluoroscopic examination of the small bowel is then attempted. This study is notoriously insensitive. It is limited by overlap of bowel loops, poor distension, flocculation of barium, intermittent barium
2111
filling, and unpredictable transit time. Visualization of the distal ileum may be improved with double-contrast technique by insufflating the colon with air (SBFT with peroral pneumocolon). Enteroclysis, or the small bowel enema, is the preferred method for detailed radiographic examination (Fig. 31.2). This study provides more uniform distension of the bowel, even distribution of barium, superior anatomic detail, and shorter overall examination time (1) . The study is performed by passing a specially designed enteroclysis catheter (12 to 14 French) through the mouth or nose and into the distal duodenum or proximal jejunum. A guidewire is used for directional control of the catheter during manipulation under fluoroscopy. The study may be performed single contrast (with approximately 600 mL of barium) or double contrast (with 200 mL of barium followed by 1,000 mL of methylcellulose to advance the barium and distend the bowel). The small bowel lumen and mucosal surface are best demonstrated by barium studies. CT complements the barium examination by demonstrating the extraluminal component of bowel disease. In addition, CT evaluates the mesentery, adjacent solid organs, the peritoneal cavity, and the retroperitoneum. CT signs that aid in the differentiation of benign and malignant disease are listed in Table 31.1 (2,3). CT enteroclysis combines the total opacification of the small bowel achieved by contrast infusion through a nasoenteric tube with thin-section multidetector CT imaging (4). MR enteroclysis is an emerging technique utilizing fast thin-section MR imaging of the small bowel after it has been well distended by nasojejunal catheter infusion of isosmotic water solution.
Anatomy The mesenteric small intestine is a tube approximately 7 m long that lies totally within the greater peritoneal cavity. The jejunum is arbitrarily defined as the proximal P.833 two fifths of the mesenteric intestine, while the ileum is the distal three fifths. The jejunum and ileum are suspended from the posterior
2112
abdominal wall by the small bowel mesentery. The small bowel mesentery is composed of connective tissue, blood vessels, and lymphatic vessels, and it is covered by peritoneum, which reflects from the posterior parietal peritoneum (5). The root of the small bowel mesentery extends obliquely from the ligament of Treitz, just left of the L2 vertebra, to the cecum, near the right sacroiliac joint. On CT the mesentery is defined by its normal vascular structures, which are outlined by fat between loops of bowel. Normal mesenteric lymph nodes may be seen as soft tissue–density nodules that are 5 mm or smaller. The concave border of the small bowel loops is the mesenteric border, where the mesentery attaches. The convex border, facing away from the mesentery, is called the antimesenteric border. Identification of the border involved by disease can be of diagnostic
value.
FIGURE 31.1. Normal Small Bowel Follow-Through. A. Prone abdominal radiograph. B . Spot-compression view of the terminal ileum. The small bowel is demonstrated on an upper GI series by
2113
having the patient ingest additional barium and by obtaining additional radiographs to document passage of barium through the small bowel into the colon. The loops of jejunum (J) have a delicate feathery appearance in the left upper abdomen, whereas the loops of ileum (I) are coarse and featureless in the right lower abdomen. Barium has filled portions of the cecum (C) and ascending and transverse colon (TC), the latter identified by its haustral folds. Colonic haustral folds extend only partway across the bowel lumen, and small bowel folds extend completely across the bowel lumen. The spot compression provides separation of bowel loops in the right lower quadrant to optimally demonstrate the terminal ileum (TI). S, stomach; D, duodenum.
On CT and barium studies, the jejunum has a feathery mucosal pattern, more prominent valvulae conniventes, a wider lumen, and a thicker wall. The ileum has a less featured mucosal pattern; thinner, less frequent folds; a narrower lumen; and a thinner wall. The transition between jejunum and ileum is gradual, and all loops are freely mobile. The ileum has larger and more numerous lymphoid follicles in the submucosa. The villi are fingerlike projections that extend from the entire mucosal surface of the small bowel. They are composed of loose connective tissue of the lamina propria. Tiny capillaries and lymphatic vessels (lacteals) extend to the submucosal vessels. The combination of valvulae conniventes and villi greatly expands the absorptive surface area of the small intestine. The caliber of the normal small bowel lumen is less than 3 cm, with normal fold thickness of less than 2 mm and normal wall thickness of 3 mm. Normal lymph nodes seen in the mesentery are smaller than 4 mm in diameter.
2114
FIGURE
31.2. Normal
Enteroclysis. The enteroclysis catheter
(small arrows) has been passed through the C-loop of the duodenum to the location of the ligament of Treitz (large arrow) with fluoroscopy used to guide catheter manipulation. The enteroclysis technique provides uniform distension of the jejunum (J) and ileum (I). Barium fills portions of the ascending colon (C). Note the small bowel folds crossing the entire diameter of the small bowel lumen. D, duodenum.
2115
TABLE 31.1 Diagnostic Findings of CT of the Gastrointestinal Tract
Benign Circumferential
Symmetric
Lesion
Neoplastic
thickening
Lesion
Eccentric thickening
thickening
Asymmetric thickening
Thickening 2 cm
Segmental or diffuse involvement
Focal soft tissue mass
Thickened
Abrupt
mesenteric
fat
transition
Wall is homogeneous soft tissue density
Lobulated
contour
“Double halo sign―: dark inner ring/bright outer ring
Spiculated contour
outer
“Target sign―: bright inner, dark
Luminal
middle, bright outer
Regional adenopathy Liver metastases
narrowing
P.834
Small
Bowel
Filling
Defects/Mass 2116
Lesions Neoplasms of the small intestine are rare, accounting for only 2% to 3% of GI tumors. Benign neoplasms are about equivalent to malignant neoplasms in overall frequency. However, when the patient presents with symptoms, malignancy is three times more common. Presenting afflictions include obstruction, pain, weight loss, bleeding, and palpable mass. Carcinoid tumors are the most common neoplasm of the small intestine, accounting for about one third of all small bowel tumors (6). They are considered a low-grade malignancy that may recur locally or metastasize to the lymph nodes, liver, or lung. They arise from endocrine cells (enterochromaffin or Kulchitsky cells) deep in the mucosa. These cells produce vasoactive substances, including serotonin and bradykinins. About 20% of all carcinoid tumors arise in the small bowel, most commonly in the ileum, where 30% are multiple. Only 7%—those with liver metastases—present with carcinoid syndrome (cutaneous flushing, abdominal cramps, and diarrhea) because the liver inactivates the vasoactive substances. The tumors grow slowly but cause a marked fibrotic response of the bowel wall and mesentery, because the serotonin produced by the tumor induces an intense local desmoplastic reaction. Complications include stricture, obstruction, and bowel infarction induced by fibrosis of the mesenteric vessels. The tumors may be pedunculated and cause intussusception. Radiographic signs of fibrosis and metastases resemble the findings of Crohn disease and overshadow demonstration of the primary tumor. Barium studies show: (1) luminal narrowing, (2) thickened and spiculated folds, (3) separation of bowel loops by mesenteric mass or (4) bowel loops drawn together by fibrosis, and (5) primary lesion appearing as a small (5 cm) and heterogeneous, with prominent areas of low-attenuation necrosis and hemorrhage (Fig. 31.10). Nodal metastases are uncommon. Calcifications are infrequent. MR shows the solid portions of the lesions to be low signal on T1WIs and high signal on T2WIs. Solid areas show distinct contrast enhancement. Hemorrhage shows characteristic MR signal dependent on its age.
FIGURE 31.10. Malignant GI Stromal Tumor of the Ileum. Contrast-enhanced CT reveals a large heterogeneous tumor (T) that envelops loops of ileum (arrowheads). Areas of low attenuation within the tumor represent hemorrhage, necrosis, and cystic change.
Adenoma accounts for about 20% of benign small bowel neoplasms. It is more common in the duodenum than in the mesenteric small intestine. The tumor is a benign proliferation of glandular epithelium and has the potential for malignant degeneration. Barium studies demonstrate an intraluminal polyp with a finely lobulated surface. Lipoma is most common in the ileum (1 0). The tumor arises from the fat of the submucosa. Lipomas account for about 17% of benign small bowel tumors. Most are asymptomatic incidental findings,
2126
although some may cause bleeding or intussusception. CT demonstration of a fat-density (–50 to –100 H) tumor is diagnostic (Fig. 31.11) . Hemangioma is usually solitary and submucosal, projecting into the lumen as a polyp. These tumors are located predominantly in the jejunum. About two thirds present with bleeding. Barium studies demonstrate a small polyp. The occasional presence of a calcified phlebolith suggests the diagnosis. They account for fewer than 10% of benign small bowel tumors. Polyposis syndromes cause multiple polypoid lesions of the small bowel. The differential diagnosis includes metastases, lymphoma, nodular lymphoid hyperplasia, Kaposi sarcoma, and carcinoid tumors. Peutz-Jeghers syndrome is an autosomal-dominant inherited condition consisting of multiple hamartomatous P.838 polyps in the small intestine (most common), colon, and stomach associated with melanin freckles on the facial skin, palmar aspects of the fingers and toes, and mucous membranes (1 1). Hamartomatous polyps are a nonneoplastic, abnormal proliferation of all three layers of the mucosa, epithelium, lamina propria, and muscularis mucosae. The polyps are most common in the jejunum, are usually pedunculated, and are variable in size up to 4 cm. Patients are at increased risk for intussusception, GI tract adenocarcinoma, and extraintestinal malignancy (breast, pancreas, ovary). Barium studies demonstrate myriad polyps in involved areas of small intestine, separated by normal bowel segments.
2127
FIGURE
31.11. Small Bowel Lipoma. A fat-density mass
(arrowhead) within a loop of proximal ileum was the cause of partial small bowel obstruction. CT demonstration of a mass of pure fat density is diagnostic of lipoma.
Cronkhite-Canada
syndrome involves the small bowel in about half of
cases with multiple inflammatory polyps. The colon and stomach are always involved. Gardner syndrome of inherited adenomatous polyposis coil usually includes a few adenomatous polyps in the small bowel. Juvenile GI polyposis is most common in the colon but occasionally involves the small bowel. Inflammatory polyps containing cysts filled with mucin develop secondary to chronic irritation. Most are round, smooth,
and
pedunculated.
Ascariasis is caused by infestation with the roundworm Ascaris lumbricoides. Ascariasis is found worldwide but is most common in Asia and Africa. Endemic areas in the United States include rural
2128
southern Appalachia and the Gulf Coast states. Infestation is acquired by ingesting food or water contaminated with Ascaris eggs. The eggs hatch in the small bowel. Larvae penetrate the wall and migrate through the vascular system to the lungs, where they molt and grow before migrating up the bronchi and trachea to the larynx, where they are again swallowed. Worms mature in the small bowel, especially in the jejunum, and may reach 15 to 35 cm in size. New generations of infective ova are excreted in feces. A large bolus of worms may obstruct the small bowel, especially in children, or cause intussusception. Worms can be identified on plain abdominal radiographs in 70% of cases (Fig. 31.12). Barium studies demonstrate worms as long linear filling defects. Barium ingested by the worms may be seen in their intestinal tracts as long, stringlike white
lines.
FIGURE 31.12. Ascaris Infestation. Plain radiograph demonstrates diffuse intestinal dilation. Roundworms in the ileum are seen as round and tubular soft tissue densities outlined by
2129
intestinal gas (white arrowheads). A large bolus of entangled worms (black arrows) plugged the distal ileum, causing small bowel obstruction.
Mesenteric
Masses
Masses arising in the small bowel mesentery frequently present as a palpable abdominal mass (1 2). The mesenteric fat may be infiltrated by edema, hemorrhage, or inflammatory cells. The disorders may be diseases of the small intestine or be primary to the mesentery itself. CT, US, and MR provide the most diagnostic information. Lymphoma causing bulky adenopathy is the most common solid mesenteric mass. Confluent adenopathy surrounds mesenteric vessels and fat, producing the “sandwich sign― (Fig. 31.8) . Adenopathy is commonly present in the retroperitoneum and elsewhere.
FIGURE 31.13. Mesenteric Desmoid. Multiple desmoid tumors are evident on this CT image. A large desmoid (D) infiltrates the mesentery, displacing bowel loops. Two smaller desmoid tumors (arrows) appear as soft tissue nodules within the mesentery. Another desmoid tumor (arrowhead) expands the linea alba in
2130
the midline of the anterior abdominal wall.
P.839 Metastases may implant in the mesentery and produce a large mesenteric mass without impingement of the bowel lumen, or they may implant adjacent to bowel, narrowing the bowel lumen. Carcinoid and small bowel adenocarcinoma metastases produce a prominent desmoplastic reaction in the mesentery, whereas melanoma produces no mesenteric retraction. Mesenteric
desmoid tumors are benign but locally aggressive,
solid, fibrous, mesenteric tumors (1 3). They may be solitary (28%) or multiple (72%) and associated with Gardner syndrome. These tumors commonly recur after surgical resection. US and CT demonstrate a homogeneous solid mass with well-defined (68%) or infiltrative
borders
(Fig. 31.13). Attenuation is similar to that of
muscle. Tumors commonly also occur within the muscles of the anterior abdominal wall or in the psoas muscles. GISTs may arise primarily in the mesentery or omentum or may be found as metastases from tumors arising elsewhere. On CT, tumors appear as large, well-defined masses with prominent areas of low density representing hemorrhage and necrosis (1 4) . Mesenteric
cysts are lymphangiomas that arise in the root of the
small bowel mesentery. Most are thin walled and multiloculated with internal fluid that may be chylous, serous, or bloody. US demonstrates a well-defined cyst with internal debris, along with fluid–debris or fluid–fat levels. CT shows a cystic mass displacing loops of small bowel anteriorly and laterally. On MR, cyst contents are hyperintense on T2WIs and hypointense on T1WIs when serous, or hyperintense on T1WIs when chylous or hemorrhagic. GI duplication cyst is a congenital, either partial or complete, replica of the small bowel. Most arise from the distal small bowel and may communicate with the normal intestinal lumen at one or both ends, or not at all. They are lined by intestinal epithelium. US, CT, and MR reveal a thick-walled cyst, usually with serous contents.
2131
Malignancies
(adenocarcinoma)
may
arise
within
duplication
cysts.
Mesenteric teratoma is heterogeneous with cystic and solid components. Demonstration of calcium or fat is a clue to radiographic
diagnosis.
Sclerosing mesenteritis is an uncommon inflammatory condition affecting the root of the mesentery, with variable inflammation, fat necrosis, and fibrosis. CT shows soft tissue infiltration of the mesentery, so-called “misty mesentery.― The cause is unknown. Patients commonly present with abdominal pain.
Diffuse
Small
Bowel
Disease
Students of radiology dread learning about diseases of the small bowel, because they are numerous, obscure, and confusing and lead to long lists of differential diagnoses (see Tables 31.3, 31.4, 31.5). A few common diseases cause the majority of small bowel abnormalities that most radiologists will encounter in routine practice (1). The rest of the list must be known to pass board examinations. Five rules, learned well, help to simplify the problem: Rule #1. Dilatation of the small bowel lumen indicates small bowel obstruction or dysfunction of small bowel muscle. Rule #2. Thickening of small bowel folds indicates infiltration of the submucosa. Rule #3. Uniform, regular, straight thickening indicates infiltration by fluid (edema or blood). Rule #4. Irregular, distorted, nodular thickening indicates infiltration by cells or other nonfluid material. Rule #5. The specific diagnosis requires matching the small bowel pattern with the clinical data. The normal values for small bowel luminal diameter and fold anatomy are given in Table 31.2. Dilated Small Bowel Lumen (Table 31.3). The hallmark of
2132
mechanical bowel obstruction is a point of transition between dilated bowel and nondilated bowel at the site P.840 of obstruction. With muscle dysfunction, the small bowel dilatation is diffuse, with no transition point. If no coexisting mucosal disease is present, the small bowel folds are straight and regular (Fig. 31.14) . See Chapter 26 for an expanded discussion of this topic.
TABLE 31.2 Normal Small Bowel Measurements
Normal Feature
Ileum
Values Jejunum
Diameter of lumen
3.0 cm
2.5 cm
Diameter of lumen during enteroclysis
4.5 cm
3.5 cm
Thickness of folds
2 mm
2 mm
No. of folds
4 to 7 per
2 to 4 per
inch
inch
Depth of folds
8 mm
8 mm
Thickness of bowel wall
3 mm
3 mm
TABLE 31.3 Causes of Dilated Small Bowel (>3 cm)
2133
Obstruction (transition zone between dilated and nondilated bowel) Adhesions (75% of small bowel obstruction) Postsurgical Postperitonitis Incarcerated hernia Volvulus Extrinsic tumor Congenital stenosis Intraluminal lesion Tumor: usually malignant Intussusception Foreign body Gallstone
ileus
Bezoar Ascaris (bolus of worms) Meconium Muscle dysfunction (no transition zone) Adynamic ileus Surgery Trauma Peritoneal
inflammation
Ischemia Drugs Opiates Barbiturates Anticholinergics Vagotomy Diabetic neuropathy Metabolic disorders Electrolyte imbalance Collagen diseases Scleroderma Dermatomyositis Malabsorption syndromes Celiac disease Chronic
idiopathic
pseudoobstruction
2134
Thickened Folds: Straight and Regular (Table 31.4). Infiltration of edema fluid or hemorrhage into the submucosa results in uniform, straight thickening of the folds (Fig. 31.15). Hemorrhage usually causes thicker folds than edema and may result in scalloping or “thumbprinting― of some folds. Thickened Folds: Irregular and Distorted (Table 31.5). This is the most difficult category of abnormality because many conditions are unusual. The distribution of fold abnormality helps to limit the differential diagnosis (Fig. 31.16) .
FIGURE 31.14. Dilated Small Bowel, Normal Folds. Small bowel follow-through examination reveals dilation of the small bowel lumen (>5 cm between large arrows) with normal thickness of well-defined folds (small arrows). The reason was small bowel obstruction caused by adhesions.
Some conditions are included in several categories. Early Crohn
2135
disease is characterized by edema and regular folds. More advanced Crohn disease has inflammatory cell infiltrate and irregular folds. Lymphoma in the mesentery obstructs lymphatics and causes edema, and lymphoma in the bowel wall causes nodular, irregular folds. Lymphoma and Crohn disease are the two most commonly encountered small bowel diseases. Scleroderma produces atrophy of the muscularis of the small bowel by the process of progressive collagen deposition, resulting in flaccid, atonic, dilated bowel. The valvulae conniventes are normal or thinned (Fig. 31.17). A “hide-bound― appearance of thinned folds tethered together is produced by contraction of the longitudinal muscle layer to a greater extent than the circular muscle layer. Excessive contraction of the mesenteric border of the small bowel results in formation of mucosal sacculations along the antimesenteric border. The jejunum and duodenum are more severely involved than the ileum. The diagnosis is confirmed by skin changes and characteristic eventually
involvement
of
the
esophagus.
Malabsorption
occurs.
Adult celiac disease (nontropical sprue) presents with malabsorption, steatorrhea, and weight loss. Gluten, an insoluble protein found in wheat, rye, oats, and barley, acts as a toxic agent to the small bowel mucosa. The mucosa becomes flattened and absorptive cells decrease in number; villi disappear. The submucosa, muscularis, and serosa remain normal. Patients with long-term sprue have an increased risk of lymphoma and GI carcinoma. The P.841 classic radiographic findings are as follows: (1) dilated small bowel (Fig. 31.18), (2) normal or thinned folds, (3) a decreased number of folds per inch in the jejunum, and (4) an increased number of folds per inch in the ileum (≥5). Findings are best demonstrated by enteroclysis. Detection of five or more folds per inch in the jejunum makes the diagnosis unlikely. Fluid excess is often evident in the ileum. Transient intussusceptions may be observed. Typical CT findings are distal small bowel dilatation, dilution of intraluminal contrast by high intraluminal fluid content, and enlarged lymph nodes in the mesentery (1 5) .
2136
TABLE 31.4 Thickened Small Bowel Folds: Straight and Regular
Intestinal edema (diffuse) Hypoproteinemia Congestive heart failure Portal hypertension Lymphatic obstruction Tumor infiltration (lymphoma) Radiation Fibrosis of the mesentery Lymphangiectasis Zollinger-Ellison syndrome Lactase
deficiency
Intestinal edema (short segment) Crohn disease Eosinophilic
gastroenteritis
Hemorrhage into bowel wall (long segment) Trauma Ischemia Anticoagulant therapy Bleeding disorders Vasculitis Henoch-Schönlein Connective tissue
syndrome disease
Radiation Thromboangiitis obliterans Stomach and small bowel involved Ménétrier disease Zollinger-Ellison syndrome Crohn disease Lymphoma Eosinophilic gastroenteritis
2137
Implies submucosal infiltration by fluid.
Tropical
sprue has similar clinical and radiographic findings as
nontropical sprue but is confined to East Asia and Puerto Rico. The disease responds to administration of folate and antibiotics.
Lactase
Deficiency
Lactase is required within the absorptive cells of the jejunum to properly digest disaccharides. Several population groups, including Chinese, Arabs, Bantu, and Eskimos, may become totally deficient in lactase during adult life. Secondary lactase deficiency may develop with alcoholism, Crohn disease, and use of drugs such as neomycin. The nondigested lactose in the P.842 small bowel causes increased intraluminal fluid and dilated small bowel with normal folds.
TABLE 31.5 Thickened Small Bowel Folds: Irregular and Distorted
Proximal (predominantly duodenum + jejunum) Giardiasis Strongyloides Whipple disease Eosinophilic gastroenteritis Zollinger-Ellison syndrome Distal (predominantly ileum) Lymphoma Crohn disease Yersinia/Campylobacter Salmonella Tuberculosis Behçet disease
2138
Cystic fibrosis AIDS-related infections Diffuse Lymphoma Polyposis syndromes Amyloidosis Histoplasmosis Systemic mastocytosis Waldenström macroglobulinemia Lymphoma Stomach and small bowel involved Lymphoma Crohn disease Eosinophilic
gastroenteritis
Whipple disease Tuberculosis Mastocytosis
FIGURE
31.15. Thickened
Folds,
2139
Regular:
Intestinal
Ischemia. Barium examination demonstrates a striking separation of multiple loops of ileum (arrows), indicating thickening of the bowel walls. The folds in involved loops are thickened and nodular because of edema and hemorrhage resulting from ischemia. A repeat study 1 month later documented complete resolution of all findings. C, colon; J, jejunum.
FIGURE 31.16. Thickened Folds, Irregular: Crohn Disease. Crohn disease of the ileum causes thickened folds (straight arrow) that are irregular and distorted. A more proximal segment of jejunum (open arrow) is effaced and narrowed. The transverse colon (curved arrow) is narrowed and stiffened and has multiple inflammatory polyps producing filling defects. This is an excellent example of the “skip lesions― that are characteristic of
2140
Crohn
disease.
Intestinal ischemia may result from embolism or thrombosis of the superior mesenteric artery or vein (1 6). Patients may present with acute abdomen or vague symptoms. Arterial occlusion may be caused by embolus, vasculitis, trauma, or adhesions. Venous thrombosis results from hypercoagulability states (neoplasms, oral contraceptives); inflammation (pancreatitis, peritonitis, abscess); or stasis (portal hypertension, congestive heart failure). Plain films demonstrate gaseous distention, thickened mucosal folds (“thumbprinting―) (Fig. 31.15), and, in some cases, intramural or portal venous gas. CT is preferred to barium studies to demonstrate characteristic abnormalities: dilated lumen, thickened bowel wall (Fig. 31.19), absent or poor enhancement of the bowel wall, engorged mesenteric vessels, thrombus in the mesenteric arteries or veins, and intramural or venous gas (1 7) . Radiation enteritis occurs when large doses of radiation are given to adjacent organs. The small bowel is the most radiosensitive organ in the abdomen. Long segments of bowel may be involved, with thickening of folds and bowel wall. Peristalsis is impaired. Progressive fibrosis leads to tapered strictures, commonly involving long segments. The bowel may be kinked and obstructed by adhesions. Fistulas to the vagina or other organs may also result. CT demonstrates wall thickening and increased density of the mesentery, along with fixation of bowel loops (Fig. 31.20) .
2141
FIGURE
31.17. Scleroderma. Radiograph from a small bowel
follow-through examination demonstrates dilatation of the jejunum with thin normal folds, an appearance commonly seen with scleroderma. Luminal dilatation is caused by smooth muscle dysfunction in the bowel wall.
Lymphangiectasia refers to gross dilation of the lymphatic vessels in the small bowel mucosa and submucosa. The primary form is a congenital lymphatic blockage, which is often associated with asymmetric edema of the extremities. Despite being congenital, symptoms often do not occur until young adulthood. Patients present with protein-losing enteropathy, diarrhea, steatorrhea, and recurrent
2142
infection. Secondary lymphangiectasia refers to lymphatic obstruction caused by radiation, congestive heart failure, or mesenteric node involvement by malignancy or inflammation. The diagnosis is confirmed by jejunal biopsy. Barium study findings include diffuse fold thickening that is most pronounced in the jejunum, increased intraluminal fluid, and groups of tiny (1-mm) nodules owing to distended villi. The pattern closely resembles Whipple disease. CT helps the differentiation by revealing thickening of the bowel wall and mesenteric adenopathy in secondary lymphangiectasia. P.843 Eosinophilic gastroenteritis virtually always affects the gastric antrum as well as all or part of the small bowel. Intense infiltration of eosinophils in the lamina propria causes thickening of the bowel wall and mucosal folds, often with luminal narrowing. Barium studies show thickened and straightened folds. Thickening of the bowel wall is evidenced by wide separation between bowel loops. CT shows thickened distorted folds in the distal stomach and proximal small bowel (1 5). Most patients have a history of allergic disorders. The disease is self-limited, but recurrences are frequent.
2143
FIGURE
31.18. Adult Celiac Disease. Small bowel follow-
through examination demonstrates dilation of the lumen of the small bowel (4 cm between long arrows). The folds are of normal thickness (small arrows), ie, less than 3 mm. This patient with malabsorption became asymptomatic on a gluten-free diet.
Amyloidosis is a disease complex associated with extracellular infiltration of an amorphous protein material in body tissues. disease may be primary or associated with multiple myeloma to 15%), rheumatoid arthritis (20% to 25%), or tuberculosis Most cases are systemic, but 10% to 20% are localized. The
The (10% (50%). small
bowel is the most common site of GI involvement. Amyloid deposits are seen throughout the wall of the small bowel, especially within the
2144
walls of small blood vessels, resulting in ischemia and infarction. Deposits in the muscularis impair motility. Diffuse, irregular thickened folds may be seen throughout the small bowel. Nodules are sometimes present. CT demonstrates symmetric wall thickening of affected bowel without luminal dilatation or hypersecretion (1 5) . Small mesenteric lymph nodes may be evident. Diagnosis is confirmed by biopsy.
FIGURE 31.19. Intestinal Ischemia. CT demonstrates circumferential thickening of small bowel loops (large arrows) caused by intestinal ischemia owing to portal vein thrombosis. A characteristic, benign, “target― appearance of bowel wall
2145
thickening is evident. The mesentery is edematous and congested (small arrow). c, descending colon; k, kidney.
Systemic mastocytosis is a proliferation of mast cells in the skin, bones, lymph nodes, and GI tract. Urticaria pigmentosa is the characteristic skin manifestation. Osteoblastic bone changes are found in 70% of cases. Lymphadenopathy and hepatosplenomegaly are often present. The bowel wall and mucosal folds are thickened, and mucosal nodules up to 5 mm are often evident (Table 31.6) . Whipple disease is an uncommon systemic disorder affecting the GI tract, joints, CNS, and lymph nodes. The disease is caused by Whipple bacilli—gram-positive, rod-shaped bacteria that are found within macrophages in many organs and tissues. Patients may present with arthritis, neurologic symptoms, or steatorrhea. Generalized lymphadenopathy is usually present. Enteroclysis P.844 demonstrates irregularly thickened folds that are most prominent in the jejunum. Demonstration of tiny (1-mm) sandlike nodules spread diffusely over the mucosa or in small groups is strong evidence of the disease. Increased luminal fluid is usual. CT reveals thick folds, especially in the jejunum, without significant dilatation. Low-density or fat-density nodes in the mesentery are characteristic (1 5) .
2146
FIGURE
31.20. Radiation
Enteritis. CT image through the
pelvis in a patient with cervical carcinoma treated with radiation reveals long segments of small bowel (arrows) with wall thickening and infiltrated mesentery.
AIDS
Enteritis. In addition to lymphoma and Kaposi sarcoma, AIDS
patients are predisposed to multiple opportunistic infections of the GI tract. Infective agents usually occur in combination and in multiple GI sites. Cryptosporidium and Isospora
belli are protozoans that may infest
the proximal intestine and cause a cholera-like diarrhea, with lifethreatening fluid loss. Barium studies show thickened folds and marked increased fluid. Cytomegalovirus causes disease in the small bowel and colon as well as the lungs, liver, and spleen. Mucosal ulceration with bleeding and perforation are the major intestinal manifestations. Barium studies show thickened folds, loop separation, ulcers, and fistulae.
2147
TABLE 31.6 Tiny Small Bowel Nodules
Nodular lymphoid hyperplasia (2–4 mm) Lymphoma (>4 mm) Amyloidosis Whipple disease (1–2 mm) Mycobacterium avium-intracellulare Lymphangiectasia Systemic mastocytosis ( Table of Contents > Section VII - Gastrointestinal Tract > Chapter 32 - Colon and Appendix
Chapter
32
Colon and Appendix William E. Brant
COLON Imaging
Methods
The primary imaging methods for detection and characterization of colon abnormalities have continued to evolve over time. The persistently expanding availability of colonoscopy has continued to reduce the role of barium enema in imaging the colon. On the other hand, the use of CT to image the abdomen and pelvis continues to increase, making CT often the preferred method for initial detection of colon disease. CT (virtual) colonography challenges the role of traditional colonoscopy for polyp and cancer detection. Once a possible neoplastic lesion is discovered, however, colonoscopy or proctoscopy is usually used for biopsy. The single-contrast barium enema is still occasionally used for the evaluation of colonic obstruction and fistulas and in old, seriously ill, or debilitated patients. The double-contrast (air-contrast) barium enema (Fig. 32.1) is favored for detection of small lesions (10 mm in short axis) as the primary criterion for involvement. This is inherently inaccurate, because cervical cancer is known to involve nodes without enlarging them. Central necrosis within a lymph node is highly predictive of tumor involvement, regardless of node size. Lymphatic spread involves internal and external iliac, presacral, and paraaortic nodes. Distant metastases most commonly involve liver, lung, and bone. Imaging studies should include the kidneys to assess for obstruction.
TABLE 35.2 Cervical Cancer Staging (FIGO)
Stage
Description
0
Carcinoma in situ
I
Tumor confined to cervix
Ia
Preclinical invasive carcinoma diagnosed by microscopy only; invasion no deeper than 5 mm and
2342
no wider than 7 mm
Ib
II
Clinical lesions confined to cervix
Tumor invades beyond cervix but not to pelvic wall or lower third of vagina
IIa
Without
IIb
With
III
parametrial
parametrial
invasion
invasion
Tumor extends to pelvic wall and/or involves lower third of vagina and/or causes hydronephrosis
IIIa
No extension to pelvic side wall
IIIb
Extension to pelvic sidewall or hydronephrosis
IV
IVa
Tumor invades mucosa of bladder or rectum and/or extends to pelvic side walls
IVb
Distant
metastases
FIGO, International Federation of Gynecology and Obstetrics 1995.
Montreal.
2343
FIGURE
35.7. Cervical Carcinoma, Stage IA, on MR. This
T2WI was obtained in an oblique coronal plane to the patient to image the cervix in transverse orientation. The tumor (T), appearing dark gray, has nearly completely replaced the normal cervix, seen only as a black rim (arrowheads). No parametrial invasion is evident. Free intraperitoneal fluid (ff) is seen in the cul-de-sac. B, bladder.
2344
FIGURE
35.8. Cervical Carcinoma, Stage IIb, on CT.
Heterogeneous tumor (T) has completely replaced the cervix on this CT scan. Stranding densities (arrowheads) into the paracervical fat indicate parametrial invasion by tumor.
Endometrial
carcinoma is now the most common invasive
gynecologic malignancy. Histologically, it is 95% adenocarcinoma and 5% sarcoma. The peak age at onset is 55 to 62 years, with postmenopausal vaginal bleeding as the key symptom. The tumor spreads initially by invasion into the myometrium and cervix, followed by lymphatic spread to the pelvic and retroperitoneal nodes, then continued direct spread into the broad ligaments, parametrium, and ovaries. Peritoneal seeding will occur with penetration of the uterine serosa. Hematogenous spread to the lung, bone, liver, and brain occurs late in the course of the disease. Prognosis and treatment depend on stage of the disease (Table 35.3), with the most critical factors being the depth of myometrial invasion and the involvement of lymph nodes. Lymph node metastases are unlikely if myometrial invasion is less than 50%. MR
2345
staging is more accurate than CT staging (1 3). On MR, the signal from tumor is similar to that of endometrium (1 4). Tumor is isointense to myometrium on T1WIs and hyperintense to myometrium on T2WIs (Fig. 35.9). Evidence of tumor includes thickening and poor definition of the endometrium. Large tumors appear as a polypoid mass that expands the uterine cavity. Tumor enhancement with gadolinium is variable and may be less than or greater than enhancement of P.915 myometrium and endometrium. Invasion of myometrium is determined on postcontrast T2WIs. An intact junctional zone myometrium is evidence of the absence of myometrial invasion (stage Ia). Pitfalls for myometrial invasion include thinning of the myometrium by rapidly expanding tumors. Cervical invasion is determined on sagittal T2WIs and postcontrast sequences, with enhancing tumor seen within the dark tissue of the cervix. T1WIs show parametrial invasion into fat. Invasion of the bladder and rectum is evidenced by disrupted tissue planes and tumor signal with bladder or rectal wall on T2WIs. On CT, the depth of myometrial invasion is determined on postcontrast images. The tumor enhances less than myometrium. Obstruction of the cervix results in filling of the uterine cavity with fluid of variable density (Fig. 35.10). Cervical involvement (stage II disease) appears as heterogeneous enlargement of the cervix. Parametrial invasion appears as irregular margins of the uterus, parametrial soft tissue stranding, or parametrial mass. CT and MR evidence of nodal metastases are lymph nodes larger than 10 mm in short axis.
TABLE 35.3 Endometrial Cancer Staging (FIGO)
Stage 0
Description Carcinoma in situ
2346
I
Ia
Tumor limited to endometrium
Ib
Tumor invasion of 50% thickness of myometrium
IIa
Tumor invades cervical mucosa
IIb
Tumor invades cervical mucosa and stroma
II
III
IIIa
Tumor invades uterine serosa and/or adnexa, and/or
positive
peritoneal
cytology
IIIb
Metastases to vagina
IIIc
Metastases to pelvic and/or paraaortic lymph nodes
IV
IVa
Tumor invades bladder and/or bowel mucosa
IVb
Distant metastases, including and/or inguinal lymph nodes
intra-abdominal
FIGO, International Federation of Gynecology and Obstetrics.
2347
FIGURE 35.9. Endometrial Carcinoma, Stage Ic, on MR. Axial-plane T2WI with fat saturation shows endometrial carcinoma (T) invading more than 50% of the thickness of the myometrium (arrow). This tumor is distinctly high signal compared to myometrium.
2348
FIGURE 35.10. Endometrial Carcinoma, Stage Ib, on CT. Postcontrast CT image shows enhancing tumor nodules (arrows) outlined by low-attenuation hemorrhagic fluid (H) within the uterine cavity. Tumor invasion is difficult to assess because the tumor is nearly isointense with enhanced myometrium. The stage was proven to be Ib at surgery.
Uterine
sarcomas are the most aggressive of the uterine tumors
(1 5). Sarcomas may be suspected when uterine masses are large and heterogeneous. Malignant mixed müllerian tumors are large solid tumors with prominent necrosis and hemorrhage that expand the uterine cavity and invade the myometrium. Lymphatic and peritoneal spread are common. Leiomyosarcomas usually present as a rapidly growing pelvic mass. The uterus is enlarged with a markedly heterogeneous mass with extensive necrosis, hemorrhage, and frequently calcifications (Fig. 35.11). Imaging differentiation from a degenerated benign leiomyoma is not possible unless signs of malignant spread of tumor are evident. Endometrial stromal sarcomas appear as polypoid endometrial masses that invade the
2349
myometrium.
FIGURE 35.11. Leiomyosarcoma. T2WI shows a huge heterogeneous tumor mass (arrowheads) arising from the anterior wall of the retroflexed uterus (arrow). Note that the uterine cavity is intact. The exophytic myometrial origin and heterogeneity of the mass are indicative of either a degenerated leiomyoma or a leiomyosarcoma. The latter diagnosis was confirmed
at
surgery.
P.916
Benign
Conditions
Leiomyomas are the most common uterine tumor, affecting 50% of women of reproductive age. Most women are asymptomatic, but the
2350
tumors may cause excessive bleeding, pelvic pain, mass symptoms, and infertility. Tumors are benign and made up of smooth muscle and a variable amount of fibrous tissue. Tumors with scant fibrous tissue enhance brightly, while those with abundant fibrous tissue enhance poorly. Most tumors are intramural (within the myometrial wall), while others are submucosal (beneath the endometrium) or subserosal (beneath the serosa). Subserosal or submucosal tumors may be pedunculated on long stalks. Submucosal tumors are prone to ulcerate, resulting in severe menorrhagia. MR provides the best characterization of size, number, and location (1 6). Leiomyomas are usually low signal compared to myometrium on both T1WIs and T2WIs, although visualization is best on T2WIs (Fig. 35.12). Areas of degeneration and cystic change cause inhomogeneous high internal signal. The tumors are well demarcated from adjacent myometrium by a discrete rim of low signal. Contrast enhancement does not improve leiomyoma detection or characterization. On CT, leiomyomas appear as homogeneous or heterogeneous masses that may be hypodense, isodense, or hyperdense relative to enhanced myometrium. Coarse calcifications within the mass are common and characteristic (Fig. 35.13). Cystic degeneration produces interior low density. Diffuse enlargement of the uterus and lobulation of its contour are common. Pedunculated leiomyomas may appear as adnexal rather than uterine masses.
2351
FIGURE
35.12. Multiple
Leiomyomas. A midsagittal T2WI of
the pelvis demonstrates multiple leiomyomas (L), which greatly enlarge and distort the uterus. The endometrial cavity (e) of the uterus and the cervix (c) are clearly demonstrated. B, bladder; V, vagina.
2352
FIGURE 35.13. Leiomyoma Calcifications. A radiograph of the pelvis obtained as part of an excretory urogram demonstrates a leiomyoma (L) causing a mass impression on the bladder (B). Multiple characteristic “popcorn― calcifications (open arrows) are evident. u, ureters.
2353
FIGURE 35.14. Adenomyosis. Sagittal T2WI shows marked widening of the junctional zone myometrium (between arrowheads), a key finding of adenomyosis. Small cysts seen as bright round foci (long arrow) are also characteristic.
P.917 Adenomyosis is a benign disease of the uterus characterized by the presence of ectopic endometrial glands and stroma within the myometrium, eliciting surrounding myometrial hypertrophy (1 7) . Patients present with dysmenorrhea or menorrhagia. The disease may be focal or diffuse. MR provides the best detection of the disease (1 8). Diffuse disease is indicated by regular or irregular thickening of the junctional zone myometrium >12 mm. The low signal abnormality corresponds to myometrial hypertrophy. Half of patients also demonstrate high-signal foci within the myometrium corresponding to islands of endometrial glands with cystic change or hemorrhage (Fig. 35.14). Focal disease is evidenced by low-signal
2354
masses within the myometrium on T2WIs. These masses are isointense to myometrium on T1WIs. High-signal foci, occasionally seen on T1WIs, represent hemorrhage. Differentiation from leiomyomas is difficult. Leiomyomas are characteristically well circumscribed, while adenomyomas are poorly defined with vague margination. Adenomyosis is not routinely evident on CT. US findings are commonly subtle and nonspecific. Nabothian cysts are retention cysts of the mucous-secreting glands of the cervical epithelium. They are common and are seen on MR as bright, round, well-defined structures in the cervix on T2WIs. On T1WIs, they are isointense to urine or muscle. Physiologic
ovarian
cysts contain simple fluid that is low signal on
T1WIs and high signal on T2WIs. A uniform, thin, dark wall is evident on T2WIs. Gadolinium enhancement of the cyst wall is common but not constant. On CT they are well defined, thin walled, and have homogeneous internal density near water. Size under 2.5 cm is indicative of physiologic ovarian follicle.
2355
FIGURE 35.15. Hemorrhagic Follicular Cysts. Fat-suppressed T1WI shows high signal in two left ovarian cysts (arrows) , indicating internal hemorrhage. The cysts are well-defined, homogeneous, and lack any solid component. U, uterus; B, bladder.
Hemorrhagic functional ovarian cysts appear high signal on T1WIs if a large amount of methemoglobin is present (Fig. 35.15). If predominantly intact red blood cells are present, the cyst appears low signal on T2WIs. Thus, hemorrhagic cysts may be low signal on both T1WIs and T2WIs, high signal on T1WIs and low signal on T2WIs, or low signal on T1WIs and high signal on T2WIs. Layering of blood products may be present. The absence of gadolinium enhancement differentiates blood clot that is adherent to the cyst wall from clot that has formed a solid nodule. On CT, hemorrhagic cysts appear as thin-walled cysts with internal density near water or higher, depending on the physical state of the blood products. Atypical cysts can be followed with US to determine if they resolve after one or two menstrual cycles. Endometriosis is the presence of endometrial tissue outside of the uterus (1 9). The endometrial implants respond to hormonal stimulation, resulting in recurrent bleeding, inflammation, and fibrosis. Hallmarks of disease include numerous tiny implantations of endometrial tissue on peritoneal surfaces, development of endometriomas (endometrial cysts filled with hemorrhage), and formation of adhesions between surrounding tissues. The most P.918 common sites of involvement are the ovaries, the cul-de-sac, and peritoneal reflections over the uterus, fallopian tubes, bladder, and rectosigmoid
colon.
2356
FIGURE
35.16. Endometrioma on MR. A. T1WI. B . T2WI. A
cystic mass (arrows) in the cul-de-sac is high signal on the TIWI and shows characteristic loss of signal on the T2WI (T2 shading). The loss of signal is caused by the presence of methemoglobin within the cyst resulting from multiple episodes of internal hemorrhage.
All imaging modalities have high sensitivity for detection of endometriomas, but they all lack the ability to reliably detect tiny endometrial implants, which are commonly smaller than 3 mm. Endometriomas (“chocolate cysts―) contain blood products of various ages, reflecting recurrent episodes of bleeding and corresponding to the menstrual cycle. They are characteristically multiple and bilateral. MR shows the cysts to be homogeneous high intensity on T1WIs and characteristically low signal on T2WIs, a finding termed “T2 shading― (Fig. 35.16). Loss of signal on T2WIs is caused by the presence of methemoglobin within the cysts (2 0). Iron concentration and viscosity increase within the cysts as water is resorbed. Cysts may appear heterogeneous because of the varying age of contained blood products. The cyst wall is usually low in signal, representing fibrous tissue or hemosiderin. Fat-saturation T1WIs improve visualization of small implants on peritoneal surfaces.
2357
On CT, endometriomas appear as complex cystic pelvic masses, frequently with relatively high-density fluid components. Inflammation and fibrosis are prominent. Multiple pelvic organs may be incorporated into a mass. Hydrosalpinx is a common associated finding (30%). Hydrosalpinx is a common finding on HSG performed for infertility (Fig. 35.17). Occlusion of the fallopian tube, caused by infection, surgery, or endometriosis, results in fluid accumulation and dilatation of the tube. The most common cause is pelvic infection. CT and MR demonstrate hydrosalpinx as a sausage-, C-, or S-shaped adnexal structure distended with fluid of variable character. Pelvic
inflammatory
disease is a common affliction of women of
reproductive age. Endometritis and myometritis are treated medically. Imaging is performed to detect tubo-ovarian abscess and pyosalpinx, complications that are treated surgically. Early findings include pelvic edema P.919 and stranding in the parametrium and paraovarian tissues. Pyosalpinx appears as a thick-walled hydrosalpinx that contains complex fluid. Tubo-ovarian abscess appears as a thick-walled, fluidfilled adnexal mass that incorporates the ovary and commonly a dilated fallopian tube (2 1). Gas bubbles are occasionally present within the collection and are highly indicative of abscess.
2358
FIGURE 35.17. Hydrosalpinx. Hysterosalpingography demonstrates a retroflexed uterus (U), with the fundus (f) directed posteriorly and inferiorly. The left fallopian tube is occluded at the isthmus (black arrow). The right fallopian tube (open arrow) is massively dilated at its distal end, forming a hydrosalpinx (HS). Occlusion of the right fallopian tube is confirmed by the absence of peritoneal spill. The curved arrow indicates the cervical cannula.
Benign
cystic
teratoma, or dermoid cyst, is the most common
germ cell neoplasm of the ovary (2 2). Lesions contain mature elements derived from ectoderm, mesoderm, or endoderm, resulting in a broad range of appearances. Mean patient age at discovery is 30 years. Most lesions are discovered incidentally while the patients are asymptomatic. The cysts are filled with liquid sebaceous material that is fat density on MR and CT. Internal contents include the Rokitansky nodule, which commonly includes hair, teeth, bone, or cartilage. US features are usually characteristic, but lesions may be discovered or further characterized by MR or CT. MR shows the sebaceous material as very high intensity on T1WIs. Signal is usually decreased on T2WIs, approximating fat signal. Fat content is confirmed by in-phase and out-of-phase gradient recall images or frequency-selective fat-saturation images. CT demonstration of fat density within a cystic adnexal mass is definitive (Fig. 35.18). CT and plain radiographs show bone and tooth formation within the mass (Fig. 35.19) . Fibrotic
neoplasms account for 4% of ovarian tumors (1). Because
they are solid masses and are commonly (40% of cases) associated with ascites, they may mimic ovarian cancers. Tissue types include fibromas, thecomas, and fibrothecomas arising from ovarian stroma. MR shows a well-defined ovarian mass that is predominantly low signal on both T1WIs and T2WIs. Scattered high-signal areas within the mass on T2WIs represent focal edema or cystic change.
2359
FIGURE
35.18. Benign
Cystic
Teratoma. CT performed without
contrast reveals a fat-density mass (arrow) in the pelvis of a 28year-old woman. The appearance is diagnostic of a benign cystic teratoma.
FIGURE 35.19. Benign Cystic Teratoma. A plain radiograph of the pelvis in a young woman demonstrates several well-formed
2360
teeth (black arrow). A subtle, well-defined mass of fat density is also present (white arrows). These findings are diagnostic of benign cystic teratoma.
Adnexal
torsion may be initially diagnosed, or confirmed, by CT or
MR examination (2 3). Key findings include a smooth-walled adnexal mass, which serves as a nidus for twisting. The torsed mass demonstrates concentric wall thickening. The ipsilateral fallopian tube appears as an amorphous mass or tube with thickened walls. The uterus is deviated toward the torsed adnexa. Signs of hemorrhagic infarction of the torsed adnexa include marked thickening of the wall of the adnexal mass (>10 mm), hemorrhage within the mass and within the twisted tube, and hemoperitoneum.
MALE Testes
GENITAL and
TRACT
Scrotum
US, supplemented by color Doppler, is the imaging method of choice to evaluate the testes and scrotal contents. MR using surface coils offers excellent spatial resolution, greater tissue contrast, and wider field of view but has the disadvantages of greater cost and lesser availability.
Radionuclide
imaging
provides
useful
information
about
perfusion but with limited anatomic detail. CT is useful in the staging of testicular tumors and in locating undescended testes that are not found by US. This chapter reviews MR and CT imaging. US imaging of scrotal contents is reviewed in Chapter 37.
Normal
MR
Anatomy
Because of its high fluid content, the testes are of uniform intermediate signal on T1WIs and uniform high signal—slightly less than water—on T2WIs (Fig. 35.20). The tunica albuginea forms a well-defined P.920 1-mm-thick rim. Septations are often visualized radiating from the
2361
mediastinum to the tunica albuginea. A small amount of fluid is normally present in the scrotum between the layers of the tunica vaginalis. The epididymis is isointense to the testes on T1WIs and brightens on T2WIs, though to a lesser extent than the testis. The scrotum is intermediate signal, reflecting the dartos muscle. The spermatic cord appears as numerous tubular structures, representing the arteries and veins, with MR signal determined by blood flow.
FIGURE 35.20. Normal MR Anatomy: Male. Coronal T2WI shows both testes (T) and the penis in cross section. The testes are high in signal because of their high fluid content. The epididymis (e) is also high signal on T2WIs but less than that of the testes. The paired corpora cavernosa (white arrows) are well demonstrated. The corpus spongiosum contains the urethra (black arrow) .
Undescended
Testis
CT and MR are used to localize undescended testes not demonstrated
2362
by US to be within the inguinal canal. The testis, if present, will be seen between the lower pole of the kidney and the internal inguinal ring. In 3% to 5% of cases, the testis is congenitally absent. The undescended testis appears as an oval soft tissue mass up to 4 cm in size. Because the undescended testis is usually atrophic, MR may show low or intermediate signal instead of high signal on T2WIs.
Neoplasms Germ cell tumors, stromal tumors, lymphoma, and leukemia all appear on T2WIs as low-signal areas of tumor within the high-signal normal testicular parenchyma (2 4). Normal parenchymal septations are disrupted by the tumor. High-signal areas within tumors correspond to areas of hemorrhage. MR cannot reliably differentiate benign from malignant tumors. CT and MR are used for staging of tumors. Lymphatic spread of tumor is most common, with a usual pattern of orderly ascending nodal involvement. Initial spread is along gonadal lymphatic vessels, following the testicular veins to the renal hilar nodes. Lymphatic metastases may also follow the external iliac chain to the paraaortic nodes. The internal iliac and inguinal nodes are rarely involved. Extensive metastatic involvement of the lymph nodes mimics lymphoma in young males. A primary tumor in the testis may be clinically occult, yet is effectively demonstrated by US. Hematogenous spread to the lungs usually follows lymphatic spread, except in choriocarcinoma, which spreads hematogenously early.
Scrotal
Fluid
Collections.
Simple hydroceles show signal characteristics of water: low signal on T1WIs and high signal on T2WIs (2 5). Hematoceles and pyoceles show high signal on T1WIs, reflecting complex fluid or high protein content. Epididymal cysts show the signal of simple fluid. Spermatoceles commonly contain fat and protein, causing high signal on T1WIs, and layering debris may be evident. Varicoceles appear as serpiginous tubular structures in the spermatic cord. Signal intensity corresponds to slow blood flow.
2363
Epididymitis/Orchitis. Orchitis causes inhomogeneous signal on both T1WIs and T2WIs, indistinguishable from tumor. With epididymitis, the epididymis is enlarged, but signal intensity on T2WIs is unpredictable and may be increased, decreased, or normal. Dilated vessels in the spermatic cord reflect hypervascularity. Hydrocele is usually present. Testicular torsion is best evaluated with Doppler US or scintigraphy. With acute torsion, MR may demonstrate a characteristic twisted pattern of torsion of the spermatic cord, with impaired blood flow evident. The testis appears heterogeneous on all image sequences.
Prostate
and
Seminal
Vesicles
No imaging modality can reliably demonstrate the presence or absence of cancer in the prostate. That diagnosis relies on biopsy, which is best performed using transrectal US for guidance. MR with endorectal coils and transrectal US offer the best promise for staging of local disease. Either CT or MR may be used to demonstrate evidence of nodal and distant tumor spread.
Normal
MR
Anatomy.
The prostate is divided into three glandular zones surrounding the urethra (Fig. 35.21). The peripheral zone contains approximately 70% of the prostate tissue and is draped around the remainder of the gland like a glove holding a baseball. Most prostate cancers (70%) arise in the peripheral zone. The transitional zone consists of two small areas of periurethral glandular tissue. Although it contains only 5% of the prostatic tissue in the normal young man, it is the site of benign prostatic hypertrophy and may enlarge greatly in the older man. The central zone consists of the glandular tissue at the base of the prostate, through which course the ducts of the vas deferens and seminal vesicles and the ejaculatory ducts. P.921 Although the central zone makes up 25% of prostate glandular
2364
tissue, only 10% of cancers arise there. The anterior portion of the prostate is occupied by nonglandular tissue called the anterior fibromuscular stroma. The base of the prostate is that portion adjacent to the base of the bladder and the seminal vesicles. The apex of the prostate rests on the urogenital diaphragm.
FIGURE 35.21. Zonal Anatomy of the Prostate. The anatomy is illustrated in the midsagittal plane (left) and the axial plane (right) at the level of the vertical dashed line on the left.
The seminal vesicles are symmetrically sized, lobulated, teardropshaped, coiled ducts that occupy the groove between the base of the bladder and the base of the prostate posteriorly. Prominent veins are frequently visualized in the periprostatic tissues. Lymphatic drainage of the prostate goes to regional pelvic lymph nodes, with channels to paraaortic and inguinal nodes. Periprostatic venous connections to vertebral veins offer a route for the hematogenous spread of tumor to the axial skeleton. On T1WIs, the prostate gland is uniform intermediate to low signal, similar to skeletal muscle. The high-signal periprostatic fat defines the margin of the prostate. Periprostatic veins and neurovascular bundles are low signal. T2WIs demonstrate the internal structure (zonal anatomy) of the prostate (Fig. 35.22). The peripheral zone is
2365
high in signal because of its higher water content and looser acinar structure. The central zone is lower in signal because of its more compact muscle fibers and acinar structure. The central and transitional zones become heterogeneous with age and the development of benign prostatic hyperplasia. The anterior fibromuscular stroma is low in signal and has poorly defined margins. The seminal vesicles are low to intermediate signal on T1WIs and brighten greatly on T2WIs because of fluid within the tubules. The normal size of the seminal vesicles varies widely, and slight asymmetry is common (2 6) .
Normal
CT
Anatomy
The prostate gland is seen at the base of the bladder, just posterior to the symphysis pubis, as a homogeneous rounded soft tissue organ up to 4 cm in maximal diameter. Prostate zonal anatomy is not demonstrated by CT. A well-defined plane of fat separates the prostate from the obturator internus. The paired seminal vesicles produce a characteristic “bowtie―-shaped soft tissue structure in the groove between the bladder base and the prostate. Prostate
carcinoma is the third leading cause of cancer death in
men. Approximately 10% of men over age 50 will develop clinical prostate carcinoma in their lifetime. Despite the high prevalence and importance of prostate disease, treatment remains extremely controversial. One of the difficulties of dealing with prostate cancer is differentiating tumors with biologic aggressiveness from those that are incidental findings. Nearly 50% of men older than 75 years of age will have prostate carcinoma on biopsy or autopsy. However, many of these cancers will not affect the patient's lifespan. The tumor is uncommon before P.922 age 50 and increases in incidence thereafter. The Gleason histologic grading system is used to assess the degree of differentiation of the tumor. A grade 1 is well differentiated, and a grade 5 is anaplastic. The Gleason score varies from 2 to 10 and adds the Gleason grade for the predominant and the secondary portions of the tumors. Tumor staging is by the American Urological Association system
2366
(Table 35.4) .
FIGURE
35.22. Normal Prostate on MR. Axial plane T2WI of a
normal prostate in a 40-year-old man demonstrates the highintensity peripheral zone (arrowheads), the urethra (long arrow) , and the surrounding lower-intensity transitional zone. B, bladder; r, rectum; oi, obturator internus muscle.
TABLE 35.4 Prostate Cancer Staging: 1992 Revision of TNM Classification
Stage Primary
Description tumor
2367
TX
Not
assessable
T0
Not
evident
T1
Clinically apparent, not palpable or visible by imaging
T1a
Found incidentally in 5% or less of tissue resected
T1b
T1c
Found incidentally in more than 5% of tissue resected
Identified by needle biopsy because of elevated PSA
T2
T2a
T2b
T2c
T3
Palpable or visible by imaging
Involves half of one lobe or less
Involves more than half of one lobe, but not both lobes
Involves
Extends
both
lobes
through
prostate
T3a
Found
unilaterally
T3b
Found
bilaterally
2368
capsule
T3c
T4
Invades
seminal
vesicles
Fixed, or invades other structures
T4a
Invades bladder neck, external sphincter, or rectum
T4b
Invades levator muscles or is fixed to pelvic wall
Regional lymph nodes
NX
Regional lymph nodes cannot be assessed
N0
No regional lymph nodes metastases
N1
Metastases to single lymph node 2 cm or smaller in
N2
greatest
dimension
Metastases to single lymph node 2 to 5 cm in greatest dimension, or to multiple lymph nodes all 5 cm in greatest
Distant
dimension
metastases
MX
Metastases cannot be assessed
M0
No
distant
metastases
2369
M1
Distant
metastases
present
Adapted from Schroeder FH, Hermanek P, Denis L, et al. The TNM classification of prostate cancer. Prostate 1992;4(suppl):129–138.
Most tumors are adenocarcinoma (95%). Prostate cancer spreads by local extension, from lymphatic vessels to regional nodes, and by hematogenous dissemination. Penetration of tumor through the capsule or into the seminal vesicles greatly worsens the prognosis. Involvement of the axial skeleton by hematogenous metastases is common. Metastases to the lungs, liver, and kidneys occur in the terminal phases of the disease.
FIGURE 35.23. Prostate Carcinoma. Proton density–weighted axial plane MR image demonstrates a lowintensity prostate carcinoma (large open arrow) in the peripheral zone (curved arrows). The tumor is confined to the prostate
2370
gland and measures approximately 2 cm. The urethra (long arrow) and dark transitional zone (small open arrow) are evident. r, rectum.
On MR T2WIs, cancers appear as areas of low signal within the highsignal peripheral zone (Fig. 35.23). Cancer is isointense with prostate tissue on T1WIs, which are best used for assessing invasion of periprostatic fat and for detecting nodal involvement. Recent biopsy limits the specificity of MR because areas of hemorrhage may mimic tumor. Criteria for extracapsular extension of tumor include: (1) asymmetry of neurovascular bundles, (2) tumor envelopment of neurovascular bundle, (3) angulated contour of the prostate gland, (4) irregular, spiculated margins of the prostate gland, and (5) obliteration of the retroprostatic angle (2 6). CT is limited to the demonstration of adenopathy and distant spread of tumor, because it cannot differentiate tumor from benign hyperplasia within the gland. Some (~50%) cancers are detectable as a focus of contrast enhancement in the peripheral zone on multidetector CT. Benign prostatic hyperplasia begins at approximately age 40 and eventually occurs in all men. Hypertrophy and hyperplasia occur in glandular tissue in the transitional and periurethral zones, accompanied by proliferation of supporting smooth muscle and stromal cells. The end result is focal or diffuse enlargement of the prostate. Pressure on the urethra obstructs bladder outflow and results in symptoms of hesitancy, decreased force and caliber of the urine stream, dribbling, frequency, nocturia, and postvoid residual. This progressive process is combatted by hypertrophy of the bladder wall musculature. Advanced symptoms require medical therapy balloon P.923 dilatation, stents, or transurethral resection. CT findings include: (1) enlargement of the prostate, commonly with lobulated contour and visible high- and low-density nodules; (2) coarse calcifications; (3) cystic degeneration; and (4) bladder wall thickening and trabeculation. MR shows prostate enlargement with heterogeneous
2371
central gland on T2WIs (Fig. 35.24). Areas of cystic degeneration are low signal on T1WIs and high signal on T2WIs.
FIGURE
35.24. Benign
Prostatic
Hypertrophy. T2WI in the
axial plane shows marked diffuse enlargement of the prostate gland (arrows) with heterogeneous signal and cystic change. The normal zonal anatomy of the prostate is not evident. B, bladder.
REFERENCES 1. Jeong YY, Outwater EK, Kang HK. Imaging evaluation of ovarian masses. Radiographics 2000;20:1445–1470. 2. Siegelman ES, Outwater EK. Tissue characterization in the female pelvis by means of MR imaging. Radiology 1999;212:5–18.
2372
3. Saini A, Dina R, McIndoe GA, et al. Characterization of adnexal masses with MRI. AJR Am J Roentgenol 2005;184:1004–1009. 4. Berridge DL, Winter TC. Saline infusion sonohysterography: technique, indications, and imaging findings. J Ultrasound Med 2004; 23:97–112. 5. Saksouk FA, Johnson SC. Recognition of the ovaries and ovarian origin of pelvic masses with CT. Radiographics 2004;24:S133–S146. 6. Outwater EK, Talerman A, Dunton C. Normal adnexa uteri specimens: anatomic basis of MR imaging features. Radiology 1996;
201:751–755.
7. Jung SE, Lee JM, Rha SE, et al. CT and MR imaging of ovarian tumors with emphasis on differential diagnosis. Radiographics 2002;22:1305–1325. 8. Troiano RN, McCarthy SM. Müllerian duct anomalies: imaging and
clinical
issues.
Radiology
2004;233:19–34.
9. Woodward PJ, Hosseinzadeh K, Saenger JS. Radiologic staging of ovarian carcinoma with pathologic correlation. Radiographics 2004;24:225–246. 10. Pannu HK, Corl FM, Fishman EK. CT evaluation of cervical cancer: spectrum of disease. Radiographics 2001;21:1155–1168. 11. Kaur H, Silverman PM, Iyer RB, et al. Diagnosis, staging, and surveillance of cervical carcinoma. AJR Am J Roentgenol 2003;180:1621–1632.
2373
12. Okamoto Y, Tanaka YO, Nishida M, et al. MR imaging of the uterine cervix: imaging-pathologic 2003;23:425–445.
correlation.
Radiographics
13. Kinkel K, Kaji Y, Yu KK, Segal MR. Radiologic staging in patients with endometrial cancer: a meta-analysis. Radiology 1999;212:711–718. 14. Lee EJ, Byun JY, Kim B, et al. Staging of early endometrial carcinoma: assessment with T2-weighted and gadoliniumenhanced T1-weighted 1999;19:937–945.
MR
imaging.
Radiographics
15. Rha SE, Byun JY, Jung SE, et al. CT and MRI of uterine sarcomas and their mimickers. AJR Am J Roentgenol 2003;181:1369–1374. 16. Murase E, Siegelman ES, Outwater EK, et al. Uterine leiomyomas: histopathologic features, differential diagnosis, treatment. Radiographics 1999;19:1179–1197. 17. Tamai K, Togashi K, Ito T, et al. MR imaging findings of adenomyosis: correlation with histopathologic features and diagnostic pitfalls. Radiographics 2005;25:21–40. 18. Byun JY, Kim SE, Choi BG, et al. Diffuse and focal adenomyosis: MR imaging findings. Radiographics 1999;19:S161–S170. 19. Woodward PJ, Sohaey R, Mezzetti TP, Jr. Endometriosis: radiologic-pathologic correlation. Radiographics 2001;21:193–216. 20. Gougoutas CA, Sigelman ES, Hunt J, Outwater EK. Pelvic
2374
and
endometriosis: various manifestations and MR imaging findings. AJR Am J Roentgenol 2000;175:353–358. 21. Sam JW, Jacobs JE, Birnbaum BA. Spectrum of CT findings in acute pyogenic pelvic inflammatory disease. Radiographics 2002;22:1327–1334. 22. Outwater EK, Siegelman ES, Hunt JL. Ovarian teratomas: tumor types and imaging 2001;21:475–490.
characteristics.
Radiographics
23. Rha SE, Byun JY, Jung SE, et al. CT and MR imaging features of adnexal torsion. Radiographics 2002;22:283–294. 24. Woodward PJ, Sohaey R, O’Donoghue MJ, Green DE. Tumors and tumorlike lesions of the testes: radiologic-pathologic correlation. Radiographics 2002;22:189–216. 25. Woodward PJ, Schwab C, Sesterhenn IA. Extratesticular scrotal masses: radiologic-pathologic correlation. Radiographics 2003;23: 215–240. 26. Claus FG, Hricak H, Hattery RR. Pretreatment evaluation of prostate cancer: role of MR imaging and 1H MR spectroscopy. Radiographics 2004;24:S167–S180.
2375
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section IX - Ultrasonography > Chapter 36 Abdomen Ultrasound
Chapter
36
Abdomen
Ultrasound
William E. Brant Ultrasound (US) is firmly established as a primary imaging modality for comprehensive evaluation of the abdomen, including the abdominal organs, the peritoneal cavity, and the retroperitoneum (1). Its role includes screening for disease, evaluation and follow-up of known abnormalities, and guidance of biopsy, aspiration, and catheter drainage procedures. Comprehensive examination commonly includes the use of Doppler and color flow imaging, as well as specialized techniques of transvaginal or transrectal US to demonstrate pelvic extension of disease. This chapter provides the basic information needed to understand the effective use of US in examining the abdomen.
PERITONEAL Normal
US
CAVITY Anatomy
The normal peritoneal cavity is a potential space best appreciated when fluid is present (2). The peritoneal membrane lines the abdominal cavity and covers—in whole or in part—the intraabdominal organs. Numerous peritoneal ligaments, folds, and recesses are visualized when outlined by fluid within the peritoneal cavity. US examination for the presence of fluid includes inspection of the subdiaphragmatic and subhepatic regions, the pericolic
2376
gutters, and the pelvic cul-de-sac. Tiny volumes of intraperitoneal fluid are best detected by transvaginal US examination of the cul-desac. Firm transducer pressure and changes in patient position are needed to inspect between bowel loops for fluid collections. Solid organs and fluid serve as sonographic windows to the abdomen, whereas gas in the bowel and the ribs, spine, and bony pelvis serve as
obstacles.
Intraperitoneal
Fluid
Fluid within the peritoneal cavity flows, under the effect of gravity, along peritoneal reflections to peritoneal recesses (Fig. 36.1) (2) . The hepatorenal recess (Morison pouch) and the pelvic cul-de-sac are the two most dependent recesses in the supine patient. They connect via the paracolic gutters. Fluid outlining the intraperitoneal organs provides an opportunity to evaluate organ surface abnormalities, such as the fine nodularity of cirrhosis. Transudative ascites, urine, and bile are anechoic. Fluid with echogenic particles, layering debris, or septations may be hemorrhage, pus, malignant ascites, or spilled GI contents. Free intraperitoneal fluid outlines recesses and compartments that retain their normal shape. Loops of bowel float and sway freely within free fluid. Loculated fluid collections, abscesses, and cystic masses create their own space, displace bowel and adjacent organs, and are usually more round and tense.
Intraperitoneal
Abscess
Although CT is commonly preferred for detection of small intraperitoneal abscesses, US readily demonstrates most abscesses and is effectively used to guide aspiration and catheter drainage (Fig. 36.2). Because abscesses most commonly form in the dependent recesses, the pelvis must be included in every examination. Abscesses appear as loculated collections of fluid that may be anechoic to densely echogenic. Fluid levels, internal debris, septations, thick walls, and gas within the abscess are common. Gas is brightly echogenic and associated with reverberation artifact and
2377
acoustic shadowing. An abscess containing extensive gas may be mistaken for gas-filled bowel and overlooked. Some abscesses appear solid. Changes in patient position show shifting of the particle pattern when liquid. Doppler and color flow US P.928 show the absence of internal blood vessels within echogenic fluid collections or the presence of blood vessels within solid tissue. Abscesses have mass effect and displace adjacent structures.
FIGURE
36.1. Ascites. A. Longitudinal US image shows
anechoic ascites (a) surrounding the spleen (S). Fluid outlines the gastrosplenic ligament (white arrow). Note the small bare area of the spleen (black arrow), where reflections of the peritoneum from the spleen to the diaphragm prevent access of intraperitoneal fluid. A left pleural effusion (e) is seen above the diaphragm (curved arrow). B . US image of the right lower quadrant of the abdomen reveals ascites (a) containing echogenic particulate matter. This exudative ascites resulted from a bowel perforation.
Intraperitoneal
Tumor
2378
Metastases are the most common tumor of the peritoneal surface. Fluid and gravity distribute malignant cells throughout the peritoneal cavity, where they implant upon visceral or parietal peritoneal surfaces. The greater omentum is fertile ground and thickens with tumor implantation to form “omental cake,― a layer of solid tissue separating bowel from contact with the anterior abdominal wall (Fig. 36.3).
Metastatic
implants P.929
appear as hypoechoic solid masses of varying size on peritoneal surfaces. Ascites is usually present, with echogenic debris and septations common. The most common tumors of origin are ovarian, colon, pancreas, and gastric carcinoma.
FIGURE
36.2. Left Subphrenic Abscess. A. CT scan
demonstrates a loculated fluid collection (Ab) in the left subphrenic space following gastric bypass surgery. The stomach (arrow) is displaced posteriorly. L, liver; S, spleen. B . US in the same patient demonstrates internal septations (arrowhead) within the fluid collection (Ab) that are not apparent on the CT study. A pleural effusion (e) is seen above the diaphragm (curved arrow). The abscess contained gram-negative organisms.
2379
FIGURE 36.3. Peritoneal Metastases. US image shows solid tumor implanted on the omentum creating “omental cake― (OC). Solid tumor causes lumpy thickening of the peritoneal surfaces (arrows). Malignant ascites (a) contains floating echogenic debris. The primary tumor was ovarian carcinoma.
Primary
peritoneal
tumors
include
mesothelioma,
desmoids,
carcinoids, primary peritoneal serous papillary carcinoma, and lymphoma. These appear as predominantly hypoechoic solid masses. Acoustic shadows may arise from dense fibrous tissue or calcifications.
RETROPERITONEUM Normal
US
Anatomy
The retroperitoneum is that portion of the abdomen behind the posterior parietal peritoneum. The anatomy of its three compartments is described in Chapter 26. US of the abdominal aorta
2380
and inferior vena cava (IVC) is discussed in Chapter 40. The crura of the diaphragm must not be mistaken for retroperitoneal adenopathy. Both are hypoechoic linear bands of muscle. The right crus is larger, more lobular, and inserts lower, extending to the L3 vertebral body. The left crus is more uniform in thickness, inserting on the L1 and L2 vertebral bodies. The crura serve as landmarks for identification of the adrenal gland. The psoas and quadratus lumborum muscles show the typical hypoechoic pattern of muscle, with longitudinally oriented echogenic fibrous strands dividing muscle bundles. Echogenic retroperitoneal fat surrounds and defines organs, vessels, and other structures.
FIGURE 36.4. Adenopathy Caused by Lymphoma. An axialplane US demonstrates multiple enlarged hypoechoic lymph nodes (n) surrounding and displacing the aorta (A) and celiac axis (open arrow). The adenopathy extends into the hilum of the right kidney (K). L, liver.
2381
Retroperitoneal
Adenopathy
Enlarged individual lymph nodes are homogeneous, hypoechoic, and round or oval (Fig. 36.4). Accentuated sound transmission may be present, and some enlarged solid nodes are so hypoechoic they appear cystic. A solitary node larger than 1.5 cm in short axis diameter, or multiple nodes larger than 1.0 cm each, are considered to be pathologically enlarged. Lymphoma is characterized by confluence of enlarged nodes to form a solid mass that surrounds vessels and organs. Causes of retroperitoneal adenopathy are lymphoma (most common); tumor metastases (testicular, renal, pelvic, GI malignancies, and melanoma); and infection, especially in AIDS
patients.
Retroperitoneal
tumors are most commonly of mesenchymal origin
and include liposarcoma, leiomyosarcoma, and malignant fibrous histiocytoma. These are aggressive tumors that invade organs and muscles and are difficult to remove surgically. Most are large, heterogenous, and partially cystic. Germ cell tumors in the retroperitoneum may be primary or secondary and either benign or malignant. The sonographic features of the various tumors overlap, and US examination does not yield a specific diagnosis. Benign lipoma may be suggested when the tumor is isoechoic to retroperitoneal
fat.
Retroperitoneal fluid collections include hemorrhage, infection, urinoma, pancreatic fluid collections, and cystic masses (lymphoceles, lymphangiomas, renal cysts, and teratomas). Portosystemic collaterals and other enlarged blood vessels are differentiated by Doppler US. As within the peritoneal cavity, retroperitoneal fluid may be anechoic P.930 or echogenic, with particulate cellular debris and layering fluid levels. Echogenic clotted blood may appear as a solid mass. Absence of internal vascularity on Doppler examination and change in appearance with time are distinguishing features.
LIVER 2382
US is an efficient imaging method to screen patients for diffuse and focal hepatic disease (3, 4, 5). For focal liver metastases, its sensitivity approaches that of CT and MR; however, its images are more difficult to reproduce for follow-up comparisons, and benign and malignant nodules cannot usually be distinguished. Color Doppler US is valuable for the assessment of liver vasculature, for the diagnosis of portal and hepatic vein thrombosis and portal hypertension, and in evaluating the vascularity of liver tumors.
Normal
US
Anatomy
The echogenicity of the liver parenchyma is homogeneous and equal to or slightly greater than that of the kidney (Fig. 36.5A). The surface of the liver is normally smooth, and the inferior margin of the liver is sharp-edged. The lobar and segmental anatomy of the liver is described and illustrated in Chapter 27. The hepatic veins are seen as echolucent tubes with thin walls that converge into the IVC. The portal veins, hepatic arteries, and bile ducts, encompassed by fibrofatty tissue, form the portal triads, which are normally visualized as echogenic foci throughout the liver. Doppler US is used to differentiate blood vessels from bile ducts and small hepatic cysts and to confirm blood flow. Fatty
infiltration causes an increase in echogenicity of the liver,
making affected areas distinctly more echogenic than normal renal parenchyma (4). Fatty infiltration also increases the attenuation of the US beam, diminishing visualization of the diaphragm and commonly requiring a lower-frequency transducer to examine deep portions of the liver (Fig. 36.5B). The hepatic echotexture appears coarsened, and visualization of the portal triads is decreased. The various patterns of fatty infiltration are reviewed in Chapter 27. The “flip-flop― pattern of fatty infiltration as seen on US compared with CT is useful in confirming a diagnosis of focal fatty infiltration and focal fat sparing. Fat-infiltrated areas are bright on US and dark on CT. Focally sparred areas within diffuse fatty infiltration are dark on US and bright on CT. Acute
hepatitis results in diffuse hepatic edema, which reduces the
2383
echogenicity of the liver, resulting in a “starry sky― appearance. The portal triads appear unusually bright on the darkened background of edematous parenchyma. The starry sky appearance has also been described with diffuse leukemic or lymphomatous infiltrate, toxic shock syndrome, and diffuse decrease in glycogen stores in the liver. Passive hepatic congestion refers to stasis of blood in the liver owing to congestive heart failure. US findings P.931 include hepatomegaly, distention of the IVC and hepatic veins, and pulsatile portal vein flow seen on Doppler caused by transmission of right atrial activity through congested sinusoids. Ascites, pleural effusion, and pericardial effusion are often present.
FIGURE 36.5. Normal and Diffuse Fatty Liver. A. Longitudinal US image demonstrates normal liver (L) and right kidney (K). The liver parenchyma is of uniform echogenicity, approximately equal to the parenchymal echogenicity of the kidney. The liver is well visualized to the level of the diaphragm (arrowhead). Small portal triad structures (arrow) are seen throughout the liver parenchyma. B . Diffuse fatty infiltration of the liver (L) markedly increases liver parenchymal echogenicity compared to that of the kidney (K). No portal triads are seen, and the diaphragm
2384
(arrowhead) is less well visualized.
FIGURE
36.6. Cirrhosis.
A. Longitudinal US image of the liver
(L) shows coarsening of the echotexture, loss of visualization of portal triads, and nodularity characteristic of cirrhosis. The deep surface of the liver (arrow) shows the fine nodular contour typical of alcoholic cirrhosis. Note that the echogenicity of the cirrhotic liver is close to the echogenicity of the normal kidney (K). Cirrhosis coarsens hepatic echotexture. Fatty infiltration increases
hepatic
echogenicity. B . A linear array transducer
produces a detailed image of the liver (L) surface showing the nodular contour (arrow) of cirrhosis. SC, subcutaneous tissues. This technique is helpful in revealing the morphologic changes of cirrhosis.
Cirrhosis US reflects the morphologic changes in the liver associated with cirrhosis (4). Hepatic echotexture is usually coarsened and heterogeneous, with numerous vague nodules commonly evident (Fig. 36.6). When examined with high-frequency transducers, the
2385
surface of the liver shows fine or coarse nodularity. Echogenicity is increased in proportion to the degree of fatty infiltration. With alcoholic cirrhosis, the right lobe is shrunken, and the left lobe and caudate lobe are enlarged. Advanced cirrhosis results in a small liver with nodular contour. The normal triphasic Doppler waveform of the hepatic veins is flattened in cirrhosis, with loss of the reverse-flow component caused by atrial systole. US is insensitive (13 mm), dilatation of the splenic and superior mesenteric veins (>10 mm), splenomegaly, and ascites. The hepatic artery may be enlarged and tortuous. Doppler demonstration of reversed (hepatofugal) flow in the portal vein is diagnostic of portal hypertension (Fig. 36.7) . Flow in a dilated paraumbilical vein traversing the falciform ligament and anterior abdominal wall is also highly specific for portal hypertension. Color Doppler US is very useful in the detection of splenorenal, retroperitoneal, and coronary vein collaterals.
2386
FIGURE 36.7. Reversed Flow in the Portal Vein. The spectral waveform below the baseline indicates venous flow away from the transducer. The anatomic image confirms that the flow direction is out of, instead of into, the liver, indicating advanced portal hypertension.
2387
FIGURE 36.8. Portal Vein Thrombosis. The portal vein (PV) is enlarged and partially filled with tumor thrombus (arrows) from hepatocellular carcinoma. L, liver.
P.932 Portal
vein
thrombosis is evidenced by the presence of echogenic
clot within an enlarged portal vein (Fig. 36.8). Color Doppler will confirm complete occlusion or demonstrate residual flow around the thrombus. The thrombus itself varies in appearance from anechoic to hyperechoic, depending upon the age of the thrombus. Tumor thrombus from invasion of the portal vein by hepatoma is confirmed by spectral Doppler demonstration of arterial waveforms in the thrombus within the portal vein. Cysts are common and easily identified and characterized by US (Fig. 36.9). Benign hepatic cysts contain anechoic fluid, have thin walls, and demonstrate posterior acoustic enhancement. Most are septated and have a lobulated, rather than spherical, contour. They vary in size from tiny to huge and are commonly multiple, producing a “bunch of grapes― appearance. Small cysts may mimic
2388
vessels on quick inspection. Doppler is useful to confirm their avascular nature. Cavernous
hemangiomas are commonly identified on hepatic
sonograms. The classic US appearance is a well-defined, homogeneous, hyperechoic mass (Fig. 36.10). Doppler usually shows no internal blood flow, although on occasion with slow-flow, highsensitivity settings, very low-velocity flow is detected. Large lesions may contain hypoechoic thrombosis, fibrosis, and calcification. Most lesions remain stable in size over time, but about 2% show enlargement. Classic-appearing lesions in patients with normal liver function tests usually require no follow-up. Atypical lesions should have a 6-month follow-up US or be confirmed with other imaging modalities, as discussed in Chapter 27.
FIGURE
36.9. Benign Hepatic Cyst. A hepatic cyst (arrow) has
sharply defined walls and anechoic contents. Benign hepatic cysts are rarely simple. They tend to occur in clusters, have thin septations, and lobulated contours. No solid nodular component is evident.
2389
Metastases vary greatly in appearance from hypoechoic to hyperechoic and from homogeneous to heterogeneous to calcified (Fig. 36.11). Metastatic disease must be considered in the differential diagnosis of all solid and atypical cystic lesions in the liver. In 90% of cases, metastatic disease is multifocal in the liver.
FIGURE
36.10. Cavernous
Hemangioma. A small, incidentally
discovered, cavernous hemangioma is seen as a well-defined echogenic mass in the liver (L). Acoustic enhancement, caused by the mostly fluid (blood) content of the lesion, is best appreciated in the mirror image (MI) of the liver projected above the diaphragm.
2390
FIGURE 36.11. Metastases. Metastases from a retroperitoneal sarcoma are seen as innumerable heterogeneous hypoechoic solid masses in the liver.
P.933 Hepatocellular carcinoma may be solitary, multifocal, or diffuse (Fig. 36.12). Detection in the diseased liver is commonly difficult with US. Most are hypervascular, with prominent vascularity shown by color Doppler. Tumor invasion of the portal and hepatic veins is common. Tumors may be hyperechoic with internal fat to hypoechoic and heterogeneous because of nonliquefactive necrosis. Any solid mass detected by US in a diseased liver is suspicious for hepatocellular carcinoma. Abscesses usually appear as complex fluid collections containing echogenic fluid, fluid–fluid layers, or gas (Fig. 36.13). Healed abscesses commonly calcify.
2391
Microabscesses occur most commonly in immunocompromised patients with fungal or parasitic septicemia. Target lesions with central echogenic spot and peripheral hypoechoic halo are common. The differential diagnosis of multiple small (10
2479
to 12 follicles per ovary). In 30% of cases, the ovaries are completely normal in size and appearance. Patients with anovulatory menstrual cycles, especially young female athletes, may have ovaries with multiple follicles but lack the clinical features of polycystic ovary syndrome. Nonovarian
cysts in the pelvis include abscess from appendicitis or
diverticulitis, urachal cysts in the midline P.963 above the bladder, lymphocele in patients with prior pelvic node dissection, and paraovarian cysts in the mesosalpinx arising from wolffian duct remnants. Sonographic demonstration of a separate ovary on the same side as the adnexal mass suggests the diagnosis of nonovarian mass.
FIGURE
37.18. Hydrosalpinx. Transvaginal US demonstrates
the tubular nature of an adnexal mass, confirming hydrosalpinx.
Hydrosalpinx can produce a large complex cystic mass. US shows a thin-walled or thick-walled tubular mass that is commonly elongated and folded on itself (Fig. 37.18) (1 7). Folds in the dilated fallopian
2480
tube may simulate septa in an ovarian tumor. Fluid within the mass is commonly echogenic. Transvaginal US is best for demonstrating the tubular nature of the mass. Other findings of pelvic inflammatory disease may be present. Carcinoma of the fallopian tube is rare. US shows features of a malignant adnexal mass. Peritoneal
inclusion
cysts are relatively common inflammatory
cysts of the peritoneal cavity that result from adhesions that envelop an ovary (1 8). The diseased peritoneum loses its ability to absorb fluid. Secretions from an active ovary confined by adhesions produces an expanding pelvic mass. Patients present with pain or a pelvic mass. Most have a history of previous pelvic surgery, infection, trauma, or endometriosis. US demonstrates a complex fluid collection occupying pelvic recesses and containing the ovary (Fig. 37.19). Septations, loculations, and particulate matter within the contained fluid are common.
FIGURE 37.19. Peritoneal Inclusion Cyst. Sonogram reveals a fixed pelvic fluid collection with angulated boundaries and fluid occupying the peritoneal recesses. The collection encloses the ovary (arrowhead), which is identified by the presence of follicles.
2481
MALE
GENITAL
Testes
and
Normal
US
TRACT
Scrotum Anatomy
The normal testis is ovoid and smooth, measuring approximately 3.5 cm in length and 2.0 to 3.0 cm in diameter (Figs. 37.20, 37.21). It is covered by a dense, fibrous capsule called the tunica albuginea. The testis consists of 250 lobules made up of seminiferous tubules, which are the site of spermatozoa development. The seminiferous tubules unite to form the tubuli recti, rete testes, and finally efferent ductules, which exit the testis at the mediastinum. The mediastinum is an invagination of the tunica albuginea on the posterior surface of the testes that provides access for the testicular vessels P.964 and exit for efferent ductules. The efferent ductules carry seminal fluid to the epididymis.
FIGURE 37.20. Normal Testis. US image demonstrates normal scrotal anatomy. The parenchyma of the testis (T) is of uniform
2482
midlevel echogenicity. The mediastinum of the testis (arrowhead) forms a bright, echogenic line caused by the infolding of the tough fibrous capsule, the tunica albuginea. The epididymis exits the testis through the mediastinum, forming the nodular head ( Table of Contents > Section IX - Ultrasonography > Chapter 38 Obstetric Ultrasound
Chapter
38
Obstetric
Ultrasound
William E. Brant
Imaging
Methods
US is the imaging method of choice for dating the pregnancy, monitoring fetal growth, assessing fetal well-being, and evaluating fetal anatomy and maternal pelvic organs (1,2). Transvaginal sonography is particularly useful in the assessment of first-trimester pregnancy and in the demonstration of fetal anatomic structures deep in the pelvis. Modern US offers superb anatomic detail in real time, keeping up with the frequently vigorous motion of the fetus. MR is used occasionally as a supplement to US imaging when the US examination is equivocal. MR offers excellent detail of maternal pelvic organs, unobscured by bone, gas, or fat (3). Demonstration of fetal anatomy is limited by fetal motion but may be overcome by fetal sedation and fast scanning techniques. CT is the method of choice for pelvimetry, now a rarely used technique. The obstetric US examination consists of a survey of the uterus and maternal pelvic organs, measurements of the fetus to date the pregnancy and assess fetal growth, and a survey of fetal anatomy. Standards for the performance of obstetric US examinations have been published by the American Institute of Ultrasound in Medicine (AIUM) and endorsed by the American College of Radiology (ACR) and the American College of Obstetricians and Gynecologists (ACOG) (4). In the first trimester, the location and appearance of the
2521
gestational sac are documented. The presence or absence of a yolk sac and embryo is confirmed. If an embryo is present, the crownrump length (CRL) is measured, and fetal cardiac activity is documented. Fetal number is determined and the uterus and adnexa are examined. Second- and third-trimester sonography includes assessment of fetal life and number; fetal position; amount of amniotic fluid; placental location and appearance; fetal measurements (biparietal diameter, head circumference, abdominal circumference, femur length); and evaluation of the uterus and adnexa. Assessment of fetal anatomy includes the cerebellum, cisterna magna, lateral cerebral ventricles, choroid plexus, midline falx, cavum septi pellucidi, a four-chamber view of the heart and ventricular outflow tracts, and images of the entire spine, stomach, kidneys, bladder, umbilical cord insertion site, umbilical cord vessel number, and arms and legs. The literature refers to “level 1― obstetric US as routine or standard examinations and “level 2― (specialized or detailed) examinations as targeted to scrutinize fetal anatomy and detect anomalies.
FIRST
TRIMESTER
The first trimester covers the period from conception to the end of the 13th menstrual week. This includes the entire embryonic period (0 to 10 weeks) and is a time of dynamic growth and the differentiation and development P.977 of most organ systems. The embryo and fetus have the greatest risk of maldevelopment, injury, and death during this period because of external factors (infection, drugs, radiation, etc.) or chromosome abnormalities. About 40% of implanted zygotes are menstrually aborted, and another 25% to 35% of surviving embryos will threaten to abort during the first trimester.
Normal
Gestation
The presence of a pregnancy is confirmed by a positive serum βhuman chorionic gonadotropin (β-hCG) test or by a positive enzyme-
2522
linked immunoassay (ELISA) urinary pregnancy test. Radioimmunoassay for serum β-hCG allows pregnancy to be detected within 2 weeks of conception (as early as 23 menstrual days) and before a normal gestational sac can be detected by either transabdominal or transvaginal US. The early gestational sac can be seen by transvaginal sonography at 3.5 to 4.5 menstrual weeks as a tiny cystic structure implanted within the echogenic decidua: the intradecidual sign (Fig. 38.1). This sign is not specific for early intrauterine pregnancy and may be mimicked by fluid collections or decidual cysts in the presence of ectopic pregnancy. A normal gestational sac is visualized by the transabdominal approach by 5 menstrual weeks. The normal gestational sac appears on US as a smoothly contoured, round or oval, fluid-containing structure positioned in the endometrial cavity near the fundus of the uterus (Table 38.1). The normal sac has an echogenic border greater than 2 mm thick, which represents the choriodecidual reaction. A double decidual sac sign is evident in about 85% of normal pregnancies. The double sac sign is produced by visualization of three layers of decidual reaction early in pregnancy (Fig. 38.2). The term decidua refers to the endometrium of the pregnant uterus. The decidua vera lines the endometrial cavity, and the decidua capsularis covers the gestational sac. The decidua basalis contributes to the formation of the P.978 placenta at the site of implantation. A small amount of fluid in the endometrial cavity separates the decidua vera from the decidua capsularis, enabling visualization of the “double sac.― The free margin of the gestational sac consists of chorion and decidua capsularis and is normally at least 2 mm thick. The double sac is not complete because visualized double An absent double pregnancy or an
of placental attachment to the uterine wall. A wellsac is excellent evidence of intrauterine pregnancy. sac sign is evidence of an abnormal intrauterine ectopic pregnancy.
2523
FIGURE
38.1. Intradecidual
Sign. Transvaginal US image of
the uterus in a transverse plane demonstrates a tiny gestational sac (arrow) implanted within the thickened decidual (between arrowheads). The size of the sac corresponds to a pregnancy of approximately 4 weeks’ menstrual age.
2524
FIGURE 38.2. Double Decidual Sac Sign. A longitudinal image of the uterus obtained transabdominally through a filled bladder (B) demonstrates a gestational sac with two decidual layers in the uterine cavity. The two echogenic lines (arrow) are formed by the decidua vera lining the endometrial cavity and the decidua capsularis covering the gestational sac. The placental implantation site on the anterior aspect of the uterus (arrowhead) has a single echogenic stripe owing to the decidua basalis.
TABLE 38.1 US Characteristics of a Normal Gestational Saca
Intradecidual sign—before 5 weeks’ GA Double decidual sac sign—after 5 weeks’ GA (>98% of IUP) Well-defined round/or oval anechoic sac Echogenic decidua >2 mm thick Position in upper uterine body midway between uterine walls Growth in MSD >1.2 mm/day Yolk sac 2 to 6 mm in diameter: Always present when MSD ≥20 mm on transabdominal US Always present when MSD ≥8 mm on transvaginal US Embryo: Always present when MSD ≥25 mm on transabdominal US Always present when MSD ≥16 mm on transvaginal US a The
gestational sac diameter is measured in three
orthogonal planes, and the measurements are averaged to calculate MSD. IUP, intrauterine pregnancy; MSD, mean sac diameter.
2525
Adapted from Nyberg DA, Laing FC, Filly R A, et al. Ultrasonographic differentiation of the gestational sac of early intrauterine pregnancy from the pseudogestational sac of ectopic pregnancy. Radiology 1983;146:755–759; and from Levi CS, Lyons EA, Lindsay DJ. Early diagnosis of nonviable pregnancy with endovaginal US. Radiology 1988;167:383–385.
The yolk sac is a 2- to 6-mm-diameter, spherical, cystic structure (Fig. 38.3) that is connected to the midgut of the embryo by a thin stalk, the vitelline duct. A Meckel diverticulum is a remnant of the connection of the vitelline duct (also called the omphalomesenteric duct) to the distal ileum. The yolk sac is the earliest site of blood cell formation in the embryo. It floats freely in fluid between the amniotic and chorionic membranes. It is generally the earliest structure visualized within the gestational sac and serves as definitive evidence of early pregnancy. The yolk sac should always be visualized in normal pregnancy in gestational sacs of 20-mm mean sac diameter by transabdominal US or 8-mm mean sac diameter by transvaginal US. The earliest demonstration of the embryo is the double bleb sign, which is produced by the amniotic sac and the yolk sac with the embryonic disc between them (Fig. 38.4). Embryos as small as 2 mm can be detected by transvaginal US. The earliest embryonic cardiac activity can be detected by careful inspection of the embryonic disc by real-time US. Transvaginal sonography may demonstrate tiny normal embryos (8 mm is indicative of cervical incompetence. Membranes may be seen bulging into the cervical canal. Sutures associated with cervical cerclage used to treat cervical incompetence are seen on US as echogenic linear structures with acoustic
shadowing.
2551
FIGURE
38.16. Cervical
Incompetence. The cervix is best
evaluated with a translabial view with the bladder (B) empty. The transducer is aimed down the long axis of the vagina (V). The cervix, measured between the internal os and the external os (arrowheads), is shortened to 9 mm in this patient with a history of multiple spontaneous abortions in the second trimester. The cervix is also dilated, allowing amniotic fluid (asterisk) to enter the endocervical canal. The fetal head (H) is presenting at the internal
cervical
Placenta
os.
and
Membranes
Normal placenta is first apparent on US at about 8 weeks as a focal thickening at the periphery of the gestational sac (2). The disclike shape of the placenta becomes evident by 12 weeks, and by 18 weeks the placenta is finely granular and homogeneous, with a smooth covering chorionic membrane along its fetal surface. The
2552
retroplacental complex of decidual and myometrial veins forms a prominent sonographic landmark (Fig. 38.17). As the gestation advances, the placenta becomes more heterogeneous, with focal echolucencies owing to venous lakes and areas of fibrin deposition. Septations become prominent P.987 US features throughout the placenta and cause undulations of the placental surface. Calcifications occur along the septations and are dispersed randomly throughout the placenta. These are normal changes of placental aging and should not be interpreted as indicators of disease. The normal placenta has a maximum thickness of 4 cm and a minimal thickness of 1 cm. Thick placentas are associated with maternal diabetes, maternal anemia, hydrops from immune and nonimmune causes, and chronic uterine infections. Thin placentas are associated with preeclampsia, placental insufficiency, IUGR, and trisomies 13 and 18.
FIGURE 38.17. Normal Placenta. A transabdominal scan demonstrates a normal placenta (P) and the insertion site of the cord onto the placenta (arrowhead). The retroplacental complex of veins (arrows) appears as a network of tubular lucencies
2553
beneath the placenta. A, amniotic cavity.
Placenta previa is present when part or all of the placenta covers the internal cervical os. Placenta previa is present at term in 0.3% to 0.6% of live births. Placenta previa is suggested by US in as many as 45% of pregnancies examined in the first and second trimesters. These cases are the result of low implantation of the placenta and filling of the bladder, which distorts the lower uterine segment and cervix. As the pregnancy progresses, the muscular portion of the cervix elongates and increases the distance from the margin of the placenta to the cervical os. Risk factors for placenta previa include scarring of the lower uterine segment associated with previous cesarean section, previous placenta previa, surgical scars, and multiple previous pregnancies. Patients usually present with painless vaginal bleeding in the third trimester. Bleeding is initiated by the effacement of the cervix and dilation of the cervical os, which disrupts the vascular bed of the placenta. US confirmation of placenta previa is performed transperineally, with the bladder empty to allow optimal identification of both the edge of the placenta and the internal os of the cervix. When the placenta covers the entire cervical os, the previa is complete (Fig. 38.18). When an edge of the placenta covers a portion of the cervical os, the previa is partial or marginal.
2554
FIGURE 38.18. Placenta Previa. Transabdominal US shows a normal cervix (between cursors) measuring 34 mm. The placenta (P) covers the internal os. A, amniotic cavity; B, bladder; V, vagina.
Vasa previa is present when placental blood vessels, or the umbilical cord, are adherent to the membranes that cover the cervix. The vessels tear as the cervix dilates, resulting in fetal hemorrhage and death. Color Doppler is used to identify blood vessels fixed in place over the internal cervical os. Placental abruption is defined as the premature separation of a normally positioned placenta from the myometrium. Separation is associated with hemorrhage from the maternal vessels at the base of the placenta. Abruption complicates 0.5% to 1.3% of pregnancies and is implicated in 15% to 25% of perinatal deaths. Risk factors include maternal hypertension, smoking, cocaine abuse, and previous history of abruption. Subchorionic hemorrhage (marginal abruption) occurs because of a separation at the edge of the placenta. Bleeding
2555
is usually venous and preferentially accumulates beneath the chorionic membrane adjacent to the placenta. Retroplacental hemorrhage occurs with more central abruption. Bleeding is usually arterial and accumulates beneath the placenta as an anechoic or mixed hypoechoic mass (Fig. 38.19). The hemorrhage may be isoechoic and difficult to differentiate from the placental tissue. The diagnosis is suggested by demonstrating disruption of the retroplacental complex of veins and thickening of the placenta (>4 cm). Placenta accreta is an abnormal adherence of the placenta to the uterine wall. Invasion of the uterine wall by the placenta is referred to as placenta increta, and penetration of the uterine wall is placenta percreta. The decidua basalis P.988 and retroplacental complex of veins are completely or partially absent. Failure of the abnormally adherent placenta to separate completely from the myometrium after delivery results in copious hemorrhage. Risk factors include prior cesarean section, prior placenta accreta, and prior placenta previa. Scarring of the uterus results in the defective formation of decidua. US findings include absence of normal vascular channels in the retroplacental region, increased echogenicity of tissues deep to the placenta, and visualization of retroplacental vessels within the bladder lumen. Placenta previa is usually also present.
2556
FIGURE
38.19. Placental
Abruption. The placenta (P) is
displaced away from the wall of the uterus (U) by an echogenic hematoma (H). Note the absence of visualization of the retroplacental complex of veins. A, amniotic cavity.
Chorioangioma is a benign vascular placental mass supplied by the fetal circulation. It appears on US as a solid hypoechoic, sometimes septated, mass in the placenta, usually close to the chorionic surface. Doppler demonstration of arterial waveforms at the fetal heart rate in vessels supplying the tumor is diagnostic. Vascular shunting may cause fetal high-output cardiac failure and fetal hydrops.
Umbilical
Cord
The normal umbilical cord consists of two arteries and one vein surrounded by Wharton jelly. It has a normal diameter of 1 to 2 cm. A single-artery umbilical cord is found in about 1% of pregnancies and has a 10% to 20% association with congenital malformations. Associated anomalies include cardiac, urinary tract, and CNS malformations, omphalocele, trisomy 13, and trisomy 18. Masses in the umbilical cord include allantoic cysts, hematomas, hemangiomas,
2557
and
teratomas.
Placental membranes consist of an outer layer (chorion) and an inner layer (amnion). These membranes commonly remain separated by a layer of fluid until 14 to 16 weeks’ GA, when the two membranes fuse. The amnion is visualized on US as a thin membrane floating in fluid. The chorion is identified as the membrane that confines fluid within the gestational sac. Occasional persistence of chorioamnionic separation into the third trimester is believed to be of no clinical significance.
FIGURE
38.20. Amniotic
Band
Syndrome. The forearm
(arrowhead) of a fetus at 15 weeks’ GA is entangled within fibrous bands (arrows) that extend across the chorionic cavity (C).
Amniotic band syndrome is caused by the early (before 10 weeks’ GA) disruption of the amnion, which enables the fetus to enter the chorionic cavity (Fig. 38.20). The fetus becomes entangled in fibrous bands that cross the chorionic cavity. Entrapment of fetal parts results in amputation deformities that range from mild to
2558
incompatible with life. Typical abnormalities include asymmetric absence of the cranium resembling anencephaly, encephaloceles, gastroschisis and truncal defects, spinal deformities, and extremity amputations. The amniotic bands trapping the fetus may be visualized. Amniotic
sheets (uterine synechia) are membranous structures that
project into the uterine cavity. They demonstrate a characteristic appearance with a bulbous-free edge, thinner midportion, and a thickened base (Fig. 38.21). The fetus is able to move freely about the sheet of tissue. No fetal deformities are associated with this condition, which makes it distinct from the amniotic band syndrome. The amniotic sheets arise from folding of the chorioamnionic membranes over an intrauterine adhesion. Patients at increased risk for amniotic sheets include those with prior history of dilation and curettage, therapeutic abortion, or endometritis. An increased rate of cesarean section because of fetal malpresentation has been reported.
Amniotic
Fluid
Normal amniotic fluid is essentially a dialysate of maternal serum in early pregnancy. As the pregnancy advances, fetal urine becomes the major source of amniotic fluid. The P.989 composition of amniotic fluid is dynamic, with turnover of the entire volume every 3 hours. The fetus swallows amniotic fluid at a rate up to 450 mL per 24 hours. Transudate from the fetal lungs contributes a small volume. Water crosses placental membranes in response to osmotic gradients. Amniotic fluid is essential in promoting normal development and maturation of the fetal lungs. Suspended particles in amniotic fluid visualized by US are attributable to normal vernix (desquamated fetal skin), blood, or meconium.
2559
FIGURE 38.21. Amniotic Sheet. A. A fibrous band covered by chorioamnionic membranes (arrow) extends across the amniotic cavity. The uterine synechia forms a shelflike structure that partially compartmentalizes the uterine cavity. The fetus has free access to both compartments. B . The characteristic free edge (arrow) of the amniotic sheet is demonstrated.
Amniotic fluid index is a rough US measurement of amniotic fluid volume obtained by measuring the vertical diameter of the deepest pockets of fluid in the four quadrants of the uterus and adding these values together. Pockets are selected that do not include fetal parts or umbilical cord. Normal values are 5 to 20 cm. Polyhydramnios is an excessive amount of amniotic fluid, traditionally defined as greater than 2 L of fluid at delivery. US is used to confirm excessive fluid any time in pregnancy. Because amniotic fluid volume is difficult to measure accurately, the diagnosis is usually made subjectively by visual inspection. The visual proportion of fluid relative to the size of the fetus is greatest early in the second trimester and decreases progressively to term. Polyhydramnios is suggested by large pockets of fluid relative to the size of the fetus and the age of the pregnancy. An amniotic fluid index greater than 20 cm or a single fluid pocket greater than 8 cm
2560
deep is strongly suggestive of polyhydramnios. Another clue is failure of the fetal abdomen to be in contact with both anterior and posterior uterine walls after 24 weeks’ GA. Excessive fluid is associated with preterm labor, premature rupture of membranes, and substantial maternal discomfort. About 60% of cases are idiopathic; 15% to 20% are related to maternal disease (diabetes mellitus, preeclampsia, anemia, obesity); and 20% to 25% are associated with fetal anomalies. About half of all fetuses with anomalies will have polyhydramnios. Gross polyhydramnios has a higher association with fetal anomalies than mild polyhydramnios. Associated anomalies include anencephaly, encephalocele, GI obstructions, abdominal wall defects, achondroplasia, and hydrops (isoimmunization). Oligohydramnios refers to an abnormally low amniotic fluid volume. Fluid pockets are small or absent, fetal parts are crowded, fetal surface features such as the face are difficult to visualize, and the amniotic fluid index measures less than 5 cm. Measurement of the largest fluid pocket in the vertical direction of less than 1 cm is indicative of severe oligohydramnios. Causes of oligohydramnios include premature rupture of membranes, IUGR, renal anomalies (lack of urine output), fetal death, eclampsia, and postdate pregnancies. A major complication of severe oligohydramnios is fetal lung immaturity.
Multiple
Pregnancy
Twins occur in 1 of every 90 births. Morbidity and mortality are significantly increased in twin pregnancy compared with singleton pregnancy. Twins account for 12% to 13% of all neonatal deaths. Morbidity associated with multiple pregnancy includes prematurity, polyhydramnios, increased incidence of congenital anomalies, P.990 discordant growth, and cord accidents. Relative risk is increased if the fetuses share a placenta (monochorionic twins, 20%) as opposed to each fetus having its own placenta (dichorionic twins, 80%). Twins that share a single amniotic cavity (monoamniotic twins) have the highest risk for morbidity, including conjoined twinning and
2561
intertwining of the umbilical cords. Visualization of two separate placentas, or determination that the twins are of different sex, is definitive proof of lower-risk dichorionic twinning. Unfortunately, about half of dichorionic twins will have a fused placenta. Visualization of a membrane separating the twins confirms diamniotic twins. Monochorionic twins usually have vascular anastomoses at the placental level, making them at risk for twin transfusion syndrome and twin embolization syndrome. Twin
transfusion
syndrome results from shunting of blood from
one twin to the other through vascular connections in the placenta. The abnormality ranges in severity from minor discordance in growth to severe IUGR in one twin, with hydropic fluid overload in the other twin. Severe disparity in amniotic fluid volume may be present, with one twin experiencing polyhydramnios while the other twin is virtually anhydramniotic (“a stuck twin― compressed against the uterine wall by the amnion). The mortality rate is as high as 70%. Twin
embolization
syndrome is an uncommon complication of the
death of one twin in utero. Blood products from the dead twin are shunted through placental interconnections to the live twin, resulting in disseminated intravascular coagulopathy and multifocal tissue infarction.
FETAL
ANOMALIES
General Fetal
hydrops refers to the pathologic accumulation of fluid in body
cavities and tissues. US demonstrates ascites, pleural and pericardial effusions, and subcutaneous edema (Fig. 38.22). Immune hydrops is caused by blood group incompatibility between mother and fetus. Current treatment, including fetal transfusion, is highly successful. Nonimmune hydrops is caused by a host of conditions, including cardiac disorders, infections, chromosomal anomalies, twin pregnancy, urinary obstruction, and umbilical cord complications. The cause of many cases is not identified. The prognosis for nonimmune
2562
hydrops
remains
poor.
Alpha-Fetoprotein Screening
(AFP)
and
Triple
Marker
AFP is a protein produced by the fetal liver. Concentrations of AFP are highest in the fetal serum, with small amounts present in the amniotic fluid (AF-AFP) and minute amounts detectable in maternal serum (MS-AFP). Open neural tube and other skin defects in the fetus allow AFP to leak into the amniotic fluid and maternal serum in abnormally large quantities. Routine MS-AFP screening is performed to aid in detection of neural tube defects and other fetal anomalies. Triple marker screening refers to expanded maternal serum screening programs that have added β-hCG and unconjugated estriol (uE3) determinations to MS-AFP screening (1 1). Results are reported as multiples of the median (MOM) values. AFP is considered elevated when it is greater
than 2.50 MOM. Low values for the triple marker
screen are correlated with maternal age to yield a risk for chromosome abnormalities, especially trisomy 21 and trisomy 18. The MS-AFP or triple marker screening blood tests are performed at 16 to 18 weeks’ GA as determined by menstrual history. The normal values for MS-AFP vary with GA, reaching maximum values at 30 to 32 weeks’ gestation. Patients with abnormal MS-AFP or triple marker screening are routinely referred for US dating, detailed fetal examination, and consideration of amniocentesis for karyotyping. The differential diagnosis for elevated AFP is listed in Table 38.6.
2563
FIGURE 38.22. Fetal Hydrops. A transverse image through the fetal thorax at the level of a four-chamber view of the heart (black arrowhead) demonstrates bilateral pleural effusions (white arrows) outlining the lungs (L). The fetal chest is viewed from above, with the spine (S) posterior. This fetus also had ascites.
Chromosome
abnormalities are suspected when multiple or major
fetal anomalies are detected by US. Advanced maternal age (>35 years at delivery) or a parent or previous child with aneuploidy or chromosomal translocation anomalies are risk factors for fetal chromosome abnormalities. Fetuses with structural anomalies detected on US have an 11% to 35% risk of associated chromosome abnormality. Fetal conditions with significant high risk of associated chromosome abnormality include syndrome, cystic hygroma,
holoprosencephaly,
Dandy-Walker P.991
cardiac malformations, omphalocele, duodenal atresia, facial anomalies, and early symmetric IUGR. Chromosome analysis is
2564
performed on samples obtained by amniocentesis or chorionic villous sampling.
TABLE 38.6 Causes of Elevated MS-AFP
Erroneous
gestational
dating
Multiple pregnancy Fetal demise Neural tube defects: Anencephaly Spina bifida Encephaloceles Abdominal
wall
defects:
Gastroschisis Omphalocele Amniotic
band
syndrome
Cystic hygroma Placental abnormalities: Subchorionic
hemorrhage
Chorioangioma Unexplained—fetus is at high risk for: IUGR Fetal death Preterm delivery Preeclampsia Oligohydramnios
MS-AFP, maternal serum level of alpha-fetoprotein; IUGR, intrauterine growth retardation.
Trisomy 21, Down syndrome, is the most common chromosome abnormality, occurring in 1 of 660 births. Although women older than age 35 have a 1 in 250 risk of carrying a fetus with trisomy 21, 80% of fetuses with Down syndrome are born to younger women (1 1) .
2565
Triple marker serum screening will detect about 60% of affected fetuses. A variety of US findings serve as markers for the condition (1 2). Major structural defects found in Down fetuses include congenital heart disease (endocardial cushion defect), duodenal atresia, and hydrocephalus. Nuchal fold thickening greater than 3 mm in the first trimester or greater than 6 mm in the second trimester is strongly associated. The skinfold is measured from the occipital bone to the external skin surface on the transcerebellar plane view. If two or more of the following findings are present, trisomy 21 is likely: short femur, short humerus, echogenic bowel, mild fetal renal pyelectasis, intracardiac echogenic focus, hypoplastic middle phalanx of fifth finger. Trisomy 18 is the second most common chromosome anomaly, occurring in 1 of 3,000 births. A large number of structural abnormalities may occur, but the most common identified by US are IUGR (74%), complex congenital heart disease (52%), choroid plexus
cysts
(47%),
congenital
diaphragmatic
hernia,
omphalocele,
neural tube defects, Dandy-Walker syndrome, clenched hands, and single umbilical artery (1 2) .
Central Nervous System, Face, and Neck Anomalies of the CNS occur in 1 of 1,000 live births (2).
Survivors
are often severely disabled and require long-term care. Effective US screening for CNS anomalies can be performed by examination of three crucial axial planes through the fetal brain. The transthalamic plane is used to measure the BPD and HC (Fig. 38.12). Abnormalities of head shape, microcephaly, macrocephaly, and major structural abnormalities are evident in this plane. The third ventricle varies in appearance from a single echogenic line to a slitlike structure narrower than 3.5 mm. The transventricular plane is an axial plane at the level of the ventricular atria (Fig. 38.23). The dominant landmark is the echogenic choroid plexus, which normally fills the atrium nearly completely. Measurements of atrial diameter made perpendicular to the walls do not normally exceed 10 mm. The transcerebellar plane is an axial scan at approximately 10° to 15° of inclination from the canthomeatal line. The anatomic landmarks
2566
include the inferior portion of the third ventricle and the cerebellar hemispheres, which are outlined by fluid in the cisterna magna (Fig. 38.24). The normal cisterna magna measures 2 to 11 mm in width. A small cisterna magna (11 mm) may be a normal variant (mega–cisterna magna) or indicate Dandy-Walker malformation, arachnoid cyst, or cerebellar hypoplasia. When these three planes are anatomically normal, the risk of CNS anomaly is minute (0.005%). An algorithm for sorting out fetal CNS anomalies is given in Table 38.7.
FIGURE 38.23. Transventricular Plane: Early Ventriculomegaly. The choroid plexus (thin arrow) hangs dependently in the atria of the lateral ventricle (between cursors). The choroid plexus (arrowhead) in the near ventricle also hangs dependently. The ventricular atrium is measured from its medial wall to its lateral wall (between cursors). The normal ventricular atrium does not exceed 10 mm in width at any time during pregnancy. The diameter of the atrium in this case
2567
measures 12 mm, indicating ventriculomegaly. This fetus has a spina bifida defect with associated Arnold-Chiari II malformation as the cause of ventriculomegaly. Note the bossing of the frontal bones (thick arrows), which gives the outline of the cranium an appearance similar in shape to a lemon (lemon head).
FIGURE 38.24. Transcerebellar Plane. Landmarks for the transcerebellar plane are the thalami (t), third ventricle (arrow) , and cerebellar hemispheres (c). The cisterna magna (between arrowheads) is measured from the vermis (white arrowhead) to the occiput (black arrowhead). The normal cisterna magna measures 2 to 11 mm throughout pregnancy.
Ventriculomegaly is an anatomic finding with many causes that can be grouped into the categories of obstructive hydrocephalus (obstruction to flow of CSF), cerebral atrophy (ex vacuo), and maldevelopment (such as agenesis of the corpus callosum). Ventriculomegaly detected in utero carries a poor prognosis. Up to 80% of fetuses with ventriculomegaly have associated anomalies.
2568
The US signs of ventriculomegaly include diameter of the ventricular atrium >10 mm, separation of choroid plexus from the ventricular wall by >3 mm, and a “dangling choroid.― The choroid plexus hangs dependently in the ventricle and marks the position of the lateral ventricular wall. The most common causes of ventriculomegaly in the fetus are Chiari II malformation and aqueductal stenosis (Fig. 38.25) .
TABLE 38.7 Algorithm for Diagnosis of Congenital Brain Abnormalitiesa
2569
FIGURE
38.25. Ventriculomegaly. An axial image of the fetal
brain in fetus with aqueduct stenosis demonstrates marked enlargement of the lateral ventricles (V). The falx (arrowhead) is seen as an echogenic stripe in the midline. A rind of cortex (arrow) is present. These latter two findings differentiate ventriculomegaly from hydranencephaly and holoprosencephaly.
P.993 Anencephaly is the most common neural tube defect. US findings include absence of the cranial vault and cerebral hemispheres above the level of the orbits (Fig. 38.26). The cerebral hemispheres may be replaced by an amorphous neurovascular mass (area cerebrovasculosa).
The
condition
is
inevitably
2570
fatal.
FIGURE 38.26. Anencephaly. A sagittal image through the head of a fetus demonstrates absence of the cranial vault (thick arrow) above the level of the eye (thin arrow). The mouth and lips are evident (arrowhead). The volume of amniotic fluid is increased. Polyhydramnios is common in the presence of anencephaly. Arm, fetal arm.
2571
FIGURE 38.27. Encephalocele. Axial US image through the fetal skull demonstrates herniation of brain tissue (B) through a large defect in the skull, forming an occipital encephalocele (between arrows). The intracranial contents are reduced and the biparietal diameter (between cursors) is less than expected for gestational age because of the encephalocele.
Cephaloceles are fluid-filled and/or brain tissue–filled sacs that protrude through a defect in the bony calvaria. They are found in the occipital (75%), frontoethmoid (13%), and parietal (12%) regions. Meningoceles contain only CSF, whereas encephaloceles contain brain tissue (Fig. 38.27) . Spina
bifida refers to a spectrum of spinal abnormalities resulting
from failure of the complete closure of the neural tube. The condition ranges from simple nonfusion of the vertebral arches with intact skin (spina bifida occulta); to protruding sacs containing CSF, spinal cord, or nerve roots (myelomeningocele); to a totally open spinal defect (myeloschisis). Spina bifida may occur anywhere in the spine but most often occurs in the lumbosacral region. US findings (Fig. 38.28) include: outward splaying, rather than inward convergence, of the laminae; defect in the soft tissues overlying the bony abnormality; and a protruding sac containing fluid and often neural tissues. The associated functional neuromuscular defect often results in club foot deformities and dislocated hips. Associated cranial abnormalities of the Chiari II malformation provide clues to the presence of the spinal defect. Ventriculomegaly is present in 75% of cases. The “lemon sign― refers to bossing of the frontal bones, causing a lemonshaped appearance to the head in the axial plane (Fig. 38.23). The “banana sign― is produced by compression of the cerebellar hemispheres into a banana shape. The cisterna magna is small or obliterated. Chiari
II
malformation is associated with 95% of
myelomeningoceles. The cranial abnormality consists of caudal displacement of the cerebellar tonsils, pons, and medulla. The fourth ventricle is elongated, the posterior fossa is small, and the cisterna
2572
magna
is
FIGURE
obliterated.
38.28. Normal Spine and Spina Bifida. A. Normal
spine. Posterior transverse image through a normal fetal spine at the L4-L5 level demonstrates the ossified portion of the vertebral body (b) anteriorly and the converging ossified portions of the lamina (arrows) posteriorly. The skin overlying the posterior aspect of the vertebra is intact (arrowhead). The spine normally casts an acoustic shadow (S). B . Spina bifida. Posterior transverse image through a spina bifida defect demonstrates the ossified portion of the vertebral body (b) anteriorly and the diverging ossified portions of the lamina (arrows) posteriorly. A defect is evident in the overlying skin (arrowhead) .
P.994 Holoprosencephaly refers to a spectrum of disorders characterized by a failure of the prosencephalon to divide and form separate right and left hemispheres and thalami. Associated facial anomalies including hypotelorism, cyclopia, and proboscis are common. Alobar holoprosencephaly is the most severe form and demonstrates absence of the falx and interhemispheric fissure with a single midline ventricle (Fig. 38.29). The semilobar and lobar forms demonstrate greater degrees of midline separation.
2573
FIGURE
38.29. Holoprosencephaly. Image through the
cranium of a fetus reveals a single large midline ventricle (V) and fused thalami (arrow). A thin rim of cortex (arrowhead) is present. These findings are characteristic of alobar holoprosencephaly. The fetal face should be examined for associated defects such as midline cleft and proboscis.
Hydranencephaly refers to total destruction of the cerebral cortex, believed to be caused by the occlusion of the internal carotid arteries. The cranial vault contains fluid, but no cortical mantle of brain tissue is visible (Fig. 38.30). The falx may be present but is usually incomplete. The brainstem and structures supplied by the vertebral arteries appear normal. Dandy-Walker malformation results from the maldevelopment of the roof of the fourth ventricle. The cisterna magna is enlarged and communicates directly with the fourth ventricle through its absent roof. The posterior fossa is enlarged, and the tentorium is elevated. The cerebellar hemispheres are usually hypoplastic (Fig. 38.31) . Hydrocephalus is usually present. The condition varies in severity
2574
across a broad spectrum. Less severe abnormalities are usually called Dandy-Walker variants. Arachnoid cysts and large cisterna magna are differentiated by their lack of communication with the fourth
ventricle.
Choroid plexus cysts are found in 1% to 3% of normal fetuses during the second trimester. The cysts themselves cause no clinical problem and nearly always resolve. Because they are present in up to 47% of fetuses with trisomy 18, their discovery causes concern for the presence of chromosome abnormality. In nearly all cases, detailed US examination, which should include echocardiography and examination of the fetal hands, will demonstrate additional structural abnormalities that justify amniocentesis for karyotyping. Trisomy 18 is unlikely and amniocentesis is not indicated if detailed US examination of the fetus is normal. Cleft lip and cleft palate account for 13% of all congenital anomalies found in the United States. Lateral cleft is most common and involves both lip and palate in 50% of P.995 cases, the lip alone in 25%, and the palate alone in 25%. The condition is bilateral in 20% to 25% of cases. Up to 60% of affected fetuses have additional anomalies, including polydactyly, heart disease, and trisomy 21. US diagnosis is made on
congenital
demonstration of a groove extending from one of the nostrils through the lip (Fig. 38.32) (1 3). Median cleft is a completely different entity associated with holoprosencephaly and accounting for fewer than 0.7% of all cases of cleft lip. A coronal plane sonogram of the face demonstrates a wide central defect in the upper lip and palate.
2575
FIGURE
38.30. Hydranencephaly. Axial sonogram through the
brain of a near-term fetus demonstrates two massive ventricles (V), a well-defined midline falx (arrowhead) and total absence of detectable cortical tissue (arrow). These findings are characteristic
of
hydranencephaly.
2576
FIGURE 38.31. Dandy-Walker Malformation. Coronal plane image demonstrates cystic enlargement of the posterior fossa (arrow). The lateral ventricles (V) are enlarged, indicating associated hydrocephalus.
Cystic hygroma is a fluid collection in the fetal neck caused by failure of the lymphatic system to develop normal connections with the venous system in the neck. US demonstrates a bilateral nuchal cystic mass with a prominent midline septum that represents the nuchal ligament (Fig. 38.33). Up to 70% have abnormal karyotypes, usually Turner syndrome or Down syndrome. Generalized lymphangiectasia and fetal hydrops may occur and are always fatal when they do.
Chest and Heart Congenital
diaphragmatic
hernia is a disorder in which abdominal
contents protrude into the thorax through defects in the diaphragm. The most common type involves the foramen of Bochdalek at the posterolateral aspect of the diaphragm. The majority (75%) occur on
2577
the left side (Fig. 38.34). Anteromedial defects at the foramen of Morgagni also occur. US findings include: fluid-filled, solid, or multicystic mass in the chest; displacement of the heart and mediastinum; absence of the stomach in the abdomen; and polyhydramnios. Associated defects, especially cardiac and CNS, are common. Mortality is high (50% to 80%) because of pulmonary hypoplasia. Cystic adenomatoid malformation is a congenital hamartomatous lesion of the lung that usually affects one lobe. The lesion consists of single or multiple cysts that vary in size from microscopic to larger than 2 cm. Type 1 lesions appear on US as single or multiple cysts larger than 2 cm. Type 2 lesions consist of multiple smaller cysts of uniform size ( Table of Contents > Section IX - Ultrasonography > Chapter 39 Chest, Thyroid, Parathyroid, and Neonatal Brain Ultrasound
Chapter
39
Chest, Thyroid, Neonatal Brain
Parathyroid, Ultrasound
and
William E. Brant
CHEST US is an excellent supplement to plain film radiography and CT for the problem-solving evaluation of the chest and to guide interventional procedures in the thorax (1, 2, 3, 4, 5, 6). US can image into and through pleural effusions and lung consolidation to evaluate the thorax that is opacified on plain radiographs. Its portability allows evaluation of critically ill patients who are impractical to move for a CT. US examination of the chest must always be correlated with available chest radiography.
Pleural Normal
Space US
Anatomy
Air in the lungs completely reflects the US beam and prohibits examination deeper into the chest. However, when pleural fluid displaces air-filled lungs away from the chest wall, disease in the pleural space can be optimally evaluated with US. The pleural space is examined by a direct intercostal approach, with the US transducer applied directly to the chest, or by an abdominal approach, imaging through the diaphragm from the abdomen. The ribs are used as
2596
sonographic landmarks for direct chest imaging (Fig. 39.1). A linear array transducer applied to the chest wall shows the ribs as curving echoes that cast acoustic shadows. The visceral pleura/air-filled lung interface is seen within 1 cm of the rib echo as a bright echogenic surface that moves with respiration (the “gliding sign―). The moving lung surface is well visualized when the transducer is turned to parallel the intercostal space. The tiny normal amount of fluid in the pleural space is seen just superficial to the gliding pleura. From the abdomen, the diaphragm is seen as a bright curving interface because of complete sound reflection from the air-filled lung above it (Fig. 39.2). Organs beneath the diaphragm (liver, spleen) are artifactually reproduced above the diaphragm because of multipath sound reflection (the “mirror image― artifact). Pleural
fluid displaces the lung away from the chest wall, allowing
visualization of the pleural space (Figs 39.1C, 39.2B). Most pleural fluid is anechoic, or hypoechoic with floating particulate matter (7) . The fluid separates the visceral and parietal pleural surfaces. From an abdominal approach, hypoechoic fluid is seen above the diaphragm, the inside of the thorax is visualized, and the mirrorimage artifact is not present. Septations not evident on CT are commonly visualized by US. Collapsed or consolidated lung moves with respiration within the fluid in the pleural space. Fluid that is echogenic, contains floating particles or layering debris, or is septated is an exudate (Fig. 39.3). Fluid that is anechoic may be a transudate, P.1003 P.1004 exudate, or even empyema. Loculations of pleural fluid and suspected empyemas can be localized and evaluated, with US visualization used to guide needle aspiration and drainage catheter placement.
2597
FIGURE 39.1. Pleural Space: Intercostal Scan. A. Longitudinal US image of the chest shows a rib (R) and its acoustic shadow (between arrowheads). The pleural space is approximately 1 cm deep to the surface of the rib (arrow) . Intercostal muscle (m) is seen between the ribs. B . Aligning the transducer parallel to the ribs in the intercostal space enables improved visualization of the pleural space (arrow). The visceral pleura/air-filled lung interface (black arrowhead) is identified by its movement with respiration: the “gliding sign.― The visceral pleura is separated from the parietal pleura (white arrowhead) by a thin layer of pleural fluid in the pleural space (arrow). The air-filled lung is obscured by reverberation artifact
2598
(Rev). C . A pleural effusion (e) separates the visceral pleura (black arrowhead) from the parietal pleura (white arrowhead). m, intercostal muscle; S, subcutaneous fatty tissue.
FIGURE 39.2. Pleural Space: Abdominal Scan. A. Examination of the chest can be performed from an abdominal approach, using the liver or spleen (Sp) as a sonographic window. The diaphragm is seen as a bright curving line (arrowhead). Normal air-filled lung causes the spleen to be reproduced as a mirror-image artifact (MI) above the diaphragm. LK, left kidney. B . A pleural effusion (e) eliminates the mirrorimage artifact and allows visualization of the chest, wall characterized by ribs and rib shadows (arrow) through the diaphragm (arrowhead) and pleural space. L, Liver.
2599
FIGURE 39.3. Echogenic Pleural Effusion. An empyema associated with a right lower lobe pneumonia appears on US as an echogenic effusion (e). Innumerable moving floating particles were observed within the fluid on real-time US examination. The liver (L) is very similar in echogenicity. The diaphragm (arrowhead) is seen as a curving, brightly echogenic line.
Pleural thickening complicates inflammatory and malignant disease of the thorax. US demonstrates uniform, undulating, or plaquelike thickening of the pleura (Fig. 39.4). The visceral pleura is easily evaluated. The parietal pleura is partially obscured by reverberation artifact in the near field.
Pleural
Masses
Pleural metastases or tumors such as mesotheliomas are seen as
2600
nodular pleural thickening or hypoechoic soft tissue masses in the pleural space projecting from the pleural surface. Pneumothorax can be diagnosed by US. Pneumothorax produces a highly echogenic reflective line very similar to that of air-filled lung but lacking the “gliding sign― associated with respiratory movement. Pneumothorax is also indicated by loss of visualization of a previously visualized lung lesion that occurs during an invasive procedure.
Lung Normal
Parenchyma US
Anatomy
The normal air-filled lung with its covering visceral pleura completely blocks transmission of US into the thorax. The gliding visceral surface of the lung is easily seen, but reverberation artifact is displayed deep to that surface. However, consolidation, atelectasis, or tumor that extends to the visceral pleural surface produces a window for US examination. When scanning the thorax from the abdomen, the normal air-filled lung produces a mirror-image artifact.
2601
FIGURE 39.4. Pleural Thickening. Intercostal US image demonstrates a moderate volume pleural effusion (e). The visceral pleura (between arrowheads) is thickened because of chronic inflammation. The parietal pleura is obscured in the near field by reverberation artifact (Rev). The air-filled lung (Lu) is brightly echogenic.
Consolidation refers to filling of the air spaces of the lung with fluid and inflammatory cells. This process “solidifies― the lung and provides a medium for sound transmission (Fig. 39.5). The consolidated lung appears solid and hypoechoic, with echogenicity similar to that of liver tissue. Sonographic air bronchograms and sonographic air alveolograms may be seen within the consolidated
2602
lung. Air-filled bronchi produce bright branching linear reflections. Air trapped in alveoli surrounded by consolidated lung produces globular bright echoes with comet-tail artifacts. Sonographic fluid bronchograms appear as anechoic fluid-filled tubes extending from the hilum of the lung. Color flow US demonstrates pulmonary vessels extending through the consolidated lung.
FIGURE
39.5. Lung
Consolidation. US image obtained using
the spleen (Sp) as a sonographic window in a patient with left upper quadrant pain reveals an unsuspected pneumonia in the left lower lobe of the lung (Lu). Inflammatory fluid and cells solidify the lung, replacing air and allowing visualization of the chest wall (black arrow) through the airless lung. Sonographic fluid bronchograms (white arrow) are seen within the pneumonia. The diaphragm (arrowhead) produces a bright, curving echo.
P.1005
2603
Atelectasis Collapse of the airspaces with absorption of air also results in solidification of the lung. With atelectasis, the lung volume is decreased and bronchi and pulmonary blood vessels are crowded together. Collapsed lung always accompanies large pleural effusions (Fig. 39.6). The atelectatic lung is wedge-shaped and sharply defined by its covering pleura. Lung
masses surrounded by air-filled lung are not visualized by US,
but those that extend to the visceral pleura or are accompanied by peripheral consolidation or atelectasis may be seen and evaluated (Fig. 39.7). US guidance may be effectively used to aspirate or biopsy lung masses in areas that are difficult to access with CT or fluoroscopy. Central tumor necrosis, hemorrhage within tumors, and lung abscesses are effectively evaluated. Pulmonary
sequestration is a congenital partition of lung tissue
that does not communicate with the bronchial tree. Most occur at the lung base. Intralobar sequestrations are within the visceral pleura. Extralobar sequestrations are invested by their own separate pleura. US is used to confirm the diagnosis by demonstration of a feeding artery arising from the aorta. Extralobar sequestrations drain via a systemic vein, whereas intralobar sequestrations connect to the pulmonary veins.
2604
FIGURE
39.6. Atelectasis. A transverse image through the liver
(L) reveals a pleural effusion (e) surrounding a tongue of collapsed lung (Lu). The patient also has ascites (a). The diaphragm (arrowhead) produces a thin, curving, bright echo. The chest wall (arrow) produces a thick, curving, bright echo.
Mediastinum Normal
US
Anatomy
The superior and anterior mediastinum are effectively evaluated with US using a parasternal or supramanubrial approach. The posterior mediastinum is less accessible because of spine and lung. Large lesions create sonographic windows to the mediastinum. Imaging downward into the superior mediastinum from just above the sternal manubrium demonstrates the innominate veins and the arteries arising from the aortic arch. Doppler US assists in the identification of vessels.
2605
Vascular
Lesions. Elongation and tortuosity of the brachiocephalic
artery are common causes of mediastinal widening in older adults. This diagnosis is easily confirmed by US, which can also exclude other masses of the superior mediastinum.
Mediastinal
Masses
Thymic masses, substernal extension of thyroid enlargement, adenopathy, and other mediastinal masses are effectively demonstrated by US, which can confirm their cystic or solid nature and vascularity. Lesions that can be visualized by US can usually be biopsied using US guidance to avoid critical structures (Fig. 39.8) . Continuation of thyroid tissue into the mediastinum is a straightforward diagnosis. Enlarged lymph nodes are usually homogeneous and hypoechoic. Confluent adenopathy caused by lymphoma produces a solid, homogeneous, hypoechoic mass that encompasses and displaces blood vessels.
2606
FIGURE 39.7. Peripheral Lung Mass. Intercostal US scan shows a 3-cm lung mass (m) abutting the pleural surface. Note how the bright echo from the visceral pleura/air-filled-lung interface (arrowhead) is obliterated over the mass (arrow). The hypoechoic mass stands out because it is surrounded by echogenic, air-filled lung (Lu). Fine-needle biopsy, which was precisely guided by US visualization into the mass without transgressing air-filled lung, confirmed squamous cell carcinoma. This patient has abundant subcutaneous fatty tissue (S).
P.1006
THYROID Imaging of the thyroid gland is a controversial topic (8,9). Thyroid nodules are exceedingly common, although thyroid cancer is uncommon and death from thyroid malignancy is rare. Highresolution US is extremely sensitive in detecting thyroid nodules; however, imaging signs to differentiate benign from malignant lesions are nonspecific and unreliable (1 0). This creates a recurring clinical problem of what to do with the many nodules detected sonographically. Because of this difficulty, the indications for US of the thyroid are debated. US is used to precisely guide percutaneous biopsy of thyroid nodules, to screen patients at high risk for thyroid cancer, to identify recurrent disease in patients with known thyroid cancer, and to determine whether palpable nodules arise from the thyroid gland. CT and MR supplement US by staging of invasive thyroid cancers, evaluating for postoperative recurrence of thyroid cancer, and demonstrating extension of goiter into the thorax. Radionuclide imaging, discussed in a subsequent chapter, evaluates the physiologic function of the gland and determines the activity of nodules.
2607
FIGURE 39.8. Mediastinal Mass. A left parasternal US image shows a large solid mediastinal mass (T). US-guided fine-needle and core biopsies were easily performed and confirmed a malignant thymoma.
Normal
US
Anatomy
The thyroid gland consists of paired lobes of nearly equal size (5×2×2 cm) connected across the trachea by a thin thyroid isthmus (Fig. 39.9). The thyroid parenchyma is homogeneous, with fine medium-level echogenicity greater than muscle. Anatomic landmarks include the midline air-filled trachea, which casts an air shadow; the common carotid artery and internal jugular vein, which parallel the lateral edge of the thyroid lobes; the longus colli muscles posteriorly; and the sternohyoid, sternothyroid, and sternocleidomastoid muscles anteriorly. Small pools of colloid (colloid cysts) are routinely visualized within the normal gland. The thyroid P.1007 lobes are often mildly asymmetric in size. The esophagus may
2608
protrude from behind the trachea on the left side and must not be mistaken for a thyroid or parathyroid mass or lymph node (see Fig. 39.15). The superior thyroid artery and vein are imaged between the upper pole of the thyroid and the longus colli. The recurrent laryngeal nerve and inferior thyroid artery and vein are seen posterior to the lower poles. The thyroid is easily imaged with the patient in a supine position, with the neck extended by placement of a pillow beneath the shoulders.
FIGURE
39.9. Normal
Thyroid. Symmetric lobes of
homogeneous thyroid tissue (t) are seen on either side of the trachea (T) on this transverse image. Anatomic landmarks include the carotid arteries (c) and longus colli muscle (m).
TABLE 39.1 Differentiation of Benign and Malignant Thyroid Nodules
2609
Indeterminate Evidence Found Evidence Favoring Benign Nodule
Evidence Favoring Malignant Nodule
Extensive cystic component
Irregular
Sharply
Poor
defined
contour
margination
With Both Benign and Malignant Nodules Hypoechoic nodule
Isoechoic
nodule
margin
Peripheral calcification
Microcalcifications
Solid
Homogeneously hyperechoic
Size >4 to 5 cm
Amorphous dense calcification
Comet tail
Single cold nodule
Echolucent
artifacts
on radionuclide scan (15%
Multiple
nodules
nodule
halo
malignant)
Age Table of Contents > Section IX - Ultrasonography > Chapter 40 Vascular Ultrasound
Chapter
40
Vascular
Ultrasound
Raymond S. Dougherty William E. Brant Spectral Doppler US and color flow vascular imaging supplement grayscale US by identifying blood vessels, confirming the presence of blood flow and its direction, detecting vessel stenosis and occlusion, assessing the perfusion of organs and tumors, and characterizing blood flow dynamics to detect physiologic abnormalities (1 , 2 , 3 , 4 , 5 ). This chapter reviews the basics of vascular US examination and Doppler interpretation.
DOPPLER
BASICS
Doppler effect refers to the change in the frequency of sound waves that occurs on account of the motion of a sound source, a sound reflector, or a sound receiver. Johann Doppler of Salzburg, Austria, described this phenomenon in 1842. In medical diagnosis, the Doppler effect is used to confirm blood flow by detecting the change in frequency of US waves that occurs when sound is reflected from moving clumps of red blood cells (RBCs). The echoes reflected from RBCs are very weak, with a signal intensity up to 10,000 times less than that of contiguous soft tissue; thus, Doppler US instruments require a high sensitivity to weak signals, and instrument settings must be routinely optimized. Doppler
shift is the change in frequency between the US waves
2639
emitted by the transducer and the US waves returning to the transducer after reflection from moving RBCs (Fig. 40.1 ). This shift in sound frequency results from the Doppler effect. The reflected sound frequency increases when blood flow direction is toward the Doppler signal and decreases when the direction is away from the Doppler signal. An increase in frequency is termed a positive Doppler shift; the sound waves are compressed by encountering RBCs moving toward the sound source. A decrease in frequency is termed a negative Doppler shift , because the reflected sound waves are stretched by RBCs moving away from the sound source. The presence of a Doppler shift within a blood vessel confirms the presence of blood flow. The direction of the Doppler shift toward higher or lower frequency indicates the direction of blood flow. Doppler shift frequencies are within the range of human hearing and produce distinctive audible sound patterns that characterize normal and abnormal arterial and venous blood flow.
Doppler
Equation
The Doppler equation describes, in mathematical form, the relationship between the Doppler P.1020 frequency shift (ΔF) and the velocity (V) of the moving RBCs that produce the shift.
2640
FIGURE 40.1. Doppler Frequency Shift. The transmitted Doppler US beam (Ft) encounters red blood cells moving toward it within a visualized blood vessel. The red blood cell motion causes an increase in frequency of the returning echo (Fr) because of the Doppler effect. The US instrument detects and measures the frequency of the returning Doppler signal, confirming the presence of blood flow and its direction by the presence and direction of the Doppler frequency shift.
ΔF = (Fr - Ft) = the Doppler frequency shift Ft = frequency of the transmitted Doppler US beam (the transducer frequency) Fr = frequency of the reflected US beam (shifted by RBC motion) V = RBC velocity (blood flow velocity) θ = the Doppler angle = the angle between the direction of blood flow and the direction of the Doppler US beam C = speed of sound in tissue (assumed to be constant at 1,540 m/s) The frequency shift (ΔF) is proportional to the following: (1 ) the velocity (V) of the moving RBCs; (2 ) the frequency of the transmitted Doppler US beam (Ft); and (3 ) the cosine of the angle between the incident Doppler US beam and the direction of blood flow. This angle is called the Doppler angle and is symbolized by the Greek letter theta (θ). The direction of blood flow is assumed to be parallel to the walls of the visualized blood vessel being interrogated (Fig. 40.2 ). The Doppler US beam can be steered by controls on the US unit. The direction of the Doppler beam is indicated on the US image by a dotted or dashed line. The fact that the Doppler frequency shift is directly proportional to the cosine of the Doppler angle has important implications (Table 40.1 ). First, the largest frequency shift—that is, the largest Doppler signal—will be obtained when the Doppler US beam is
2641
directed straight down the barrel of the vessel (θ = 0°, cosine 0° = 1). Second, no Doppler shift will occur when the Doppler US beam is directly perpendicular to blood flow (θ = 90°, cosine 90° = 0). Small errors in Doppler angle estimation cause only small errors in velocity calculations at small Doppler angles, but small errors in Doppler angle estimation cause large errors in velocity calculations at angles close to 90°. As a general rule, Doppler scanning should be performed to keep Doppler angles at 60° or less.
FIGURE
40.2. Doppler
Angle. The Doppler angle (θ) is defined as
the angle between the Doppler US beam and the direction of blood flow, which is assumed to be parallel to the walls of the blood vessel. The Doppler sample volume is indicated by two parallel lines. The Doppler angle indicator is displayed as a dashed line within the sample volume. The US unit has a control knob that is used to align the Doppler angle indicator with the blood vessel walls.
By algebraic manipulation we can rewrite the Doppler equation as follows:
0°
2642
1 10° 0.98 20° 0.93 30° 0.87 40° 0.77 50° 0.64 60° 0.50 70° 0.34 80° 0.17 90° 0 Angle
Cosine
TABLE 40.1 Cosine Values
2643
FIGURE
40.3. Duplex Doppler US. US image shows the Doppler
spectrum of the common carotid artery. The vertical scale shows blood flow velocity in meters per second. The horizontal scale shows time in seconds. The Doppler trace demonstrates peak velocities in systole (S) and low flow velocities in diastole (D). A 2-mm Doppler sample volume (curved arrow ) is placed by the sonographer in the midportion of the artery visualized by real-time US. Only Doppler shifts originating from this sample volume are analyzed for display. An estimated Doppler angle of 50° is communicated to the US unit computer by aligning the angle indicator (open arrow ) parallel to the vessel walls.
P.1021 The US unit detects and measures the frequency of the Doppler beam reflected from moving RBCs (Fr) and calculates the Doppler frequency shift (ΔF = Ft - Fr). The transmission frequency (Ft) is determined by the transducer chosen to perform the examination. The speed of sound in human tissue is assumed to be constant (C). The operator communicates the Doppler angle to the US unit by aligning the Doppler angle “wings― to be parallel with the walls of the vessels examined (Figs. 40.2 , 40.3 ). Because the depth of a structure in an US image is measured by the
2644
time delay between transmission of the US into tissue and the return of the echo from the structure, Doppler information can be limited to a selected Doppler “sample volume― by use of a “time window.― The length of the time window determines the size of the sample volume, and the time delay of the time window determines its depth. Thus, we can restrict Doppler information to a small portion of a single visualized vessel. On most Doppler US units, the size and location of the Doppler sample volume is indicated by two short parallel lines along the Doppler beam indicator line (Figs. 40.2 , 40.3 ). Simultaneous grayscale imaging plus Doppler scanning is called duplex US. Both spectral and color Doppler imaging are examples of duplex imaging.
Doppler
Spectral
Display
Returning Doppler signals are processed using a fast Fourier transform spectrum analyzer that sorts the range and mixture of Doppler frequency shifts into individual components and displays them as a function of time on a velocity (or frequency shift) scale (Fig. 40.3 ). Analysis is performed rapidly enough to be displayed in real time. The horizontal scale (x axis) of the Doppler spectrum represents time in seconds. The vertical scale (y axis) represents blood flow velocity in m/s or cm/s. Because velocity and Doppler frequency shift are directly related mathematically, Doppler frequency shift may alternatively be used on the vertical scale without changing the appearance of the Doppler spectrum. Since blood flow velocity provides the most diagnostically useful information, velocity is the usual choice for the vertical axis. Each pixel (dot) in the spectral display represents a group of RBCs moving at a specific velocity at a given moment in time. The more RBCs that are moving at that specific velocity and time, the brighter the pixel. Flow toward the Doppler beam (positive frequency shift) is displayed above beam (negative baseline. Peaks and periods of
the zero baseline, and flow away from the Doppler frequency shift) is displayed below the zero of higher velocity occur during ventricular systole, lower velocity represent ventricular diastole.
2645
Spectral
Waveforms
Different blood vessels have unique flow characteristics that can be recognized by the Doppler spectral waveform (Doppler “signature―) that they produce (6 ,7 ). Factors that affect the appearance of the spectral waveform include cardiac contraction, vessel compliance, and downstream vascular resistance. Cardiac arrhythmias are reflected in the periodicity of the systolic peaks and the velocities reached during each cardiac contraction. A major determinant of the spectral waveform's appearance is the resistance to blood flow offered by the vascular bed supplied by the artery being studied. Arteries can be categorized as high resistance or low resistance, based upon their Doppler spectral waveform. Highresistance spectral waveforms are characterized by velocities that increase sharply with systole, decrease rapidly with cessation of ventricular contraction, and show little or no forward flow during diastole (Fig. 40.4 ). Blood flow direction may reverse briefly during early diastole, P.1022 producing a triphasic waveform. Blood flow in high-resistance arteries is always under considerable pressure and encounters constricted arterioles that impede forward blood flow. Pulse pressures traveling down the arterial tree are highly reflected, which results in minimal flow to the capillary bed during diastole. Diastolic flow velocity is low, absent, or reversed, and pulse pressure is high. The ratio of systolic velocity to diastolic velocity (pulsatility) is high. Arteries that normally show a high-resistance Doppler waveform include arteries that supply primarily skeletal muscle at rest, including the iliac, femoral, popliteal, subclavian, and brachial arteries. The external carotid artery waveform is relatively high resistance in appearance.
2646
FIGURE
40.4. High-Resistance
Doppler
Spectrum. A high-
resistance waveform is characterized by rapid systolic upstroke (straight arrow ), low flow velocities during diastole (curved arrow ), and, commonly, reversal of flow direction (arrowhead ) in early diastole. This Doppler spectrum was obtained from the common femoral artery.
FIGURE 40.5. Low-Resistance Doppler Spectrum. A lowresistance waveform is characterized by relatively high flow velocities throughout diastole (curved arrow ). The narrow spectrum and clean systolic window (straight arrow ) are characteristic of laminar blood flow. This Doppler spectrum was obtained from the internal
carotid
artery.
Low-resistance spectral waveforms are characterized by a slower increase in flow velocity with onset of systole and a gradual decrease in velocity during diastole, with continued forward flow throughout the cardiac cycle (Fig. 40.5 ). Arteries that supply vital organs characteristically have a low-resistance waveform. These include the mesenteric fasting and opening of
internal carotid, hepatic, and renal arteries. The superior artery waveform has a high-resistance pattern during a low-resistance pattern after eating, reflecting the intestinal tract arterioles and increased intestinal blood
2647
flow induced by food in the gut. The common carotid artery, with 70% of its blood flow going to the internal carotid artery, has a lowresistance spectral waveform.
Laminar
Blood
Flow
Most normal arteries and large veins have a laminar pattern of blood flow. Blood flow velocity is highest at the center of the vessel and progressively diminishes closer to the vessel wall (Fig. 40.6 ). The Doppler waveform of laminar flow is characterized by a “narrow spectrum―—a narrow band of blood flow velocities throughout the cardiac cycle with a “window― beneath the spectral trace in systole (Fig. 40.4 ). Large arteries such as the aorta have “plug― flow characterized by a uniform flow velocity extending from the center to near the vessel wall. At vessel bifurcations, the division of blood flow results in a small area of normal reversed blood flow near the vessel wall, opposite the flow divider (Fig. 40.7 ). Tortuous blood vessels demonstrate normal slowing of blood flow on the inner aspect of the curve, with acceleration of blood flow on the outer aspect of the curve. The highest velocities are seen at the outer aspect of the curving vessel, rather than at midlumen. Blood flow velocity returns to a laminar distribution a short distance downstream from the curve.
FIGURE 40.6. Laminar Blood Flow. Blood flow in most normal arteries is arranged in an orderly layering pattern, with the highest velocity at midstream and the lowest velocity near the vessel wall.
2648
Disturbed
Blood
Flow
Turbulent and disturbed spectral waveforms are usually, but not always, indicative of pathologic changes in blood flow. Disturbed blood flow is a loss of the normal orderly laminar flow pattern. Characteristic spectral Doppler signs of disturbed blood flow are: increased velocity, spectral broadening, simultaneous P.1023 forward and reverse flow, and fluctuations of flow velocity with time (5 ). Peak systolic velocity increases with severity of vessel stenosis. Spectral broadening is widening of the spectral waveform that reflects a broader range of flow velocities within the Doppler sample volume. Spectral broadening increases with the severity of flow disturbance. However, normal spectral broadening occurs when the size of the Doppler sample volume is large compared to the size of the vessel, or when the sample volume is placed near the vessel wall instead of midlumen. Flow velocity fluctuation and simultaneous forward and reverse flow characterize turbulence. Turbulence is most pronounced just downstream from a severe vessel stenosis, where eddy currents are produced as the high-velocity flow slows and occupies a larger vessel area.
FIGURE
40.7. (Color Plates) Normal Flow Reversal at
Bifurcation. Flow in the internal carotid artery is shown in red, with
2649
areas of higher flow velocity shown in yellow. A normal area of blood flow reversal (short arrow ) is seen in the carotid bulb. Note how the true color change is outlined in black. The color Doppler interrogation area is marked on the image by the box outlined in white (arrowhead ). Higher velocity in the middle of the vessel, shown in yellow (long arrow ), is indicative of laminar blood flow in the
artery.
Velocity
Ratios
Blood flow velocity calculations are dependent upon accurate estimation of the Doppler angle. When the Doppler angle cannot be determined because of poor visualization of the interrogated blood vessel or the vessel's tortuosity (as with the umbilical artery in the cord), velocity cannot be accurately calculated. When the Doppler angle indicator is not displayed, the US instrument calculates Doppler velocities using Doppler equation by assuming that the Doppler angle is 0° (cosine 0° = 1). Velocity ratios can be calculated from the spectral waveform and can be used to estimate vascular resistance and hemodynamics. The ratios are independent of absolute velocity measurements. The velocity ratios in common use are listed in Table 40.2 .
Assessing
Arterial
Stenosis
Acute narrowing of the blood vessel lumen disturbs laminar flow. Doppler characterization of vessel stenosis is based upon changes in blood flow pattern and velocity. To assess the degree of stenosis, Doppler spectra are routinely obtained in three areas of the vessel lumen (Fig. 40.8 ): (1 ) proximal to stenosis, (2 ) at the point of maximal stenosis, and (3 ) 1 to 2 cm downstream from the stenosis. Laminar flow is generally present proximal to the stenosis. Within the stenotic zone, velocity is increased but usually remains laminar. The severity of stenosis correlates best with the highest blood flow velocity during peak systole. The highest velocity may be in a very small region, and a careful search of the vessel is necessary. In the
2650
poststenotic zone, flow spreads out, causing turbulence and eddy currents to occur and produce broadening of the Doppler spectrum. Downstream from severe stenosis (>50%) the Doppler signals are dampened, producing the parvus-tardus waveform. Flow velocities are low (parvus) with a slow systolic upstroke (tardus) (see Fig. 40.22 ) (8 ,9 ).
TABLE 40.2 Doppler Velocity Ratios
FIGURE 40.8. Assessment of Arterial Stenosis. To assess a vessel plaque for stenosis, Doppler spectra are obtained: (1 ) proximal to the plaque, where blood flow velocity is normal and flow is laminar; (2 ) in the area of the plaque where flow usually remains laminar but where flow velocity is at maximum; and (3 ) downstream from the plaque, where turbulence and eddy currents
2651
are
detected.
Color Flow US Currently, three different techniques are used to produce color flow US images. Color Doppler imaging (CDI) superimposes Doppler flow information on a standard grayscale B-mode real-time US image (1 0 ,1 1 ). The B-mode image is displayed in shades of gray, and the Doppler flow information is displayed on the same image in color (Fig. 40.9 ). Most of the same principles and limitations of spectral Doppler apply to color Doppler imaging. Color Doppler energy (CDE), also termed power
Doppler ,
displays color flow information obtained from integration of the power of the Doppler signal rather the Doppler frequency shift itself (1 , 1 2 , 1 3 , 1 4 ). Power Doppler displays information more directly related to the number of moving RBCs than to their velocity (Fig. 40.10 ). CDE is relatively angle independent and is more sensitive to slow flow than CDI. The third method of color flow imaging is called color
velocity
imaging (CVI) (1 5 ). Color velocity imaging is not a Doppler technique; instead, CVI tracks the movement of structure in the US image from one moment to the next. In effect, CVI images blood like a soft tissue and stores a unique texture pattern for comparison with a subsequent image. Movement of the blood flow speckle pattern is displayed as a color flow image.
2652
FIGURE 40.9. (Color Plates) Power Doppler (Color Doppler Energy [CDE]) Image. Power Doppler image of a transplant kidney in the transverse plane shows the increased sensitivity of CDE. Blood vessels are detected more peripherally in the kidney. However, power Doppler lacks the capability to show blood flow direction. Arteries and veins are displayed in the same color shades. The CDE area of interrogation is indicated by the white box (thin arrow ). The CDE color map is shown at the left (squiggly arrow ).
P.1024 On the usually usually coloring
color Doppler, image flow directed toward the transducer is colored red, whereas flow away from the transducer is colored blue. The operator may arbitrarily change the of the Doppler information. The color map used is displayed
as part of the color US image. Faster blood flow velocities are colored in lighter shades, whereas slower blood flow is colored in darker shades. Color shading is dependent on mean velocities, not peak velocities. Thus, peak velocities cannot be estimated from the color image alone and must be determined from spectral Doppler. A normal laminar flow pattern will demonstrate lighter shades in the midstream and darker shades near the vessel walls, reflecting rapid
2653
flow in the middle of the vessel and slower flow near its walls. Disturbed flow, such as turbulence, is indicated by a wide range of colors in a scrambled pattern. Changes in color within a blood vessel on a color flow US image may be caused by: (1 ) change in the Doppler angle, (2 ) change in blood flow velocity, (3 ) aliasing, or (4 ) artifact. A change in Doppler angle causes a change in Doppler frequency shift, which, on a color flow image, produces a change in the color displayed. Variations in the Doppler angle may be caused by divergence of US beams emanating from sector or curved array transducers, a blood vessel curving through the color image, or a combination of both. Color flow images are used to detect changes in blood flow velocity for further analysis by spectral Doppler. To interpret a color flow image, inspect the color map for color display orientation, then analyze the image for variations in Doppler angle and blood flow velocity.
FIGURE
40.10. (Color Plates) Aliasing on Spectral Doppler.
The high-velocity peaks (arrow ) of the spectral Doppler display are cut off at the top, “wrapped around,― and displayed at the bottom of the spectral display. The spectral Doppler scale on the left is set with a Nyquist limit of 0.30 m/s, too low for the peak velocities encountered within the interrogated blood vessel.
2654
Doppler
Artifacts
A variety of artifacts distort Doppler information and limit the information
provided.
Aliasing is a limitation of pulsed Doppler US that occurs with both spectral and color flow Doppler (1 1 ,1 6 ). Aliasing happens with high-velocity blood flow and improper velocity scale and baseline settings. Aliasing on spectral displays is seen as a “wraparound― of peak velocities to the opposite end of the scale (Fig. 40.11 ). The highest velocities are cut off one side of the scale and artifactually displayed on the opposite side of the scale. Aliasing on color Doppler “wraps around― high velocities onto the opposite color scale (Fig. 40.12 ). For example, velocities too high for the red-scale setting are artifactually displayed as shades of blue. Color aliasing must be distinguished from true color changes caused by flow reversal or changes in the Doppler angle. True color changes are always surrounded by a black border, whereas color shifts related to aliasing lack this black border. Aliasing occurs when the pulse Doppler sampling rate is too low for a given Doppler signal frequency, thus resulting in an inaccurate frequency measurement. The US instrument measures the frequency of returning Doppler signal, piece by piece, by a series of pulses. The rate at which pulses can be transmitted (the pulse
repetition
frequency P.1025 [PRF]) is limited by the depth of the vessel interrogated. Deeper vessels require more time for the US beam to travel to the vessel and for the echo to return. To avoid aliasing, the PRF must be at least twice the frequency of the signal to be detected. The maximum frequency that can be accurately detected without aliasing is called the Nyquist limit and is equal to half the PRF. The Nyquist limit is displayed at the top and bottom of the spectral Doppler scale and the color map. On color Doppler images, aliasing may be helpful and serve as a tag for high velocities associated with significant stenosis. Aliasing may be eliminated by proper adjustment of the Doppler scale and baseline settings, by using a lower Doppler
2655
transmission frequency, or by increasing the Doppler angle.
FIGURE
40.11. (Color Plates) Aliasing on Color Doppler
Imaging. The color map (squiggly arrow ) is set for red/yellow color at the top of the scale to indicate flow toward the Doppler beam and for blue/green color at the bottom of the scale to indicate flow away from the Doppler beam. The direction of the Doppler beam is indicated by the parallel sides (thin arrow ) of the color Doppler sample volume box, shown in white. The dominant color within the visualized blood vessel is red/yellow, indicating that the direction of blood flow is from right to left. The higher velocity of blood flow in the center of the blood vessel (wide arrow ) exceeds the lowvelocity scale setting (Nyquist limit = 0.020 m/s) and is displayed in green, the high-velocity color on the opposite end of the color scale. The lack of a black border around the color shift is a sign of aliasing.
Incorrect
Doppler
Gain
When the Doppler gain is set too low, Doppler information may be lost and blood flow may not be demonstrated. The color Doppler image with gain that is too high demonstrates color in non-flow
2656
areas and random color noise. Correct gain settings are attained by turning up the gain setting until noise appears on the image and then slightly lowering the setting.
Velocity
Scale
Errors
Velocity range settings that are too high may obscure low-velocity flow, which is lost in noise and the wall filter near the baseline. Vessels that are patent but with very slow flow may be considered thrombosed. When velocity scale settings are too low, aliasing occurs. Such aliasing is corrected by adjusting scale and baseline settings.
FIGURE
40.12. (Color Plates) Common Carotid Artery (CCA)
Occlusion . A . Longitudinal color Doppler image of the right CCA showing echogenic thrombus (arrow ) with absence of color, indicating no blood flow. B . Spectral Doppler confirms no CCA flow. The irregular signal near the baseline is due to noise. C . Spectral Doppler of the right external carotid artery demonstrates retrograde flow (toward the heart). D . Spectral Doppler of the right internal carotid artery shows antegrade flow (toward the head). See text for explanation.
Color
Flash
2657
Any motion of a reflector relative to the transducer produces a Doppler shift. Rapid movement of the transducer itself may produce a Doppler shift and a flash of color projected over the grayscale image. Most instruments incorporate motion discriminators that suppress color flash in hyperechoic but not in hypoechoic areas. Color flash is accentuated in cysts, the gallbladder, and other hypoechoic nonvascular structures. accentuate color flash.
Tissue
Vibration
High
color
sensitivity
settings
Artifact
Tissue vibration may produce color display in perivascular tissues, indicating flow where none is present. Tissue vibration artifact is produced in non-flow areas by bruits, arteriovenous fistulas, and shunts.
Fluid
Motion
Color signal can be produced during CDI by motion of fluids other than blood. Motion of fluid within cysts and bowel may be misinterpreted as blood flow. Ureteral peristalsis produces a jet of color in the bladder that confirms patency of the ureter.
CAROTID
ULTRASOUND
Stroke Approximately 700,000 strokes occurred in the United States in 2002, resulting in 163,000 deaths. It is projected that the direct and indirect public healthcare cost of P.1026 stroke in 2005 will approach $56 billion. Atherosclerotic lesions of the extracranial vessels are estimated to cause 75% of strokes that have either a thrombotic or embolic cause (1 7 ).
2658
FIGURE
40.13. (Color Plates) Color Doppler Image. This color
Doppler image of blood flow in the kidney displays arteries entering the kidney in red (wide arrow ) and the veins leaving the kidney in blue (arrowhead ). The color Doppler area of interrogation (sample volume) is etched by thin green lines (thin arrow ). Any detected Doppler shift within this area is displayed in color. The kidney and surrounding tissue are displayed in grayscale. The color Doppler map on the left is set to show blood flow toward the Doppler beam in red/yellow at the top of the scale and blood flow away from the Doppler beam in blue/green at the bottom of the scale. The zero flow baseline is represented by the black bar (squiggly center of the color map.
arrow ) in the
During the last 15 years, the management of atherosclerotic carotid disease has changed significantly. In 1991, two randomized prospective multicenter studies, the North American Symptomatic Carotid Endarterectomy Trial (NASCET) (1 8 ) and the European Carotid Stenosis Trial (ECST) (1 9 ), demonstrated a clear benefit of carotid endarterectomy in patients with an internal carotid artery (ICA) stenosis ≥70% of the diameter, as defined by conventional
2659
angiography. An important difference between the NASCET and ECST studies is the method of measuring the stenosis (Fig. 40.13 ). NASCET measured the degree of narrowing as the ratio of the diameter of the stenosis to the diameter of the normal ICA distal to the stenosis on catheter angiography. The ECST measured carotid stenosis in the more traditional fashion, by comparing the residual lumen diameter to an approximation of the original vessel diameter. A 70% ECST stenosis is approximately equal to a 50% NASCET stenosis. NASCET found an unequivocal difference between the risk of stroke in patients receiving the best medical care alone and those undergoing endarterectomy. The risk of ipsilateral stroke at 2 years is 26% for those treated medically and 9% for those treated surgically. Additional data from NASCET published in 1998 showed modest benefits of endarterectomy for stenoses measuring 50% to 69% as long as the rate of the institution's serious surgical complication rate was 40 cm/s Increase >60 cm/s Main portal vein velocity Peak Table of Contents > Section X - Musculoskeletal Radiology > Chapter 43 - Skeletal Trauma
Chapter
43
Skeletal
Trauma
Clyde A. Helms Most of the differential diagnoses in skeletal radiology that I use are geared to be 95% inclusive; that is, the correct diagnosis will be mentioned 95% of the time. The yield can be increased by lengthening the list, but if the list gets too long, it will become unwieldy and less useful for the clinician. In trauma cases, however, being right 95% of the time is not good enough. Missing the correct diagnosis 5% of the time is unacceptable. Fractures simply should not be missed. Before reviewing specific examples, the reader should keep a few key points in mind concerning radiology of trauma. First, have a high index of suspicion. Every radiologist in the world has missed fractures on radiographs because he or she was not sufficiently attuned to the possible presence of a fracture. Often, the history is either nonexistent or misleading, so that the anatomic area of concern is overlooked. When in doubt, examine the patient. Orthopaedic surgeons rarely miss seeing fractures on radiographs because they have examined the patient, they know where the patient hurts, and they have a high index of suspicion. Second, always get two radiographs at 90° to each other in every trauma case. A high percentage of fractures are seen only on one view (the anteroposterior [AP] or the lateral) and will therefore be missed unless two views are routinely obtained. Third, once a fracture is identified, do not forget to look at the rest of the film. About 10% of
2849
all cases have a second finding that often is as significant or even more so than the initial finding. Many fractures have associated dislocation, foreign bodies, or additional fractures, so be sure to examine the entire film. Finally, do not hesitate to obtain a CT scan or an MR study if the plain films fail to confirm what is believed to be present clinically. MR imaging is being used more frequently as a primary imaging tool for trauma, replacing CT or radionuclide studies in cases in which the plain films are negative or equivocal. Make sure that an expensive examination such as CT or MR is truly going to affect patient care rather than just show an abnormality and then result in the same treatment whether positive or negative. For example, there is no reason to perform a CT scan or an MR study to find a subtle or occult fracture of the radial head in the elbow, because the patient is going to have a posterior splint regardless of the results of the advanced study (assuming the patient had trauma to the elbow, has pain, and the plain film shows a displaced fat-pad, indicative of fluid in the joint). On the other hand, an elderly patient who has hip pain after a fall and has a negative plain film would benefit from an MR study because his treatment will vary based on whether an occult fracture is
present.
SPINE The cervical spine (C-spine) is one of the most commonly filmed parts of the body in a busy emergency department and can be one of the most difficult examinations to interpret. One of the most important pieces of information for the radiologist to have is the clinical history. If the patient has been involved in an automobile accident and has no neck pain, it is extremely unlikely that a fracture is present (1). So-called precautionary radiographs are not justified. On the other hand, if the plain films are negative in a trauma victim who has neck pain or neurologic deficits, a CT scan should be obtained. Usually, a cross-table lateral view of the C-spine is obtained first to avoid unduly moving the patient who might
2850
P.1103 have a cervical fracture. If the lateral C-spine appears normal, the remainder of the C-spine series, including flexion and extension views (if the patient can cooperate) is obtained.
FIGURE 43.1. Shoulders Obscuring C5–C6 Dislocation. This patient presented to the emergency department after an injury suffered while diving into a shallow swimming pool. He had neck pain but no neurologic deficits. A . The initial radiograph obtained of the cervical spine was interpreted as within normal limits. Only five cervical vertebrae are visible, however, because of high-riding shoulders. B . A repeat examination with the shoulders lowered reveals a dislocation of C5 on C6. To visualize C7, the shoulders were lowered even further. The C7 vertebral body must be visualized on every lateral cervical spine examination in a trauma setting.
2851
What does one look for on the lateral C-spine? First, make certain that all seven cervical vertebral bodies can be visualized. A large number of fractures are missed because the shoulders obscure the lower C-spine levels (Fig. 43.1). If the entire cervical spine is not visualized, repeat the film with the shoulders lowered. Next, evaluate five parallel (more or less) lines for step-offs or discontinuity as follows (Fig. 43.2): Line 1 is the prevertebral soft tissue and extends down the posterior aspect of the airway; it should be several millimeters from the first three or four vertebral bodies and then moves further away at the laryngeal cartilage. Line 1 should be less than one vertebral body width from the anterior vertebral bodies from C3 or C4 to C7, and it should be smooth in its contour. Line 2 follows the anterior vertebral bodies and should be smooth and uninterrupted. Anterior osteophytes can encroach on this line and extend beyond it and should therefore be ignored in drawing this line. Interruption of the anterior vertebral body line is a sign of a serious injury (Fig. 43.1B) . Line 3 is similar to the anterior vertebral body line (line 2) except that it connects the posterior vertebral bodies. Like line 2, it should be smooth and uninterrupted, and any disruption signifies a serious injury. Line 4 connects the posterior junction of the lamina with the spinous processes and is called the spinolaminal line. The spinal cord lies between lines 3 and 4; therefore, any offset of either of these lines could mean a bony structure is impinging on the cord. It takes very little force against the cord to cause severe neurologic deficits, and any bony structure lying on the cord must be recognized as soon as possible. Line 5 is not really a line so much as a collection of points—the tips of the spinous processes. They are quite variable in their size and appearance, although C7 is consistently the largest. A fracture of one of the spinous processes by itself is not a serious
2852
injury, but it occasionally heralds other, more serious injuries.
FIGURE 43.2. Normal Lateral Cervical Spine (C-spine). A. Lateral radiograph of a normal cervical spine. B . Diagrammatic representation of a lateral C-spine showing four parallel lines that should be observed in every lateral C-spine examination. Line 1 is the soft tissue line that is closely applied to the posterior border of the airway through the first four or five vertebral body segments; it then widens around the laryngeal cartilage and runs parallel to the remainder of the cervical vertebrae. Line 2 demarcates the anterior border of the cervical
2853
vertebral bodies. Line 3 is the posterior border of the cervical vertebral bodies. Line 4 is drawn by connecting the junction of the lamina at the spinous process, which is called the spinolaminal line. It represents the posterior extent of the central canal that contains the spinal cord itself. These lines should be generally smooth and parallel, with no abrupt stepoffs.
P.1104 After visually inspecting these five lines on the lateral C-spine film, then inspect the that the anterior (Fig. 43.3). Any whom up to 5.0
C1–C2 area a little more closely. Make certain arch of C1 is no more than 2.5 mm from the dens greater separation than this (except in children, for mm can be normal) is suspicious for disruption of the
transverse ligament between C1 and C2 (Fig. 43.4) .
FIGURE 43.3. Normal C1 and C2. A lateral radiograph (A) and drawing (B) of the upper cervical spine showing the normal distance of less than 2.5 mm from the anterior arch of C1 to the odontoid process (dens) of C2 (arrows) .
2854
The disk spaces are examined next to see that there is no inordinate widening or narrowing, either of which could indicate an acute traumatic injury. If a disk space is P.1105 P.1106 narrowed, it will usually be secondary to degenerative disease, but the clinician must make certain that associated osteophytosis and sclerosis are present before diagnosing degenerative disease.
FIGURE
43.4. C1–C2
Dislocation. A lateral radiograph (A)
and drawing (B) of the upper cervical spine in a patient who suffered trauma to the neck shows that the anterior arch of C1 is 9 mm anterior to the odontoid process of C2 (arrows). This is diagnostic of a dislocation of C1 on C2 and indicates rupture of the transverse ligaments that normally hold these vertebral segments together.
2855
FIGURE
43.5. Jefferson
Fracture.
A. An anteroposterior open-
mouth odontoid view is suspicious for the lateral masses of C1 being laterally displaced on the body of C2. Because of overlying structures, however, this is difficult to appreciate. B . A CT examination was obtained and shows multiple fracture sites in the C1 ring (arrows). This is called a Jefferson fracture. CT should be routinely used in spinal trauma because of frequent shortcomings of plain films.
The examination of the lateral C-spine as described here can be done in less than 1 minute. If it is normal, then the remainder of the examination can be completed, including flexion and extension views. It is imperative that the patient initiate the flexion and extension without help from the technician or anyone else. A patient, if conscious and alert, will not injure himself or herself with voluntary flexion and extension and will have muscle guarding preventing P.1107 motion if there is an injury present. Even gentle pressure to aid in flexion or extension can cause severe injury if a fracture or dislocation is present.
2856
FIGURE
43.6. Rotatory Fixation of the Atlantoaxial Joint.
This patient presented to the emergency department with pain and decreased motion in the cervical spine. A . An anteroposterior open-mouth odontoid view shows that the space on the left side of the odontoid between the odontoid and the lateral mass of C1 (arrows) is wider than the corresponding space on the right side. This is often the result of rotation. Therefore, open-mouth odontoid views with right and left obliquities were obtained. B . This view shows rotation of the patient's head to the left, which causes the space on the left side of the odontoid process (arrows) to be wider than that on the right, which is appropriate.
2857
C . This view, however, shows that when the patient turns the head to the right, the space on the right (arrows) does not get wider than the space on the left. This is diagnostic of rotatory fixation of the atlantoaxial joint.
A few examples of fractures, dislocations, and other abnormalities are illustrated in the following paragraphs.
Jefferson
Fracture
A blow to the top of the head, such as when an object falls directly on the apex of the skull, can cause the lateral masses of C1 to slide apart, splitting the bony ring of C1. This is called a Jefferson fracture (Fig. 43.5). It nicely illustrates how a bony ring will not break in just one place, but must break in several places. This is a rule that is seldom violated. All the vertebral rings, when fractured, must fracture in two or more places. The bony rings of the pelvis behave similarly. CT is excellent at demonstrating the complete bony ring of C1 and shows the fractures, as well as any associated soft tissue mass, much better than plain films do. For correct diagnosis of a Jefferson fracture on plain film, the lateral masses of C1 must extend beyond the margins of the C2 body (Fig. 43.5A). The presence of asymmetry of the spaces on either side of the dens is not enough to make the diagnosis, as this can be normally asymmetric with rotation or with rotatory fixation of the atlantoaxial joint. Rotatory fixation of the atlantoaxial joint is a somewhat controversial, little-understood process in which the atlantoaxial joint becomes fixed and the C1–C2 bodies move en mass instead of rotating on one another. It is easily diagnosed with open-mouth odontoid views. In the normal odontoid view, the spaces lateral to the dens (odontoid) are equal. With rotation of the head to the left, the space on the left widens, and with rotation to the right the space on the right widens. With rotatory fixation, one of the spaces is wider than the other and stays wider, even with rotation of the head to the
2858
opposite side (Fig. 43.6). This is a P.1108 P.1109 P.1110 relatively innocuous malady that by itself is usually treated with a soft cervical collar, gentle traction, or both. However, on rare occasions it is associated with disruption of the transverse ligaments at C1–C2 (diagnosed by an increase of more than 2.5 mm in the space between the anterior arch of C1 and the dens), and when it is, it is then a serious problem. It usually presents spontaneously or after very mild trauma such as an unusual sleeping position.
FIGURE
43.7. Clay
Shoveler's
Fracture. A nondisplaced
fracture of the C7 spinous process (arrow) is noted that is diagnostic of a clay shoveler's fracture.
2859
FIGURE 43.8. Hangman's Fracture. A. Lateral films of a patient with a hangman's fracture shows an obvious example of
2860
the posterior elements of the C2 vertebral body fractured and displaced inferiorly (arrow). B . This view shows a very subtle fracture through the posterior elements of C2 (arrow) in another patient. A line drawn through the spinolaminal lines of the posterior elements shows that the C2 spinolaminal line is offset posteriorly in this example.
FIGURE 43.9. Flexion Teardrop Fracture. This patient suffered a hyperflexion injury in an automobile accident and presented to the emergency department with severe neurologic deficits. A lateral radiograph of the lower cervical spine shows wedging anteriorly of the C7 vertebral body, with some displacement of the posterior vertebral line at C7 into the central canal. A small avulsion fracture off the anterior body is also noted.
2861
FIGURE 43.10. Unilateral Locked Facets. The C6-C7 disk space is abnormally widened, and the C7 vertebra is posteriorly located in relation to C6. Also note the C7 facets, which are dislocated and locked on the C6 facets (arrow). When the facets are perched in this manner, it is termed locked facets, which are unilateral in this example.
2862
FIGURE 43.11. Seatbelt Fracture. Hyperflexion at the waist can cause anterior wedging of the vertebral body in the lower thoracic or upper lumbar region as shown in (A). By itself, although painful, it is somewhat innocuous; however, (B) shows a horizontal fracture through the right transverse process and pedicle (arrow) caused by extreme traction during the flexion injury. When fracture of the posterior elements occurs, this injury is considered to be unstable and potentially debilitating. Any anterior wedging injury to a vertebral body should have the posterior elements of that level closely inspected.
2863
FIGURE
43.12. Spondylolysis.
A. An oblique plain film of the
lumbar spine shows a defect in the neck of the “Scottie dog― at L5 (arrow), which is diagnostic of a spondylolysis. B . A drawing of an oblique view of the lumbar spine shows how a spondylolysis appears as a “collar― around the Scottie dog's
neck.
2864
FIGURE
43.13. Spondylolisthesis.
A. A lateral plain film of the
lumbar spine shows that the L5 vertebral body is slightly anteriorly offset on the S1 body, as noted by the posterior margins (arrows). B . The drawing illustrates this more clearly. Because this offset is less than 25%, as measured by the length of the S1 endplate, it is termed a grade 1 spondylolisthesis. A grade 2 offset is more than 25% but less than 50% of the length of the S1 endplate.
“Clay-Shoveler's―
Fracture
Another relatively innocuous injury is a fracture of the C6 or C7 spinous process, which is called a “clay-shoveler's― fracture. Supposedly, workers shoveling sticky clay would toss shovels full of clay over their shoulders; once in a while, the clay would stick to the
2865
shovel, causing the ligaments attached to the spinous processes (supraspinous ligaments) to undergo tremendous force, pulling on the spinous process and avulsing it. This can occur at any of the lower cervical spinous processes (Fig. 43.7) .
“Hangman's―
Fracture
A “hangman's― fracture is an unstable, serious fracture of the upper C-spine that is caused by hyperextension and distraction (such as hitting one's head on a dashboard). This is a fracture of the posterior elements of C2 and, usually, displacement of the C2 body anterior to C3 (Fig. 43.8). These patients actually do better than one might think. They often escape neurologic impairment because the fractured posterior elements of C2, in effect, cause decompression and take pressure off the injured area.
Flexion-Teardrop
Fracture
Severe flexion of the C-spine can cause a disruption of the posterior ligaments with anterior compression of a vertebral body. This is called a flexion “teardrop― fracture (Fig. 43.9). A teardrop fracture is usually associated with spinal cord injury, often a result of displacement of the posterior portion of the vertebral body into the central canal.
Unilateral
Locked
Facets
Severe flexion associated with some rotation can result in rupture of the apophyseal joint ligaments and facet joint dislocation. This can result in locking of the facets in an overriding position that, in effect, causes some stabilization to protect against further injury. This is called unilateral locked facets (Fig. 43.10). It occasionally occurs bilaterally.
“Seatbelt
Injury―
“Seatbelt injury― is seen secondary to hyperflexion at the waist (as occurs in an automobile accident while restrained by a lap
2866
belt). This causes distraction of the posterior elements and ligaments and anterior compression of the vertebral body. It usually involves the T12, L1, or L2 level. Several variations of this injury can occur: a fracture of the posterior body is called a Smith fracture, and a fracture through the spinous process is called a Chance fracture. Horizontal fractures of the pedicles, laminae, and transverse processes can also occur (Fig. 43.11) .
Spondylolysis A somewhat controversial spinal abnormality that may or may not be caused by trauma is spondylolysis. Spondylolysis is a break or defect in the pars interarticularis portion of the lamina (Fig. 43.12). On oblique views, the posterior elements form the figure of a “Scottie dog,― with the transverse process being the nose, the pedicle forming the eye, the inferior articular facet being the front leg, the superior articular facet representing the ear, and the pars interarticularis (the portion of the lamina that lies between the facets) equivalent to the neck of the dog. If a spondylolysis is present, the pars interarticularis, or the neck of the dog, will have a defect or break. It often looks as if the dog has a collar around the neck. The cause of spondylolysis is controversial but is thought to be congenital and/or posttraumatic. Many believe this is a stress-related injury from infancy that develops when toddlers try to walk and repeatedly fall on their buttocks, sending stress to their lower lumbar spine. The significance of spondylolysis is just as controversial as its etiology. More and more clinicians are coming to the viewpoint that a spondylolysis is an incidental finding with no P.1111 clinical significance in most cases. It has been reported in up to 10% of the asymptomatic population. Certainly some patients have pain related to a spondylolysis and get relief after rest or immobilization, and some require surgical stabilization. It is important to identify spondylolysis preoperatively in patients undergoing lumbar discectomy, so that the possibility of clinical symptoms from the
2867
spondylolysis that can mimic disk symptoms can be evaluated. Although plain films can usually show spondylolysis, CT will show it to better advantage and will demonstrate any associated disk disease. MR will show spondylolysis, but it can be difficult to see and is easily overlooked with this method.
FIGURE 43.14. Anterior Wedge Compression Fracture. Anterior compression of this lower thoracic spine vertebral body (arrow) is present, which may or may not be acute. If the patient
2868
has pain in this area, it is most likely acute and must be protected with a back brace until the symptoms abate.
If spondylolysis is bilateral and the vertebral body in the more cephalad position slips forward on the more caudal body, spondylolisthesis is said to be present (Fig. 43.13). Spondylolisthesis may or may not be symptomatic and by itself has no clinical significance. If severe, it can cause neuroforaminal stenosis and can impinge on the nerve roots in the central spinal canal. If it is symptomatic, it can be stabilized surgically.
Other
Abnormalities
Anterior wedge compression fractures of the spine are commonly seen (Fig. 43.14), especially at the thoracolumbar junction, because of an old injury; they are passed off by the radiologist, if they are mentioned at all, as incidental findings. The problem with this assumption is that you cannot tell from a plain film if the fracture is old or new, even if degenerative changes are present (which are often not related to the fracture). If acute and left unprotected, a wedge compression fracture can proceed to delayed further collapse, with resulting severe neurologic deficits (Fig. 43.15). This is called Kümmel disease and typically occurs 1 to 2 weeks after the initial trauma. Multiple lawsuits have been filed against radiologists who failed to mention minor anterior wedging of a vertebral body that went on to further collapse, with associated paraplegia. All that needs to be mentioned is that a fracture is present that is of indeterminate age and requires clinical correlation. If the patient has pain in that location, a back brace must be worn until they are pain free. Old films can help determine whether it is an old fracture. If no pain is present on physical exam, it can be safely assumed to be an old fracture. It is not necessary to obtain a CT or MR even if pain is present, because the treatment will be the same regardless of what the CT or MR reveals. No spine surgeon will operate on a stable spine fracture without kyphosis or neurologic deficits, so the CT or MR adds nothing but time and expense.
2869
FIGURE
43.15. Kümmel Disease. A. Very minimal anterior
wedging of the L1 vertebral body is noted by comparing the height of the anterior body versus the posterior height. This patient had been in an automobile accident and complained of back pain. No treatment for his back was given. B . After several weeks of continuing pain, he presented with leg weakness, which proceeded to paraplegia. A spine film showed progression of the vertebral body collapse of L1. This almost certainly could have been avoided with simple bracing of the spine after the initial injury.
2870
FIGURE
43.16. Spine Fracture in Ankylosing Spondylitis. A.
A lateral plain film of the spine following trauma shows fusion anteriorly, which was secondary to ankylosing spondylitis. Minimal anterior wedging of the L1 vertebral body is present, which was overlooked. B . Two weeks later, a CT of the spine was performed because of the sudden onset of paralysis. This axial image through L1 shows a fracture of the posterior elements, which was undoubtedly present on the initial visit to the emergency room. Patients with ankylosing spondylitis need to be examined closely for any back pain following trauma and imaged with CT or MR if any pain is present.
P.1112 Patients who have fusion of their spine from ankylosing spondylitis, and, to a lesser extent, from diffuse idiopathic skeletal hyperostosis, are at a very high risk of spinal fractures from even relatively minor trauma. Patients with ankylosing spondylitis typically have marked osteoporosis, which further magnifies their risk of fracture. A fused spine is more likely to fracture than a normal spine in a manner similar to a long glass pipette breaking more easily than a short one
2871
because it has a long lever arm. A small force at one end is greatly magnified further down the lever arm. For that reason, a patient with ankylosing spondylitis should be treated as though a spinal fracture is present if they have back pain following trauma. CT and/or MR are mandatory if plain films are negative (Fig. 43.16) .
HAND AND WRIST Several seemingly innocuous fractures in the hand require surgical fixation rather than just casting and should therefore be recognized by the radiologist as serious injuries.
Bennett
Fracture
One such fracture is a fracture at the base of the thumb into the carpometacarpal joint—a Bennett fracture (Fig. 43.17). Because of the insertion of the strong thumb adductors at the base of the thumb, it is almost impossible to keep the metacarpal from sliding off its proper alignment. It almost always requires internal fixation. The radiologist occasionally has to remind a nonorthopaedic practitioner of this, as well as closely examine the alignment of a Bennett fracture in plaster that has not been internally fixed with wires. A comminuted fracture of the base of the thumb that extends into the joint has been termed a Rolando fracture (Fig. 43.18), and a fracture of the base of the thumb that does not involve the joint has been called a pseudo-Bennett
fracture.
Mallet finger or baseball finger is an avulsion injury at the base of the distal phalanx (Fig. 43.19) where the extensor digitorum tendon inserts. With the extensor tendon inoperative, the distal phalanx flexes without opposition, which can result in a flexion deformity and inability to extend the distal phalanx if not properly treated.
Volar
Plate
Fracture
A fracture at the volar aspect of the base of the interphalangeal and metacarpophalangeal joints from an avulsion of the volar plate can appear innocent but often requires surgical intervention. The volar
2872
plate is a dense fibrocartilaginous band that covers the joint on the volar aspect and can get interposed in the joint once it is torn, often requiring surgical removal.
FIGURE 43.17. Bennett Fracture. A small corner fracture of the base of the thumb is noted that involves the articular surface of the base of the thumb (arrow); this is a serious injury that almost always requires internal fixation.
P.1113
“Gamekeeper's
Thumb―
Another innocent-appearing fracture that often requires fixation is an avulsion on the ulnar aspect of the first
2873
internal
metacarpophalangeal joint (Fig. 43.20); this is where the ulnar collateral ligament of the thumb inserts. If the ulnar collateral ligament is torn, normal function of the thumb can be impaired, and this can have a serious result if not properly treated. This injury is called a “gamekeeper's thumb― because of the propensity of English game wardens to acquire it from breaking rabbits’ necks between their thumb and forefinger. A more current scenario is falling on a ski pole and having the pole jam into the webbing between the thumb and index finger. This avulsion injury usually requires pinning to fix the ligament securely.
Lunate/Perilunate
Dislocation
A fall on the outstretched arm can result in any number of wrist fractures and dislocations. One serious such injury is the lunate/perilunate dislocation. This occurs when the ligaments between the capitate and the lunate are disrupted, allowing the capitate to dislocate from the cup-shaped articulation of the lunate. This is best seen on lateral views. Ordinarily, on the lateral view the capitate should be seen seated in the cup-shaped lunate (Figs. 43.21, 43.22A). In a dorsal dislocation (the capitate occasionally dislocates volarly, but this is uncommon), the capitate and all of its surrounding bones, including the metacarpals, come to lie dorsal to a line drawn through the radius and the lunate (Figs. 43.22B, 43.23) . If the capitate then pushes the lunate volarly and tips it over, a line drawn up through the radius shows that the lunate is volarly displaced, and the line goes through the capitate. This has been termed a lunate dislocation (Figs. 43.22C, 43.24). Failure to diagnose and treat this disorder can result in permanent median nerve impairment, as the nerve can get impinged by the volarly displaced lunate.
2874
FIGURE 43.18. Rolando Fracture. A comminuted fracture of the base of the thumb that extends into the articular surface is a more serious type of Bennett fracture, which has been termed a Rolando fracture.
A lunate or perilunate dislocation can be diagnosed on an AP view of the wrist by noting a triangular or pie-shaped lunate (Fig. 43.24B) . Ordinarily, the lunate has a rhomboid P.1114 shape on the AP view, with the upper and lower borders parallel.
2875
FIGURE 43.19. Mallet Finger. A small avulsion injury is noted at the base of the distal phalanx, which is where the extensor digitorum tendon inserts. This is termed a mallet finger or baseball finger because it is often caused by a baseball striking the distal phalanx and causing the avulsion.
2876
FIGURE
43.20. Gamekeeper's
Thumb. A small avulsion injury
on the ulnar aspect of the first metacarpophalangeal joint (arrow) is diagnostic of a gamekeeper's thumb. This is the insertion site for the ulnar collateral ligament and usually requires internal fixation.
Several fractures are known to be associated with a perilunate dislocation, the most common of which is a transscaphoid fracture. The capitate, radial styloid, and triquetrum are also known to fracture frequently when a perilunate dislocation occurs.
2877
Hook of the Hamate Fracture One of the most difficult wrist fractures to identify radiologically is a fracture of the hook of the hamate. A special view, the carpal tunnel view, should be obtained when trying to see the hook of the hamate. This view is obtained with the wrist (palm down) flat on an x-ray plate and the fingers pulled dorsally. The x-ray beam is angled about 45°—parallel to the palm of the hand so that the carpal tunnel is in profile. The hook of the hamate is seen as a bony protuberance off the hamate on the ulnar aspect of the carpal tunnel. A fractured hook of the hamate is often identified with the carpal tunnel view (Fig. 43.25) but occasionally can be very difficult to visualize. A CT scan will often show an obvious fracture that the plain film does not (Fig. 43.26) and should be considered P.1115 in any possible carpal fracture when plain films are not diagnostic.
2878
FIGURE
43.21. Normal Lateral Radiograph of the Wrist. The
normal lateral view should show the lunate seated in the distal radius and the capitate seated in the lunate. A line drawn up through the radius should connect all three structures. Compare this radiograph with the drawing in Fig. 43.22A.
2879
FIGURE 43.22. Perilunate and Lunate Dislocations. Schematic depiction of normal lateral wrist (A), perilunate dislocation (B), and lunate dislocation (C). (Dorsal is to the right.)
2880
FIGURE
43.23. Perilunate
Dislocation. Although the lunate (L)
is in a normal relationship to the distal radius, the capitate (C) and the remainder of the wrist are dorsally displaced in relation to the lunate. Compare this radiograph with the drawing in Fig. 43.22B.
A fracture of the hook of the hamate most commonly occurs from a fall on the outstretched hand. A clinical setting that has gained attention in sports medicine circles is that of a professional athlete who participates in an activity in which the butt of a club, bat, or racket is held in the palm of the hand. Overswinging can result in the butt of the club levering off the hook of the hamate. This has been seen in professional baseball players, tennis players, and golfers. It is not seen as often in amateurs because they usually are not strong
2881
enough to exert enough force to lever the hook off and, if they do, will usually terminate that activity, allowing healing, whereas a professional will continue participation, which can lead to a nonunion of the fracture.
FIGURE
43.24. Lunate
Dislocation.
A. The lateral radiograph
of the wrist shows the lunate (L) tipped off of the distal radius, whereas the capitate (C) seems to be normally aligned in relation to the radius yet is dislocated from the lunate. Compare this with the drawing in Fig. 43.22C. B . Anteroposterior view shows a pieshaped lunate (L) rather than a lunate with a more rhomboid shape. A pie-shaped lunate on an anteroposterior view is diagnostic of a perilunate or lunate dislocation.
2882
FIGURE
43.25. Fracture of the Hook of the Hamate. The
hook of the hamate is seen on a carpal tunnel view in this patient and has an area of sclerosis with a faint cortical break (arrow) . This represents a fracture at the base of the hook of the hamate.
Rotary subluxation of the navicular is another wrist injury seen after a fall onto the outstretched hand. This results from rupture of the scapholunate ligament, which allows the scaphoid (navicular) to rotate dorsally. On an P.1116 AP wrist plain film, a space is seen between the navicular and the lunate (Fig. 43.27), where ordinarily they are closely apposed. This has been called the “Terry Thomas― sign after a famous British actor (circa 1950s) who had a gap between his two front teeth.
2883
FIGURE
43.26. Fractured
Hamate. A CT scan through the wrist
in this patient shows a faint lucency surrounded by sclerosis in the left hamate (arrow), which represents a fracture through the base of the hook of the hamate with moderate reactive sclerosis. This could not be seen in the plain films, even in retrospect.
2884
FIGURE 43.27. Rotatory Subluxation of the Navicular. An anteroposterior view of the wrist shows a gap or space between the navicular and the lunate (arrow). This is abnormal and represents the “Terry Thomas― sign, which means that the scapholunate ligament is ruptured. This is diagnostic of a rotatory subluxation of the navicular.
2885
FIGURE
43.28. Scaphoid
Fracture. A coronal T1WI of the wrist
in a patient with snuffbox tenderness and a normal plain film shows a fracture of the mid-waist of the scaphoid (arrow) .
2886
FIGURE
43.29. Avascular Necrosis of the Navicular. An
anteroposterior view of the wrist shows a fracture through the waist of the navicular (arrow). The proximal half of the navicular is slightly sclerotic in relation to the remainder of the carpal bones, which indicates avascular necrosis of the proximal half.
P.1117
Navicular
Fracture
A fracture of the navicular is a potentially serious injury because of the high rate of avascular necrosis (AVN) that occurs with this injury. When AVN occurs, it usually requires surgical intervention with a metallic screw and bone grafting to obtain healing. This fracture can be very difficult to detect initially; therefore, whenever a fracture of the navicular is clinically suspect (trauma with pain over the snuffbox of the wrist), the wrist should be casted and repeat radiographs obtained in 1 week. Often, the fracture is then visualized because of
2887
the disuse osteoporosis and hyperemia around the fracture site. Thus, in the acute setting, a negative film does not exclude a fractured scaphoid. Instead of casting the wrist and repeating the films in a week, many patients now get immediate MR to determine whether a fracture is present (Fig. 43.28). This has been shown to be less expensive overall than having the patient casted and reexamined in a week (2) . If AVN of the navicular develops, it is the proximal fragment that undergoes necrosis because the blood supply to the navicular begins distally and runs proximally. A fracture with disruption of the blood supply thus leaves the proximal pole without a vascular supply; hence, it dies. AVN is diagnosed by noting increased density of the proximal pole of the navicular compared with the remainder of the carpal bones (Fig. 43.29) .
2888
FIGURE
43.30. Kienböck
Malacia. An anteroposterior view of
the wrist reveals the lunate to be sclerotic and abnormal in shape. The lunate has collapsed because of aseptic necrosis. This is known as Kienböck malacia. Note that the ulna is shorter than the radius; this is termed negative ulnar variance, which is often associated with Kienböck malacia.
AVN can occur in other carpal bones, most commonly the lunate. This is called Kienböck malacia and is most often caused by trauma; however, it is also thought to be idiopathic. It is diagnosed by noting the increased density in the lunate, which may or may not go on to collapse and fragmentation (Fig. 43.30). It often requires surgical bone grafting and, occasionally, removal or proximal carpal row
2889
fusion. It has a high association with a discrepancy between the length of the radius and the ulna as seen at the radiocarpal joint. If the ulna is shorter than the radius, it is termed negative ulnar variance and there is an increased incidence of Kienböck malacia (Fig. 43.30). If the ulna is longer than the radius, it is termed positive ulnar variance and there is an increased incidence of triangular
fibrocartilage
tears.
A common avulsion fracture in the wrist is a triquetral fracture. It is best seen on a lateral film, which shows a small chip of bone off the dorsum of the wrist (Fig. 43.31). This is virtually pathognomonic of an avulsion from the triquetrum.
FIGURE 43.31. Triquetral Fracture and Perilunate Dislocation. A perilunate or lunate dislocation is present (it is
2890
difficult to classify exactly which has occurred, because both the lunate and the capitate are out of their normal position). A small avulsion is seen on the dorsum of the wrist (arrow), which is virtually diagnostic of an avulsion off the triquetrum. It is often associated with a lunate or perilunate dislocation.
FIGURE 43.32. Colles Fracture. A fracture of the distal radius with dorsal angulation is noted, which has been termed a Colles fracture.
2891
FIGURE
43.33. Smith
Fracture. A fracture of the distal radius
with volar angulation such as this is called a Smith fracture. This is a much less common injury than the Colles fracture, shown in Fig. 43.32.
P.1118
ARM Colles
Fracture
One of the most common fractures of the forearm is a fracture of the distal radius and ulna after a fall on an outstretched arm. This results in a dorsal angulation of the distal forearm and wrist and is called a
2892
Colles fracture (Fig. 43.32). When the fracture angulates volarly, it is called a Smith fracture (Fig. 43.33). A Smith fracture is much less common than a Colles fracture. Sometimes the radius and ulna suffer a traumatic insult, and the force on the bones causes bending instead of a frank fracture. This has been termed a plastic bowing deformity of the forearm (Fig. 43.34) and is often treated by breaking the bones while the patient has undergone anesthesia and resetting them. Left untreated, a plastic bowing deformity can result in reduced supination and pronation.
Monteggia
Fracture
The forearm is a two-bone system that has some of the same properties as a ring bone. As mentioned previously, a solid ring cannot break in only a single place; it must break in at least two places. In the forearm, a fracture of one bone should be accompanied by a fracture of the other. If the second fracture is not present, P.1119 a dislocation of the nonfractured bone usually occurs. The most common example of this is a fracture of the ulna with a dislocation of the proximal radius (Fig. 43.35). This is called a Monteggia fracture. The dislocated radial head can be missed clinically and develop into AVN, with subsequent elbow dysfunction. Whenever the forearm is fractured, the elbow must be examined to exclude a dislocation.
2893
FIGURE 43.34. Plastic Bowing Deformity of the Forearm. These anteroposterior and lateral views of the forearm of a child show the radius to be abnormally bowed anteriorly. This has been termed a plastic bowing deformity of the forearm and occurs only in children.
2894
FIGURE 43.35. Monteggia Fracture. A blow to the forearm such as with a police officer's nightstick can result in a fracture of the ulna. Although the head of the radius appears normally placed in an anteroposterior view (A), the lateral examination (B) reveals that the head of the radius is displaced. Failure to recognize this abnormality can result in death of the radial head, with subsequent elbow dysfunction. This illustrates the importance of always obtaining two views of a bone after trauma.
Galeazzi
Fracture
A fracture of the radius with dislocation of the distal ulna is called a Galeazzi fracture (Fig. 43.36). This is less common than a Monteggia
2895
fracture. A helpful indicator of a fracture about the elbow is a displaced posterior fat-pad. Ordinarily, the posterior fat-pad is not visible on a lateral view of the elbow because it is tucked away in the olecranon fossa of the distal humerus. When the joint becomes distended with blood secondary to a fracture, the posterior fat-pad is displaced out of the olecranon fossa and is visible on the lateral view (Fig. 43.37A). Therefore, in the setting of trauma, a visible posterior fatpad indicates a fracture. In an adult (epiphyses closed), the fracture site is almost always the radial head (Fig. 43.37B). In a child (epiphyses open), it is usually indicative of a supracondylar fracture (Fig. 43.38) .
FIGURE 43.36. Galeazzi Fracture. A. A fracture of the distal radius in this patient is seen on the anteroposterior view without a definite fracture of the ulna. B . This view shows an obvious dislocation of the distal ulna, which would almost certainly not be missed clinically. This has been termed a Galeazzi fracture and is much less common than the Monteggia fractures.
2896
Often, the fracture itself is not visualized, and extraordinary steps are taken by clinicians and radiologists alike to demonstrate the fracture. These steps include oblique views, special radial head views, tomograms, and even CT or MR. These are absurd attempts to document pathology that will be treated identically whether or not it is radiographically recorded. As long as there is no obvious deformity or loose body, it does not matter whether the fracture is definitely identified or not in a patient with a posttraumatic painful elbow and a visible posterior fat-pad. An infection, an arthritide, or any elbow effusion could cause a joint effusion and a displaced posterior fatpad, but the clinical setting would not be to rule out a fracture. The anterior fat-pad also gets displaced with a joint effusion. Ordinarily it is visible as a small triangle just anterior to the distal humeral diaphysis on a lateral film (Fig. 43.39). With an effusion, it gets displaced superiorly and outward from the humerus and has been called a “sail sign― because it resembles a spinnaker sail (see Figs. 43.37, 43.38) . Shoulder
dislocations are generally easily diagnosed, both
clinically and radiographically. The most common shoulder dislocation is the anterior dislocation. It is at least 10 times more common than a posterior dislocation. For all practical purposes, anterior and posterior dislocations are the only two types of shoulder dislocations about which to be concerned.
2897
FIGURE
43.37. Displaced Elbow Fat-Pads. A. On the lateral
view of this elbow, the posterior fat-pad is faintly visible (arrow) and the anterior fat-pad is elevated and anteriorly displaced (curved arrow). These findings indicate a fracture about the elbow that in an adult should be in the radial head. B . An oblique view shows the fracture of the radial head (arrow). Even without seeing the fracture on the radiographs, it should be surmised to be present when the posterior fat-pad is visualized in the setting of trauma. The elevated and displaced anterior fat-pad has been termed a sail sign.
P.1120 An anterior dislocation occurs when the arm is forcibly externally rotated and abducted. This is commonly seen when football players
2898
“arm tackle,― when kayakers “brace― with the paddle above their heads and allow their arms to get too far posterior, when skiers plant their uphill pole and get it stuck, and from other similar athletic positions. Radiographically, the diagnosis is easily made on an AP shoulder film: the humeral head is seen to lie inferiorly and medial to the glenoid (Fig. 43.40). The humeral head often impacts on the inferior lip of the glenoid, causing an indentation on the posterosuperior portion of the humeral head; this is called a HillSachs deformity. The presence of a Hill-Sachs deformity is said to indicate a greater likelihood of recurrent dislocation, and some surgeons use it as an indicator to intervene surgically to prevent a recurrence. A bony irregularity or fragment off the inferior glenoid, which occurs from the same mechanism as the Hill-Sachs deformity, is called a Bankart
deformity. It is not seen radiographically as often
as the Hill-Sachs deformity. A posterior
dislocation can be a difficult diagnosis to make, both
clinically and radiographically. An AP view may look completely normal, or nearly so. On the AP view of a normal shoulder, the humeral head should slightly overlap the glenoid (Fig. 43.41) , forming what has been called the “crescent sign.― In a patient with a posterior dislocation, this crescent of bony overlap is usually absent and a small space is seen between the glenoid and the humeral head (Fig. 43.42) . The best way to unequivocally diagnose a dislocated shoulder is to obtain a transscapular view. An axillary view will show basically the same thing but requires the patient to move the arm and shoulder, which can be painful and may even re-dislocate the shoulder if it has spontaneously reduced itself. The transscapular view is obtained by angling the x-ray beam across the shoulder in the same plane as the blade of the scapula. This gives an en face view of the glenoid, and the humeral head can easily be related to it as either normal, anterior (Fig. 43.43), or posterior. P.1121 P.1122 P.1123 Because of frequently overlapping ribs and clavicles, the exact
2899
anatomy is often difficult to discern on the transscapular view. To find the glenoid, one has to find the coracoid, the spine of the acromion, and the blade of the scapula. These three structures all lead to the glenoid and form a “Y― around it. All that is necessary to find the center of the glenoid is to find two of those bony landmarks, usually the coracoid and the blade of the scapula. The humeral head can then be found and its position determined.
FIGURE 43.38. Displaced Elbow Fat-Pads. A lateral view of the elbow in this child shows a posterior fat-pad (arrow) and a sail sign anteriorly (curved arrow). This is indicative of a fracture about the elbow, which in a child (epiphyses are open) usually means a supracondylar fracture.
2900
FIGURE
43.39. Normal Anterior Fat-Pad of the Elbow. Note
the lucency just anterior to the humerus of this normal elbow and compare this with the sail sign of the anterior fat-pads in Figs. 43.37 and 43.38.
2901
FIGURE
43.40. Anterior
Shoulder
Dislocation. An
anteroposterior view of the right shoulder shows the humeral head to lie medial to the glenoid and inferior to the coracoid process (C). This is diagnostic of an anterior dislocation of the shoulder.
2902
FIGURE
43.41. Normal Anteroposterior View of the
Shoulder. Note in this example of a normal shoulder that the humeral head slightly overlaps the glenoid, which has been termed the crescent sign.
2903
FIGURE 43.42. Posterior Shoulder Dislocation. Note that the humeral head in this patient is slightly displaced from the glenoid on the anteroposterior view. This is termed absence of the crescent sign and is often seen with a posterior dislocation. Compare this with the normal shoulder in Fig. 43.41.
2904
FIGURE 43.43. Transscapular View of an Anterior Dislocation. This transscapular view of the shoulder is obtained by aiming the x-ray beam parallel to the shoulder blade. The coracoid process (C) can be seen anteriorly, and the spine of the acromion (A) can be seen posteriorly. Both of these structures extend inwardly and meet at the glenoid (G). The humeral head is seen in this example to lie anterior to the glenoid.
2905
FIGURE
43.44. Pseudodislocation of the Shoulder. A. An
anteroposterior view of the shoulder in this patient who had trauma to the shoulder shows the humeral head to be inferiorly placed in relation to the glenoid with absence of the normal crescent sign. A dislocation was suspected. B . The transscapular lateral film, however, reveals that the humeral head is normally placed over the glenoid. This is a pseudodislocation owing to a hemarthrosis. A search for an occult fracture should be made. In this case, a fracture can be seen in (A) (arrow), which caused bleeding into the joint.
2906
FIGURE 43.45. Fracture of the Glenoid. A. An anteroposterior view of the shoulder demonstrates a faint lucency, indicative of a fracture of the glenoid (arrows), with a fragment of bone seen inferior to the joint. B . The full extent of the fracture cannot be appreciated until the CT is examined. On the CT scan, the fracture can be seen to extend fully through the scapula and is seen to be slightly displaced in the articular portion.
An entity that can be mistaken for a dislocated shoulder is a traumatic hemarthrosis, which displaces the humeral head inferolaterally on the AP film (Fig. 43.44). Because the anterior dislocation displaces inferomedially, it should not be confused with this. The posterior dislocation will easily be excluded by looking at a transscapular view. This has been termed a pseudodislocation. It should be recognized so that attempts to “reduce― the “dislocation― are not made. Also, it can suggest a subtle or occult humeral head fracture.
2907
FIGURE
43.46. Dislocation of the Hip. A. An anteroposterior
plain film of the left hip shows dislocation of the femoral head, which lies slightly superior to the acetabulum. B . Fractures are easily identified on CT. A cortical break through the articular surface of the posterior acetabulum, as well as the dislocation, is identified.
If a fracture is suspected about the shoulder and the plain films are negative or equivocal, a CT scan should be performed. A complex joint such as the shoulder or hip is best examined with CT scanning when the full extent of the fracture needs to be identified (Fig. 43.45) .
PELVIS Fractures of the pelvis, especially those involving the acetabulum, can be difficult to evaluate completely with plain films alone. CT scanning should be considered in almost all acetabular fractures because of the possibility of free fragments and subtle fractures that do not show on plain films (Fig. 43.46) .
2908
FIGURE
43.47. Fracture of the Sacrum. An anteroposterior
view of the sacrum in this patient shows normal arcuate lines on the left side of the sacrum that are interrupted on the right side (arrows). Interruption of these lines indicates a fracture through this portion of the sacrum.
2909
FIGURE
43.48. Sacral Stress Fracture. A. Faint sclerosis is
noted in the left part of the sacrum as compared with the right in this patient complaining of pelvic pain. A radionuclide bone scan showed increased isotope uptake on the left half of the sacrum, and metastatic disease was postulated. B . A CT scan through this region demonstrates a cortical disruption (arrow), indicative of a fracture. These are characteristic plain film and CT appearances of a stress fracture of the sacrum.
P.1124 Sacral
fractures are said to occur in half of the cases of pelvic
fractures. They can be difficult to see on even the best of films because the sacrum is often hidden by bowel gas. In looking for sacral fractures, one should examine the arcuate lines of the sacrum bilaterally to see whether they are intact. Fractures often interrupt these lines and, because of the side-to-side asymmetry, can therefore be easily identified (Fig. 43.47) . Sacral stress fractures in patients who are osteoporotic or who have undergone radiation therapy can present as patchy or linear sclerosis on the sacral alae that may or may not show cortical disruption on plain films (Fig. 43.48A). These should be differentiated from metastatic disease because of their characteristic location,
2910
appearance, and history of prior radiation and by visualization of a cortical break. CT will usually, but not always, demonstrate cortical disruption (Fig. 43.48B). These fractures have a characteristic appearance on radionuclide bone scans (Fig. 43.49A), which is termed the “Honda sign― because of its appearance to the logo of the car. The Honda sign is seen only with bilateral stress fractures; unilateral fractures will have increased radionuclide uptake throughout one sacral ala. MR will demonstrate an area of diffuse low signal on T1WIs corresponding to the area of involvement (Fig. 43.49B). Sacral stress fractures have also been termed insufficiency fractures, indicating that the underlying bone is abnormal, similar to a pathologic fracture.
FIGURE
43.49. Sacral Stress Fracture. A. A radionuclide bone
scan in an osteoporotic patient with pelvic pain shows a classic “Honda sign― seen with bilateral sacral stress fractures. B . A T1W coronal MR in this patient shows diffuse low signal throughout the sacrum adjacent to the sacroiliac joints bilaterally. This represents edema and hemorrhage in the fractures and corresponds to the bone scan Honda sign.
2911
FIGURE 43.50. Avulsion Off the Ischium. An anteroposterior view of the pelvis shows an area of cortical disruption and periostitis at the right ischium (arrow) in a patient complaining of pain at this site. These findings are characteristic for an ischial avulsion and should not undergo biopsy.
2912
FIGURE
43.51. Rectus
Femoris
Avulsion. An anteroposterior
plain film of the left hip shows a faint calcific density superior to the acetabulum (arrow), which is characteristic for an avulsion of the rectus femoris muscle from the anterior inferior iliac spine.
2913
FIGURE
43.52. Osteoarthritis of the Symphysis Pubis.
Sclerosis with erosion is noted at the symphysis in this ultramarathoner complaining of severe pubic pain. This is characteristic of degenerative joint disease (DJD) or osteoarthritis at this site in such an overuse setting. Erosions are ordinarily not seen in DJD, except in certain joints such as the symphysis pubis, sacroiliac, and the acromioclavicular.
P.1125 P.1126 Avulsion
injuries affect the pelvis quite often and should be easily
recognized by radiologists. On occasion, an avulsion injury can have an aggressive appearance and, if not diagnosed radiographically, a biopsy might be performed. This can be calamitous, as avulsion injuries have been known to mimic malignant lesions histologically, with a misdiagnosis leading to radical treatment (Fig. 43.50) . Therefore, when an avulsion injury is a consideration, it becomes a “do not touch― lesion (see Chapter 46). Common sites for
2914
pelvic avulsions include the ischium, the superior and inferior anterior iliac spines (Fig. 43.51), and the iliac crest. These injuries are said to be fairly common in long jumpers, sprinters, hurdlers, gymnasts,
FIGURE
and
cheerleaders.
43.53. Osteoarthritis of the Sacroiliac Joint.
Sclerosis and erosions (arrow) are seen in the left sacroiliac joint in this young professional dancer. Although this has the appearance of an inflammatory arthritis, this is also seen in degenerative joint disease or osteoarthritis secondary to overuse.
2915
FIGURE
43.54. Sacroiliac
Osteophytes.
A. An anteroposterior
view of the pelvis in this marathoner shows dense sclerosis over both sacroiliac joints. B . A CT through this area demonstrates dense, bridging osteophytes, characteristic of degenerative joint disease.
Another area in the pelvis that can demonstrate radiologic findings as a result of stress is the symphysis pubis. In ultramarathoners, crosscountry skiers, soccer players, and other athletes, the symphysis can be affected by degenerative joint disease (DJD) or osteoarthritis (Fig. 43.52). The hallmarks of DJD are sclerosis, joint space narrowing, and osteophytosis (see Chapter 44). In certain joints, however, erosions can occur as a result of DJD. These joints include the temporomandibular joint, the acromioclavicular joint, the symphysis pubis, and the sacroiliac joint. When the sacroiliac joints are involved with DJD, this can closely resemble a human leukocyte antigen–B27 spondyloarthropathy (Fig. 43.53) and lead to erroneous diagnosis and treatment. Large osteophytes can develop across the sacroiliac joints and mimic sclerosis or even a tumor (Fig. 43.54) . P.1127
2916
LEG Overt fractures in the femur and lower leg are, for the most part, straightforward and deserve no special radiologic treatment for fear of missing subtle abnormalities. Stress
fractures, however, need to be considered in anyone with
hip or leg pain, because overlooking the diagnosis can lead to a complete fracture. The most serious stress fracture, and fortunately, one of the rarest, is the femoral neck stress fracture (Fig. 43.55) . Many of these progress to complete fractures (Fig. 43.56) that, with continued weight bearing, can displace; these are very serious lesions. Stress fractures also occur in the distal diaphysis of the femur and in the proximal, middle, and distal thirds of the tibia. All of these stress fractures need to be treated with the utmost caution, because complete fractures are not uncommon with continued stress (Fig. 43.57). Sclerosis in a weight-bearing bone that has a horizontal or oblique linear pattern should be considered a stress fracture until proved otherwise. A history of repetitive stress is not always associated, so the diagnosis should not depend solely on the history. A stress fracture occasionally will appear somewhat aggressive, with aggressive periostitis and no definite linearity to the sclerosis (Fig. 43.58A). If this is mistaken for a tumor and undergoes biopsy, it can be confused with a malignancy, with subsequent radical therapy. Therefore, these should not undergo biopsy under any circumstance. If the clinical presentation is unusual for a stress fracture and the plain films are not diagnostic, additional films should be obtained 1 or 2 weeks later. CT and MR sometimes will better delineate the lesion (Fig. 43.58B). Stress fractures can be difficult to diagnose radiologically early on but should be straightforward after several weeks.
2917
FIGURE sclerosis
43.55. Femoral Stress Fracture. An area of linear (arrows) is seen at the base of the femoral neck in a
runner with hip pain. This is diagnostic of a stress fracture of the femur.
2918
FIGURE 43.56. Stress Fracture of the Femoral Neck. A linear lucency with surrounding sclerosis is seen in the femoral neck in this jogger with hip pain. This is a severe femoral neck stress fracture.
One final stress fracture that deserves mention because it is frequently misdiagnosed clinically and overlooked radiographically is the calcaneal stress fracture (Fig. 43.59). It is often clinically misdiagnosed as a “heel spur― or plantar fasciitis and can be a somewhat subtle radiographic finding.
Hip
Fracture
Overt fractures in the lower extremity are uncommonly missed on radiographs; however, a few exceptions should be noted. Hip fractures in the elderly population can be very difficult to detect (Fig.
2919
43.60), and a high index of suspicion should be maintained. A negative plain film in an elderly patient with hip pain after trauma (even relatively mild trauma) does not exclude a femoral neck fracture. MR has been shown to be very useful in demonstrating femoral neck fractures that are occult (Fig. 43.61) .
Tibial
Plateau
Fracture
Another fracture that can be difficult to exclude on routine plain films is a tibial plateau P.1128 P.1129 fracture. A cross-table lateral plain film should be obtained in cases of knee trauma to look for a fat–fluid level (Fig. 43.62); this indicates a fracture that allows fatty marrow to leak into the knee joint. In the appropriate clinical setting, MR or CT may be necessary to make the diagnosis.
2920
FIGURE 43.57. Stress Fracture of the Proximal Tibia. A. A faint linear sclerotic area (arrow) is seen, which is characteristic for a stress fracture of the proximal tibia. B . This view shows the result of continued exercise in this patient: a complete fracture of the tibia and of the proximal fibula.
2921
FIGURE
43.58. Stress Fracture of the Tibia. A. An irregular
focus of sclerosis is seen in the posterior proximal tibia with adjacent periostitis. There was concern that this might represent a primary bone tumor, and the surgeons recommended a biopsy. B . An MR scan was performed, however, which shows a linear low signal area running obliquely across the tibia on this T1W coronal image, which is characteristic for a stress fracture. No significant soft tissue mass was found. The patient's recent history included an increase in his jogging. A stress fracture was diagnosed on the basis of these images.
2922
FIGURE 43.59. Calcaneal Stress Fracture. A linear band of sclerosis is seen in the posterior calcaneus (arrows), which is diagnostic for a stress fracture of the calcaneus.
2923
FIGURE
43.60. Fracture of the Hip. A. An anteroposterior view
of the hip of an elderly man was obtained following a fall. It was interpreted as normal, and the patient was dismissed from the emergency department. Two weeks later, the patient returned to the emergency department unable to walk, and another radiograph (B) was obtained. It shows a complete fracture through the femoral neck. In retrospect, the fracture can be faintly seen in (A) and should have been picked up initially. Fractures of the hip in the elderly can be very difficult to see and should be diligently searched for with additional views when the clinical setting is appropriate.
Lisfranc
Fracture
A serious fracture in the foot that can be missed radiographically when little or no displacement occurs is the so-called Lisfranc fracture (Fig. 43.63). It is named after a surgeon in Napoleon's army who would perform forefoot amputations in patients with gangrenous toes caused by frostbite. The Lisfranc fracture is a fracture-
2924
dislocation of the tarsometatarsals. If the dislocation is slight, it can be easily overlooked. A key to normal alignment is that the medial border of the second metatarsal should always line up with the medial border of the second cuneiform. If it does not, a Lisfranc fracture-dislocation should be suspected. This fracture is seen most commonly in patients who catch the forefoot in something such as a hole in the ground, or a horseback rider who falls and hangs by the forefoot in the stirrups. It is commonly seen as a neurotrophic or Charcot joint in diabetics.
FIGURE 43.61. Occult Fracture of the Hip. A. An anteroposterior plain film in an elderly patient with hip pain after a fall appears normal. B . A coronal T1WI was obtained because of the clinical suspicion of a fracture and shows linear low signal in the intertrochanteric region (arrow), confirming the fracture.
2925
FIGURE
43.62. Tibial Plateau Fracture. A. A cross-table
lateral plain film of the knee reveals a fat–fluid level (arrows) , which indicates a fracture with fatty marrow leaking into the joint. B . An anteroposterior view shows a barely discernible fracture (arrow) near the tibial spines, indicative of a tibial plateau
fracture.
2926
FIGURE 43.63. Lisfranc Fracture. An anteroposterior view of the foot in this patient shows a space between the first and second metatarsals with the base of the second metatarsal displaced off the second cuneiform. This is indicative of a Lisfranc fracture dislocation.
2927
FIGURE 43.64. Böhler Angle in a Normal Calcaneus. This drawing depicts the normal calcaneus with a line across the anterior process extending to the apex of the calcaneus intersecting with a line from the posterior portion of the calcaneus to the apex. This is termed the Böhler angle, and when it becomes flattened or smaller than 20°, a calcaneal fracture should be diagnosed.
2928
FIGURE
43.65. Calcaneal
Fracture. The Böhler angle in this
calcaneus is smaller than 20°, which is indicative of a fracture of the calcaneus.
P.1130 P.1131 Fracture of the calcaneus can be difficult to appreciate on routine radiographs. The Böhler angle is a normal anatomic landmark that should be looked for in every foot film when trauma has occurred (Fig. 43.64). An angle narrower than 20° indicates a compression of the calcaneus, as seen in jumping injuries (Fig. 43.65) . This is a fairly simplified overview of some commonly overlooked fractures and dislocations and should not be interpreted as a substitute for the more complete texts listed in the references (3, 4, 5) .
REFERENCES 1. Mirvis S, Diaconis J, Chirico P, Reiner B, Joslyn J, Militello P. Protocol-driven radiologic evaluation of suspected cervical injury: efficacy study. Radiology 1989;170:831–834.
spine
2. Dorsay TA, Major NM, Helms CA. Cost-effectiveness of immediate MR imaging versus traditional follow-up for revealing radiographically occult scaphoid fractures.― AJR Am J Roentgenol 2001;177(6):1257–1263. 3. Rogers LF. Radiology of Skeletal Trauma. 3rd ed. New York: Churchill
Livingstone,
2002.
4. Rockwood CA Jr, Green DP. Fractures in Adults. 5th ed. Philadelphia: Lippincott Williams & Wilkins, 2001.
2929
5. Harris JH Jr, Harris WH. The Radiology of Emergency Medicine. 4 t h ed. Baltimore: Lippincott Williams & Wilkins, 2000.
2930
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section X - Musculoskeletal Radiology > Chapter 44 - Arthritis
Chapter
44
Arthritis Clyde A. Helms
OSTEOARTHRITIS Osteoarthritis, or degenerative joint disease (DJD), is the most common arthritide. It is believed to be caused by trauma—either overt or as an accumulation of microtrauma over years, although there is also a hereditary form called primary osteoarthritis that occurs primarily in middle-aged women. The hallmarks of DJD are joint space narrowing, sclerosis, and osteophytosis (Table 44.1 and Fig. 44.1). If all three of these findings are not present on a radiograph, another diagnosis should be considered. Joint space narrowing is the least specific finding of the three, yet it is virtually always present in DJD. Unfortunately, it is also seen in almost every other joint abnormality. Sclerosis should be present in varying amounts in all cases of DJD unless severe osteoporosis is present. Osteoporosis will cause the sclerosis to be diminished. For instance, in long-standing rheumatoid arthritis in which the cartilage has been destroyed, DJD often occurs with very little sclerosis. Osteophytosis will also be diminished in the setting of osteoporosis. Otherwise, sclerosis and osteophytosis should be prominent in DJD. The only disorder that will cause osteophytes without sclerosis or joint space narrowing is diffuse idiopathic skeletal hyperostosis (1) . This is a common bone-forming disorder that at first glance
2931
resembles DJD, except that there is no joint space narrowing (or disk space narrowing in the spine) and there is no sclerosis (Fig. 44.2) . Diffuse idiopathic skeletal hyperostosis (DISH) is not believed to be caused by trauma or stress, as is DJD, and is not painful or disabling, as DJD can be. Millions of dollars per year are awarded to federal employees upon retirement, representing disability payments for supposed DJD acquired during their employment, when in fact, these retirees have diffuse idiopathic skeletal hyperostosis and have been misdiagnosed. Osteoarthritis is divided into two types: primary and secondary. Secondary osteoarthritis is what radiologists refer to when speaking of DJD. It is, as mentioned, secondary to trauma of some sort. It can occur in any joint P.1133 in the body but is particularly common in the hands, knees, hips, and spine.
TABLE 44.1 Hallmarks of Degenerative Joint Disease
Joint space narrowing Sclerosis Osteophytes
Primary
osteoarthritis is a familial arthritis that affects middle-
aged women almost exclusively and is seen only in the hands. It affects the distal interphalangeal joints, the proximal interphalangeal joints, and the base of the thumb in a bilaterally symmetric fashion (Fig. 44.3). If it is not bilaterally symmetric, the diagnosis of primary osteoarthritis should be questioned. A type of primary osteoarthritis that can be very painful and
2932
debilitating is erosive osteoarthritis. It has the identical distribution mentioned for primary osteoarthritis but is associated with severe osteoporosis of the hands, as well as erosions. It is uncommon, and radiologists generally see little of this disorder. It is also called Kellgren arthritis.
FIGURE 44.1. Osteoarthritis (DJD). Plain film of a finger with osteoarthritis (DJD) of the distal and proximal interphalangeal joints. Both joints demonstrate joint space narrowing, subchondral sclerosis, and osteophytosis, which are hallmarks of DJD.
2933
FIGURE
44.2. Diffuse
Idiopathic
Skeletal
Hyperostosis
(DISH). A lateral view of the lumbar spine shows extensive osteophytosis without significant disk space narrowing or sclerosis. This is a classic picture for diffuse idiopathic skeletal hyperostosis.
There are a few exceptions to the classic triad of findings seen in DJD (sclerosis, joint space narrowing, and osteophytes). Several joints also exhibit erosions as a manifestation of DJD: the temporomandibular joint, the acromioclavicular joint, the sacroiliac joints, and the symphysis pubis (Table 44.2). When erosions are seen in one of these joints, DJD must be considered, or inappropriate
2934
treatment may be instituted (Fig. 44.4) . A subchondral cyst, or geode (taken from the geologic term used when a volcanic rock has a gas pocket that leaves a large cavity in the rock), is often found in joints P.1134 affected with DJD. Geodes are cystic formations that occur around joints in a variety of disorders (including, in addition to DJD, rheumatoid arthritis, calcium pyrophosphate dihydrate crystal deposition disease (CPPD) and avascular necrosis) (Table 44.3) (2) . Presumably, one method of geode formation takes place when synovial fluid is forced into the subchondral bone, causing a cystic collection of joint fluid. Another etiology is following a bone contusion, in which the contused bone forms a cyst. They rarely cause problems by themselves but are often misdiagnosed as something more sinister (Fig. 44.5) .
TABLE 44.2 Joints That Have Erosions as a Feature of Degenerative
Joint
Sacroiliac Acromioclavicular Temporomandibular Symphysis pubis
2935
Disease
FIGURE
44.3. Primary
Osteoarthritis. Bilateral hand films in a
patient with primary osteoarthritis. Present are classic findings of osteophytosis, joint space narrowing, and sclerosis at the distal interphalangeal joints, the proximal interphalangeal joints, and at the base of the thumb. This is bilaterally symmetric, which is typical
for
primary
osteoarthritis.
2936
FIGURE
44.4. Osteoarthritis of the Sacroiliac (SI) Joint. A
young woman who is a professional dancer complained of leftsided hip pain. An anteroposterior film of the pelvis demonstrated left SI joint sclerosis, joint irregularity, and erosions. A complete workup to rule out a human leukocyte antigen (HLA) –B27 spondyloarthropathy was negative, and no laboratory or clinical evidence for infection was found. Her clinical history pointed to this being completely occupationrelated, and an aspiration biopsy to rule out infection was therefore not performed. This is not an unusual appearance for degenerative joint disease of the SI joint.
RHEUMATOID
ARTHRITIS
Rheumatoid arthritis is a connective tissue disorder of unknown etiology that can affect any synovial joint in the body. The radiographic
hallmarks
are soft tissue swelling, P.1135
osteoporosis, joint space narrowing, and marginal erosions. In the hands, it is classically a proximal process that is bilaterally symmetric (Table 44.4 and Fig. 44.6). There are so many exceptions
2937
to these rules, however, that I have come to regard them as no better than 80% accurate. Rheumatoid arthritis has a large variety of appearances, and from its radiographic appearance alone, it can be very difficult to diagnose with any degree of assurance.
TABLE 44.3 Diseases in Which Geodes Are Found
Degenerative joint disease Rheumatoid arthritis Calcium pyrophosphate Avascular necrosis
dihydrate
2938
deposition
disease
(CPPD)
FIGURE 44.5. Subchondral Cyst or Geode of the Shoulder. This patient has marked degenerative joint disease (DJD) of the shoulder, with joint space narrowing, sclerosis, and osteophytosis. A large lytic process (arrows) is seen in the humeral head, which is a subchondral cyst or geode often seen in association with DJD. Because of the DJD in the shoulder, a biopsy to rule out a more sinister lesion in the humeral head should be avoided.
Rheumatoid arthritis in large joints is fairly characteristic in that it causes marked joint space narrowing and is associated with osteoporosis. Erosions might or might not be present and tend to be marginal—that is, away from the weight-bearing portion of the joint. In the hip, the femoral head tends to migrate axially, whereas
2939
in osteoarthritis, it tends to migrate superolaterally (Figs. 44.7, 44.8). In the shoulder, the humeral head tends to be “highriding― (Fig. 44.9). Other things to think of when confronted with a high-riding shoulder are a torn rotator cuff and calcium pyrophosphate dihydrate deposition disease (CPPD) (Table 44.5) .
TABLE 44.4 Hallmarks of Rheumatoid Arthritis
Soft tissue swelling Osteoporosis Joint space narrowing Marginal erosions Proximal distribution (hands) Bilateral
FIGURE
symmetry
44.6. Rheumatoid
Arthritis. An erosive arthritis
2940
affecting primarily the carpal bones and the metacarpophalangeal joints is seen that has associated osteoporosis and soft tissue swelling (note the soft tissue over the ulnar styloid processes). It is a bilaterally symmetric process in this patient, which is classic.
FIGURE 44.7. Migration of the Femoral Head. A drawing of the hip showing routes of migration of the femoral head. Osteoarthritis of the hip tends to cause superior (S) migration of the femoral head in relation to the acetabulum, whereas rheumatoid arthritis tends to cause axial (A) migration of the femoral head in relation to the acetabulum.
2941
FIGURE 44.8. Rheumatoid Arthritis of the Hip. Note the severe joint space narrowing in this patient with rheumatoid arthritis. The femoral head has migrated in an axial direction, with fairly concentric joint space narrowing. Minimal secondary degenerative changes have occurred, as noted by the sclerosis in the superior portion of the joint; however, these have been diminished somewhat by the osteoporosis that usually accompanies
rheumatoid
arthritis.
P.1136 When rheumatoid arthritis is long-standing, it is not unusual for secondary DJD to superimpose itself on the findings one would
2942
expect with rheumatoid arthritis. This picture of DJD differs somewhat from that usually seen, in that the sclerosis and osteophytes are considerably diminished in severity as compared with the joint space narrowing (Fig. 44.10) .
HLA-B27
SPONDYLOARTHROPATHIES
A group of diseases that was formerly known as rheumatoid variants is now known as the seronegative, HLA-B27–positive spondyloarthropathies. These disorders are all linked to the human leukocyte antigen (HLA)-B27 histocompatibility antigen. Included in this group of diseases are ankylosing spondylitis, inflammatory bowel disease, psoriatic arthritis, and Reiter syndrome (also called reactive arthritis). They are characterized by bony ankylosis, proliferative formation of new bone, and predominantly axial (spinal) involvement.
FIGURE 44.9. Rheumatoid Arthritis in the Shoulder. An anteroposterior view of the shoulder in this patient with
2943
rheumatoid arthritis shows that the distance between the acromion and the humeral head is diminished (arrowheads) . Ordinarily, this space is about 1 cm in width to allow the rotator cuff to pass freely beneath the acromion. This is a common finding in rheumatoid arthritis as well as in calcium pyrophosphate dihydrate deposition disease.
One of the more characteristic findings in these disorders is that of syndesmophytes in the spine. A syndesmophyte is a paravertebral ossification that resembles an osteophyte, except that it runs vertically, whereas an osteophyte has its orientation in a horizontal axis. Sometimes it can be difficult to decide whether a particular paravertebral ossification is an osteophyte or a syndesmophyte based on its orientation alone (Fig. 44.11). Bridging osteophytes and large syndesmophytes can have a similar appearance, with an orientation halfway between vertical and horizontal. To evaluate those cases, one must look at the other vertebral bodies and use the ossifications on them to determine whether they are osteophytes or syndesmophytes. If no other level is involved, the two might not be distinguishable. Syndesmophytes are classified as to whether they are marginal and symmetric or nonmarginal and asymmetric. A marginal syndesmophyte has its origin at the edge P.1137 or margin of a vertebral body and extends to the margin of the adjacent vertebral body. They are invariably bilaterally symmetric, as viewed on an anteroposterior (AP) spine film. Ankylosing spondylitis classically has marginal, symmetric syndesmophytes (Fig. 44.12) . Inflammatory bowel disease has an identical appearance when the spine is involved. Nonmarginal, asymmetric syndesmophytes are generally large and bulky. They emanate from the vertebral body, away from the endplate or margin, and are unilateral or asymmetric as viewed on an AP spine film (Figs. 44.11, 44.13). Psoriatic arthritis and Reiter syndrome classically have this type of syndesmophyte.
2944
TABLE 44.5 Causes of High-Riding Shoulder
Rheumatoid
arthritis
Calcium pyrophosphate Torn rotator cuff
dihydrate
2945
deposition
disease
(CPPD)
FIGURE 44.10. Secondary Degenerative Joint Disease (DJD) in the Knee in a Patient With Rheumatoid Arthritis. This patient has a history of long-standing rheumatoid arthritis. An anteroposterior view of the knee shows severe osteoporosis and joint space narrowing. Secondary DJD is occurring, as evidenced by the sclerosis and osteophytosis; however, these findings are out of proportion to the severe joint space narrowing. When DJD narrows a joint to this extent, the osteophytosis and sclerosis are invariably
much
more
pronounced.
Involvement of the sacroiliac (SI) joints is common in the HLA-B27 spondyloarthropathies. The patterns of involvement, like the patterns of involvement of the spine, are somewhat typical for each disorder. Ankylosing spondylitis and inflammatory bowel disease typically cause bilaterally symmetric SI joint disease, which is initially erosive in nature and progresses to sclerosis and fusion (Figs. 44.14, 44.15) . It is extremely unusual to have asymmetric or unilateral SI joint disease in these two disorders.
2946
FIGURE
44.11. Psoriasis
With
Syndesmophytes. The large
paravertebral ossification on the left side of the T12-L1 disk space (open arrow) is difficult to differentiate between an osteophyte and a syndesmophyte. Either could have this appearance. The paravertebral ossification at the left L1-L2 disk space (large solid arrow) definitely has a vertical rather than a horizontal orientation, however, as does the faint ossification seen at the T11-T12 disk space (small solid arrow). These definitely represent syndesmophytes. It makes sense, therefore, to assume that the ossification at the T12-L1 disk space is almost certainly a syndesmophyte as well. This patient has large
2947
nonmarginal, asymmetric syndesmophytes, which are typical of psoriatic arthritis or Reiter disease. This patient has psoriatic arthritis.
Reiter syndrome and psoriatic arthritis can exhibit unilateral or bilateral SI joint involvement. It seems that it is bilateral about 50% of the time. It is often asymmetric when it is bilateral, but exact symmetry can be difficult to assess; therefore, when involvement is definitely bilateral and not clearly asymmetric, consider the SI joints to be in the bilateral symmetric category. This means that if there is bilateral, symmetric SI joint disease, it could be caused by any of the four
HLA-B27
spondyloarthropathies.
If
there P.1138 P.1139
is unilateral (or clearly asymmetric) SI joint involvement, one can exclude ankylosing spondylitis and inflammatory bowel disease and consider Reiter syndrome or psoriatic disease. In this latter example, one would also have to consider infection and DJD (remember that DJD can cause erosions in the SI joints). Although it is seen less commonly, gout can also affect the SI joints unilaterally (Table 44.6 and Figs. 44.4, 44.16) .
2948
FIGURE
44.12. Marginal,
Ankylosing
Symmetric
Syndesmophytes
in
Spondylitis. Bilateral marginal syndesmophytes are
seen bridging the disk spaces throughout the lumbar spine in this patient. This is a so-called bamboo spine and is classic for ankylosing
spondylitis
and
inflammatory
2949
bowel
disease.
FIGURE 44.13. Syndesmophytes in Psoriatic Arthritis. Large, bulky, nonmarginal, asymmetric syndesmophytes (arrows) are seen in this patient with psoriatic arthritis.
2950
FIGURE
44.14. Ankylosing
Spondylitis. Bilateral symmetric
sacroiliac joint sclerosis and erosions are seen in this patient with ankylosing spondylitis. Inflammatory bowel disease could have a similar appearance. Although this is classic for these two disorders, it would not be that unusual for psoriatic disease or Reiter syndrome to have this same appearance. Although less likely, it would be possible for infection and even degenerative joint disease to be bilateral in this fashion.
2951
FIGURE
44.15. Fusion of the Sacroiliac (SI) Joints in
Ankylosing Spondylitis. Bilateral complete fusion of the SI joints in this patient with ankylosing spondylitis makes the SI joints totally indistinguishable. Inflammatory bowel disease could have a similar appearance.
CT can be very helpful in examining the SI joints and is considered by many to be the diagnostic procedure of choice because of the unobstructed view of the entire joint (Fig. 44.17) . Large joint involvement with the HLA-B27 spondyloarthropathies is uncommon (except for ankylosing spondylitis), but when it does occur, the arthropathy will resemble rheumatoid arthritis (Fig. 44.18). The hips are involved in up to 50% of patients with ankylosing
spondylitis.
Small joint involvement, specifically in the hands and feet, is not common in ankylosing spondylitis and inflammatory bowel disease. Psoriatic arthritis causes a distinctive arthropathy that is characterized by its distal predominance, proliferative erosions, soft
2952
tissue swelling, and periostitis. Proliferative erosions are different from the clean-cut, sharply marginated erosions seen in all other erosive arthritides, in that they have fuzzy margins with wisps of periostitis emanating from them (Fig. 44.19A). The severe forms are often associated with bony ankylosis across joints (Fig. 44.19B) and arthritis mutilans deformities. A fairly common finding is a calcaneal heel spur that has fuzzy margins, as opposed to the well-corticated heel spur seen in DJD or posttraumatically (Fig. 44.20) .
TABLE 44.6 Causes of Sacroiliac Joint Disease
Ankylosing Inflammatory
spondylitis bowel
disease
Psoriatic arthritis Reiter syndrome Infection Degenerative Gout
joint
disease
Reiter syndrome causes identical changes in every respect to psoriatic arthritis, with the exception that the hands are not as commonly involved as the feet. The P.1140 P.1141 interphalangeal joint of the great toe is a commonly affected location in Reiter disease (Fig. 44.21) .
2953
FIGURE
44.16. Psoriatic Arthritis With Sacroiliac (SI) Joint
Disease. Unilateral SI sclerosis and erosions are seen in this patient with psoriatic arthritis. Ankylosing spondylitis and inflammatory bowel disease virtually never have this appearance.
2954
FIGURE 44.17. CT of the Sacroiliac (SI) Joints in Psoriasis. A CT scan through the SI joints in this patient with psoriatic arthritis shows unilateral SI joint sclerosis and erosions (arrows) , typical for psoriatic arthritis or Reiter disease.
FIGURE
44.18. Ankylosing Spondylitis With Hip Disease.
Anteroposterior view of the pelvis in this patient with ankylosing spondylitis shows bilateral complete fusion of the sacroiliac joints. Concentric left hip joint narrowing is present with axial migration of the femoral head. This would be a typical finding in rheumatoid arthritis or, as in this example, in ankylosing spondylitis. Note the secondary degenerative joint disease changes in the left hip as well.
2955
FIGURE 44.19. Psoriatic Arthritis. A. Cartilage loss at the proximal interphalangeal joints of the third, fourth, and fifth digits in this hand is apparent, with erosions noted most prominently in the fourth digit (arrow). These erosions are not sharply demarcated but are covered with fluffy new bone. These are termed proliferative erosions. Note also the periostitis along the shafts of each of the proximal phalanges. B . Advanced psoriatic arthritis. Fusion or ankylosis is apparent across the proximal interphalangeal joints of the second through the fifth digits. Several of the distal interphalangeal joints are also ankylosed. Severe joint space narrowing at the metacarpophalangeal joints is noted. This distal distribution is typical for psoriatic arthritis in advanced stages.
2956
FIGURE 44.20. Reiter Syndrome. Lateral view of a calcaneus in a patient with Reiter syndrome shows poorly defined new bone on the posteroinferior margin of the calcaneus with a calcaneal spur which is also poorly defined. This is typical of psoriatic or Reiter disease, as opposed to the well-formed calcaneal spur in degenerative
joint
disease.
CRYSTAL-INDUCED
ARTHRITIS
The crystal-induced arthritides include, primarily, gout and pseudogout (CPPD). Ochronosis and Wilson disease are so rare that they are not discussed in this chapter. Gout is a metabolic disorder that results in hyperuricemia and leads to deposition of monosodium urate crystals in various sites in the body, especially in the joint cartilage. The actual causes of the hyperuricemia are myriad and include heredity.
2957
The arthropathy caused by gout is very characteristic radiographically. It takes 4 to 6 years for gout to cause radiographically evident disease, and most patients are treated successfully long before the destructive arthropathy occurs; therefore, gouty arthritis is not commonly encountered. The classic radiographic findings in gout are well-defined erosions, often with sclerotic borders or overhanging edges; soft tissue nodules that calcify in the presence of renal failure; and a random distribution in the hands without marked osteoporosis (Table 44.7 and Fig. 44.22) . Even though erosions with overhanging edges occur with gout, they can occur in other disorders as well and are by no means pathognomonic. The sclerotic margins of the erosions are rarely seen in any other arthritide; therefore, this is a very useful differential point. Gout typically affects the metatarsophalangeal joint of the great toe (Fig. 44.23). In the advanced stages, it can be very deforming (Fig. 44.24). Patients with gout often have chondrocalcinosis because they have a predisposition for pseudogout (CPPD). As many as 40% of patients with gout concomitantly have CPPD.
2958
FIGURE 44.21. Reiter Syndrome. Anteroposterior view of the large toe in a patient with Reiter disease shows fluffy periostitis (arrow) in the erosions adjacent to the interphalangeal joint of the great toe. Marked soft tissue swelling is also present throughout the great toe. These changes are typical in appearance and location for Reiter disease or psoriatic arthritis.
2959
TABLE 44.7 Hallmarks of Gout
Well-defined
erosions
(sclerotic
margins)
Soft tissue nodules Random distribution No osteoporosis
FIGURE 44.22. Gout. Sharply marginated erosions, some with a sclerotic margin, are noted throughout the carpus and proximal metacarpals. These erosions are classic in gout. Note the absence of marked demineralization.
2960
FIGURE
44.23. Gout. A sharply marginated erosion with an
overhanging edge (arrow) and a sclerotic margin is seen in the metatarsophalangeal in the great toe in this patient with gout. This appearance and location are classic for gout, whereas psoriatic arthritis and Reiter disease usually involve the interphalangeal joint and do not have erosions that are this sharply marginated.
2961
FIGURE
44.24. Advanced
Gout. Marked diffuse and focal soft
tissue swelling is present throughout the hand and wrist in this patient with long-standing gout. Destructive, large, wellmarginated erosions, some with overhanging edges, are noted near multiple joints. The focal areas of soft tissue swelling are called tophi, some of which are calcified. These only calcify with coexistent renal disease.
2962
FIGURE
44.25. Chondrocalcinosis in the Knee. Cartilage
calcification known as chondrocalcinosis is seen in the fibrocartilage (white arrow) and in the hyaline articular cartilage (black
arrow) in this patient with calcium pyrophosphate
dihydrate
deposition
disease.
2963
FIGURE
44.26. Chondrocalcinosis in the Wrist. This patient
with calcium pyrophosphate dihydrate deposition disease exhibits chondrocalcinosis in the triangular fibrocartilage of the wrist (curved arrow). A small amount of chondrocalcinosis is also seen in the second metacarpophalangeal (small arrow). Triangular ligament calcification is one of the more common locations for chondrocalcinosis to occur.
P.1142 P.1143 Pseudogout has a classic triad: pain, cartilage calcification, and
2964
joint destruction. The patient may have any combination of one or more of this triad at any one time. Each of these is addressed individually in some detail in this chapter, but note that two of the three are radiographic findings. This is a disorder that is best diagnosed radiographically. The pain of CPPD is nonspecific. It can mimic that of gout (hence the term “pseudogout―), infection, or just about any arthritis. It typically is intermittent for a large number of years until DJD occurs and becomes the main cause of pain.
TABLE 44.8 Most Common Location of Chondrocalcinosis in Calcium
Pyrophosphate
Knee Triangular fibrocartilage Symphysis pubis
of
Dihydrate
wrist
2965
Deposition
Disease
FIGURE
44.27. Calcium
Pyrophosphate
Dihydrate
Crystal
Deposition Disease (CPPD) Arthropathy. Degenerative joint disease (DJD) of the elbow is seen in this patient with CPPD. Note the joint space narrowing, with minimal sclerosis and large osteophytes (arrows). Osteophytes of this nature are termed drooping osteophytes and are often seen in CPPD. The elbow is an unusual place for DJD to occur except in the setting of CPPD or trauma.
Cartilage calcification, known as chondrocalcinosis, can occur in any joint but tends to affect a few select sites in most patients. These are the medial and lateral compartments of the knee (Fig. 44.25), the triangular fibrocartilage of the wrist (Fig. 44.26), and the symphysis pubis (Table 44.8). Chondrocalcinosis in these areas is virtually diagnostic of CPPD (3). When CPPD crystals are deposited in the soft tissues, such as in the rotator cuff of the shoulder, a radiograph
2966
cannot differentiate between CPPD and calcium hydroxyapatite, which occurs in calcific tendinitis. Calcium hydroxyapatite does not occur in the joint cartilage except in extremely unusual cases; therefore, all chondrocalcinosis can be considered to be secondary to CPPD. The joint destruction or arthropathy of CPPD is virtually indistinguishable from that of DJD. In fact, it is DJD. It is caused by CPPD crystals eroding the cartilage. There are a few features of the DJD caused by CPPD that will help distinguish it from DJD caused by trauma or overuse, however. The main difference is location. The DJD of CPPD has a proclivity for the shoulder, the elbow (Fig. 44.27) , the radiocarpal joint in the wrist (Fig. 44.28), the patellofemoral joint of the knee, and the metacarpophalangeal joints in the hand (Table 44.9). These are areas of wear and tear not normally involved by DJD (such as in the distal P.1144 interphalangeal joints of the hand, the hip, and the medial compartment of the knee). When DJD is seen in the joints that CPPD tends to involve, a search for chondrocalcinosis should be made. If necessary, a joint aspiration for CPPD crystals may be required to confirm
the
diagnosis.
2967
FIGURE 44.28. Calcium Pyrophosphate Dihydrate Crystal Deposition Disease (CPPD) Arthropathy. Marked degenerative joint disease (DJD) at the radiocarpal joint is seen in this patient with CPPD. Severe joint space narrowing and sclerosis with large subchondral cysts or geodes are all hallmarks of DJD. This is an unusual location for DJD except in the setting of CPPD or trauma.
Occasionally, the arthropathy of CPPD causes such severe destruction that a neuropathic or Charcot joint is mimicked on the radiograph. This has been termed a pseudo-Charcot joint. It is not a true Charcot joint because of the presence of sensation (4) . There are three diseases that have a high degree of association with CPPD. These are primary hyperparathyroidism, gout, and hemochromatosis(Table 44.10). This is not a differential diagnosis for chondrocalcinosis; rather, these are diseases that tend to occur at the same time as CPPD. If the patient has one of these three disorders, he or she is more likely to have CPPD than is an unaffected person. There is probably no good reason to work up
2968
every patient with chondrocalcinosis for one of the three associated diseases because they are so uncommon and CPPD is extremely common.
TABLE 44.9 Most Common Location of Arthropathy in Calcium
Pyrophosphate
Shoulder Radiocarpal
Dihydrate
Deposition
Disease
joint
Patellofemoral joint Elbow Metacarpophalangeal joints hand
in
TABLE 44.10 Diseases With High Association With Calcium Pyrophosphate
Primary
Dihydrate
Deposition
Disease
hyperparathyroidism
Gout Hemochromatosis
COLLAGEN-VASCULAR Scleroderma,
systemic
lupus
DISEASES
erythematosus,
2969
dermatomyositis,
and
mixed connective tissue disease are all grouped together as collagenvascular diseases. The striking abnormality in the hands in each of these disorders is osteoporosis and soft tissue wasting. Systemic lupus erythematosus characteristically has severe ulnar deviation of the phalanges (Fig. 44.29). Erosions are generally not a feature of these disorders. Soft tissue calcifications are typically present in scleroderma (Fig. 44.30) and dermatomyositis. The calcifications in scleroderma are typically subcutaneous, whereas in dermatomyositis they are intramuscular in location. Mixed connective tissue disease is an overlap of scleroderma, systemic lupus erythematosus, polymyositis, and rheumatoid arthritis. It has a myriad of radiographic findings.
SARCOID Sarcoidosis is a disease that causes deposition of granulomatous tissue in the body—primarily in the lungs, but also in the bones. In the skeletal system, it has a predilection for the hands, where it causes lytic destructive lesions in the cortex. These often have a socalled lacelike appearance, which is characteristic (Fig. 44.31) . Sarcoidosis can have associated skin nodules inthe hands.
HEMOCHROMATOSIS Hemochromatosis is a disease of excess iron deposition in tissues throughout the body that leads to fibrosis and eventual organ failure. Twenty percent to 50% of patients with hemochromatosis have a characteristic arthropathy in the hands that should suggest the diagnosis. The classic radiographic changes are essentially DJD that involves the second through the fourth metacarpophalangeal joints (Fig. 44.32). Up to 50% of patients with hemochromatosis also have CPPD; therefore, a search should be made for chondrocalcinosis. Another finding that is often seen in hemochromatosis is called “squaring― of the metacarpal heads, which appear enlarged and blocklike as a result of the large osteophytes commonly seen in this P.1145
2970
disorder. The osteophytes are often said to be “drooping― because of the unusual way they hang off the joint margin.
FIGURE
44.29. Systemic
Lupus
Erythematosus. Marked soft
tissue wasting, as noted by the concavity in the hypothenar eminence, and ulnar deviation of the phalanges, seen primarily in the right hand, are hallmarks of systemic lupus erythematosus.
NEUROPATHIC
OR
CHARCOT
JOINT
The radiographic findings for a Charcot joint are characteristic and almost pathognomonic. A classic triad has been described that consists of destruction, 44.11 and Fig. 44.33) .
dislocation, and heterotopic new bone (Table
Joint destruction is seen in every type of arthritis and therefore seems very nonspecific; however, nothing causes such severe destruction in a joint as a Charcot joint. Progressive joint destruction
2971
occurs in a neuropathic joint because the joint is rendered unstable by inaccurate muscle action and is unprotected by intact nerve reflexes. Early in the development of a Charcot joint, the joint destruction may appear to be merely joint space narrowing. It is extremely difficult to make the diagnosis this early. In the spine, instead of joint space destruction, there is disk space destruction (Fig. 44.34) .
FIGURE 44.30. Scleroderma. Diffuse subcutaneous soft tissue calcification is seen throughout the hands and wrists in this patient with scleroderma. Soft tissue wasting and osteoporosis are also present, as well as bone loss in multiple distal phalanges secondary to the vascular abnormalities often present in this disease.
2972
Dislocation, like joint destruction, can be present in varying degrees. Early on, the joint may have subluxation instead of dislocation.
FIGURE 44.31. Sarcoid. Anteroposterior view of the hand in this patient with sarcoid demonstrates classic changes of bony involvement with this granulomatous process. Note the lacelike pattern of destruction, which is seen most prominently in the proximal phalanges and in the distal third phalanx. Soft tissue swelling and some areas of severe bony dissolution are also noted, which occur in more advanced patterns of sarcoid. These changes are typically limited to the hands but can rarely occur in other parts of the skeleton.
P.1146 Heterotopic new bone has also been termed debris or detritus and
2973
consists of soft tissue calcification or clumps of ossification adjacent to the joint. It, too, can be present in varying amounts. The most commonly seen Charcot joint today is in the foot of a diabetic. The disease typically affects the first and second tarsometatarsal joints in a fashion similar to a Lisfranc fracture (Fig. 44.35) . Tabes dorsalis from syphilis is rarely seen today. More commonly seen is a Charcot joint in a patient with paralysis who continues to use the affected limb for support. A Charcot joint that is also seen on occasion is the so-called pseudo-Charcot joint in CPPD.
HEMOPHILIA, JUVENILE RHEUMATOID ARTHRITIS, AND PARALYSIS Why would clinically disparate entities like paralysis, juvenile rheumatoid arthritis (JRA), and hemophilia be covered in the same section?
Because
they
are
usually
radiographically
2974
indistinguishable.
FIGURE 44.32. Hemochromatosis. Anteroposterior view of the hand in this patient with hemochromatosis shows severe joint space narrowing throughout the hand, which is most marked at the metacarpophalangeal joints. Associated sclerosis at the metacarpophalangeal joints with large osteophytes seen off the metacarpal heads suggests degenerative joint disease (DJD). These are very unusual joints for DJD to occur in, yet this is the classic appearance of hemochromatosis. No chondrocalcinosis is seen in the triangular cartilage in this patient; however, a small amount of chondrocalcinosis can be seen at the second metacarpophalangeal joint (arrow). Fifty percent of patients with hemochromatosis also have calcium pyrophosphate dihydrate deposition disease.
The classic findings for JRA and hemophilia are overgrowth of the ends of the bones (epiphyseal enlargement) associated with gracile
2975
diaphyses (Fig. 44.36). Joint destruction might or might not be present. A finding that is purported to be classic for JRA and hemophilia is widening of the intercondylar notch of the knee. This sign can be quite variable and difficult to apply. It is rarely present when the other classic signs are not also present and obvious. Another process that can mimic the findings of JRA and hemophilia is a joint that has undergone disuse from paralysis (Fig. 44.37). It has always been said that the reason the epiphyses are overgrown in JRA and hemophilia is P.1147 P.1148 because of the hyperemia; however, many other things cause hyperemia without affecting the size of the epiphyses (such as rheumatoid arthritis and infection). The common denominator shared by JRA, hemophilia, and paralysis is disuse. This is most likely what causes the overgrowth of the ends of the bones that is seen in all three of these disorders.
TABLE 44.11 Hallmarks of a Neuropathic Joint
Joint destruction Dislocation Heterotopic new bone formation
2976
FIGURE
44.33. Charcot
Joint. Anteroposterior view of the knee
in this patient with tabes dorsalis shows the classic changes of a neuropathic or Charcot joint. Note the severe joint destruction, the subluxation, and the heterotopic new bone (arrow) .
2977
FIGURE
44.34. Charcot
Spine. Anteroposterior view of the
spine in this paraplegic patient shows severe destruction of the L2 and L3 vertebral bodies and the intervening disk space, heterotopic new bone (arrow), and malalignment or dislocation. Numbers
indicate
lumbar
vertebrae.
2978
FIGURE
44.35. Lisfranc
Charcot
Joint. Dislocation of the
second and third metatarsals along with joint destruction and large amounts of heterotopic new bone are present in the foot of this diabetic patient. These findings are classic for a Charcot joint, which has been termed a Lisfranc fracture-dislocation. It is most commonly seen secondary to trauma rather than as a Charcot joint but is the most common neuropathic joint seen today.
2979
FIGURE 44.36. Juvenile Rheumatoid Arthritis (JRA). A lateral view of the knee in this patient with JRA shows the classic findings of overgrowth of the ends of the bones and associated gracile diaphyses. These changes can also be seen in patients with hemophilia or paralysis.
2980
FIGURE
44.37. Muscular
Dystrophy
Simulating
Juvenile
Rheumatoid Arthritis (JRA) or Hemophilia. Anteroposterior view of the ankle in this patient with muscular dystrophy shows subtle changes of overgrowth of the distal tibia and fibular epiphyses. Marked tibiotalar slant, which can also be present in JRA or hemophilia, is also present.
SYNOVIAL
OSTEOCHONDROMATOSIS
Synovial osteochondromatosis is a relatively common disorder caused by a metaplasia of the synovium that results in deposition of foci of cartilage in the joint. Most of the time, these cartilaginous deposits calcify and are readily seen on a radiograph (Fig. 44.38). It is most commonly seen in the knee, hip, and elbow. Up to 30% of the time,
2981
the cartilaginous deposits do not calcify. In these cases, all that is seen on the radiograph is a joint effusion, unless erosions or joint destruction occur (Fig. 44.39) . The calcifications begin in the synovium and then tend to shed into the joint, where they can cause symptoms of free fragments, or “joint mice.― They then embed into the synovium and tend not to be free in the joint after a while. It is usually necessary to perform a complete synovectomy to relieve the symptoms.
FIGURE 44.38. Synovial Osteochondromatosis. Anteroposterior view of the hip in this patient with left hip pain shows multiple calcified loose bodies in the hip joint, which is virtually diagnostic of synovial osteochondromatosis.
2982
FIGURE
44.39. Synovial
Osteochondromatosis
Without
Calcification. Anteroposterior view of the hip in this patient shows that the femoral neck is eroded, with the femoral head having an “apple core― appearance. This has occurred from the pressure erosion of multiple nonossified loose bodies in the joint. This is nonossified synovial osteochondromatosis (probably more properly termed synovial chondromatosis). It usually does not cause this degree of bony erosion and is indistinguishable from pigmented villonodular synovitis.
2983
FIGURE 44.40. Tumefactive Synovial Osteochondromatosis. Plain film of the shoulder (A) shows a partially calcified mass which is eroding the medial aspect of the humerus. Coronal proton-density MR (B) and T2WI (C) of the shoulder reveal a large mass encircling the humeral head, which was interpreted as a sarcoma. A biopsy was performed and called “chondrosarcoma,― which resulted in a forequarter amputation. The intra-articular nature of the mass was not appreciated until after the radical surgery, when it was correctly
2984
recognized
as
synovial
chondromatosis.
P.1149 An uncommon presentation that can lead to diagnostic confusion is when the loose bodies are tightly packed in a joint, giving it the appearance of a tumor on MR (Fig. 44.40). This has been termed tumefactive synovial chondromatosis. If a biopsy is performed, it can get interpreted as a chondrosarcoma and treated with resultant radical surgery. Because no malignant tumors arise in joints, this should not present a problem in diagnosis.
PIGMENTED
VILLONODULAR
SYNOVITIS
Pigmented villonodular synovitis (PVNS) is a rare chronic inflammatory process of the synovium that causes synovial proliferation. A swollen joint with lobular masses of synovium occurs and causes pain and joint destruction (Fig. 44.41). It rarely, if ever, calcifies. It has been termed giant cell tumor of tendon sheath and tendon sheath xanthoma when it occurs in a tendon sheath, which is not unusual. Joints with PVNS look radiographically identical to noncalcified synovial osteochondromatosis, yet they are much less common. Therefore, whenever PVNS is a consideration, synovial chondromatosis should be mentioned. PVNS has a characteristic appearance on MR, with low-signal hemosiderin seen lining the synovium on both T1WIs and T2WIs (Fig. 44.42) .
SUDECK
ATROPHY
Also known as shoulder-hand syndrome and reflex sympathetic dystrophy, Sudeck atrophy is a poorly understood joint affliction that typically occurs after minor trauma to an extremity, resulting in pain, swelling, and dysfunction. Severe, patchy osteoporosis and soft tissue swelling are seen radiographically (Fig. 44.43). It typically affects the distal part of an extremity, such as a hand or foot, yet intermediate joints such as the knee and hip are believed by some to be occasionally involved. The pain usually
2985
P.1150 subsides, but the osteoporosis may persist. With time, the swelling will subside and the skin may become atrophic. It is important for the radiologist to recognize the aggressive osteoporosis in this disorder and differentiate it from disuse osteoporosis so that the treating physician can begin aggressive physical therapy.
FIGURE 44.41. Pigmented Villonodular Synovitis. Anteroposterior view of the hip in this patient shows joint space destruction and bony erosions throughout the femoral head and neck. Pigmented villonodular synovitis or synovial chondromatosis could have this appearance.
2986
FIGURE Sagittal
44.42. Pigmented
Villonodular
Synovitis
(PVNS).
T1W (A) and fast spin-echo T2W (B) images of an ankle
with PVNS show a soft tissue mass emanating from the ankle joint, which is low signal on both sequences and has very low signal hemosiderin lining parts of the synovium, which is characteristic
JOINT
for
PVNS.
EFFUSIONS
Most joint effusions are clinically obvious and do not radiographic validation. The elbow is an exception. In trauma to the elbow, an effusion indicates a fracture. radiographic signs of an elbow effusion are generally
require the setting of The clearly seen
(displaced fat-pads, as described in Chapter 43) and have been proven valid. Clinical determination of an elbow effusion can be difficult; therefore, the radiologist can be very helpful in this area. Clinical determination of a hip effusion is also very difficult. The presence of a hip effusion can be valuable in certain clinical settings. For instance, a patient with pain in the hip and an effusion should have the joint aspirated to rule out an infection. If only pain is
2987
present, an aspiration would probably not be performed. The radiology literature mentions displacement of the fat stripes about the hip as an indicator for an effusion, but this has been proved to be unfounded. The only fat-pad around the hip that gets displaced with an effusion is the obturator internus, and it is seen uncommonly. The radiographic sign for a knee effusion that seems to be the most reliable is the measurement of the distance between the suprapatellar fat-pad and the anterior femoral fat-pad (Fig. 44.44). A distance between these two fat-pads of more than 10 mm is definite evidence for an effusion. A distance of less than 5 mm is normal. A distance of 5 to 10 mm is equivocal. It does not make any difference if there is an effusion in the knee—treatment is the same, regardless. If it were vital to the patient, one could aspirate the joint or perform an MR study to find out. I should P.1151 P.1152 point out that an MR should never be performed just to see whether there is fluid in the joint.
2988
FIGURE 44.43. Sudeck Atrophy. Diffuse soft tissue swelling and marked osteoporosis that is so aggressive it has a spotty or permeative appearance is noted around all of the joints in the hand. This patient experienced severe hand pain and dysfunction following minor trauma. This is characteristic of Sudeck atrophy.
2989
FIGURE
44.44. Knee Joint Effusion. This patient has joint fluid
in the knee, with widely displaced fat-pads. The suprapatellar fat-pad (left arrow) is more than 5 mm from the anterior suprafemoral
fat-pad
(right
arrow), which indicates a joint
effusion. The patella is fractured.
2990
FIGURE 44.45. Early Avascular Necrosis of the Hip. Patchy sclerosis is present in the femoral head in this patient with a renal transplant and avascular necrosis of the right hip. No subchondral lucency or articular surface irregularity in the weight-bearing region is yet present, with the exception of a small
cortical
irregularity
seen
laterally.
2991
FIGURE
44.46. Avascular Necrosis (AVN) of the Hip.
Subchondral lucency (arrows) is seen in the weight-bearing portion of this hip with AVN. Patchy sclerosis throughout the femoral head is also noted.
2992
FIGURE 44.47. Avascular Necrosis (AVN) of the Shoulder. Articular surface collapse is present in this shoulder with longstanding AVN. Dense bony sclerosis is also present.
Shoulder effusions are very difficult to detect unless they are massive enough to displace the humeral head inferiorly, as with a fracture and hemarthrosis (see Chapter 43). Fortunately, as with most other joints, treatment is not based solely on the presence or absence of an effusion, so it hardly matters. The same is true in the ankle, wrist, and smaller joints.
2993
FIGURE
44.48. Avascular Necrosis of the Hip. An axial T1WI
(time of repetition 600; time of echo 20) of the hips shows a focal area of abnormality in the left femoral head (arrow), which is characteristic for AVN. The low-signal, serpiginous border is a typical finding, as is the anterior location.
FIGURE 44.49. Avascular Necrosis (AVN) of the Hip. Coronal T1WI (A) (time of repetition 600; time of echo 20) and coronal STIR (B) images show bilateral AVN.
2994
AVASCULAR
NECROSIS
Avascular necrosis (AVN), or osteonecrosis, can occur around almost any joint for a host of reasons, including steroid use, trauma, a variety of underlying disease states, and even idiopathically. It is often seen in renal transplant patients. The hallmark of AVN is increased bone density at an otherwise normal joint. Increased density at a narrowed joint usually indicates DJD; however, if either osteophytes or joint space narrowing are absent, another disorder should be considered. The earliest sign of AVN is a joint effusion. This often is not visible radiographically or is so nonspecific that it does not help with the diagnosis unless the clinical setting had already raised suspicion for AVN. The next sign for AVN is a patchy or mottled density (Fig. 44.45). In the P.1153 P.1154 knee, this density increase can occur throughout an entire condyle, whereas in the hip, it often involves the entire femoral head. Next, a subchondral lucency often develops that forms a thin line along the articular surface (Fig. 44.46). This lucent line has been described as being an early indicator for AVN, whereas, in fact, it is a late finding. Also, the lucent line stage is often not present in the evolution of AVN. Therefore, use of the lucent line as one of the main criteria for AVN can lead to missing early findings in some cases and missing the diagnosis completely in others.
2995
FIGURE
44.50. Osteochondritis
Dissecans.
A small focal area
of avascular necrosis (AVN) in the medial epicondyle of the femur (black arrows) is present, which is an area of osteochondritis dissecans. Part of the area of AVN has shed a bony fragment (white arrow) that is loose in the joint, which is known as a loose body or “joint mouse.―
2996
FIGURE 44.51. Osteochondritis Dissecans of the Talus. A focal area of avascular necrosis in the talus, as seen here (arrows), is called osteochondritis dissecans. The talus is the second most common site after the knee and, as in the knee, can cause a joint mouse, or loose body in the joint.
2997
FIGURE
44.52. Osteochondritis Dissecans of the Elbow. The
third most common site for osteochondritis dissecans is in the capitellum of the elbow. The faint lucency seen in this capitellum (arrows) was at first believed to be a chondroblastoma or an area of infection.
2998
FIGURE
44.53. Geode in the Hip. A large cystic lesion (arrows)
is seen in this patient with avascular necrosis (AVN) of the hip. Note the adjacent patchy sclerosis, indicative of AVN. A subchondral cyst or geode should be considered any time a lytic lesion is found around a joint.
The final sign in AVN is collapse of the articular surface and joint fragmentation (Fig. 44.47). I must stress that these changes all occur on only one side of a joint, which makes for an easy diagnosis because almost everything else around joints involves both sides of the joint. MR is extremely useful in evaluating AVN. It is the most sensitive imaging study available, often showing AVN when plain films or radionuclide scans are normal (5). In the hip, AVN typically has an area of low or mixed signal on T1WIs that is located in the anterosuperior portion of the femoral head (Figs. 44.48, 44.49). If the anterior portion of the femoral head is not involved, the diagnosis
of P.1155
2999
AVN should be questioned, as it is uncommon for this condition to present otherwise. Posterior femoral head AVN can occasionally be found after posterior dislocation of the hip because of impaction of the femoral head on the posterior column of the acetabulum.
FIGURE
44.54. Kienböck
Malacia. Avascular necrosis (AVN)
of the lunate, or Kienböck malacia, is demonstrated in this patient's wrist. The increased density and partial fragmentation of the lunate are characteristic for AVN. Also, note the slightly shortened ulna (in comparison with the radius), which is called negative ulnar variance. Negative ulnar variance is said to have a high association with Kienböck malacia.
3000
FIGURE 44.55. Köhler Disease. Flattening and sclerosis of the tarsal navicular (arrow) in children is thought by many to be avascular necrosis and is called Köhler disease. Others have found this to be an asymptomatic normal variant and believe that it is an incidental finding.
3001
FIGURE 44.56. Freiberg Infraction. Flattening, collapse, and sclerosis of the second metatarsal head, as seen in this patient, is typical of avascular necrosis or Freiberg infraction. It can also involve the second, third, or fourth metatarsal heads. Note the compensatory hypertrophy of the cortex of the second metatarsal, which is invariably found with this disorder.
A form of AVN that is smaller and more focal than that just described is osteochondritis dissecans. It is most likely caused by trauma; however, this is controversial, with one school of thought believing the cause is idiopathic. It occurs most often in the knee at the medial epicondyle (Fig. 44.50). It also is frequently seen in the dome of the talus (Fig. 44.51) and occasionally in the capitellum (Fig. 44.52) . Osteochondritis dissecans frequently leads to a small fragment of bone being sloughed off and becoming a free fragment in the joint,
3002
i.e., a “joint mouse― (see Fig. 44.50) . AVN is one of the disorders around joints in which subchondral cysts or geodes can occur. It is the only one of the four disorders (rheumatoid arthritis, DJD, and CPPD being the others) that can have an essentially normal joint and have a geode (Fig. 44.53). The other abnormalities will have any or a combination of joint space narrowing, findings.
osteophytes,
osteoporosis,
chondrocalcinosis,
or
other
FIGURE 44.57. Scheuermann Disease. Avascular necrosis of the apophyseal rings of the vertebral bodies is called Scheuermann disease. He originally described a painful kyphosis
3003
with multiple vertebral bodies involved. It is most commonly seen without kyphosis or pain and with only a few vertebral bodies involved.
FIGURE
44.58. Kienböck
Malacia. A coronal T1WI (time of
repetition 600; time of echo 20) of the wrist shows low signal throughout the lunate, which is characteristic for avascular necrosis of the lunate, or Kienböck malacia.
A host of names have been ascribed to certain bones with AVN, usually with the eponym being the first person to describe the disorder. These have been called osteochondroses. They are believed to be idiopathic for the most part but can also occur secondary to trauma. A few of the more common bones involved are the following: the carpal lunate in Kienböck malacia (Fig. 44.54); the tarsal navicular in Köhler disease (Fig. 44.55); the metatarsal heads in Freiberg infraction (Fig. 44.56); the femoral head in Legg-Perthes disease; the ring epiphyses of the spine in Scheuermann disease (Fig. 44.57); and the tibial tubercle in Osgood-Schlatter disease, also
3004
called surfer's knees. MR can be very useful in identifying AVN in these sites. It shows diffuse low signal on T1WIs that involves the entire area of AVN (Fig. 44.58) .
REFERENCES 1. Resnick D, Shaul S, Robins J. Diffuse idiopathic skeletal hyperostosis with extraspinal 1975;115:513–524.
manifestations.
Radiology
2. Resnick D, Niwayama G, Coutts R. Subchondral cysts (geodes) in arthritic disorders: pathologic and radiographic appearance of the hip joint. AJR Am J Roentgenol 1977;128:799–806. 3. Resnick D, Niwayama G, Goergen T, et al. Clinical, radiographic and pathologic abnormalities in calcium pyrophosphate dihydrate deposition disease (CPPD): pseudogout. Radiology 1977;122:1–15. 4. Helms CA, Chapman GS, Wild JH. Charcot-like joints in calcium pyrophosphate 1981;7:
dihydrate
deposition
disease.
Skeletal
Radiol
55–58.
5. Mitchell D, Kressel H, Arger P, et al. Avascular necrosis of the femoral head: morphologic assessment by MR imaging, with CT correlation.
Radiology
1986;161:739–742.
3005
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section X - Musculoskeletal Radiology > Chapter 45 - Metabolic Bone Disease
Chapter
45
Metabolic
Bone
Disease
Clyde A. Helms
OSTEOPOROSIS Osteoporosis is defined as diminished bone quantity in which the bone is otherwise normal. This contrasts to osteomalacia, in which the bone quantity is normal but the quality of the bone is abnormal in that it is not normally mineralized. Osteomalacia results in excess nonmineralized osteoid. It is not possible in most cases to distinguish between osteoporosis and osteomalacia on plain films; hence, many prefer the term “osteopenia― for the plain film finding of diminished mineralization. There are myriad causes of osteoporosis, the most common of which is primary osteoporosis (so-called senile osteoporosis or osteoporosis of aging). This is seen most commonly in postmenopausal women and is a major health concern because of the increase in vertebral body and hip fractures in this patient population. Secondary osteoporosis implies that an underlying disorder, such as thyrotoxicosis or renal disease, has caused the osteoporosis. Only about 5% of osteoporosis cases are of the secondary type. The differential diagnosis for secondary osteoporosis is quite long and probably should not be memorized. One cannot even be sure whether it is osteoporosis or osteomalacia on the basis of plain films; therefore, the differential for presumed osteoporosis would
3006
have to include the causes of osteomalacia. The main radiographic finding in osteoporosis is thinning of the cortex. Although this can be seen in any bone, it is most reliably demonstrated in the second metacarpal at the middiaphysis. Normal metacarpal cortical thickening should be approximately one fourth to one third the thickness of the metacarpal (Fig. 45.1). In osteoporosis, this cortical thickness is decreased (Fig. 45.2). The metacarpal cortex (and all bony cortices, for that matter) decreases in thickness normally with age and is thinner in women than in men of the same age. Several tables have been published that give normal metacarpal cortical measurements, with age and sex adjustments to assist in the determination of normal. Unfortunately, these only determine the mineralization of the peripheral skeleton and do not seem to relate to whether vertebral body or hip fractures will
occur.
Measurement of the bone mineral content in the axial skeleton can be done by one of several methods that use CT to assess bone quantity in the spine. There is much debate about which method is superior and even about whether knowledge of the bone mineral content is clinically more helpful than just knowing the age and sex of the patient, which is fairly accurate for predicting the bone mass quantity. Exercise and proper diet seem to help delay the onset of primary osteoporosis. Calcium additives have not been shown to reverse the process of primary osteoporosis. Estrogen clearly plays a role in alleviating postmenopausal osteoporosis, yet its use in a widespread manner
is
somewhat
controversial.
A type of osteoporosis that can be seen in a patient of any age is disuse osteoporosis. It results from immobilization from any cause, most commonly following treatment P.1157 of a fracture. The radiographic appearance of disuse osteoporosis differs from that of primary osteoporosis in that it occurs somewhat more rapidly and gives the bone a patchy appearance (Fig. 45.3) . This is from osteoclastic resorption in the cortex, which causes
3007
intracortical holes. If the disuse were to continue, the bone would resemble any bone with marked osteoporosis, that is, severe cortical thinning.
FIGURE 45.1. Normal Mineralization. The cortical width (arrows) at the mid–second metacarpal in this patient with normal mineralization is greater than one third of the total width of the metacarpal.
3008
FIGURE
45.2. Osteoporosis. Severe cortical narrowing
(arrows) at the midsecond metacarpal cortex is seen in this patient with severe osteoporosis. Note the intracortical tunneling, which occurs in more aggressive forms of osteoporosis.
3009
FIGURE
45.3. Disuse
Osteoporosis. A mottled, patchy
appearance is present in the proximal right femur in this patient with aggressive disuse osteoporosis secondary to an amputation. Note the mottled, irregular cortex seen in the femoral shaft, which is representative of cortical holes that can be seen in aggressive
osteoporosis.
Occasionally, aggressive osteoporosis from disuse can mimic a permeative lesion, such as a Ewing sarcoma or multiple myeloma, because of the multiple cortical holes that project over the medullary space, thus resembling a medullary permeative process (Fig. 45.4). The way to differentiate a true intramedullary permeative process from an intracortical process such as osteoporosis is to observe the cortex and see whether it is solid or
3010
riddled with holes (Fig. 45.5). If the cortex is solid, one can assume the permeative process is emanating from the medullary space (Fig. 45.6); if the cortex has multiple small holes, the clinician can assume the permeative pattern is from the cortical process. I call a permeative appearance that is secondary to cortical holes a “pseudopermeative― process to distinguish it from a true permeative process (1) . Other causes for a pseudopermeative process include hemangioma and radiation. A hemangioma can cause cortical holes in two ways: from focal hyperemia causing P.1158 focal osteoporosis, or by the blood vessels themselves tunneling through the cortex (Fig. 45.7). Radiation can cause cortical holes in bone and mimic a permeative pattern because of the death of cortical osteocytes, which can result in large lacunae in the cortex (Fig. 45.8). The cortical holes from radiation can be large, in which case they would not be confused with a true permeative process, but they can also be small and resemble an aggressive lesion.
3011
FIGURE
45.4. Aggressive
Osteoporosis. Multiple small holes
are seen in the cortex and overlying the medullary space in the proximal humerus of this patient who has suffered a stroke. This represents aggressive osteoporosis from disuse and is mimicking an aggressive permeative process. These holes are, however, almost entirely within the cortex of the bone.
3012
FIGURE
45.5. Differentiation of Permeative Process. A.
Schematic of a permeative lesion. A true permeative process has multiple small holes secondary to endosteal involvement, with sparing of the cortex. This represents a marrow process. B . Schematic of a pseudopermeative process. A pseudopermeative process such as osteoporosis has multiple small cortical holes that are then superimposed over the marrow, giving an appearance similar to that of a permeative process.
3013
FIGURE
45.6. Myeloma Causing a Permeative Process. A
diffuse permeative process throughout the femur is seen in this patient with myeloma. Note that the cortex is solid, although the endosteum has some scalloping. This is a true permeative process.
If a permeative lesion is found, the differential diagnosis is usually an aggressive process such as Ewing sarcoma, infection, or eosinophilic granuloma in a young person ( Table of Contents > Section X - Musculoskeletal Radiology > Chapter 46 - Skeletal “Don’t Touch― Lesions
Chapter
46
Skeletal Lesions
“Don’t
Touch―
Clyde A. Helms Skeletal “don’t touch― lesions are those processes that are so radiographically characteristic that a biopsy or additional diagnostic tests are unnecessary. Not only does the biopsy result in unnecessary morbidity and cost, but in some instances, as discussed in this chapter, a biopsy can also be frankly misleading and lead to additional unnecessary surgery. Most radiology training stresses giving a differential diagnosis of a lesion, leaving it up to the clinician to decide between the various entities. For the “don’t touch― lesions, however, a differential list is inappropriate, the decision-making process a need to undergo biopsy for a should be made without a list
as that often makes the next step in biopsy. Because these lesions do not final diagnosis, a radiologic diagnosis of differential possibilities. These
lesions can be classified into three categories: (1) posttraumatic lesions, (2) normal variants, and (3) lesions that are real but obviously
benign.
POSTTRAUMATIC Myositis
LESIONS
ossificans is an example of a lesion that should not
undergo biopsy because its aggressive histologic appearance can often mimic a sarcoma (1). Unfortunately, radical surgery has been
3044
performed based on the histologic appearance of myositis ossificans when the radiologic appearance was diagnostic. The typical radiologic appearance of myositis ossificans is circumferential calcification with a lucent center (Fig. 46.1). This is often best appreciated on a CT scan (Fig. 46.2). A malignant tumor that mimics myositis ossificans has an ill-defined periphery and a calcified or ossific center (Fig. 46.3). Periosteal reaction can be seen with myositis ossificans or with a tumor. Occasionally, the peripheral calcification of myositis ossificans can be too faint to appreciate; in these cases, a CT scan should help, or delayed films 1 or 2 weeks later are recommended. Biopsy should be avoided when myositis ossificans is a clinical consideration. MR can be misleading because the peripheral calcification is not as well seen, and edema in the soft tissues can extend beyond the calcific rim (Fig. 46.4) .
Avulsion
Injury
Another posttraumatic entity in which a biopsy can be misleading is any avulsion injury (2,3). These injuries can have an aggressive radiographic appearance, but because of their characteristic location at ligament and tendon insertion sites (e.g., anteroinferior iliac spine or ischial tuberosity), they should be recognized as benign. (Figs. 46.5, 46.6). As with myositis ossificans, delayed films of several weeks will usually allow the problem case to become more radiographically clear. Biopsy can lead to the mistaken diagnosis of a sarcoma and should therefore be avoided. Any area undergoing healing can have a high nuclear-to-chromatin ratio and a high mitotic figure count, thereby occasionally simulating a malignancy histologically. Cortical desmoid is a process on the medial supracondylar ridge of the distal femur that is considered by many to be the result of an avulsion of the adductor magnus muscle. It occasionally simulates an aggressive lesion radiographically, and histologically, it can look malignant (4). In many instances, biopsy has led to amputation for this benign, radiographically characteristic lesion (Figs. 46.7, 46.8) . Cortical desmoids occur only on the posteromedial epicondyle of the femur. They might or might not be associated with pain and can have
3045
increased radionuclide uptake on a bone scan. They might or might not exhibit periosteal new bone and usually occur in young people. Biopsy should be avoided in all cases. Painful cortical P.1169 P.1170 desmoids should become asymptomatic with rest. They are often seen as an incidental finding on MR of the knee and have a characteristic
appearance
(Fig. 46.9) .
FIGURE 46.1. Myositis Ossificans. A plain film of the femur (A) in this patient who presented with a soft tissue mass shows a calcific density adjacent to the posterior cortex of the femur, which is calcified primarily in its periphery. If it is difficult to determine from the plain film alone that this is definitely peripheral, circumferential calcification, a CT scan (B) can be helpful in showing that the calcification is unequivocally
3046
peripheral in nature. This is virtually diagnostic of myositis ossificans.
FIGURE 46.2. Myositis Ossificans. A. Hazy calcification is seen adjacent to the humeral shaft, with underlying periosteal reaction noted. It is difficult to ascertain whether the calcification is circumferential. B . A CT scan through this mass shows that the calcification is unequivocally circumferential in nature, making the diagnosis of myositis ossificans a certainty.
3047
FIGURE
46.3. Osteogenic
Sarcoma. Hazy, ill-defined
calcification is seen adjacent to the iliac wing in this patient, which can be ascertained from the plain film as definitely not circumferential in nature. Although a prior history of trauma was obtained in this case, myositis ossificans is not a consideration with this appearance of calcification. Biopsy showed this to be an osteogenic
sarcoma.
Trauma can lead to large, cystic geodes or subchondral cysts near joints and can be mistaken for other lesions, resulting in a biopsy being ordered. Although the biopsy specimen is not likely to mimic a malignant process, it is nevertheless avoidable. Because geodes from degenerative joint disease are almost always associated with additional findings such as joint space narrowing, sclerosis, and osteophytes, a diagnosis should be made radiographically (Fig.
3048
46.10). On occasion, however, the additional findings are subtle and can be missed (Fig. 46.11). Geodes can also occur in the setting of calcium pyrophosphate dihydrate crystal disease, rheumatoid arthritis, and avascular necrosis (5,6) .
Discogenic
Vertebral
Sclerosis
An entity that is often confused for metastatic disease to the spine is discogenic vertebral disease. It can mimic metastatic disease very closely, and unless the radiologist is familiar with this process, it can lead to an unnecessary biopsy (7). Discogenic vertebral disease most often is sclerotic and focal (Fig. 46.12). It is always adjacent to the endplate, and the associated disk space should be narrow. Osteophytosis is invariably present. It really is a variant of a Schmorl node and should not be confused with a metastatic focus. On occasion it can be lytic or even mixed lytic-sclerotic. The typical clinical setting is a middle-aged woman with chronic low back pain. Old films often confirm the benign nature of this process. In the setting of disk space narrowing and osteophytosis, focal sclerosis adjacent to an endplate should not undergo biopsy (8) .
Fracture Occasionally, a fracture will be the cause of extensive osteosclerosis and periostitis, which can mimic a primary bone tumor (Fig. 46.13) . Lack of immobilization can result in exuberant callus, which can be misinterpreted as aggressive periostitis or even new tumor bone. Results of a biopsy in such a case might resemble a malignant lesion; therefore, any case associated with trauma should be carefully reviewed for a fracture.
Pseudodislocation
of
the
Humerus
Another traumatic process that can be misdiagnosed radiologically, leading to inappropriate treatment and morbidity, is a pseudodislocation of the humerus (Fig. 46.14). This results from a fracture with hemarthrosis, which causes distension of the joint and
3049
migration of the humeral head inferiorly (9). An axial or transscapular view shows it is not anteriorly or posteriorly dislocated (the usual forms of shoulder dislocation) but merely inferiorly subluxated. On an anteroposterior view, it can mimic a posterior dislocation in that the normal superimposition of the humeral head and the glenoid is missing. Often, attempts are made to “relocate― the humeral head, which, of course, are both fruitless (because it is not dislocated) and painful. A fracture is invariably present, and if not seen on the initial films, it should be sought with additional views. The P.1171 transscapular or the axial view is the key to making the diagnosis of a pseudodislocation. If necessary, the joint can be aspirated to confirm the presence of a bloody effusion and to show the normal position of the humeral head when fluid has been removed from the joint.
3050
FIGURE 46.4. Myositis Ossificans. A. A plain film of the humerus in this 30-year-old man shows a calcific mass adjacent to the diaphysis of the humerus. The calcification is not clearly peripheral in nature, although the central portion is less wellmineralized. B . An axial T2WI through the mass shows only a high-signal mass without evidence of calcification. C . A CT scan
3051
through the mass demonstrates the typical peripheral calcification that is virtually pathognomonic for myositis ossificans.
NORMAL
VARIANTS
Dorsal Defect of the Patella A normal variant that has been described in the patella that can be mistaken for a pathologic process is a lytic defect in the upper outer quadrant called a dorsal defect of the patella (Fig. 46.15) (1 0). It can mimic a focus of infection, osteochondritis dissecans, or a chondroblastoma. It is a normal developmental anomaly, however, and because of its characteristic location, it should not undergo biopsy. On MR, it will have an appearance similar to that of many other bony lesions—that is, low signal on T1WIs and high signal on T2WIs (Fig. 46.16) . Pseudocyst of the humerus is another entity that is often mistaken for a lytic pathologic lesion (Fig. 46.17). This is merely an anatomic variant caused by increased P.1172 P.1173 cancellous bone in the region of the greater tuberosity of the humerus, giving this region a more lucent appearance on radiographs (1 1,1 2). With hyperemia and disuse caused by rotator cuff problems or any other shoulder disorder, this area of lucency may appear strikingly more lucent and mimic a lytic lesion. Many of these have mistakenly undergone biopsy, and several have even had repeat biopsies after the initial pathology report stated “normal bone, no lesion in specimen.― Because of the associated hyperemia from the shoulder disorder (be it rotator cuff injury or another condition), a bone scan can show increased radionuclide uptake and thus sway the surgeon to perform a biopsy of this normal variant. It is radiographically characteristic in its location and appearance and should not undergo biopsy. Although other lesions, such as a
3052
chondroblastoma, an infection, or even a metastatic focus, could occur in a similar location, they do not have quite the same appearance as a pseudocyst of the humerus.
FIGURE
46.5. Avulsion
Injury. Cortical irregularity (arrows) at
the ischial tuberosity in this patient with pain over this region raises the question of possible tumor. This is a classic appearance, however, for an avulsion injury from this region, and a biopsy should be avoided.
3053
FIGURE 46.6. Avulsion Injury. Cortical irregularity with a Codman triangle of periostitis is seen along the ischial tuberosity, which was at first believed to represent a malignancy. However, because of the characteristic location, an avulsion injury was considered and the lesion was observed. It healed without sequelae.
3054
FIGURE
46.7. Cortical
Desmoid. A focal cortical irregularity in
this patient is seen in the posterior aspect of the femur (arrow) , with adjacent periostitis noted. Although a tumor such as an early parosteal osteosarcoma could perhaps have this appearance, this is a characteristic location and appearance for a cortical desmoid and should not undergo biopsy. Pain will disappear with rest.
3055
FIGURE
46.8. Cortical
Desmoid. A well-defined cortical defect
is seen in the posterior distal femur (arrow), which is a common appearance for a fairly well-healed cortical desmoid.
3056
FIGURE 46.9. Cortical Desmoid. Anteroposterior film of the knee (A) in a child shows a faint lytic lesion (arrows) in the medial aspect of the distal femur. Axial T1WI (B) and T2WI (C) through the lesion show a cortically based process (arrows) in the medial supracondylar ridge, which is characteristic for a cortical desmoid.
3057
Os odontoideum is a normal variant of the cervical spine that may, in fact, be posttraumatic (1 3). It is an unfused dens that may move anterior to the C2 body with flexion and can mimic a fractured dens (Fig. 46.18). Many of these require surgical fixation; some surgeons fuse every case, believing that they are all unstable. Radiologists should recognize that this process is not acute, thus, saving the patient from halo fixation and possible immediate surgical intervention. Most of these cases are seen after trauma, and if no neurologic deficits are present, these patients can be seen electively and spared the morbidity P.1174 P.1175 P.1176 P.1177 P.1178 P.1179 associated with treatment of the acutely fractured cervical spine. The radiologic signs of an os odontoideum are the smooth, often wellcorticated, inferior border of the dens and the hypertrophied, densely corticated anterior arch of C1 (1 4). This latter finding presumably represents compensatory hypertrophy and indicates a long-standing condition.
3058
FIGURE
46.10. Geode. A large cystic lesion was found in the
shoulder in this middle-aged weightlifter, and the possibility of a metastatic process was considered. Because the humeral head has sclerosis and osteophytosis as well as a loose body in the joint (arrow), degenerative disease of the shoulder was diagnosed; this makes the cystic lesion almost certainly a geode or subchondral cyst.
3059
FIGURE
46.11. Geode. A. A cystic lesion was noted in the
femoral head (solid arrows) of a young man with a painful hip. B . A CT scan through this area shows the subarticular nature and adjacent sclerosis. The differential diagnosis of infection, eosinophilic granuloma, and chondroblastoma was given. A ring of osteophytes (open arrowheads) was noted in retrospect on the plain film (A) in the subcapital region, which indicates degenerative disease of the hip. Degenerative joint disease is extremely unusual in a 20-year-old healthy man; however, it makes the lytic lesion in the femoral head almost certainly a subchondral cyst or geode. This was an active soccer player who had been playing with pain in his hip for several years following an injury that had caused the degenerative disease. Unfortunately, a biopsy was performed anyway and a subchondral cyst or geode was confirmed.
3060
FIGURE
46.12. Discogenic
Vertebral
Sclerosis. This patient
has sclerosis on the inferior portion of the L4 vertebral body associated with minimal osteophytosis and joint space narrowing at the adjacent disk space. This is the classic appearance for discogenic vertebral sclerosis, and a biopsy to rule out metastatic disease should not be performed.
3061
FIGURE
46.13. Fracture
Mimicking
Osteosarcoma.
A. This
16-year-old patient had experienced pain around the knee for 2 weeks before these radiographs. The knee films showed diffuse sclerosis and extensive periostitis about the distal femur, which is thought to be characteristic for an osteogenic sarcoma. The periosteal reaction, however, was thought to be much too thick, dense, and wavy to represent a malignant type of periostitis. B . A small offset of the epiphysis can be seen (arrow), which indicates an epiphyseal slippage consistent with a Salter epiphyseal fracture. This teenager had fallen off a bicycle and fractured the femur, yet continued to be active. The lack of immobility caused exuberant periostitis or callus with a large amount of reactive sclerosis, all of which mimicked an osteogenic sarcoma.
3062
FIGURE
46.14. Pseudodislocation of the Shoulder. A. This
patient experienced trauma to the shoulder, with resultant pain and immobility, and was thought to have a dislocation of the shoulder after the anteroposterior film was seen. The humeral head is inferiorly placed in relation to the glenoid; however, this is not the characteristic location of an anterior or posterior dislocation. B . The transscapular view shows that the humeral head is situated normally over the glenoid, without anterior or posterior dislocation. These findings are characteristic for a pseudodislocation caused by hemarthrosis, or blood in the joint, which allows the shoulder to be subluxed rather than dislocated. When a pseudodislocation is seen, as in this example, search for an occult fracture should ensue. In this case, as seen on (A), a fracture (arrowhead) was initially missed.
3063
FIGURE
46.15. Dorsal Defect of the Patella. A lytic defect in
the upper outer quadrant of the patella was seen in this patient on the anteroposterior film (A) and the axial or sunrise view (B) (arrows), which is characteristic for a normal variant called dorsal defect of the patella. It occurs only in the upper outer quadrant and should be asymptomatic.
3064
FIGURE 46.16. Dorsal Defect of the Patella. A. Axial T1WI shows a focal area of low signal in the patella in a subarticular location in the lateral facet of the patella. B . The axial T2WI shows high signal in the lesion. This is typical in location and appearance for a dorsal defect of the patella.
3065
FIGURE
46.17. Pseudocyst of the Humerus. A well-defined
lytic process is seen in the greater tuberosity, which was thought to represent a lytic lesion. This patient was symptomatic and had increased radionuclide uptake on an isotope bone scan. However, this is a characteristic location and appearance for a pseudocyst of the humerus, which merely represents decreased cortical bone in this region. This becomes more pronounced when pain in the shoulder is present and hyperemia or disuse osteoporosis occurs.
3066
FIGURE
46.18. Os
Odontoideum. Flexion (A) and extension
(B) views show the anterior arch (A) of the C1 vertebrae has moved markedly anterior in relation to the body of C2 in flexion. The odontoid or dens is difficult to see but appears to be separated from the body of C2. Because of the smooth borders of the separated dens and because of the cortical hypertrophy of the anterior arch of C1, this can safely be called an os odontoideum, which is a congenital or long-standing posttraumatic abnormality rather than an acute fracture. Obviously, patients with this condition should have no neurologic problems, yet in many instances are still believed to be unstable and undergo surgically fusion. This, however, can be done on an elective
basis.
3067
FIGURE 46.19. Nonossifying Fibroma. A well-defined, slightly expansile, lytic lesion is seen in the fibula (lower curved arrow); this is characteristic for a nonossifying fibroma. A second lytic lesion is seen in the posterior distal femur (upper curved arrow) , which is also typical in appearance for a nonossifying fibroma.
OBVIOUSLY
BENIGN
LESIONS
3068
Multiple real lesions exist that should be recognized radiographically as benign and left alone. These are lesions that should be diagnosed by the radiologist, not the pathologist. Listing a differential in these cases often spurs the surgeon to a biopsy, when, in fact, no biopsy should be necessary. Nonossifying fibroma (NOF) is perhaps the most often encountered lesion in this category. NOF is identical to a fibrous cortical defect, but the term is usually reserved for defects larger than 2 cm. They are, classically, lytic lesions located in the cortex of the metaphysis of a long bone and have a well-defined, often sclerotic, scalloped border with slight cortical expansion (Fig. 46.19). They are found almost exclusively in patients younger than the age of 30 years; hence, the natural history of the lesion is involution. As they involute, they fill in with new bone, giving it a sclerotic appearance (Fig. 46.20) thus, they can have some increased radionuclide activity on bone scans. They are most often mistaken for an area of infection, eosinophilic granuloma, fibrous dysplasia, or aneurysmal bone cyst. They are asymptomatic and have never been reported to P.1180 be associated with malignant degeneration. On occasion, a pathologic fracture can occur through these lesions, but most surgeons do not advocate prophylactic curettage to prevent fracture, as with unicameral bone cysts. NOFs can be quite large but invariably have a benign appearance (Fig. 46.21), and biopsy should be avoided. Their asymptomatic nature should help differentiate them from most of the other lesions in the differential diagnosis and thereby preclude even giving a differential diagnosis. On occasion, they are found to be multiple, yet each lesion is so characteristic as to be easily diagnosed.
3069
FIGURE
46.20. Healing
Nonossifying
Fibroma. A minimally
sclerotic process is seen in the proximal tibia (arrows), which was thought by the surgeons to represent a focus of infection or an osteoid osteoma, even though the patient was asymptomatic. This is a characteristic appearance for a disappearing or healing nonossifying fibroma and should not undergo biopsy.
3070
FIGURE
46.21. Nonossifying
Fibroma.
Anteroposterior (A)
and lateral (B) films of the tibia show a large, well-defined, minimally expansile lytic lesion of the proximal tibia, which is characteristic for a nonossifying fibroma. Even though the patient was asymptomatic, biopsy was performed and the diagnosis confirmed.
3071
FIGURE 46.22. Giant Bone Island. A large sclerotic focus is seen in the right iliac wing (arrow). Note how the lesion is somewhat spherical or oblong and in the lines of trabecular stress, which is characteristic for a bone island. This patient was asymptomatic and had no evidence of a primary carcinoma.
Bone
islands are not a radiographic dilemma when they are 1 cm or
smaller. Occasionally, however, they grow to golf ball size or larger and mimic sclerotic metastases (Fig. 46.22). They are always asymptomatic. Radiographically, two signs can be found to help distinguish giant bone islands from metastases. First, bone islands usually are oblong, with their long axis in the axis of stress on the bone: for example, in a long bone they align themselves along the axis of the diaphysis. Second, the margins of a bone island, when examined closely, will show bony trabeculae extending from the lesion into the normal bone in a spiculated fashion (1 5). This is characteristic of a bone island and helpful in differentiating it from a more
aggressive
process.
3072
Unicameral bone cysts are often prophylactically curettaged and packed so as to prevent fracture with subsequent deformity. When these cysts occur in the calcaneus, however, they should be left alone. They always occur in the anteroinferior portion of the calcaneus (Fig. 46.23), an P.1181 area that does not receive undue stress. In fact, a pseudotumor of the calcaneus is seen in the identical position because of the absence of stress and resulting atrophy of bony trabeculae (Fig. 46.24) . These lesions are asymptomatic, only rarely fracture, and should not suffer the same fate as their counterparts in long bones—that is, surgical
removal.
FIGURE
46.23. Unicameral Bone Cyst. A well-defined lytic
lesion on the anteroinferior portion of the calcaneus, as in example, is virtually pathognomonic for a unicameral bone or simple bone cyst. Because this is an area of diminished it is thought to be unnecessary to curettage and pack this
this cyst stress, lesion
prophylactically in an effort to avoid a pathologic bone fracture,
3073
which is often done in the femur and humerus with unicameral bone cysts.
FIGURE
46.24. Pseudocyst of the Calcaneus. An area of
radiolucency is seen on the anteroinferior portion of the calcaneus (arrows) similar to the example in Fig. 46.23, but it is not as well defined. This is a pseudocyst similar to the pseudocyst of the humerus that results from diminished stress through this region.
3074
FIGURE 46.25. Early Bone Infarct. Patchy demineralization is seen in the distal femur and proximal tibia in this patient with systemic lupus erythematosus. The opposite leg was similarly involved. This is characteristic for early bone infarcts and should not be confused with infection or metastatic disease.
3075
FIGURE
46.26. Bone Infarct. A. A plain film of the knee shows
a permeative pattern in the proximal tibia, which was thought at first to be infection or a primary tumor. B . Coronal T1WI shows the characteristic serpiginous border seen with bone infarct in the tibia and in the femur. MR can on occasion better characterize the ill-defined early bone infarct, as in this example. This patient has systemic lupus erythematosus.
Bone
Infarction
Early in the course of its development, a bone infarct can have a patchy or mixed lytic-sclerotic pattern or even resemble a permeative process (Fig. 46.25) (1 6). In a patient with bone pain and a permeative bone lesion, many aggressive disorders head the differential list and a biopsy soon ensues. If this process can be noted to be multiple and in the diametaphyseal region of a long bone, especially if the patient has an underlying disorder such as
3076
sickle cell anemia or systemic lupus erythematosus, areas of early bone infarction should be considered. In some cases, the characteristic MR appearance of an infarct may save a patient from biopsy when the plain films are equivocal (Fig. 46.26) . These are but a few of the many examples in skeletal radiology in which the well-trained radiologist can be of invaluable assistance to the clinician and the patient by helping avert a needless biopsy. Dozens of other examples are nicely shown in normal variant textbooks, which are widely available. Because of the potential harm in performing a needless biopsy, the examples described in this chapter are stressed. When these lesions are encountered by the radiologist, a differential diagnosis should not be offered, as it will often lead the surgeon to a biopsy in an attempt to get a diagnosis. A biopsy in many of these entities is not only unnecessary but can be misleading. P.1182
REFERENCES 1. Murray R, Jacobson H. The Radiology of Skeletal Disorders. 2nd ed. New York: Churchill Livingstone, 1977:603. 2. Wootton J, Cross M, Holt K. Avulsion of the ischial apophysis. J Bone Joint Surg 1990;72B:625–627. 3. Schneider R, Kaye J, Ghelman B. Adductor avulsive injuries near
the
symphysis
pubis.
Radiology
1976;120:567–569.
4. Barnes G, Gwinn J. Distal irregularities of the femur simulating malignancy. AJR Am J Roentgenol 1974;122:180–185. 5. Ostlere S, Seeger L, Eckardt J. Subchondral cysts of the tibia secondary to osteoarthritis of the knee. Skeletal Radiol 1990;19: 287–289.
3077
6. Resnick D, Niwayama G, Coutts R. Subchondral cysts (geodes) in arthritic disorders: pathologic and radiographic appearance of the hip joint. AJR Am J Roentgenol 1977;128:799–806. 7. Martel W, Seeger J, Wicks J, et al. Traumatic lesions of the discovertebral junction in the lumbar spine. AJR Am J Roentgenol 1976;127:457–464. 8. Lipson S. Discogenic vertebral sclerosis with calcified disc. New Engl J Med 1991;325:794–799. 9. Helms C, Richmond B, Sims R. Pseudodislocation of the shoulder: a sign of an occult fracture. Emerg Med 1986;18:237–241. 10. Johnson JF, Brogdon BG. Dorsal effect of the patella: incidence and distribution. AJR Am J Roentgenol 1982;139:339–340. 11. Helms C. Pseudocyst of the humerus. AJR Am J Roentgenol 1979;131:287–292. 12. Resnick D, Cone R. The nature of humeral pseudocysts. Radiology 1984;150:27–28. 13. Minderhoud J, Braakman R, Penning L. Os odontoideum: clinical, radiological, and therapeutic aspects. J Neurol Sci 1969;8:521–544. 14. Holt RG, Helms CA, Munk PL, et al. Hypertrophy of C-1 anterior arch: useful sign to distinguish os odontoideum from acute
dens
fracture.
Radiology
1989;173:207–209.
15. Onitsuka H. Roentgenologic aspects of bone islands. Radiology
3078
1977;124:607–612. 16. Munk PL, Helms CA, Holt RG. Immature bone infarcts: findings on plain radiographs and MR scans. AJR Am J Roentgenol 1989;152:547–549.
3079
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section X - Musculoskeletal Radiology > Chapter 47 - Miscellaneous Bone Lesions
Chapter
47
Miscellaneous
Bone
Lesions
Clyde A. Helms There are a host of bony conditions, diseases, and syndromes that do not fit conveniently into any of the preceding chapters, yet should be given some mention in an attempted overview of musculoskeletal radiology. These are listed alphabetically for lack of a more scientific basis.
Achondroplasia The most common cause of dwarfism is achondroplasia, a congenital, hereditary disease of failure of endochondral bone formation. The femurs and humeri are more profoundly affected than the other long bones, although the entire skeleton is abnormal. A characteristic finding is that the spine typically has narrowing of the interpedicular distances in a caudal direction (Fig. 47.1), the opposite of normal, in which the interpedicular distances get progressively wider as one proceeds down the spine. The long bones are short but have normal width, giving them a thick appearance.
Avascular The terms avascular
Necrosis
(Osteonecrosis)
necrosis (AVN) and osteonecrosis refer to a lack
of blood supply with subsequent bone death and ensuing bony collapse in an articular surface. The etiology of AVN is an extensive differential that most commonly includes trauma, steroids, aspirin,
3080
collagen vascular diseases, alcoholism, and idiopathic causes (Table 47.1) (1). The radiographic appearance ranges from patchy sclerosis (Fig. 47.2A) to articular surface collapse and fragmentation (Fig. 47.3). Just before collapse, a subchondral lucency is occasionally seen (Fig. 47.4); however, this is a late and inconstant sign of AVN. MR is extremely valuable in demonstrating the presence and extent of AVN (Fig. 47.2B), even when plain films are apparently normal. MR is currently considered the most efficacious means to evaluate a joint for AVN (2). It is useful not only in AVN of the hip but also in the knee, wrist, foot, and ankle.
Hypertrophic Pulmonary Osteoarthropathy Hypertrophic pulmonary osteoarthropathy is manifested by clubbing of the fingers and periostitis, usually in the extremities (Fig. 47.5) , which might or might not be associated with bone pain. It is most commonly seen in patients with lung cancer, but many other etiologies have been reported, including bronchiectasis, GI disorders, and liver disease. The actual mechanism of formation of periostitis secondary to a distant malignancy or other process is unknown. The differential diagnosis for periostitis in a long bone without an underlying bony abnormality would include hypertrophic osteoarthropathy, venous stasis, thyroid acropachy, pachydermoperiostosis, and trauma (Table 47.2) .
pulmonary
Melorheostosis Melorheostosis is a rare, idiopathic disorder characterized by thickened cortical new bone that accumulates near the ends of long bones, usually only on one side of the bone, P.1184 and has an appearance likened to “dripping candle wax― (Fig. 47.6). It can affect several adjacent bones and can be symptomatic.
3081
FIGURE
47.1. Achondroplasia. An anteroposterior plain film of
the spine in this patient with achondroplasia demonstrates narrowing of the interpedicular distance (arrows) in a caudal direction, which is characteristic of this disorder. Ordinarily, the interpedicular distance widens in each vertebra in a caudal direction.
3082
FIGURE
47.2. Avascular Necrosis (AVN). A. A plain film of
the hip in this patient with AVN shows faint, patchy sclerosis throughout the femoral head. This is a relatively early plain film finding for AVN. B . Coronal T1WI shows typical findings in AVN. Diffuse low signal in the left hip is noted, which has more extensive involvement than the right. The right hip has a low signal serpiginous rim which is characteristic for AVN.
TABLE 47.1 Common Causes of Avascular Necrosis
Trauma Steroids Collagen vascular Alcoholism Idiopathic
diseases
causes
3083
Mucopolysaccharidoses Hurler,
and
Hunter
(Morquio, Syndromes)
The mucopolysaccharidoses are a group of inherited diseases characterized by abnormal storage and excretion in the urine of various mucopolysaccharides, such as keratan sulfate (Morquio) and heparan sulfate (Hurler). These patients have short stature, primarily from shortened spines, and characteristic plain film findings. In the spine, patients with Morquio have platyspondyly (generalized flattening of the vertebral bodies) with a central anterior projection or “beak― off the vertebral body, as viewed on a lateral plain film (Fig. 47.7). Hurler and Hunter syndromes show P.1185 platyspondyly, with a beak that is anteroinferiorly positioned (Fig. 47.8). The pelvis in these disorders is similar in appearance to that seen in patients with achondroplasia, with wide, flared iliac wings and broad femoral necks. A characteristic finding in the hands is a pointed proximal fifth metacarpal base that has a notched appearance to the ulnar aspect (Fig. 47.9) .
3084
FIGURE
47.3. Avascular
Necrosis
(AVN). An anteroposterior
plain film of the shoulder reveals articular surface collapse in this patient who was treated with steroids for systemic lupus erythematosus. This is an advanced stage of AVN.
3085
FIGURE
47.4. Avascular
Necrosis
(AVN). An anteroposterior
frog-leg lateral view of the hip in this patient with sickle cell disease shows a subchondral lucency (arrows) and patchy sclerosis in the femoral head, indicative of AVN. This is a relatively advanced stage of AVN. The subchondral lucency is often better demonstrated with the frog-leg lateral view.
3086
TABLE 47.2 Differential Diagnosis for Periostitis Without Underlying
Trauma Hypertrophic pulmonary Venous stasis
Bony
Lesions
osteoarthropathy
Thyroid acropachy Pachydermoperiostosis
3087
FIGURE
47.5. Hypertrophic
Pulmonary
Osteoarthrosis.
Periostitis can be seen along the shafts of the distal tibia and fibula (arrows) in this patient with bronchogenic carcinoma and leg pain. This is characteristic for hypertrophic pulmonary osteoarthrosis.
3088
FIGURE 47.6. Melorheostosis. Dense, wavy, new bone is seen adjacent to the lateral tibial cortex, which has a dripping candle wax appearance, which is classic for melorheostosis. A similar pattern can be seen in the medial aspect of the distal femur.
P.1186
Multiple
Hereditary
Exostosis 3089
Also known as diaphyseal
aclasia, this is a not uncommon hereditary
disorder that seems to affect multiple members of a family with multiple osteochondromas, or exostoses. An osteochondroma is a cartilage-capped bone outgrowth that may be pedunculated or sessile in appearance. In the multiple hereditary form, the knees are virtually always involved (Fig. 47.10). Undertubulation (a widened diameter of the bone) is invariably present at the site of the exostosis. The incidence of malignant degeneration in this population has been reported to be as high as 20%, but this is a gross overestimation; malignant degeneration is in fact extremely rare. As with solitary osteochondromas, the more axially situated lesions are more prone to undergo malignant degeneration, whereas the more peripheral lesions are less likely to do so. The proximal femurs are P.1187 frequently involved and have a characteristic appearance (Fig. 47.11) .
3090
FIGURE
47.7. Morquio
Syndrome. A lateral plain film of the
spine reveals a central beak or anterior bony projection off the vertebral bodies in this patient with Morquio syndrome.
3091
FIGURE
47.8. Hurler
Syndrome. A lateral plain film of the
spine in this patient with Hurler syndrome shows an inferiorly placed bony projection extending anteriorly off the vertebral bodies (arrow) .
3092
FIGURE 47.9. Hurler Syndrome. An anteroposterior plain film of the hand in this patient with Hurler syndrome shows a notch (arrow) at the base of the fifth metacarpal, which is a characteristic finding in all of the mucopolysaccharidoses.
Osteoid
Osteoma
The etiology of osteoid osteoma is unknown. It is a painful lesion that occurs almost exclusively in patients younger than 30 years of age and is treated successfully with surgical excision or thermal ablation. Radiographically, an osteoid osteoma is said to have a classic appearance, but it has many different appearances, which can make diagnosis difficult (3). The classically described radiographic appearance is a cortically based sclerotic lesion in a long bone that has a small lucency within it called the nidus (Fig. 47.12A). It is the nidus that causes the pain and the surrounding reactive sclerosis. If
3093
the nidus is surgically removed or thermally ablated, P.1188 complete cessation of pain is the rule. CT is often very helpful in demonstrating the exact location of the nidus (Fig. 47.12B) .
FIGURE 47.10. Multiple Hereditary Exostosis. The knees are involved in virtually every case of multiple hereditary exostosis. They typically show not only multiple exostoses (arrows) but marked undertubulation in the metaphyses.
3094
FIGURE
47.11. Multiple
Hereditary
Exostosis. The femoral
necks are often involved in multiple hereditary exostosis. They will show undertubulation, as in this example, and usually have one or more exostoses (arrows) .
3095
FIGURE
47.12. Osteoid Osteoma. A. An anteroposterior plain
film of the femur in a child with hip pain shows an area of sclerosis medially near the lesser trochanter with a small lucency (arrow), which is the nidus of an osteoid osteoma. Osteomyelitis could have this identical appearance. B . CT scan of the femur shows the sclerosis medially and the lucent nidus (arrow) to better advantage. The CT scan gives the surgeon a more precise anatomic location of the nidus than the plain film.
If the nidus of an osteoid osteoma is located in the medullary rather than the cortical portion of a bone, or if it is located in a joint, there is much less reactive sclerosis present. This gives the lesion a different overall appearance than the more common cortical lesion, in that it does not appear as sclerotic. Up to 80% of osteoid osteomas are located intracortically, with the remainder located in the intramedullary part of a bone. Rarely, an osteoid osteoma will
3096
present in the periosteum, causing exuberant periostitis. The nidus itself is usually lucent but often develops some calcification within it. It then has the appearance of a sequestrum, as is seen in osteomyelitis. If the nidus calcifies completely, it blends in with the surrounding sclerosis and cannot be seen on most radiographs. Therefore, the diagnosis of an osteoid osteoma should not depend on visualization of a nidus. Because an osteoid osteoma resembles osteomyelitis, regardless of the appearance of the nidus, it can be difficult to differentiate the two radiographically. It cannot be reliably done with plain films, CT, or MR. However, because the nidus is extremely vascular, it avidly accumulates radiopharmaceutical bone-scanning agents. An osteoid osteoma will have an area of increased uptake corresponding to the area of reactive sclerosis and will additionally demonstrate a second area of increased uptake corresponding to the nidus (Fig. 47.13) . This has been termed the double-density sign (4). In contrast, osteomyelitis has a photopenic area corresponding to the plain film lucency that represents an avascular focus of purulent material. The natural history of an osteoid osteoma is presumed to be spontaneous regression, as they are rarely seen in patients over the age of 30.
Osteopathia
Striata
Also known as Voorhoeve disease, this disorder is manifested by multiple 2- to 3-mm-thick linear bands of sclerotic bone aligned parallel to the long axis of a bone (Fig. 47.14). It usually affects multiple long bones and is asymptomatic; hence, it is usually an incidental finding.
3097
FIGURE 47.13. Osteoid Osteoma. A. A lateral plain film of the tibia in this child with leg pain shows cortical thickening in the posterior diaphysis. No lucency in the sclerotic area could be
3098
identified. B . A radionuclide bone scan reveals uptake corresponding to the area of sclerosis in the tibia, with a more marked area of uptake centrally (arrow), which is the doubledensity sign of an osteoid osteoma. C . The surgical specimen shows the nidus (arrow) as a faint lucency within the sclerotic bone.
FIGURE 47.14. Osteopathia Striata. Multiple dense linear streaks are seen in the distal femur, which are characteristic of osteopathia striata.
P.1189 P.1190
Osteopoikilosis Osteopoikilosis is an hereditary, asymptomatic disorder that is
3099
usually an incidental finding of multiple small (3 to 10 mm) sclerotic bony densities affecting primarily the ends of long bones and the pelvis (Fig. 47.15). It has no clinical significance other than that it can be confused for diffuse osteoblastic metastases.
FIGURE
47.15. Osteopoikilosis. An anteroposterior view of the
pelvis reveals multiple small, round sclerotic foci throughout the pelvis and femurs. This is diagnostic of osteopoikilosis. This is occasionally mistaken for metastatic disease.
Pachydermoperiostosis Pachydermoperiostosis is a rare familial disease that is manifested by thickening of the skin of the extremities and face, clubbing of the fingers, and widespread periostitis. It seems to be more common in black patients. The periosteal reaction is similar to that of hypertrophic pulmonary osteoarthropathy, but pachydermoperiostosis is only occasionally painful.
3100
FIGURE 47.16. Sarcoid. An anteroposterior plain film of the hands in a patient with sarcoidosis shows multiple lytic lesions, many of which demonstrate a lacelike pattern.
3101
FIGURE
47.17. Idiopathic Transient Osteoporosis of the
Hip. A. A plain film of a 40-year-old man with left hip pain shows osteoporosis involving the left hip, with no other abnormalities seen. B . A coronal T1WI done at the same time as the plain film shows low signal in the superior portion of the left femoral head. This is a characteristic appearance for avascular necrosis (AVN) but is a nonspecific finding. Clinically, this patient had no underlying causes for AVN, and he was treated conservatively. C . Seven months later, after near total cessation of the hip pain, a repeat T1WI shows no abnormality in the hip. This is consistent with idiopathic transient osteoporosis of the hip.
P.1191
Sarcoidosis
3102
Sarcoidosis is a noncaseating granulomatous disease that primarily affects the lungs. When the musculoskeletal system is involved, the hands are most often affected, with the spine and long bones only infrequently involved. Sarcoid causes a characteristic lacelike pattern of bony destruction in the hands (Fig. 47.16). Multiple phalanges are typically affected in either one or both hands. It is so radiographically characteristic that there is almost no differential diagnosis for this pattern.
Transient
Osteoporosis
of
the
Hip
This poorly understood disorder is an idiopathic process that begins with a painful hip with no underlying disorder or other findings other than osteoporosis, which is limited to the painful hip. Some believe it is early AVN; however, this has not been proved. Its appearance on MR is similar to that of early AVN (5) in that low signal on T1WIs is seen throughout the femoral head and neck (Fig. 47.17). Transient osteoporosis of the hip invariably is self-limited with full resolution. It tends to occur more often in men.
REFERENCES 1. Mankin H. Nontraumatic necrosis of bone (osteonecrosis). N Engl
J
Med
1992;326:1473–1479.
2. Mitchell D, Kressel H, Arger P, et al. Avascular necrosis of the femoral head: morphologic assessment by MR imaging, with CT correlation.
Radiology
1986;161:739–742.
3. Marcove R, Heelan R, Huvos A, et al. Osteoid osteoma. Diagnosis, localization, and treatment. Clin Orthop 1991;267:197–201. 4. Helms CA, Hattner RS, Vogler JB III. Osteoid osteoma: radionuclide diagnosis. Radiology 1984;151:779–784.
3103
5. Takatori Y, Kokubo T, Ninomiya S, et al. Transient osteoporosis of the hip. Magnetic resonance imaging. Clin Orthop 1991;271: 190–194.
3104
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section X - Musculoskeletal Radiology > Chapter 48 - Magnetic Resonance Imaging of the Knee
Chapter
48
Magnetic Resonance the Knee
Imaging
of
Clyde A. Helms MR of the knee has developed into one of the most frequently requested examinations in radiology. This is because of its inherent accuracy in depicting internal derangements and its ability to allow orthopaedic surgeons to use it as a road map for subsequent therapeutic arthroscopic procedures. Also, MR has a very high negative predictive value; therefore, a normal MR knee examination is highly accurate in excluding an internal derangement (1,2) .
TECHNIQUE The proper imaging protocol is essential for a high diagnostic accuracy rate. If the appropriate sequences are obtained, an accuracy of 90% to 95% can be expected. A sagittal T1W (or protondensity) sequence is essential for examining the menisci, and 4- or 5-mm-thick slices with a relatively small field of view and at least a 256×192 matrix are recommended. The knee should be imaged using a dedicated knee coil and externally rotated about 5° to 10° (should not exceed 10°) to put the anterior cruciate ligament in the plane of imaging. T2 fast spin-echo (FSE, also called turbo spin echo) or T2* GRASS (gradient-recalled acquisition in the steady state) sagittal images are obtained primarily to examine the cruciate ligaments and cartilage.
3105
FSE sequences are particularly poor for examining the menisci. Even when performed as fast proton-density images with a short echo train length, they have too much blurring to provide an accurate demonstration of meniscal tears. Conventional spin-echo images have consistently given a sensitivity for meniscal tears in the 90% to 95% range, whereas FSE proton-density sequences have been reported in multiple papers to be only around 80% sensitive for meniscal tears. Coronal images are obtained to examine the collateral ligaments and cartilage and to look for meniscocapsular separations. These abnormalities can generally only be seen with T2WIs. Coronal T1WIs are therefore a waste of time, because nothing can be seen on these images that cannot be seen equally as well on the sagittal images or the T2 or T2* coronal images. The coronal images are rarely useful for seeing a meniscal tear that cannot be appreciated on the sagittal images. Axial images are used for viewing the patellofemoral cartilage, identifying bursal fluid collections, and examining a medial patellar plica. As for the coronal images, to afford an opportunity to see any pathology, T2WIs must be obtained.
MENISCI The normal meniscus is a fibrocartilaginous, C-shaped structure that is uniformly low in signal on both T1WIs and T2WIs. Many centers have found that the menisci are more easily examined if they fatsuppress the T1 or proton-density sequences (Fig. 48.1). With T2* sequences, the menisci will usually demonstrate some internal signal. With T1WIs, any signal within the meniscus is abnormal, P.1193 except in children, in whom some signal is normal and represents normal
vascularity.
3106
FIGURE
48.1. Normal Meniscus. A. A sagittal T1WI (time of
repetition [TR] 600; time of echo [TE] 30) through a normal lateral meniscus demonstrates uniform low signal in the meniscus. This is a section through the body of the meniscus, as it has a bow-tie configuration. With 4- or 5-mm-thick slices, two sections of the body should be seen in each meniscus. B . In the same T1W sequence, this sagittal image demonstrates uniform low signal in the anterior and posterior horns of this normal lateral meniscus. C . This sagittal proton-density image (TR 2,000; TE 20) shows how fat suppression accentuates the menisci.
Meniscal
Degeneration 3107
Meniscal signal that does not disrupt an articular surface is representative of intrasubstance degeneration (Fig. 48.2), which is myxoid degeneration of the fibrocartilage. It most likely represents aging and normal wear and tear. It is not thought to be symptomatic and cannot be diagnosed clinically or with arthroscopy. Some choose, therefore, not to mention intrasubstance degeneration in the radiology interpretation. A grading scale for meniscal signal that is widely used is the following (Fig. 48.3): grade 1, rounded or amorphous signal that does not disrupt an articular surface; grade 2, linear signal that does not disrupt an articular surface; and grade 3, rounded or linear signal that disrupts an articular surface. Grades 1 and 2 are intrasubstance degeneration and should not be reported as “grade 1 or 2 tears,― since the term “tear― can lead to unnecessary arthroscopy (arthroscopy is not indicated for intrasubstance degeneration). Grade 3 is a meniscal tear.
Meniscal
Tear
When high signal in a meniscus disrupts the superior or inferior articular surface, a meniscal tear is diagnosed (Fig. 48.4). Meniscal tears have many different configurations and locations; an oblique tear extending to the inferior surface of the posterior horn of the medial meniscus is the most common type. In a small but significant percentage of cases, it can be virtually impossible to be certain whether meniscal high signal disrupts an articular surface. In these cases it is recommended that the surgeon be advised that it is too close to call. The surgeon can then rely on his or her clinical expertise to decide if arthroscopy is warranted, and if it is, the MR will guide the surgeon to the location of the questionable tear. If these equivocal cases are excluded, the remaining cases will have an extremely high accuracy rate. It has been shown that MR imaging sensitivity for meniscal tears decreases significantly when the anterior cruciate ligament (ACL) is torn (3). These frequently overlooked tears occur in the periphery of the meniscus and in the posterior horn of the lateral meniscus. Hence,
great
3108
P.1194 care must be used in examining these areas of the menisci in patients with ACL tears.
FIGURE
48.2. Intrasubstance
Degeneration. Faint
intermediate signal can be seen in the posterior horn of this meniscus (arrow) that does not disrupt the articular surface of the meniscus. This is intrasubstance degeneration.
Bucket-Handle
Tear
Another very common meniscal tear is a bucket-handle tear. This is a vertical longitudinal tear that can result in the inner free edge of the meniscus becoming displaced into the intercondylar notch (Fig. 48.5). It is most easily recognized by observing on the sagittal images that only one image is present that has the bow-tie appearance of the body segment of the meniscus (4) (Fig. 48.6) .
3109
Normally, two contiguous sagittal images with a bow-tie shape are seen, because the normal meniscus is 9 to 12 mm wide and the sagittal images are 4 to 5 mm in thickness. On the coronal images, a bucket-handle tear may reveal the meniscus to be shortened and truncated; however, the torn meniscus often remodels and truncation cannot be appreciated. The displaced inner edge of the meniscus (the “handle― of the bucket) is often seen in the intercondylar notch on sagittal or coronal views (Fig. 48.7) .
FIGURE 48.3. Grading Scale for Menisci. A schematic of the MR grading scale for meniscal abnormalities. Grade 1 is rounded or amorphous signal in the meniscus that does not disrupt an articular surface. Grade 2 is linear signal that does not disrupt an articular surface. Grades 1 and 2 represent intrasubstance degeneration. Grade 3 is signal that does disrupt an articular surface and indicates a meniscal tear.
3110
FIGURE
48.4. Meniscal
Tear. This T1W sagittal image (time of
repetition 600; time of echo 30) shows linear high signal in the posterior horn of the meniscus that disrupts the inferior articular surface. This is the appearance of a meniscal tear.
Discoid
Meniscus
A discoid meniscus is a large meniscus that can have many different shapes: lens-shaped, wedged, flat, and others. Whether it is congenital or acquired is not known, but most are found in children and P.1195 young adults. It is seen laterally in up to 3% of the population, with a discoid medial meniscus being much less common. A discoid meniscus is thought to be more prone to tear than a normal meniscus, and it can be symptomatic Although they are easily identified on extensions of meniscal tissue into the intercondylar notch (Fig. 48.8), they
even without being torn. coronal images by noting tibial spines at the are most reliably diagnosed by
noting more than two consecutive sagittal images that show the meniscus with a bow-tie appearance (Fig. 48.9) (5) .
3111
FIGURE 48.5. Bucket-Handle Tear. This drawing illustrates a bucket-handle tear, with the torn free edge of the meniscus displaced as the handle of the bucket.
FIGURE
48.6. Bucket-Handle
Tear. Sagittal T1WIs (time of
3112
repetition 600; time of echo 30) through the medial meniscus at its most medial aspect reveal one bow-tie, indicative of the body of the meniscus (A), with the adjacent image (B) showing apparently normal anterior and posterior horns. However, since there should be two consecutive sagittal images with a bow-tie configuration, this suggests a bucket-handle tear.
Meniscal
Cysts
Meniscal cysts occur in about 5% of cases and can cause pain even if the meniscus is not torn. The etiology is unknown, but they occur more frequently in discoid menisci. If the meniscus is not torn, the surgical approach used by some is percutaneous decompression and packing, whereas if a meniscal tear is associated with the cyst, it is approached intra-articularly. Hence, accurate diagnosis of a tear is imperative. The intrameniscal portion of the cyst typically does not get as bright as fluid in signal on T2 sequences (Fig. 48.10), which has misled many P.1196 radiologists into discounting the presence of a cyst. A meniscal cyst will enlarge the meniscus and give it a swollen appearance unless it decompresses into the soft tissues (called a parameniscal cyst) or into the joint via a meniscal tear. Decompression into a parameniscal cyst does not indicate a meniscal tear. A meniscal tear, by definition, has to disrupt the articular surface of the meniscus.
3113
FIGURE
48.7. Displaced Fragment in Bucket-Handle Tear. A
sagittal T1WI (time of repetition 600; time of echo 30) through the intercondylar notch in a patient with a bucket-handle tear reveals the displaced free fragment or handle (arrow) just anterior to the PCL.
3114
FIGURE
48.8. Discoid
Lateral
Meniscus. A coronal gradient-
recalled acquisition in the steady state image (time of repetition 500; time of echo 30; θ 30°) through the intercondylar notch shows a large lateral meniscus with meniscal tissue extending into the notch medially (arrow) .
3115
FIGURE 48.9. Discoid Lateral Meniscus. Three consecutive 5mm-thick T1WIs (time of repetition 600; time of echo 30) through the lateral meniscus, beginning with the most lateral (A) and extending medially (B, C), each show the meniscus to have a bow-tie configuration. Because only two images should have a bow-tie shape, indicative of the body of the meniscus, this is diagnostic of a discoid lateral meniscus. (Fig. 48.8 is an image of the same knee.)
Transverse
Ligament
The lateral meniscus often has what appears to be a tear on the anterior horn near its upper margin, which is a pseudotear from the insertion of the transverse ligament (Fig. 48.11). This can easily be
3116
differentiated from a real tear by following it medially across the knee in the Hoffa fat-pad, where it inserts into the anterior horn of the medial meniscus.
CRUCIATE Anterior
LIGAMENTS Cruciate
Ligament
The normal ACL is seen in the intercondylar notch as a linear, predominantly low-signal structure on T1WIs; it often shows some linear striations near its insertion onto the medial tibial spine when viewed on sagittal images (Fig. 48.12). When torn, the ACL is most often simply not visualized, although sometimes the actual disruption will be seen (Fig. 48.13). T2WIs are imperative for obtaining the highest accuracy in diagnosing ACL tears, because fluid and hemorrhage will often obscure the ligament on T1WIs. Partial tears or sprains of the ACL are manifested by high signal within an otherwise intact ligament. MR is highly accurate in diagnosing a torn ACL, with sensitivities reported in the literature approaching 100%.
Posterior
Cruciate
Ligament
The normal posterior cruciate ligament (PCL) is a gently curved, homogeneously low-signal structure (Fig. 48.14) that is infrequently torn and even less frequently repaired by surgeons. When torn, it takes on diffuse intermediate signal throughout (Fig. 48.15). This increased signal usually does not get brighter with T2WIs and is therefore often overlooked. Most orthopaedic surgeons do not even inspect the PCL at arthroscopy and do not repair it when torn, because it is rarely a cause of instability.
3117
FIGURE
48.10. Meniscal
Cyst. A sagittal proton-density
weighted image (A) through the medial meniscus shows a swollen anterior horn filled with increased signal (arrow). A T2WI (B) shows high signal similar to fluid in the parameniscal portion, whereas the intrameniscal signal is only intermediate.
P.1197
Meniscofemoral
Ligament
A low-signal, round structure is often seen just anterior or posterior to the PCL, as seen in the sagittal views. A loose body or a free fragment of a piece of torn meniscus can have this appearance (Fig. 48.16), but it is most commonly caused by a meniscofemoral ligament that extends obliquely across the knee from the medial femoral condyle to the posterior horn of the lateral meniscus. If it passes in front of the PCL, it is called the ligament of Humphry, and if it passes behind the PCL, it is called the ligament of Wrisberg (Fig. 48.17). Either one of these ligaments is present in up to 72% of all knees.
3118
FIGURE
48.11. Pseudotear From a Transverse Ligament. A
sagittal T1WI (time of repetition 600; time of echo 30) through the lateral meniscus shows linear high signal through the upper anterior horn (arrow), which resembles a tear. This is the insertion of the transverse ligament onto the meniscus.
3119
FIGURE
48.12. Normal Anterior Cruciate Ligament (ACL). A
sagittal T1WI (time of repetition 600; time of echo 30) through the intercondylar notch shows the normal appearance of the ACL (arrows) .
3120
FIGURE
48.13. Torn Anterior Cruciate Ligament (ACL). A
sagittal T2WI through the intercondylar notch shows fibers of the ACL that are disrupted centrally (arrow).This is a common MR appearance of a torn ACL.
P.1198 The insertion of the ligament of Humphry or Wrisberg onto the lateral meniscus can produce a pseudotear similar to that caused by the transverse ligament on the anterior horn of the lateral meniscus (Fig. 48.18). Prior to diagnosing a tear on the upper aspect of the posterior horn of the lateral meniscus, care must be taken to look for a meniscofemoral ligament to be certain it is not a pseudotear from the ligament's insertion. Similarly, prior to diagnosing a loose body in front of or behind the PCL, care must be taken to try to follow the structure across to the lateral meniscus to determine whether it is a meniscofemoral ligament.
3121
FIGURE
48.14. Normal Posterior Cruciate Ligament (PCL).
A sagittal T1WI (time of repetition 600; time of echo 30) through the intercondylar notch shows the appearance of the normal PCL, with its characteristic uniform low signal (arrow) .
3122
FIGURE
48.15. Torn Posterior Cruciate Ligament (PCL). A
sagittal T1WI (time of repetition 600; time of echo 30) through the intercondylar notch reveals the PCL to have diffuse intermediate signal throughout. This is typical for a torn PCL.
COLLATERAL Medial
LIGAMENTS
Collateral
Ligament
The medial collateral ligament (MCL) originates on the medial femoral condyle and inserts on the tibia. It is closely applied to the joint and is intimately associated with the medial joint capsule and
3123
the medial meniscus. The MCL is uniformly low in signal on T1 and T2 or T2* sequences. Injuries to the MCL usually occur from a valgus stress to the lateral part of the knee (such as a “clipping― injury in football). A grade 1 injury represents a mild sprain and is diagnosed on MR by the presence of fluid or hemorrhage in the soft tissues medial to the MCL. The ligament is otherwise normal. A grade 2 injury is a partial tear and is seen as high signal in and around the MCL on T2 or T2* coronal sequences. The ligament is intact, although the deep or superficial fibers may P.1199 show minimal disruption (Fig. 48.19). A grade 3 injury is a complete disruption of the MCL. It is best appreciated on T2 or T2* images (Fig. 48.20) .
FIGURE
48.16. Free Fragment of a Torn Meniscus. A sagittal
T1WI (time of repetition 600; time of echo 30) through the intercondylar notch in this patient with a torn meniscus shows two rounded, low-signal structures (arrows) that are free fragments of meniscal tissue. A meniscofemoral ligament of Wrisberg could have the appearance of either of these loose
3124
bodies.
A meniscocapsular separation occurs when the medial meniscus is torn from its attachment to the joint capsule. This occurs most commonly at the site of the MCL and often occurs concomitantly with an MCL injury. It is easily recognized on a T2 or T2* coronal image by noting joint fluid extending between the medial meniscus and the capsule (Fig. 48.21). It is essential to use T2 or T2* sequences, as a T1WI will not detect the fluid between the meniscus and the capsule.
Lateral
Collateral
Ligament
The lateral collateral ligament (LCL) consists of three parts. The most posterior P.1200 structure is the tendon of the biceps femoris, which inserts onto the head of the fibula. Next, anterior to the biceps, is the true lateral collateral ligament, also called the fibular collateral ligament, which extends from the lateral femoral condyle to the head of the fibula. The biceps and the fibular collateral ligament usually join and insert onto the head of the fibula in a conjoined fashion. Anterior to the fibular collateral ligament is the iliotibial band, which extends into the fascia more anteriorly and blends into the lateral retinaculum on the patella. The LCL is torn P.1201 infrequently in comparison to the MCL, but a tear can require surgery if instability is present. A torn LCL is seen as disruption of the ligamentous fibers on coronal images (Fig. 48.22) .
3125
FIGURE 48.17. Ligament of Wrisberg. A sagittal T1WI (time of repetition 600; time of echo 30) through the intercondylar notch shows a rounded, low-signal structure posterior to the posterior cruciate ligament, which is the meniscofemoral ligament of Wrisberg (arrow) .
FIGURE
48.18. Pseudotear From Ligament of Humphry
Insertion. A. A sagittal proton-density fat-suppressed image (time of repetition 2,000; time of echo 20) through the lateral
3126
meniscus reveals an apparent tear of the posterior horn (arrow) . (The “speckled― appearance in the anterior horn of the lateral meniscus is a frequently seen normal variant and should not be confused for a torn meniscus.) B . On the image through the intercondylar notch, a ligament of Humphry (arrow) is seen anterior to the posterior cruciate ligament (PCL). The ligament of Humphry could be followed on adjacent images, from anterior to the PCL to its insertion on the posterior horn of the lateral meniscus.
FIGURE 48.19. Partial Tear of the Medial Collateral Ligament (MCL). A gradient-recalled acquisition in the steady state coronal image (time of repetition 500; time of echo 15; θ 30°) reveals high signal adjacent to the MCL (arrows), which represents edema and hemorrhage from a partial tear or sprain of the MCL. The MCL is clearly intact; hence, a complete tear is easily excluded.
3127
FIGURE
48.20. Torn Medial Collateral Ligament (MCL). A
coronal gradient-recalled acquisition in the steady state image (time of repetition 500; time of echo 15; θ 30°) shows a large joint effusion, with the MCL disrupted proximally (arrow). In addition, joint fluid can be seen extending between the medial meniscus and the MCL, which indicates a meniscocapsular separation. Neither of these diagnoses could be made on the T1W coronal
images.
3128
FIGURE 48.21. Meniscocapsular Separation. A. A T1W coronal image (time of repetition [TR] 600; time of echo [TE] 30) reveals a contusion of the lateral femoral condyle (arrow) , indicative of a valgus strain, which is often associated with a medial collateral ligament (MCL) tear. The MCL appears normal on this image; however, the linear low signal in the soft tissues just adjacent to the MCL is suggestive of fluid. This would indicate a partial tear or sprain of the MCL. B . A coronal gradient-recalled acquisition in the steady state image (TR 500; TE 15; θ 30°) in the same knee reveals fluid between the medial meniscus and the MCL (arrow), which is diagnostic for a meniscocapsular separation. Faint high signal in the MCL and adjacent to it indicates a partial tear. A T2 or T2* sequence in the coronal plane is necessary to see these abnormalities.
PATELLA Chondromalacia
Patella
The patellar cartilage commonly undergoes degeneration, causing exquisite pain and tenderness. This is called chondromalacia patella. It can be diagnosed on sagittal images but is more easily identified on axial images. Because hyaline articular cartilage has the same signal intensity as joint fluid on T1W sequences, T2 or T2* sequences
3129
are necessary to diagnose chondromalacia patella in most instances. Chondromalacia patella begins with focal swelling and degeneration of the cartilage. This can be seen as low- or high-signal foci in the cartilage. Its progression causes thinning and irregularity of the articular surface of the cartilage; finally, underlying bone is exposed. This final stage occurs more commonly from trauma than from wear and tear. A frequent cause of a patellar cartilage defect is dislocation of the patella, in which the patella strikes the lateral femoral condyle and displaces a piece of patellar articular P.1202 cartilage (Fig. 48.23). The lateral femoral condyle invariably has a contusion following a dislocated patella.
FIGURE 48.22. Lateral Collateral Ligament Tear. A coronal fast spin-echo T2WI with fat suppression reveals a tear of the lateral collateral ligament (fibular collateral ligament) (arrow) . The normal ligament should be a low-signal structure between
3130
the femur and the fibula.
FIGURE
48.23. Chondral Defect in Patella. An axial fast spin-
echo T2WI (time of repetition 3,000; time of echo 108) through the patella shows a large cartilage defect on the apex and medial facet of the patella (white arrow) in this patient who suffered a dislocated patella. Note the high signal throughout the medial retinaculum (curved dislocation.
arrow), a frequent finding after a patella
3131
FIGURE
48.24. Plica. An axial gradient-recalled acquisition in
the steady state image (time of repetition 500; time of echo 15; θ 30°) through the patella shows a low-signal linear structure (arrow) extending from the medial capsule toward the medial facet of the patella. This is a medial patellar plica that is not abnormally thickened. Without the joint effusion or the T2WI, the plica would not be visualized.
3132
FIGURE 48.25. Contusion. A coronal fast spin-echo T2WI with fat suppression shows a focus of high signal in the lateral femoral condyle, which is a characteristic appearance for a severe bone contusion. This occurred from a patella dislocation.
Patellar
Plica
A normal structure seen in over half of the population is the medial patellar plica. It is an embryologic remnant from when the knee was divided into three compartments. It is a thin, fibrous band that extends from the medial capsule toward the medial facet of the
3133
patella (Fig. 48.24). Suprapatellar and infrapatellar plicae also exist. The medial patellar plica can, on rare occasions, thicken and cause clinical symptoms that are indistinguishable from those of a torn meniscus; this has been termed “plica syndrome.― An abnormal plica can be removed arthroscopically quite easily.
BONY
ABNORMALITIES
Contusions The most frequently encountered bony abnormality seen with MR is a contusion. A contusion represents microfractures from trauma (6) . They are also called bone bruises. They are easily identified on T1WIs P.1203 as subarticular areas of inhomogeneous low signal. With T2 weighting, a contusion will show increased signal for several weeks, depending on its severity (Fig. 48.25). Visualization of increased signal with T2* images can be difficult because of the susceptibility artifacts of the bone seen with T2* images. Contusions can progress to osteochondritis dissecans if they are not treated with decreased weight bearing; hence, an isolated bone contusion, with no other internal derangement, is a serious finding that requires protection.
3134
FIGURE
48.26. Contusion. A sagittal T1WI through the lateral
compartment shows irregular low signal in a subarticular location of the posterior tibial plateau and in the anterior part of the lateral femoral condyle. These findings are characteristic for bone contusions. This distribution of contusions in the posterior lateral tibial plateau and anterior in the lateral femoral condyle is almost always associated with a torn anterior cruciate ligament.
3135
FIGURE 48.27. Pes Anserinus Bursitis. A coronal T2* gradient-echo image shows a fluid collection below the medial joint line near the insertion of the pes anserinus tendons. This is pes anserinus bursitis.
FIGURE
48.28. Semimembranosus
3136
Tibial
Collateral
Ligament Bursa. A. A sagittal fast spin-echo (FSE) T2WI with fat suppression through the medial aspect of the knee shows a fluid collection (arrows) at the joint line that is adjacent to the posterior horn of the medial meniscus. This is characteristic of a semimembranosus medial collateral ligament bursa. B . A coronal FSE T2WI with fat suppression shows that this bursa has a comma-shaped appearance at the joint line (arrow) .
A commonly seen contusion is one that occurs on the posterior part of the lateral tibial plateau (Fig. 48.26). It is invariably associated with a torn ACL. Acute ACL tears have been reported to have this type of contusion in over 90% of cases (7) .
Fractures MR is useful in examining fractures about the knee. Tibial plateau fractures can be imaged precisely with CT; however, MR allows the soft tissues to be seen in addition to any bony abnormalities. A fracture that is almost always associated with an internal derangement is the Segond fracture. A small, bony fragment pulled off the posterior lateral tibial joint line by an avulsion of the lateral joint capsule, it is almost always associated with an ACL tear.
BURSAE An abnormality that can cause joint pain and clinically mimic plica syndrome or a torn meniscus is bursitis. Two bursae typically are identified medially that can become symptomatic. The first is the pes anserine bursa, which is somewhat uncommon. Three tendons—the sartorius, the gracilis, and the semitendinosus—insert onto the anteromedial aspect of the tibia in a fan-shaped been likened to a goose's foot, hence the name bursa lies beneath the insertion site, which can cause medial joint line or patellar pain; this can
manner that has pes anserinus. A become inflamed and be confused with
plica syndrome or a torn medial meniscus (Fig. 48.27). A second and much more common medial bursa is the semimembranosus tibial
3137
collateral ligament bursa. It occurs at the medial joint line and often mimics a meniscal cyst. It has a characteristic comma shape as it drapes over the semimembranosus tendon (Fig. 48.28). Making the diagnosis of pes anserinus or semimembranosus tibial collateral ligament bursitis with MR imaging can prevent an unnecessary arthroscopy procedure—one in which the bursae would be overlooked,
since
they
are
extracapsular
structures.
REFERENCES 1. Crues JI, Mink J, Levy T, et al. Meniscal tears of the knee: accuracy of MR imaging. Radiology 1987;164:445–448. 2. Mink JH, Deutsch AL. Magnetic resonance imaging of the knee. Clin Orthop 1989;244:29–47. 3. De Smet A, Graf B. Meniscal tears missed on MR imaging: relationship to meniscal tear patterns and anterior cruciate ligament tears. AJR Am J Roentgenol 1994;162:905–911. P.1204 4. Helms CA, Laorr A, Cannon WD Jr. The absent bow tie sign in bucket-handle tears of the menisci in the knee. AJR Am J Roentgenol
1998;170(1):57–61.
5. Silverman J, Mink J, Deutsch A. Discoid menisci of the knee: MR
imaging
appearance.
Radiology
1989;173:351–354.
6. Mink JH, Deutsch AL. Occult cartilage and bone injuries of the knee: detection, classification, and assessment with MR imaging. Radiology
1989;170:823–829.
7. Murphy B, Smith R, Uribe J, et al. Bone signal abnormalities in the posterolateral tibia and lateral femoral condyle in complete tears of the anterior cruciate ligament: a specific sign? Radiology
3138
1992;182:221–224.
3139
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section X - Musculoskeletal Radiology > Chapter 49 - Magnetic Resonance Imaging of the Shoulder
Chapter
49
Magnetic Resonance the Shoulder
Imaging
of
Clyde A. Helms MR of the shoulder is well accepted for its diagnostic utility for abnormalities of the rotator cuff and the glenoid labrum. It has been shown to have a high shoulder has replaced examining the rotator majority of diagnostic
degree of accuracy (1, 2, 3, 4). MR of the standard arthrography and CT arthrography for cuff and the glenoid labrum in the vast imaging centers.
ANATOMY The rotator cuff is comprised of the tendons of four muscles that converge on the greater and lesser tuberosities of the humerus: the supraspinatus, infraspinatus, subscapularis, and teres minor (Fig. 49.1). Of these, the supraspinatus most commonly causes clinically significant problems, and it is the most commonly surgically treated. The supraspinatus tendon lies just superior to the scapula and inferior to the acromioclavicular (AC) joint and acromion. It inserts onto the greater tuberosity of the humerus. Two to three centimeters proximal to its insertion is a section of the tendon called the “critical zone.― This area is reported to have decreased vascularity and is therefore less likely to heal following trauma. The critical zone of the supraspinatus tendon is a common location for rotator cuff tears. Most cuff tears, however, begin at the
3140
bone/tendon interface on the greater tuberosity. The glenoid labrum is a fibrocartilaginous ring that surrounds the periphery of the bony glenoid of the scapula. It serves as an attachment site for the capsule and broadens the base of the glenohumeral joint to allow increased stability. Tears or detachments of the glenoid labrum most commonly occur from, and result in, dislocations or instability of the humerus.
ROTATOR
CUFF
The rotator cuff commonly suffers from what has been termed impingement syndrome. Impingement of the supraspinatus tendon occurs from abduction of the humerus, which allows the tendon to be impinged between the anterior acromion and the greater tuberosity. The tendon can also be impinged by the undersurface of the AC joint if downward-pointing osteophytes or a thickened capsule are present. Other theories exist for rotator cuff disease, including natural degeneration from aging and a predisposition for the cuff to undergo degeneration as a result of decreased blood supply. Most investigators agree that whatever the cause, the natural course of impingement syndrome or cuff degeneration is a complete, or fullthickness, tear of the rotator cuff. The rotator cuff is best seen on oblique coronal images that are aligned parallel to the supraspinatus muscle (Fig. 49.2) and on oblique sagittal images (Fig. 49.3). Both T1W (or proton-density) and T2W sequences are typically performed, although little diagnostic information is P.1206 P.1207 present on the T1WIs and they are not obtained by many radiologists. Multiple acceptable variations of imaging sequences are available to demonstrate normal and abnormal structures. A fatsuppressed fast spin-echo (FSE) T2W oblique coronal sequence is gaining popularity as the primary sequence for imaging the rotator cuff in many imaging centers. The slice thickness should be no greater than 5 mm, and 3 mm is preferable. As with most joint
3141
imaging, a small field of view (16 to 20 cm) is recommended. A dedicated shoulder coil or a surface coil placed anteriorly over the shoulder is necessary.
FIGURE
49.1. Normal Shoulder Anatomy. A. Drawing showing
the rotator cuff muscles in a sagittal plane (anterior is on the left). C, Coracoid; A, acromion; H, humeral head. B . This sagittal T1WI through the glenoid shows the normal cuff musculature. SUB, Subscapularis; SUP, supraspinatus; IS, infraspinatus; T, teres minor; C, coracoid process; G, glenoid.
3142
FIGURE
49.2. Oblique Coronal Image of Normal Rotator
Cuff. This oblique coronal fast spin-echo T2WI with fat suppression through the supraspinatus shows a normal supraspinatus with a broad footprint where the tendon inserts onto the greater tuberosity.
3143
FIGURE
49.3. Oblique Sagittal Image of a Torn Rotator
Cuff. An oblique coronal image fast spin-echo T2WI with fat suppression shows the rotator cuff inserting onto the greater tuberosity in a normal fashion, except at the far anterior portion (arrow). This indicates a partial tear of the articular surface fibers of the cuff.
3144
FIGURE
49.4. Internal Rotation Hiding Partial Tear of the
Supraspinatus Tendon. A. An oblique coronal fast spin-echo T2WI with fat suppression shows an apparently normal supraspinatus inserting onto the greater tuberosity (arrow). B . One slice anterior to that shown in (A), the bicipital groove can be identified, with the fibers of the supraspinatus just lateral to the groove lifted off of the greater tuberosity (arrow). This is a partial tear of the rotator cuff at its anteriormost portion.
In examining the rotator cuff, the most anterior oblique coronal images will show the supraspinatus tendon. A useful landmark for noting the anterior portion of the supraspinatus tendon is the bicipital groove, which has the anteriormost fibers of the supraspinatus immediately lateral to the groove. This is where most cuff tears begin and can be overlooked if the patient's shoulder is internally rotated, which is common (Fig. 49.4). The infraspinatus tendon is seen on the more posterior images and can easily be mistaken for the tendon of the supraspinatus. The supraspinatus tendon can be differentiated from that of the infraspinatus by noting the more horizontal course of the supraspinatus as compared with the infraspinatus, which runs obliquely inferiorly to superiorly.
3145
The normal supraspinatus tendon is said to be uniformly low in signal on all pulse sequences. Unfortunately, this is not always the case. In fact, it usually has some intermediate to high signal in the tendon, which causes much confusion. If the signal in the tendon gets brighter on T2WIs, it is abnormal and represents either tendonitis (many investigators prefer the terms tendinosis or tendinopathy over tendonitis, as no inflammatory cells are found histologically) or a partial tear. A partial tear can be diagnosed by noting thinning of the tendon itself (Fig. 49.5). However, if it disappears or has the same signal P.1208 intensity as the adjacent muscle on T2WIs, it may represent one of many different processes:
FIGURE 49.5. Partial Cuff Tear. An oblique coronal fast spinecho T2 image with fat suppression shows thinning of the supraspinatus tendon (arrow), which is a partial tear of the articular side of the cuff.
Partial volume averaging of peritendinous fat can cause some high signal in the supraspinatus tendon on the oblique coronal images; this does not get brighter on T2WIs.
3146
If the plane of the oblique coronal images is slightly off of the plane of the tendon, muscle slips can be partially volume averaged, which may appear as relatively high signal on T1WIs but will not get brighter with T2WIs. The so-called “magic angle― effect can cause apparent high signal in a tendon that lies at 55° to the bore of the magnet (as does the critical zone of the supraspinatus tendon) (5). This high signal will not be seen on T2WIs (or any sequence with a long time of echo). This is believed to be a common cause of high signal in the critical zone on T1WIs. Myxoid and fibrillar degeneration of the supraspinatus tendon are commonly found in autopsy specimens in older patients. The majority of asymptomatic shoulders in patients over the age of 50 are believed to have some degeneration of the supraspinatus; this has been termed tendinopathy. This is seen as high signal in the critical zone on T1WIs that does not increase with T2 weighting (6) . Myxoid degeneration is felt by many surgeons to be more significant than anatomic impingement as a source of cuff pathology (7). Rather than decompressing the coracoacromial arch by removing bony structures and the coracoacromial ligament, which might be sources of impingement, these surgeons resect the areas of myxoid degeneration in the cuff tendons. Tendon degeneration (tendinopathy) can be seen in asymptomatic shoulders of all ages; hence, it needs to be correlated with the clinical picture. If the signal gets brighter on T2WIs, it must be considered pathologic—either tendonitis or a partial tear. If disruption of the supraspinatus tendon is seen, obviously a fullthickness tear is present. In these cases, fluid is invariably present in the subacromial bursa (Fig. 49.6). It should be noted that fluid in the subacromial bursa can also occur from isolated subacromial bursitis or for several days following a therapeutic injection into the bursa. Care should be made to look for retraction of the supraspinatus muscle, as marked retraction will obviate some types of surgery.
3147
Three basic categories exist for the appearance of the supraspinatus tendon: Normal—high signal on T1WIs that does not get brighter on T2WIs. This can represent one of several processes in a normal tendon or myxoid degeneration, also called tendinopathy. This should basically be considered “normal― unless it is an inordinate amount, as it has not been proved that tendinopathy is symptomatic.
FIGURE
49.6. Complete Tear of the Supraspinatus
Tendon. An oblique coronal fast spin-echo T2WI with fat suppression shows disruption of the supraspinatus tendon (arrow) with fluid in the torn tendon.
Tendonitis (or tendinosis)—high signal on T1WIs that gets brighter on T2WIs. This represents tendonitis or a partial tear. Cuff tear—tendon disruption; a gap must be seen to call a full-
3148
thickness cuff tear. It is generally easy to place the MR appearance of the rotator cuff into one of these three categories, and a high degree of accuracy in diagnosing the state of the rotator cuff can be expected. Partial cuff tears have marked clinical significance, since most agree that they will not heal on their own if they are greater than 25% of the cuff thickness (8). Although we generally cannot be so precise as to what percentage of the cuff is involved, we can usually identify partial cuff tears. If there is an irregularity or thinning of the cuff on either the bursal side of the cuff (Fig. 49.5) or on the joint side I will describe it as small, medium, or large (near full thickness). A particular type of articular-sided partial tear has been described that is commonly seen. This has been termed a “rim rent― tear (Fig. 49.7). It occurs at the insertion of the fibers of the cuff onto the greater tuberosity. It most commonly occurs anteriorly, at the insertion of the supraspinatus, and, as mentioned previously, can be easily overlooked if the patient's arm is internally rotated.
3149
FIGURE
49.7. Rim Rent Tear. A. An oblique coronal fast spin-
echo (FSE) T2 image with fat suppression shows increased signal at the insertion of the supraspinatus onto the greater tuberosity (arrow). B . An oblique sagittal FSE T2 image with fat suppression shows linear high signal anteriorly between the cuff fibers and the greater tuberosity (arrow). This is an articular-sided partial tear called a rim rent tear.
P.1209
BONY
ABNORMALITIES
The undersurface of the anterior acromion and the AC joint should be examined for osteophytes or irregularities that can be responsible for impingement syndrome (Fig. 49.8). In the proper clinical setting, an anterior acromioplasty will relieve the symptoms of impingement syndrome and prevent a more serious full-thickness rotator cuff tear.
3150
Many believe it is imperative that the surgeon also remove any AC joint undersurface irregularity, if present, or failed surgery can occur, although more recently this theory has been challenged (7) . Abnormalities of the humeral head include sclerosis and cystic changes about the greater tuberosity, which are commonly present in patients with impingement syndrome and rotator cuff tears. Bony impaction on the posterosuperior aspect of the humeral head can be seen in patients with anterior instability of the humeral head. This is called a Hill-Sachs lesion and is best identified on the two or three superiormost axial images (Fig. 49.9). The normal humeral head should be round on the superior slices; any irregularity seen posteriorly is abnormal.
FIGURE 49.8. Acromioclavicular Joint Osteophytes. An oblique coronal T1WI (time of repetition 600; time of echo 30) reveals osteophytes extending inferiorly off the acromioclavicular
3151
joint (arrow). This is a common source of impingement on the supraspinatus tendon.
FIGURE 49.9. Hill-Sachs Lesion. An axial T1WI (time of repetition 600; time of echo 30) through the superior portion of the humeral head shows a posterior impaction (arrow) caused by the glenoid labrum during an anterior dislocation of the humerus. This has been termed a Hill-Sachs lesion.
P.1210
GLENOID
LABRUM
Tears or detachments of the glenoid labrum cause glenohumeral joint instability. They are commonly caused by dislocations, but less traumatic episodes, such as repeated trauma from throwing, can also result in labral tears. Torn or detached labra are often repaired arthroscopically
with
good
results.
3152
The glenoid labrum is best imaged on axial T2WIs or T2*WIs. Axial T1WIs are not necessary to diagnose labral abnormalities and can be omitted from the shoulder protocol. Fluid in the joint makes for easier assessment of the labrum; hence, MR arthrography has evolved into a routine exam in many centers. It is performed by injecting a 1:200 dilution of gadolinium/diethylenetriamine pentaacetic acid with saline into the joint using fluoroscopic guidance. The normal labrum is a triangular-shaped, low-signal structure as viewed on an axial image, with the anterior labrum usually larger than the posterior labrum (Fig. 49.10). The anterior labrum is much more commonly involved with tears than the posterior, and the superior labrum is even less commonly involved. The superior labrum is evaluated on the oblique coronal views. If no joint effusion is present, a labral tear can be difficult to see unless it is quite severe. If joint fluid extends between the bony glenoid and the base of the labrum, a detached labrum is present. Tears or detachments of the labrum are diagnosed by noting fluid extending between the labrum and the bony glenoid or by truncation of the labrum (Fig. 49.11). Superior labral tears are called SLAP lesions (superior labrum anterior
to posterior)
(Fig. 49.12). They are
seen most frequently in throwing athletes secondary to the pull of the long head of the biceps, which inserts on the superior labrum. They are also seen in older patients in association with cuff tears.
3153
FIGURE 49.10. Normal Labrum. An axial T2* gradient-recalled acquisition in the steady state (GRASS) image (time of repetition 600; time of echo 30; θ 20°) shows a normal anterior (black arrow) and posterior (white arrow) glenoid labrum. The anterior labrum is usually larger than the posterior labrum.
Several normal variants in the labrum that can mimic a torn or detached labrum have been described. Two P.1211 occur solely in the anterosuperior portion of the labrum, an area in which tears are uncommon. The first is a sublabral foramen, which is an opening beneath the anterosuperior labrum and the bony glenoid that mimics a detachment (Fig. 49.13). This is seen in up to 20% of the population. A second variant is called a Buford complex. It consists of an absent anterosuperior labrum in association with a thickened, cordlike middle glenohumeral ligament. This is seen in about 3% of the population (9). A sublabral recess is often seen on the oblique coronal images, which can mimic a SLAP tear. It is found in up to 70% of shoulders. A sublabral recess should be seen only on the anterior part of the superior labrum, should be thin and smooth (Fig. 49.14), and extends medially, whereas a SLAP tear is typically more irregular and extends superiorly or laterally.
3154
FIGURE 49.11. Torn Labrum. An axial fast spin-echo T2WI with fat suppression shows a tear of the anterior labrum (arrow) .
3155
FIGURE
49.12. SLAP (Superior Labrum Anterior
to
Posterior) Lesion. An oblique coronal T2WI with fat suppression shows a torn superior labrum (arrow) .
FIGURE 49.13. Sublabral Foramen. This axial fast spin-echo T2WI with fat suppression reveals fluid between the glenoid and the anterior labrum (white arrow), which appears to be a detached labrum; however, this is a sublabral foramen, which is a normal variant seen only in the anterosuperior labrum. Note the normal middle glenohumeral ligament (black arrow) anterior to the labrum.
3156
FIGURE
49.14. Sublabral
Recess. A coronal T1W gadolinium
arthrogram with fat suppression shows fluid between the superior labrum and the cartilage of the bony glenoid (arrow), which is thin and smooth. This is a sublabral recess.
BICEPS
TENDON
The long head of the biceps tendon runs in the bicipital groove between the greater and lesser tuberosities and inserts onto the superior labrum. It can be impinged by an abnormal acromion in the same way the supraspinatus tendon is impinged, resulting in tenosynovitis or tendonitis. In tenosynovitis, fluid can be seen in the tendon sheath surrounding an otherwise normal tendon. Because fluid in the glenohumeral joint can normally fill the biceps tendon sheath, this diagnosis is difficult to make with MR alone. If the tendon is enlarged and/or has signal within it, tendonitis or a partial tear is present (Fig. 49.15). If the tendon is not seen on one or more of the axial images, it is disrupted or dislocated. Dislocation is uncommon, but when it occurs, the tendon can be seen to lie
3157
anteromedial to the joint. A subscapularis tear must be present if the biceps is dislocated.
FIGURE
49.15. Biceps
Tendonitis. An axial T2* gradient-
recalled acquisition in the steady state (GRASS) image (time of repetition 600; time of echo 30; θ 30°) shows the biceps tendon (arrow) to be swollen and filled with high signal, indicating tendonitis.
P.1212
SUPRASCAPULAR
NERVE
ENTRAPMENT
The suprascapular nerve is made up of branches from the C4, C5, and C6 roots of the brachial plexus. It runs superior to the scapula, from anterior to posterior, just medial to the coracoid process. It
3158
gives off a branch that innervates the supraspinatus muscle as it courses posteriorly in the suprascapular notch, and then innervates the infraspinatus muscle after it runs through the spinoglenoid notch in the posterior scapula. It can easily be entrapped by a tumor or a ganglion as it runs above the scapula, because it is bounded superiorly by a transverse ligament both anteriorly and posteriorly. A fairly common finding is a ganglion in the spinoglenoid notch that impresses the infraspinatus portion of the nerve, with resultant pain and atrophy of the infraspinatus muscle (Fig. 49.16). This is most commonly seen in men who are athletic, particularly weightlifters. The ganglion can be percutaneously drained with CT guidance or surgically removed. They can also spontaneously rupture, which results in cessation of symptoms (1 0). There is a 100% association of a torn posterior labrum with these cysts.
FIGURE
49.16. Ganglion in Spinoglenoid Notch. An axial
T2WI reveals a high-signal mass posterior to the scapula in the spinoglenoid notch (arrow). This is a ganglion that has impressed the suprascapular nerve, causing shoulder pain and atrophy of the infraspinatus muscle.
3159
QUADRILATERAL
SPACE
SYNDROME
Oblique sagittal T1WIs are useful to observe fatty atrophy in any of the cuff muscles. If the infraspinatus is smaller than the other muscles and/or has fatty infiltration, P.1213 the aforementioned suprascapular nerve entrapment secondary to a ganglion in the spinoglenoid notch is the likely diagnosis. If the teres minor has fatty atrophy (Fig. 49.17), the diagnosis is quadrilateral space syndrome. This most commonly occurs from fibrous bands or scar tissue in the quadrilateral space impinging on the axillary nerve. The quadrilateral space lies between the teres minor superiorly, the teres major inferiorly, the long head of the triceps medially, and the diaphysis of the humerus laterally. The axillary nerve traverses the quadrilateral space and innervates the teres minor and deltoid muscles; however, the deltoid is rarely involved in quadrilateral space syndrome. Quadrilateral space syndrome is found in about 1% of shoulder MR images. The presentation of these patients is clinically similar to that of a rotator cuff tear, and many patients have had needless surgery for presumed cuff pathology when the real problem was quadrilateral space syndrome. Generally no surgery is necessary, as physical therapy is usually successful in breaking up the fibrous bands or scar tissue that cause this entity.
3160
FIGURE
49.17. Quadrilateral
Space
Syndrome. This oblique
sagittal T1WI shows fatty atrophy of the teres minor muscle (arrow), which is diagnostic of quadrilateral space syndrome.
PARSONAGE-TURNER
SYNDROME
Oblique sagittal FSE T2W fat-suppressed images are useful for identifying muscle edema. In about 1% of cases, neurogenic edema is found in muscle groups that corresponds to a particular nerve (i.e., supraspinatus/infraspinatus = suprascapular nerve; teres minor/deltoid = axillary nerve). This is characteristic for ParsonageTurner syndrome (Fig. 49.18) but not pathognomonic, because a traumatic nerve injury (such as a brachial plexus injury) could have a similar appearance. It becomes pathognomonic once the clinical presentation is provided. If there is no history of trauma or of an
3161
insidious onset, and if the onset is sudden, with severe pain followed in a day or two by profound weakness, the edema pattern is virtually pathognomonic for Parsonage-Turner syndrome. The etiology of Parsonage-Turner syndrome is unknown, but it seems to have an association with prior vaccinations, viral illness, or general anesthesia in about one third of cases. It is bilateral in about 10% to 15% of cases. It affects all ages of both sexes and is selflimited. It can affect either the axillary or suprascapular nerves, or both simultaneously. Unnecessary shoulder, brachial plexus, and cervical spine surgeries have been performed on patients with Parsonage-Turner syndrome before the correct diagnosis was made. Parsonage-Turner syndrome was first described in the radiology literature in 1998 (1 1), indicating that we all missed it on MR for over 15 years. This is because fat suppression on shoulder images was not routinely done until the early 1990s, and the edema in the muscles was not conspicuous enough to be picked up on non–fatsuppressed sequences.
FIGURE 49.18. Parsonage-Turner Syndrome. An oblique sagittal T2WI with fat suppression shows edema in the supraspinatus (S) and the infraspinatus (I) muscles, consistent
3162
with involvement of the suprascapular nerve. The sudden onset with no history of trauma is characteristic for Parsonage-Turner syndrome.
REFERENCES 1. Palmer W, Brown J, Rosenthal D. Rotator cuff: evaluation with fat-suppressed MR arthrography. Radiology 1993;188:683–688. 2. Singson RD, Hoang T, Dan S, Friedman M. MR evaluation of rotator cuff pathology using T2-weighted fast spin-echo technique with and without fat suppression. AJR Am J Roentgenol 1996;166(5):1061–1065. 3. Rafii M, Firooznia H, Sherman O, et al. Rotator cuff lesions: signal patterns at MR imaging. Radiology 1990;177(3):817–823. 4. Zlatkin MB, Iannotti JP, Roberts MC, et al. Rotator cuff tears: diagnostic performance of MR imaging. Radiology 1989;172:223–229. 5. Erickson S, Cox I, Hyde J, et al. Effect of tendon orientation on MR imaging signal intensity: a manifestation of the “magic angle― phenomenon. Radiology 1991;181:389–392. 6. Kjellin I, Ho CP, Cervilla V, et al. Alterations in the supraspinatus tendon at MR imaging: correlation with histopathologic findings in cadavers. Radiology 1991;181:837–841. 7. Budoff JE, Nirschl RP, Guidi EJ. Debridement of partialthickness tears of the rotator cuff without acromioplasty. Long-
3163
term follow-up and review of the literature. J Bone Joint Surg Am 1998;80:733–748. 8. Fukuda H. The management of partial-thickness tears of the rotator cuff. J Bone Joint Surg Br 2003;85:3–11. 9. Carroll KW, Helms CA. Magnetic resonance imaging of the shoulder: a review of potential sources of diagnostic errors. Skeletal
Radiol
2002;31(7):373–383.
10. Fritz R, Helms CA, Steinbach L, Genant H. Suprascapular nerve entrapment: evaluation with MR imaging. Radiology 1992;182:437–444. 11. Helms CA, Martinez S, Speer KP. Acute brachial neuritis (Parsonage-Turner-syndrome): MR imaging appearance—report of three cases. Radiology 1998;207:255–259.
3164
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section X - Musculoskeletal Radiology > Chapter 50 - Magnetic Resonance Imaging of the Foot and Ankle
Chapter
50
Magnetic Resonance the Foot and Ankle
Imaging
of
Clyde A. Helms MR is playing an increasingly important role in the examination of the foot and ankle (1). Orthopaedic surgeons and podiatrists are learning that critical diagnostic information can be obtained in no other way and are relying on MR for many therapeutic decisions.
TENDONS One of the more common reasons to perform MR of the foot and ankle is to examine the tendons. Although multiple tendons course through the ankle, only a few are routinely affected pathologically. These are primarily the flexor tendons, located posteriorly in the ankle. The extensor tendons, located anteriorly, are rarely abnormal. Only those tendons that are more commonly seen to be abnormal will be discussed in detail. Tendons can be traumatized directly or be injured from overuse. Either etiology can result in (1) tenosynovitis, which is seen on MR as fluid in the tendon sheath with the underlying tendon appearing normal; (2) tendonitis or a partial tear, which is seen as focal or fusiform swelling of the tendon with signal within the tendon that gets bright on T2W or T2*W images; thinning or attenuation of the tendon is a more severe form of tendonitis that can be recognized on MR; tendinosis is seen as increased signal within a tendon that does
3165
not get fluid-bright on T2WI and represents myxoid degeneration; and (3) tendon rupture, which is best identified on axial images by noting the absence of a tendon on one or more images. Complete tendon disruption can be difficult to see on sagittal or coronal images because of the tendency for tendons to run oblique to the plane of imaging. An exception to this is the Achilles tendon, which is usually best seen on a sagittal image (2) . It is important to distinguish between tendonitis (or partial tear) and a complete disruption, because surgical repair is often warranted for the latter and not for the former. Making the distinction clinically is often difficult.
Achilles
Tendon
The Achilles tendon does not have a sheath associated with it; therefore, tenosynovitis does not occur. Tendonitis (or a partial tear) is commonly seen in the Achilles tendon; however, it is such an easy clinical diagnosis that MR is usually not necessary. Complete disruption is commonly seen in athletes and in men who are approximately 40 years of age. It is also commonly associated with other systemic disorders that cause tendon weakening, such as rheumatoid arthritis, collagen vascular diseases, and hyperparathyroidism.
diseases,
crystal
deposition
Achilles tendon disruption can be treated surgically or by placing the patient in a cast with equinus positioning (marked plantar flexion) for several months. Which treatment is superior is a controversial issue, with both methods of treatment seemingly working well. MR is used by many surgeons to help decide whether surgery should be performed. If a large gap is present (Fig. 50.1), some surgeons feel that surgery should be performed for reapposition of the torn ends of the tendon; on the other hand, if the ends of the tendon are not retracted, nonsurgical P.1215 treatment is preferred. However, no published papers have shown that this is, in fact, valid.
3166
FIGURE
50.1. Torn Achilles Tendon. A sagittal T1WI (time of
repetition 600; time of echo 30) reveals the Achilles tendon to be torn, with a 2-cm gap. Only a thin remnant of the tendon remains intact across the gap (arrow). Note the high signal in the swollen ends of the separated tendon, indicative of hemorrhage and edema.
3167
FIGURE 50.2. Normal Ankle Anatomy. A. This drawing of the tendons around the ankle at the level of the tibiotalar joint shows the relationship of the flexor tendons posteriorly and the extensor tendons anteriorly. B . An axial T1WI through the ankle just above the tibiotalar joint shows the normal anatomy. A, Achilles tendon; T, posterior tibial tendon; D, flexor digitorum tendon; H, flexor hallucis tendon; P, peroneus tendons; TA, tibialis anterior tendon.
Posterior
Tibial
Tendon
The flexor tendons are easily remembered and identified by using the mnemonic “Tom, Dick, and Harry,― with Tom representing the posterior tibial tendon, Dick the flexor digitorum longus tendon, and Harry the flexor hallucis longus tendon. The posterior tibial tendon (PTT) is the most medial and the largest, except for the Achilles, of the flexor tendons (Fig. 50.2). The posterior tibial tendon inserts
3168
onto the navicular, second and third cuneiforms, and the bases of the second to fourth metatarsals. As it sweeps under the foot, it provides some support for the longitudinal arch; hence, problems in the arch or plantar fascia can sometimes lead to stress on the posterior tibial tendon, with resulting tendonitis or even rupture. Posterior tibial tendonitis and rupture are commonly encountered in patients with rheumatoid
arthritis.
Differentiation of tendonitis from tendon rupture can be difficult, and MR has become very valuable for making this distinction (3). Most surgeons will operate on a disrupted posterior tibial tendon; however, nonoperative therapy is usually preferred for tendonitis. Posterior tibial tendinosis is seen on axial T1WIs as swelling and/or signal within the normally low-signal tendon on one or more images (Fig. 50.3). T2WIs or T2*WIs show the signal in the tendon getting brighter but not fluid-bright. Tendon disruption is diagnosed by noting the absence of the tendon on one or more axial images (Fig. 50.4) . P.1216 This typically occurs just at or above the level of the tibiotalar joint.
3169
FIGURE
50.3. Posterior
Tibial
Tendon
Tendinosis. A proton-
density (time of repetition 2,000; time of echo 20) axial image through the ankle at the midlevel of the calcaneus shows the posterior tibial tendon (arrow) swollen and containing high signal. This is the appearance of marked tendinosis.
3170
FIGURE
50.4. Torn Posterior Tibial Tendon. Axial T1WI (A)
and T2WI (B) through the ankle in this patient with chronic pain reveal a distended posterior tibial tendon sheath (arrows), with no low-signal tendon identified within. This is a tear of the posterior
tibial
tendon.
Rupture of the PTT results clinically in a flat foot as a result of the loss of arch support given by this tendon. The spring ligament runs just deep to the PTT and then goes underneath the neck of the talus, which it supports in a sling-like fashion. When the PTT tears, stress is then placed on the spring ligament to support the talus and the arch. The spring ligament has a high incidence of disruption when the PTT tears. The spring ligament is identified on axial and coronal images just deep to the PTT. When it is stressed, it typically gets scarred and thickened (Fig. 50.5). A tear can be diagnosed by noting a gap in the ligament. After the PTT and the spring ligament tear, the next structures to fail are the subtalar joint ligaments in the sinus tarsi. In a report of 25
3171
patients with PTT tears, it was found that 92% of the cases had abnormal spring ligaments (thickened or torn), and 72% had an abnormal sinus tarsi (4). It's clear that these structures are linked, and injury or stress to one can affect the others. Flexor hallucis longus tendon (FHL) tendon is easily identified near the tibiotalar joint because it is usually the only tendon at that distal level that has muscle still attached. In the foot, the FHL can be seen beneath the sustentaculum talus, which it uses as a pulley to plantar flex the foot. The FHL is known as the Achilles tendon of the foot in ballet dancers because of the extreme flexion positions they employ. Ballet dancers often will have tenosynovitis of the FHL, seen on MR as fluid in the sheath surrounding the tendon. Care must be taken to have clinical correlation, because fluid can be seen in the FHL tendon sheath from a connection to the ankle joint, which has an effusion in P.1217 as many as 20% of normal patients. Rupture of the FHL is rare.
3172
FIGURE
50.5. Abnormal
Spring
Ligament. An axial T2WI
through the ankle shows a markedly thickened spring ligament (arrows) just deep to the posterior tibial tendon.
3173
FIGURE
50.6. Dislocated
Peroneus
Longus
Tendon. An axial
T1WI in this rock climber who injured his ankle in a fall shows a low-signal rounded structure (arrow) lateral to the lateral malleolus. This is a dislocated peroneus longus tendon.
Peroneus
Tendons
The peroneus longus and peroneus brevis tendons can be seen posterior to the distal fibula, to which they are bound by a thin fibrous structure, the superior retinaculum. The fibula serves as a pulley for the tendons to work as the principal everter of the foot. The tendons course close together adjacent to the lateral aspect of the calcaneus until a few centimeters below the lateral malleolus, where they separate, with the peroneus brevis tendon inserting onto the base of the fifth metatarsal and the peroneus longus tendon
3174
crossing under the foot to the base of the first metatarsal. Avulsion of the base of the fifth metatarsal from a pull by the peroneus brevis tendon is known as a “dancer's fracture― or a Jones fracture. Disruption of the superior retinaculum, often seen in skiing accidents (5), can result in displacement of the peroneus tendons (Fig. 50.6) and must be surgically corrected. P.1218 It often occurs with a small bony avulsion, called a flake fracture, off the fibula.
3175
FIGURE 50.7. Longitudinal Split Tear of the Peroneus Brevis. This axial T1WI shows the peroneus brevis (arrow) with a “V― or chevron shape, which is characteristic for a longitudinal split tear of the brevis.
Entrapment of the peroneus tendons in a fractured calcaneus or fibula can occur and is easily diagnosed with MR. This can be a difficult diagnosis to make clinically. Complete disruption of the peroneus tendons is uncommon but is easily noted with MR.
FIGURE 50.8. Unstable Osteochondral Lesion of the Talus. A . A proton-density (time of repetition [TR] 2,000; time of echo [TE] 20) coronal image through the talus shows a focus of low signal in the medial subarticular part of the talus (arrow). This is a characteristic appearance for an osteochondral lesion. B . A T2WI (TR 2,000; TE 80) shows high signal throughout the focus of the osteochondral lesion, which indicates an unstable fragment.
3176
Longitudinal split tears of the peroneus brevis are commonly seen in patients following an inversion ankle sprain with associated dorsiflexion. The peroneus brevis gets trapped against the fibula by the peroneus longus, and a longitudinal split tear of the peroneus brevis results. These patients have chronic lateral ankle pain, often associated with ankle instability because of the lateral collateral ligament disruption that also occurs with the inversion trauma. A split tear of the peroneus brevis is easily identified on MR images by noting either a chevron or “V― shape to the tendon distal to the fibula (Fig. 50.7), or by noting a division of the tendon into two parts. There is an 80% association with lateral ligament tears, so close attention should be paid to the ligaments when a split tear of the peroneus brevis is found.
AVASCULAR
NECROSIS
Avascular necrosis commonly occurs in the foot and ankle. The talar dome is the second most common location for an osteochondral lesion (OCL), formerly called osteochondritis dissecans. (The knee is the most common site.) MR is useful in identifying and staging an OCL. Even when not apparent on plain films, MR can show an OCL as a P.1219 focal area of low signal in the subarticular portion of the talar dome on T1WIs. On T2WIs or T2*WIs, if high signal is seen surrounding the dissecans fragment in the bone at the bed of the fragment or throughout the fragment (Fig. 50.8), the fragment is most likely unstable. If the fragment has become displaced and lies in the joint as a loose body, MR can sometimes be useful to localize it; however, loose bodies in any joint can be exceedingly difficult to find.
3177
FIGURE 50.9. Avascular Necrosis of the Tarsal Navicular. A T1W (time of repetition 600; time of echo 30) sagittal image of the ankle in this patient with pain on the dorsum of the foot shows diffuse low signal throughout the tarsal navicular. This is a characteristic appearance for avascular necrosis and will often precede any plain film findings.
Diffuse low signal throughout a tarsal bone on a T1WI avascular necrosis. If the signal is increased on T2WIs, may not be reversible. This occasionally occurs in the (Fig. 50.9). MR can be useful in making this diagnosis films are normal or equivocal.
3178
is typical for it may or tarsal navicular when plain
FIGURE 50.10. Giant Cell Tumor of Tendon Sheath. Axial proton-density (A) and T2W (B) images reveal a mass surrounding the flexor hallucis longus tendon (arrows), which is confined by the tendon sheath. Although high-signal fluid is present, large amounts of low-signal material lines the distended tendon sheath. This low signal is hemosiderin, which is typically found in a giant cell tendon of tendon sheath. Pigmented villonodular synovitis in a joint has an identical appearance.
TUMORS A few tumors have a predilection for the foot and ankle (6). Up to 16% of synovial sarcomas occur in the foot. Desmoid tumors are commonly seen in the foot. Giant cell tumors of tendon sheath are often found in the tendon sheaths of the foot and ankle (Fig. 50.10) . They are characterized by marked low signal in the synovial lining and in the tendons on T1WIs and T2WIs, just as pigmented
3179
villonodular synovitis appears in a joint. The differential diagnosis for calcaneal tumors is similar to that of the epiphyses—giant cell tumor, chondroblastoma, and infection—with a unicameral bone cyst added. Soft tissue tumors in the medial aspect of the foot and ankle can press on the posterior tibial nerve, resulting in tarsal tunnel syndrome (7). Clinically, patients with tarsal tunnel syndrome present with pain and paresthesia in the plantar aspect of the foot. In the aforementioned P.1220 mnemonic, “Tom, Dick, and Harry,― the “and― is for artery, nerve, and vein. It is the position of the posterior tibial nerve. The nerve is easily compressed in the tarsal tunnel, which is bounded medially by the flexor retinaculum, a strong fibrous band that extends across the medial ankle joint for approximately 5 to 7 cm in a superior-to-inferior direction. Ganglions and neural tumors, which can look similar on T1WIs and T2WIs, often lie in the tarsal tunnel (Fig. 50.11) and compress the posterior tibial nerve, resulting in pain and paresthesia on the plantar aspect of the foot, extending into the toes. Tarsal tunnel syndrome often occurs secondary to trauma or fibrosis, or it can occur idiopathically. Regardless, this syndrome may not respond to surgical intervention; hence, MR is valuable in delineating a treatable lesion in many cases.
3180
FIGURE 50.11. Ganglion Causing Tarsal Tunnel Syndrome. A fast spin-echo T2W axial image of the ankle in a patient complaining of pain and paresthesia on the plantar aspect of the foot shows a homogeneous, high-signal mass (arrow) lying adjacent to the flexor hallucis longus tendon. This is the position of the tarsal tunnel, which contains the tibial nerve, that can be impinged by a mass, such as in this case, resulting in tarsal tunnel syndrome. This was a ganglion.
3181
FIGURE
50.12. Anomalous
Muscle. An axial T1WI of both
ankles in this patient complaining of a mass in the right ankle shows an anomalous muscle (arrow) lateral to the flexor hallucis longus muscle that is responsible for the mass the patient feels.
3182
FIGURE
50.13. Schematic of Lateral Collateral Ligaments.
A . This drawing of the ankle in a lateral view shows how the anterior (ant.) and posterior (post.) tibiofibular ligaments extend off the fibula and course superiorly to the tibia. B . A drawing in the axial plane shows that the fibula has a flat or convex surface at the origin of these ligaments.
3183
FIGURE
50.14. Schematic of Lateral Collateral Ligaments.
A . This drawing of the ankle in a lateral view shows how the anterior (ant.) and posterior (post.) talofibular ligaments and the calcaneofibular ligament extend off the fibula and course inferiorly. These ligaments arise off of the fibula more distally than the anterior and posterior tibiofibular ligaments. B . A drawing in the axial plane shows that the anterior and posterior talofibular ligaments arise from the level of the distal fibula, which has a concave medial surface, the malleolar fossa.
3184
FIGURE
50.15. Anterior
Talofibular
Ligament.
A. An axial
T2WI (time of repetition [TR] 4,000; time of echo [TE] 76) through the distal fibula at the level of the malleolar fossa (the concave medial surface of the fibula) shows an intact anterior talofibular
ligament
(arrow) that makes up part of the joint
capsule at this level. Note the high-signal joint fluid adjacent to the ligament. B . This axial T2WI (TR 3,200; TE 100) at the level of the malleolar fossa reveals a thickened anterior talofibular ligament that has a disruption (arrow). The marked thickening of the ligament indicates a chronic process.
P.1221 P.1222 Anomalous muscles in the foot or ankle are reported to be present in
3185
up to 6% of the population. These can be mistaken for a tumor and biopsy may be performed unnecessarily. MR will show these “tumors― to have imaging characteristics identical to those of normal muscle (Fig. 50.12) and to be sharply circumscribed. Accessory soleus and peroneus quartus muscles are the most common accessory muscles encountered around the foot and ankle.
LIGAMENTS MR is not the best way to diagnose acute ankle ligament abnormalities. The clinical evaluation is usually straightforward and no diagnostic imaging of any type is necessary. Nevertheless, in clinically equivocal cases or when the examination is ordered for other reasons, the ligaments can be clearly evaluated with highquality MR in most instances (8) . The deltoid ligament lies medially as a broad band beneath the tendons. Although often seen on coronal images deep to the posterior tibial tendon, it has a variable anatomic appearance. Injury to the deltoid ligament accounts for only 5% to 10% of ankle ligament sprains. The lateral ligaments are injured in more than 90% of ankle sprains. The lateral complex is made up of two parts: P.1223 a superior group, i.e., the anterior and posterior tibiofibular ligaments that make up part of the syndesmosis (Fig. 50.13); and an inferior group, i.e., the anterior and posterior talofibular ligaments and the calcaneofibular ligament (Fig. 50.14). The anterior and posterior tibiofibular ligaments can be seen on axial images at or slightly below the tibiotalar joint. The anterior and posterior talofibular ligaments are seen on axial images just below the tibiotalar joint and emanate from a concavity in the distal fibula called the malleolar fossa (Fig. 50.14B). The most commonly torn ankle ligament is the anterior talofibular ligament. It is easily identified when a joint effusion is present, because it makes up the anterior capsule of the joint (Fig. 50.15). The anterior talofibular ligament is usually torn without other ligaments being involved;
3186
however, if the injury is severe enough, the next ligament to tear is the calcaneofibular ligament. Even with very severe trauma, the posterior talofibular ligament will rarely tear.
FIGURE
50.16. Sinus Tarsi Syndrome. A sagittal T1WI in a
patient with chronic lateral ankle pain shows absence of the normal fat in the sinus tarsi (arrows). This is virtually diagnostic of sinus tarsi syndrome, except in the setting of an acute ankle sprain.
3187
FIGURE 50.17. Anterolateral Impingement Syndrome. This axial T1WI (A) through the ankle reveals absence of the anterior talofibular ligament (arrow). The corresponding T2WI (B) shows low-signal scar tissue deep to the expected location of the anterior talofibular ligament (arrow), which indicates anterolateral
impingement
syndrome.
3188
FIGURE 50.18. Tarsal Coalition. An axial T1WI in a patient with painful flat feet shows bilateral talocalcaneal coalition (arrows), which is primarily fibrous. The normal joint space is irregular and widened bilaterally. In cases of suspected coalition, both ankles should be imaged, as coalition often occurs bilaterally.
Several entities have a high association with chronic tears of the lateral ligaments. These include chronic lateral ankle instability, sinus tarsi syndrome, and anterolateral impingement syndrome. Patients with sinus tarsi syndrome present with lateral ankle pain and tenderness and a perception of hindfoot instability. The sinus tarsi is the cone-shaped space between the talus and the calcaneus that opens up laterally. It is a fat-filled space through which traverse several important ligaments that provide subtalar stability. In sinus tarsi syndrome, these ligaments are torn and the fat is replaced with granulation tissue or scar tissue. Hence, on T2WIs there may be high
3189
signal (granulation tissue) or low signal (scar), but on T1WIs there is low signal in the sinus tarsi (Fig. 50.16). In the acutely sprained ankle, the sinus tarsi may undergo replacement of the fat because of hemorrhage and edema, which will resolve. Anterolateral impingement syndrome results from hypertrophy and scarring of the synovium in the lateral P.1224 gutter of the ankle. The lateral gutter is the space between the tibia and fibula and is bound by the lateral ankle ligaments. Patients with anterolateral impingement syndrome present with lateral ankle pain and inability to dorsiflex normally. They often have a click on dorsiflexion. Arthroscopic resection of the scar tissue has been reported, with good results. T2WIs show low-signal tissue in the lateral gutter (Fig. 50.17). The anterior talofibular ligament is commonly torn or fibrosed in this condition.
FIGURE 50.19. Calcaneal Stress Fracture. A. A 70-year-old patient with a prior history of lung cancer presented with heel pain and a normal plain film. A bone scan showed diffuse increased radionuclide uptake throughout the posterior calcaneus. B . A sagittal T1WI revealed a linear area of low signal (arrows), which is characteristic for a stress fracture. Metastatic disease would not have this appearance.
3190
BONY
ABNORMALITIES
Tarsal coalition is a common cause of a painful flatfoot. It occurs most commonly at the calcaneonavicular joint and the middle facet of the talocalcaneal joint (Fig. 50.18). Up to 50% of patients with tarsal coalition have bilateral coalition. It can be difficult (or impossible) to see the coalition on plain films; however, CT and MR will show bony coalition with a high degree of accuracy. The coalition can also be fibrous or cartilaginous. In these cases, secondary findings, such as joint space irregularity at the affected joint or degenerative joint disease at nearby joints that are subjected to accentuated stress, can be seen. Fractures of the foot and ankle are usually well documented with plain films. Stress fractures, however, can be difficult to diagnose radiographically or clinically, and they can mimic more sinister abnormalities. MR will show stress fractures as linear low signal on T1WIs and high signal on T2WIs (Fig. 50.19) . MR has had mixed reviews when used for diagnosing osteomyelitis in the foot. In diabetic patients with foot infections, diagnosing osteomyelitis is important because the treatment is often much more aggressive—including amputation—than if the bone is not involved. If the marrow appears normal, MR is highly accurate in predicting no osteomyelitis; however, if low signal is present in the marrow around a joint, osteomyelitis may or may not be present. Low signal can be caused by edema or hyperemia without infection. The only definitive MR findings for osteomyelitis are cortical disruption, a bony abscess (not a common finding), or a sinus tract (an even less common finding). MR is therefore very sensitive but not very specific in diagnosing osteomyelitis in the foot and ankle (9) .
REFERENCES
3191
1. Anzilotti K, Schweitzer ME, et al. Effect of foot and ankle MR imaging on clinical decision making. Radiology 1996;201:515–517. 2. Quinn S, Murray W, Clark R, Cochran C. Achilles tendon: MR imaging at 1.5 T. Radiology 1987;164:767–770. 3. Rosenberg Z, Cheung Y, Jahss M, et al. Rupture of posterior tibial tendons: CT and MR imaging with surgical correlation. Radiology 1988;169:229–236. 4. Balen PF, Helms CA. Association of posterior tibial tendon injury with spring ligament injury, sinus tarsi abnormality, and plantar fasciitis on MR imaging. AJR Am J Roentgenol 2001;176:1137–1143. 5. Oden R. Tendon injuries about the ankle resulting from skiing. Clin
Orthop
1987;216:63–69.
6. Keigley B, Haggar A, Gaba A, et al. Primary tumors of the foot: MR
imaging.
Radiology
1989;171:755–759.
7. Erickson S, Quinn S, Kneeland J, et al. MR imaging of the tarsal tunnel and related spaces: normal and abnormal findings with anatomic correlation. AJR Am J Roentgenol 1990;155:323–328. 8. Erickson S, Smith J, Ruiz M, et al. MR imaging of the lateral collateral ligament of the ankle. AJR Am J Roentgenol 1991;156:131–136. 9. Erdman W, Tamburro F, Jayson H, et al. Osteomyelitis: characteristics and pitfalls of diagnosis with MR imaging. Radiology
1991;180:533–539.
3192
3193
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XI - Pediatric Radiology > Chapter 51 - Pediatric Chest
Chapter
51
Pediatric
Chest
Susan D. John Leonard E. Swischuk
ABNORMAL
LUNG
OPACITY
Pulmonary opacities in children are classified in the same way as in adults: as primarily alveolar or interstitial, focal or diffuse, and unilateral or bilateral. Some abnormalities occur in a central or parahilar distribution, whereas others are predominantly peripheral or basal in location. Mixed patterns also occur. An understanding of the causes of these various patterns is necessary to provide a useful interpretation of abnormal lung opacities in children.
Alveolar
Patterns
Alveolar consolidation occurs when the alveolar airspace is replaced by a substance, usually fluid. Focal consolidations most often represent exudates associated with bacterial pneumonia (Table 51.1). Bacterial consolidation begins as an oval, round, ill-defined, or fluffy area of solid opacification, often more peripheral than central in location. The pneumonia may progress to involve an entire lobe, but involvement of an entire lung is uncommon. Bacterial pneumonia is a space-occupying process within the lung and, therefore, little or no volume will be lost in the affected lung during the acute stage of infection (Fig. 51.1) .
3194
Streptococcus pneumoniae is the most common cause of lobar pneumonia throughout childhood. The incidence of Haemophilus influenzae pneumonia has dramatically decreased in the United States and other developed countries because of the use of the H . influenzae type b vaccine. Mycoplasma infections may also occasionally produce focal consolidating pneumonia. Consolidations with viral infections are not particularly common but can occur with more serious viral infection, such as adenovirus, influenza, parainfluenza, and respiratory syncytial virus. There is some question as to whether these consolidations represent true airspace consolidations. It is likely that they represent intense interstitial disease causing compression of the alveoli and mimicking the findings of airspace consolidations. The same may be true of Mycoplasma infections. Pneumonia caused by gram-negative bacilli is uncommon in children; it occurs primarily in infants and P.1228 immunocompromised children. Primary tuberculosis should be considered when the infiltrate is accompanied by hilar lymphadenopathy. Other causes of isolated lung consolidation in children include fungal infection, pulmonary infarction, lung contusion, and focal pulmonary hemorrhage.
TABLE 51.1 Causes of Focal Alveolar Consolidation
3195
Bacterial pneumonia Streptococcus pneumoniae Mycobacterium Staphylococcus Haemophilus influenzae Nonbacterial infection Tuberculosis Actinomycosis Pulmonary infarction Pulmonary
contusion
Atelectasis is a common occurrence in children, especially those with bronchial disease such as acute viral respiratory tract infections, reactive airway disease, and asthma. Atelectasis can sometimes resemble a bacterial consolidation. The findings of volume loss, such as shift of the fissures or the mediastinum, help to distinguish atelectasis from bacterial consolidation. Generally, volume loss will not be seen with a bacterial pneumonia until it begins to resolve. A flattened or linear shape in a pulmonary opacity should also suggest that it represents atelectasis rather than consolidation (Fig. 51.2) . Atelectasis is particularly problematic in children with asthma, who are also at increased risk for bacterial pneumonia. Clinical information may be necessary to help distinguish atelectasis from pneumonia in such children. Opacities seen in a child with acute asthmatic exacerbation but without high fever, chest pain, or leukocytosis are much more likely to be caused by atelectasis than pneumonia. Multiple patchy lung opacities is a pattern seen in a wide variety of conditions (Table 51.2). Such opacities reflect filling of the alveolar space with exudates, edema, or blood. Multiple bilateral alveolar infiltrates suggest bacterial infection (most commonly staphylococcal) (Fig. 51.3) or fungal disease. Opportunistic infections in immunocompromised patients are much more likely to be multiple and bilateral. Aspiration pneumonia also tends to present with multiple patchy pulmonary opacities. The
3196
P.1229 pneumonitis associated with hydrocarbon ingestion typically occurs in the medial portions of the lung bases (Fig. 51.4). Other less common causes of patchy alveolar opacities include milk allergy, hypersensitivity pneumonitis, uremic lung disease, near and pulmonary hemorrhage (i.e., idiopathic pulmonary hemosiderosis).
FIGURE
51.1. Bacterial
Pneumonia.
drowning,
A. Frontal view. B .
Lateral view. A typical alveolar consolidation in the right upper lobe. Note that the fissures are not displaced, indicating that there is little volume loss. A right pleural effusion is also present.
3197
FIGURE
51.2. Acute Asthma. A. Opacity silhouettes the right
heart border on the posteroanterior view. B . Lateral view shows displacement of the horizontal and oblique fissures (arrows) , indicating right middle lobe atelectasis.
TABLE 51.2 Sources of Multiple Patchy Lung Opacities
3198
Infection Staphylococcus Mycoplasma Fungal Opportunistic Aspiration
organisms
Hydrocarbon ingestion Near drowning Immune-mediated pneumonitis Milk allergy Hypersensitivity pneumonitis Pulmonary hemorrhage Pulmonary edema
Peribronchial
and
Interstitial
Patterns
The vast majority of upper respiratory tract infections in childhood are viral in nature and primarily bronchial in location. Such infections may result in pulmonary opacities that differ significantly from those seen with bacterial pneumonia.
3199
FIGURE 51.3. Staphylococcal Pneumonia. Note the typical multiple, bilateral alveolar opacities.
FIGURE
51.4. Hydrocarbon
Aspiration. Typical patchy alveolar
opacities are seen in the lung bases bilaterally in this child who ingested
kerosene.
P.1230 Parahilar peribronchial opacities are sometimes seen and are the result of peribronchial inflammation and edema associated with bronchitis (Table 51.3) (Fig. 51.5A). The pattern consists of bilateral, ill-defined, hazy soft tissue opacity in the hilar region of the lungs. When extensive, these opacities may cause a “shaggy― appearance to the cardiac borders (Fig. 51.5B). Acute peribronchial opacities are most often caused by viral respiratory infections (1,2) . Bilateral hilar adenopathy and scattered areas of subsegmental atelectasis are common associated findings (Fig. 51.6). This pattern is very different from the more peripheral alveolar opacification that is usually seen with bacterial pneumonias. However, it should be noted that a superimposed consolidating bacterial pneumonia can
3200
develop later in the course of a viral lower respiratory tract infection. M. pneumoniae and pertussis infections also commonly produce this pattern (3). Follicular bronchitis, associated with proliferation of lymphoid follicles along the airways, is indistinguishable radiographically. Chlamydia trachomatis infection has a similar appearance and usually occurs just after the newborn period (Fig. 51.7). Chronic bronchial inflammation associated with conditions such as asthma, cystic fibrosis (4) (see Fig. 51.18), immunologic deficiency diseases, and recurrent aspiration may result in persisting patterns of parahilar peribronchial opacity and may eventually lead to bronchiectasis.
FIGURE 51.5. Viral Infection. A. Bilateral parahilar peribronchial opacities are typical of viral lower respiratory tract infections. B . More pronounced inflammatory edema produces dense parahilar regions, leading to the “shaggy heart― appearance.
3201
TABLE 51.3 Causes of Parahilar Peribronchial Opacity
Acute
(infection)
Viral Mycoplasma Chlamydia Pertussis Chronic Asthma Cystic fibrosis Immunologic deficiency Chronic aspiration
disease
3202
FIGURE 51.6. Viral Lower Respiratory Tract Infection With Atelectasis. A. Note the ill-defined bilateral parahilar peribronchial opacities and vague focal opacity at the right heart border. B . On the lateral view, shift of the fissures (arrows) toward the right middle lobe opacity indicates volume loss (atelectasis) C . The common peribronchial opacities are accompanied by elevation of the horizontal fissure (arrows) .
3203
FIGURE
51.7. Chlamydia
Pneumonitis. Prominent bilateral
peribronchial opacities with slight nodularity are seen in the lung bases. The appearance is similar to that seen with viral infections.
P.1231 Hazy, reticular, or reticulonodular opacities that occur diffusely in the lungs indicate interstitial lung pathology, and the causes include many of the same conditions that cause parahilar peribronchial opacities (Table 51.4). The most common cause of an interstitial pattern in the lungs of a child is viral or Mycoplasma infection (Fig. 51.8). In general, bacterial infections of the lung do not have this appearance, except in the neonate, when bacterial pneumonia can present as diffuse haziness or reticulonodularity. Infections with fungi, such as Histoplasma capsulatum and Coccidioides pattern.
immitis, can also occasionally result in an interstitial
Pulmonary edema, when it is confined to the interstitial space, often produces a hazy or reticular pattern in the lungs. Cardiogenic pulmonary edema occurs when the pulmonary venous pressures are
3204
elevated because of left-sided myocardial failure or congenital lesions that impede blood flow through the left side of the heart (e.g., pulmonary vein atresia, cor triatriatum, hypoplastic left heart syndrome). P.1232 Noncardiogenic causes of pulmonary edema predominate in children. One of the most common causes of pulmonary edema in children is acute glomerulonephritis (Fig. 51.9). Sodium and fluid retention leads to hypervolemia, which can then result in cardiomegaly and pulmonary vascular P.1233 congestion with edema. The radiographic appearance can be indistinguishable from that of edema caused by cardiac failure. Other noncardiogenic causes of pulmonary edema in children include near drowning, increased intracranial pressure, inhalation injuries, drug overdose, and ARDS.
TABLE 51.4 Conditions Causing Hazy, Reticular, or Reticulonodular Patterns
Infection Viral Mycoplasma Fungal Pulmonary Heart Acute
edema
disease renal failure
Near drowning Increased intracranial Inhalation injury Drug overdose
pressure
“Acute― respiratory distress syndrome (ARDS) Pulmonary lymphangiectasia/hemangiomatosis Idiopathic pulmonary hemosiderosis Interstitial pneumonitis
3205
Langerhans cell histiocytosis Tuberous sclerosis Connective tissue diseases Lymphocytic infiltrative Malignancy Leukemia/lymphoma Lymphangitic
disease
metastasis
FIGURE 51.8. Viral Lower Respiratory Tract Infection. A. A mild, diffuse, reticulonodular pattern is present in this patient with respiratory syncytial virus infection. B . Note the prominent reticulonodular pattern caused by herpes pneumonia in an immunosuppressed
patient.
3206
FIGURE
51.9. Acute
Glomerulonephritis. Note marked
bilateral vascular congestion edema. The heart is mildly enlarged.
Pulmonary lymphangiectasia is a rare condition that consists of dilated lymphatic channels secondary to either abnormal embryonic development of the lymphatic system or obstruction of lymphatic drainage. The dilated lymphatics cause a coarsely nodular or reticular pattern in the lungs, usually developing early in infancy (Fig.
51.10A)
(5,6). Pulmonary hemangiomatosis is a similar rare condition. Recurrent hemorrhage into the lungs in patients with idiopathic pulmonary hemosiderosis eventually leads to chronic diffuse haziness or reticula in the lungs, representing pulmonary fibrosis (Fig. 51.10B, C). Langerhans cell histiocytosis (LCH) causes an interstitial pattern that often is more prominent in the upper lung zones. The lung volumes in LCH are normal or increased, which differs from fibrotic conditions, in which lung volumes tend to be decreased. HRCT may show cysts and nodules (7) .
3207
FIGURE 51.10. Interstitial Patterns. A. Pulmonary lymphangiectasia. Note the diffuse reticulonodular pattern throughout both lungs caused by dilated lymphatics in the interstitium. Dextrocardia is also present. B . Another patient with a fine reticular pattern in the lungs, caused in this case by idiopathic pulmonary hemosiderosis. C . The same patient after an episode of acute hemorrhage.
Interstitial lung disease that predominates in the lower lobes can be seen with tuberous sclerosis, connective tissue diseases, and primary interstitial pneumonitis. Leukemia, lymphoma, and lymphatic
3208
metastases to the lungs can also cause a reticular or reticulonodular infiltrative pattern. Mycoplasma pneumonitis sometimes presents as an interstitial pattern confined to one lobe of the lung (3) .
FIGURE 51.11. Miliary Tuberculosis. The numerous tiny nodules in the lungs of this immunosuppressed patient represent hematogenous dissemination of tuberculosis.
P.1234
3209
Miliary nodules usually consist of tiny nodules (smaller than 5 mm) that are randomly distributed throughout the lungs. This pattern in children is most often caused by hematogenous dissemination of tuberculosis or histoplasmosis (Fig. 51.11), although viral pneumonitis, idiopathic pulmonary hemosiderosis, and metastatic disease can also have this appearance (Table 51.5). The tiny nodules can be difficult to see on radiographs in some cases, and CT can better define the nodules and other associated abnormalities, such as lymphadenopathy (8). Acute disseminated tuberculosis in infants and young children can sometimes produce larger nodules, and CT may show larger areas of opacity caused by coalescent nodules and interstitial thickening.
TABLE 51.5 Causes of Miliary Nodules
Infection Tuberculosis Histoplasmosis Viral Idiopathic Metastatic
pulmonary disease
hemosiderosis
3210
FIGURE 51.12. Lymphocytic Interstitial Pneumonitis. Note the diffuse interstitial pattern in the lungs of an HIV-positive child.
Opportunistic infections may occur in children with HIV infection and other forms of congenital or acquired immunodeficiency. Infection with common viral, bacterial, and fungal organisms creates a pattern similar to that seen in immunocompetent children, but the findings tend to be more rapidly progressive and more pronounced. Lymphocytic infiltrative disease (LIP) produces a reticulonodular pattern that is indistinguishable from infection (Fig. 51.12), except for its chronicity.
ABNORMAL
LUNG
VOLUME
Pulmonary aeration abnormalities are best evaluated on the chest radiograph by observing the following criteria: (1) the relative size of
3211
a lung or hemithorax, (2) the degree of radiolucency of the lung, and (3) the pulmonary vascularity or blood flow to the lung. Bilateral smallness of the lungs is commonly caused by less than complete inspiration. The technical difficulties of obtaining good inspiratory chest films in children are significant. The lungs may appear small if the diaphragm is elevated, because of either neuromuscular abnormality or the presence of large masses, fluid collections, or bowel distension in the P.1235 abdomen. Infrequently, inspiratory obstruction of the trachea can lead to bilateral underaeration of the lungs. Causes of such obstruction include intratracheal masses or foreign bodies, or extrinsic compression of the trachea by anomalous vascular structures. A hyperlucent but small hemithorax usually signifies some degree of pulmonary hypoplasia, either congenital or acquired.
Pulmonary
Hypoplasia
or
Agenesis
Congenital pulmonary hypoplasia is associated with hypoplasia or absence of the ipsilateral pulmonary artery (9); thus, pulmonary vascular markings will be diminished in size on radiographs. Congenital lung hypoplasia is sometimes associated with congenital heart disease, most often tetralogy of Fallot or persistent truncus arteriosus. In cases of tetralogy of Fallot, the left lung is hypoplastic. A hypogenetic lung is one of the features of congenital pulmonary venolobar (scimitar) syndrome. Other variable components of this syndrome include partial anomalous venous return, hypoplasia or absence of the pulmonary artery, pulmonary sequestration, systemic arterialization of the lungs, accessory diaphragm, and absent inferior vena cava (1 0). Pulmonary agenesis is a rare anomaly that results from an insult during the fourth week of fetal life. The right and left lung are affected with equal frequency. Right pulmonary agenesis has an increased association with other congenital malformations involving the heart, skeleton, GI tract, and genitourinary tract. Chest radiographs or CT demonstrate severe volume loss and opacity on the side of agenesis, often with close spacing of the ribs. The bronchus and pulmonary artery to the affected lung are absent.
3212
Pulmonary hypoplasia in the neonate can be unilateral or bilateral. Bilateral pulmonary hypoplasia is most often the result of compression of the lungs during fetal development. Congenital dysplasias and syndromes associated with short ribs and a small thoracic cage (asphyxiating thoracic dystrophy, thanatophoric dwarfism, Ellis–van Creveld syndrome) compress the lungs and cause hypoplastic lungs (Fig. 51.13). The degree of hypoplasia is often severe and leads to the demise of these infants. Chromosomal abnormalities such as the trisomies are associated with hypoplastic lungs, and in some infants, hypoplasia is “primary― and unexplained. The most common cause of intrathoracic compression of the fetal lungs is congenital diaphragmatic hernia. Although the hernia itself is most often unilateral, the increased volume of the thorax on the side of the hernia causes compression of the contralateral lung, resulting in bilateral and asymmetric lung hypoplasia (Fig. 51.14). The degree of hypoplasia varies in severity; the earlier in gestation that the hernia occurs, the more severe the lung hypoplasia. Pulmonary insufficiency is the most significant cause of morbidity and mortality in these infants. Infants with less severely hypoplastic lungs can be supported with artificial ventilation or extracorporeal membranous oxygenation (ECMO) until their lungs develop enough to permit survival. Other causes of intrathoracic compression leading to bilateral pulmonary hypoplasia include bilateral chylothorax, large intrathoracic cysts or tumors (neuroblastoma, teratoma, cystic adenomatoid
malformation),
or
marked
3213
cardiomegaly.
FIGURE 51.13. Lung Hypoplasia. The very small thoracic cage caused by rib shortening in this thanatophoric dwarf is associated with marked lung hypoplasia.
3214
FIGURE
51.14. Congenital
Diaphragmatic
Hernia. Multiple
air- and fluid-filled loops of bowel in the left hemithorax displace the mediastinum into the right hemithorax. Note the small, hypoplastic left lung (arrows) .
3215
FIGURE
51.15. Swyer-James
Lung. Note that the left lung is
small and relatively hyperlucent when compared with the right lung. The left pulmonary vascularity is decreased.
P.1236 Extrathoracic compression of the fetal lungs is most often caused by oligohydramnios secondary to fetal urinary tract abnormalities or by abnormal amniotic fluid production or leakage. Potter syndrome, associated with bilateral renal agenesis, congenital renal cystic disease, or obstructive uropathy (posterior urethral valves, prune belly syndrome), commonly results in hypoplastic lungs. Additional causes include neuromuscular abnormalities with persistent elevation of the diaphragm or prolonged distension of the abdomen by large abdominal masses or ascites. Swyer-James Syndrome is an acquired hypoplastic lung that develops following severe obliterative bronchiolitis, leading to
3216
bronchiolar obstruction, bronchiectasis, and distal airspace destruction (Fig. 51.15). Bronchiectasis is not present in all cases (1 1). Air enters the lung by air drift phenomenon but becomes trapped because of the bronchiolar obstruction. Air trapping results in a lung that changes very little in size between inspiration and expiration. This important feature helps to distinguish the hypoplastic Swyer-James lung from the congenitally hypoplastic lung. Radionuclide ventilation/perfusion studies can be used to verify the expiratory airway obstruction as well as the diminished perfusion of the hypoplastic lung. CT is more sensitive than radiographs in detecting areas of air trapping and helps to exclude other causes of central bronchial obstruction (1 2). Although the Swyer-James lung is classically clear and hyperlucent, some patients show a fibrotic reticular pattern in the hypoplastic lung. Other causes of a unilateral small reticular lung include scarring after radiation therapy or congenital unilateral pulmonary vein atresia or stenosis (Fig. 51.16) . The reticula of the lung in pulmonary vein atresia are caused by a combination of interstitial pulmonary edema, fibrosis, and dilated interstitial lymphatics.
3217
FIGURE 51.16. Unilateral Pulmonary Vein Atresia. The right lung is small with diffuse reticula, very likely caused by fibrosis from prolonged pulmonary edema and/or infection in this patient with pulmonary vein atresia on the right.
Bilateral
Lung
Hyperinflation
Bilateral overaeration of the lungs is most often caused by airway obstruction that can be central or diffuse and peripheral (Table 51.6) .
Small
Airway
Obstruction
Widespread obstruction of the peripheral airway is a common cause of obstructive emphysema and is most often the result of viral bronchitis and bronchiolitis or asthma. Acute bronchiolitis in infants often is accompanied by severe air trapping and overinflation of the lungs, with little or no other visible pulmonary abnormality. Hyperinflation tends to be less severe in older children with viral lower respiratory tract infections, but the mechanism (i.e., mucosal edema and bronchospasm secondary to inflammation) is the same. Infants with cystic fibrosis can present with an appearance identical to that of bronchiolitis. Cystic fibrosis should be considered in any infant who presents with multiple episodes of bronchiolitis (Fig. 51.17). Peripheral small airway obstruction with parahilar peribronchial opacities is seen with certain immunologic deficiency diseases, chronic aspiration, and graft versus host disease. Central airway obstruction leading to bilateral overaeration of the lungs is less common than peripheral P.1237 obstruction.
Intratracheal
foreign
bodies,
neoplasms,
granulomas,
and intrinsic stenoses of the trachea are all rather rare. More commonly, tracheal obstruction is the result of extrinsic compression caused by cysts, neoplasms, adenopathy, and congenital vascular abnormalities.
3218
TABLE 51.6 Possible Causes of Bilateral Lung Hyperinflation
Diffuse Viral
peripheral obstruction bronchitis/bronchiolitis
Asthma Cystic fibrosis Immunologic deficiency diseases Chronic aspiration Graft versus host disease Central obstruction Extrinsic Vascular
anomalies
Mediastinal masses Intrinsic Tracheal foreign body Tracheal
neoplasm/granuloma
A right-sided aortic arch is the key radiographic clue to the presence of an obstructing vascular ring (Fig. 51.18). Most cases of double aortic arch consist of a large posterior, right-sided arch and a small anterior, left-sided arch that encircle the esophagus and trachea. The diagnosis may be verified by barium esophagram, which shows a reverse S configuration caused by the bilateral vascular impressions on the esophagus (Fig. 51.19). Lateral radiographs demonstrate increased retrotracheal opacity, tracheal narrowing, and anterior tracheal bowing (1 3). Similar radiographic abnormalities are seen when the vascular ring consists of a right ascending aortic arch, an aberrant left subclavian artery that passes posterior to the esophagus, and a ligamentum arteriosum or persistent ductus arteriosus stretching from the left subclavian artery to the pulmonary artery anterior to the trachea. Definitive evaluation of the vascular anatomy is best accomplished with MR or helical CT (1 4,1 5) (see Fig. 51.18) .
3219
FIGURE 51.17. Cystic Fibrosis. A. The early stages of this disease are often manifest only by bilateral peribronchial opacities and hyperinflation of the lungs. This appearance resembles that seen with viral lower respiratory tract infections or bronchiolitis. B . Here, the characteristic changes of late-stage cystic fibrosis are seen, including bronchial wall thickening, bronchiectasis, and persistent atelectasis.
The pulmonary sling anomaly is a rare condition that may also result in tracheal compression and bilateral hyperaeration of the lungs. The left pulmonary artery arises from the right pulmonary artery, and as it courses to the left lung, the left pulmonary artery passes between the trachea and esophagus and compresses the trachea posteriorly (Fig. 51.20). The right lung may be either underaerated or overaerated.
Asymmetric/Unilateral Abnormalities
Aeration
Pulmonary aeration abnormalities are frequently asymmetric or unilateral. A large, hyperlucent hemithorax most often indicates overinflation of an entire lobe or lung. Such hyperaeration may represent
obstructive
emphysema
(Table 51.7) or compensatory
3220
overinflation resulting from decreased volume of the contralateral lung. Pulmonary vascularity is the key to differentiation. Obstructive emphysema generally results in diminished size of the pulmonary vessels because of compression and hypoxia-induced P.1238 reflex arterial spasm. With compensatory hyperinflation, the pulmonary vessels are normal or even increased in size. If doubt remains, inspiratory and expiratory frontal views of the chest or fluoroscopy can be helpful. The lung that changes the least in volume between inspiration and expiration is the abnormal lung (Fig. 51.21) . This holds true whether the lung is obstructed and overinflated or is small because of atelectasis or hypoplasia. The only exception to this rule is mild congenital pulmonary hypoplasia, which is associated with relatively normal lung dynamics.
FIGURE 51.18. Double Aortic Arch. A. Note the large rightsided aortic arch displacing the trachea to the left (arrows). A smaller nodular opacity is seen to the left of the tracheal shadow. B . Coronal MR clearly defines the large right aortic arch (arrow) and the smaller left arch (arrowhead), which encircle and compress the trachea.
3221
FIGURE
51.19. Double Aortic Arch. The encircling aortic
arches compress the esophagus, creating a reverse S configuration (arrows). Note that the right arch is higher and more prominent than the left arch.
Congenital lobar emphysema consists of obstructive emphysema of a single lobe of the lung, most commonly the left upper, right middle, or right upper lobe. Usually emphysema is the result of underdevelopment of the segmental bronchial cartilage, which leads
3222
to expiratory airway collapse and a ball valve type of obstruction. Early in the newborn period, the obstructed lobe may be opaque because of delayed clearance of fluid distal to the obstruction. Gradually, the fluid clears, and the involved lobe is filled with air and overinflated (Fig. 51.22). A severely enlarged emphysematous lobe can occupy the entire hemithorax and erroneously suggest a pneumothorax or emphysema of the entire lung. Careful inspection reveals the collapsed, compressed adjacent lobes of the lung, confirming the diagnosis. A similar appearance can occur in the rare case of bronchial stenosis or atresia. Classically, an oval opacity is seen adjacent to the overinflated lung, near the hilum, representing a collection of mucus (mucocele) within the obstructed bronchus. Acquired lobar emphysema can occur as a result of bronchial damage associated
with
inflammatory
conditions
or
bronchopulmonary
dysplasia (BPD) (Fig. 51.23) .
Endobronchial
Lesions
In older infants and children, obstructive emphysema is most often caused by an P.1239 endobronchial foreign body or a mucous plug (see Fig. 51.21) . Mucous plugs occur most commonly in asthmatics and in children with viral lower respiratory tract infections. Other less common causes of unilateral obstructive emphysema include endobronchial masses such as tuberculous granulomas (Fig. 51.24) and extrinsic compressing lesions such as anomalous blood vessels and mediastinal tumors and cysts. Pneumothorax may cause a large hyperlucent hemithorax that mimics obstructive emphysema. In supine patients, the pleural air may lie entirely along the anterior surface of the lung and no free lung edge will be visible. Clues to the presence of an anterior pneumothorax include increased radiolucency of the hemithorax and increased sharpness of the mediastinal border (Fig. 51.25). In newborns, a pneumothorax can compress the normal thymus gland, creating a mediastinal “pseudomass.― Rarely, an air-filled lung cyst or pneumatocele, or a markedly dilated stomach in a diaphragmatic hernia, can occupy an entire hemithorax,
3223
rendering
FIGURE
it
hyperlucent.
51.20. Aberrant Left Pulmonary Artery (Pulmonary
Sling). A. Note slight smallness of the right lung. B . MR study demonstrates the typical course of the aberrant left pulmonary artery (arrow). Ascending and descending aorta (a), pulmonary artery (p). (From Swischuk LE. Imaging of the Newborn, Infant, and Young Child. 5th ed. Baltimore: Williams & Wilkins, 2004:315; used with permission.)
TABLE 51.7 Causes of Unilateral Obstructive Emphysema
Bronchial foreign Mucous plug Congenital Bronchial
body
lobar emphysema stenosis/atresia
Tuberculosis Vascular anomalies Mediastinal masses
3224
PULMONARY
CAVITIES
Cavities in the lungs of children are most often inflammatory or postinflammatory. Lung abscesses usually develop as a complication of a bacterial pneumonia and can be solitary or multiple. The wall of an abscess is thick and irregular, and some contain air–fluid levels (Fig. 51.26) . CT is valuable for distinguishing intrapulmonary abscess from loculated empyema in the pleural space (Fig. 51.27). An abscess may form distal to a bronchial obstruction, which in children is often caused by a retained foreign body. Cavitary tuberculosis is rare in childhood, and echinococcal cysts are rare outside of endemic areas. Pneumatoceles are thin-walled lung cavities that commonly occur with pulmonary infections in children. Staphylococcal pneumonia is classically associated with pneumatocele formation, although they can occur with other infections, including tuberculosis (Fig. 51.28) . Pneumatoceles develop from a bronchiolar obstruction that leads to air trapping and alveolar rupture. Pneumatoceles can become large and cause significant mass effect. More often, the pneumatoceles remain relatively small and resolve spontaneously. Occasionally, pneumatoceles rupture, leading to pneumothorax or pneumomediastinum. Other causes of pneumatoceles in children include blunt chest trauma, hydrocarbon pneumonitis, and LCH. Congenital lung cysts are uncommon and may be indistinguishable from pneumatoceles. Congenital cysts are usually thin walled and more commonly occur in the lower lobes. Most are asymptomatic unless they become infected P.1240 or undergo rapid expansion with the development of a tension phenomenon.
3225
FIGURE 51.21. Bronchial Foreign Bodies on the Right. A. On inspiration, the right lung is slightly larger and more radiolucent than the left lung. B . With expiration, the left lung decreases in size, but the right lung remains overinflated. This indicates obstructive emphysema on the right, in this case caused by pieces of walnut lodged in the right mainstem bronchus.
Cystic adenomatoid malformation is a congenital lesion of the lung characterized by abnormal lung tissue containing dysplastic
3226
adenomatous tissue within communicating cysts of variable sizes. The radiographic appearance can vary from a predominantly solid lesion with multiple tiny cysts to multiple, large, thin-walled cysts that mimic congenital lobar emphysema (Fig. 51.29) (1 6). In the first days of life, the cysts are fluid filled and the lesion has the appearance of a solid mass. With time, air replaces the fluid, and small radiolucent cysts become apparent. The cysts gradually enlarge and can cause enough mass effect to lead to respiratory distress. The malformation is usually unilateral and can affect any portion of the lung.
Congenital
Diaphragmatic
Hernia
Air-filled loops of bowel in a congenital diaphragmatic hernia can resemble the multiple cysts of cystic adenomatoid malformation. An important clue to the correct diagnosis of diaphragmatic hernia is the absence or paucity of gas-filled bowel loops within the abdomen (Fig. 51.30). Congenital diaphragmatic hernias most often occur through the foramen of Bochdalek, which lies posteriorly and medially in each hemidiaphragm. Left-sided hernias are more common and more frequently involve bowel herniation. Solid abdominal viscera are more likely to herniate into the chest through right-sided hernias (Fig. 51.30B). Hernias through the foramen of Morgagni, which lies anteriorly, are less common and usually are less severe.
3227
FIGURE 51.22. Congenital Lobar Emphysema. A. Initially, the left upper lobe was large and hazy because fluid was trapped in the obstructed lung. B . A later film shows that the fluid has cleared, leaving a typical overinflated left upper lobe, with compressive atelectasis of the left lower lobe.
Infants with large diaphragmatic hernias usually present with severe respiratory distress immediately after birth. Compression of the
3228
ipsilateral lung in utero causes it to be hypoplastic, and often the contralateral lung is also P.1241 small. The patients are profoundly hypoxic, and persistent fetal circulation caused by hypoxia-induced pulmonary hypertension usually further compromises the infant's condition. Even with early diagnosis and surgery, the mortality of this condition remains high. ECMO has improved the survival of some patients by circumventing the problem of pulmonary hypertension and the right-to-left shunting of blood away from the lungs. Congenital diaphragmatic hernia may be minimally symptomatic at birth and can present later in life.
FIGURE 51.23. Acquired Lobar Emphysema. CT shows marked hyperinflation of the right lower lobe secondary to severe bronchopulmonary dysplasia. The infant was able to be extubated after
lobectomy.
3229
FIGURE
51.24. Primary
Tuberculosis. Note the left hilar
adenopathy and obstructive emphysema of the left lung, caused by tuberculous granulomas of the left main bronchus.
FIGURE 51.25. Bilateral Pneumothorax in a Neonate. The anterior pneumothoraces cause increased radiolucency along the
3230
mediastinum and increased sharpness of the heart border.
FIGURE 51.26. Pulmonary Abscess. A. A typical bacterial pneumonia is seen in the right middle lobe. B . Later, cavitation has developed in the central portion of the pneumonia, and an air–fluid level is evident. This is a typical appearance for a lung abscess.
3231
FIGURE 51.27. Pulmonary Abscess. A. Radiograph shows a solid-appearing mass in the right lower lobe. B . CT demonstrates the intrapulmonary location of this large abscess (arrows) .
P.1242
3232
LUNG DISEASE IN THE NEONATE The conditions leading to respiratory distress in the newborn infant are numerous and can be divided into those that can be treated medically and those that require surgical intervention. Surgical conditions consist primarily of congenital and developmental abnormalities that result in a space-occupying lesion within the chest (diaphragmatic hernia, congenital lobar emphysema, chylothorax, pneumothorax, cystic adenomatoid malformation). This deal with diffuse pulmonary disease of the newborn. Surfactant
deficiency
section
will
disease (hyaline membrane disease) is one
of the most common causes of respiratory distress in the newborn (1 7,1 8). It is most common in premature infants; however, it occasionally occurs in full-term infants of diabetic mothers. In both cases, lung immaturity is the main predisposing factor. The primary abnormality is a lack of surfactant normally produced by the type II alveolar cells. This substance is responsible for decreasing the surface tension of the alveoli. When absent, the alveoli are poorly distensible and remain collapsed. A cycle of hypoxia, acidosis, and diminished perfusion results. Clinically, these infants present with respiratory distress within the first few hours after birth. The classic radiographic findings of surfactant deficiency disease consist of lungs that are small in volume and have a finely granular pattern, with air bronchograms that extend into the lung periphery (Fig. 51.31). The granular pattern reflects the histologic findings of distended alveolar ducts and terminal bronchioles superimposed over generalized alveolar collapse. When the alveoli and terminal bronchioles overdistend, small, round, 1- to 2-mm bubbles result. During expiration, the air bronchograms and granular pattern disappear and the lungs become totally opaque. With surfactant therapy, these changes are very transient. Similar lung opacities can be seen with neonatal pneumonia, pulmonary lymphangiectasia, neonatal retained fluid syndrome, and congenital heart abnormalities associated with severe pulmonary venous obstruction. However, unlike patients with hyaline membrane disease, the lung volumes in these conditions are normal to
3233
increased (Fig. 51.32). In a few cases of neonatal pneumonia, the lung pattern is indistinguishable from that seen in surfactant deficiency. Until recently, the primary form of therapy of this condition consisted of positive pressure–assisted ventilation, which attempts to force air deeper into the respiratory tree and alveoli. Although in some patients the use of assisted ventilation significantly improves oxygenation, in others, the elevated airway pressures result in complications caused by air leakage from the distended terminal airways. Air dissects through the interstitium and lymphatics (pulmonary interstitial emphysema), creating a radiographic pattern of serpiginous bubbles that extends all the way to the lung periphery (Fig. 51.33). Unlike the air bronchograms of uncomplicated hyaline membrane disease, the bubbles of pulmonary interstitial emphysema do not collapse upon expiration. Pneumomediastinum and pneumothorax are other common complications of positive pressure ventilation. Air also can dissect into the pericardium and peritoneum, and occasionally air embolism can develop, with devastating consequences. Early surfactant therapy has significantly reduced the incidence of these complications. The hypoxemia associated with the respiratory distress syndrome sometimes leads to persistent patency of the ductus arteriosus. Often, radiographic changes are the first clue to this complication. Suggestive findings include lungs that are large and increasingly opaque with loss of the granular pattern, cardiomegaly, and pulmonary vascular congestion (Fig. 51.34). The P.1243 increased lung opacity represents pulmonary edema. Poor renal function and neurogenic pulmonary edema resulting from cerebral hypoxic injury and hemorrhage are common noncardiac causes of pulmonary edema in the premature infant.
3234
FIGURE
51.28. Postinflammatory
Pneumatocele.
A. Note the
multiple thin-walled cysts within this nearly resolved pneumococcal pneumonia. B . Multiple large pneumatoceles followed a viral infection in this HIV-positive child. C . Multiple small pneumatoceles are seen within this resolving pneumonia.
Bronchopulmonary
Dysplasia
Continued use of positive pressure–assisted ventilation and high
3235
oxygen concentration damages the lung parenchyma and results in the condition known as BPD. Inflammation probably also plays a role in the development of BPD. Initially described in four stages, now most authors recognize an edematous phase and a bubbly phase. The initial edematous phase results from oxygen toxicity and hypoxia. Damage to the basement membrane of the capillaries causes them to leak fluid into the interstitium of the lungs. The lungs become hazy and in some cases even reticular (Fig. 51.35). The hazy pattern is the most common pattern encountered in premature infants and may persist for weeks or months. Because there is no dysplasia in this phase, it has been suggested that this phase be termed “leaky lung syndrome― (1 7,1 8). Pathophysiologically, this phase of the disease resembles ARDS. The pulmonary edema pattern can precede or occur simultaneously with the bubbly phase of BPD. However, in most cases, the conditions are somewhat separated. The bubbly phase results from the overdistension of some alveolar groups, while others remain atelectatic. Originally believed to be exclusively a late stage, it is now known that P.1244 bubbly lungs can occur early, even within days after birth (1 9). The problem can be seen even in patients who are born with clear lungs and who do not have surfactant deficiency disease. BPD is generally considered to be a result of damage to the structurally immature lung by oxygen and positive pressure ventilation.
3236
FIGURE
51.29. Cystic
Adenomatoid
Malformation.
A.
Multiple air-filled cysts of widely variable size expand the right lung and shift the mediastinum to the left. B . Another malformation that consists of a single thin-walled cyst is seen on CT. C . and D . Chest radiograph and CT in another patient reveal that the congenital malformation is predominantly solid or fluid filled, with multiple, small, air-filled cysts of varying sizes.
The edematous phase of BPD is treated with fluid restriction and diuretics. Vitamin A supplementation and low-dose dexamethasone may also decrease BPD; however, routine use of steroids is not advocated because of the potential risk of neuromotor and cognitive dysfunction later in life. The major predisposing factors for BPD are
3237
sepsis, very low birth weight, and young gestational age. The radiographic findings of advanced BPD consist of overaerated lungs with bubbles of varying sizes (Fig. 51.36). In other cases, the bubbly pattern is less pronounced, but pulmonary fibrosis and scattered areas of segmental atelectasis are seen. CT findings include reticular opacities, air trapping, and architectural distortion (2 0) (see Fig. 51.23) . Retained fetal lung fluid is the result of delayed clearance of the fluid normally present in the fetal lung. This condition, also known as wet lung disease, transient tachypnea of the newborn, and transient respiratory distress of the newborn, causes grunting and tachypnea in otherwise healthy term infants. The condition is particularly common in infants delivered by cesarean section, presumably caused by the lack of squeezing of the chest as it passes through the vaginal canal. In some cases, P.1245 the radiographic findings of retained fetal lung fluid are minimal, but commonly diffuse haziness or reticula is seen within the lungs. The symptoms and radiographic findings are transient and resolve within 24 to 48 hours (Fig. 51.37) .
3238
FIGURE
51.30. Congenital
Diaphragmatic
Hernia.
A. The left
hemithorax is filled with multiple air-filled loops of bowel, displacing the mediastinum to the right. The course of the nasogastric tube and the absence of normal bowel loops in the abdomen are additional clues to the diagnosis. B . Another infant with right-sided diaphragmatic hernia. What appears to be an elevated right hemidiaphragm is actually the liver extending into the chest.
Other conditions can produce these patterns (Table 51.8). A streaky parahilar appearance can occur with retained fluid that is similar to that seen with neonatal pneumonia; however, pneumonia characteristically progresses during the first few days of life. In other cases, the lung fluid may cause a granular pattern in the lungs that resembles surfactant deficiency disease. However, the lung volumes
3239
are generally normal to large with retained lung fluid, versus the small lung volumes seen with surfactant deficiency.
Meconium
Aspiration
Intrauterine fetal distress can lead to the passage of meconium, which can be aspirated into the tracheobronchial tree. Aspirated meconium particles cause obstruction of small peripheral bronchioles, resulting in unevenly distributed areas of subsegmental atelectasis with alternating areas of overdistension. This creates a coarse reticulonodular or nodular appearance of the lungs (Fig. 51.38). In severe cases, progressive air trapping results in complications such as pneumothorax and pneumomediastinum. The resultant hypoxia can lead to persistent fetal circulation, with right-to-left shunting across the foramen ovale. Treatment consists of endotracheal suctioning and the administration of humidified oxygen. ECMO may be required in severe cases. Pulmonary
lymphangiectasia is a rare condition that can occur as
an isolated abnormality or be associated with congenital heart disease or generalized lymphangiectasia. The isolated form is caused by abnormal pulmonary lymphatic development, resulting in dilated and obstructed lymphatic channels. Lymphangiectasia associated with congenital heart disease usually occurs with conditions leading to severe pulmonary venous obstruction (e.g., hypoplastic left heart syndrome, total anomalous pulmonary venous return type III, or pulmonary vein atresia). In both forms of the condition, the dilated lymphatics course through the lung interstitium, causing a diffuse reticular or reticulonodular pattern on radiographs (see Fig. 51.10) . The lungs often are hyperinflated and pleural effusions may occur.
3240
FIGURE 51.31. Surfactant Deficiency Disease. A. Shortly after birth, the lungs are small and diffusely opaque with air bronchograms that extend into the periphery of the lung. This is a typical appearance for surfactant deficiency disease. B . After treatment with endotracheal surfactant, lung volumes have dramatically improved and lung opacity has virtually disappeared.
P.1246 Extracorporeal membranous oxygenation is a widely used therapy to support infants with life-threatening respiratory disease.
3241
The technique consists of a bypass of the pulmonary blood flow through a semipermeable silicon membrane. The procedure interrupts the cycle of pulmonary hypertension and persistent fetal circulation (right-to-left shunting) and diminishes the damaging effect of high oxygen concentrations and barotrauma to the lungs. ECMO is commonly used in patients with congenital diaphragmatic hernia, meconium aspiration syndrome, neonatal sepsis, and pneumonia. Premature infants with surfactant deficiency disease are often too small for the large-caliber ECMO catheters; therefore, use of ECMO is limited for this condition. While on the extracorporeal circuit, the lungs invariably become opaque because the ventilator settings are reduced, allowing the lungs to collapse. Often, pleural effusions are present but may be obscured on chest radiographs by the opacity of the lungs. In such cases, the lungs often fail to reexpand despite increasing ventilator pressures. Shifting of the position of the ECMO catheters on radiographs should suggest an increased pleural fluid collection (2 1). US can be used to identify the pleural fluid and help distinguish blood from serous fluid.
3242
FIGURE
51.32. Neonatal
Pneumonia.
A. The lungs are
diffusely hazy, with a granular appearance that is similar to that seen with hyaline membrane disease. Note, however, that the lungs are normal in volume. B . Pneumonia in a different neonate has a more reticular appearance, with central alveolar opacities.
PLEURAL
THICKENING
AND
EFFUSIONS
Generalized thickening of the pleural space because of the
3243
accumulation of fluid has the same configurations in children as in adults. The most easily recognized pattern is thickening along the lateral and apical portions of the lung. Subpulmonic collections can mimic an elevated diaphragm, but characteristic flattening and laterally displaced curvature of the dome are clues to the presence of subpulmonic pleural fluid (Fig. 51.39). A totally opaque hemithorax of normal or increased volume nearly always P.1247 indicates a large collection of pleural fluid. Opacification of an entire lung by pneumonia is very unusual in children. However, occasionally a large cyst or intrathoracic mass can occupy most of the hemithorax. In such cases one should look for residual radiolucency in the costophrenic angle, which is not present when pleural fluid is the cause of total opacification of a hemithorax. The presence of pleural fluid is easily verified by US. The type of fluid in the pleural space (serous effusion, inflammatory exudate, chyle, or blood) cannot
be
reliably
determined
radiographically.
3244
FIGURE 51.33. Pulmonary Interstitial Emphysema. Serpiginous bubbles of interstitial air extend to the periphery of the left lung. The interstitial air causes the lung to be stiff and hyperexpanded, even during expiration.
Unilateral
pleural
effusions are most commonly associated with
pneumonia in the ipsilateral lung (Table 51.9). Such effusions are often transudates, but empyema is likely if the collection is large.
3245
Empyemas most often occur with staphylococcal, Haemophilus, and pneumococcal pneumonias. Empyema is characteristically loculated, P.1248 and the internal septations are easily verified with US (Fig. 51.40) . Prompt diagnosis of loculated empyema is important for successful treatment by video-assisted thoracoscopic surgery (VATS). Serous effusions may be seen with a variety of infections, including Mycoplasma. Inflammation below the diaphragm, particularly abscesses or pancreatitis, can also result in pleural effusions. Rarely, a unilateral effusion will accompany an intrathoracic tumor that involves the pleura (Fig. 51.41; see Fig. 51.50) .
3246
3247
FIGURE
51.34. Patent Ductus Arteriosus in a Premature
Infant With Hyaline Membrane Disease. A. Early films showed the typical small granular lungs seen in hyaline membrane disease. B . Following surfactant therapy, the lungs increased in volume and became clear. C . A few days later, the heart has enlarged and, although the lungs have increased in volume, they have also become more opaque. The lung opacity represents pulmonary edema because of the development of a patent ductus arteriosus.
FIGURE
51.35. “Leaky Lung― Syndrome. A. This
premature infant was born with clear lungs. B . A few days later the lungs, although well expanded, are hazy to opaque. The opacity represents capillary leak pulmonary edema.
Bilateral serous pleural effusions are most commonly seen in patients with renal diseases such as acute glomerulonephritis or nephrotic syndrome, lymphoma (usually non-Hodgkin), or neuroblastoma. Congestive heart failure, collagen vascular and fluid overload may also result in pleural effusions.
3248
diseases,
FIGURE
51.36. Bronchopulmonary
Dysplasia. Typical bubbly
appearance with bubbles of various sizes is seen in advanced bronchopulmonary
dysplasia.
3249
FIGURE
51.37. Retained Fluid Syndrome. A. On the first day
of life, the lungs of this term newborn show diffuse haziness, streaky parahilar opacities, and bilateral pleural effusions. B . The following day, all the abnormalities have resolved, which is the typical sequence of events in an infant with retained lung fluid.
3250
TABLE 51.8 Sources of Diffusely Hazy or Reticular Lungs in the
Neonate
Decreased lung volumes Poor inspiration Hyaline membrane disease Normal to increased lung volumes Retained fluid Aspiration (amniotic fluid/meconium) Pneumonia Pulmonary edema Pulmonary lymphangiectasia
P.1249 Hemothorax is usually the result of trauma, either direct chest wall trauma with or without rib fractures or aortic rupture from deceleration injury. Occasionally, bleeding disorders can result in hemothorax. Rarely, an aneurysm of the ductus arteriosus can rupture and bleed into the pleural space. Chylothorax is the most common cause of massive pleural effusion in the neonate. Chylous effusions are usually unilateral and are somewhat more common on the right (Fig. 51.42). The cause of chylothorax is uncertain, but hypotheses include traumatic tear or congenital defect of the thoracic duct. Chylous effusions that occasionally result from superior vena cava thrombosis are more difficult to manage. Pulmonary lymphangiectasia is a rare cause of chylothorax. Most chylothoraces resolve following thoracentesis, although occasionally chest tube drainage or pleuroperitoneal shunting
is
required.
Complications of indwelling catheters in thoracic vessels are a relatively common iatrogenic cause of pleural fluid (2 2) .
3251
FIGURE 51.38. Meconium Aspiration. A coarse, reticulonodular pattern throughout both lungs is typical of meconium
aspiration.
3252
FIGURE 51.39. Right Pleural Effusion in a Patient With Nephrotic Syndrome. The flattened and laterally displaced curvature of the right hemidiaphragm (arrows) indicates the presence of subpulmonic pleural fluid.
LUNG
MASSES
The most common pulmonary “mass― in children is a pseudomass caused by a spherical pneumonia (Fig. 51.43). Such an appearance is not uncommon at certain stages of pneumonia in children. Pulmonary abscess has a similar masslike appearance but usually contains central cavitation with air–fluid levels. Postinflammatory granulomas caused by tuberculosis or fungal infections are the most common true lung masses. Such granulomas are usually small and are very often calcified (Fig. 51.44). Plasma cell granuloma, or postinflammatory pseudotumor, is a reactive lesion that develops from a healing pneumonia. Calcification is uncommon, and the lesion gradually resolves over a period of years.
TABLE 51.9 Possible Causes of Pleural Effusions
Unilateral Pneumonia/empyema Chylothorax Iatrogenic Trauma Intra-abdominal inflammation Intrathoracic neoplasm Ruptured aneurysm of ductus arteriosus Bilateral Renal disease Lymphoma Neuroblastoma Congestive heart
failure
3253
Collagen vascular Fluid overload
FIGURE
diseases
51.40. Empyema.
A. Radiograph of an 18-month-old
child shows a large area of opacity in the left lung. B . US localizes the fluid to the pleural space and shows multiple septations, which is characteristic of empyema.
3254
FIGURE
51.41. Pleural Effusion Associated With a Thoracic
Neoplasm.
A. Pleural fluid causes complete opacification of the
right hemithorax and shift of the mediastinum to the left. B . Upon drainage of the effusion, a large right intrathoracic mass (teratoma) becomes apparent. C . CT of another child shows a large pleural effusion associated with mediastinal neuroblastoma.
3255
FIGURE 51.42. Chylothorax in Newborn Infant. A large right pleural fluid collection compresses the right lung (arrows) and displaces the mediastinum to the left.
P.1250 P.1251 Bronchogenic cysts are lined with respiratory epithelium and filled with mucoid liquid. They occur in the lung parenchyma or in the mediastinum. A subcarinal location is very common. Some are connected to the bronchial tree and are air filled. The cystic nature of these lesions is readily demonstrable with CT or MR (Fig. 51.45) . Pulmonary sequestration is a mass of lung tissue that lacks a connection to the bronchial tree and is supplied by abnormal vessels from the descending aorta. Sequestrations are classified as extralobar (covered by their own pleura) or intralobar (covered by the pleura of the adjacent normal lung). Most appear as a triangular
3256
or oval-shaped mass in the medial and basal portions of a lung, more commonly on the left (Fig. 51.46A). Air is sometimes present within sequestrations because of collateral air drift. Most are clinically silent until they become infected and present as pneumonia. The diagnosis is made by demonstrating the abnormal blood supply with US (2 3) , CT (2 4), or MR angiography (Fig. 51.46B) (2 5,2 6) .
Rare
Pulmonary
Masses
Other rare causes of a pulmonary mass usually have few distinguishing features. A mass connected to an unusually large vessel is likely to be a pulmonary arteriovenous malformation. A central, oval-shaped nodule associated with overaeration of the involved lobe suggests the diagnosis of a mucocele in a patient with bronchial atresia. Primary lung tumors are rare, and the majority are benign. Pulmonary hamartoma is a benign congenital tumor that occasionally contains characteristic flocculent calcifications. Rarely, laryngeal papillomas can spread into the trachea and the lungs. Primary pulmonary malignancies are very rare and include sarcomas, primitive neuroectodermal tumors, and squamous cell carcinoma. Pleuropulmonary blastoma is a rare neoplasm composed of both epithelial and mesenchymal elements. These neoplasms often arise from a congenital lung cyst. Pleuropulmonary blastoma can be either solid or cystic and is usually accompanied by a pleural effusion (2 7) .
Multiple
Nodules
By far the most common malignant neoplasm in the lung during childhood is metastasis, whether single or multiple. The most common childhood tumors to metastasize to the lungs are Wilms tumor, Ewing sarcoma, osteosarcoma, and rhabdomyosarcoma. Other masses and nodules that can be multiple include granulomas (most often fungal), abscesses, hemangiomas, and Wegener granulomatosis. Cavitary nodules are characteristic of septic emboli, Wegener granulomatosis, laryngeal papillomatosis, sarcoidosis, and metastases. Staphylococcal pneumonia can also be associated with multiple cavitating lesions (Fig. 51.47) .
3257
MEDIASTINAL
AND
HILAR
MASSES
The division of the mediastinum into anterior, middle, and posterior compartments is the most useful scheme for categorizing mediastinal masses in both children and adults. This discussion will use an arbitrary system based on the division of the chest into rough thirds on the lateral view. The thymus gland is the primary normal structure in the anterior mediastinum and is also the most common cause of an apparent anterior mediastinal mass. The normal thymus gland varies widely in its appearance, sometimes causing considerable confusion during the interpretation of an infant or young child's chest radiograph. The gland is commonly very prominent at birth, remains easily visible up to about 2 years of age, and may be seen in an older child. On posteroanterior chest radiographs, the thymus gland causes smooth bilateral widening of the superior mediastinum. The gland overlies and silhouettes the upper cardiac borders, and sometimes a small notch is visible at the junction between the thymus and the heart (Fig. 51.48A, B). The border of the thymus gland may have a wavy contour caused by compression by the overlying ribs (Fig. 51.48C) . One thymic lobe may appear more prominent than the other and have a triangular configuration called the “sail― sign (Fig. 51.48D). This appearance is more commonly seen on the right and may be mistaken for lung consolidation, particularly if the patient is in a slightly right-sided rotated position. On the lateral view of the chest, the thymus lies over the anterosuperior portion of the cardiac silhouette in the retrosternal space (Fig. 51.48E). The normal thymus gland can have more unusual configurations, such as extensions high into the superior P.1252 mediastinum, the lower neck, or posteriorly between the innominate and left brachiocephalic arteries. Only rarely do these atypical positions result in symptoms.
3258
FIGURE 51.43. Round Pneumonia. A. and B . This pneumonia (arrows) of the right upper lobe has a round, masslike configuration on frontal and lateral views. C . Another child with left lower lobe pneumonia that has a round configuration on CT.
Stress atrophy of the thymus is an interesting phenomenon that occurs secondary to almost any type of illness or to the use of steroids. The thymus rapidly shrinks in size during illness, only to return to normal size after the infant has recovered. Occasionally, rebound hypertrophy follows stress atrophy. When stress atrophy is severe, the mediastinum appears very narrow, suggesting absence or
3259
hypoplasia of thymus gland (Fig. 51.49). The distinction between thymic atrophy and aplasia becomes important in infants who are suspected of having certain immunologic disorders. The best known of these is the DiGeorge syndrome, which consists of thymic aplasia, absence of the parathyroid glands, and cardiovascular anomalies. This syndrome is caused by faulty development of the third and fourth pharyngeal pouches. US is helpful in identifying a small thymus gland that is not apparent radiographically. Thymic tissue has a characteristic texture on US, with multiple linear connective tissue septa (2 8). Ectopic cervical thymic tissue can mimic a neck mass but can be distinguished from other neck masses sonographically (2 9,3 0) . A large thymus gland is nearly always a normal gland in an infant. Leukemia or lymphoma can infiltrate the thymus gland, sometimes causing massive enlargement (Fig. 51.50). Thymic cysts are uncommon developmental lesions that can be seen with US, appearing as well-defined, round lesions that are anechoic unless complicated by hemorrhage or infection. Spontaneous hemorrhage into the thymus gland has been described in newborn infants. When pneumothorax is present in a neonate, the thymus P.1253 gland can become compressed and elevated by the free air, creating a pseudomass in the superior mediastinum (Fig. 51.51). This masslike compression of the thymus gland may be a clue to a subtle anterior
pneumothorax.
3260
FIGURE
51.44. Histoplasma
Granuloma. The small, round,
well-defined mass in the right costophrenic angle (arrow) is a granuloma.
On cross-sectional imaging, the lobes of the normal thymus gland have a smooth, somewhat triangular shape with homogeneous texture. Bulging or convexity of the borders of the thymus gland suggests pathologic enlargement, particularly if the trachea or the great vessels are displaced or compressed. Primary neoplasms and cysts produce focal alterations of attenuation or signal intensity, whereas infiltration by leukemia or lymphoma or hemorrhage results in a more diffuse and heterogenous parenchymal pattern. Overall, MR probably is the best examination for defining whether thymic tissue is normal or abnormal.
Anterior
Mediastinal
Masses
The majority of pathologic masses in the anterior mediastinum in children are neoplasms. Benign germ cell tumors (i.e., teratoma, dermoid) are common in this location. Dermoids are benign tumors
3261
comprised only of ectodermal elements, whereas teratomas contain elements from all dermal layers. Mature teratomas characteristically contain calcifications, fluid, and fat, which are best demonstrated by CT (Fig. 51.52) (3 1). Malignant germ cell tumors also occur in the anterior mediastinum (3 2). Other neoplasms in the anterior mediastinum include thyroid tumors, hemangiomas, and cystic hygromas. Cystic hygroma is a congenital malformation of lymphatic origin that commonly arises in the neck. Cystic hygroma tends to be locally invasive and often extends into the mediastinum. US reveals the multiloculated, cystic nature of this mass (Fig. 51.53). MR is usually used to evaluate the extent of cystic hygroma prior to resection. Hemangiomas tend to be solid and echogenic, with small cystic areas representing vascular lakes in the cavernous type of hemangioma. Blood flow is demonstrable with color Doppler imaging.
Middle
Mediastinal
Masses
Normal structures in the middle mediastinum from which masses can arise include lymph nodes, the airway, the esophagus, and the heart and great vessels. Lymphadenopathy is by far the most common middle mediastinal mass. Inflammatory lymphadenopathy is much more common than neoplastic disease, but when massive adenopathy is seen, lymphoma or leukemia should be considered (Fig. 51.54) . Hilar lymph node enlargement may accompany mediastinal adenopathy or may occur alone. Bilateral hilar adenopathy commonly is the result of viral lower respiratory tract infections in children, but mycoplasmal, fungal, and tuberculous infections are also common causes. Other causes of lymphadenopathy include LCH, metastatic disease, sarcoidosis, and Wegener granulomatosis. Unilateral lymphadenopathy is a common radiographic finding of primary tuberculosis in children (3 3) and is often associated with a small area of opacity in the ipsilateral lung (the Ghon complex) (Fig. 51.55A) (2 3). Unilateral lymphadenopathy is seen frequently with mycoplasma or fungal infections of the lung and occasionally with a bacterial pneumonia (Fig. 55B). Unilateral lymphadenopathy is uncommon with viral infections. Neoplastic lymph node enlargement can be unilateral or bilateral.
3262
Cystic masses in the middle mediastinum can be associated with the airway or the esophagus. Bronchogenic cysts are sharply marginated, fluid-filled masses that may be lobulated and commonly occur around the carina (see Fig. 51.45) (3 4). Occasionally, the cysts will appear solid on CT, and MR may better demonstrate the cystic nature of the lesion. GI duplication cysts are caused by abnormal development of the posterior division of the primitive foregut. Duplication cysts that reside in the thorax tend to arise from the esophagus, the stomach, or the duodenum (Fig. 51.56). These cysts usually do not communicate with the esophagus, but they displace and compress the esophagus on contrast studies. In some cases, gastric mucosa will be present in the cyst lining, leading to ulceration and hemorrhage. Enlarged vascular structures may present as a middle mediastinal mass. Aortic aneurysms are rare in childhood, except for those associated with trauma or with connective tissue disorders such as the Marfan or Ehlers-Danlos syndromes. In the newborn infant, a small bump may be visible along the upper descending aorta, caused by a dilated infundibulum of the ductus arteriosus after closure (Fig. 51.57). Normally this “ductus bump― disappears in the first weeks of life. Enlargement or persistence of this bump in later infancy suggests an aneurysm of the ductus arteriosus. Enlargement of the aorta and the main pulmonary artery may occur with congenital
cardiac
anomalies
(see “Congenital
3263
Heart
Disease―) .
FIGURE
51.45. Bronchogenic Cyst. A. A rounded opacity
peeks out from behind the right heart border (arrows). B . A lateral view more clearly shows the round, well-defined cyst (arrows). C . Coronal T2WI shows the high-intensity bronchogenic cyst in a right paraspinal location.
3264
FIGURE 51.46. Pulmonary Sequestration. A. A poorly defined opacity is seen through the cardiac silhouette on the left (arrows). B . Coronal MR image in another patient shows a left lower lobe sequestration. The sequestration appears as a highintensity mass associated with a large abnormal vessel (arrows) .
3265
FIGURE 51.47. Staphylococcal nodules are seen on CT.
Pneumonia. Multiple cavitary
P.1254 P.1255 A mass along the upper left cardiac border can be caused by herniation of the left atrial appendage through a partial pericardial defect or by a coronary artery aneurysm. Coronary artery aneurysms occur in children with periarteritis nodosa or the mucocutaneous lymph node syndrome. Enlargement of the azygos vein presents as a mass in the right paratracheal region. In children, this most often occurs with total anomalous pulmonary venous return to the azygos vein or absence of the inferior vena cava with azygos continuation.
3266
Posterior mediastinal masses are largely of neurogenic origin (3 5). Close inspection of the vertebra and posterior ribs may reveal pedicle erosion, interpedicular or rib space widening, or bone erosions, which are clues to a mass extending into the spinal canal. In such cases, the mass is most often a neoplasm of the neuroblastoma-ganglioneuroma group (Fig. 51.58). These tumors are probably congenital in origin and arise in paraspinal sympathetic nerve tissue. Primary thoracic neuroblastoma has a more favorable prognosis than neuroblastoma that originates in the abdomen. Ganglioneuroma is the benign counterpart of neuroblastoma, and the two lesions cannot be reliably distinguished from one another radiologically. Calcifications may be seen in both lesions, and it is believed that some neuroblastomas can mature to ganglioneuromas. MR is valuable to assess the extent of tumor, especially intraspinal extension
(3 6,3 7) .
Neurofibromas also occur in the posterior mediastinum and cause widening of intervertebral foramina. These tumors can be solitary but more often occur with the neurofibromatosis syndromes. Anterior thoracic meningoceles also occur in patients with neurofibromatosis and have a similar radiographic appearance. Neurenteric cysts are a form of enteric duplication cysts that communicate with the spinal canal. The cysts lie in the posterior mediastinum and are almost always associated with vertebral anomalies (Fig. 51.59). Spinal cord anomalies may also be present. MR is the procedure of choice for evaluating this condition. A posterior mediastinal inflammatory mass occasionally accompanies inflammatory conditions of the spine. Rare causes of a posterior mediastinal mass include lymphangioma, teratoma, lymphoma, and sarcoma. Diaphragmatic hernias through the foramen of Bochdalek and pulmonary sequestration often present as masses in the inferior portion of the posterior mediastinum, adjacent to the diaphragm.
CHEST
WALL
MASSES
Most masses that involve the chest wall of children arise from the cartilage or bones of the thoracic cage. Many such lesions are
3267
malignant and are often quite large at the time of presentation (Fig. 51.60A). Ewing sarcoma and primitive neuroectodermal tumor (Askin tumor) are the most common malignancies to involve the chest wall in children. Radiologically, both lesions appear as large extrapleural soft tissue masses, usually with evidence of rib destruction and ipsilateral pleural effusion. CT is the most helpful way to characterize the extent of these tumors (3 8). Rhabdomyosarcoma may also involve the chest wall. Chondrosarcoma is rare in childhood. Metastatic rib lesions are common in infants and children with neuroblastoma. Chest wall involvement in leukemia and lymphoma is also common. Pleural thickening adjacent to such lesions can help identify subtle metastases. A variety of benign masses may occur in the chest wall (Table 51.10). Osteochondromas of the ribs are common. Mesenchymal hamartoma is a rare benign neoplasm of the ribs that occurs primarily in infants under 1 year of age (3 9). The neoplasm is composed of solid elements of proliferating cartilage, bone, and fibroblasts. Cystic areas of hemorrhage are commonly seen within the mass. These tumors are noninvasive, but they often cause erosion of adjacent ribs and compression of other adjacent tissues. Complete resection is curative. Intrathoracic infection (empyema) may extend to involve the chest wall, creating the appearance of a mass (Fig. 51.60B). Staphylococcus and Fusobacterium are common organisms.
3268
FIGURE 51.48. Normal Thymus Configurations. A. In a young infant, the thymus is often quite prominent, causing bilateral
3269
superior mediastinal widening. Note the right-sided buckling of the trachea, which is a normal finding during expiration in infants and should not be mistaken for mass effect. B . In older infants and young children, the lobes of the thymus gland become less prominent (arrows). Note the subtle notch at the junction of the thymic and cardiac shadows (arrowheads). C . The wavy thymic contour on the left is caused by rib compression (arrows). D . The right thymic lobe is prominent in this patient, with a configuration that has been likened to a sail. E . The straight inferior border of the thymus gland, which lies in the retrosternal space, is seen on the lateral view.
FIGURE 51.49. Stress Atrophy of the Thymus. The narrow superior mediastinum is caused by absence of a thymic shadow in this infant suffering from failure to thrive.
P.1256
3270
P.1257
CONGENITAL
HEART
DISEASE
A wide variety of imaging modalities is now available for evaluation of congenital heart disease in children. Many congenital cardiac abnormalities that previously required angiocardiography can now be diagnosed noninvasively by echocardiography, MR, and helical CT (4 0, 4 1, 4 2, 4 3). Radiographs continue to play a role in the initial evaluation for congenital cardiac anomalies in infants and children. Although the specific diagnosis often will not be apparent on the plain films, a systematic approach to plain film interpretation will allow categorization into one of several groups of disorders. This section will provide a framework for an organized scheme for radiographic evaluation of congenital heart disease.
FIGURE
51.50. Hodgkin
Lymphoma. The mediastinal
enlargement is caused by lymphomatous involvement (arrows) of
3271
mediastinal lymph nodes. A right pleural effusion is present.
FIGURE
51.51. Thymic
Pseudomass. The thymus gland is
elevated and compressed by the bilateral anterior pneumothoraces, creating the appearance of a superior mediastinal mass (arrows) .
Assessment begins with the pulmonary vascularity. Vascular patterns are placed in one of three broad groups: increased (congested), decreased, and normal (Table 51.11). If the vascularity is increased, one should attempt to distinguish active congestion from passive congestion. Active congestion occurs whenever the amount of blood flowing through the pulmonary vasculature has increased. This occurs in
3272
conditions with left-to-right shunts and with preferential blood flow into the lower-pressure pulmonary circulation. Left-to-right shunts do not become radiographically apparent until the output of the RV is approximately two and a half times greater than that of the LV. At this point, the pulmonary vessels become increased in diameter and are visible farther than usual into the periphery of the lungs (Fig. 51.61A). The vessels may appear tortuous, but their margins remain relatively distinct. In borderline cases, if the diameter of the right descending pulmonary artery (PA) is less than that of the trachea, a left-to-right shunt is unlikely.
FIGURE 51.52. Mediastinal Teratoma. A. A superior mediastinal mass displaces the trachea to the right. A toothlike
3273
calcification (arrow) is seen within the mass. B . Lateral view shows the anterior location of the mass and the calcification (arrow). C . A CT scan reveals the heterogeneous nature of this teratoma (arrows), which contains dense calcifications and hypodense areas representing fat.
P.1258 Passive congestion reflects elevation of pulmonary venous pressure, which can result from obstruction or dysfunction of the left side of the heart. As venous pressure increases and the veins dilate, edema fluid leaks into the perivascular interstitial tissues, causing the margins of the vessels to become less distinct on the chest radiograph (Fig. 51.61B). As pulmonary venous hypertension increases, alveolar pulmonary edema and pleural effusions develop. In patients with large left-to-right shunts and left heart failure, a mixed pattern of passive and active congestion occurs. Decreased pulmonary vascularity indicates diminished blood flow to the lungs, most often caused by obstruction of the right ventricular outflow tract and associated right-to-left shunts. Oligemia causes the lungs to appear more radiolucent, and the vessels appear uniformly thin and wispy (Fig.
51.61C). A diminished caliber of the
peripheral two thirds of the PAs combined with prominence of the central PAs is characteristic of pulmonary arterial hypertension and increased
pulmonary
vascular
resistance.
Normal pulmonary vascularity is usually seen in patients with uncomplicated valvular disease, coarctation of the aorta, and mild forms of cardiomyopathy. The vessels retain a normal contour and diameter until congestive heart failure develops. Asymmetry of pulmonary blood flow is most commonly seen in tetralogy of Fallot, persistent truncus arteriosus, and valvular pulmonic stenosis. In tetralogy of Fallot, the blood flow to the left lung tends to be diminished. Blood flow to either lung, or occasionally only one lobe, can be decreased in persistent truncus arteriosus. In valvular pulmonic stenosis, the abnormal valve tends
3274
to direct the blood flow preferentially into the left pulmonary P.1259 arterial system, although an enlarged left PA and increased blood flow to the left lung is seldom apparent on the radiographs of children. Unilateral or asymmetric decreased pulmonary vascularity can also be seen in association with obstructive emphysema of the lungs. The next step in radiographic interpretation of congenital heart disease is assessment of the main PA and the aorta.
FIGURE
51.53. Pericardial
Lymphangioma. The pericardial
mass has the characteristic multiseptated, cystic appearance of lymphangioma on contrast-enhanced CT (arrows) .
Pulmonary
Artery
An enlarged PA may represent generalized increased pulmonary blood flow, poststenotic dilation caused by valvular pulmonic stenosis (Fig. 51.62), or pulmonary valve insufficiency caused by increased
3275
right P.1260 ventricular output. An enlarged PA will often be higher in position than a normal PA and can be mistaken for a large aortic knob. A small or absent PA shadow occurs with decreased blood flow caused by pulmonary outflow obstruction or with an abnormal position of the PA, such as with persistent truncus arteriosus or transposition of the great vessels.
FIGURE
51.54. Bilateral Hilar Adenopathy. A. The prominent
nodular hilar opacities represent lymph nodes in this patient with primary tuberculosis. B . The bilateral hilar and paratracheal masses represent lymphadenopathy caused by Hodgkin lymphoma.
3276
FIGURE 51.55. Unilateral Hilar Adenopathy. A. Right hilar adenopathy and adjacent parenchymal opacity comprise the typical Ghon complex of primary tuberculosis. B . Right hilar adenopathy is associated with a bacterial pneumonia.
Aorta Evaluation of the aorta includes estimation of size, position, and contour abnormalities. The size of the aorta is assessed in the region
3277
of the aortic knob. The aorta may appear small because of hypoplasia (as in hypoplastic left heart syndrome) and with certain left to-right shunts (e.g., atrial septal defect [ASD], ventricular septal defect [VSD]). Because the aorta in children is normally small relative to adults, a truly small aorta may be difficult to recognize. Enlargement of the ascending aorta and the aortic knob most often represents poststenotic dilation resulting from valvular aortic stenosis; it may also be caused by increased aortic blood flow seen with aortic valve insufficiency, left-to-right shunting at the great vessels level (patent ductus arteriosus, persistent truncus arteriosus), or severe tetralogy of Fallot. Generalized aortic enlargement can occur with systemic hypertension. The most common abnormality of the contour of the aorta is the notching that occurs at the site of coarctation of aorta. Dilation of the aorta proximal and distal to the coarctation results in the characteristic “figure 3― sign (Fig. 51.63) . Right-sided
aortic
arch is most often an isolated anomaly.
However, it can also accompany congenital heart disease, especially persistent truncus arteriosus or tetralogy of Fallot. A right aortic arch is seen as a bulge or fullness in the right paratracheal region, slightly above the usual level of a left-sided aortic arch. The trachea will be displaced to the left by a right aortic arch (see Fig. 51.61C). In addition, a right descending aorta can be visualized just to the right of the spine in many cases. A right-sided aortic arch may be a clue to the presence of a vascular ring, such as a double aortic arch or aberrant left subclavian artery with an encircling ligamentum arteriosum. In such cases, barium swallow reveals opposing indentations in the barium-filled esophagus in a reverse S configuration (see Fig. 51.20) . Cardiomegaly is an important indicator of cardiac disease in children and often accompanies congenital heart disease. Unfortunately, the estimation of cardiac enlargement in children is somewhat subjective, and measurements such as the cardiothoracic ratio are usually not helpful. Beware of the normally prominent thymus gland overlying the heart and of films obtained during a poor degree of inspiration that can erroneously suggest cardiac enlargement. The configurations of enlargement of specific cardiac
3278
chambers are the same as those seen in adults. Generalized cardiomegaly with a globular appearance suggests pericardial fluid, which can easily be confirmed with US (Fig. 51.64). Pericardial effusions in children commonly accompany viral infections or rheumatic fever. Other causes include acute or chronic renal failure, collagen vascular diseases, bacterial infections, and, rarely, tuberculosis, fungal infections, and pericardial metastases. Blood in the pericardial space is usually the result of trauma. Generalized cardiac enlargement also occurs with conditions that cause increased blood volume and elevated cardiac outputs, including renal diseases, inappropriate secretion of antidiuretic hormone, large arteriovenous fistulae, chronic anemias (especially sickle cell disease and thalassemia), and hyperthyroidism.
Acyanotic Heart Disease Pulmonary Vascularity
With
Increased
Actively increased pulmonary vascularity in the absence of cyanosis most often occurs when a defect allows oxygenated blood from the left side of the heart or the aorta to be shunted back to the right side of the heart or the pulmonary circulation. Because no desaturated blood is shunted into the systemic circulation, cyanosis does not occur. The increased blood volume recirculating through the right heart and pulmonary circulation results in cardiac enlargement and increased size of the pulmonary vessels. The most common conditions in this category are VSD, ASD, and patent ductus arteriosus (PDA). Ventricular
septal
defect is the most common congenital heart
abnormality after bicuspid aortic valve. VSD occurs frequently as an isolated anomaly, although it may accompany many of the cyanotic forms of congenital heart disease. The defect is categorized according to its location within the ventricular septum. Most are perimembranous defects in the portion of the septum near the fusion of the membranous and muscular portions. Defects in the muscular septum are less common and tend to be smaller and less hemodynamically significant than a perimembranous defect. The third
3279
type of defect is uncommon and develops high in the membranous septum because of abnormal development of the conus portion of the truncus arteriosus. This type is most often seen with persistent truncus arteriosus or tetralogy of Fallot. Newborns with VSD are usually asymptomatic. A murmur will often not be detected until after the newborn period. This delay in manifestation of left-to-right shunting is the result of the normal phenomenon of postnatal pulmonary vascular involution. In the fetus, the walls of the PAs are thicker than in postnatal life, resulting in increased pulmonary vascular resistance. Elevated pulmonary vascular pressure inhibits blood flow through the lungs and through the septal defect. During the early postnatal period, the pulmonary vascular resistance diminishes, allowing left-to-right shunting through the septal defect. In patients with moderate to large VSDs, symptoms
usually P.1261
develop within the first 2 years of life. Small defects can close spontaneously.
3280
FIGURE
51.56. Intestinal Duplication Cyst. A. Radiograph of
this newborn infant shows a large soft tissue mass in the right chest (arrows), displacing the mediastinum to the left. B . Coronal T2WI shows the cystic nature of the mass, which arises from the duodenum.
3281
FIGURE 51.57. Ductus Bump. The prominent “bump― that is seen along the upper descending aorta represents the dilated infundibulum of the ductus arteriosus in this newborn infant.
The characteristic findings in VSD are PA enlargement, increased pulmonary vascularity, and cardiomegaly that is predominantly leftsided (Fig. 51.65). Increased pulmonary venous return results in volume overload of the LA and LV, leading to dilation. LV dilation causes a “drooping― shape of the left cardiac border. LA enlargement is best seen on the lateral or left anterior oblique view as a bulge along the upper posterior cardiac border that causes posterior displacement of the esophagus and the left mainstem bronchus. If the shunt is large, biventricular enlargement occurs. Atrial septal defect is much less common than VSD. The most common type of ASD occurs centrally at the foramen ovale, the ostium secundum defect. Because ASD is associated with a low-
3282
pressure shunt, these children seldom develop symptoms in infancy or early childhood. If the shunt persists as the child grows older, the risk of developing pulmonary hypertension increases. As the pressure in the right side of the heart rises, the shunt becomes balanced and eventually reverses. This phenomenon is P.1262 referred to as Eisenmenger physiology and can be seen with any leftto-right shunt.
FIGURE 51.58. Thoracic Neuroblastoma. A. Note the large mediastinal mass. Thinning of the posterior and medial portion of the left second rib and widening of the T2-T3 rib space (arrow) localizes this mass to the posterior mediastinum and suggests possible intraspinal extension. B . Coronal MR of a different patient with a left posterior mediastinal mass demonstrates extension of the tumor into the spinal canal (arrow) .
The LA is not enlarged with ASD because of rapid shunting of blood away from the LA into the right side of the heart. Typically, the RA is enlarged, causing prominence of the right cardiac border on the frontal view (Fig. 51.66). On the lateral view, RV enlargement
3283
produces fullness in the retrosternal space. In both ASD and VSD, the aorta is rather small, as the shunt is below the level of the great vessels. The ostium primum type of ASD (endocardial cushion defect) is caused by abnormal development of the primitive endocardial cushions that form the interatrial and interventricular septa and atrioventricular valves. This condition commonly occurs in trisomy 21. The specific malformation ranges from two separate atrioventricular valves with a low ASD and a VSD to the complete form, with a common atrioventricular ring and a five-leaflet valve. The mitral valve is clefted and abnormally positioned, resulting in elongation of the left ventricular outflow tract, which creates a “gooseneck― appearance on angiography (Fig. 51.67). The partial form behaves hemodynamically as a simple atrial left-to-right shunt, with only mild degrees of mitral or tricuspid insufficiency. The clinical course of the complete form is much more severe. The shunts are large and usually bidirectional because of the abnormal valve development. The patients are cyanotic and tend to develop pulmonary hypertension and congestive heart failure early. Radiographically, these patients present with marked cardiomegaly with right atrial and right ventricular predominance and pulmonary vascular engorgement (Fig. 51.68) .
Patent
Ductus
Arteriosus
The ductus arteriosus connects the PA and aorta in fetal life. Normally, this structure begins to close immediately after birth, but in some infants, closure is delayed. Prolonged patency is a common complication of hypoxia in the premature infant. In many infants, the cause of persistent patency is unknown. Symptoms develop within the first 2 years of life. Blood is shunted from the aorta through the ductus to the PA, resulting in increased blood volumes flowing through the lungs and the left side of the heart. Consequently, the LA, LV, and PA become dilated with active pulmonary vascular engorgement. An enlarged proximal aorta differs from the small aorta that is seen with ASD and VSD (Fig. 51.69). In young infants with large shunts, cardiomegaly tends to be more generalized, and
3284
the size of the aorta is difficult to evaluate because of overlying thymus gland. PDA is easily diagnosed by echocardiography or angiography. Aortopulmonary window is a rare condition that is very similar to PDA, both hemodynamically and radiographically. The abnormality results from failure of complete division of the primitive truncus arteriosus, which leaves a communication between the aorta and the PA just above the valves. Other rare conditions that result in shunting of blood from the aorta to the PA include rupture of an aneurysm of the sinus of Valsalva and fistulas between the coronary arteries and the coronary sinus, PA, or right cardiac chambers.
3285
FIGURE 51.59. Neurenteric Fistula. A. Sagittal T2WI shows a small anterior cyst with a fistula extending into the T2 vertebra (arrow). B . Axial CT image shows the obstructed esophagus (E) and the air-filled fistula (arrow) extending into a defect in the
3286
vertebral
body.
P.1263
Cyanotic
Heart
Pulmonary
Disease
With
Increased
Vascularity
This category consists of a group of complex heart abnormalities whose common feature is the admixture of oxygenated and deoxygenated blood that is circulated systemically, resulting in cyanosis. Transposition anomalies of the great vessels are lesions with abnormal anteroposterior positioning of the aorta and PA and with abnormalities of the relationship between the atria and ventricles and their connections to the great vessels. Complete transposition of the great vessels (D-transposition) is the most common form of cyanotic congenital heart disease with increased pulmonary blood flow. In this condition, the positions of the aorta and PA are reversed. The ventricles lie in their normal positions, so that the aorta arises anteriorly from the RV and the PA arises posteriorly from the LV. This results in two separate circulations, one through the pulmonary circulation and the other systemic. Communications that allow the infant to survive are most commonly a VSD, an ASD, or PDA. Bidirectional shunting through these communications allows adequate mixing of the blood if the shunts are large enough. If pulmonary stenosis is not present, blood flows preferentially into the low-resistance pulmonary circulation, resulting in increased pulmonary venous return and pronounced volume overloading. Congestive heart failure develops in the first weeks of life. The prognosis is more favorable with associated pulmonary stenosis. Cardiomegaly with an oval configuration develops in the first few days of life (Fig. 51.70A). The superior mediastinum and base of the heart are narrow because of thymic atrophy and the abnormal alignment of the aorta and PA. Both active and passive pulmonary
3287
vascular congestion can be seen. On lateral views of the chest, the anteriorly placed aorta causes increased opacity in the retrosternal region (Fig. 51.70B). The corrective arterial switch procedure is performed early in many infants, and therefore, the classic radiographic findings may not develop.
Corrected
transposition
of
vessels
(L-transposition)
the
great
In L-transposition, ventricular inversion (left to right reversal) accompanies the transposed positions of the aorta and PA, resulting in functional correction of the transposition. Blood circulates through the RA to LV to PA to the pulmonary circulation, and LA to RV to aorta to the systemic circulation. The anatomic RV functions as an LV and vice versa. Because the aorta lies anteriorly and to the left, this condition is often called L-transposition, while complete transposition is called D-transposition. Patients with the simple form of corrected transposition tend to be asymptomatic, but patients with coexisting cardiac defects (VSD, pulmonary stenosis, conduction defects) have an unfavorable prognosis. The diagnosis is suggested radiographically by a characteristic prominence along the upper left cardiac border that represents the right ventricular outflow tract and the left-sided aorta (Fig.
51.70C) .
Double-outlet right ventricle is characterized by an aorta that is anterior to or lateral to the PA and arises from the RV (Fig. 51.71) . The PA also empties the RV, originating entirely from the RV (type I), or overriding a high VSD and draining both the LV and RV (type II or the Taussig-Bing anomaly). The hemodynamics are similar to complete transposition of the great vessels. Radiographic findings are also similar. However, because the aorta and PA are oriented in a more side-to-side fashion in some cases, the P.1264 cardiac waist may be of normal or even increased width, unlike the narrow waist seen in transposition.
3288
FIGURE
51.60. Chest Wall Masses. A. A Ewing sarcoma arises
from a posterior rib in a child. Note the rib destruction. B . CT shows a large posterior chest wall abscess that arose following staphylococcal empyema (empyema necessitatis).
Total anomalous pulmonary venous return (TAPVR) is a condition in which the pulmonary veins, instead of emptying into the LA, return blood to the right side of the heart via the RA, coronary sinus, or a systemic vein. major cardiac isolated form. was described
This anomaly can occur in conjunction with other defects, but this discussion will refer only to the The best known classification of the types of TAPVR by Craig et al (4 4). In all types, the pulmonary veins
converge into a single common vein before emptying into the anomalous site. In type 1 TAPVR, the most common form, the abnormal vein empties into a large supracardiac vein (a persistent left superior vena cava, the left brachiocephalic vein, the right superior vena cava, or the azygos vein). In the type 2 anomaly, the common vein drains into the coronary sinus or directly into the RA. In the type 3 anomaly, the common vein travels through the esophageal hiatus to empty into the portal vein or, less commonly,
3289
an abdominal systemic vein.
TABLE 51.10 Chest Wall Masses
Malignant Ewing sarcoma Primitive neuroectodermal Neuroblastoma Leukemia
tumor
(Askin
tumor)
Lymphoma Rhabdomyosarcoma Benign Osteochondroma Aneurysmal bone cyst Mesenchymal hamatoma Langerhans cell histiocytosis Fibrous dysplasia Hemangioma Lymphangioma Teratoma Abscess Calcifying Osteoid
fibrous
pseudotumor
osteoma
TAPVR types 1 and 2 overload the right side of the heart, causing dilation of the RA, RV, and PA and engorgement of the pulmonary vessels. Communication with the left side of the heart is mandatory for survival and usually occurs as an ASD or patent foramen ovale. The classic radiographic configuration of the type 1 anomaly is the “snowman― heart, so named because of prominence of the superior mediastinum caused by a large, inverted U-shaped vessel that empties into the superior vena cava (Fig. 51.72). This configuration is only present when the abnormal common pulmonary vein enters the persistent left superior vena cava or vertical vein. In
3290
the other forms of the type 1 anomaly and in the type 2 anomaly, the cardiac configuration is less specific. Type 2 findings resemble P.1265 those of the transposition complex of lesions. In the type 1 anomaly, when the abnormal vein empties into the azygos vein, the azygos vein will be dilated.
TABLE 51.11 Pulmonary Vascular Patterns
Increased vascularity (active) Atrial septal defect Ventricular septal defect
without
cyanosis
Patent ductus arteriosus Aortic-pulmonary window Ruptured aneurysm of sinus of Valsalva Coronary artery fistula Partial anomalous pulmonary venous return Increased
vascularity
(active)
with
cyanosis
Total anomalous pulmonary venous return (types 1, 2) Persistent truncus arteriosus Complete endocardial cushion defect Transposition of the great vessels complex Single ventricle (without pulmonary stenosis) Increased vascularity (passive) Total anomalous pulmonary venous return (type 3) Pulmonary vein atresia Hypoplastic left heart syndrome (in failure) Cor triatriatum Decreased vascularity Tetralogy of Fallot Pseudotruncus arteriosus Hypoplastic right heart Tricuspid atresia Pulmonary atresia Tricuspid stenosis
syndrome
3291
(right-to-left
shunt)
Hypoplastic R V Ebstein anomaly Uhl anomaly Trilogy of Fallot Single ventricle or transposition of great vessels with pulmonary stenosis or atresia Tricuspid or pulmonary insufficiency with right-to-left shunt Normal vascularity Left heart lesions Coarctation of the aorta Interrupted aortic arch Hypoplastic left heart syndrome (before failure develops) Endocardial fibroelastosis Cardiomyopathy Aberrant left coronary artery Mitral stenosis and insufficiency Aortic
stenosis
and
insufficiency
Cor triatriatum Right heart lesions (without right-to-left shunt) Pulmonary Tricuspid
stenosis or insufficiency
insufficiency
Type 3 TAPVR is hemodynamically and radiographically distinct from the other forms. Like the other forms of TAPVR, blood is directed incorrectly to the right side of the heart. However, the length and small caliber of the common vein with type 3 TAPVR increases the resistance to flow and creates pulmonary venous obstruction. The pulmonary vessels appear thin, with hazy margins caused by pulmonary interstitial edema and passive vascular engorgement (Fig. 51.73). The heart does not enlarge. The differential diagnosis includes hypoplastic left heart syndrome and pulmonary vein atresia. These three diagnoses are the most common causes of passive vascular congestion in the first 3 days of life. Persistent truncus arteriosus occurs when the primitive truncus arteriosus fails to divide normally into the aorta and PA. Both vessels
3292
are fed by a single vessel that overrides a high VSD. The CollettEdwards classification is based on the site of origin of the PA (Fig. 51.74). The degree of cyanosis is variable and symptoms depend largely on the amount of pulmonary blood flow. Most often, the chest radiograph shows cardiomegaly and active pulmonary vascular congestion. In most forms of persistent truncus, concavity is seen at the usual site of the main PA and strongly suggests the diagnosis. A right aortic arch is present in 30% of the cases (Fig. 51.75). The aorta (truncus) is often dilated, with a high arch and an elevated left PA. Single ventricle refers to a group of anomalies in which one ventricle is rudimentary, leaving the other large ventricle as the only functional ventricle. An underdeveloped RV is most common. The connections between ventricles and the atrioventricular valves, aorta, and PA are variable. Associated lesions include pulmonary valve stenosis, PA atresia, and transposition of the great vessels. If pulmonary stenosis is not present, mixing of saturated and unsaturated blood occurs in the single chamber, and the radiographs show cardiomegaly and pulmonary vascular engorgement. When pulmonary stenosis is present, blood flow to the lungs is diminished and cyanosis is more severe. Echocardiography is usually diagnostic; however, angiocardiography or MR is sometimes needed for complete demonstration of the anatomy.
Decreased
Pulmonary
Vascularity
A decreased pulmonary vascular pattern usually indicates a condition in which the flow through the right side of the heart is obstructed. This obstruction can occur anywhere from the tricuspid valve to the PA. Often, an intracardiac right-to-left shunt, which varies with the severity of right ventricular outflow obstruction, is also present. Tetralogy of Fallot is the most common anomaly to cause diminished pulmonary vascularity and is the most common cause of cyanotic congenital heart disease. The classic components are (1) a high VSD, (2) pulmonary stenosis (usually infundibular, with or without valvular stenosis), (3) right ventricular hypertrophy, and (4)
3293
an aorta that overrides the VSD. A right aortic arch occurs in 25% of cases, and PA coarctation, hypoplasia, or absence are common.
FIGURE
51.61. Pulmonary Vascular Patterns. A. Active
congestion. Large but distinct pulmonary vessels extend into the periphery of the lung as the result of left-to-right shunting in a patient with a large ventricular septal defect. B. Passive congestion. Passive vascular congestion is caused by mitral insufficiency and results in indistinctness of the pulmonary vascular markings. C. Decreased vascularity in a patient with tetralogy of Fallot. Note the right aortic arch (A), concave pulmonary artery segment (arrow), and the characteristic “boot― cardiac configuration caused by right ventricular hypertrophy.
3294
FIGURE
51.62. Pulmonary
Artery
Enlargement.
Poststenotic
dilation of the pulmonary artery is seen in this patient with valvular pulmonic stenosis (arrow) .
3295
FIGURE 51.63. Coarctation of the Aorta. Prestenotic and poststenotic dilation of the aorta creates the characteristic figure-3 sign (arrows) .
3296
FIGURE
51.64. Pericardial
Effusion.
A. The cardiac silhouette
is markedly enlarged and has a rounded, globular appearance caused by a pericardial effusion that developed following open heart surgery. B . US is the best method for verifying a pericardial
effusion.
P.1266 P.1267
3297
The degree of pulmonary stenosis is the most critical component of this anomaly. Severe stenosis leads to marked right-to-left shunting and aortic enlargement, causing greater overriding. Greater right-toleft shunting results in more severe cyanosis. Patients with mild pulmonary stenosis are usually acyanotic and asymptomatic (“pink― or “balanced― tetralogy of Fallot). Patients with moderate to severe forms of tetralogy of Fallot have a characteristic radiographic appearance. The pulmonary vascularity is decreased, with a shallow or concave PA shadow. Right ventricular hypertrophy causes lateral and superior displacement of the cardiac apex without overall enlargement of the cardiac silhouette, creating the classic “boot-shaped heart― (Fig. 51.76). Unequal pulmonary blood flow caused by PA hypoplasia or atresia (usually on the left) is a common finding. The combination of a right aortic arch and decreased pulmonary vascularity is highly suggestive of tetralogy of Fallot or persistent truncus arteriosus.
3298
FIGURE
51.65. Ventricular Septal Defect (VSD). A. Cardiac
3299
enlargement that is predominantly left-sided and increased pulmonary vascularity are characteristic of a VSD. B . Lateral view demonstrates left atrial enlargement (arrows) .
Hypoplastic
right
heart syndrome consists of tricuspid atresia,
usually with pulmonary atresia or stenosis and an underdeveloped RV. Isolated hypoplasia of the RV is rare. The common features are a small RV with right-to-left shunting through an ASD, resulting in cyanosis. A VSD P.1268 or PDA can also be present. Nonspecific cardiomegaly and diminished pulmonary vascularity are seen radiographically. The PA shadow is flat or concave, and the RA is enlarged (Fig. 51.77). The smaller the ASD, the larger the RA. Tricuspid atresia may be accompanied by transposition of the great vessels. When this occurs, the PA drains the LV and, if no pulmonary stenosis is present, the pulmonary vascularity is engorged.
3300
3301
FIGURE 51.66. Atrial Septal Defect (ASD). A. Cardiomegaly, mild right atrial enlargement, and increased pulmonary vascularity are characteristic of an ASD. B . Lateral view shows a normal LA and fullness in the retrosternal region (arrow) caused by right ventricular enlargement.
FIGURE 51.67. Endocardial Cushion Defect. Angiocardiography demonstrates the cleft mitral valve (straight arrow) and the elongated left ventricular outflow tract (gooseneck
deformity)
(curved
arrow) .
Pulmonary atresia is considered a part of the hypoplastic right heart syndrome when accompanied by an intact ventricular septum. In most cases, the pulmonary valve is atretic and the RV and tricuspid valve are hypoplastic. Less commonly, the RV and tricuspid valve are nearly normal, and blood that enters the RV is regurgitated back into the RA, resulting in marked RA enlargement (Fig. 51.78) .
3302
In either case, survival requires a PDA to shunt blood into the P.1269 pulmonary circulation. Prostaglandin E1 is used to help maintain ductal patency until surgery can be performed.
FIGURE
51.68. Endocardial
Cushion
Defect. Marked
cardiomegaly, right atrial enlargement, and increased pulmonary vascularity are typical of this condition.
3303
FIGURE 51.69. Patent Ductus Arteriosus. The heart is enlarged, with left-sided prominence and increased pulmonary vascularity. Note the prominent aorta (arrows) .
In patients with pulmonary atresia with a VSD, the RV is not hypoplastic. This situation occurs in severe tetralogy of Fallot and in pulmonary atresia with VSD and systemic collaterals (the CollettEdwards type IV truncus arteriosus). The most significant difference between these two conditions is the derivation of pulmonary blood flow. In severe tetralogy of Fallot, blood reaches the lungs through a long, tortuous, “wandering― PDA. The other form of pulmonary atresia with VSD relies on primitive systemic collaterals that transport blood from the aorta to the PA branches. Ebstein that is portion portion causes
anomaly consists of a malformed, enlarged tricuspid valve displaced downward, resulting in atrialization of a large of the RV. The remaining RV is very small. The atrialized has abnormal musculature, contracts ineffectively, and functional obstruction of RA emptying. Atrial right-to-left
shunting results in cyanosis in the more severely affected patients. Clinical symptoms and radiographic findings depend on the degree of
3304
downward displacement of the tricuspid valve. Cardiomegaly is mainly right-sided, with decreased pulmonary vascularity and a flattened PA shadow (Fig. 51.79). The right atrial contour is often very prominent. Occasionally, the small displaced RV is seen as a bulge along the upper left cardiac border, causing a squared cardiac appearance. Uhl anomaly is rare and consists of focal or complete absence of the RV myocardium. The very thin, poorly contractile RV functionally impairs the flow of blood through the right side of the heart. The clinical and radiographic findings are similar to those of the Ebstein anomaly.
Normal
Pulmonary
Vascularity
Congenital cardiac anomalies with normal pulmonary vascularity are predominantly abnormalities of the cardiac valves and great vessels. Most have left-sided obstruction. When a concomitant left-to-right shunt is present, cardiac failure develops early and the pulmonary vascularity becomes congested. In the absence of a left-to-right shunt, patients can live for many years without left ventricular failure. Congenital
cardiac
valve
stenosis most commonly affects the
aortic or pulmonary valves. The radiographic findings in children are similar to those seen in adults and consist of hypertrophy of the ventricle that ejects blood through the stenotic valve and poststenotic dilation of the affected artery (Fig. 51.80). Ventricular hypertrophy alters the shape of the heart, with little change in the size of the heart. LV hypertrophy causes a more rounded appearance of the left cardiac border. RV hypertrophy produces fullness in the retrosternal region on the lateral view and upward displacement of the cardiac apex on the posteroanterior view. In valvular pulmonic stenosis, poststenotic dilation of the main PA is often accompanied by prominence of the left PA and increased pulmonary blood flow to the left lung. This phenomenon most likely results from preferential flow through the stenotic valve into the left PA. However, it is less commonly observed on the radiographs of children than on those of
3305
adults because the findings take time to develop. Aortic and pulmonary stenosis may also occur above or below the valve. Subvalvular aortic stenosis is more common than supravalvular stenosis, and the subvalvular narrowing can be caused by a discrete diaphragm or disproportionate hypertrophy of the intraventricular septum in the subaortic region. Supravalvular aortic stenosis is most often associated with Williams syndrome (idiopathic hypercalcemia of infancy) (Fig. 51.81) The features of this syndrome include supravalvular aortic stenosis and other systemic and pulmonary vascular stenoses, facial dysmorphism, mental and growth retardation, and hypercalcemia that may be the result of abnormal regulation of vitamin D metabolism. Subvalvular (infundibular) pulmonary stenosis is the most frequently seen form of stenosis in tetralogy of Fallot. Supravalvular pulmonary stenosis usually consists of multiple areas of narrowing in the peripheral pulmonary artery. Unlike the valvular forms of aortic and pulmonary stenosis, the subvalvular and supravalvular forms are usually not associated with poststenotic dilation of the vessel.
Congenital
Valvular
Insufficiency
Isolated congenital insufficiency of any of the cardiac valves is a very rare occurrence; however, sometimes valvular insufficiency accompanies other cardiac anomalies. In general, valvular insufficiency causes dilation of the cardiac chambers or vessels on both sides of the involved valve. The resulting P.1270 cardiac configurations are the same as those seen in adults with valvular insufficiency.
3306
FIGURE
51.70. Transposition of the Great Vessels. A. In the
more common D-transposition, the heart is enlarged and has an oval “egg― shape. Note the narrow superior mediastinum and
increased
pulmonary
vascularity. B . Angiocardiography in
lateral projection shows the aorta (arrow) arising anteriorly from the RV. C . L-transposition. The transposed aorta arising from the inverted RV on the left causes characteristic prominence along the upper left cardiac border (arrows) .
Coarctation of the aorta occurs in two distinct forms: the juxtaductal (adult) type, which lies at or just distal to the level of the ductus arteriosus, and the rarer preductal (infantile) form, which generally is a long-segment narrowing. Coarctation of the aorta often is associated with other cardiac anomalies, most commonly bicuspid
3307
aortic valve, PDA, or a VSD. Patients with the preductal form of coarctation undergo a more severe clinical course, frequently developing congestive heart failure during the first month of life. Patients with the juxtaductal form usually remain asymptomatic until later in childhood, except in those cases with an associated left-toright shunt. Older children usually present with hypertension, discrepancies between blood pressure in the upper and lower extremities, or a heart murmur. In juxtaductal coarctation of the aorta, the aortic narrowing leads to pressure overloading and hypertrophy of the LV. Usually the heart is normal in size; however, eventually some rounding and prominence of the left cardiac border can develop. Prestenotic and poststenotic dilation of the aorta commonly occurs and is responsible for the “figure-3― sign (see Fig. 51.63). In some cases, the poststenotic dilatation can extend along the entire thoracic portion of the descending aorta. Progressive collateral circulation develops, usually involving the intercostal arteries. It is the dilation of these arteries that eventually causes a notching along the inferior edge of the posterior ribs, most often from T4 to T8 (Fig. 51.82A). This finding usually is not visible until the patient is at least 7 or 8 years of age. Coarctation of the aorta is now frequently diagnosed by echocardiography, P.1271 and further definition of the anatomy is achieved by MR (Fig. 51.82B) .
3308
FIGURE 51.71. Double-Outlet RV With D-Transposition. Sagittal MR shows that both the aorta and pulmonary artery arise from the RV, with the aorta anterior (arrow) .
Hypoplastic left heart syndrome consists of a variety of lesions characterized by some degree of underdevelopment of the left side of the heart. The anomalies range from isolated atresia of the ascending aorta or aortic or mitral valves, to aortic and mitral valve atresia combined with marked hypoplasia of the LA, LV, and ascending aorta. In all cases, blood flow through the left heart is severely impaired and a PDA is necessary to allow blood to reach the systemic circulation. Although the heart size and pulmonary vascularity can appear normal in the first few hours of life, cardiomegaly and congestive heart failure usually develop within the first 2 days. At this point, the pulmonary vasculature becomes passively congested and often a diffusely hazy or reticular pattern develops in the lungs (Fig. 51.83). This pattern signifies interstitial pulmonary edema and resembles that which is seen in other causes of severe pulmonary venous obstruction such as
3309
P.1272 pulmonary vein atresia and TAVPR type 3. The diagnosis can usually be accomplished by echocardiography.
FIGURE
51.72. Total Anomalous Pulmonary Venous Return,
Type 1. A. The characteristic snowman (arrows) or figure 8 configuration results from cardiomegaly combined with prominence of the superior mediastinum because of the anomalous pulmonary vein. B . Cardioangiogram demonstrates the inverted, U-shaped vessel (arrows), which constitutes the upper portion of the snowman.
3310
FIGURE 51.73. Total Anomalous Pulmonary Venous Return, Type 3. The heart is normal in size, with thin and somewhat indistinct pulmonary vessels owing to passive vascular congestion. A prominent interstitial pattern in the lungs represents pulmonary edema.
FIGURE
51.74. Persistent
Truncus
Arteriosus. These
diagrams illustrate the basic forms of the original CollettEdwards classification of persistent truncus arteriosus.
3311
Cor
triatriatum is a rare anomaly that also presents in early infancy
with pulmonary venous obstruction. In this anomaly, the pulmonary veins empty into a common vein, which is abnormally incorporated into the LA. A partial membrane at this site creates an extra chamber along the superior and dorsal aspect of the LA and variably obstructs venous emptying into the LA. The usual radiographic findings consist primarily of cardiomegaly and passive venous congestion without evidence of left atrial enlargement. Primary abnormalities of the myocardium can also present with a normal pulmonary vascular pattern. Cardiomyopathies may accompany a variety of conditions in children. These include bacterial or viral infections, autoimmune diseases, toxic insults, and hereditary neuromuscular diseases. Asymmetric septal hypertrophy is an unusual form of cardiomyopathy that is associated with subvalvular hypertrophic aortic stenosis. Radiographic findings of cardiomyopathy include cardiomegaly that may be generalized or predominantly left sided (Fig. 51.84). The pulmonary vascularity remains normal until congestive heart failure develops. Endocardial fibroelastosis is a condition in which the left ventricular myocardium becomes markedly thickened and contains increased amounts of elastic and fibrous tissue, resulting in marked enlargement of the LV and LA. Radiographically, the heart has a rounded configuration because of the thickened myocardium (Fig. 51.85). The enlarged LV often encroaches on the RV and impairs right ventricular function as well. The LV also can cause left lower lobe atelectasis by compression of the left lower lobe bronchus. Congestive heart failure usually occurs early in infancy in these patients.
3312
FIGURE 51.75. Persistent Truncus Arteriosus. Note oval cardiomegaly, increased pulmonary vascularity, a concave pulmonary artery segment (arrow), and a right aortic arch (A).
3313
FIGURE 51.76. Tetralogy of Fallot. The upwardly displaced cardiac apex caused by right ventricular hypertrophy and the concave pulmonary artery shadow are characteristic of Tetralogy of Fallot.
Cardiac
Malpositions
Cardiac malpositions are a confusing group of abnormalities, and a detailed description of these conditions will not be attempted here. Nevertheless, mastery of the terminology used to describe these conditions can help provide a basic understanding of the anatomy involved. Dextrocardia implies a cardiac apex that lies to the right of the spine because of primary malpositioning during development. Levocardia is the normal position of the cardiac apex, to the left of the spine. When faced with cardiac P.1273 malpositioning, one must then determine the position of the abdominal organs (i.e., the abdominal situs). In general, the RA will lie on the same side as the liver and the left atrium will lie on the
3314
side opposite the liver. Situs solitus refers to the “normal― position of the viscera—that is, the liver on the right and the stomach on the left. The reversed position is referred to as visceral situs
inversus.
FIGURE 51.77. Tricuspid Atresia. A. Typical findings of left ventricular and right atrial enlargement with a concave pulmonary artery segment and decreased pulmonary vascularity are present. B . Angiocardiography shows contrast in the RA, LA, and enlarged LV. A bare area is seen because of lack of filling of a normal RV (arrow) .
Inversion refers to positioning of anatomic structures, usually from right to left and vice versa. The cardiac chambers of a dextroposed heart can be inverted or not. The atria and ventricles can be inverted simultaneously or separately they remain normally related to the LV and the RA to the vice versa, the condition is
and are referred to as concordant if to each other (i.e., the LA is connected RV). If the LA is connected to the RV or referred to atrioventricular discordance.
3315
FIGURE
51.78. Pulmonary Atresia With Intact Ventricular
Septum. In this patient, a nearly normal-sized RV is present. Regurgitation of blood from the RV results in marked right atrial enlargement (arrows) .
Mirror-Image
Dextrocardia
The most common type of cardiac malposition is referred to as mirror-image dextrocardia. In these patients, the cardiac chambers are completely inverted and the cardiac apex points to the right. P.1274 Normal anteroposterior chamber relationships are preserved and there is no discordance. Visceral situs inversus is present, and the incidence of congenital heart disease in such patients is only slightly greater than that in patients with complete situs solitus.
3316
FIGURE 51.79. Ebstein Anomaly. Note the severe cardiomegaly with marked right atrial enlargement. Pulmonary vascularity typically is decreased.
3317
FIGURE 51.80. Aortic Valve Stenosis. The ascending aorta and aortic arch are prominent (arrows) in this 5-year-old child with congenial aortic stenosis.
Dextroversion refers to right-sided rotation of the cardiac position so that the RA and RV become more posterior and the LA and LV lie anterior. Chamber inversion does not occur in this condition. Visceral situs solitus or inversus can be present, and in either case congenital cyanotic heart disease is frequent. Other variations of dextroposition are relatively rare.
FIGURE
51.81. Supravalvular
Aortic
Stenosis
(Williams
Syndrome). Coronal MR demonstrates the short-segment narrowing of the aorta (arrows) just above the sinus of Valsalva.
3318
FIGURE 51.82. Coarctation of the Aorta. (arrows) along the inferior edges of some of bilaterally are caused by enlarged collateral oblique sagittal MR image clearly shows the (arrow) .
3319
A. The small notches the upper ribs vessels. B . A slightly area of coarctation
FIGURE
51.83. Hypoplastic
Left
Heart
Syndrome.
Cardiomegaly and passive pulmonary vascular congestion usually develop within the first few days of life.
FIGURE
51.84. Cardiomyopathy. Note the marked
3320
cardiomegaly with left-sided predominance and early passive vascular congestion in this child with idiopathic cardiomyopathy.
P.1275
Asplenia-Polysplenia
Syndromes
Visceral heterotaxia and congenital heart disease are common components of the cardiosplenic (asplenia or polysplenia) syndromes. Asplenia (Ivemark) syndrome is usually associated with more severe forms of congenital heart disease than the polysplenia syndrome. The liver often lies in the midline, and intestinal malrotation commonly occurs. For simplicity's sake, the asplenia syndrome can be thought of as bilateral right-sidedness (i.e., absent spleen, bilateral threelobed lung, bilateral superior vena cava) (Fig. 51.86), and polysplenia
resembles
bilateral
left-sidedness
(i.e.,
multiple
spleens,
bilateral bilobed lungs, interrupted inferior vena cava with azygos continuation, biliary atresia). Situs inversus with levocardia is an uncommon
occurrence.
Systemic
venous
abnormalities
and
congenital heart disease with right ventricular outflow tract obstruction are common associated abnormalities. US is useful to verify the presence or absence of splenic tissue and to evaluate venous anatomy (4 5) .
3321
FIGURE
51.85. Endocardial
Fibroelastosis. Both the LA
(arrows) and LV are enlarged.
3322
FIGURE
51.86. Asplenia
Syndrome. The liver has a midline
configuration and bilateral horizontal lung fissures are faintly seen (arrows) .
Positional abnormalities of the aorta and great vessels are common, and the components of these anomalies are highly variable. The most important vascular anomalies are those that produce symptoms, mainly those of airway obstruction. These conditions were discussed previously.
REFERENCES
3323
1. Swischuk LE, Hayden CK Jr. Lower respiratory tract infection in children (Is roentgenographic differentiation possible?). Pediatr Radiol 1986;16:278–284. 2. Khamapirad RT, Glezen WP. Clinical and radiographic assessment of acute lower respiratory tract disease in infants and children.
Semin
Respir
Infect
1987;2:130–151.
3. John SD, Ramanathan J, Swischuk LE. The clinical and radiological spectrum of mycoplasma pneumonia. Radiographics 2001;21: 121–131. 4. Brody AS, Klein JS, Molina PL, et al. High-resolution computed tomography in young patients with cystic fibrosis: distribution of abnormalities and correlation with pulmonary function tests. J Pediatr 2004;145:6–7. P.1276 5. Scalzetti EM, Heitzman ER, Groskin SA, et al. Developmental lymphatic disorders of the thorax. Radiographics 1991;11:1069–1085. 6. Nobre LF, Muller NL, de Souza AS Jr, et al. Congenital pulmonary lymphangiectasia: CT and pathologic findings. J Thorac Imaging 2004;19:56–59. 7. Lynch DA, Hay T, Newell JD Jr, et al. Pediatric diffuse lung disease: diagnosis and classification using high-resolution CT. AJR Am
J
Roentgenol
1999;173:713–718.
8. Kim WS, Moon WK, Kim IO, et al. Pulmonary tuberculosis in children: evaluation with CT. AJR Am J Roentgenol 1997;168:1005–1009.
3324
9. Daltro P, Fricke BL, Kuroki I, et al. CT of congenital lung lesions in pediatric patients. AJR Am J Roentgenol 2004;183:1497–1506. 10. Konen E, Raviv-Zilka L, Cohen RA, et al. Congenital pulmonary venolobar syndrome: spectrum of helical CT findings with emphasis on computerized 2003;23:1175–1184.
reformatting.
Radiographics
11. Lucaya J, Gartner S, Garcia-Pena P, et al. Spectrum of manifestations of Swyer-James-MacLeod syndrome. J Comput Assist
Tomogr
1998;22:592–597.
12. Moore AD, Godwin JD, Dietrich PA, et al. Swyer-James syndrome: CT findings in eight patients. AJR Am J Roentgenol 1992;158:1211–1215. 13. Pickhardt PJ, Siegel MJ, Gutierrez FR. Vascular rings in symptomatic children: frequency of chest radiographic findings. Radiology 1997;203:423–426. 14. Katz ML, Konen E, Rozenman J, et al. Spiral CT and 3D image construction of vascular rings and associated tracheobronchial anomalies. J Comput Assist Tomogr 1995;19:564–568. 15. Beckman RP, Hazekamp MG, Sobotka MA, et al. A new diagnostic approach to vascular rings and pulmonary slings: the role of MRI. Magn Reson Imaging 1998;16:137–145. 16. Rosado-de-Christenson ML, Stocker JT. From the archives of the AFIP congenital cystic adenomatoid malformation. Radiographics 1991;11:865–886. 17. Swischuk LE, Shetty B, John SD. The lungs in immature
3325
infants; how important is surfactant therapy in preventing chronic lung problems?. Pediatr Radiol 1996;26:508–511. 18. Swischuk LE, John SD. Immature lung problems: can our nomenclature be more specific? AJR Am J Roentgenol 1996;166:917–918. 19. Fitzgerald P, Donoghue V, Gorman W. Bronchopulmonary dysplasia: a radiographic and clinical review of 20 patients. Br J Radiol 1990;63:514–517. 20. Aquino SL, Schechter MS, Chiles C, et al. High-resolution inspiratory and expiratory CT in older children and adults with bronchopulmonary dysplasia. AJR Am J Roentgenol 1999;173:963–967. 21. Gross GW, Cullen J, Kornhauser MS. Thoracic complications of extracorporeal membrane oxygenation: findings on radiographs and sonograms. AJR Am J Roentgenol
chest
1992;158:353–358. 22. Amodio JB, Abramson SJ, Berdon WE, Solar C, Markowitz R, Kasznica J. Iatrogenic causes of large pleural fluid collections in the premature infant: ultrasonic and radiographic findings. Pediatr Radiol
1987;17:104–108.
23. West MS, Donaldson JS, Shkolnik A. Pulmonary sequestration: diagnosis by ultrasound. J Ultrasound Med 1989;8:125–130. 24. Lee EY, Siegel MJ, Sierra LM, et al. Evaluation of angioarchitecture of pulmonary sequestration in pediatric using 3D MDCT angiography. AJR Am J Roentgenol 2004;183–188.
3326
patients
25. Berrocal T, Madrid C, Novo S, et al. Congenital anomalies of the tracheobronchial tree, lung, and mediastinum: embryology, radiology, and pathology. Radiographics 2004;24:e17. Epub 2003 Nov 10. 26. Lehnhardt S, Winterer JT, Uhrmeister P, et al. Pulmonary sequestration: demonstration of blood supply with 2D and 3D MR angiography. Eur J Radiol 2002;44:28–32. 27. Naffaa LN, Donnelly LF. Imaging findings in pleuropulmonary blastoma. Pediatr Radiol 2005;35:387–391. 28. Han BK, Suh YL, Yoon HK. Thymic ultrasound. I. Intrathymic anatomy
in
infants.
Pediatr
Radiol
2001;31:474–479.
29. Han BK, Yoon HK, Suh YL. Thymic ultrasound. II. Diagnosis of aberrant
cervical
thymus.
Pediatr
Radiol
2001;31:480–487.
30. Koumanidou C, Vakaki M, Theophanopoulou M, et al. Aberrant thymus in infants: sonographic evaluation. Pediatr Radiol 1998;28:987–989. 31. Moeller KH, Rosado-de-Christenson ML. Mediastinal teratoma: imaging features. AJR Am J Roentgenol 1997;169:985–990.
mature
32. Ueno T, Tanaka YO, Nagata M, et al. Spectrum of germ cell tumors: from head to toe. Radiographics 2004;24:387–404. 33. Leung AN, Müller NL, Pineda PR, et al. Primary tuberculosis in childhood: radiographic manifestations. Radiology 1992;182:87–91. 34. McAdams HP, Kirejczyk WM, Rosado-de-Christenson ML, et al.
3327
Bronchogenic cyst: imaging features with clinical and histopathologic correlation. Radiology 2000;217:441–446. 35. Saenz NC, Schnitzer JJ, Eraklis AE, et al. Posterior mediastinal masses. J Pediatr Surg 1993;28:172–176. 36. Lonergan GJ, Schwab CM, Suarez ES, et al. Neuroblastoma, ganglioneuroblastoma, and ganglioneuroma: radiologic-pathologic correlation.
Radiographics
2002;22:911–934.
37. Slovis TL, Meza MP, Cushing B, et al. Thoracic neuroblastoma: what is the best imaging modality for evaluating extent of disease? Pediatr Radiol 1997;27:273–275. 38. Sallustio G, Pirronti T, Lasorella A, et al. Diagnostic imaging of primitive neuroectodermal tumour of the chest wall (Askin tumour). Pediatr Radiol 1998;697–702. 39. Groom KR, Murphey MD, Howard LM, Lonergan GJ, Rosadode-Christensen ML, Torop AH. Mesenchymal hamartoma of the chest wall: radiologic manifestations with emphasis on crosssectional imaging and 2002;205–211.
histopathologic
comparison.
Radiology
40. Boxt LM. Magnetic resonance and computed tomographic evaluation of congenital heart disease. J Magn Reson Imaging 2004;19: 827–847. 41. Gutierrez FR, Siegel MJ, Fallah JH, et al. Magnetic resonance imaging of cyanotic and noncyanotic congenital heart disease. Magn Reson Imaging Clin North Am 2002;10:209–235. 42. Haramati LB, Glickstein JS, Issenberg HJ, et al. MR imaging and CT of vascular anomalies and connections in patients with
3328
congenital heart disease: significance in surgical planning. Radiographics 2002;22:337–347. 43. Choe YH, Kim YM, Han BK, et al. MR imaging in the morphologic diagnosis of congenital heart disease. Radiographics 1997;17: 403–422. 44. Craig JM, Darling RC, Rothney WB. Total pulmonary venous drainage into the right side of the heart: report of 17 autopsied cases not associated with other major cardiovascular anomalies. Lab Invest 1957;6:44–65. 45. Hernanz-Schulman M, Ambrosino MM, Genieser NB, et al. Pictorial essay. Current evaluation of the patient with abnormal visceroatrial situs. AJR Am J Roentgenol 1990;797–802.
3329
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XI - Pediatric Radiology > Chapter 52 - Pediatric Abdomen and Pelvis
Chapter
52
Pediatric Abdomen and Pelvis Susan D. John Leonard E. Swischuk
GASTROINTESTINAL Gastrointestinal
TRACT Obstruction
GI obstruction is a relatively common problem in infants and children and must be distinguished from numerous other causes of vomiting and abdominal distension. The causes of intestinal obstruction in children are widely varied in urgency and management implications, and imaging plays a critical role in the appropriate treatment of such conditions. The most likely causes of obstruction in children shift in importance according to age; therefore, it is helpful to consider the various causes of obstruction within four age-group categories (Table 52.1). Determination of the level of obstruction also helps to infer the possible etiology. Abdominal radiographs continue to be a useful initial screening examination for assessing the site of obstruction and determining the need for further imaging. Hypopharyngeal/upper esophageal obstruction is uncommon in infants and children but may be caused by a spasm of the cricopharyngeus muscle. Cricopharyngeal spasm may be related to neurologic dysfunction (e.g., Chiari malformation, cerebral palsy) or inflammation resulting from gastroesophageal reflux (GER). Lack of normal cricopharyngeus muscle relaxation disturbs the well-
3330
coordinated swallowing mechanism and can lead to aspiration. In refractory cases, surgical division of the muscle may be required. Difficulties with swallowing can also occur with inflammatory processes such as epiglottitis, retropharyngeal abscess, tonsillar abscess, or a number of tumors or cysts that occur in this area. A large pharyngeal diverticulum may produce obstruction. These diverticula can be congenital or iatrogenic owing to perforation of the hypopharynx during intubation. They are best demonstrated with barium swallow.
Esophageal Esophageal
Obstruction Atresia
and
Tracheoesophageal
Fistula The most common congenital obstruction of the esophagus is esophageal atresia (Table 52.2). This anomaly is a result of faulty development and separation of the embryonic foregut early in gestation. The site of atresia is usually in the upper third of the esophagus. The air-filled upper esophageal pouch is often visible on radiographs (Fig. 52.1). The distended proximal esophageal pouch may cause pressure on the trachea during fetal development, resulting in focal tracheomalacia. Esophageal
atresia
is
frequently
associated
with
tracheoesophageal
fistula. Most commonly, the fistula extends obliquely from the trachea, just above the carina, to the distal esophageal pouch. The fistula allows air to enter the stomach and intestines, in some cases in large volumes. Air in the GI tract differentiates this type of esophageal P.1278 atresia from isolated esophageal atresia without tracheoesophageal fistula, in which the stomach and intestines remain gasless. The tracheoesophageal fistula may also develop without esophageal atresia. Such fistulas should be carefully sought on esophagrams performed on infants with choking or respiratory difficulty during feeding (Fig. 52.2). Much less commonly, esophageal atresia may be
3331
accompanied by a fistula from the proximal esophageal pouch to the trachea. A small amount of barium may be placed in the proximal esophageal pouch with an end-hole catheter to demonstrate such fistulas.
TABLE 52.1 Most Common Causes of GI Tract Obstruction by Age
Age 0–1
month
Cause of Obstruction Congenital
anomalies
Atresia/stenosis
Malrotation/volvulus
Hirschsprung
disease
Meconium plug/small left colon syndrome
Meconium
1–5
5
months
months–3
ileus
Hernias
Intussusception
years
3 years and older
Perforated
Adhesions
3332
appendicitis
Regional
enteritis
Esophageal atresia is more common in infants with trisomy 21 and may be associated with vertebral anomalies, duodenal atresia, anal rectal malformations, and other features of the VACTERL syndrome. The prognosis of infants with esophageal atresia depends primarily on the severity of associated anomalies and on the length of the atretic segment. Surgical repair is more challenging with long gap atresia. Common complications of esophageal atresia repair include anastomotic strictures (40%), anastomotic leakage (14% to 21%), and recurrent fistula (3% to 14%). Esophageal peristalsis is abnormal in patients with esophageal atresia, and GER is very common.
FIGURE
52.1. Esophageal
Atresia.
3333
A. Frontal chest radiograph
demonstrates the blind, air-filled upper esophageal pouch (arrows). Note the gas within the stomach and intestines, which indicates the presence of an associated lower tracheoesophageal fistula. B . Lateral view in another patient demonstrates the typical air-filled proximal esophageal pouch (arrows). Esophageal atresia prevents passage of the catheter beyond the pouch. The trachea is compressed and displaced by the dilated pouch.
TABLE 52.2 Causes of Esophageal Obstruction
Congenital
atresia/stenosis
Web/diverticulum Foreign body Stricture
(peptic,
caustic)
Extrinsic compression Achalasia
(cysts,
neoplasms,
vascular)
A variety of communications may exist between the esophageal pouch and the spine, ranging from fibrous bands to actual fistulae. If the communication involutes at both ends and only the central portion remains, a neurenteric cyst results. In other cases, an esophageal communication persists and a diverticulum is formed that extends into the spine or spinal canal (see Fig. 51.59). Faulty separation of the trachea and esophagus may also result in fistulae, fibrous bands, or diverticula. Tracheoesophageal fistula located high in the esophagus without esophageal atresia is the third most common abnormality in this group of lesions, following esophageal atresia with a distal tracheoesophageal fistula, and isolated esophageal atresia. The fistula is usually identified with barium swallow (Fig. 52.2) .
3334
FIGURE 52.2. Tracheoesophageal Fistula. The trachea (T) and esophagus (E) are connected by a fistula (arrow) .
P.1279 Congenital esophageal stenosis is a far less common cause of congenital esophageal obstruction. As in esophageal atresia, esophageal stenosis arises from faulty tracheal and esophageal
3335
separation, where tracheobronchial cartilage remnants remain in the wall of the esophagus. On barium swallow, small diverticula (mucous glands) can be seen in the areas of stenosis. Congenitally short esophagus with intrathoracic stomach is not a truly congenital lesion. Although it is seen at birth, this condition more likely represents the aftermath of chronic hiatal hernia during fetal life, with GER and subsequent esophageal stricture leading to shortening (Fig. 52.3) . Other uncommon congenital causes of esophageal obstruction include esophageal webs and diverticula. Esophageal obstruction can also result from a variety of extrinsic lesions that produce pressure on the esophagus. Acquired esophageal obstructive lesions are primarily strictures or foreign bodies. Esophageal neoplasms are extremely rare in infants and children, and malignant neoplasms are virtually nonexistent.
FIGURE 52.3. Intrathoracic Stomach. A large portion of the stomach lies above the diaphragm (arrows). Note reflux into a
3336
dilated, shortened esophagus with thickened mucosa, suggestive of inflammation.
Peptic esophagitis is associated with GER and can be seen with or without hiatus hernia. Although GER is very common in infants, peptic esophagitis with stricture is a relatively uncommon complication. GER may be primary (chalasia), caused by a lax gastroesophageal sphincter, or secondary to a gastric outlet obstruction. Causes of gastric obstruction (pylorospasm, pyloric stenosis, gastric diaphragm, gastric ulcer disease) must be excluded in infants with severe GER. GER is most reliably identified with 24hour esophageal pH monitoring, but this procedure can be cumbersome and gives no direct information about obstruction. Nuclear gastric reflux studies are also quite sensitive, but barium upper GI series give more anatomic information and are often preferred. US with color Doppler can be used to detect GER, but the technique has not gained widespread popularity (1) . Usually, peptic esophageal strictures are short and located in the distal third of the esophagus. The occasional case of Barrett esophagus with a high stricture may also be encountered. Peptic strictures may be irregular or surprisingly smooth, mimicking the findings of achalasia (Fig. 52.4). Achalasia is uncommon as a cause of distal esophageal obstruction in children. Caustic esophagitis with stricture usually results from accidental ingestion of alkaline substances such as sodium hydroxide, potassium hydroxide (lye), or alkaline disk batteries. Disk batteries can become lodged in the P.1280 esophagus and leak their alkaline contents, producing deep burns of the mucosa and submucosa. All alkaline burns cause deep penetrating injury that commonly results in stricture. Acids, even when swallowed in significant quantities, produce more superficial burns. While mucosal injury may be extensive, deep mural injury with fibrotic stricture is less common. Lye strictures lead to long areas of irregular narrowing (Fig. 52.5). Esophageal burns caused by
3337
an ingested battery or medication (aspirin, tetracycline, Clinitest tablets) result in a more focal stricture.
FIGURE 52.4. Peptic Stricture. The beaklike narrowing of the distal esophagus (arrow) caused by gastroesophageal reflux and stricture
mimics
achalasia.
Epidermolysis bullosa is a hereditary condition characterized by inflammatory skin and mucosal lesions that can heal with fibrosis, resulting
Acute
in
esophageal
stricture.
Esophagitis
3338
Other forms of esophageal inflammation are uncommon in children. Acute inflammation with spasm occurs with infectious esophagitis, caused by organisms such as Candida or herpes. These types of esophagitis are more common in immunocompromised children. Eosinophilic esophagitis is thought to be allergic in nature and is most commonly seen in children with asthma.
FIGURE
52.5. Caustic
(Lye)
Stricture. The long-segment,
irregular configuration of the stricture (arrows) is characteristic of caustic ingestion.
3339
Gastric
Obstruction
Congenital obstructing lesions of the stomach are far less common than congenital obstructing lesions elsewhere in the GI tract (Table 52.3). Gastric distension on radiographs does not always indicate obstruction in infants. A large, gas-filled stomach is commonly seen in normal infants, and persistent asymptomatic gastric distention occurs in infants receiving prostaglandins for ductal-dependent congenital heart disease (2) .
TABLE 52.3 Causes of Gastric Obstruction
Atresia/antral
diaphragm
Duplication cyst Pylorospasm Hypertrophic pyloric Gastritis/ulcer
stenosis
disease
Volvulus Microgastria
3340
FIGURE
52.6. Gastric
Atresia. Typical findings consist of a
dilated, air-filled stomach, with no air distal to the pylorus.
P.1281 Gastric atresia is believed to result from a vascular insult to the stomach in utero. In the newborn infant, if obstruction of the stomach is complete, radiographs show no air distal to the stomach (Fig. 52.6). Gastric atresia usually occurs at the level of the pylorus. In some cases, atresia takes the form of a gastric diaphragm or membrane, which if incomplete, will allow some gas distal to the obstructing web. US or an upper GI series can be used to identify the incomplete
diaphragm.
3341
FIGURE 52.7. Pylorospasm US Features. A. The antrum is contracted. The muscle is slightly thickened but measures less than 3 mm (arrows). B . After treatment with antispasm medication, the antrum opens, peristaltic activity is present, and the muscle has returned to normal thickness (arrows) .
Gastric atresia may occur in infants with congenital epidermolysis bullosa because of inflammatory stricture. Microgastria occurs with other GI atresias or VACTERL syndrome and is commonly associated with the polysplenia/asplenia syndromes. Gastric duplications must be critically located in the antrum or be very large to result in obstruction. They are best demonstrated with US, where they appear sonolucent with a wall that demonstrates both mucosal and muscular layers (see Fig. 52.68) . Gastric volvulus is an uncommon cause of gastric obstruction in children. The volvulus may be idiopathic or may be associated with congenital conditions that involve abnormal position of the stomach, such as diaphragmatic hernia, diaphragmatic eventration, or asplenia syndrome. Gastric volvulus may be classified as organoaxial, in which the stomach rotates along its longitudinal axis, or mesoaxial, in which it rotates about a line perpendicular to the cardiopyloric line.
3342
Gastric volvulus should be considered an acute surgical emergency; however, in some children volvulus may be chronic. Pylorospasm is a reactive problem secondary to insult to the gastric mucosa or muscle contraction from other causes of stress. Mucosal inflammation may be caused by milk allergy or peptic disease with ulceration. Real-time US demonstrates persistent contraction of the antropyloric region and poor emptying of liquids from the stomach. Mild ( Table of Contents > Section XII - Nuclear Radiology > Chapter 56 - Pulmonary Scintigraphy
Chapter
56
Pulmonary
Scintigraphy
David K. Shelton Rhonda A. Wyatt Although CT angiography (CTA) has taken a central role in diagnosing pulmonary embolism (PE), ventilation perfusion (V/Q) scans remain important. A scintigraphic lung scan is a physiologic map that evaluates the primary functions of the lung, pulmonary vasculature perfusion, and segmental bronchioalveolar tree ventilation. Most commonly, V/Q scans are used to evaluate patients suspected of having PE. In an attempt to provide more accurate results, the criteria for interpreting V/Q studies have been constantly revised. Different schema that compare defects present on the perfusion scan with those found on the ventilation scan and/or chest radiograph (CXR) have been developed to estimate the probability of PE. This chapter describes radiopharmaceuticals used, examination technique, imaging protocols, and criteria for the interpretation of V/Q scans.
ANATOMY
AND
PHYSIOLOGY
An understanding of the segmental anatomy of the lungs (Fig. 56.1) is vital to the interpretation of lung scans. The three-dimensional location of individually anatomy of distribution
ventilation or perfusion defects must be determined and correlated with the segmental or subsegmental the lung. PE will have a segmental or subsegmental pattern, usually peripheral and wedge shaped in nature.
3564
Although pulmonary ventilation occurs primarily via the branching bronchial system, other pathways exist by which distal alveoli can be aerated. The pores of Kohn connect adjacent alveoli, and the canals of Lambert connect alveoli with respiratory, terminal, and preterminal bronchioles. These canals and pores permit collateral ventilation of alveoli whose conducting airways have become blocked. Collateral air drift is dynamic and is mediated by neurohormonal control, which can be altered by pathologic events, atmospheric/alveolar gas tension, and drugs. Ventilation and pulmonary blood flow both demonstrate marked gravity effects. When a patient is in an upright posture, the gradient for blood flow is from the apices to the lung bases; the apex receives only one third of the blood volume that the base receives. A corresponding ventilation gradient exists when the patient sits upright. Because intrapleural pressure is greater at the bases, the differential negative intrapleural pressure at the apices causes alveoli there to remain more open at expiration than the basilar alveoli. Therefore, basilar alveoli undergo greater respiratory cycle changes in size. This results in greater gas exchange occurring in the base and greater oxygen tension in the apices. On average, ventilation at the base is 1.5 to 2 times that at the apex. When a patient is supine, the ventilation gradient shifts from superoinferior to anteroposterior, and perfusion is increased to the dependent posterior portions of the lungs. Normally, capillary perfusion and alveolar ventilation are matched to maximize gas exchange. Diseases that produce localized hypoxia invoke autoregulatory mechanisms that divert blood flow away from the hypoxemic pulmonary segments. These dynamic changes prevent P.1372 nonventilated lung segments from being perfused. Conversely, localized hypoperfusion rarely induces localized bronchoconstriction. Primary vascular disorders such as PE, if unassociated with parenchymal consolidation or pulmonary infarction, usually have normal ventilation. Thus an anatomic perfusion deficit with normal ventilation is referred to as a V/Q mismatch and is the hallmark of PE diagnosis.
3565
FIGURE 56.1. Pulmonary Segment Anatomy. Bronchopulmonary segments of the right lung: 1, apical; 2, posterior; 3, anterior; 4, lateral; 5, medial; 6, superior; 7, medial basal; 8, posterior basal; 9, lateral basal; 10, anterior basal. Bronchopulmonary segments of the left lung: 11, apical posterior; 12, anterior; 13, superior lingual; 14, inferior lingual; 15, superior; 16, anterior medial basal; 17, lateral basal; 18, posterior basal. LPO, left posterior oblique; POST, posterior; RPO, right posterior oblique; RAO, right anterior oblique, ANT, anterior; LAO, left anterior oblique; RLAT, right lateral; LLAT, left
3566
lateral. (Adapted with minor modifications from Sostman HD, Gottschalk A. Diagnostic Nuclear Medicine. 2nd ed. Baltimore: Williams & Wilkins, 1988:513; used with permission.)
VENTILATION
LUNG
SCAN
Radiopharmaceuticals Xenon-133 (Xe-133) is a radioisotope widely used to perform ventilation lung scans. A noble gas produced by fission of uranium235 in a nuclear reactor, Xe-133 has a half-life of 5.3 days and is a beta emitter. The principle photon energy is 81 keV, which results in significant attenuation effects. Xe-133 ventilation scans should be performed before perfusion lung scans, because Compton scatter from
the
higher-energy
technetium-99m
(Tc-99m)
macroaggregated
albumin (MAA) (140 keV) down-scatters into the region of the 81-keV photopeak of the Xe-133 and would thus interfere with ventilation images. The usual adult dose of Xe-133 for a ventilation scan is 10 to 20 mCi (370 to 540 MBq). Xenon-127 (Xe-127) is a cyclotron-produced isotope with a physical half-life of 36.4 days and principle photon energies of 203 keV, 172 keV, and 365 keV. Because it has higher-energy photons, Xe-127 ventilation scans can be performed, if needed, after the perfusion scan, since down-scatter image deterioration is not a significant problem. Unfortunately, because Xe-127 is cyclotron-produced, it is both expensive and of limited availability. The usual adult dose is 8 to 15 mCi (296 to 555 MBq). Krypton-81m (Kr-81m) is the other noble gas used for ventilation scans. With an extremely short half-life of only 13 seconds, Kr-81m is produced from a rubidium-81-Kr-81m generator. Kr-81m decays by isomeric transition and has a photon energy of 191 keV. The higher energy allows the ventilation scan to be acquired, if needed, after an abnormal perfusion scan. Unfortunately, the generator is expensive and therefore Kr-81 is limited in use. The usual adult dose is 10 to
3567
20 mCi (370 to 740 MBq). P.1373
Technetium-99m
Aerosols
Ventilation scans can be performed using aerosolized rather than gaseous agents. Radioisotope-labeled aerosols are produced by nebulizing radiopharmaceuticals into a fine mist that is inhaled. Tc99m diethylenetriaminepentaacetic acid (DTPA) is the most commonly used radioaerosol. The advantages of Tc-99m aerosols are that they are widely available and inexpensive, and they have a 140keV photopeak, which is ideal for gamma camera imaging. The nebulizer-produced mist is passed through a settling bag, which traps larger particles. The mist is delivered to the patient via a nonrebreathing valve and is inhaled. The process is inefficient; only 2% to 10% of the aerosolized radioisotope is deposited within the lungs. Of 30 mCi of nebulized Tc-99m DTPA, only 1 to 2 mCi actually are deposited within the lungs. The site of deposition of the aerosolized particles depends on the size of the inhaled particle. The larger the particle, the greater the gravitational effect, which results in more central deposition. Particles larger than 2 µm localize in the trachea and pharynx. Current aerosol nebulizers can produce microaerosols of less than 0.5 µm. Thus, microaerosol particles are small enough to reach the distal tracheobronchial tree and reflect regional ventilation. Patients with narrowed airways caused by asthma, bronchitis, or chronic obstructive pulmonary disease (COPD) have more central deposition of the particles than normal patients because of airway turbulence. This results in poor visualization of the peripheral lung fields. The deposited Tc-99m DTPA is absorbed across the alveolar membrane, with a clearance half-life of 60 to 90 minutes. The half-life is approximately 20 minutes shorter in tobacco smokers, owing to their increased alveolar permeability.
Dosimetry The critical organ for Xe-133 is the trachea, which receives a dose of
3568
0.64 rad/mCi. The lung dose is 0.01 to 0.04 rad/mCi, whereas the whole-body absorbed dose is 0.001 rad/mCi. For Tc-99m aerosols, the lungs receive an absorbed dose of 0.1 rad/mCi, the bladder wall a dose of 0.18 rad/mCi, and the entire body a dose of 0.01 rad/mCi.
Ventilation
Scan
Technique
Ventilation scanning using radioactive gases requires special equipment to prevent leakage of the gas into the imaging room. Gas delivery systems consist of a shielded spirometer, oxygen delivery system, and a Xe charcoal trap to capture most of the exhaled Xe. Because Xe is heavier than air, loose Xe pools at floor level; thus the room should be well ventilated and have negative pressure flow.
Xenon-133
Ventilation
Scanning
The patient is initially fitted with an airtight facemask. While the patient takes a maximal inspiration, Xe-133 is injected into the mask intake tubing. The patient is instructed to hold his or her breath as long
as
possible.
A
posterior-projection,
100,000-count,
first-breath
image of the lungs is then obtained. The ventilation system is then switched so that the patient rebreathes the air/Xe-133 mixture. After 5 minutes of rebreathing, a posterior 100,000-count equilibrium image is obtained. The distribution of Xe-133 activity on the equilibrium image represents aerated lung volume. The ventilation system is then readjusted so that the patient breathes in fresh air and exhales the Xe-133 mixture into the trap. Serial posterior 30second washout images are obtained over a 5-minute interval. The Xe-133 normally washes out of the lungs within 3 to 4 minutes. Because the lung bases are better ventilated than the apices, the Xe133 washes out of the bases faster than the apices in a normal patient. If possible, all images should be performed with the patient in an upright position. Xenon-127 Ventilation Scanning is performed in the same manner as Xe-133 ventilation lung scans. The Xe-127 ventilation scan need be performed only if the perfusion scan is abnormal.
3569
Krypton-81m
Ventilation
Scanning
The high-photon energy of Kr-81m allows ventilation scans to follow perfusion scans. Immediately after each perfusion image and without moving, the patient inhales Kr-81m and the corresponding ventilation image is obtained. This process is repeated until ventilation and perfusion images are obtained in all six matching positions.
Technetium-99m
Aerosol
Ventilation
Scanning The patient inhales the nebulized aerosol while in the supine position to avoid the normal apex-to-base gravity gradient. After inhaling the Tc-99m aerosol for 3 to 5 minutes, the patient sits upright and is imaged in the same projections as for the perfusion lung scan. The exhaled aerosol is trapped in a filter that is stored until decay is sufficient for safe disposal. A Tc-99m aerosol ventilation lung scan can be performed either before or after the perfusion lung scan. If the perfusion scan is performed first, a small dose (0.5 mCi) of Tc-99m MAA is used with a large dose (30 mCi) of Tc-99m DTPA. If the ventilation scan is performed first, 5 to 10 mCi of Tc-99m DTPA and 5 mCi of Tc-99m MAA are administered.
PERFUSION
LUNG
SCAN
Radiopharmaceuticals Perfusion lung scanning is based on the principle of capillary blockade. Particles slightly larger than the pulmonary capillaries (>8 µm) are injected intravenously and travel to the right heart, where venous blood is uniformly mixed. Radiolabeled particles in the pulmonary arterial blood pass into the distal pulmonary circulation. Because the radioactive particles are larger than the capillaries, they lodge in the precapillary arterioles. Their distribution in the lung reflects the relative blood flow to pulmonary P.1374
3570
segments. Pulmonary segments with decreased or absent blood flow show diminished radioactivity. Tc-99m-MAA is the radiopharmaceutical used to perform most perfusion lung scans. MAA is prepared by heat denaturation of human serum albumin. The MAA particles are irregularly shaped molecules, with the size range and number of particles in commercially available kits tightly controlled. Most particles are in the 20- to 40-µm size range, with 90% of the particles between 10 and 90 µm. Particles larger than 150 µm should not be injected, because they can obstruct arterioles. The size and number of particles in a kit are checked by counting a sample volume in a light microscopy hemocytometer. Tc-99m-MAA is prepared by adding Tc-99mpertechnetate (Tc-99mO4 ) to the MAA kit. The MAA leaves the lungs by breaking down into smaller particles that pass through the alveolar capillaries into the systemic circulation, where they are removed by the reticuloendothelial system. The biologic half-life of MAA particles in the lung is 2 to 9 hours. The physical half-life of Tc99m-MAA is 6 hours. A minimum of 60,000 Tc-99m MAA particles must be injected to ensure reliable count statistics and image quality. Typically, 200,000 to 500,000 particles are injected, and fewer than 0.1% of capillaries are temporarily and safely occluded. However, several types of patients should receive a reduced number of particles during a perfusion scan. Patients with pulmonary hypertension and right-toleft shunts should be given only 100,000 particles. Children should also be injected with only 100,000 particles because they have fewer pulmonary arterioles. To perform reduced-count imaging, each perfusion view is imaged for a longer time interval, allowing for nearly equivalent count statistics. Alternatively, the kit can be reconstituted with higher-than-usual Tc-99m activity per particle. The normal 5-mCi dose can be administered but with fewer particles. Contraindications to perfusion lung scanning include severe pulmonary hypertension and allergy to human serum albumin products.
3571
Dosimetry The normal adult dose is 3 to 5 mCi (111 to 185 MBq). The lung is the critical organ and receives an absorbed dose of 0.15 to 0.5 rad/mCi. The whole-body and gonadal absorbed dose is 0.15 rad/mCi.
Perfusion
Scan
Technique
The syringe containing the Tc-99m MAA should be gently agitated prior to injection to resuspend all particles. The patient is injected in the supine position while taking slow, deep breaths to minimize the pulmonary perfusion gravitational gradient. Blood should not be drawn into the syringe because aspirated blood may form clots, which then become labeled by the Tc-99m MAA. Injection of clumped Tc-99m MAA particles or labeled clot can result in multiple small focal hot spots scattered through the lungs. The patient is usually imaged in the upright position using a large field of view and a high-resolution gamma camera. Images (500,000 count) are obtained in the anterior, posterior, right lateral, left lateral, right posterior oblique, left posterior oblique, right anterior oblique, and left anterior oblique positions. Supplemental or decubitus views can be added to clarify findings on the standard views.
VENTILATION/PERFUSION
SCANS
Indications The most common indication for V/Q scans is diagnosis of suspected PE. This examination has also been used to monitor pulmonary function of lung transplants, to provide preoperative estimates of lung function in lung carcinoma patients in whom pneumonectomy is planned (split lung function study), to evaluate right-to-left shunts, and to conduct serial assessment of inflammatory lung disease. The CXR should be evaluated prior to obtaining a V/Q scan. Infiltrates, effusions, pulmonary edema, or pneumothorax may explain sudden
3572
respiratory deterioration and eliminate the need for a V/Q scan.
CT
Angiography
Versus
Ventilation/Perfusion
Scans
There are situations in which a V/Q scan should be the first option to diagnose PE and times when CTA should be the first option. The sensitivity and accuracy of CTA have increased with the use of thincut helical CT and multidetector CT (Fig. 56.2). It should be considered
when
P.1375 the patient is in intensive care or has an abnormal CXR, high clinical probability for PE, or a relative contraindication for anticoagulation. A V/Q scan is highly sensitive and avoids the use of iodinated contrast. It should be considered when the clinical probability is low, the CXR is normal, and the patient is pregnant or has a relative contraindication for iodinated contrast.
FIGURE
56.2. Pulmonary Emboli on MDCT Pulmonary
Angiogram. Scan demonstrates multiple bilateral pulmonary emboli (arrows) and a left pleural effusion (e).
3573
FIGURE 56.3. Normal Ventilation/Perfusion Scan. A. Normal Xenon-133 ventilation lung scan (top two rows). post 1b, posterior initial breath; pos eq, posterior equilibrium; eq, equilibrium; eq3, equilibrium after 3 minutes; wo/1, 1 minute after start of washout; wo/2, 2-minute washout; wo/3, 3-minute washout; wo/4, 4-minute washout. B. Normal Technetium-99m macroaggregated albumin perfusion lung scan (bottom two rows). post, posterior; lpo, left posterior oblique; lt lat, left
3574
lateral; lao, left anterior oblique; rao, right anterior oblique; rt lat, right lateral; rpo, right posterior oblique.
Normal ventilation scans (Fig. 56.3A) have homogeneous radiopharmaceutical distribution throughout all lung fields on all three phases of the scan (initial breath, equilibrium, and washout). A subtle base-to-apex gradient may be seen because more lung parenchyma is located at the base than the apex. The first-breath Xe-133 image is often grainy because it has relatively poor count statistics. However, it still reflects regional lung volume. The equilibrium images have greater activity and will fill in P.1376 areas of restricted lung disease. The washout phase of the study demonstrates rapid clearance of the Xe-133 from the lungs. Normal half-time for xenon washout is less than 1 minute. Washout is complete within 3 minutes. Retention (trapping) of xenon in the lungs in a focal or diffuse pattern is an indication of obstructive lung disease. Normal Tc-99m DTPA aerosol scans resemble normal Tc-99m MAA perfusion scans. However, activity is frequently present within the trachea and mainstem bronchi, especially in smokers. Swallowed Tc99m DTPA aerosol is sometimes seen within the esophagus and stomach. Normal perfusion scans show well-defined margins of both lungs on all views, with sharply defined costophrenic angles. A mild baseto-apex count activity gradient is present because of the physical differences in lung thickness of the base compared to the apex. Tracer distribution should otherwise be homogeneous (Fig. 56.3B) . The heart causes a smoothly defined defect along the left medial lung border that is curvilinear in all projections. A prominent, focal triangular margin suggests the presence of a perfusion defect abutting the heart. The hila are usually seen, even in normal patients. Focal asymmetric hilar perfusion defects are abnormal. Cardiomegaly, tortuosity of the aorta, and mediastinal or hilar
3575
enlargement cause defects along the medial border of the lung associated with less well-defined corresponding defects on the ventilation scan. The size and shape of any mediastinal structure on the V/Q scan should match its appearance on the CXR.
Abnormal
Scans
Focal defects or inhomogeneous tracer distribution are abnormal on either ventilation or perfusion scans. Focal perfusion defects should be compared with corresponding areas on the ventilation scan and vice versa. The relative size and shape of V/Q defects should then be correlated with corresponding areas on a recent CXR. Ideally, the correlative CXR should have been performed no more than 6 to 12 hours prior to the V/Q scan since acute findings may change rapidly. Ventilation scans are abnormal if areas of delayed Xe wash-in or washout are present. Restrictive changes or defects on the singlebreath image may disappear on the equilibrium images when Xe bypasses obstructed pulmonary bronchioles through the pores of Kohn and canals of Lambert. Movement by collateral air drift proceeds more slowly than through the bronchioles, resulting in delayed wash-in and washout. Focal areas of abnormal retention therefore suggest obstructive lung disease (Fig. 56.4) .
PULMONARY
EMBOLISM
PE is a common cause of death in the United States. Dahlen and Alpert estimated that 30% of untreated patients with PE die as a consequence of their emboli, in comparison to 10% to 16% mortality for patients treated with anticoagulant therapy. Anticoagulants, however, place patients at significant risk for life-threatening bleeding and should not be prescribed without high probability for the diagnosis of venous thrombosis or PE. PE usually originate from thrombi within the deep venous system of the legs and pelvis. Predisposing factors include prolonged immobilization, surgery (particularly intrapelvic or hip surgery), history of prior PE, preexisting cardiac disease, estrogen therapy,
3576
smoking, hypercoagulable states such as cancer, and congenital defects of thrombolysis. PE can be difficult to diagnose clinically. In 70% of patients who survive PE, the emboli may not be clinically suspected. The classic triad of dyspnea, hemoptysis, and pleuritic chest pain occurs in fewer than 20% of patients with PE. Larger emboli increase the likelihood of symptoms. Symptoms associated with PE, however, are nonspecific. Pulmonary infection or inflammation, pneumothorax, cancer, and cardiac disease may produce similar symptoms. An electrocardiogram should be performed in patients suspected of having PE to detect cardiac causes for chest pain or dyspnea. If a patient develops acute cor pulmonale because of pulmonary emboli, the electrocardiogram will show signs of right heart strain.
Radiographic Embolism
Findings
of
Pulmonary
The CXR is normal in 12% of patients with PE. The classic findings are a wedge-shaped, pleural-based infarct (Hampton hump) or a wedge-shaped area of oligemia (Westermark sign). The most common but nonspecific CXR finding of PE is atelectasis or opacities in the region with emboli. An elevated diaphragm, small pleural effusion, and/or prominent hilum are also frequently seen. Recently, spiral CT and MR have been used to diagnose pulmonary emboli. The sensitivity of spiral CT is 73% to 95%, with a specificity of 87% to 97%. Spiral CT and MR accurately detect emboli in the segmental or larger pulmonary arteries but may not display more peripheral emboli.
Scintigraphic Thrombosis
Findings
of
Deep
Venous
A radionuclide venogram may be performed in conjunction with a perfusion lung scan. Tc-99m MAA is divided between two syringes and injected into the veins on the dorsa of the feet instead of into
3577
the arm. The nuclear venogram is most sensitive for thrombi occurring above the knees. Deep venous thrombosis (DVT) is indicated by obstruction of the veins, which show cutoff of activity and multiple collateral vessels. Tracers are also utilized to detect acute DVT, such as antifibrin monoclonal antibodies and Tc-99m–labeled peptides. Acute thrombi demonstrate focal areas of asymmetric increased uptake within the deep venous system and could be helpful in differentiating chronic from acute DVT.
3578
FIGURE 56.4. Chronic Obstructive Pulmonary Disease. A. Ventilation scan, posterior projection (top two rows). Obstructive changes in middle and upper lobes cause retention of xenon-133 on 4- minute washout image (post wo/4). post 1b, posterior initial breath; post eq, posterior equilibrium; lpo eq, left posterior oblique equilibrium; rpo eq, right posterior oblique equilibrium; second row: post-washout images at 1 to 4 minutes. B. Technetium-99m macroaggregated albumin perfusion scan (bottom two rows). Labeling is the same as in Fig. 56.3. Patchy, inhomogeneous uptake is seen primarily in the middle and upper lung zones. Perfusion defects match those seen on the initial-breath image of the ventilation scan.
P.1377
VENTILATION/PERFUSION INTERPRETATION
SCAN
Multiple, bilateral perfusion defects with a normal ventilation scan are the classic diagnostic findings of PE (Fig. 56.5).
Pulmonary
emboli that occlude pulmonary arteries produce segmental perfusion defects that extend to the pleural surface. However, pneumonia, COPD, tumors, and prior infarcts may also produce perfusion defects. The ventilation scan is performed to improve the low specificity P.1378 of the perfusion scan. The bronchial tree is unaffected by vascular embolization; thus, ventilation of the embolized region remains normal. Most nonembolic lung diseases have both ventilation and perfusion abnormalities, which are typically matched defects. Pulmonary emboli are more common in the lower lobes because more pulmonary blood flow goes to the basilar pulmonary segments.
3579
FIGURE 56.5. High-Probability Ventilation/Perfusion Scan. A. Xenon-133 ventilation scan (top two rows) is normal. B . Technetium-99m macroaggregated albumin perfusion scan (bottom two rows). The perfusion scan demonstrates absence of perfusion to most segments of the right lung with multiple subsegmental defects in the left lung. Labeling is the same as in Fig. 56.3.
3580
V/Q scan findings are categorized according to the likelihood that emboli will be demonstrated on pulmonary angiography. All interpretation schemas are based on careful analysis of perfusion scan defects to determine whether they correspond to anatomic segments or subsegments of the lung. An understanding of the segmental anatomy of the lung is essential. The shape, location, and size of any defect are analyzed for fit to a specific pulmonary segment on all views. The size of a segmental defect must be assessed. By definition, a defect of less than 25% of a pulmonary P.1379 segment is a small defect, 25% to 75% is a moderate defect, and greater than 75% is a large defect. Subsegmental defects are summed to provide full-segment equivalents. Two moderate or four small perfusion defects are equivalent to a full-segment defect. Even experienced readers tend to underestimate the size of segmental defects.
FIGURE Pleural
56.6. Low-Probability Perfusion Scan With Bilateral Effusions. Scan demonstrates bilateral wedge-shaped
defects that correspond to pleural effusions within the major
3581
fissures bilaterally (open arrows) and the minor fissure on the right (closed arrow). Ant, anterior; post, posterior, rao, right anterior oblique; lpo, left posterior oblique, lao, left anterior oblique; rpo, right posterior oblique; lt lat, left lateral; rt lat, right lateral.
Interpretation schemes compare defects visualized on the perfusion scan with corresponding regions of the ventilation scan and CXR. A perfusion defect that demonstrates normal ventilation is termed a mismatched defect. A perfusion defect the same size and location as a ventilation defect is called a matched defect. Perfusion defects that match ventilation and CXR abnormalities in size and location are called triple match defects. The size and number of matched and/or mismatched segmental defects are used to estimate the likelihood that the defects represent emboli. Nonsegmental defects should be compared to CXRs to determine whether a mass, an effusion, or a mediastinal or hilar structure is responsible for the perfusion scan finding. Non–wedge-shaped defects, or wedge-shaped defects that do not correspond to segmental anatomy, are usually not caused by pulmonary emboli. Common nonsegmental defects include cardiomegaly, pleural effusions (Fig. 56.6), adenopathy, hilar and parenchymal masses, cardiac pacemakers (Fig. 56.7), pneumonia, large bullae, atelectasis, pulmonary hemorrhage, and aortic aneurysm or tortuosity.
Diagnostic
Criteria
The Biello criteria originally categorized V/Q scans as normal, low probability, intermediate, or high probability. The PIOPED (Prospective Investigation of Pulmonary Embolism Diagnosis) study used a modified Biello schema, with more detailed categorizations of V/Q scan patterns. The PIOPED classification has undergone several revisions after retrospective analysis of the data pointed out subcategories of incorrectly classified scan patterns. The amended PIOPED criteria are listed in Table 56.1, with examples provided in Figs. 56.3, 56.6, 56.7, and 56.8.
3582
Stripe
and
Fissure
Signs
Two types of perfusion defects not listed in either the original PIOPED or Biello criteria have been found to strongly correlate with a normal pulmonary angiogram. Central perfusion defects that have a rim or stripe of increased activity around them have a less than 10% probability of being caused by PE. The defect should be seen in different views to not extend to the pleural surface. The surrounding stripe of perfused lung is called the stripe sign. PE perfusion defects extend to the pleural surface and have no overlying stripe of perfused
lung.
Perfusion defects that match the location and shape of the major or minor fissures of the lung usually represent P.1380 pleural effusions tracking up the fissures (Fig. 56.6). When this defect is seen, the lateral view can be repeated with the patient in the supine or decubitus position to demonstrate layering of the fluid. The fissure sign usually correlates with the presence of a pleural effusion on CXR.
3583
FIGURE 56.7. Chronic Obstructive Pulmonary Disease. A. Technetium-99m macroaggregated albumin perfusion (top two rows). Moderate to large bilateral perfusion defects match the ventilation scan defects. A nonsegmental defect is also present over the left upper lobe, representing an artifact secondary to a cardiac pacemaker (arrow). B. Xenon-133 ventilation scan (bottom two rows). Patchy defects are seen in the mid and lower lung zones on the right on the initial breath image (post 1b). The defects partially fill in on the equilibrium images (eq, rpo, lpo). Persistent retention of xenon-133 is seen
3584
in these same regions on the washout images (wo/1, wo/2, wo/3, wo/4). Labeling is the same as in Fig. 56.3.
PIOPED
Findings
The PIOPED study was designed to evaluate the usefulness of V/Q scans for diagnosing acute PE. In the original study, 13% of patients had high-probability V/Q scans, 39% had intermediate-probability scans, 34% had low-probability scans, and 14% showed P.1381 normal or near-normal scans. The interobserver agreement in classifying scans was very good (92% to 95%) for normal/nearnormal scans and high-probability scans, but it was significantly worse for low- and intermediate-probability scans (25% to 30%). The prevalence of thromboembolism in patients who underwent angiography was 33%. The sensitivity of a high-probability scan was 41%, with a specificity of 97%.
TABLE 56.1 Amended PIOPED Criteria
V/Q Scan
Criteria
Category High
Two or more mismatched
Likelihood
Prevalence
of PE
of PE
≥80%
perfusion segments or segmental equivalents without corresponding ventilation or CXR abnormalities:
3585
87%
a. ≥2 large segmental perfusion defects
b. 1 large and 2 moderate segmental defects
c. ≥4 moderate segmental defects
Intermediate
1. One
20%–79%
moderate to 2 large mismatched segments or segmental equivalents without corresponding ventilation or CXR abnormalities
2. Triple matched defects in the lower lung
3586
35%
zone
3. Single moderate matched V/Q defects with normal CXR
4. Corresponding V/Q defects and small pleural
5.
effusion
Findings
difficult
to
classify as normal, high, or low
Low
1.
Multiple
≤19%
matched V/Q defects with a normal CXR
2. Corresponding V/Q defects and CXR opacities (triple matched defects) in the middle or upper lung zones
3587
12%
3. Corresponding V/Q defects and large pleural effusions (more than one third of the hemithorax)
4. Any perfusion defect with substantially larger CXR abnormality
5. Any defect with a rim of surrounding normally perfused lung (stripe
sign)
6. >3 small perfusion defects with normal CXR
7. Nonsegmental perfusion defects
3588
Very low
Table of Contents > Section XII - Nuclear Radiology > Chapter 57 - Cardiovascular System Scintigraphy
Chapter
57
Cardiovascular
System
Scintigraphy
David K. Shelton Michael F. Hartshorne Nuclear medicine applications in the cardiovascular system include gated or nongated myocardial perfusion imaging, myocardial viability studies, infarction imaging, gated ventricular function studies of the blood pool in the ventricles, and detection and quantitation of intracardiac shunts.
MYOCARDIAL
PERFUSION
SCANS
Technique Each of the perfusion agents may be imaged with planar techniques or with SPECT. Meticulous quality control of the stress and rest images is essential. The comparison of images between stress and rest requires identical repositioning so that the same areas of myocardium are visualized. Poor positioning will lead to false-positive interpretations of ischemia and infarct. The three principle coronary artery distributions of the LV are the left anterior descending artery (LAD), the left circumflex artery (LCX), and the posterior descending artery (PDA). Each artery normally provides an equal intensity of myocardial labeling at any given level of cardiac work. Perfusion of the thinner right ventricular wall is considerably less than that of the LV, but it can be imaged using the same techniques (Figs. 57.1 ,
3602
57.2 , 57.3 ). Exercise on a treadmill, or simulation of exercise by infusion of dipyridamole or adenosine, is used in conjunction with perfusion agents to increase radionuclide delivery to the normal myocardium. Stepwise increases in physical exercise are monitored by sequential electrocardiogram (ECG) and blood pressure and pulse measurements while the patient is queried for symptoms of angina. The radiopharmaceutical is injected under conditions of maximal exercise, which should be continued for 30 to 60 seconds after injection to obtain optimal mapping of stress perfusion. Exercise should reach at least 85% of the maximum predicted heart rate to achieve adequate stress. Exercise may also be stopped because of chest pain and ischemic changes on the ECG. Adequacy of the exercise challenge can be more thoroughly estimated simply from a calculation of the “double product― (DP) (systolic pressure × heart rate = DP). The DP correlates with an individual's myocardial work performed, whereas the duration of exercise and heart rate alone may not. For exercise to be judged as adequate, the DP should double, or preferably triple, from rest to peak exercise and should rise to above 20,000. For those patients who cannot perform physical exercise, coronary vasodilatation can be pharmacologically induced. IV dipyridamole or adenosine will vasodilate normal coronary arteries but does not effectively increase flow through vessels with 50% stenosis or greater. Those areas, which cannot dilate normally, will appear to have decreased myocardial perfusion when compared with the rest of the myocardium. IV dobutamine can also be used when dipyridamole or adenosine are contraindicated, such as active bronchospasm. Dobutamine has direct inotropic and chronotropic effects that result in increased coronary flow P.1388 similar to true exercise. Areas of relative hypoperfusion result from coronary stenosis.
3603
FIGURE
57.1. Normal
Exercise/Rest
Planar
Technetium-99m
Sestamibi Myocardial Scan. Anterior (ANT), left anterior oblique (LAO 40, LAO 70), and left lateral (L LAT) planar views of a 380-pound patient, with the upper row representing stress and the lower row representing rest injections of the radiopharmaceutical. Note the superb image quality in spite of the patient's large size.
Image
Acquisition
Planar imaging has largely been replaced by SPECT imaging with reconstruction of the LV myocardium into short-axis, vertical long-axis, and horizontal long-axis planes. A 180° acquisition is generally preferred over 360° acquisition because of the asymmetry of the heart in the thorax and owing to spine attenuation effects in the posterior projections. ECG gated acquisitions are readily accomplished for technetium and thallium radiotracers, allowing evaluation of wall motion, brightening, and thickening from diastole to systole. Functional data acquisition has also become routine, allowing accurate calculations of end-diastolic volume, end-systolic volume, and left ventricular ejection fraction (LVEF). ECG gated planar imaging can still be accomplished for patients who cannot be imaged on the SPECT table (often because of weight).
3604
FIGURE
57.2. Normal
SPECT
Projections.
Short-axis (A) , vertical long-
axis (B) , and horizontal long-axis (C) images in standard projections show the walls of the LV. In the short-axis images, an apical “button― starts the series, which extends back to the base of the ventricle. The names of the walls for the short-axis images are best given by the diagram in Fig. 57.3 . In the vertical long-axis images, the anterior and inferior (or posterior) walls are seen. In the horizontal long-axis images, the short septum and long lateral or “free― walls are well seen. The long-axis images also show the apex very well.
The tomographic images from SPECT have improved the accuracy of myocardial perfusion imaging (MPI) and provide better correlation to other imaging modalities such as echocardiography, CT, and MR. The addition of ECG gated SPECT allows wall motion analysis and functional information, which has improved interpretation and made MPI a more complete examination. Prone imaging is frequently accomplished after the standard supine,
3605
post–stress acquisition, and may help reduce false-positive results caused by breast or diaphragm attenuation, hot bowel loops, or motion artifacts. Attenuation correction can also be accomplished with emission P.1389 sources or on the new SPECT-CT devices. The SPECT camera itself can be a single-, dual-, or triple-headed camera.
FIGURE 57.3. LV Short Axis Vascular Distributions and Wall Names. The schematic diagram locates the expected position of the principle coronary arteries. The left anterior descending artery (LAD) usually serves the apex. The names of the wall segments are listed in a clockwise fashion as anterior, anterior-lateral, lateral, inferior-lateral, inferior, inferiorseptal, septal, and anterior-septal. The LAD sends diagonal vessels (numbered with digits in the order in which they leave the LAD, e.g., D1, D2, etc.) onto the anterior and anterior-lateral walls and septal perforators down into the septum. The left circumflex (LCX) sends obtuse marginal
3606
(OM) branches along the free wall, numbered in sequence (OM1, OM2, etc.) The posterior descending artery (PDA) arises from the right coronary artery (RCA) 85% of the time and serves the inferior wall and the inferior-septal wall.
Radiopharmaceuticals Three gamma-emitting radiopharmaceuticals are readily available for mapping the flow of blood to the myocardium. Each has advantages and some disadvantages. Thallium-201 (Tl-201), an analog of the potassium ion (K+ ), is delivered to capillary beds by regional blood flow and is actively pumped into viable cells by the sodium/potassium (Na+ /K+ ) adenosine triphosphatase pump. Cyclotron production at a remote site (requiring shipping); a long physical half-life (73 hours); low-energy, poorly penetrating photons (mostly 69- to 83-keV γ rays); and a relatively high absorbed dose (0.24 rad/mCi for the whole body at the usual dose of 2 to 5 mCi) combine to make Tl-201 a less-than-ideal agent for imaging. However, because of its active transport into cells, it is a more physiologic radionuclide than the technetium-99m (Tc-99m)–labeled
agents.
A widely used technique utilizes Tl-201 with exercise stress or a pharmacologic challenge. Images are usually acquired as soon after injection as possible. However, some authors advocate waiting for 5 to 10 minutes to allow the exercised patient to stop breathing heavily so that movement of the heart heaving up and down with the diaphragm will be minimized. This slight delay also limits an artifact caused by the “upward creep― of the heart. As the lungs decrease in volume slowly after exercise, the average level of the diaphragm is raised, shifting the heart upward. This shift in location of the heart produces an artifactual shift in radionuclide activity that may be misinterpreted as ischemia. The effective half-life, or 50% washout, of Tl-201 from the normal myocardium is about 4 hours. A complex “redistribution― of the isotope within the myocardium is governed by rates of washout from myocardial cells, renal P.1390
3607
excretion, and shifts of the isotope between muscle, viscera, and other compartments. Rest or redistribution imaging is usually done 3 to 4 hours after the stress injection. Because Tl-201 has significant blood pool activity, it can slowly redistribute into the myocardium and thus slowly fill in ischemic-type defects. In addition to clinical data (ECG, angina, etc.), the initial Tl-201 images of the chest and heart may help assess the heart's performance. High lung activity immediately after exercise usually indicates that left ventricular failure occurred during exercise. Poststress dilation of the heart compared to the resting images is another indicator of failure. Both phenomena have a severe prognosis for subsequent cardiac events (angina, infarction, arrhythmia, and sudden death) (Fig. 57.4 ). Another imaging strategy for improving the visual detection of ischemic myocardium by Tl-201 scintigraphy calls for a “reinjection― of 1 mCi of Tl-201 just before delayed rest imaging. This technique is especially important to view defects caused by very high-grade stenoses, resulting in more accurate diagnosis of ischemia versus infarction.
FIGURE
57.4. Abnormal
Thallium-201
Lung/Heart
Ratio. This frame is
an anterior projection acquired immediately after the start of a stress
3608
SPECT study. The lung:heart ratio of 0.77 is markedly elevated, indicating that the patient experienced heart failure during exercise.
Tc-99m is used to label two commercially available myocardial perfusion agents. Tc-99m
Sestamibi (trade name Cardiolite) is taken up by the perfused
myocardium by passive diffusion and is bound in the myocyte, mostly within myocardial mitochondria. There is no significant redistribution effect with this agent. Washout is negligible. Imaging of the 15- to 20-mCi dose is delayed for 30 minutes to 1 hour after stress to allow for biliary and background clearance. Because there is neither redistribution nor significant washout of Tc-99m sestamibi, a repeat injection of 15 to 20 mCi for resting images is commonly performed on a different day. With this 2day protocol, stress imaging is usually done first. An alternative 1-day approach uses a small dose (8 mCi) for the initial rest scan, followed 4 hours later by the stress scan, with a larger dose of 20 to 25 mCi. Tc-99m
Tetrofosmin (trade name Myoview) is rapidly extracted from the
blood by perfused myocardium in a fashion that resembles Tc-99m sestamibi. The manufacturer claims that it clears background faster and therefore can be imaged sooner than Tc-99m sestamibi. The two agents have proven to act clinically in a very similar manner, but availability and pricing
make
important
considerations.
Both of the Tc-99m–labeled agents are prepared from Tc-99m pertechnetate and stocked pharmaceutical kits. Both are easy-to-image radiopharmaceuticals with good soft tissue penetration (140 keV gamma energy) and a high photon flux from typical doses of 8 to 25 mCi. In addition, the Tc-99m agents also provide perceptibly improved image quality and an opportunity with the same injection to better perform gated first-pass or gated SPECT studies, which can be used to evaluate wall motion, and left ventricular functional parameters such as LVEF.
Dual-Isotope
Myocardial
Scans
An innovative way to maximize the logistical patient throughput involves the use of a Tl-201 and a Tc-99m agent for sequential scans. The most
3609
widely used dual-isotope scan technique uses a resting Tl-201 scan, which can be immediately or subsequently followed by a Tc-99m (sestamibi or tetrofosmin) stress scan. Because the energy and photon flux of the subsequent Tc-99m scan are higher than those of the Tl-201 scan there is no problem with cross talk between the rest and stress images. Excellent scan quality can be combined with 1-day convenience, or a delayed 24-hour thallium scan can be accomplished if needed.
Interpretation Myocardial
Ischemia
Interpretation of myocardial perfusion scans is difficult but important. Subtle abnormalities can signal serious coronary artery disease. Observer knowledge and experience are essential for an accurate diagnosis. Parametric methods of perfusion image analysis have been employed in attempts to standardize diagnosis. Circumferential profiles of isotope distribution and analyses of regional rates of Tl-201 washout, compared with normal databases, make interpretation more sensitive in the detection of ischemia. Displayed as graphic data, “bull's-eye― maps of SPECT images, and three-dimensional reconstruction of SPECT data, these aids in interpretation may be overly sensitive. If an abnormality is truly present, it should also be visible in the planar or SPECT images. Depending on the statistical assumptions used and the population studied, the sensitivity and specificity
for
detecting
P.1391 myocardial ischemia are in the percent range of the high 80s or low 90s. It is important to remember that according to Bayes theorem, the positive and negative predictive values of a test will vary according to the prevalence of disease in the population being tested. The myocardial perfusion scan also detects other causes of ischemia (including left bundle branch block, coronary vasculitis, and small vessel disease) that cannot be seen on coronary arteriography and thereby reduces its apparent specificity. In addition to detecting significant coronary artery disease, the presence and severity of ischemic myocardium correlates strongly with the prognosis for adverse cardiac events, including angina and cardiac death (Fig. 57.5 ).
3610
Myocardial perfusion imaging demonstrates relative regional perfusion. Areas of myocardium with poor blood supply, usually because of atherosclerosis, fail to increase radiotracer uptake during the stress component. The most important feature of the myocardial perfusion test is comparison of the stress and rest images to detect areas of ischemia that are inadequately perfused at exercise yet still viable. These areas are redundantly called reversibly ischemic. Ischemia detected by exercise or pharmacologic dilation of normal vessels usually corresponds with angiographic abnormalities in coronary arteries. Correction of the anatomic abnormality by angioplasty, laser atherectomy, or coronary artery bypass surgery is expected to relieve the ischemia. A frequent location of ischemic tissue is immediately adjacent to an area of infarct. This is called periinfarct ischemia and does not portend the same clinical significance as an ischemic or reversible zone. Some patients will have naturally recruited coronary collaterals or bypass grafts that produce apparent discrepancies between angiographic and scintigraphic studies. Abnormal anatomy in a coronary artery may not produce hemodynamically significant changes in blood flow to the myocardium, and not all ischemia is produced by large vessel atherosclerosis. Capillary disease in diabetics, left bundle branch block, vasospasm, vasculitis, or cardiomyopathy (dilated or hypertrophic) may produce ischemic myocardium even with normal arteries. Ischemia may not P.1392 be detected if there is inadequate exercise, inadequate pharmacologic challenge, or balanced triple-vessel disease. Fortunately, it is uncommon for all three coronary arteries to be hemodynamically compromised equally, and poststress dilation will usually be present.
3611
FIGURE
57.5. SPECT of Left Anterior Descending Artery Reversible
Ischemia. The row of short-axis stress images (A) has a perfusion defect in the anterior wall, which perfuses normally in the rest of the short-axis images (B). This is also visible in the horizontal long-axis stress (C) and vertical long-axis stress images (E) , which have the same perfusion defect. At rest the matched images (D, E, F) show normal perfusion.
3612
FIGURE
57.6. Hibernating
Myocardium. Two vertical long-axis thallium-
201 images at rest (A) are compared with matched images at 24 hours (B). The large anteroapical defect (arrows ) partly fills in over time, indicating that some hibernating (viable) myocardium is present in the midst of what looked initially to be infarcted tissue.
Hibernating
Myocardium
Severe ischemia with high-grade stenosis may be so slow to “reverse― on Tl-201 imaging that it will not be detected by rest or redistribution images at 3 to 4 hours after stress. Imaging at 24 hours, or a Tl-201 second injection rest study, may be required to detect extreme ischemia. The Tc-99m–labeled agents are routinely given as two separate injections, but evidence suggests that rest-injected Tl-201 with delayed imaging may be best for detecting the severe ischemia that leads to a phenomenon known as hibernating myocardium. Hibernating myocardium is important to diagnose, as it simulates infarction by not contracting at rest.
3613
It remains viable, however, and will return to normal function after revascularization (Fig. 57.6 ). Myocardial
infarction produces layers of nonperfused scar tissue that are
detected as areas of thin myocardium with decreased radiotracer uptake at both stress and rest imaging. The extent of an infarct, from subendocardial to transmural, is reflected by the size and degree of the perfusion defect (Fig. 57.7 ). Technical artifacts from attenuation of the perfusion agent's radiation may be produced by SPECT tables, breast tissue, and subdiaphragmatic structures. These may appear as fixed defects superimposed on planar or SPECT images, which may lead to a falsepositive reading of infarction. A false-positive interpretation of ischemia should not occur as long as the artifact does not change between stress and rest imaging. Three techniques are in use to reduce the artifactual appearance of fixed defects seen with myocardial perfusion scanning. The simplest relies on repeat poststress scanning of the patient in the prone position with a Tc agent. This changes the position of the heart, breasts, diaphragm, and subdiaphragmatic organs and reduces the appearance of fixed defects in the appropriate distribution, which may be misinterpreted as infarctions. Repeat prone positioning scans may also help with motion artifacts. Unfortunately, obese patients in whom there is plenty of breast or subdiaphragmatic attenuation may not be able to lie prone for this scan (Fig. 57.8 ). Another technique that avoids misinterpretation of fixed artifactual defects as infarctions relies on gating the acquisition. It is a very simple technique with planar scans and somewhat more involved with SPECT scans. A cine replay of the gated study allows assessment of wall motion. The normal wall moves inward during systole, thickens as it contracts, and becomes brighter on the display. An area in question that demonstrates normal wall motion, brightening, and thickening is probably not infarcted. The elegant solution to the problem of attenuation artifacts has only recently become available. This attenuation correction relies on the simultaneous SPECT acquisition of an emission scan and a transmission scan performed with a radioactive source of a different energy than that used for the emission scan. The transmission scan can also be acquired
3614
with the CT component of SPECT-CT. With a transmission scan, allowance for the emission P.1393 photons lost because of attenuation can be made, and the resulting SPECT scans are surprisingly free of artifacts. A related improvement on this scheme incorporates correction for photons scattered from the emission source but still accepted by the imaging system. A combination of attenuation and scatter correction promises truly quantitative imaging in the future (Fig. 57.9 ).
FIGURE 57.7. Resting Images of Infarcts of the Left Anterior Descending Artery (LAD). Short-axis (1A) and horizontal long-axis (1B) SPECT images show a small anterior LAD infarct (arrows ). This is compared with another patient who has a much larger LAD infarct (2A, 2B) in the same vascular distribution (arrows ). Note that the second patient's infarct extends from the anterolateral wall to and including the septum. The ventricle is also dilated at rest.
3615
Stunned
Myocardium
A single myocardial perfusion scan cannot determine the age of an infarct. Acute infarcts usually appear larger than old infarcts when imaged with Tl201. Temporarily damaged cells around infarcted cells, referred to as “stunned myocardium,― will be hypokinetic or akinetic and will not hold on to the Tl-201 until recovery several weeks later. Thus the defect can appear worse on the rest imaging compared to the stress imaging, in a so-called “reverse redistribution.― The abnormality may revert to normal or shrink as repair occurs.
Infarct-Avid
Scans
Acute infarcts may also be detected with Tc-99m pyrophosphate labeling. Ionized calcium released from myocytes forms dystrophic calcifications with phosphates, and a “hot spot― is formed, marking the infarcted tissue. Antimyosin antibodies labeled with Tc-99m or indium-111 also localize on the fringes of acute infarctions. The need for imaging of acute infarction is clinically infrequent, usually when the patient has left bundle branch block. Contused myocardium is also detected with these techniques.
Positron
Emission
Tomography
Technique PET is more expensive than standard myocardial perfusion imaging but offers the advantages of coincidence imaging, higher-energy photons, efficient attenuation correction, and different radiopharmaceuticals. PET agents can also be imaged on hybrid SPECT cameras or SPECT cameras with heavy collimators. PET scanning with coincidence detection allows high photon flux because collimators are not required. PET scans have higherresolution images and fewer attenuation artifacts than standard MPI. Thus, PET scans may be the gold standard for MPI.
Radiopharmaceuticals
3616
For PET scans, stress testing is usually done with pharmacological agents. Perfusion is usually evaluated with rubidium-82 or ammonia-13 P.1394 P.1395 ( 13 NH 3 ), with comparisons of rest imaging with stress imaging, as in standard MPI. Viability or hibernating myocardium is evaluated with resting injection of fluorine-18 fluorodeoxyglucose (FDG).
FIGURE
57.8. Stress Versus Prone Imaging.
Technetium-99m
sestamibi
imaging with a single headed SPECT camera shows a defect in the inferior wall during stress imaging (arrowheads ), which is not present when the patient is imaged again in the prone position.
3617
FIGURE
57.9. Attenuation
Correction. An apparent anterior wall defect
caused by large breasts is corrected by simultaneous transmission and emission scans. An uncorrected vertical long-axis scan (A) shows an apparent anterior wall defect, which disappears when the transmission scan (B) is used to correct for the asymmetric attenuation (C) . The anterior wall (arrows ) is normal.
3618
FIGURE 57.10. Fluorodeoxyglucose (FDG) PET Myocardial Viability Scan. Tc-tetrofosmin resting scan (A) shows defects in the anterior and inferior walls on male for potential bypass surgery. Fluorine-18 FDG resting PET scan (B) demonstrates normal uptake consistent with fully viable, hibernating myocardium.
Interpretation In evaluation of coronary artery disease, rubidium or ammonia-13 pharmacologic stress imaging is accomplished, and defects, which reverse on rest imaging, are indicative of coronary stenosis. Fixed defects on stress and rest usually identify infarcted myocardium or hibernating myocardium. With FDG imaging, hibernating myocardium will show normal or even
3619
relatively increased FDG uptake as the result of a shift from free fatty acid metabolism to glucose metabolism (Fig. 57.10 ). True infarction will show no significant FDG uptake.
GATED BLOOD POOL SCANS The radionuclide ventriculogram (RVG) is a study that uses circulating, Tc99m–labeled red blood cells to evaluate the size, wall motion, and functional parameters of the LV. RV evaluation is better accomplished by the first-pass study, to be discussed later.
Technique The red blood cells are labeled with Tc-99m, utilizing one of several techniques, and make an excellent blood pool imaging agent. Doses of 20 to 30 mCi are commonly used for typical adult patients. ECG leads are placed to obtain a suitable gating signal (the R wave) for the computer. With the ECG as a measure of the cardiac cycle length, the cardiac cycle is divided into a minimum of 16 frames for analysis of systolic function. Higher temporal resolution of 32 frames per cardiac cycle is required for good measurement of diastolic function. The result of this acquisition is a composite, “averaged― series of images representing the patient's cardiac cycle. Data from a sufficient number of cardiac cycles (several hundred) must be obtained to make the images statistically significant for analysis. Typical acquisition time is 5 to 20 minutes per view (Fig. 57.11 ). Analysis of the functional parameters of the LV, including the LVEF and first derivative (dV/DT; where V is LV volume and T is time) of the LV volume curve, is most accurate from images obtained in the “best septal― left anterior oblique view. This view produces the greatest separation of the activity of the LV from that of the RV.
3620
FIGURE 57.11. Normal Gated Blood Pool Image. An end-diastolic image is shown with a computer-generated region of interest around the LV blood pool. The RV is adjacent.
3621
FIGURE 57.12. Left Ventricular Time Activity Curve. The graph shows this curve (from the patient in Fig. 57.11 ) that displays the ventricle's relative volume during the cardiac cycle. The vertical dashed line represents the relative stroke volume, expressed as an ejection fraction of 62%. The curve begins at end diastole, A marks end systole, B marks the start of diastolic filling, C marks the peak filling rate, D is the end of rapid filling, and E represents the beginning of the atrial “kick.― The horizontal dashed line shows the interval of the first third of diastole, during which more than half of the stroke volume is recovered.
P.1396 Computer processing of the image data by spatial and temporal smoothing algorithms improves both the visual analysis of wall motion and the accuracy and precision of the derived functional parameters. Outlining the edge of the ventricular blood pool in each frame of the study with computerized second-derivative edge detection methods is superior to threshold detection or manually drawn regions of interest. The volume curve is generated by plotting the number of counts in the ventricle versus time during the cardiac cycle. This curve generates the LVEF, which measures the change in volume between end diastole and end systole. The LVEF is the single best parameter of LV function (Fig. 57.12 ). Arrhythmias such as frequent premature beats and atrial fibrillation tend to falsely lower the LVEF. The R-R (R-wave) interval histogram from the ECG can demonstrate the presence of arrhythmias. Most nuclear medicine computer systems allow analysis of selected populations of beats of the same R-R interval to yield a more accurate LVEF. Additional functional parameters are easily obtained. The dV/DT of the LV volume curve gives important information on the rates (average or maximal) of systolic emptying and diastolic filling. Cardiac output (CO) in liters per minute may be calculated if the heart rate, the LVEF, and the left ventricular end-diastolic volume (LVEDV) are known. The product of all three is CO. The LVEDV can be measured by comparing the count rate of a blood sample of known volume with the count rate of the ventricle at end diastole and end systole. Another simpler method to measure CO uses a count-based ratio method.
3622
The ratio of the total counts to the maximum counts in the diastolic frame is entered into an equation that also requires a calibration of the voxel size for the acquisition and depends on constants derived from the formula for a sphere. The resulting measurement of the LVEDV has about the same error as more complicated methods and allows a rapid estimate of CO (Fig. 57.13 ). The exact range of normal for functional parameters of the RVG will depend on multiple factors, such as number of frames acquired, counts within each image, method of computer filtering of the data, and methods of background correction and edge detection. In general, the clinically established normal resting LVEF is approximately 65%, with a standard deviation of 5%. (The normal range of 2 standard deviations is 55% to 75%.)
Interpretation Left
Ventricular
Ejection
Fraction
The most common causes of elevated LVEF values include mitral or aortic valvular regurgitation, hypertrophic cardiomyopathy, and high cardiac output states such as those found in hyperthyroidism. Low LVEF values are usually seen in patients with prior myocardial infarction, ischemia (with congestive heart failure), or cardiomyopathy of any cause. A common application of the RVG is monitoring for the development of cardiotoxicity from
chemotherapeutic
End-Diastolic
drugs.
Volume
The relative end-diastolic size and shape of the RV and LV chambers (RVEDV and LVEDV) should be always noted. Although they appear roughly equal in a normal best-septal left anterior oblique view, the RVEDV is normally greater than the LVEDV. If no intracardiac shunts are present, the stroke volumes of the ventricles are equal because the RV ejection fraction is smaller than the LVEF. As the LV fails for any reason, it dilates and usually becomes rounder in shape. (See Fig. 57.13 for an example of a dilated LV.)
3623
Wall motion of various regions of the LV can be assessed from an overlay of end-diastolic and end-systolic edge images. This is best evaluated by visually observing a cine display of the beating heart in orthogonal views. The left anterior oblique or best septal view is the critical view, but the anterior and left posterior oblique views are complementary. As the ventricular wall is damaged or infarcted, the progression of wall motion abnormality is from normal to hypokinetic to akinetic. If an aneurysm forms, the wall will become dyskinetic. This analysis is true for gated SPECT as well as RVG. To determine the degree of abnormality, it is important to concentrate on the margins of the LV chamber, which is the interface of the endocardial surface and blood. The observer should attempt to correlate a P.1397 suspicion of abnormal wall motion in one view with this same area on the orthogonal view. Color computer displays that enhance the margins of the chambers may make subtle wall motion abnormalities more easily detectable.
3624
FIGURE
57.13. Sample Calculations of Cardiac Output (CO).
Calculation of left ventricular end-diastolic volume (LVEDV) can be done using the count-based ratio method, which requires measurement of the total counts in the end-diastolic region of interest, the maximum pixel counts in the same region of interest, and measurements of the size of a pixel in centimeters. In this case, the dilated LV has a LVEDV of 275 cm3 . Multiplication of the LVEDV by the LVEF and heart rate gives a global CO of 10.97 L/min.
Fourier phase analysis provides powerful additional information on the amount of motion (amplitude) of various LV wall segments and also their relative timing (phase). The amplitude image is especially useful for confirming areas suspected to be hypokinetic or akinetic on the cine display. Damaged areas of myocardium contract with less vigor than normal areas. The phase display may help detect such areas because damaged areas contract slowly (tardive kinesis). Dyskinetic, aneurysmal areas are dramatically displayed using Fourier amplitude and phase images. There is
3625
wall motion of the segment displayed on the amplitude image, but it is opposite (180° out of phase) compared with undamaged areas (Figs. 57.14 , 57.15 ).
Valvular
Regurgitation
Another use of Fourier amplitude images is in the calculation of valvular regurgitation. Each pixel in an amplitude image is coded with a number proportional to the blood volume change under that pixel during the cardiac cycle. A simple total of the pixel values in all the LV and RV pixels outlined with region-of-interest markers will produce a ratio of the LV-to-RV stroke volume. The ratio can be used to calculate the regurgitant fraction. This method works only when there are regurgitant valves on one side of the septum. It cannot distinguish aortic regurgitation from mitral regurgitation, however (Fig. 57.16 ).
Exercise
Radionuclide
Ventriculogram
The RVG study can also be done repeatedly while the patient is exercising on a bicycle ergometer at various workload levels. This is an excellent method to monitor cardiac functional response to exercise. The 2- to 3minute periods for each exercise level usually supply a minimally acceptable amount of statistical counts in the gated images. Finally, a large amount of data must be processed and reviewed because each stage of the study is compared with the resting study (Table 57.1 ). The relative cardiac output can be measured from one stage to another and rises with increasing workload. Normal patients increase or augment their LVEF and dV/DT significantly while decreasing their LV end-systolic volume. Abnormal exercise RVG response can be seen in several ways, such as an increase of LVEDV by more than 10%, lack of increase or even a fall in LVEF with greater workloads, and development of wall motion abnormalities caused by ischemia brought on by the exercise.
RIGHT
VENTRICULAR
First-Pass
Function
STUDIES Studies
3626
Right ventricular function is more difficult to assess by the RVG study than is LV function. This is because labeled activity in the RV cannot be isolated as well from other chambers as can LV activity. RV function is best assessed by analyzing images from the first pass of a radionuclide P.1398 P.1399 bolus through the right-sided chambers and lungs before the overlapping left-sided chambers are seen. The patient is usually imaged in the right anterior oblique projection. A bolus of up to 30 mCi of high-specific-activity isotope must be very rapidly injected, followed immediately by a nonradioactive flush dose. This activity will pass through the RV in three to eight heartbeats. A region of interest is established around the RV and a time-activity curve allows an RV ejection fraction to be measured for each beat. An average RV ejection fraction is then calculated (Fig. 57.17 ).
FIGURE
57.14. Normal Fourier Phase and Amplitude Images. These
3627
are from the same patient shown in Figs. 57.11 and 57.12 . The lower (amplitude) image shows the relative displacement of blood in each chamber of the heart. The pixel brightness depicts the relative degree of motion. The upper (phase) image shows the relative timing of contraction of each chamber. The histogram summarizes the number of pixels with a given phase angle. The cardiac cycle is represented on an arbitrary scale of –90° to 270°. Note that the gray pixels representing ventricular motion are tightly grouped around –30°, indicating synchronous contraction. Approximately 180° up the time scale, there is a cluster of white pixels corresponding to atrial motion.
FIGURE 57.15. Fourier Phase and Amplitude in Left Bundle Branch Block. Two separate populations of phase values are seen in the RV and LV. The lighter-colored RV contracts before the darker-colored LV. This is much easier to see in color.
3628
FIGURE
57.16. Mitral Regurgitation Calculated From a Fourier
Amplitude
Image. The total counts in the LV and RV regions of interest
yield a 1.5:1 LV/RV stroke ratio with a 0.33 regurgitant fraction. This is the same patient imaged in Fig. 57.12 . The global cardiac output (CO) of 10.97 is multiplied by the complement of the regurgitant fraction (0.67) to generate a forward CO of 7.35 L/min.
Another good RVEF technique uses xenon-133 in saline solution for injection. During a slow venous infusion, the xenon-133 passes through the right side of the heart and into the lungs, where it immediately fills the alveoli and is exhaled. In this way, overlapping activity never enters the left side of the heart. A gated study over many seconds is acquired and processed in a manner identical to the standard RVG of the LV. In general, an average RVEF is 42%, with a standard deviation of 5% and a normal range of 32% to 52% (Fig. 57.18 ).
First-Pass
Flow
Studies
3629
The first-pass study in an anterior projection can also be used to detect abnormalities of blood flow to one lung compared with the other. The effect of extrinsic compression on a pulmonary artery by a mediastinal or hilar mass can be easily detected. Abnormal blood flow to a lung segment P.1400 such as is seen in pulmonary sequestration can be detected. The first-pass study can be used to measure the transit time of an injected bolus between ventricles. There is a delay in the passage of blood from the RV to the LV, which typifies congestive heart failure. Obstruction of the superior vena cava is also easily diagnosed in a matter of seconds (Fig. 57.19 ). Rest 78 150 11.7 0.56 1.00 1.00 1.00 1.00 EX1 101 152 15.4 0.60 0.83 0.76 0.89 1.16 EX2 102 158 16.1 0.64 0.90 0.74 1.03
3630
1.35 EX3 106 158 16.7 0.64 0.96 0.79 1.10 1.50 EX4 115 170 19.6 0.68 0.91 0.66 1.11 1.63 EX5 133 190 25.3 0.72 0.77 0.49 1.00 1.70 EX6 153 192 29.4 0.71 0.59 0.39 0.75 1.47
3631
Post-EX 162 162 26.2 0.70 0.69 0.47 0.86 1.79 This exercise radionuclide ventriculogram shows that the patient worked hard during six levels of bicycle exercise (EX1 to EX6) as the heart rate (HR) rose from 78 to 153, while systolic blood pressure (BP) rose from 150 to 192, with a resultant rise in double product (DP) from 11,700 to 29,400. The left ventricular ejection fraction (LVEF) rose appropriately from 0.56 to 0.71 at peak exercise. However, the ventricle filled poorly during exercise as relative end-systolic volume (rLVESV) went progressively down (normal) and relative end-diastolic volume (rLVEDV) also declined (abnormal). The stroke volume (rLVSV) changed little and the only improvement in relative cardiac output (rCO) was the result of increased heart rate. Poor diastolic function (poor compliance) limited exercise endurance in this otherwise healthy individual. This is a superb example of the quantitative data inherent in nuclear medicine images. Level
HR
BP
DP (×1,000)
LVEF
rLVEDV
rLVESV
rLVSV
rCO
TABLE 57.1 Example of Poor Diastolic Function as Measured by Exercise Radionuclide Ventriculogram-Relative Volumes and Cardiac Output
3632
FIGURE 57.17. Right Ventricular First-Pass Function Study. A. Fast dynamic right ventricular ejection fraction by first pass. The acquisition totaled 512 frames taken at 40-ms intervals in the right anterior oblique (RAO) projection as a radioactive bolus traversed the RA and RV. An image of the RV is made by summing dozens of individual frames. A fixed region of interest (ROI) is drawn around the RV. SVC, Superior vena cava; PA, pulmonary artery. B . A time-activity curve from the ROI in (A) shows the relative volume of the ventricle rising and falling with diastole and systole. Peaks and valleys in the curve are flagged, and beat by beat ejection fractions are averaged.
3633
FIGURE 57.18. Xenon-133 Right Heart Gated First-Pass Study. Xenon-133 in saline is slowly infused for a gated first-pass study. It is shown with right anterior oblique end-diastolic and end-systolic frames and their respective regions of interest around the RV blood pool. The RV ejection fraction in this patient with congestive heart failure is only 20%.
Left-to-right intracardiac shunts can be detected and quantified using a first-pass imaging technique. Instead of using a region of interest over the RV for analysis, an area P.1401 of lung is used. In a normal person, the bolus of activity passes into and out of the lung exponentially in a way that can be mathematically described by a gamma function. If a left-to-right shunt is present, some blood that has gone through the lungs to the left side of the heart reenters the right side of the heart and is pumped back into the lungs. This causes a prolongation of the washout of activity from the lung region of interest. A gamma-variate curve-fitting method can be used to detect and quantify the
3634
amount of the left-to-right shunt. The method is sensitive to detect P.1402 shunts with a ratio as low as 1.2:1, far below the 2:1 shunt that can be detected by chest radiograph (Fig. 57.20 ).
FIGURE
57.19. Superior Vena Cava (SVC) Obstruction. A first-pass
study with 1-second frames is shown in the anterior projection after injection in the right antecubital vein. Serpiginous collateral veins on the chest wall probably communicate with the intercostal and azygos veins. Very little flow courses through the SVC into the RA and RV (arrow ). The patient required stenting of the SVC to relieve obstruction caused by an encircling tumor.
3635
FIGURE 57.20. Abnormal Left-to-Right Shunt Study. A. Regions of interest are drawn around the superior vena cava (SVC, square box ) and the right lung (R lung) on image data from a first-pass flow study. Note lack of activity in the LV (arrow ) in this summary of images from the right heart phase of the flow. B . Graph showing time-activity curve of the activity within the two regions shown in (A). A is the sharp bolus injection passing through the superior vena cava. B is the right lung time activity curve, which rises exponentially but does not follow the fitted gamma variate curve (C) on the way down. This indicates early recirculation owing to a left-to-right shunt. The shunt is quantified by comparing the area under C with the area under the fitted recirculation gamma variate (D).
3636
FIGURE 57.21. Abnormal Right-to-Left Shunt Study. A significant portion of the injected technetium-99m macroaggregated albumin (MAA) particles are seen in capillary beds outside the lungs in the brain and kidneys. This indicates and measures the amount of shunted blood.
Right-to-left
shunts can be detected by using an IV injection of
macroaggregated albumin particles. In a normal person, less than 10% of the injected dose should pass through normal arteriovenous shunts in the lungs and be found in the systemic circulation. After injection, static images of the patient's whole body are obtained. Regions of interest are taken over the lungs, head, neck, abdomen, and extremities. The amount of radioactivity outside the lungs in the systemic circulation is then quantified. The study can be repeated at a later date to check progression (Fig. 57.21 ).
SUGGESTED
READINGS
3637
Anagnostopoulos C, Harbinson M, Kelion A. et al. Procedure guidelines in radionuclide myocardial perfusion imaging. Heart 2004;90:1–10. Bax JJ, Cornell JH, Visser FC, et al. Comparison of fluorine-18-FDG with rest-redistribution thallium SPECT to delineate viable myocardium and predict functional recovery after revascularization. J Nucl Med 1998;39:1481–1486. Bax JJ, Patton JA, Poldermans D, et al. 18-Fluorodeoxyglucose imaging with position emission tomography and single photon emission computer tomography: cardiac applications. Semin Nucl Med 2000;30:281–298. Beller GA. Clinical value of myocardial perfusion imaging in coronary artery disease. J Nucl Cardio 2003;10:529–542. Berman DS, Kiat H, Friedman J, et al. Separate acquisition restthallium-201/stress Tc-99m sestamibi dual isotope myocardial SPECT: a clinical validation study. J Am Coll Cardiol 1993;23:1455–1464.
perfusion
Berman DS, Kiat H, van Train K, et al. Myocardial perfusion imaging with technetium-99m-sestamibi: comparative analysis imaging protocols. J Nucl Med 1994;35:683–688.
of
available
Bolli R. Myocardial “stunning― in man. Circulation 1992;86:1671–1691. Bonow RO. Identification of viable myocardium. Circulation 1996; 94:2674–2680. Crean A, Dutka D, Coulder R. Cardiac imaging using nuclear medicine and position emission tomography. Radiol Clin North Am 2004;42(3):619–634.
3638
Demer LL, Gould KL, Goldstein RA, et al. Assessment of coronary artery disease severity by positron emission tomography. Comparison with quantitative arteriography in 193 patients. Circulation 1989;79:825–835. DePuey EG, Berman DS, Garcia EV. Cardiac SPECT Imaging. 2nd ed. Philadelphia: Lippincott Williams and Wilkins, 2001. Dilsizian V, et al. Myocardial viability in patients with chronic coronary artery disease. Comparison of 99mTc-sestamibi with thallium reinjection and [18F] fluorodeoxyglucose. Circulation 1994;89:578–587. Eitzman D, Al-Aouar Z, Kanter H, vom Dahl J, et al. Clinical outcome of patients with advanced coronary artery disease after viability studies with positron emission tomography. J Am Coll Cardiol 1992;20:559–565. Gibbons RJ, Balady GJ, Bricker JT, et al. ACC/AHA 2002 guidelines update for exercise testing. Circulation 2002;106; 1883–1892. Go RT, Marwisk TH, MacIntyre WJ, et al. A prospective comparison of rubidium-82 PET and thallium-201 SPECT myocardial perfusion imaging utilizing a single dipyridamole stress in the diagnosis of coronary artery disease. J Nucl Med 1990;31:1899–1905. Green MV, Bacharach SL. Functional imaging of the heart: methods, limitations, and examples from gated blood pool scintigraphy. Prog Cardiovasc Dis 1986;28:319–348. Hayes SW, De Lorenzo A, Hachamovitch R, et al. Prognostic implications of combined prone and supine acquisitions in patients with equivocal or abnormal supine myocardial perfusion SPECT. J Nucl Med 2003;44:1633–1640.
3639
He ZX, Iskandrian AS, Gupta NC, et al. Assessing coronary artery disease with dipyridamole technetium-99m tetrofosmin SPECT: a multicenter trial. J Nucl Med 1997;38:44–48. Jones RH. Use of radionuclide measurements of left ventricular function for prognosis in patients with coronary artery disease. Semin Nucl Med 1987;17:95–103. Kapur A, Latus KA, Davies G, et al. A comparison of three radionuclide myocardial perfusion tracers in clinical practice: the ROBUST study. Eur J Nucl Med 2002;29:1608–1616. Kiat H, Berman DS, Maddahi J. Myocardial perfusion imaging using technetium-99m
radiopharmaceuticals.
Radiol
Clin
North
Am
1993;31:795–815. Kiat H, Van Train KF, Friedman JD, et al. Quantitative stressredistribution thallium-201 SPECT using prone imaging: methodologic development and validation. J Nucl Med 1992;33:1509–1515. Leppo JA. Dipyridamole myocardial perfusion imaging. J Nucl Med 1994;35:730–733. Lette J, Lapointe J, Waters D, Cerino M, Picard M, Gagnon A. Transient left ventricular cavitary dilation during dipyridamole-thallium imaging as an indicator of severe coronary artery disease. Am J Cardiol 1990;66:1163–1170. Loong CY, Anagnostopoulos C. Diagnosis of coronary artery disease by radionuclide myocardial perfusion imaging. Heart 2004;90(suppl V)V2:V9. Miller DD, Younis LT, Chaitman BR, et al. Diagnostic accuracy of dipyridamole technetium-99m-labeled sestamibi myocardial tomography
3640
for detection of coronary artery disease. J Nucl Cardiol 1997;4:18–24. Parker DA, Karvelis KC, Thrall JH, Froelish JW. Radionuclide ventriculography: methods. In: Gerson MC, ed. Cardiac Nuclear Medicine. New York: McGraw Hill, 1991:81–125. Robinson VJB, Corley JH, Marks DS, et al. Causes of transient dilatation of the left ventricle during myocardial perfusion imaging. AJR Am J Roentgenol
2000;174:1349–1352. P.1403
Rozanski A. Referral bias and the efficacy of radionuclide stress tests: problems and solutions. J Nucl Med 1992;33:2074–2079. Santoro GM, Sciagra R, Buonamici P, et al. Head-to-head comparison of exercise
stress
testing,
pharmacologic
stress
echocardiography
and
perfusion tomography as first-line examination for chest pain in patients without history of coronary artery disease. J Nucl Cardiol 1998;5:19–27. Schelbert HR. Assessment of myocardial viability with positron emission tomography. In: Iskandrian AE, Van der Wall EE, eds. Myocardial Viability.
Dordrecht:
Kluwer
Academic
Publishers,
2000:47–72.
Sharir T, Germano G, Kavanagh PB, et al. Incremental prognostic value of post-stress left ventricular ejection fraction and volume by gated myocardial perfusion single photon emission computed tomography. Circulation 1999;100:1035–1042. Schwaiger M. Myocardial perfusion imaging with PET. J Nucl Med 1994;35:693–698. Smanio OE, Watson DD, Segalla DL, et al. Value of gating of technetium-99m sestamibi single-photon emission computed
3641
tomographic imaging. J Am Coll Cardiol 1997;30:1687–1692. Stewart R. Schwaiger M, Molina E, et al. Comparison of rubidium-82 positron emission tomography and thallium-201 SPECT imaging for detection of coronary artery disease. Am J Cardiol 1991;67:1303–1310. Tamaki N, Ohtani H, Yamashita K, et al. Metabolic activity in the areas of new fill-in after thallium-201 reinjection: comparison with positron emission tomography using fluorine-18 deoxyglucose. J Nucl Med 1991;32:673–678. Tamaki N, Takahashi N, Kawamoto M, et al. Myocardial tomography using technetium-99m tetrofosmin to evaluate coronary artery disease. J Nucl Med 1994;35:594–600. Treves S, Parker JA. Detection and quantification of intracardiac shunts. In: Strauss HW, Pitt B, eds. Cardiovascular Nuclear Medicine. St. Louis: CV Mosby, 1979:148–161. Van Train KF, Garcia EV, Maddahi J, et al. Multicenter trial validation for quantitative analysis of same-day rest-stress technetium-99m-sestamibi myocardial tomograms. J Nucl Med 1994;35:609–618. Watanabe K, Sekiya M, Ikeda S, et al. Comparison of adenosine triphosphate and dipyridamole in diagnosis by thallium-201 myocardial scintigraphy. J Nucl Med 1997;38:577–581. Wijns W, Vatner S, Camici P. Hibernating myocardium. N Engl J Med 1998;339:173–181. Williams B, Millani N, Jansen D, Anderson B. A retrospective study of the diagnostic accuracy of a community hospital-based PET center for the detection of coronary artery disease using rubidium-82. J Nucl Med
3642
1994;35:1586–1592. Yamagishi H, Shirai A, Yoshiyama M, et al. Incremental value of left ventricular ejection fraction for detection of multivessel coronary artery disease in exercise 201 Tl gated myocardial perfusion imaging. J Nucl Med 2002;43:131–139. Zaret BL, Beller GA. Clinical Nuclear Cardiology. 3rd ed. Philadelphia: Elsevier
Mosby,
2005.
Zaret BL, Wackers FJT. Nuclear cardiology (1). N Engl J Med 1993;329: 775–783. Zaret BL, Wackers FJT. Nuclear cardiology (2). N Engl J Med 1993;329: 855–863.
3643
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XII - Nuclear Radiology > Chapter 58 - Endocrine Gland Scintigraphy
Chapter
58
Endocrine
Gland
Scintigraphy
Marc G. Cote
THYROID Imaging
Methods
Diagnosis and treatment of thyroid disease require the evaluation of thyroid function, anatomy including palpatory findings with or without thyroid US, and tissue characterization of thyroid lesions. Radionuclide scintigraphy and measurement of radioactive iodine uptake (RAIU) form the basis of functional assessment of the thyroid. Radionuclide scintigraphy is used to assess the physiologic function of the gland and to determine the presence or functional status of thyroid nodules post–fine-needle aspiration (FNA). Thyroid imaging is most commonly indicated to evaluate hyperthyroidism. Solitary palpable thyroid nodules are best evaluated initially with FNA. Radionuclide scintigraphy utilizing iodine123 (I-123) is useful in differentiating substernal thyroid from thymus glands. Normal thyroid parenchyma appears relatively homogeneous with technetium-99m-pertechnetate (Tc-99m-O4 ) or I-123 scintigraphy. Iodine is trapped via active transport and organified onto the tyrosine contained in intrathyroidal thyroglobulin within thyroid follicles. Tc-99m-O4 is only trapped and will subsequently wash out of the gland since it is not organified. I-123 is the agent of choice for thyroid imaging, especially when
3644
imaging nodules (Table 58.1 ). Tc-99m-O 4 is best reserved for imaging hyperthyroid patients in conjunction with an I-131 RAIU measurement, in which the percentage of the administered dose present in the thyroid gland is measured at a specific time after oral administration, usually at 4 and 24 hours. The functional status of a thyroid nodule may be categorized as hyperfunctioning (“hot―), hypofunctioning (“cold―), or indeterminate (sometimes called “warm―) relative to the normal parenchymal uptake of radioiodine. The term “warm― is misleading to clinicians and should not be used. Hot nodules usually represent hyperfunctioning adenomatous tissue and are rarely malignant. Although solitary cold nodules are hypofunctioning adenomatous tissue in approximately 40% of cases, they may harbor malignancy in up to 15% of cases. Indeterminate nodules have the same significance as cold nodules. The term “warm― should be avoided since it is easily misunderstood by the referring physician to have the same clinical significance as a “hot― nodule. Indeterminate nodules are caused by normal activity overlying or surrounding a hypofunctioning cold nodule. Tc-99m-O4 is inexpensive but has the disadvantage of a lower target-tobackground ratio. Also, if a nodule is hot with Tc-99m-O4 , an additional I123 study must be performed to exclude a discordant nodule. A discordant nodule demonstrates increased Tc-99m-O4 uptake but decreased I-123 uptake and thus potentially harbors malignancy. Discordant nodules still have the ability to trap Tc-99m-O4 but have lost their ability to organify iodine. Because pertechnetate imaging is performed 4 to 6 hours after administration, initial trapping of the radiopharmaceutical may reveal uptake that is isointense or increased relative to normal parenchyma. Imaging of I123 is performed at 18 to 24 hours after administration; therefore, any iodine that may have been trapped has time to wash out of the gland prior P.1405 to imaging, thereby revealing the true nature of the nodule. I-123 13 hr 159 Physiologic
3645
Good organ-to-background ratio Same dose can be used for imaging and uptake Expensive Image 4 hours after administration I-131 8 days 364 Inexpensive Widely available Long half-life High radiation dose per mCi High-energy photon Unsuitable for gamma camera imaging Whole-body scans used for evaluation of residual thyroid and metastatic disease in patients with thyroid cancer Tc-99m 6 hr 140 Inexpensive Excellent imaging qualities Requires separate dose of I-123 or I-131 for uptake measurements Must repeat imaging with I-123 if hot nodule found I, Iodine; Tc, technetium. Isotope
Halflife
Principal γ ray (keV)
Advantages
Disadvantages
Comments
TABLE 58.1 Radiopharmaceuticals Used for Thyroid Imaging It is possible to detect nonpalpable abnormalities using a gamma camera with a pinhole collimator. Abnormalities smaller than 1 cm cannot be reliably resolved because of the inherent limitations of the Anger camera. Thyroid US is rapidly replacing this use of scintigraphy because studies that compared palpation of the gland to US have demonstrated the relative insensitivity of
3646
palpation to small nodules. RAIU measurement served as a measure of thyroid function for many years prior to the development of laboratory assays. The development of accurate serologic methods of measuring serum levels of thyroid hormones and ultrasensitive third- and fourth-generation thyroid-stimulating hormone (TSH) assays has provided superior methods of evaluating thyroid function. Serum TSH is the single best test for screening thyroid function. Only in cases of suspected pituitary or hypothalamic disease is the TSH assay alone insufficient for screening thyroid functional status. Measurement of the RAIU is usually indicated for one of three reasons: (1) differentiation of Graves disease (uptake high, usually >35% at 24 hours) from subacute or factitious hyperthyroidism (uptake usually 50% cystic) or well-defined peripheral calcification as seen at US are unlikely to be malignant. Regression of nodule size following thyroid hormone therapy is a sign of a benign nodule. Large, predominantly solid nodules with irregular contour and poor margination on US examination are likely to be malignant. Five-year survival rates with treatment are ~90% to 95%. The histologic types of thyroid malignancy are as follows: Extensive cystic component Multiple nodules Hot on radionuclide scan Peripheral calcification Shrinkage in size following levothyroxine suppression hormone therapy Sudden onset Female gender Older patient TABLE 58.2 Signs Suggesting Benign Etiology of Thyroid Nodules
3658
Imaging findings Solid nodule Cold on radionuclide scan Irregular contour Poor margination Size >4 to 5 cm Clinical findings Hard on palpation History of neck irradiation Age 1.5 cm. Some disagreement now exists in the literature on the treatment of patients with I-131 if the primary tumor is Table of Contents > Section XII - Nuclear Radiology > Chapter 59 - Gastrointestinal, Liver/Spleen, and Hepatobiliary Scintigraphy
Chapter
59
Gastrointestinal, Liver/Spleen, and Hepatobiliary Scintigraphy David K. Shelton Michael F. Hartshorne
GASTROINTESTINAL
STUDIES
Nuclear medicine imaging studies can provide considerable information in the functional evaluation of the GI system. Routine studies include hepatobiliary, GI bleeding studies, and gastric emptying measurements. Other procedures that are less frequently ordered provide clinically valuable information.
Salivary
Scanning
A quick look at the salivary glands of the mouth is frequently performed in conjunction with the technetium-99m pertechnetate (Tc-99mO4 ) scan of the thyroid gland. The salivary glands can be scanned intentionally with Tc-99mO4 . A 5- to 10-mCi dose is injected intravenously with planar images performed immediately and after a delay, during which lemon juice is washed around the mouth. In the past, this study has been used to grade the severity of Sjögren syndrome, as inflammation degrades the secretory function of the glands. The salivary scan can demonstrate salivary obstruction and can be used as an adjunct to or replacement for sialography. Stimulation of the glands with lemon juice is important to document
3671
the drainage of saliva through the duct system to the mouth (Fig. 59.1). A salivagram can also be accomplished by swabbing a child's mouth with radiotracer. Delayed images of the lungs can be used to check for aspiration in children suspected of swallowing disorders.
Esophageal
Imaging
The esophageal transit study, performed with swallowed solutions or solid boluses labeled with Tc-99m sulfur colloid (Tc-SC), is an examination that can be done in lieu of esophageal manometry. It has been reported to detect esophageal dysmotility in 50% of symptomatic patients with an otherwise normal evaluation for dysphagia. In the supine or upright position with a gamma camera placed in the anterior projection, the patient swallows a radiolabeled bolus, and dynamic data are obtained via computer. The esophagus is divided into three regions of interest (ROIs): upper, middle, and lower. Transit times are then calculated from time-activity curves representing the ROIs. The normal esophagus demonstrates sequential activity from proximal to distal, with no visualized esophageal activity remaining after 10 seconds. Regional analysis may differentiate between achalasia and scleroderma. It is important to remember that esophageal scintigraphy is functional and does not provide detailed anatomic information. A barium or endoscopic study is necessary to exclude the possibility of neoplasm or infection as the cause of impaired esophageal function (Fig. 59.2) .
Gastroesophageal
Reflux
The evaluation of heartburn and atypical chest pain in the adult commonly raises the clinical question of gastroesophageal reflux (GER) disease. In the pediatric population, failure to thrive and recurrent pneumonia often elicit the same question. A common diagnostic tool currently used in the diagnosis of GER disease is acid reflux monitoring. This examination unfortunately P.1417 requires
nasoesophageal
intubation
and
3672
24-hour
continuous
recording. It is invasive and unwieldy, especially in the pediatric age group.
FIGURE 59.1. Warthin Tumor by Salivary Scan. Immediate (A) and delayed (B) images after IV administration of technetium-99m pertechnetate show extra uptake in a palpable mass in the right parotid gland. Retention of the pertechnetate is prolonged in the mass (arrows), even after lemon juice stimulation of the salivary glands. This finding is characteristic of a functioning Warthin tumor, which is not drained by salivary ducts.
GER scintigraphy is performed with acidified orange juice mixed with Tc-SC. The acid decreases the lower esophageal sphincter pressure and also delays gastric emptying. ROIs are established via computer to correspond to the stomach and the segments (upper, middle, and lower) of the esophagus. In the pediatric population, ROIs over the lungs detect aspiration by the end of the study or on delayed imaging 1 to 3 hours later. Images may be recorded in adults with an abdominal binder that increases abdominal
3673
P.1418 pressure sequentially in 10-mm Hg increments to a maximum of 100 mm Hg. Normal patients have no detectable GER activity. This examination is reported to have a 90% sensitivity in the detection of GER (Fig. 59.3) .
FIGURE
59.2. Normal Esophageal Transit Study. A. A
composite image of the esophagus and stomach is used to generate regions of interest around the upper, middle, and lower esophagus. B . Time-activity curves for each region are displayed for 10 seconds after the swallow. Inspection of the curves allows calculation of the transit time.
Gastric
emptying is a complex physiologic process directed not only
by neuroendocrine processes but also by a host of local factors. Food type, pH and fatty content, as well as food osmolality affect the rate of gastric emptying. Impaired gastric emptying can be caused by many disease states, such as diabetes mellitus, electrolyte disturbances, postvagotomy syndromes, and by some medications. Exclusion of mechanical obstruction is important in diagnosing the cause of the patient's symptoms. Endoscopy or barium studies are superior in the detection of gastric ulcers, tumors, or bezoars. Gastric emptying scintigraphy has become the gold standard in the
3674
clinical evaluation of gastric motility. It is a simple test to perform, although interpretation is based upon complicated mathematical models. Solid food, liquids, or both are labeled with a radiotracer and consumed by the patient. Digital images of the stomach are acquired, and a time-activity curve is generated for graphic analysis of the rate of emptying (Fig. 59.4) . The normal half-emptying time (T1 / 2) of radioactive solids and liquids varies with the technique employed. In general, the normal T1 / 2 is less than 90 minutes for solids and less than 60 minutes for liquids. Each laboratory should establish its own normal T1 / 2 values. Liquid gastric emptying usually follows an exponential curve, whereas P.1419 solid emptying is biphasic, with an initial lag phase followed by a linear curve. Visual interpretation of the stomach images is invaluable in understanding the “number― generated for T1 / 2. The gastric emptying technique can be extended to study and time the transit between stomach and colon to characterize disorders of the bowel's smooth muscle or enteric nervous system.
3675
FIGURE
59.3. Abnormal
Gastroesophageal
Reflux
Study.
Three regions of interest (ROIs) are established over the esophagus. A time-activity curve corresponding to the upper ROI for 60 minutes shows refluxed activity after 3 minutes, which continues for about 30 minutes.
3676
FIGURE 59.4. Normal Solid Gastric Emptying Study. In the image at upper left, a region of interest (arrow) is established around the stomach. The next two graphics show the stomach's time-activity curve. The bottom two graphics show the linear best-fit to this data. The slope of the curve allows calculation of the half-emptying time (T1 / 2) of 48 minutes.
Carbon-14
Urea
Breath
Test
The carbon-14 urea breath test (C-14 UBT) is an inexpensive, accurate, noninvasive test for active infection with Helicobacter pylori. H. pylori is a gram-negative, spiral-shaped bacterium that is found in the gastric mucosa or adherent to the epithelial lining of the stomach. H. pylori is known to cause more than 90% of duodenal ulcers and up to 80% of gastric ulcers. It has also been associated with an increased incidence of gastric carcinoma and gastric lymphoma. For the UBT, a patient needs to be fasting and off antibiotics, bismuth, and proton pump inhibitors. A 1-µCi capsule of C-14–labeled urea is administered by mouth, and 10 minutes later a breath sample is collected. H. pylori contains an enzyme, urease,
3677
which hydrolyzes the urea into ammonia and C-14–labeled carbon dioxide, which is subsequently detected in the breath sample using liquid scintillation techniques. This can be accomplished locally, or the sample can be sent into a central laboratory. The UBT has a sensitivity and specificity of 94% to 98%. Alternative tests include endoscopic biopsy to identify the bacterium or serology titers, which will remain positive even after treatment.
FIGURE 59.5. Upper GI Bleeding Caused by Gastric Varices. Sequential 5-minute images of the abdomen, in a patient with negative endoscopy, show tagged red cells filling the lumen of the stomach (arrow). This diagnosis cannot be made if there is free technetium-99-pertechnetate (Tc-99mO4 ) mixed with the red cells; Tc-99mO4 is excreted physiologically in the stomach. The tagged red blood cells used had no free Tc-99mO4 .
3678
Gastrointestinal
Bleeding
Scintigraphy
Patients who present with clinically suspected upper GI bleeding are usually evaluated and often treated by endoscopy. Scintigraphy is not usually needed. The patient with suspected lower GI bleeding presents different problems in diagnosis and therapy. Proctosigmoidoscopy can exclude hemorrhoidal bleeding, and colonoscopy may identify the cause. An emergent GI bleeding study with in vitro Tc-99m–tagged red blood cells is very sensitive and can locate the bleeding site. Continuous 1-minute dynamic frames are acquired over the abdomen and pelvis in the anterior projection for at least 90 minutes or until the patient has bled enough to locate the source of hemorrhage. This technique can detect bleeding rates as low as 0.1 mL/min versus contrast angiography, whic h detects 1 mL/min of GI bleeding. P.1420 The study should be done emergently during the clinical period of suspected active bleeding. Repeated imaging can be done at any time up to 24 hours after red cell labeling, but care should be taken because blood may have moved into the colon.
3679
FIGURE
59.6. Splenic Flexure Bleeding in the Colon Caused
by Diverticular Disease. Sequential 5-minute images show a small focus of bleeding (arrow) in the left upper quadrant that varies with intensity as it accumulates and moves through the colon. Note that the patient has a large, static blood pool in the penis. Unchanging blood pools such as the aorta (A), the inferior vena cava (IVC), and penis (P) should not be misinterpreted as hemorrhage.
Accurate hemorrhage localization is required for resection of a bleeding site. If angiographic therapy for a bleeding site is proposed, the more sensitive GI bleeding study should be done first. The GI
3680
bleeding study will guide selection of appropriate vessels for embolization or infusion of vasoactive drugs. Positive GI bleeding studies demonstrate three cardinal findings: (1) an abnormal hot spot of radiotracer activity appears “out of nowhere― as it enters the bowel lumen; (2) this activity persists and may increase with time; and (3) the activity moves with peristalsis antegrade, retrograde, or in both directions (Figs. 59.5, 59.6, 59.7) .
Meckel
Scan
A Meckel diverticulum, which contains ectopic gastric mucosa, may ulcerate and bleed. Tc-99mO4 is given intravenously and the abdomen is imaged immediately and for 1 hour's worth of dynamic images. Tc-99mO4 localizes in the gastric mucosa and can be used to detect the acid-producing mucosa in the diverticulum. A focus of activity representing the ectopic gastric mucosa in the middle or right lower quadrant of the abdomen is detected as it concentrates the Tc-99mO4 in synchrony with the stomach. Detection may be enhanced by the use of pentagastrin to stimulate uptake or cimetidine to block the outflow of Tc-99m-O4 from the diverticulum (Fig. 59.8) .
LIVER
AND
Liver/Spleen
SPLEEN
STUDIES
Scan
Liver/spleen scanning is performed by IV injection of Tc99m–radiolabeled albumin or sulfur colloid. Colloid imaging provides information based upon organ perfusion and the distribution of reticuloendothelial cells, which phagocytize the colloid particles. Kupffer cells in the liver and reticuloendothelial cells in the spleen are normally imaged. Reticuloendothelial cells in the bone marrow are minimally seen. The liver/spleen scan is an inexpensive and easy means to evaluate for focal or diffuse hepatic disease, but it lacks disease specificity. Radiotracer uptake may be abnormal in a multitude of diseases. To make matters worse, hepatic lesions smaller than 1 cm in diameter are routinely missed, even with
3681
SPECT. MR, CT, P.1421 P.1422 and US have better resolution for hepatic masses. Tc-SC liver SPECT, however, can be very specific in diagnosing focal nodular hyperplasia (FNH). Lesions that are large enough to be identified on SPECT and are isointense or hotter than liver parenchyma on uptake may be confidently diagnosed as FNH. This is because of the presence or increased concentration of reticuloendothelial cells within the lesion. Tc-SC scanning can also be useful in diagnosing masses outside of the liver as myelolipomas or extramedullary hematopoiesis.
FIGURE 59.7. Cecal Bleeding Caused by Angiodysplasia. Images show a right lower quadrant hemorrhage (arrow) .
3682
FIGURE
59.8. Meckel
Diverticulum. A small focus (arrow) of
technetium-99-pertechnetate uptake gradually becomes visible the ectopic gastric mucosa of a Meckel diverticulum in the
in
midabdomen.
Liver/spleen radionuclide imaging remains accurate and easy for evaluation of liver and spleen size, configuration, and position. This helps in the evaluation of suspected hepatomegaly in patients with obstructive lung disease causing diaphragmatic flattening or in patients with anatomic variants, such as a large left liver lobe or a Riedel lobe on the right (Fig. 59.9). Alterations in perfusion and reticuloendothelial system function caused by cirrhosis and hepatitis are seen as a “shift― of activity to the spleen, bone marrow, and lungs. The liver/spleen scan provides information that helps monitor the disease process and efficacy of therapy (Fig. 59.10) .
3683
FIGURE 59.9. Normal Liver/Spleen Scan. begin with an anterior projection with a lead dots) on the right costal margin. Subsequent anterior, right anterior oblique, right lateral,
Sequential images marker (row of cold images are right posterior
oblique, posterior, left posterior oblique, left lateral, and left anterior oblique from left to right, top to bottom. Note the
3684
homogeneous labeling of the liver and spleen and the relative size and position of these two organs in various projections.
Liver/spleen scans can be “subtracted― from other nuclear medicine studies to provide spatial information about the liver or spleen in relation to a suspected abnormality. Indium-111 leukocyte scans (for infection); gallium-67 scans (for inflammation, lymphoma, or hepatoma); indium-111 octreotide scans (for neuroendocrine tumors); and labeled antibody scans have physiologic uptake in the liver and/or spleen. Subtracting the liver/spleen scan from any of these scans confirms “hot― abnormalities adjacent to the liver or spleen (Fig. 59.11). This may be particularly useful in cirrhotic livers with regenerating nodules.
Heat-Damaged Red Blood Cell Scan for Splenic Tissue Tc-99m–labeled red blood cells that have been damaged by heating are preferentially extracted from circulation by splenic tissue. Applications include diagnosis P.1423 of polysplenia and splenosis and confirmation of accessory splenic tissue.
3685
FIGURE 59.10. Abnormal Liver/Spleen Scan in a Patient With Cirrhosis. The liver is small and labels poorly. The left lobe of the liver (L) is better seen than the right lobe (R). Note the “colloid― shift of the radiopharmaceutical to the bone marrow and spleen. Ascites separate the liver from the right ribs (arrowheads). Compare these images with Fig. 59.9.
3686
FIGURE 59.11. Liver/Spleen Subtraction From a Gallium-67 Scan in a Patient With a Hepatoma. A. Image of gallium-67 (Ga-67) distribution in the anterior projection at 48 hours. B . Matched image of colloid distribution (Tc-99mSC). Careful selection of the gamma camera energy windows allows simultaneous imaging of the two radiopharmaceuticals. C . The subtraction image shows the gallium-avid hepatoma, which does not label with the colloid.
HEPATOBILIARY
IMAGING
Nuclear imaging of the gallbladder and biliary system is easily performed with Tc-99m–labeled iminodiacetic acid compounds. Only two of the numerous iminodiacetic acid radiotracers that have been developed are still commercially available: Tc-disofenin and Tcmebrofenin. Tc-HIDA (lidofenin) is actually no longer commercially available. These radiopharmaceuticals are excreted unchanged into the biliary system and work even in the presence of elevated serum
3687
bilirubin.
Acute
Cholecystitis
Hepatobiliary scans are most commonly used to evaluate suspected acute cholecystitis. A minimum of 2 hours’ fasting is recommended in preparation for this scan. The anterior dynamic images of normal hepatobiliary scans show prompt and homogeneous uptake of the radiopharmaceutical by the liver. Liver activity decreases progressively as the radiotracer is excreted into the biliary system and drains into the small bowel. The activity should be seen in the major extrahepatic ducts, P.1424 gallbladder, and small bowel within 1 hour (Fig. 59.12). Most patients with acute cholecystitis have a stone or stones obstructing the cystic duct. A small minority of patients, usually the chronically ill, have acalculous cholecystitis. The hallmark of acute cholecystitis by cholescintigraphy is nonvisualization of the gallbladder at both 1- and 4-hour intervals after IV injection of the biliary agent, or 30 minutes after morphine administration. Chronic cholecystitis is diagnosed when the gallbladder is not visualized at 1 hour but is seen by 4 hours. When properly done, the nuclear medicine hepatobiliary examination has a sensitivity and specificity of 98% and better than a 95% accuracy rate in the diagnosis of acute cholecystitis. Small doses of IV morphine (1 to 2 mg) can be used during the scan to raise the pressure on the sphincter of Oddi. This helps push radiolabeled bile into the gallbladder. It is a handy way to speed up a “normal scan,― because the diagnosis of acute cholecystitis is excluded as soon as the gallbladder is seen. Morphine administration may also allow true negative scans to be performed in patients who have stimulated their gallbladders to contract by eating before the scan.
3688
FIGURE 59.12. Normal Hepatobiliary Scan. Images of the liver immediately after injection and at subsequent 5-minute intervals show rapid clearance of the blood pool followed promptly by central biliary duct and gallbladder (straight arrow)
3689
activity. Activity continues to fill the common bile duct (arrowheads) at 20 minutes and the small bowel (curved at 25 minutes.
arrow)
Increased blood flow on radionuclide angiograms of the gallbladder fossa aids in the diagnosis of acute cholecystitis. A “rim sign― on hepatobiliary scan images is seen as a band of increased activity around the gallbladder fossa, which represents poor excretion of radiotracer from inflamed hepatocytes. The rim sign is usually associated with gangrenous cholecystitis (Fig. 59.13). A pitfall in interpretation of acute cholecystitis may be caused by prolonged P.1425 fasting with gallbladder distension. The radiopharmaceutical will not enter the completely full, atonic gallbladder. This can be avoided by pretreating the patient with analogs of cholecystokinin (CCK). CCK is a short-acting, natural hormone that causes prompt gallbladder contraction. After emptying, the gallbladder refills, allowing entry of the biliary agent. A false-positive diagnosis of acute cholecystitis may also occur with previous cholecystectomy, tumor obstructing the cystic duct, and agenesis of the gallbladder. Mirizzi syndrome can be suspected when there is evidence of acute cholecystitis (nonvisualization of gallbladder) and common bile duct obstruction.
3690
FIGURE
59.13. Acute
Cholecystitis
Diagnosed
by
Hepatobiliary Scan. The first hepatobiliary scan in this patient was positive for acute cholecystitis but the referring service did not believe the diagnosis. The scan was repeated on the next day. In it, the first image shows radioisotope left over from that first scan. A dim line of activity in the transverse colon is marked (arrowheads). Cholecystokinin (CCK) was used as a pretreatment for the second scan. Starting with the second image, the dose is rapidly concentrated and excreted by the liver into bile ducts and small bowel. The gallbladder never fills. As the liver clears, a “hot rim― of activity is seen around the gallbladder fossa (arrows), which indicates inflammation caused by the severe acute
cholecystitis.
3691
Acalculous biliary disease includes chronic acalculous cholecystitis, cystic duct syndrome, and gallbladder dyskinesis. These patients present with similar complaints of right upper quadrant pain, fatty food intolerance, and epigastric distress. Routine cholescintigraphy and US may be normal. CCK-assisted cholescintigraphy in acalculous biliary disease demonstrates decreased gallbladder contraction and decreased gallbladder ejection fraction. The normal gallbladder ejection fraction is greater than 35%. Other uses for the hepatobiliary scan include the detection of postoperative complications and bile leaks in trauma (Fig. 59.14) . The excretion phase of the scan is important in evaluating hepatic and common bile duct patency. A delay of more than 1 hour in visualization of the bile ducts suggests obstruction. Caution must be exercised to differentiate severe hepatocellular disease from obstruction, as both present with delays in biliary visualization. The hepatobiliary scan maintains a niche in differentiating neonatal hepatitis from biliary atresia. In the latter case, radiolabeled bile will never enter the bowel or gallbladder. Unfortunately, if the hepatitis is bad enough, the radiopharmaceutical may not leave the liver. When bowel activity is seen (and this may take 4 to 24 hours), the diagnosis of biliary atresia is excluded.
HEPATIC Cavernous
BLOOD
POOL
SCINTIGRAPHY
hemangioma is the most common benign hepatic tumor
and the second most common hepatic tumor after metastatic disease. Frequently subcapsular in location, this tumor is often found incidentally by US, CT, or MR. Although there are specific criteria for diagnosis of hepatic hemangioma by these techniques, no study is 100% specific. Because a significant risk of hemorrhage exists with biopsy, a noninvasive diagnostic approach is preferred. Scintigraphy with Tc-99m–labeled red blood cells using an in vitro labeling technique has proved both sensitive and specific for cavernous hemangioma. A flow study should be performed initially and will demonstrate normal or decreased early uptake if the suspected lesion is a hemangioma. Tumors and inflammatory lesions tend to
3692
have increased arterial flow. Subsequent delayed SPECT imaging will reveal foci of increased activity within the P.1426 liver that are hotter than surrounding parenchyma. Sensitivity decreases with lesions smaller than 1.5 cm, greater organ depth, and single-detector SPECT as opposed to multidetector SPECT. Correlation with a second imaging technique is always advised because these lesions may be seen concomitant with malignancy. Specificity is generally high, although isolated cases of increased activity on delayed imaging with both colon and lung carcinoma metastases have been reported. In general, if two of the four imaging techniques demonstrate characteristic features of cavernous hemangioma, no further evaluation is warranted (Fig. 59.15) .
FIGURE 59.14. Biliary Leak After Cholecystectomy Detected by a Hepatobiliary Scan. Images (left to right) obtained immediately, 30 minutes, and 1 hour after administration of the biliary agent show accumulation of bile in the area around the right lobe of the liver (arrows) .
3693
FIGURE
59.15. Hepatic
Hemangioma. At first glance, this
study could be confused with a biliary scan. However, note the vascular blood pool and that these are four coronal slices from a SPECT study of the liver performed 1 hour after the injection of 25
mCi
of
technetium-99m-pertechnetate–labeled
red
blood
cells. The area of increased activity (arrow), which simulates a gallbladder, is actually a hemangioma in the liver of this hepatic transplant
recipient.
PET FLUOROXYDEGLUCOSE IN GASTROINTESTINAL CANCERS PET with fluorodeoxyglucose (FDG) is an extremely powerful tool in the evaluation and staging of GI tumors but is covered more fully in chapter 64. It has a higher sensitivity (95% to 100%) than CT (81% to 92%) for detecting esophageal cancer and helps in radiotherapy planning. Likewise, PET has better sensitivity and specificity in nodal staging and in detecting distant metastatic disease. PET has also
3694
been approved for the initial staging and restaging of colorectal carcinoma, with sensitivity and specificity of 95% to 99% for recurrent colorectal cancer. It may prove useful in pancreatic and gastric carcinoma evaluation as well. PET-CT has the ability to improve the anatomic mapping of functional data and improves the accuracy of detection of metastatic abdominal adenopathy, adrenal lesions, and hepatic involvement. In general, the positive predictive value of PET-CT for metastatic disease in the liver is so high as to militate against the necessity for biopsy in most cases. “Incidental― findings of focal uptake in the colon on other PET studies usually indicate cancerous lesions or premalignant lesions such as adenomatous polyps, which require further evaluation.
SUGGESTED
READINGS
Balon HR, Fink-Bennett DM, Brill DR, et al. Procedure guideline for hepatobiliary scintigraphy. Society of Nuclear Medicine. J Nucl Med 1997;38:1654–1657. Charron M. Pediatric inflammatory bowel disease imaged with Tc99m white blood cells. Clin Nucl Med 2000;25:708–715. Charron M, Di LC, Kocoshis S. CT and 99mTc-WBC vs colonoscopy in the evaluation of inflammation and complications of inflammatory bowel diseases. J Gastroenterol 2002;37:874–875. Chatziioannou SN, Moore WH, Ford PV, Dhekne RD. Hepatobiliary scintigraphy is superior to abdominal ultrasonography in suspected acute cholecystitis. Surgery 2000;127:609–613. Connolly LP, Treves St, Bozorgi F, O’Connor SC. Meckel's diverticulum: demonstration of heterotopic gastric mucosa with technetium-99m-pertechnetate SPECT. J Nucl Med 1998;39:1458–1460.
3695
P.1427 Donohoe KJ, Maurer AH, Ziessman HA, Urbain JL, Royal HD. Procedure guideline for gastric emptying and motility. Society of Nuclear Medicine. J Nucl Med 1999 Jul;40:1236–1239. Ford PV, Bartold SP, Fink-Bennett DM, et al. Procedure guideline for gastrointestinal bleeding and Meckel's diverticulum scintigraphy. Society of Nuclear Medicine. J Nucl Med 1999 Jul;40(7):1226–1232. Hustinx R. PET imaging in assessing gastrointestinal tumors. Radiol Clin North Am 2004;42:1123–1139. Kamel EM, Thumshirn M, Truninger K, et al. Significance of incidental 18F-FDG accumulations in the gastrointestinal tract in PET/CT: correlation with endoscopic and histopathologic results. J Nucl Med 2004;45:1804–1810. Klein HA. Esophageal transit scintigraphy. Semin Nucl Med 1995;25:306–317. Krishnamurthy
S,
Krishnamurthy
GT.
Cholecystokinin
and
9 9 mTc-HIDA
morphine pharmacological intervention during cholescintigraphy: a rational approach. Semin Nucl Med 1996;26:16–24. Mariani G, Boni G. Barreca M, et al. Radionuclide gastroesophageal motor studies. J Nucl Med 2004;45:1004–1028.
Mauree AH, Krevsky B. Whole-gut transit scintigraphy in the evaluation of small-bowel and colon transit disorders. Semin Nucl Med 1995;25:326–338.
3696
Maurer AH. Gastrointestinal bleeding Nucl Med 1996;26:43–50.
and
cine-scintigraphy.
Semin
Nadel HR. Hepatobiliary scintigraphy in children. Semin Nucl Med 1996;26:25–42. Szepes A. Bertalan V, Varkonyi T, et al. Diagnosis of gallbladder dyskinesia by quantitative hepatobiliary scintigraphy. Clin Nucl Med
2005;30:302–307.
Tripathi M, Chandrashekar N, Kumar R, et al. Hepatobiliary scintigraphy: an effective tool in the management of bile leak following laparoscopic cholecystectomy. Clin Imaging 2004;28:40–43. Urbain J-L, Charkes ND. Recent advances in gastric emptying scintigraphy.
Semin
Nucl
Med
1995;25:318–325.
Zuckier LS, Freeman LM. Selective role of nuclear medicine in evaluating the acute abdomen. Radiol Clin North Am 2003;41:1275–1288.
3697
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XII - Nuclear Radiology > Chapter 60 - Genitourinary System Scintigraphy
Chapter
60
Genitourinary Scintigraphy Mike
System
McBiles
Peter W. Blue
RENAL
IMAGING
Renal radionuclide imaging has always been an important part of the practice of nuclear medicine. US, CT, and MR have the clear advantage of superior anatomic images. Despite advances in US functional imaging with the refinement of Doppler techniques, and in MR with newer dynamic protocols, there remain many areas in which scintigraphy remains the easiest, least expensive, most accurate test, or the least radiation-burdensome test available. Radiopharmaceuticals are well suited for evaluating blood flow, nephron function and mass, and collecting system excretory function and
drainage.
Radionuclide renal studies are safe, minimally invasive, and expose patients to radiation doses comparable to or lower than those used in competing radiologic procedures. They are used to study such disorders as renal collecting system obstruction, vesicoureteral reflux, renovascular hypertension, renal masses, and pyelonephritis. Renal function parameters such as renal flow, differential function, glomerular filtration rate (GFR), and effective renal plasma flow (ERPF) can be determined with radioisotopes. In patients with
3698
radiographic contrast sensitivity or contraindications to MR, radionuclide studies can frequently answer the clinical question, despite their limited spatial resolution.
Radiopharmaceuticals Radionuclide renal imaging involves the assessment of the four major renal functions: blood flow, glomerular filtration, tubular function, and excretory system function. The nuclear physician will choose a radiopharmaceutical tailored for the specific function or combination of functions to be studied. Only with an understanding of the sequential dynamic movement of the tracer from the blood to the bladder is accurate interpretation possible. The following radiopharmaceutical doses should be adjusted for body surface area for pediatric patients, with minimum doses as per published recommendations.
Blood
Flow
Renal perfusion can be assessed using any radioisotope that is given intravenously in sufficient dosage to acquire a usable series of 3- to 5-second images. Flow images are usually obtained using the commonly accepted doses of technetium-99m (Tc-99m) renal agents: diethylenetriamine pentaacetic acid (Tc-99m-DTPA) (380 to 860 MBq), mercaptoacetyltriglycine (Tc-99m-MAG3 ) (380 to 570 MBq), or glucoheptonate (Tc-99m-GH) (570 to 860 MBq). Glomerular filtration is a passive pumping of plasma components through the semipermeable glomerular membrane into the Bowman space. Plasma water and those molecules small enough to pass the membrane (e.g., electrolytes, creatinine, urea, etc.) are filtered and, if not fully reabsorbed, reach the renal collecting system. In general, proteins are too large to be filtered. Tc-99m-DTPA is small enough and not significantly protein bound (5% to 10%) and so is almost completely filtered. Its lack of protein binding increases its extravascular distribution and consequently its “background― presence, causing a lower target-to-background ratio than Tc-99m-
3699
MAG3 or Tc-99m-dimercaptosuccinic acid (Tc-99m-DMSA), especially at lower levels of renal function. Because there is no significant reabsorption of Tc-99m-DTPA, and because Tc-99m-DTPA does not reach the urine by any other mechanism than glomerular filtration, it traces this function well P.1429 and can be used to accurately measure the GFR. Although diffusion across the glomerulus is passive, the glomerular permeability and preglomerular pressure must be maintained for filtration to continue. Renal artery stenosis, microvascular disorders, and cardiac pump failure decrease the preglomerular pressure. Renal obstruction and acute tubular necrosis increase the postglomerular (Bowman space) pressure. In either case, the net pressure across the glomerulus drops, decreasing the GFR and the amount of tracer (Tc-99m-DTPA) passing through the kidney. Intrinsic diseases such as glomerulonephritis
(chronic
renal
failure)
disrupt
glomerular
membrane permeability, resulting in diminished filtration and tracer movement. Tc-99m-GH is a glucose analog that is mostly glomerularly filtered; 90% of the total dose reaches the urine, whereas 10% is retained by the proximal tubular cell, allowing for delayed cortical imaging and some functional imaging. Tubular
secretion is an active process by which certain molecules are
removed from the peritubular capillaries and secreted into the glomerularly filtered tubular urine that is passing by. For a radionuclide to effectively trace tubular function, most or all of the tracer must be removed in one pass by the tubule; that is, it must be secreted into the urine and not be reabsorbed. Until recently, iodine131-orthoiodohippurate (I-131-OIH) was used because of its high tubular extraction efficiency and lower protein binding, and there is an extensive literature on its use. However, because of inferior spatial resolution owing to its I-131 component, and its high kidney dosimetry in situations of prolonged cortical retention, it has been replaced by Tc-99m-MAG3 . Tc-99m-MAG3 , with a 5% glomerular/55% tubular extraction efficiency, a high (90%) protein binding, which limits
extravascular
distribution,
and
3700
relatively
high
target-to-
background ratio, has become the agent of choice for most tubular functional imaging. Because the clearance of Tc-99m-MAG3 (300 to 400 mL/min) (extraction efficiency = 60%) is so much greater than that of Tc99m-DTPA (80 to 140 mL/min) (extraction efficiency = 20%), Tc99m-MAG3 is the agent of choice for imaging kidneys in moderate to severe renal failure, immature kidneys, and transplant kidneys, in which renal function is often in flux. In practice, because glomerular function and tubular function generally parallel each other, Tc-99mMAG3 has replaced Tc-99m-DTPA in many clinics, unless glomerular function analysis is specifically requested or comparison with legacy Tc-99m-DTPA exams is needed.
Cortical
Imaging
While both Tc-99m-GH and Tc-99m-DMSA can be used, Tc-99mDMSA has minimal urinary excretion ( Table of Contents > Section XII - Nuclear Radiology > Chapter 61 - Scintigraphic Diagnosis of Inflammation and Infection
Chapter
61
Scintigraphic Diagnosis of Inflammation and Infection Christopher Charito
J.
Palestro
Love
The scintigraphic evaluation of infection and inflammation is extremely broad in scope, encompassing numerous radiopharmaceuticals, imaging techniques, and diseases. In addition to reviewing the roles of gallium-67 (Ga-67) citrate and radiolabeled leukocytes (WBC), this chapter also addresses the potentials of the recently approved antigranulocyte antibody, technetium-99m (Tc99m) -fanolesomab, and of fluorine-18-fluorodeoxyglucose PET (F18-FDG-PET) for imaging inflammation and infection.
GALLIUM-67 Ga-67, which has been used for localizing infection for more than three decades, is produced with a cyclotron. Decay is by electron capture and the half-life is 78.1 hours. With principal photon energies of 93, 184, and 296 keV used for imaging and a poor photon yield per disintegration, Ga-67 is a suboptimal imaging agent (1) . Several factors govern uptake of this tracer in inflammation and infection. About 90% of circulating Ga-67 is in the plasma, nearly all transferrin bound. Increased blood flow and increased vascular
3729
membrane permeability result in increased delivery and accumulation of transferrin-bound Ga-67 at inflammatory foci. Ga-67 also binds to lactoferrin, which is present in high concentrations in inflammatory foci. Direct bacterial uptake may also account for some Ga-67 accumulation in infection. Siderophores, low-molecular-weight chelates produced by bacteria, have a high affinity for Ga-67. The siderophore–Ga-67 complex is presumably transported into the bacterium, where it eventually is phagocytosed by macrophages. Although some Ga-67 may be transported bound to leukocytes, it is important to note that, even in the absence of circulating leukocytes, Ga-67 accumulates in infection (1) . Imaging is usually performed 18 to 72 hours after injection of 185 to 370 MBq (5 to 10 mCi) of Ga-67 citrate. A gamma camera equipped with a medium-energy collimator and capable of imaging multiple energy peaks is used. The normal biodistribution of Ga-67 is variable and includes bone, bone marrow, liver, genitourinary and GI tracts, and soft tissues (2) (Fig. 61.1). Nasopharyngeal and lacrimal gland activity can be very prominent, even in the absence of disease. Intense breast uptake is associated with hyperprolactinemic states, including pregnancy, lactation, certain drugs, and hypothalamic lesions. In patients who have undergone multiple transfusions, increased renal, bladder, and bone/marrow activity, together with decreased hepatic and colonic activity, is often observed, presumably because of iron receptor saturation by exogenous iron from the transfused cells. The MR contrast agent gadolinium can cause similar alterations in the biodistribution of Ga-67 (3, 4, 5, 6, 7, 8) . Although labeled leukocyte imaging is generally considered the radionuclide test of choice for imaging infection in the immunocompetent population, Ga-67 imaging remains both popular and useful, providing information that is complementary to, and at times not available from, other tests. Indications for gallium imaging include the following.
Opportunistic
Infection
Nuclear medicine plays an important role in the detection of
3730
infections unique to the immunocompromised patient, and for most of these, P.1441 Ga-67 imaging is the radionuclide procedure of choice. Many opportunistic infections affect the lungs, and a normal Ga-67 scan of the chest excludes infection with a high degree of certainty, especially in the setting of a negative chest radiograph. In the HIVpositive patient, lymph node uptake of Ga-67 is most often caused by mycobacterial disease or lymphoma. Focal, or localized, pulmonary parenchymal Ga-67 uptake is usually associated with bacterial pneumonia. Diffuse pulmonary gallium uptake is indicative of Pneumocystis jiroveci pneumonia, especially when the uptake is intense (Fig. 61.2). In addition to its value for diagnosis, Ga-67 can be used for monitoring response to therapy. Kaposi sarcoma, a malignancy often found in patients with AIDS, is not Ga-67 avid (1,9,1 0) .
FIGURE
61.1. Normal Pediatric and Adult Gallium Studies.
A . Anterior and posterior whole-body gallium-67 (Ga-67) images performed on an 11-year-old child. Prominent skeletal uptake is normal in children. The distal femoral and proximal tibial growth plates are easily identified. B . Anterior and posterior whole-body
3731
Ga-67 images performed on a 20-year-old woman. In this patient there is more soft tissue and less skeletal activity than in the patient illustrated in A . Note the physiologic breast activity, which can be confused with abnormal pulmonary uptake. This can be resolved by obtaining oblique and lateral views, or by SPECT.
Interstitial
Lung
Disease
Ga-67 is an extremely sensitive indicator of pulmonary inflammation. Uptake of this tracer occurs in sarcoidosis, interstitial pneumonitis in virtually all its forms and etiologies, drug reactions, collagen vascular disease, and pneumoconioses. Although the degree of activity generally correlates with the severity of the underlying illness, a normal study does not exclude very mild inflammation. Furthermore, neither the intensity nor the pattern of uptake is diagnostic of specific illnesses (1 1,1 2) . In sarcoidosis, pulmonary uptake of Ga-67 correlates with disease activity and response to therapy. Ga-67 scintigraphy has been reported to be up to 97% sensitive P.1442 for detection of active sarcoidosis when considering both pulmonary and extrapulmonary sites (Fig. 61.3). Whether Ga-67 imaging can provide prognostic information or therapeutic insight for other inflammatory lung diseases is not known.
3732
FIGURE 61.2. Pneumocystis jiroveci Pneumonia. Gallium-67 image of the chest demonstrates intense diffuse bilateral pulmonary activity that, in an AIDS patient with a normal chest radiograph,
strongly
suggests P.
jiroveci pneumonia.
3733
FIGURE
61.3. Sarcoidosis. Anterior and posterior whole-body
gallium-67 images demonstrate bilateral hilar activity, a pattern that is characteristic for sarcoidosis. Prominent parotid and submandibular gland activity is often seen in sarcoidosis as well. Moderately intense activity in the descending colon is normal.
3734
FIGURE
61.4. Diffuse Pulmonary Uptake of Gallium-67 (Ga-
67). Pulmonary uptake of Ga-67 is often compared to hepatic activity and here is graded on a numeric scale from 0 (normal) to 4 (abnormal, with pulmonary activity more intense than hepatic activity). Details are provided in the text.
Because the determination of relative pulmonary Ga-67 activity may be helpful for assessing the level of inflammatory activity present, an objective index of Ga-67 activity has been sought. Some authors have chosen to compare pulmonary activity with sternal activity, others have compared pulmonary activity with hepatic activity, and still others have used semiquantitative techniques involving computer acquisitions, SPECT, and whole-body imaging to report activity ratios. As a result, there is no standard method for quantifying Ga-67 pulmonary activity. Therefore, when reporting Ga-67 activity, the specific scale and reference standard should be stated. We use a
3735
scale of 0 to 4, in which pulmonary activity is compared to liver activity. In this schema, 0 (normal) represents pulmonary activity that is indistinguishable from background, and 1 represents equivocally increased activity. Grades 2, 3, and 4 represent definitely abnormal pulmonary activity; grade 2 is less than, grade 3 is equal to, and grade 4 is greater than, hepatic activity. A photopenic or “cold― cardiac silhouette is present in patients with grades 2, 3, and 4 uptake (Fig. 61.4) (1) . Interstitial
nephritis, a well-recognized cause of acute renal
failure, is characterized by interstitial edema and a mononuclear cellular infiltrate. Although definitive diagnosis is by biopsy, Ga-67 can be helpful in differentiating interstitial nephritis from acute tubular necrosis in the acute setting. Interstitial nephritis is characterized by renal uptake of Ga that is more intense than spinal uptake, while acute tubular necrosis is characterized by little or no renal uptake. The test is less reliable in patients with chronic renal failure (Fig. 61.5) (1 3) .
3736
FIGURE
61.5. Interstitial
Nephritis. Anterior and posterior
whole-body gallium-67 images. Renal activity is more intense than adjacent lumbar spine activity, which is a pattern typical of interstitial nephritis.
P.1443 Fever of undetermined origin (FUO) is an illness of at least 3 weeks’ duration with several episodes of fever exceeding 38.3°C and no diagnosis after an appropriate inpatient or outpatient evaluation. There are numerous causes of FUO, and infection accounts for only about 20% to 30% of them. Neoplasms are responsible for about 15% to 25%. Other etiologies include collagen
3737
vascular disease, vasculitis, granulomatous diseases, pulmonary emboli, cerebrovascular accidents, and drug fever. Radionuclide imaging is typically reserved for those situations in which other imaging tests fail to localize the source of the fever. Nearly 80% of FUOs are caused by an entity other than infection, and therefore Ga67, which accumulates in infection, inflammation, and tumor, is often preferred over WBC imaging for this indication (Fig. 61.6) 1,2).
Spinal
Osteomyelitis
Although it has been replaced by labeled leukocyte imaging for evaluation of osteomyelitis in most regions of the skeleton, Ga-67 remains the radionuclide procedure of choice for diagnosing spinal osteomyelitis. Ga-67 imaging is frequently performed in conjunction with bone scintigraphy, and over the years, criteria have been established for the interpretation of bone/ Ga-67 imaging. These criteria are used regardless of the area of the skeleton being evaluated
(1 4). The combined test is:
Positive for osteomyelitis when distribution of the two tracers is spatially incongruent, or when the distribution is spatially congruent and the relative intensity of uptake of Ga-67 is greater than that of the bone agent (Fig. 61.7) Equivocal for osteomyelitis when the distribution of the two radiotracers is congruent, both spatially and in terms of intensity (Fig. 61.8) Negative for osteomyelitis when the Ga-67 images are normal, regardless of the bone scan findings, or when the distribution of the two tracers is spatially congruent and the relative intensity of uptake of Ga-67 is less than that of the bone agent (Fig. 61.9) .
RADIOLABELED
LEUKOCYTES
Although a variety of in vitro leukocyte labeling techniques have been investigated, the only approved methods in the United States employ the lipophilic compounds indium-111 (In-111) oxyquinoline
3738
and 99 mTc-HMPAO (exametazine hexamethyl propyleneamine oxime). The labeling procedure takes about 2 to 3 hours. Approximately 40 to 60 mL of whole blood is withdrawn from the patient into an anticoagulant-containing syringe. All of the cellular components of the blood can be labeled, and the leukocytes must be separated from erythrocytes and platelets. After withdrawal, therefore, the syringe containing the blood is kept in an upright position for about 1 to 2 hours to promote erythrocyte sedimentation, a process facilitated by the addition of hydroxyethyl starch. This process can also be accelerated by substituting hypotonic lysis of the red cells for gravity sedimentation. After they are separated from the erythrocytes, the leukocytes are separated from platelets via centrifugation, and the leukocyte “pellet― that forms at the bottom of the tube is incubated with the radiolabel, washed, and reinjected into the patient (2) .
Indium-111–Labeled
Leukocytes
The imaging characteristics of In-111 are superior to those of Ga-67, with photopeaks of 173 keV and 247 keV. It decays by electron capture with a half-life of 67 hours. These energies require the use of a medium-energy collimator and a gamma camera capable of imaging multiple energy peaks. Energy discrimination is accomplished by using a 15% window centered on the 174-keV photopeak and a 20% window centered on the 247-keV photopeak of In-111. High target (abscess)–to-background ratios provide excellent image contrast. The spleen is the critical organ, receiving up to 20 rads per mCi of injected cells. As a result, the adult dose of In-111–labeled leukocytes is limited to about 18.5 MBq (500 µCi). Images obtained shortly after injection are characterized by intense pulmonary activity. This activity, which clears rapidly, is probably caused by P.1444 P.1445 leukocyte activation during labeling, which impedes their movement through the pulmonary vascular bed, prolonging their passage through the lungs. At 24 hours after injection, the usual imaging time
3739
for In-WBCs, the normal distribution of activity is limited to the liver, spleen, and bone marrow (2) .
FIGURE 61.6. Fever of Undetermined Origin in a Patient With Metastatic Renal Cell Carcinoma. Anterior whole-body (A) and coronal SPECT (B) images from a gallium-67 study, performed on an 81-year-old carcinoma, persistent fevers, demonstrate focally increased supraclavicular region, brain,
woman with a history of renal cell and no localizing signs, activity in the mediastinum, left and distal right femur. Mediastinal
3740
lymph node biopsy confirmed involvement by metastatic renal cell carcinoma. Brain and femoral metastases were radiographically confirmed.
FIGURE
61.7. Positive Bone/Gallium-67 (Ga-67) Study. A.
The distribution of activity in the bone (left) and Ga-67 (right) images of the pelvis and upper thighs is spatially incongruent. Irregularly increased activity in the proximal left femur is present on the bone scan, while the abnormal Ga-67 activity occupies a much smaller area. B . The distribution of activity on the bone (left) and Ga-67 (right) images of the lumbar spine is spatially congruent; increased activity involving the right pedicular region
3741
of L5 is present on both studies. The intensity of this uptake is greater on the Ga-67 image than on the bone image.
FIGURE
61.8. Equivocal
Bone/Gallium-67
(Ga-67)
Study.
Posterior bone (left) and Ga-67 (right) images from a study performed on a patient with a failed left hip arthroplasty. The spatial distribution and intensity of uptake of both tracers are virtually identical, and hence the study is equivocal for infection.
Advantages of the In-111 label are its stability and a virtually constant normal distribution of activity limited to the liver, spleen, and bone marrow. The 67-hour physical half-life of In-111 permits delayed imaging, which is particularly valuable for musculoskeletal infection. There is another advantage to the use of In-WBCs in musculoskeletal infection. Patients undergoing this test often require bone or marrow scintigraphy, which can be performed while the patient's cells are being labeled, or as simultaneous dual-isotope acquisitions, or immediately after completion of the In-WBC study. If Tc-WBCs are used, an interval of at least 48 hours is required between the white cell and bone or marrow scans (2) .
3742
FIGURE
61.9. Negative
Bone/Gallium-67
(Ga-67)
Studies.
A . Increased activity involving contiguous lower thoracic/upper lumbar vertebrae is present on the bone scan (left). The Ga-67 image is normal. B . The distribution of activity on the bone (left) and Ga-67 (right) images of the pelvis and femurs is spatially congruent; there is increased activity in the intertrochanteric region on both studies. The uptake, however, is less intense on the Ga-67 image than on the bone image.
Disadvantages of the In label include a low photon flux, less-thanideal photon energies, and the fact that a 24-hour interval between injection and imaging is generally required (2) .
Technetium-99m–Labeled
3743
Leukocytes
For Tc-WBC studies, a high-resolution, low-energy parallel hole collimator is used with a 15% to 20% window centered on the 140keV photopeak of Tc-99m. The usual adult dose of Tc-WBCs is 185 to 370 MBq (5 to 10 mCi). The normal biodistribution of Tc-WBCs is more variable than that of In-WBCs. In addition to the reticuloendothelial system and pulmonary activity seen soon after injection, activity is also normally present in the genitourinary tract, large bowel (within four hours after injection), blood pool, and occasionally, the gallbladder (Fig. 61.10). The time interval between injection of Tc-WBCs and imaging varies with the indication; in general, imaging is usually performed within a few hours after injection
(2) .
Advantages of Tc-99m-WBCs include a photon energy that is optimal for imaging using current instrumentation, a high photon flux, and the ability to detect abnormalities within a few hours after injection. Disadvantages include genitourinary tract activity, which appears shortly after injection, and colonic activity, which appears by 4 hours after injection. The instability of the label and the 6-hour half-life of Tc-99m are disadvantages when delayed 24-hour imaging is needed. This occurs in those infections that tend to be indolent in nature and for which several hours may be necessary for accumulation of a sufficient quantity of labeled leukocytes to be successfully imaged (2) .
3744
FIGURE
61.10. Normal White Blood Cell (WBC) Studies in
the same adolescent male patient, approximately 2 weeks apart. Anterior images on the left and posterior images on the right. Compare the biodistribution of technetium (Tc)-WBCs at 90 minutes postinjection with that of indium (In)-WBCs at 18 hours postinjection. Note the cardiac, femoral vessel, renal, and bladder activity on the Tc-WBC images, which is not seen on the In-WBC images. Faint early intestinal activity is superimposed on the sacrum in the anterior Tc-WBC image. Physeal plate marrow activity on both studies is normal for the patient's age. T, TcWBC; I, In-WBC.
3745
P.1446
General
Observations
Regardless of the radiolabel used, uptake of labeled WBCs depends on intact chemotaxis, the number and types of cells labeled, and the cellular component of a particular inflammatory response. Labeling of WBCs, now a routine procedure, does not affect their chemotactic response. A total white count of at least 2,000/µL is needed to obtain satisfactory images. Because the majority of leukocytes labeled are neutrophils, the procedure is most useful for identifying neutrophil-mediated inflammatory processes, such as bacterial infections. The procedure is less useful for those illnesses in which the predominant cellular response is not neutrophilic e.g., opportunistic (2,9,1 5) .
infections,
tuberculosis,
and
sarcoidosis
(Fig. 61.11)
Although pulmonary uptake of WBCs during the first few hours after injection is a normal physiologic event, by 24 hours such uptake is abnormal. Focal pulmonary uptake that is segmental or lobar in appearance is usually associated with bacterial pneumonia (Fig. 61.12). This pattern is also seen in patients with cystic fibrosis and is caused by WBC accumulation in pooled secretions in bronchiectatic regions of the lungs. Nonsegmental focal pulmonary uptake is caused by technical problems during labeling or reinfusion and is not usually associated with infection (1 6) .
3746
FIGURE 61.11. Sarcoid. Anterior indium–white blood cell (InWBC) (left) and gallium-67 (Ga-67) (right) whole-body images of a patient with sarcoid (same patient as illustrated in Fig. 61.3) . Compare the normal In-WBC image to the obviously abnormal Ga-67 image. Radiolabeled WBC studies are not useful for detecting inflammations and infections in which neutrophils are not the predominant cellular response.
3747
FIGURE
61.12. Focal White Blood Cell (WBC) Pulmonary
Activity. Focal pulmonary activity that is segmental or lobar in appearance, as shown in this indium-WBC image, is usually associated with bacterial pneumonia.
P.1447 Diffuse pulmonary uptake on images obtained more than 4 hours after reinjection of labeled cells is associated with opportunistic infection, radiation pneumonitis, pulmonary drug toxicity, and ARDS (Fig. 61.13). This pattern is almost never seen, however, in bacterial pneumonia (1 6) . Diffuse pulmonary uptake of WBCs is also seen in septic patients with
3748
normal chest radiographs and who have no clinical evidence of respiratory tract inflammation or infection. It is believed that the circulating neutrophils, activated by cytokines, pool in the pulmonary circulation because it is more difficult for them to undergo the cytoskeletal deformation required to maneuver through the pulmonary circulation. The cytokines presumably also activate pulmonary vascular endothelial cells, causing increased adherence of leukocytes to the cell walls (1 6) . In-WBCs do not accumulate in normal bowel. Such activity is always abnormal and is seen in antibiotic-associated or pseudomembranous colitis, infectious colitis, inflammatory bowel disease, ischemic colitis, and GI bleeding (Fig. 61.14) (2,9) . Radiolabeled WBCs do not accumulate in normally healing surgical wounds, so the presence of such activity indicates infection—although there are certain exceptions. Granulating wounds, which heal by secondary intention, can appear as areas of intense activity on WBC images even in the absence of infection. Examples
include
“ostomies―
(tracheostomies,
ileostomies,
feeding gastrostomies, etc.) and skin grafts (Fig. 61.15). Vascular access lines, dialysis catheters, and even lumbar punctures can all produce false-positive results in the absence of appropriate clinical history (1 7). Indications for In-labeled leukocyte imaging include the following.
3749
FIGURE
61.13. Diffuse White Blood Cell (WBC) Pulmonary
Activity. There is mild, diffuse bilateral pulmonary activity on this indium-WBC image. While this is a normal finding on images performed shortly after injection, this is an abnormal finding on later images and is associated with many entities, but not with bacterial pneumonia.
Fever
of
Undetermined
Origin
As mentioned previously, because of its diverse etiologies, it can be argued that the nonspecific tracer Ga-67 is the preferred radionuclide test for FUO. However, the data suggest that In-WBC imaging is more sensitive early in the course of an illness, whereas Ga-67 is more sensitive later in the illness, and thus the selection of the procedure might be governed by the duration of the illness. We prefer to begin with an In-WBC study and follow with Ga-67 if needed (Fig. 61.16). Our rationale is as follows. The energies of the photons emitted by, and the physical half-lives of, these two tracers are similar. The amount of activity injected for Ga-67 is typically 10 or more times the amount of activity injected for an In-WBC study.
3750
Should the In-WBC study fail to provide a diagnosis, the patient can be injected with Ga-67 and scanned 48 to 72 hours later. If In-WBC is performed after Ga-67, however, it is necessary to wait a minimum of 1 week to obtain diagnostically useful images (2) .
Postoperative
Infection
Radionuclide tests are an adjunct to anatomic imaging modalities and facilitate the P.1448 differentiation of abscess from other fluid collections, from tumor, and even from normal postoperative changes. Ga-67 can detect intra-abdominal infection, but the presence of large bowel activity can obscure foci of infection, and the need to often wait 48 hours or more between injection and imaging is another disadvantage. Furthermore, Ga-67 accumulates in both infection and tumor, as well as in normally healing surgical incisions. Labeled WBCs, in contrast, rarely accumulate in uninfected neoplasms, and do not, with the exceptions already mentioned, accumulate in normally healing surgical incisions. For these reasons, WBC imaging is the preferred radionuclide study for the evaluation of postoperative infection (Fig. 61.17) (1 7) .
3751
FIGURE 61.14. Colitis. Anterior whole body indium–white blood cell image demonstrates intense pancolonic activity. The differential diagnosis includes antibiotic-associated (pseudomembranous) colitis, infectious colitis, ischemic colitis, and inflammatory bowel disease. No conclusions about the extent of bowel involvement can be drawn from a single 24-hour image, because activity in the bowel lumen is redistributed over time by normal peristalsis.
3752
FIGURE 61.15. White Blood Cell (WBC) Activity at a Tracheostomy Site. Focally increased activity around a tracheostomy site can be seen on this anterior whole-body indium-WBC image. “Ostomies― are granulating wounds and as such recruit granulocytes. Thus WBC activity around an ostomy is normal finding. As this case illustrates, normal “ostomy― uptake can be intense.
Cardiovascular
and
CNS
3753
Infections
Echocardiography is a readily available and accurate method for diagnosing bacterial endocarditis, and radionuclide methods play a very limited role in the diagnostic workup of this entity. Echocardiography is less sensitive, however, for detecting one of the complications of bacterial endocarditis: the myocardial abscess. Both Ga-67 and WBC imaging detect myocardial abscesses in patients with infective endocarditis (2). WBC imaging is the radionuclide procedure of choice for diagnosing prosthetic vascular graft infection, with a sensitivity of more than 90% (Fig. 61.18). Neither duration of symptoms nor pretreatment with antibiotics adversely affects the study. The specificity of WBC imaging is more variable, however, ranging from 53% to 100%. Causes of false-positive results include perigraft hematoma, bleeding, graft thrombosis, pseudoaneurysms, and graft endothelialization, which occurs within the first 1 to 2 weeks after placement (2,9) .
3754
FIGURE
61.16. Pelvic Abscess in a Patient With Fever of
Unexplained Origin. Anterior and posterior whole-body indiumWBC images demonstrate a focus of intense activity in the left lower quadrant of the abdomen (arrow). A subsequent CT scan (not shown) confirmed a pelvic abscess. Faint ascending and transverse colonic activity was attributed to antibiotic-associated colitis.
P.1449 The differential diagnosis of a contrast-enhancing brain lesion identified on CT or MR includes abscess, tumor, cerebrovascular accident, and even multiple sclerosis. WBC scintigraphy provides valuable information about contrast-enhancing brain lesions. A positive study indicates that the origin of the brain lesion is almost assuredly infectious; a negative result rules out infection with a high degree of certainty. Faint uptake in brain tumors has been observed, and false-negative results in patients receiving high-dose steroids have been reported (1 8,1 9) .
Osteomyelitis Three-phase bone scintigraphy is the radionuclide procedure of choice for diagnosing osteomyelitis in bones not affected by underlying conditions. Focal hyperperfusion, focal hyperemia, and focally increased bony uptake on delayed (2 to 4 hours postinjection) images are the classic presentation of osteomyelitis (Fig. 61.19) . Bone scan abnormalities reflect the rate of new bone formation in general; consequently, fractures, orthopedic hardware, and the neuropathic joint can all produce a positive three-phase bone scan, even in the absence of infection. In these situations, often described as “complicating osteomyelitis,― the bone scan, because of decreased specificity, is less reliable (2 0,2 1) .
3755
FIGURE 61.17. Postoperative Infection. A patient with a history of multiple abdominal surgeries was noted to have a mass on CT scan of the abdomen and pelvis (not shown). The differential diagnosis included postoperative changes and tumor. On the anterior whole-body indium–white blood cell (WBC) image, abnormal accumulation of labeled leukocytes extends from the left abdomen into the thigh. Multiple abscesses were subsequently drained. WBC imaging is a useful adjunct to CT in the evaluation of postoperative infection.
3756
Except in the spine, In-WBC scintigraphy is the procedure of choice for diagnosing complicating osteomyelitis. To maximize accuracy, the test is frequently performed in conjunction with Tc-99m sulfur colloid marrow imaging. Although labeled leukocytes do not usually accumulate at sites of increased bone mineral turnover in the absence of infection, they do accumulate in the bone marrow. The normal distribution of hematopoietically active bone P.1450 marrow in adults is limited to the axial and proximal appendicular skeleton, and WBC activity outside this normal distribution is indicative of infection. Unfortunately, the “normal― distribution of hematopoietically active bone marrow is variable. Systemic conditions such as sickle cell and Gaucher diseases produce generalized alterations in marrow distribution, whereas fractures, orthopedic hardware, and the neuropathic joint cause localized alterations. The normal distribution of hematopoietically active marrow in children varies with age. Consequently, it may not be possible to determine if an area of activity on a WBC image represents infection or marrow. Performing complementary bone marrow imaging with Tc-99m sulfur colloid overcomes this problem. Both labeled leukocytes and sulfur colloid accumulate in the bone marrow. Leukocytes also accumulate in infection; sulfur colloid, however, does not. The combined study is positive for infection when activity is present on the WBC image without corresponding activity on the sulfur colloid marrow image. Any other pattern is negative for infection (Figs. 61.20, 61.21). The overall accuracy of combined WBC/marrow imaging is approximately 90% (2 1) .
3757
FIGURE 61.18. Vascular Graft Infection. Indium–white blood cell study demonstrates linearly increased activity along the medial aspect of the right thigh, from the groin to the knee, in a patient with an infected femoral popliteal prosthetic graft.
In contrast to other areas in the skeleton, WBC imaging with or
3758
without marrow imaging is not useful for detecting spinal osteomyelitis. Although increased uptake is virtually diagnostic of this entity, 50% or more of all cases of vertebral osteomyelitis present as areas of decreased or absent activity on WBC images. This photopenia is not specific for vertebral osteomyelitis and is associated with other entities such as tumor, infarction, and Paget disease (Fig. 61.22) (2 1,2 2) .
Inflammatory
Bowel
Disease
Although early studies were performed with In-WBCs, it is now agreed that Tc-WBC is the radionuclide study of choice for inflammatory bowel disease, a group of idiopathic chronic disorders that includes Crohn disease and ulcerative colitis. WBC imaging is very sensitive for detecting inflammatory bowel disease and can be used as a screening test to determine which patients need to undergo more invasive investigation. In patients thought to have ulcerative or indeterminate colitis, skip areas of activity in the colon or the presence of small bowel activity support the diagnosis of Crohn disease (Fig. 61.23). The radionuclide study is also useful in patients who refuse endoscopy or contrast radiography, as well as in those in whom these studies cannot be satisfactorily performed because of narrowing of the bowel lumen. The ability of the radionuclide study to differentiate active inflammation, which may respond to medical therapy, from scarring, which may require surgery, can have a significant impact on patient management. WBC imaging can also be used to monitor patient response to therapy. Decreasing bowel uptake on serial studies confirms that the patient is responding to treatment, whereas persistent or recurrent uptake indicates residual disease or relapse (2 3, 2 4, 2 5) . Imaging at multiple time points and SPECT increase the sensitivity of the test. The caudal, or pelvic outlet, view facilitates detection of rectal disease that might otherwise be masked by urinary bladder activity. Physiologic bowel activity, probably caused by hepatobiliary excretion of Tc-99m–labeled hydrophilic complexes, frequently appears on delayed images and must be differentiated from activity secondary to inflammation. Physiologic activity appears in the distal
3759
small bowel no sooner than 3 hours after injection, is diffuse and mild in intensity, and migrates into the cecum by 4 hours. There must be no accumulation in other bowel segments (2 6) . There are limitations to WBC imaging. It cannot be the only imaging test used for inflammatory bowel disease. It cannot define anatomic changes such as strictures, which are best delineated with endoscopy and contrast radiography. The test is less sensitive for disease of the upper GI tract versus the lower GI tract. The sensitivity of the test also may be affected adversely by concomitant administration of corticosteroids
(2 4,2 7) .
FIGURE 61.19. Acute Osteomyelitis. Three-phase bone scan in a 13-year-old girl. Radiographs were normal. A . Anterior view of flow study of the feet and ankles with technetium-99m methylene diphosphonate, 2 seconds per frame. Note the early and
3760
increased flow to the right foot and ankle. B . Blood pool image, with a marker indicating the right side. Note the markedly increased activity in the distal tibia (arrow) extending to the physeal
plate.
P.1451
TECHNETIUM-99M-FANOLESOMAB Tc-99m-fanolesomab,
a
monoclonal
murine
M-class
immunoglobulin,
binds to CD15 receptors present on leukocytes. This agent presumably binds to circulating neutrophils that eventually migrate to the focus of infection, as well as to neutrophils or neutrophil debris containing CD15 receptors, already sequestered in the area of infection. About 370 to 740 MBq (10 to 20 mCi) of the radiolabeled compound, containing 75 to 125 µg of antibody is injected. In contrast to in vitro labeled leukocytes, there is no increased retention of activity in the lungs. The normal distribution includes liver, spleen, bone marrow, and genitourinary tract. Blood pool activity, present on images obtained shortly after injection, decreases over time. Large bowel activity normally appears as early as 4 hours P.1452 after injection and is usually present on images obtained at 24 hours. Small bowel activity is also occasionally seen (Fig. 61.24). The doselimiting organ is the spleen, which receives an estimated 0.064 mGy/MBq (0.24 rads/mCi), considerably lower than the estimated 5 mGy/MBq (18 rads/mCi) for In-WBCs. Within 20 minutes after injection, there is a transient decrease in the number of circulating WBCs. There have been no clinical complaints associated with this phenomenon, and recovery usually occurs within 45 minutes. Based on available data, the agent is safe, with little toxicity (2 8,2 9) .
3761
FIGURE 61.20. Infected Orthopedic Hardware. There is slightly increased left femoral activity on the indium–white blood cell (In-WBC) image (left) performed on a patient with an intramedullary rod in the left femur. There is a photopenic area (arrow) on the marrow image (right), and the study is positive for infection. Notice also, however, that most of the left femur activity on the In-WBC image is caused by marrow, not infection.
3762
FIGURE
61.21. Aseptic Loosening of Right Total Hip
Replacement. Indium–white blood cell (In-WBC) image (left) in a patient with bilateral total hip arthroplasties shows increased activity around the femoral component of a right total hip replacement. Infection cannot be excluded. The distribution of activity on the marrow image (right) is identical to that on the In-WBC image. Therefore the combined study is negative for infection.
At the present time, Tc-99m-fanolesomab is approved for diagnosis of equivocal appendicitis in patients older P.1453 than 5 years. A large multicenter trial involving 200 patients between 5 and 86 years old was conducted to assess the efficacy of fanolesomab for diagnosing acute appendicitis in patients with an equivocal presentation to evaluate its safety and to assess its potential impact on the clinical management of these patients. Fiftynine patients had histopathologically confirmed acute appendicitis. The diagnosis of appendicitis was made in all cases within 90 minutes after injection. Images became positive within 8 minutes in 50% of
3763
patients with acute appendicitis and within 50 minutes in 90% (Fig. 61.25). Sensitivity, specificity, and accuracy were 91%, 86%, and 87%, respectively. Positive and negative predictive values of the test were 73% and 96%, respectively. The high negative predictive value P.1454 is especially important, because a negative result means that acute appendicitis is very unlikely, thereby reducing unnecessary time in the hospital for observation as well as unnecessary surgery. Finally, there was a significant improvement in making the appropriate management decision, in both patients with and without appendicitis, after the scan (3 0) .
FIGURE 61.22. Spinal Osteomyelitis. Posterior indium–white blood cell image of the abdomen demonstrates absent activity in a midlumbar vertebra (arrow). Photopenia is seen in more than 50% of all cases of spinal osteomyelitis. Although decreased activity is consistent with spinal osteomyelitis, it is not specific for it, and is also associated with numerous other entities,
3764
including tumor, infarction, compression fracture, and Paget disease.
FIGURE
61.23. Crohn
Disease. Technetium–white blood cell
activity is present in the distal jejunum/proximal ileum, distal ileum, and colon. Small bowel activity in a patient with colitis supports the diagnosis of Crohn disease. (Courtesy of Dr. Martin Charron.)
3765
FIGURE 61.24. Whole-Body Technetium-99-Fanolesomab Images. Anterior and posterior images were acquired about 2 (left) and 24 (right) hours after injection. In addition to reticuloendothelial activity, genitourinary tract activity appears soon after injection. Colonic activity may be seen as early as 4 hours after injection and is usually present at 24 hours.
3766
FIGURE
61.25. Appendicitis.
A. The “appendicitis zone―
(oval) extends approximately from the pubic symphysis to the lower pole of the right kidney. Any activity in this region is considered indicative of appendicitis (courtesy of Dr. Samuel Kipper). B . Image in another patient, taken about 30 minutes after injection of technetium-99-fanolesomab, shows a linear area of increased activity overlying the right iliac vessels (arrow) indicative of appendicitis. The diagnosis was confirmed at surgery.
Planar imaging is carried out over 90 minutes. Abnormal right lower quadrant activity in the “appendicitis zone― that persists is considered positive for appendicitis (3 0) . Preliminary investigations suggest that Tc-99m-fanolesomab may a satisfactory alternative to In-WBC imaging for diagnosis of osteomyelitis (3 1,3 2) .
FLUORINE-18-FLUORODEOXYGLUCOSE FDG, a glucose analogue, is transported into cells via glucose transporters. Increased uptake of FDG in inflammation is
3767
be
related—at least in part—to an increased number of glucose transporters. In addition, the affinity of glucose transporters for deoxyglucose presumably is increased by various cytokines and growth factors. The normal distribution of FDG includes brain, myocardium, and genitourinary tract. Bone marrow, gastric, and bowel activity are variable. Thymic uptake, especially in children, can be prominent. Liver and spleen uptake are generally low-grade and diffuse, although in infection, splenic uptake may be intense. Imaging is performed about 1 hour after injection of 370 to 740 MBq (10 to 20 mCi) F-18-FDG (3 3) . Although it is not approved for evaluation of inflammation and infection, FDG-PET is an intriguing and exciting alternative to the conventional radionuclide approach to the FUO. FDG is similar to Ga; i.e., although it is not specific, it is exquisitely sensitive, ideally suited to the evaluation of an entity with diverse etiologies. The short half-life of F-18, moreover, does not delay the performance of any additional radionuclide studies that might be contemplated (Fig. 61.26). Recent investigations have found that the test is sensitive and compares favorably to and could potentially replace Ga-67 for the evaluation of patients with FUO (3 3,3 4). The value of FDG-PET is further enhanced by data that suggest that vasculitis and bacterial endocarditis, both of which can be the source of an FUO and which are not amenable to detection with other radionuclide studies, may be identified with this test. Other entities including thromboembolic disease, sarcoidosis, and chronic granulomatous disease, which can all present as an FUO, are also associated with increased FDG up take (3 3) . The negative predictive value of FDG-PET in the patient with an FUO is apparently very high; i.e., a negative test makes it very unlikely that a morphologic origin of the fever will be identified. If confirmed in future investigations, FDG-PET, by reducing the number of imaging studies performed, may prove to be a very cost-effective method of investigating the FUO (3 4) .
3768
FIGURE 61.26. Recurrent Lung Carcinoma. An 81-year-old man with chronic lymphocytic leukemia and a remote history of lung carcinoma presented with fever of unknown origin and an elevated leukocyte count. The indium–white blood cell study (In-WBC) (left) was negative. The fluorodeoxyglucose (FDG) PET study (right) demonstrated focal hypermetabolism in the right paratracheal region that corresponded to lymph nodes identified on CT (not shown). The final diagnosis was recurrent lung carcinoma. This case demonstrates the importance of sensitivity when evaluating patients with fever of undetermined origin (FUO). In-WBC correctly excluded infection as the source of the fever, but provided no information about what the source of the fever was. Although FDG did not provide a diagnosis, it did correctly localize the source of the FUO, facilitating the diagnosis with other studies.
FDG-PET also appears to be a potentially useful test for spinal osteomyelitis. Although most of the series reported P.1455
3769
to date involve small patient samples, this test accurately diagnoses spinal osteomyelitis, with an accuracy comparable to that of Ga-67 (Fig. 61.27) (3 3,3 5) .
FIGURE
61.27. Spinal
Osteomyelitis. Intense
fluorodeoxyglucose accumulation in the lower lumbar spine is seen in this patient with spinal osteomyelitis (left), similar to the abnormality on the coronal gallium SPECT image (right) .
Radionuclide imaging plays a pivotal role in the diagnosis of infection and inflammation and will continue to do so for the foreseeable future. Optimal diagnosis requires careful consideration of the patient, indications for the study, and the imaging modalities at one's disposal.
REFERENCES 1. Palestro CJ. The current role of gallium imaging in infection. Semin
Nucl
Med
1994;24:128–141.
2. Love C, Palestro CJ. Radionuclide imaging of infection. J Nucl Med Technol 2004;32:47–57.
3770
3. Palestro CJ, Malat J, Collica CJ, et al. Incidental diagnosis of pregnancy on bone and gallium scintigraphy. J Nucl Med 1986;27:370–372. 4. Lopez OL, Maisano ER. Ga-67 uptake post cesarean section. Clin Nucl Med 1984;9:103–104. 5. Desai AG, Intenzo C, Park C, et al. Drug-induced gallium uptake in the breasts. Clin Nucl Med 1987;12:703–704. 6. Vasquez R, Oates E, Sarno RC, et al. Gallium-67 breast uptake in a patient with hypothalamic granuloma (sarcoid). J Nucl Med 1998;19:118–121. 7. Engelstad B, Luks S, Hattner RS. Altered 67Ga citrate distribution in patients with multiple red blood cell transfusions. AJR Am J Roentgenol 1982;139:755–759. 8. Hattner RS, White DL. Gallium-67/stable gadolinium antagonism. MRI contrast agent markedly alters the normal biodistribution of Gallium–67. J Nucl Med 1990;31:1844–1846. 9. Palestro CJ, Torres MA. Radionuclide imaging of nonosseous infection. Q J Nucl Med 1999;43:46–60. 10. Palestro CJ, Goldsmith SJ. The use of gallium and labeled leukocyte scintigraphy in the AIDS patient. Q J Nucl Med 1995;39:221–230. 11. Waxman AD. An update on the role of nuclear medicine in pulmonary disorders. In: Freeman LM, Weissman HS, eds. Nuclear Medicine Annual 1985. New York: Raven Press, 1985:199–231.
3771
12. Kramer EL, Divgi CR. Pulmonary applications of nuclear medicine. Clin Chest Med 1991;12:55–75. 13. Linton AL, Richmond JM, Clark WF, et al. Gallium 67 scintigraphy in the diagnosis of acute renal disease. Clin Nephrol 1985;24:84–87. 14. Palestro CJ, Torres MA. Radionuclide imaging in orthopedic infections.
Semin
Nucl
Med
1997;27:334–345.
15. Fineman DS, Palestro CJ, Kim CK, et al. Detection of abnormalities in febrile AIDS patients with In-111-labeled leukocyte and Ga-67 scintigraphy. Radiology 1989;170:677–680. 16. Love C, Opoku-Agyemang P, Tomas MB, et al. Pulmonary activity on labeled leukocyte images: physiologic, pathologic, and imaging correlation. Radiographics 2002;22:1385–1393. 17. Palestro CJ, Love C, Tronco GG, et al. Fever in the postoperative patient: role of radionuclide imaging in its diagnosis. Radiographics 2000;20:1649–1660. 18. Schmidt KG, Rasmussen JW, Frederiksen PB, et al. Indium111-granulocyte scintigraphy in brain abscess diagnosis: limitations and pitfalls. J Nucl Med 1990;31:1121–1127. 19. Palestro CJ, Swyer AJ, Kim CK, et al. Role of In-111 labeledleukocyte scintigraphy in the diagnosis of intracerebral lesions. Clin Nucl Med 1991;16:305–308. 20. Schauwecker DS. The scintigraphic diagnosis of osteomyelitis. AJR Am J Roentgenol 1992;158:9–18.
3772
21. Palestro CJ, Torres MA. Radionuclide imaging in orthopedic infections. Semin Nucl Med 1997;27:334–345. 22. Palestro CJ, Kim CK, Swyer AJ, et al. Radionuclide diagnosis of vertebral osteomyelitis: indium-111-leukocyte and technetium99m-methylene diphosphonate bone scintigraphy. J Nucl Med 1991;32:1861–1865. 23. Martin-Comin J, Prats E. Clinical applications of radiolabeled blood elements in inflammatory bowel disease. J Nucl Med 1999;43:74–82. 24. Del Rosario MA, Fitzgerald JF, Siddiqui AR, Chong SK, Croffie JM, Gupta SK. Clinical applications of technetium Tc 99m hexamethyl propylene amine oxime leukocyte scan in children with inflammatory bowel disease. J Pediatr Gastroenterol Nutr 1999;28:63–70. 25. Charron M. Pediatric inflammatory bowel disease imaged with Tc-99m white blood cells. Clin Nucl Med 2000;25:708–715. 26. Charron M, Del Rosario JF, Kocoshis S. Comparison of the sensitivity of early versus delayed imaging with Tc-99m HMPAO WBC in children with inflammatory bowel disease. Clin Nucl Med 1998;23:649–653. 27. Granquist L, Chapman SC, Hvidsten S, et al. Evaluation of 99 m Tc-HMPAO leukocyte scintigraphy in the investigation of pediatric inflammatory bowel disease. J Pediatr 2003;143:48–53. 28. Thakur ML, Marcus CS, Henneman P, et al. Imaging inflammatory diseases with neutrophil-specific technetium-99mlabeled monoclonal antibody anti-SSEA-1. J Nucl Med 1996;37:1789–1795.
3773
29. Love C, Palestro CJ. 99mTc-fanolesomab. IDrugs 2003;6:1079–1085. 30. Rypins EB, Kipper SL, Weiland F, et al. 99 mTc
anti-CD15
monoclonal antibody (LeuTech) imaging improves diagnostic accuracy and clinical management in patients with equivocal presentation of appendicitis. Ann Surg 2002;235:232–239. 31. Palestro CJ, Kipper SL, Weiland FL, et al. Osteomyelitis: diagnosis with 99 mTc-labeled antigranulocyte antibodies compared with diagnosis with indium-111-labeled leukocytes—initial experience. Radiology 2002;223:758–764. 32. Palestro CJ, Caprioli R, Love C, et al. Rapid diagnosis of pedal osteomyelitis in diabetics with a technetium-99m-labeled monoclonal antigranulocyte antibody. J Foot Ankle Surg 2003;42:2–8. 33. Love C, Tomas MB, Tronco GG, Palestro CJ. FDG-PET of infection and inflammation. Radiographics 2005 Sep–Oct;25(5):1357–1368. 34. Bleeker-Rovers CP, de Kleijn EMHA, Corstens FHM, et al. Clinical value of FDG PET in patients with fever of unknown origin and patients suspected of focal infection or inflammation. Eur J Nucl Med Mol Imaging 2004;31:29–37. 35. Stumpe KDM, Zanetti M, Weishaupt D, et al. FDG positron emission tomography for differentiation of degenerative and infectious endplate abnormalities in the lumbar spine detected on MR imaging. AJR Am J Roentgenol 2002;179:1151–1157.
3774
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XII - Nuclear Radiology > Chapter 62 - Molecular Imaging of Tumors
Chapter
62
Molecular Imaging of Tumors James H. Timmons
WHEN
PHYSIOLOGY
MATTERS
Radiologists tend to be intellectually oriented toward pathologic anatomy. This partially results from the historical development of the field. Anatomic imaging methods have been dominant in the field of radiology. Nuclear medicine developed originally from a confluence of the interests of internal medicine, pathology, and radiology physicians in investigating the physiology of disease. Imaging moved only gradually to the forefront of nuclear medicine, with evolution from radiotracers and radioimmunoassay to single- or dual-probe studies, then Anger camera imaging, and then SPECT. From the standpoint of a radiologist, nuclear medicine is often viewed as unappealing because the images generally have poorer resolution and less anatomic detail than traditional radiographs. This disadvantage is even more apparent when comparisons are made to CT, US, or MR. It is important for the modern radiology resident to overcome this anatomic bias. The recent rapid evolution of PET from a research tool to the dominant imaging modality in oncology stems from its ability to distinguish physiologically active masses from mere masses. Although advances continue in anatomic imaging (such as MR white matter tractography), technical advances in imaging physiology are likely to dwarf those in imaging of anatomy in the
3775
next decade. This will likely be true even without nuclear medicine, as evident from development of functional contrast agents for MR and of functional MR imaging of brain activity. At the same time, the development of hybrid devices such as PET-CT and SPECT-CT imagers has largely overcome the “unclear medicine― reputation of the nuclear image. Coregistered images obtained at a single setting allow simultaneous investigation of anatomy and physiology. There has been remarkable growth in the availability of scintigraphic methods for tumor imaging over the past few years. Many of these agents provide information that was not previously available. Dramatic improvements in both sensitivity and specificity for detection and staging of tumors have been made. These agents are unique because they provide physiologic information concerning the imaged tumor, such as metabolic activity (and by extension viability), presence of specific cell surface receptors, and presence of multidrug resistance apparatus. The development of these agents has depended heavily on the use of tracer techniques that do not significantly alter the physiologic activity of the substance being employed for P.1457 imaging. Development has also been made possible by extensive basic science investigations into polypeptide chemistry and monoclonal antibodies and by serendipitous clinical observations. It is important to note that the recent rapid evolution of clinical PET resulted entirely from a change in the regulatory climate. There was no fundamental technologic leap that opened the floodgates. Rather, the already mature imaging technology began to be applied clinically when Food and Drug Administration (FDA) rules were changed to allow regional fluorine-18 fluorodeoxyglucose (F-18-FDG) production in a cost-effective manner and when the Center for Medicare Studies recognized F-18-FDG scanning as both clinically effective and cost competitive. Hundreds of PET radiotracers and SPECT radiotracers for basic physiologic processes have been developed and clinically proven over the past two decades. What stands between these agents and rapid clinical diffusion is political recognition of the very
3776
low toxicologic risk represented by the use of tracer imaging and of the cost-effectiveness of tracer use. At the doses employed, only substances such as botulinum toxin and cyanide retain any toxicity. The gram quantities of material actually employed in scanning are lower than the accepted levels of known toxins in drinking water (Table 62.1 ). Monoclonal antibodies are an exception to the rule, because of larger material quantities and because of the inclusion of foreign proteins. Only political decisions prevent widespread adoption of multiple molecular targeted agents. The clear choice is for radiology residents to become comfortable with physiologic imaging or become competitively disadvantaged. Fortunately, the average radiology resident is well prepared for imaging physiology. Preparedness requires mainly a shift in emphasis. PET has become important enough for tumor imaging to require a separate chapter. The current chapter is devoted to single-photon agents for imaging tumors. Whereas these agents are relatively less important than PET in oncologic diagnosis, the therapeutic applications of these agents are now beginning to become significant. There are literally hundreds of agents currently under academic investigation for possible clinical use. Unfortunately, whereas physicians would generally prefer the most specific agent available, economic considerations currently favor nonspecific but generally applicable agents like F-18-FDG. It is therefore quite difficult to predict which agents may be used routinely a few years hence. Specific agents are discussed herein only as examples. It is much more important to understand the general principles behind imaging and therapy with the various classes of agents, along with their potential benefits and limitations, than to understand the specific imaging appearances for any individual agent.
IMAGING
PHYSIOLOGIC
PROCESSES
Physiology covers a broad range of activities distinguished primarily by scale. Processes such as blood flow, excretion, and motion are physiologic and are amenable to imaging. Blood flow principles
3777
applicable to time-of-flight MR, Doppler US, and timing of the liver bolus in CT are similar to those involved in three-phase bone scintigraphy. Excretion principles are identical in excretory urography and the nuclear medicine renal scan. Motion analysis is common to echocardiography, MR cardiography, and SPECT cardiac stress/rest perfusion scanning. The recent paradigm shift toward “molecular imaging― indicates imaging of biologic processes in living beings at the biochemical or molecular level. Our ability to do this is rapidly growing, but it is in no sense new. This potential is also developing in more traditionally anatomic imaging. Examples include MR spectroscopy, the blood oxygenation level-dependent (BOLD) technique in MR, and MR contrast agents that assess bile excretion or Kupffer cell function. A radiology resident should not allow themselves to be cowed simply by the concept of imaging physiology. Imaging of molecular physiology is of particular value in oncology. Tumors tend to be dedifferentiated relative to their cell of origin. This has implications for loss of contact inhibition, alteration of cell surface receptors, failure of apoptosis, disordered vessel growth with relative hypoxia, alterations in transport systems for metabolically significant substances, and unregulated energy use. The rapidly growing, relatively hypoxic tumor cell has a high rate of glucose utilization, allowing FDG PET images to detect and gauge the metabolic rate of malignancies. Because P.1458 all of the underlying processes (receptor display, apoptosis, hypoxia, transport systems, and symporters) can currently be effectively imaged in vivo, any of the underlying imaging technologies may potentially spring full-grown into clinical imaging practice with little forewarning. Each physiologic alteration in a tumor cell also provides a potential target for the therapeutic application of a radiopharmaceutical attached to an appropriate ligand. All radiology residents should be forewarned to think physiologically while they continue to image anatomically. Satumomab
3778
OncoScint TAG-72 Whole Indium-11 Cytogen Corp. Nofetumomab Verluma 40-kD glycoprotein Fab Technetium-99m DuPont Pharma Arcitumomab CEA-Scan CEA Fab’ Technetium-99m Immunomedics,
Inc.
Capromab pendetide ProstaScint Prostate-specific Whole Indium-111 Cytogen
membrane
antigen
Corp.
CEA, carcinoembryonic antigen; Fab, fragment of IgG involved in antigen binding; Fc, fragment of IgG in effector mediator region; TAG-72, Name
tumor-associated Trade Name
glycoprotein-72.
Antigen
Ab Type
Label
Source
TABLE 62.1 Monoclonal Antibody Imaging Agents
CLASSES
OF
IMAGING
ISOTOPES
Isotopes suitable for tumor imaging may be divided into positron emitters and isotopes that emit single γ photons. The former are the basis for an imaging technology known as PET. The latter are
3779
employed in standard gamma camera imaging and in SPECT. Typical γ-emitting isotopes are technetium-99m (Tc-99m) and iodine-123 (I-123). γ-Emitting isotopes are advantageous because they are available with a variety of half-lives, suitable to any purpose. Several generator systems are obtainable for convenient production of the desired isotope at the point of consumption, the most common being the molybdenum-99 (Mo-99)/Tc-99m generator system. The cost of the gamma camera and generator is significantly lower than the cost of a PET scanner and cyclotron and is similar to the cost for competing technologies such as CT and MR. PET radiopharmaceuticals are very similar chemically to an unlabeled substrate molecule. γ emitters are disadvantageous in that they are not easily incorporated into molecules without disrupting their biologic function. The “atoms of life― generally do not have γemitting isotopes, although γ isotopes are available for some less common biometallic elements, such as selenium-75. Incorporation of γ emitters into biomolecules generally involves addition of significant bulk. It is difficult to produce a γ-labeled molecule that exhibits biologic behavior identical to that of the original molecule. Positron-emitting isotopes have the disadvantage of short half-life, which makes the synthesis and use of a positron emitter–labeled substance a race against time. This does not merge well with “best manufacturing practice,― at least as currently understood by the FDA. It is possible to make PET agents with gallium-68 (Ga-68), available as a generator system product. Unfortunately, the limitations that apply to single-photon agents apply equally to Ga-68.
RADIOIODINE: THE TUMOR IMAGING
INDEX
AGENT
FOR
The first radiopharmaceutical identified as having clear medical potential was radioiodine. The natural trapping of iodine by the thyroid made radioiodine a potentially ideal agent for imaging and
3780
therapy. We examine thyroid tumor imaging here briefly because it is paradigmatic of the choices involved in the development and use of tumor imaging agents generally. Iodine is trapped within the thyroid by conversion to thyroxine, which is stored within the colloid cells of thyroid tissue. Uptake is proportional to metabolic activity. Some thyroid tumors also trap iodine by the same mechanism. However, because their metabolism is disordered, they trap iodine at a much lower rate. This leads to an interesting phenomenon: thyroid tumors typically appear as cold spots in the thyroid gland on radioiodine imaging of the intact gland but as hot spots in the body when metastases are imaged after thyroid removal or ablation. Thyroid tumors arising from cells that do not trap and metabolize iodine, such as medullary carcinoma of the thyroid, will not take up radioiodine. Multiple isotopes of iodine are available. Each is suitable for a particular purpose. Reactor-produced I-131 is relatively inexpensive. It has photons that can be imaged but extensive β particle emission. It thus images and destroys thyroid tissue and can be used to ablate both malignant and benign thyroid tissue. The location of the tissues being ablated and the extent of uptake within those tissues can be determined simultaneously. I-131 is suboptimal when only imaging is desired, however, because the significant β radiation increases the risk of development of benign and malignant tumors. Cyclotron-produced I-123 has no β radiation and is safer for routine imaging. It also has a nearly ideal γ energy of 159 keV. Cyclotron production means this agent is more expensive than I131, and its relatively short half-life means that I-123 is limited to use at scheduled times. Every imaging prescription involves issues of cost, likely tumor type, adequacy of the agent for the available equipment, availability of the agent, metabolic rate and route and their effect on timing of imaging, and patient radiation dose and safety. Every therapeutic dose must be planned to achieve adequate dose to the tissue to be ablated while minimizing whole-body and nontarget organ radiation. The choices involved are typical of those in other areas of oncologic
3781
scintigraphy.
THE IDEAL TUMOR IMAGING AGENT What are the characteristics of an ideal tumor imaging agent? Foremost is specificity for the target tissue. In practice, this is never fully realized. Radioiodine is secreted by the salivary glands and swallowed. Because bowel clearance is slow, whole-body images for metastases may be problematic in the abdomen, where swallowed radioiodine in the bowel may mimic or obscure tumor. Similarly, FDG localizes in metabolically active cardiac tissue, monoclonal antibodies are bound nonspecifically to liver tissue, and Ga-67 is bound to metabolically active bone P.1459 and also taken up by normal liver. When uptake by normal tissue occurs, it is preferable that the tissue be one that is less frequently involved by the primary tumor or by metastatic disease. Because cardiac metastases and malignant pericardial effusion are much less common than liver metastases, the normal distribution of FDG is preferable, in terms of imaging sensitivity, to that of a monoclonal antibody. Rapid clearance of background activity is essential for short-lived isotopes such as pertechnetate, which must be imaged within a few hours of administration. Clearance need not be as rapid for longerlived isotopes. For whole-body I-131 scans, imaging of the abdomen may be delayed for a day to allow administration of a cathartic to purge nonspecific bowel radioactivity. Similarly, agents that clear rapidly, such as by renal excretion, are preferable to those that clear more slowly, such as by hepatic excretion. “Metabolic trapping― of an imaging agent within a tumor by any mechanism increases conspicuity and thus imaging sensitivity by increasing target activity over background. Trapping of radioiodine, for example, leads to increasing activity in tumor relative to background over time. An agent that is metabolically trapped, such as Tc-99msestamibi (Tc-99m-MIBI), will generally be more sensitive than an agent that is easily washed out, such as thallium-201 (Tl-201). In
3782
general, the higher the target-to-background ratio at the time of imaging, the higher the sensitivity for detection of tumor. An additional desirable trait is the ability of an imaging agent to predict the effectiveness of therapy. This may be inherent in the localization of the agent. Iodine avidity indicates greater likelihood of effective radioiodine therapy, and somatostatin receptor avidity increases the likelihood of successful therapy with “cold― somatostatin in carcinoid tumors. Repeat scanning after therapy may differentiate patients who responded from those who require further therapy. Thus, increased bone scan activity suggests effective therapy for bone metastases (the flare phenomenon), and decreased glucose localization on FDG PET scans indicates improved outcome after
chemotherapy
for
lymphoma.
BLOOD FLOW AND CAPILLARY PERMEABILITY: A CAVEAT Tumor imaging based on differential blood flow or capillary permeability is of incidental interest only, since these methods lack tissue specificity. Nevertheless, it is important to consider agents that localize in tumors by these mechanisms, because the potential for differential labeling on the basis of these mechanisms exists with any imaging tracer. Tc-99m-methylene diphosphonate (Tc-99mMDP) is a typical example of a blood flow imaging agent, and Tc99m-glucoheptonate is a typical example of a capillary permeability agent. During the flow and immediate static phases of a bone scan, it is occasionally possible to detect a hypervascular tumor because of increased blood flow to the tumor and a vascular blush within the tumor. On delayed static images, increased activity in bone usually is a result of increased osteoblast activity. However, the bone may have increased activity solely on the basis of increased blood flow to that bone, whatever the underlying cause. Similarly, a cold defect on a bone scan can indicate an absence of osteoblast activity caused by rapid tumor growth, or it may point to an absence of blood flow caused by tissue necrosis or vascular obstruction. These findings
3783
based on alterations of blood flow may result in either unintentional or intentional detection of tumor but are entirely nonspecific. In the classic but obsolete brain scan, Tc-99m glucoheptonate or Tc99m diethylenetriamine pentaacetic acid (DTPA) were used to detect areas of increased capillary permeability. This breakdown of the blood-brain barrier may be caused by a number of disease processes, including primary or metastatic tumors. When positive, these studies provided excellent “hot-spot― images, which allowed detection of the tumor. Again, though, the mechanism is nonspecific, and diagnosis is limited to what can be determined through pattern recognition and clinical scenario. It is important to realize that these mechanisms still apply when using modern imaging agents. A hypervascular structure may have activity that exceeds background because of high blood flow and nonspecific binding of the antibody. A tumor with high antibody affinity but poor blood flow can incidentally display activity similar or equal to background when imaged with a monoclonal antibody imaging agent. A tumor with high blood flow can appear to have high affinity for a radiolabeled antibody on the basis of preferential exposure to the imaging agent, without having specific affinity for that agent. If a brain tumor demonstrates increased activity after administration of Tc-99m-hexametazime, for example, it is difficult a priori to determine if the localization is caused by increased blood flow, diffusion across a damaged blood-brain barrier, or specific localization within the tumor. With all agents discussed in this review, the reader is urged to consider possible mechanisms for nonspecific
localization
when
interpreting
images.
CLASSES OF NON-PET TUMOR RADIOPHARMACEUTICALS Markers
of
IMAGING
Metabolism
Tc-99m-MIBI and Tc-99m-tetrofosmin (Tc-99m-TFos) are typical metabolic agents for gamma camera imaging. Both were developed
3784
for myocardial perfusion imaging. In oncology, they are of greatest potential interest for scintimammography. These agents were initially designed to substitute for Tl-201 in P.1460 evaluating myocardial viability. Tl-201 had been serendipitously observed to localize within some tumors. Evaluation of Tc-99m-MIBI for tumor localization was a logical extension of this prior work. While the mechanism of localization is quite different, these agents fill a niche similar to that of FDG. These agents are lipophilic, cationic materials whose uptake is dependent, in part, on cellular and mitochondrial membrane potentials. Malignant tumors have negative membrane potentials and high mitochondrial content to support their high metabolic rate. Thus, unlike FDG, these agents are indirect markers of metabolic activity. Both agents also bind to P glycoprotein (P-gp), a mediator of multidrug resistance in breast cancer cells. When the multidrug resistance gene is present and active in a cell, P-gp is manufactured by the cell. P-gp traps cancer drugs by the same mechanism employed to trap these imaging pharmaceuticals: both are actively transported out of the cell. Thus, a breast cancer that localizes Tc-99m-MIBI or Tc-99m-TFos is likely to be responsive to chemotherapy. Alternatively, a breast cancer that does not localize these agents is likely to express multidrug resistance. Tl-201 apparently is not transported out of the cell by Pgp. This is reasonable, because Tl-201 is a simple ion and is not representative of the types of large molecules (anthracyclines, vinca alkaloids, actinomycin D) that are usually transported by P-gp. The greatest advantage of metabolic agents for staging is their ability to detect small or diffuse foci of primary or metastatic disease that are not detectable on anatomic images or are not pathologic by size criteria. They are also very useful for determining the presence or absence of recurrence in the bed of a treated tumor. The major disadvantage of these agents is that they are nonspecific: They image high metabolic activity rather than tumor per se. For example, FDG may localize in an active tuberculous granuloma. With firm clinical acceptance of FDG, the role of γ-emitting agents in this arena has faded rapidly. They remain one good reason to look at the
3785
projection images on myocardial perfusion studies (Fig. 62.1 )
FIGURE
62.1. Lung Cancer on Myocardial Perfusion Study.
Frontal (A) and left anterior oblique (B) views from acquisition for technetium-99m-sestamibi (Tc-99m-MIBI) myocardial perfusion study. A diffuse, mild, and unilateral increase in lung activity with vague focal uptake in right hilum is apparent.C . Coronal image from PET scan in same patient. The Tc-99m-MIBI was localizing in the right retrohilar area (non–small cell lung cancer) and throughout the lung because of lymphangitic spread of tumor.
3786
P.1461
Monoclonal
Antibodies
Polyclonal antibodies are formed in response to an antigenic stimulus. The polyclonal antibody formed is really a set of chemically different substances that share the basic antibody structure but have widely variant affinities for the intended target of the antigenic response. The variations occur because a large number of different cells respond with rapid division and antibody secretion when they are stimulated by an antigen. Each activated cell divides rapidly to form a clone of identical cells that produce one of the antibodies in the set. The set of antibodies produced by the many activated cell lines is thus termed polyclonal. Polyclonal antibodies have variable affinities for nontarget tissues. For an imaging agent, it is clearly desirable to have a single antibody of known and reliable affinity. A single antibody has the additional advantage of decreased nonspecific cross-reactivity with tissues other than the intended target. Such antibodies must be produced by a single clone, the descendants of a single cell. It is thus termed a monoclonal antibody. Monoclonal antibodies have three important advantages over polyclonal antibodies for imaging: maximum specificity, maximum sensitivity, and predictable nontarget binding. By providing a high affinity, the monoclonal antibody reduces the dose of labeled antibody necessary to produce an acceptable target (tumor)-tobackground ratio. By selecting a monoclonal antibody with high affinity, one simultaneously selects for greater specificity for that target. Nevertheless, some cross-reactivity may be inevitable, especially if the target tumor is very similar to a tissue normally present in the body. These cross-reactions need to be constant so that the appearance of a normal scan may be determined with a high degree of certainty.
Basic
Antibody
Structure
3787
A variety of antibody fragments and modified antibodies are employed for both imaging and therapy. Imaging agents to date have generally been based on the IgG antibody. This discussion will be limited to IgG for the sake of brevity, although there is no intrinsic reason that other antibody classes (IgA, IgE, IgD or IgM) could not be employed for imaging. Each IgG molecule has two binding regions for antigen (Fig. 62.2 ). An additional host binding region mediates effector functions such as blood clearance and nonspecific liver binding. When separated from the remainder of the antibody, the antigen-binding regions are termed Fab fragments and the effector mediator region is termed an Fc fragment. In the intact antibody, two Fab fragments are bound together by disulfide bonds. The Fc fragment is bound to each Fab fragment by covalent bonds. It is possible to remove the Fc region while retaining the disulfide bonds between Fab regions. The resulting fragment is termed Fab2 . The antibody also consists of constant, variable, and hypervariable regions. The constant regions are the same from antibody to antibody for a given antibody class and species of organism. The variable regions differ between antibodies within a class and contain the binding sites for antigen. The hypervariable regions have the greatest effect on antibody affinity and are those portions of the variable region that are most exposed at the antigen-binding site.
3788
FIGURE
62.2. Monoclonal
Antibody
Structure. The basic
antibody glycoprotein molecule consists of two identical heavy and light chains linked by a disulfide bridge. Each chain has a variable region that is responsible for antigenic binding and a constant region that is involved in complement fixation and antibody-dependent cell toxicity. Each variable region, in turn, contains three hypervariable regions that form unique antigen-binding sites. The antibody can be fragmented into smaller units, depending on the cleavage plane. Enzymatic digestion produces either a Fab2 fragment or two smaller Fab fragments and an Fc fragment. Fragment size dictates blood pool clearance rates, with the smaller ones clearing faster.
The behavior of antibody fragments in biologic systems is predictable in part on the basis of molecular weight alone. Smaller fragments partition more rapidly into various cellular and extracellular tissues because of their smaller size. They are also metabolized and cleared from the bloodstream more rapidly for the same reason. A small fragment is thus maximally exposed to the target tissue earlier but for a shorter period of time. Smaller fragments will be favored for imaging when binding is rapid and strong, as this will result in rapid target labeling and rapid background clearance. Local recurrences, nodal metastases, and peritoneal metastases may be better detected with fragments
3789
because their intact vasculature will insure that antigen will be exposed to the antibody. Smaller fragments will be advantageous, and are often required, for isotopic labels with relatively short halflives. The Tc-99m label, with a 6-hour half-life, requires rapid localization and background clearance, so imaging can be performed within a day. Smaller fragments are also less immunogenic and thus less likely to induce an undesirable immune response. Larger fragments or intact antibody may be preferable when a long reaction time is required, such as when binding is weak or slow, or for therapeutic
applications
in P.1462
which extended residence time in the target tissue is required to maximize radiation dose to the tumor. Because larger fragments and intact antibodies are cleared more slowly from the background, target-to-background ratios are lower. However, absolute uptake in tumor is higher with intact antibody. Slow background clearance increases whole-body and nontarget dose in both imaging and therapeutic applications. Larger fragments allow, and may require, use of isotopes with longer half-lives. Biochemical differences between fragments also affect biodistribution. The Fc region is of greatest interest because it mediates nonspecific binding to the liver. For imaging agents, nonspecific liver binding significantly decreases sensitivity for detection of liver metastases. Metastases may even appear as photopenic rather than photo-enhanced areas. High activity in the liver may also obscure activity in adjacent structures unless a lead shield is employed to cover the liver. Unfortunately, use of a shield also carries the risk of obscuring an area of tumor activity. Dualisotope imaging with simultaneous acquisition of planar or SPECT data may allow digital subtraction of normal tissue or comparison of the biologic distributions of an organ imaging agent (e.g., Tc-99m sulfur colloid [Tc-99m-SC]) with that of the labeled antibody to increase diagnostic accuracy. Finally, sequestration of an antibody or fragment in the liver reduces bioavailability of the agent for binding to the intended target. For a target outside the liver, rapid clearance of the background through nonspecific liver binding could
3790
improve imaging sensitivity. In practice, this advantage seldom outweighs the disadvantages listed. The use of antibody fragments for imaging began as an attempt to address the problem of nonspecific liver binding. Other techniques, such as pretreatment with Fc fragment to reduce nonspecific binding, have variable success. The use of smaller fragments for imaging results in significantly increased renal activity. This can be addressed by administration of unlabeled leucine, which saturates the renal binding sites for the fragments and reduces their residence time in the kidneys. Mixtures of amino acids, commercially available for total parenteral nutrition, appear to work just as well. Another approach to limiting renal activity is to use smaller fragments consisting only of the active binding site or even of fragments of this active site. It is important to recognize that monoclonal antibodies for imaging are usually produced from mouse hybridoma cells. These cells result from the fusion of antibody-producing B lymphocytes with “immortal― myeloma cells. From the numerous fusion products that result, a single clone that produces high-affinity antibody is selected. This cell line is then grown to produce the antibody. Because the mouse antibody is a foreign protein in the human bloodstream, it may incite an immune response. This results in production of human antimouse antibody (HAMA), which can cause mild to severe symptoms. Of greater concern, subsequent administrations of mouse antibody can result in mild to fatal (anaphylactic) reactions. This problem could limit the repeat use of monoclonal imaging or of therapeutic agents based on murine antibodies. A variety of attempts have been made to reduce the antigenicity of the monoclonal imaging agents. Chimeric antibodies are formed from the variable regions of mouse antibodies and the constant regions of human antibodies. Humanized antibodies are formed by inserting the hypervariable region of the mouse antibody into a human antibody using recombinant DNA techniques. These modifications improve product safety at the cost of increased complexity and expense of
3791
production. Clearly, monoclonal antibody imaging carries a significant risk of side effects. Before using these agents, physicians should thoroughly familiarize themselves with the side effect profile for the agent and be prepared to deal with any complications. Prior to administration, the risk of HAMA reaction should be assessed in those patients who have previously undergone monoclonal antibody imaging or therapy. Some current agents of interest are available with either a diagnostic or a therapeutic label (or a label suitable for both), so that the likelihood of successful therapy can be assessed in advance of administration of the therapeutic agent. The agents of greatest current interest are CD20 antibodies for therapy of lymphoma (discussed later). Figure 62.3 illustrates a normal antibody fragment scan. Figure 62.4 illustrates a positive scan.
Peptide
Receptor
Scintigraphy
Regulatory peptides are a group of polypeptides that serve as internal messengers. They bind to targets on multiple tissues, most notably those of the brain and GI system. They are expressed variably on the surfaces of a variety of human tumor cell lines. In patients in whom a stable radioactive label can be achieved without significantly decreasing the affinity of the receptors for the polypeptide, imaging and therapy are possible for specific tumor cells. The index compound, and the only one commercially available at the time of this writing, is indium-111-DTPA-D-Phe1 -octreotide (In-111-OCT). This protein binds with high affinity to the sst2 somatostatin receptor and with lesser affinity to two of the remaining five receptors for somatostatin. Somatostatin receptors are expressed by neuroendocrine tumors, some neural tumors (meningiomas, medulloblastomas, most astrocytomas and neuroblastomas), and small cell lung carcinomas. They are also highly expressed on the cell surfaces of carcinoid tumors and by cells in some nonmalignant conditions such as Wegener granulomatosis, rheumatoid arthritis, and tuberculosis. sst2 receptor with the attached ligand is internalized by the cell,
3792
The
resulting in effective metabolic trapping of the radioligand within the tumor cell. In-111-OCT has >90% sensitivity for detection of carcinoid tumors. It can also be used to detect gastrinomas, insulinomas, glucagonomas, and other (or unclassified) APUDomas (Fig. 62.5 ). It also has fair sensitivity for medullary thyroid carcinoma (65%). P.1463 It detects primary small cell lung carcinoma with 90% sensitivity and nonliver metastases from small cell lung carcinoma with 70% sensitivity. It has successfully detected recurrent small cell lung carcinoma in patients who were in complete remission by standard criteria. The agent is rapidly cleared by renal excretion, with 90% excretion at 24 hours. It should be clear that the agent lacks specificity, but this is not significant in most tumor imaging applications. The agent will also image active granulomas, which are a potential source of false-positive scans.
3793
FIGURE 62.3. Normal Carcinoembryonic Antigen Scan (Arcitumomab). Whole-body anterior and posterior images.
Tc-99m depreotide became an important agent during the relatively brief period it was available. This agent has affinity for three of the five types of somatostatin receptors and was an excellent agent for imaging non–small cell lung cancer. SPECT imaging with this agent remains the most cost-effective method for assessing solitary pulmonary nodules that are equivocal for tumor. Depreotide was achieving significant clinical acceptance when manufacturing problems resulted in a drop in chemical and radiochemical purity,
3794
causing the agent to be removed from the market. Although the “impure― product was essentially equally useful for diagnostic purposes and clearly harmless, it was removed from the market to protect the public. The irony in this age of revelations about the dangers of (for example) cyclooxygenase-2 inhibitors could not be more stark. The equivalent assessment with FDG is significantly more expensive, although the FDG scan has decreased background and improved resolution (Fig. 62.6 ). Unfortunately, past experience indicates that an agent is unlikely to again achieve wide acceptance once taken off the market. The extensive replacement of synthetic cholecystokinin with fatty meal for gallbladder ejection fraction, P.1464 resulting from temporary unavailability of the agent because of manufacturing problems, is instructive in this regard. An interesting case of depreotide labeling in an early mesothelioma is illustrated in Fig. 62.7 .
FIGURE
62.4. Metastatic
Colon
Carcinoma.
Indium-111-
OncoScint study, anterior (A) and posterior (B) views. Patient is status postresection of colon carcinoma with rising carcinoembryonic antigen. Study demonstrates paraspinous adenopathy (long arrow ). Less obvious is the fixed uptake in the pelvis (short arrow ), which remained unchanged over several days of imaging and was proven
3795
to be locally recurrent carcinoma.
Mixed
Mechanisms
The traditional gamma camera imaging agent Ga-67 depends on receptor binding but also falls within the category of metabolic marker agents. The metabolic activity imaged by Ga, at least in part, is the cellular competition for nutrient iron. Ga-67 ion, administered as the citrate salt, mimics free iron. It is rapidly bound to transferrin receptors on tumor cells, which are overexpressed in an attempt to maximize this scarce P.1465 resource required for rapid growth. Limitation of free iron is a major defense mechanism of the body, although this is directed primarily at defense from pathogens. It is mediated by free transferrin protein in serum and by transferrin receptors on white cells. Pathogens also compete for free iron through production of soluble iron chelators termed siderophores . Ga tumor activity thus depends both on the availability of transferrin receptors on the tumor and the metabolic activity of the tumor, which determines the extent to which these receptors will be expressed. Tumors for which Ga-67 imaging was historically useful are now generally evaluated with other techniques (FDG PET for lymphomas and MR or multiphase CT for hepatocellular carcinoma). The high-energy γemissions of Ga-67 result in “fuzzy― images that were instrumental in creating the “unclear― medicine reputation (Fig. 62.8 ). Tumor localization by Ga is now more significant as a confounding factor when Ga-67 is employed for infection imaging.
3796
FIGURE 62.5. Metastatic metastatic islet cell tumor anterior liver and spleen B . Technetium-99m sulfur
Islet Cell Tumor. Patient with known was evaluated for receptor status. A . The were imaged with indium-111-octreotide. colloid liver-spleen scan shows multiple
defects corresponding to the tumor. C . Subtraction study shows uptake by the tumor receptors.
3797
FIGURE 62.6. False-Positive Depreotide and Fluorodeoxyglucose (FDG) Scans. Coronal image (A) from technetium-99m-depreotide SPECT lung scan and coronal image (B) from FDG PET scan in a patient with diffuse pulmonary abnormalities. These were felt to represent granulomatous disease radiographically, based on the CT scan (C) . Granulomas were subsequently demonstrated on biopsy, and stability was demonstrated on long-term clinical and imaging follow-up. The exact etiologic factor remains unknown. Note the nearly identical pathologic distribution of the radiopharmaceutical in the two scintigraphic scans. The high lung background on the depreotide scan makes it less appealing, but the information content is identical for the two scans.
Lymphoscintigraphy has undergone a revival as the result of the
3798
sentinel node concept. Cutaneous melanoma and breast cancer tend to migrate to the nearest draining node without skip lesions in nodes further along the drainage chain. Absence of tumor within the first draining lymph node makes presence of other metastatic disease highly unlikely. If the sentinel node can be located and demonstrated to be free of disease, more extensive lymph node dissection can be avoided. This approach is now the standard of care in breast cancer in most localities because it reduces the morbidity of surgical therapy, especially postoperative lymphedema. Sentinel node biopsy is more accurate than FDG PET. This fact has limited the use of FDG PET in breast cancer role to restaging of recurrent tumor. PET retains a primary role in the diagnosis and staging of melanoma. Sentinel node biopsy is less useful in melanoma because of the relatively poor success of current melanoma therapies other than resection of an P.1466 P.1467 isolated primary tumor for cure. The role of sentinel node biopsy in other tumors, especially head and neck tumors, remains in flux.
3799
FIGURE
62.7. Technetium-99m-Depreotide
Scan
in
Mesothelioma. Axial CT image (A) 3 weeks prior to depreotide scan shows pleural thickening (arrows ) in the left thorax caused by mesothelioma. Axial depreotide image at the approximate location of CT slice (B) and coronal (C) depreotide image both show diffuse uptake in the pleura, indicating that tumor involvement is more widespread than indicated by CT. D . Axial slice from CT scan 3 months later shows marked progression (arrows ).
3800
FIGURE
62.8. Lymphoma. Anterior chest images from gallium-67
studies
pretreatment (A) and posttreatment (B) for lymphoma
demonstrate the complete resolution of mediastinal (long arrow ), parenchymal (short arrow ), and left supraclavicular nodal (open arrow ) disease.
Any colloidal agent can in principle be employed for lymphoscintigraphy. Agents with particle sizes smaller than 50 nm are ideal. Reasonable agents are Tc-99m-SC and Tc-99m human serum albumin (Tc-99m-HSA). Tc-99m-SC is somewhat suboptimal because of its large particle size, which results in significant retention at the injection site. This problem is partially overcome by filtration to obtain only the smaller particles. Tc-99m-HSA immediately enters the lymphatic system and allows rapid dynamic imaging of lymphatic channels with reduced overall imaging time. However, retention in lymph nodes is significantly less than with Tc99m-SC, which may make Tc-99m-HSA less optimal for use of an intraoperative gamma probe to localize the sentinel node. In general, Tc-99m-HSA is probably preferable for imaging and Tc99m-SC for intraoperative use, but either can be used effectively. Only Tc-99m-SC is currently available in the United States. Technical factors can significantly affect the outcome of the study. There has been extensive debate on the best method of injection;
3801
some favor axillary subdermal, some periareolar subdermal, and some intratumoral injection. At the current time, it appears that the periareolar approach is both most accurate and best tolerated. A transmission scan or other body image scan (i.e., bone scan) is required to allow effective interpretation of node location (Fig. 62.9 ). Head and neck lesions have especially short drainage chains. Since these lesions drain inferiorly, injection should occur above (cranial to) the level of the lesion to ensure that activity at the injection site does not obscure the sentinel node. Body wall lymphatics are highly variable. Both ipsilateral and contralateral imaging is required with body wall injections to ensure that all nodal groups draining the lesion are detected. The need for imaging in breast cancer sentinel node lymphoscintigraphy is hotly debated. Imaging will reveal drainage to nodes outside the axilla. This is important only in those centers in which surgeons are willing to sample the extra-axillary nodes when present. Currently, these centers are in the minority. Imaging is still mandatory for nonextremity melanoma, which can drain to multiple sentinel node sites, whereas imaging is useful, but not necessarily mandatory, for extremity melanomas. These always drain first to the axilla or groin, but the intraoperative search is facilitated by imaging. Head and neck sentinel node imaging clearly requires meticulous high-resolution imaging for success because of the complex and densely packed anatomy. Hybrid or fusion imaging is of great potential interest to provide additional anatomic information in this setting.
Other
Tumor
Imaging
Agents
Metaiodobenzylguanidine can be used for detection and therapy of neuroblastoma and pheochromocytoma (Fig. 62.10 ), and iodinated cholesterol derivatives can be used for adrenal cortical carcinoma. Bone imaging agents for both primary bone tumors and metastatic disease are discussed in Chapter 55 . In addition to localizing areas of increased or absent bone repair (“hot― and “cold― lesions, respectively), these agents have a tendency to localize
3802
within malignant effusions and P.1468 malignant ascites. The mechanism for this localization has yet to be fully explained. The use of Tc-99m-MIBI for localization of benign but clinically significant parathyroid tumors is also discussed elsewhere.
FIGURE 62.9. Malignant Melanoma. Example of lymphoscintigraphy, with filtered technetium-99m sulfur
colloid
(Tc-
99m-SC) injected intradermally in four locations around the site of malignant melanoma located on the back at the inferior aspect of the left clavicle. Transmission imaging is performed with a Tc-99m flood source, causing the patient to cast a “shadow― for positioning information. A . Image is slightly off center, true posterior with left arm held out to the side. B . A left lateral image. In both images the injection sites have been digitally subtracted (a “halo― remains in the lower portion of each image) so that axillary nodal activity is more readily apparent. Imaging of the remainder of the trunk failed to demonstrate drainage elsewhere.
3803
FIGURE 62.10. Metastatic Pheochromocytoma. Anterior (A) and posterior (B) whole-body images with iodine-131metaiodobenzylguanidine in a patient with widely metastatic pheochromocytoma. (Courtesy of Dr. Peter Blue, Columbia, South Carolina.)
Pretreatment
with
agents
designed
to
improve
target-to-background
ratios may become more common in the future. These agents may increase binding of the imaging agent to the intended target or they may decrease binding of the agent to nontarget tissues. The best current example is the use of “cold― rituximab to saturate nonspecific binding sites prior to imaging with In-111-ibritumomab prior to yttrium-90-ibritumomab tiuxetan therapy.
DUAL-AGENT SCINTIGRAPHY, HYBRID IMAGING, AND FUSION IMAGING Dual-agent
imaging may improve sensitivity for detection of
metastases from certain classes of tumors in which the metastatic foci have variable receptor expression because of dedifferentiation or origin from different cell clones. FDG and I-131 may be used in combination to simultaneously or sequentially image differentiated (I-131) and dedifferentiated (FDG) metastases from thyroid cancer. The combination of Tc-99m-MIBI and In-111-OCT has also been suggested as a means for determining multiple drug resistance in
3804
small cell lung carcinoma. Presumably, FDG or Tl-201 in combination with any of the MIBI-like agents could achieve the same effect in tumors that do not have somatostatin receptors. Fusion imaging is the exact superimposition of the physiologic information from a nuclear medicine study with the anatomic information from CT or MR using fiduciary markers. Hybrid imaging achieves the same goal more elegantly by performing both an anatomic and a functional study on a single system without patient movement between scans. PET-CT hybrid scanners have captured nearly all of the recent PET scanner market. Current research would tend to favor this approach over visual or fusion imaging of separate anatomic and functional studies. My own experience and reading of this literature suggest there is significant bias resulting from comparison of hybrid imaging to either anatomic or functional images alone. Those studies that have compared hybrid to fused or visual comparison of anatomic and functional images have not tended to use interpreters who are significantly experienced in both the anatomic and the functional technique employed—a nearly unavoidable bias in academic centers where radiology and nuclear medicine tend to be separate endeavors. It is likely that a properly performed, definitive study will be moot before it is available (if ever). Just as the speed and convenience of CT angiography for assessment of pulmonary embolus resulted in a large-scale conversion to this technique before the comparisons with ventilation/perfusion lung scanning were published, the simplicity of reading hybrid images is selling the technique long before the intellectual justification is available. The issue has already been settled de facto by sales of hybrid scanners. Conversely, hybrid scanners have the potential to salvage some γemitting agents that have been too difficult for general use. Especially notable is the improvement in sensitivity and specificity for imaging with the monoclonal antibody In-111 capromab pendetide when SPECT-CT hybrid scanners are employed (Fig. 62.11 ). Without reliable anatomic coregistration, this agent is extremely unreliable.
3805
A resident being shown a tumor imaging case in nuclear medicine at the current time may expect that anatomic P.1469 imaging for comparison will be provided if hybrid or fusion images are not provided. The current standard of care requires correlation of functional and anatomic imaging by some method. The resident has a right to expect that the test cases will allow him or her to work at the standard of care.
FIGURE
62.11. (Color Plates) SPECT-CT of Metastatic Prostate
Carcinoma With ProstaScint. Coronal images presented in standard sequence (scintigram, fusion, CT). (Courtesy of Dr. Milton Gross, VA Hospital and University of Michigan, Ann Arbor, Michigan.)
TUMOR THERAPY WITH RADIOPHARMACEUTICALS A complete discussion of therapy with radiopharmaceuticals is beyond the scope of this text, which is intended primarily to help residents learn and review information important in the board examination in radiology. The discussion will be limited to the example provided by two anti-CD20 monoclonal antibodies for lymphoma therapy: I-131 tositumomab (trade name Bextra) and Y90 ibritumomab tiuxetan (trade name Zevalin). These agents exemplify the general principles involved in tumor therapy with radiopharmaceuticals. Radiology residents should be aware that
3806
therapeutic radiopharmaceuticals are available for treatment of bone metastases (commonly 89-strontium chloride [SrCl2 ] or samarium153 ethylenediamine tetramethylene phosphoric acid). These are essentially bone scan agents with β emitters. I-131 metaiodobenzylguanidine is employed to treat neuroblastoma. Somatostatin-avid polypeptides with β-emitter labels are available to treat neuroendocrine tumors, at least under protocol. Agents are under development for most tumor types that will clearly be part of nuclear medicine and radiation oncology practice. Most radiologists will not be directly involved in administering these therapies, but most radiologists will be involved in determining the correct therapy. Both Bextra and Zevalin were developed because of the success of rituximab (Rituxan) in treating low-grade B-cell non-Hodgkin lymphoma. This anti-CD20 antibody has been remarkably successful in palliating low-grade lymphoma, for which a cure is seldom achieved because of the slow growth rate. Generally, tumor therapy other than resection is only effective for growing cells. Since rituximab can be used repeatedly in the same individual, long-term therapy selects for cells with reduced cell surface CD20 receptors that are not killed by rituximab. Radiopharmaceuticals
can
overcome
the
increasing
chemoresistance
of these tumors through the process of crossfire. A monoclonal agent with a β emitter attached will deliver radiation to cells with the CD20 receptor. By choosing a β emitter with a mean free path that is greater than a typical cell radius, the radiation damage can be extended to a few tumor cells distant to the target. This crossfire from cells retaining CD20 receptors and adequate blood supply can kill tumor cells that have dedifferentiated to lose the CD20 receptor and cells whose blood supply is partially disrupted. The therapy treats tumors that have become resistant to standard therapy. It also renews the therapeutic effectiveness of the unlabeled monoclonal antibody by increasing the relative number of CD20expressing cells in the tumor population. Bextra is labelled with I-131. Because this isotope is also a γ emitter, it can be imaged and provides its own dosimetry scan. The
3807
major disadvantage of the agent is the need to hospitalize the patient as a radiation safety precaution for bystanders. Zevalin relies on Y-90, a pure β emitter with no γ component. A patient receiving this agent does not have to be hospitalized. However, to visualize biodistribution, it is necessary to scan with antibody labeled with the γ emitter In-111 in the week prior to the therapeutic dose (Fig. 62.12 ). Bextra and Zevalin have roughly equivalent efficacy. It is likely that only one of these agents will survive the marketplace. It is also likely that the outcome will depend more on cost, convenience, and marketing than on any absolute advantage in outcome, even if such an advantage is ultimately established.
FIGURE
62.12. Zevalin Therapy of Low-Grade Non-Hodgkin
Lymphoma. Pretreatment axial CT image (A) shows innocuousappearing groin nodes (arrowheads ). Anterior (B) and posterior (C) pretreatment scans with indium-111-ibritumomab showed extensive uptake in the groin nodes (arrowheads ). Biopsy of a groin node confirmed recurrent disease. After treatment, the patient was free of
3808
recurrence for more than a year.
P.1470
A
COMPLETE
ONCOLOGY
IMAGING
APPROACH In addition to primary diagnosis, staging, and therapy, radiology residents should remain attuned to the value of ancillary nuclear medicine techniques in the oncology setting. Pretherapy and posttherapy assessment of cardiac function with tagged red blood cell equilibrium ventriculography remains important for patients receiving cardiotoxic chemotherapy. Ga scanning, tagged white blood cell agents, and the newer monoclonal antibody agent for infection imaging (Tc-99m fanolesomab; trade name NeutroSpec) will remain important to assess posttherapy infectious complications. Ga-67 may also be useful to assess lung toxicity from chemotherapeutic agents such as bleomycin. Ga-67 is especially useful to assess fever of unknown origin, when both recurrent tumor and infection are in the differential diagnosis. Effective tumor screening remains the holy grail of imaging, but successful examples in nuclear oncology are limited. Scintimammography of the dense breast, for all of its theoretical advantages, has made little headway in most practices. If the economics are right, newer breast-specific cameras may increase the use of scintimammography. Bone scan for bony metastatic disease is the only widely employed nuclear medicine screening method in oncology. The concept that bone scan is unnecessary in the absence of bone pain, while likely true for some tumors like prostate cancer, is not universally true for all tumors. Familiarity with molecular imaging as a research tool for understanding basic molecular biologic processes is valuable. The resident should also be alerted to techniques of preclinical development imaging employed by drug manufacturers. An occasional foray into the literature in these areas is of great value to maintain enthusiasm for the potential of molecular imaging, however
3809
remote the clinical applications may seem. Tumor imaging with PET, for example, labored in research laboratories for many years but has now become the standard of care. This list is not intended to be inclusive. During the board examination, as in real life, expertise in handling problems at the fringe of a field of practice often distinguishes the proficient practitioner. Oncology should be approached with a combination of all available radiologic and nuclear medicine techniques to achieve the best outcome for the patient. Residents should seek to become proficient in comparing the effectiveness of all modalities available to them. Your clinical colleagues will ultimately show their appreciation for this effort through consultation and referral.
Summary Oncologic nuclear medicine is an expanding field. The indications for radiopharmaceuticals in the P.1471 detection and treatment of primary and metastatic cancer will doubtless increase. The advances in cost-effective PET imaging technology have greatly strengthened the position of nuclear radiology in the cost-conscious care environment. Radiologists will need to understand not only the indications, contraindications, and pitfalls of this new imaging technology but also its position in the cost-effective delivery of care.
SUGGESTED
READINGS
Aboagye EO. Positron emission tomography imaging of small animals in anticancer drug development. Mol Imaging Biol 2005;7:53–58. Barrio J, Marcus C, Hung JC, Keppler JS. A rational regulatory approach for positron emission tomography imaging probes: from “first in man― to NDA approval and reimbursement. Mol Imaging
Biol
2004;6:361–367.
3810
Baulieu F, Bourlier P, Scotto B, et al. The value of immunoscintigraphy in the detection of recurrent colorectal cancer.
Nucl
Med
Commun
2001;22:1295–1304.
Britz-Cunningham SH, Adelstein S. Molecular targeting with radionuclides: state of the science. J Nucl Med 2003;44:1945–1961. Burak Z, Moretti JL, Ersoy O, et al. 99mTc-MIBI imaging as a predictor of therapy response in osteosarcoma compared with multidrug resistance-associated protein and p-glycoprotein expression. J Nucl Med 2003;44:1394–1401. Carlisle MR, Lu C, McDougall IR. The interpretation of 131 I scans in the evaluation of thyroid cancer, with an emphasis on false positive findings. Nucl Med Commun 2003;24:715–735. Dadachova E, Carrasco N. The Na+ /I - symporter (NIS): imaging and therapeutic applications. Semin Nucl Med 2004;34:23–31. Danthi SN, Pandit SD, Li KCP. A primer on molecular biology for imagers: VII. Molecular imaging probes. Acad Radiol 2004;11:S77–S84. Debray MP, Geoffroy O, Laissy JP, et al. Imaging appearances of metastases from neuroendocrine tumours of the pancreas. Br J Radiol 2001 Nov;74:1065–1070. Duet M, Sauvaget E, Petelle B, et al. Clinical impact of somatostatin receptor scintigraphy in the management of paragangliomas of the head and neck. J Nucl Med 2003;44:1767–1774.
3811
Fuster D, Munoz M, Pavia J, et al. Quantified 9 9 m Tc-MIBI scintigraphy for predicting chemotherapy response in breast cancer patients: factors that influence the level of 9 9 m Tc-MIBI uptake. Nucl Med Commun 2002;23:31–38. Higuchi T, Taki J, Sumiya H, et al. Intense 201 Tl uptake in giantcell tumor of bone. Nucl Med Commun 2002;23:595–599. Hutton BF, Braun M. Software for image registration: algorithms, accuracy, efficiency. Semin Nucl Med 2003;33:180–192. Ilias I, Pacak K. Current approaches and recommended algorithm for the diagnostic localization of pheochromocytoma. J Clin Endocrinol
Metab
2004;89:479–491.
Jani AB, Spelbring D, Hamilton R, et al. Impact of radioimmunoscintigraphy on definition of clinical target volume for radiotherapy after prostatectomy. J Nucl Med 2004;45:238–246. Keidar Z, Israel O, Krausz Y. SPECT/CT in tumor imaging: technical aspects and clinical applications. Semin Nucl Med 2003;33:205–218. Krausz Y. Normal SPECT/CT scans: artifacts and pitfalls in iodine, iodine-MIBG, and indium octreotide imaging. In von Schultess G, ed. Clinical Molecular Anatomic Imaging. Philadelphia: Lippincott Williams & Wilkins, 2003:441–446. Kushner BH. Neuroblastoma: a disease requiring a multitude of imaging studies. J Nucl Med 2004;45:1172–1188. Lewington VJ. Bone-seeking radionuclides for therapy. J Nucl Med 2005;46:38S–47S.
3812
Li, KCP. A primer on molecular biology for imagers: IX. How to become a “molecular imager.― Acad Radiol 2004;11:S97–S100. Machac J, Krynyckyi B, Kim C. Peptide and antibody imaging in lung cancer. Semin Nucl Med 2002;32:276–292. Meier DA, Kaplan MM. Radioiodine uptake and thyroid scintiscanning. Endocrinol Metab Clin 2001;30:291–313. Menda Y, Kahn D. Somatostatin receptor imaging of non-small cell lung cancer with 9 9 m Tc depreotide. Semin Nucl Med 2002;32:92–96. Mozley P. Weaving single photon imaging into new drug development. Mol Imaging Biol 2005;7:30–36. Pashankar FD, O’Dorisio M, Menda Y. MIBG and Somatostatin receptor analogs in children: current concepts on diagnostic and therapeutic use. J Nucl Med 2005;46:55S–61S. Pelosi E, Bello M, Giors M, et al. Sentinel lymph node detection in patients with early-stage breast cancer: comparison of periareolar and subdermal/peritumoral injection techniques. J Nucl Med 2004;45: 220–225. Reubi JC, Macke HR, Krenning EP. Candidates for peptide receptor therapy today and in the future. J Nucl Med 2005;46:67S–75S. Robbins RJ, Schlumberger MJ. The evolving role of 131 I for the treatment of differentiated thyroid carcinoma. J Nucl Med 2005;46:28S–37S.
3813
Schettino CJ, Kramer EL, Noz ME, Taneja S, Padmanabhan P, Lepor H. Impact of fusion of indium-111 capromab pendetide volume data sets with those from MRI or CT in patients with recurrent prostate cancer. AJR Am J Roentgenol 2004;183:519–524. Schirrmeister H, Arslandemir C, Glatting G, et. al. Omission of bone scanning according to staging guidelines leads to futile therapy in non-small cell lung Cancer. Eur J Nucl Med Mol Imaging
2004;31:
964–968.
Sharkey RM, Goldenberg DM. Perspectives on cancer therapy with radiolabeled monoclonal antibodies. J Nucl Med 2005;46:115S–127S. Silverman DH, Delpassand ES, Torabi F, Goy A, McLaughlin P, Murray JL. Radiolabeled antibody therapy in non-Hodgkins lymphoma: radiation protection, isotope comparison and quality of life issues. Cancer Treat Rev 2004;30:165–172. Stokking R, Zubal IG, Viergever MA. Display of fused images: methods, interpretation and diagnostic improvements. Semin Nucl
Med
2003;33:219–227.
Taillefer R. Clinical applications of 9 9 m scintimammography. Semin Nucl Med
Tc-sestamibi 2005;35:100–115.
Thomasson DM, Gharib A, Li KCP. A primer on molecular biology for imagers: VIII. Equipment for imaging molecular processes. Acad
Radiol
2004;11:S85–S96.
Torabi M, Aquino SL, Harisinghani MG. Current concepts in lymph node imaging. J Nucl Med 2004;45:1509–1518.
3814
Wagner JD, Evdokimow DZ, Weisberger E, et al. Sentinel node biopsy for high-risk nonmelanoma cutaneous malignancy. Arch Dermatol 2004;140:75–79. Wahl RL. Tositumomab and 131 I therapy in non-Hodgkin's lymphoma. J Nucl Med 2005;46:128S–140S. Wilczek B, Sandelin K, Eriksson S, Larsson SA, Jacobsson H. Sentinel node scintigraphy in breast cancer using a dual tracer technique. Nucl Med Commun 2004;25:135–138. Willkomm P, Bender H, Bangard M, Decker P, Grunwald F, Biersack HJ. FDG PET and immunoscintigraphy with 9 9 m Tclabeled antibody fragments for detection of the recurrence of colorectal carcinoma. J Nucl Med 2000;41:1657–1663. Yuksel M, Cermik TF, Doganay L, et al. 9 9 m Tc-MIBI SPET in nonsmall cell lung cancer in relationship with Pgp and prognosis. Eur J Nucl Med 2002;29:876–881.
3815
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XII - Nuclear Radiology > Chapter 63 - Central Nervous System Scintigraphy
Chapter
63
Central Nervous Scintigraphy
System
David H. Lewis Vivek
Manchanda
TWO-DIMENSIONAL SCANS
(PLANAR)
BRAIN
Before the advent of CT, the planar nuclear medicine brain scan, using radiopharmaceuticals like technetium-99m diethylenetriaminepentaacetic acid (Tc-99m-DTPA) or Tc-99m glucoheptonate (Tc-99m-GH), was used to detect a breakdown in the blood-brain barrier (BBB). Currently, a similar study using agents that cross the intact BBB, such as Tc-99m hexamethylpropyleneamine oxime (Tc-99m-HMPAO) or Tc-99m ethyl cysteinate dimer (Tc-99m-ECD), can detect absence of cerebral blood flow, which is characteristic of brain death. The normal BBB protects the CNS by preventing entry of harmful substances. Most materials are excluded from the CNS on the basis of molecular size and chemical characteristics. Active transport mechanisms are present for certain key nutrients such as glucose.
Radiopharmaceuticals Planar brain scanning for the detection of the BBB breakdown is typically performed with Tc-99m bound to either DTPA or GH. Any
3816
agent that does not normally cross the BBB can potentially be employed, although agents of cellular size (radiolabeled red blood cells, for example) will be excluded even by a damaged BBB.
Technique A dose of 15 to 20 mCi of Tc-99m-DTPA or Tc-99m-GH is injected into an arm vein. Flow images are typically obtained at a rate of one image every 3 seconds for a total of 60 seconds, with the camera anterior to the head. Anterior, posterior, and lateral static images are subsequently obtained; vertex images, obtained by placing the camera at the vertex of the skull, are often useful. A lead collar is employed to exclude radiation from the radiopharmaceutical localized below the neck. Immediate static images are useful to evaluate blood pool abnormalities, while delayed static images after clearance of the background activity are of greater value to detect breakdown of the BBB.
Interpretation Interpretation of the static images depends primarily upon detection or exclusion of radiopharmaceutical localization within the brain parenchyma. Some activity is invariably present from the radiopharmaceutical within the soft tissues of the scalp and within intracerebral blood vessels. Increased or asymmetric localization indicates breakdown of the BBB. This finding is entirely nonspecific; it is present in conditions as diverse as cerebral infarction, primary or metastatic tumor, and infectious processes. For this reason, clinical information is essential for interpretation. The presence of a lenticular photo-enhanced (or occasionally photopenic) rim can be used to diagnose subdural hematoma. The normal radionuclide angiogram is characterized by prompt symmetric perfusion. Asymmetric flow in the carotid arteries may indicate occlusive disease. The so-called “flip-flop sign― (decreased activity in the arterial phase and increased activity in the venous phase) may be seen in carotid occlusion. Vascular malformations, high-grade or vascular tumors (e.g., glioblastoma
3817
multiforme and meningioma), and inflammatory processes have increased flow. Low-grade or benign tumors, areas of porencephaly or edema, and occlusive processes have decreased flow. The complete absence of brain activity in the presence of prompt common carotid and scalp flow indicates brain death. The traditional brain scan with Tc-99m-DTPA or Tc-99m-GH has largely been superseded by planar perfusion brain scans in clinical practice to corroborate the impression of brain death. There is little doubt that modern brain radiopharmaceuticals such as Tc-99mHMPAO or P.1473 Tc-99m-ECD provide, with greater assurance, the status of cerebral blood flow. These scans are also much easier to interpret, although they are more costly to perform because of the higher expense of the radiopharmaceuticals employed. Total cessation of cerebral blood flow including posterior fossa structures can be demonstrated with this technique, which is required by the Uniform Determination of Death Act (passed by the U.S. Congress in 1981).
CEREBROSPINAL
FLUID
STUDIES
CSF is formed in the choroid plexus as an ultrafiltrate of plasma. It flows from the ventricles through the foramina of the fourth ventricle and ascends over the convexities of the brain to be absorbed by the arachnoid villa. Processes that impede flow over the convexities or absorption of the fluid by the villi result in communicating hydrocephalus. Tracer techniques are ideal for imaging of this process, because they are injected in small amounts and do not alter CSF flow. Processes that obstruct the outflow from a ventricle are more difficult to assess by these techniques because injection can be made directly into the ventricle. Patency and flow in therapeutic shunts and reservoirs can easily be evaluated by injecting the tracer directly into the device.
Radiopharmaceuticals
and
Technique
The standard cisternogram is performed by intrathecal injection of a
3818
sterile, pyrogen-free radiopharmaceutical. The only approved agent currently marketed for this purpose is indium-111-DTPA (half-life = 2.8 days). The injection of 0.5 mCi follows a spinal tap performed in the standard manner. Initial images may be obtained to ensure intrathecal injection. Subsequently, the radiopharmaceutical ascends to the basilar cisterns in approximately 4 hours and flows over the convexities within 24 hours in a normal individual. Images of the basilar cisterns are obtained at 4 to 6 hours. If images at 24 hours show ascent over the convexities with activity in the interhemispheric fissure and relative clearance of the basilar cisterns, imaging may be terminated. Otherwise, images should be obtained at 48 and 72 hours. CSF shunt and reservoir studies are performed by direct injection of the device with radiopharmaceutical in a small volume. Maintenance of sterile techniques during the injection is critical. It is also critical to understand the specific device being evaluated, as shunts often contain check valves and reservoir capacities are limited. A patient may also have several shunt tubes, some of which may be known to be occluded. In general, it is best to have direct input from the neurosurgeon involved in the case to ensure that the maximum amount of information is obtained. A small amount of CSF, up to 1 mL, may be withdrawn for lab studies prior to the tracer injection. Dynamic gamma camera imaging for 10 minutes is usually first performed in the supine position. If radiotracer drains from the reservoir, then in cases of ventriculoperitoneal shunts, an abdominal image is obtained. If the shunt leads elsewhere, such as the atrium, imaging of the thyroid and chest is performed.
Application
to
Hydrocephalus
Standard cisternography is performed primarily to evaluate for normal-pressure hydrocephalus and for CSF leak. Normal-pressure hydrocephalus is a form of communicating hydrocephalus clinically associated with ataxia, dementia, and urinary incontinence. Cisternography demonstrates early localization of activity within the lateral ventricles persisting beyond 24 hours and delayed clearance over the convexities (Fig. 63.1). While these findings indicate an
3819
increased likelihood of a clinical response to shunting, they are not univariate predictors of outcome. Other forms of communicating hydrocephalus (such as might result from radiation therapy or intrathecal chemotherapy) can also be evaluated with cisternography.
Application
to
Cerebrospinal
Fluid
Leak
Cisternography has high sensitivity for CSF leak and remains the procedure of choice for this condition. Its sensitivity results from the ability of tracer technique to detect very small amounts of activity that may be intermittently leaking. Imaging is performed between 1 and 3 hours after injection and also at 24 hours and perhaps 48 hours. Patient and camera positions are chosen to maximize the likelihood of detection, with lateral views for CSF rhinorrhea and anterior views for CSF otorrhea. Cotton pledgets should be placed in the nostrils after intrathecal injection of tracer. These are counted at 4 to 6 hours in a well counter. A serum sample from peripheral blood drawn concurrently is also counted. Pledget activity exceeding 1.5 times the serum concentration is evidence for CSF rhinorrhea. However, because this method does not give morphologic information about the leakage, DiChiro et al. suggested the use of imaging as well.
Application
to
Shunts
and
Reservoirs
Shunts are evaluated primarily for patency. If the proximal portion is occluded manually (or contains a check valve), flow through the distal limb can be evaluated. The tracer should flow freely into the peritoneum (for ventriculoperitoneal shunts). Delayed flow or persistent activity at the shunt tip suggests malfunction. Diffusion will typically allow determination of the level of obstruction, even when the flow is absent. Reservoir injection tests for proper placement, patency, and proper functioning of the reservoir. If the reservoir empties directly into the ventricle (such as an Omaya shunt placed for intrathecal chemotherapy), noncommunicating hydrocephalus may be excluded by normal progression of activity to
3820
the basilar cisterns and over the convexities. Ventricular spinal shunts may be evaluated only by direct injection of radiopharmaceutical into the ventricle. The half-time to clearance from the reservoir can be P.1474 calculated and should usually be less than 8 minutes. If the CSF flow is disturbed and opening pressure is low, the blockage is in the proximal limb. If opening pressure is high (more than 20 cm H2 O), the obstruction is distal to the reservoir.
FIGURE 63.1. Normal-Pres sure Hydrocephalus on Cisternogram. Selected images from an indium-111 diethylene triaminepentaacetic acid cisternogram in lateral views shown at 4 to 6 hours, 24 hours, and 48 hours. Abnormal uptake is seen in lateral ventricles that persists through 48 hours.
In instances of distal malfunction of the shunt system, the clearance
3821
curve is flat and the half-time to clearance of the reservoir indicates infinite values. For complete distal obstruction or rupture of abdominal catheter, the isotope fails to migrate through the tubing, whereas in partial distal obstruction caused by relatively high abdominal pressure or obstruction of the valve, extremely low radionuclide clearance is encountered. Peritoneal loculations are typically characterized by stagnation of radioactivity at the initial site of appearance within the abdominal cavity and absence of uniform diffusion in the peritoneum. Clearance studies of isotope are not of much help in proximal obstruction, as the time required for isotope to reach distal site of shunt depends on many factors, including the patient's CSF production, the proportion of CSF circulating through normal pathways, resting intraventricular pressure, the patient's position prior to the test, opening pressure of the shunt system, length of the tubing, and variations of intraventricular pressure caused by coughing, straining, or crying. Conversely, absence of ventricular reflux seems to be a highly reliable scintigraphic feature in proximal obstruction. Inability to measure intraventricular pressure, to aspirate CSF freely, or to inject the isotope into the ventricle were considered of some help in assessing proximal shunt obstruction. Vernet et al suggest that a shuntogram exhibiting good opacification of the distal shunt system and normal clearance of the isotope, but no ventricular reflux, cannot be considered as normal but rather as inconclusive.
FUNCTIONAL BRAIN IMAGING DIMENSIONS (TOMOGRAPHIC SCANS)
IN THREE BRAIN
Radiotracer techniques may be employed to evaluate blood flow in the cerebral microvasculature. Suitable for this purpose are diffusible radiotracers such as xenon-133, tracers that are passively taken up by neural tissues as lipophilic substances, and tracers that effectively function as “chemical microspheres.― True microspheres that
3822
lodge in and thus obstruct capillaries are contraindicated because they could cause a stroke. Therefore, these radiopharmaceuticals must cross the BBB, enter, and be retained in cells. Glucose consumption and blood flow are linked in normally functioning brain tissues and in most pathologic processes. Therefore, the relative localization of SPECT blood flow or PET metabolic tracers in various cortical tissues gives a reasonable qualitative indication of relative function.
Positron
Emission
Receptor
Imaging
Tomography
The paradigmatic technique for performance of in vivo biochemical and functional evaluation of the brain is PET. This technique allows qualitative and quantitative evaluation of receptor systems within the brain. Adrenergic, cholinergic, dopaminergic, serotonergic, benzodiazepine, and opioid receptors have P.1475 been evaluated extensively. PET allows true biochemical assessment of the properties of these receptor systems, such as affinity, saturation, and nonspecific bindings, as well as more general information about distribution and uptake kinetics. These unique capabilities of the PET technique provide an extremely valuable research tool for evaluation of brain biochemistry and development of both imaging agents and therapeutic pharmaceuticals. However, these studies are expensive, experimental, and time consuming. Experimental uses should be clearly differentiated from proven clinical indications; the associated ethical issues have been assessed and summarized by the Brain Imaging Council of the Society of Nuclear Medicine.
Metabolic Imaging With Fluorine-18Fluorodeoxy glucose (F-18-FDG) Clinical indications for PET almost exclusively involve F-18-FDG, although a tailored examination with other agents may be warranted
3823
in evaluating the status of some brain tumors. The primary clinical indications for PET are evaluation of epilepsy, dementia, and glioma. PET use in epilepsy is predominantly in the presurgical evaluation of mesial temporal lesions resulting in partial complex seizures. Approximately 85% of these can be cured surgically if the focus of the abnormal brain is limited to a single temporal lobe. F-18 scanning must be performed interictally in seizure patients, as the 1.83-hour half-life of F-18 does not allow it to be held available for ictal scanning. Interictally, the seizure focus is usually hypometabolic. This is now thought to be caused by interruptions with adjacent neurons, which reduce neural activity and thus metabolism. Loss of neural connections can also result in decreased metabolism in more distant sites (diaschisis). Sites of temporal epilepsy are identified as hypometabolic foci in 70% of interictal scans, with a false-positive rate of only 5%. Dementia evaluation is performed predominantly to distinguish Alzheimer-type dementia from other types of dementia and from pseudodementia caused by depression. While the only unequivocal test remains brain biopsy, detection of Alzheimer disease (AD) is very reliable when applied to an appropriate population at risk. Patterns of activity with F-18-FDG are similar to those described for SPECT agents subsequently in this review. The role of PET and SPECT scanning in evaluation of brain tumors has been discussed elsewhere and will not be repeated here. While PET is certainly a useful technique in each of these considerations, there are techniques available for SPECT imaging in each of these conditions that have similar sensitivity and specificity at decreased cost. Because both SPECT cameras and the appropriate imaging pharmaceuticals are more likely to occupy the predominant role in routine clinical application of functional brain imaging, with PET reserved for special circumstances. For this reason, the technique and interpretation of these studies are not discussed in detail. The interested reader is referred to one of the many excellent reviews available for PET brain imaging.
Single-Photon
Emission
3824
Computed
Tomography Radiopharmaceuticals Much of the early work in this area was performed with xenon-133. This inhaled gas dissolves in blood to an extent adequate for imaging. The rapidity of perfusion and diffusion of this agent makes rapid imaging essential. Therefore, multiple probe–type cameras have predominantly been employed. This tracer is not well suited to the rotating camera SPECT technique. For this reason, and because of difficulties in handling and recovering a gaseous agent, this agent has largely been superseded by other radiopharmaceuticals. Iodinated
amphetamines tagged with iodine-123 replaced radioxenon
for a time. These agents readily cross the BBB. Both uptake and BBB diffusion are reversible. This agent, therefore, will slowly redistribute over time. Iodoamphetamine is also immediately sequestered by and slowly released from the lung. This effectively yields slow intraarterial injection over a period of hours. Because of these phenomena, iodoamphetamine images represent integration of all brain activity from the time of injection until completion of imaging. Because iodine-123 is cyclotron produced and has a relatively short half-life (13.2 hours), availability has proven problematic. This agent also has been largely superseded and is no longer commercially available in the United States. Agents currently in widespread use include Tc-99m-HMPAO and Tc-99m-ECD. Tc-99m-HMPAO is an agent of the chemical microsphere type. This agent crosses the BBB and is trapped within the brain substance. The mechanisms proposed for trapping have included change in the ionic state, binding to glutathione, and chemical decomposition. For purposes of scan interpretation, it is necessary to understand only that this agent essentially crosses the BBB irreversibly. Unlike iodinated amphetamine, this agent provides a “snapshot― of brain activity for a short period after injection (less than 10 minutes, with peak activity usually within 1 minute after IV injection). The HMPAO is available as a kit that is combined with the generatorproduced, freshly eluted pertechnetate prior to use. Availability is
3825
thus not problematic. The initial form of this agent was unstable in aqueous solution. It had to be used within 30 minutes after preparation, which made quality control procedures difficult. A stabilized form is now available that can be used for 4 hours after aqueoust preparation. Tc-99m-ECD is also an agent of the “chemical microsphere― type. This agent, unlike Tc-99m-HMPAO, does not localize in areas of luxury perfusion (Fig. 63.2). While there are a number of subtle differences between Tc-99m-ECD and Tc-99m-HMPAO, the remaining differences are not of routine clinical relevance. Tc-99m-ECD is stable in P.1476 aqueous solution for at least 6 hours and is therefore preferable when attempting ictal injection and scanning of seizure disorders. This advantage is somewhat offset by a higher pharmaceutical cost. Both agents have their proponents and are in common use.
3826
FIGURE
63.2. (Color Plates) Ethyl Cysteinate Dimer (ECD)
Versus Luxury
Hexamethylpropyleneamine Oxime (HMPAO) in Perfusion. Transverse axial images of technetium-99m
(Tc-99m)-ECD on rows 1 and 3 and Tc-99m-HMPAO on rows 2 and 4. The photopenic defect in left frontal lobe on ECD imaging shows uptake on HMPAO imaging. This was a subacute stroke.
Technique Brain-perfusion SPECT scans are preferably performed with a multidetection rotating camera. A dual-head rotating camera can be employed, but single-head cameras are not generally recommended at this time. Cameras limited to brain work are not generally necessary because of improvements in equipment since the first edition of this book, although they may be useful in practices with a
3827
large brain scanning referral base. High-resolution collimators should be employed. The key issues for dual imaging are distance from the brain to the detector head and total counts acquired. One should try to achieve the smallest possible radius of rotation camera while remaining cognizant of patient comfort and safety. Prior to the procedure, the physician should review history, the recorded neurologic physical examination, and the mini mental status examination if available. Historical data of importance include symptoms and duration; history of stroke, head injury, or seizure; any medications (especially psychotropic anticonvulsants); and whether CT or MR scans have been performed. The neurologic examination should include cognitive and motor examination P.1477 and cranial nerve examination. The mental status examination should address orientation, registration, attention, recall and language functions. The scanning agent is injected with a patient in a controlled resting state. This usually involves a supine, resting patient with closed eyes in a quiet room (or a room with white noise) and indirect lighting (see procedure guidelines). Alternatively, tracer can be injected during seizure ictus or acute stroke. The IV line should be established in advance, and all instructions and questions should be dealt with prior to injection to avoid unintended stimulation of brain activity. The patient should remain in this controlled environment for at least 5 minutes after injection. Injection of 15 to 30 mCi of Tc99m-HMPAO or Tc-99m-ECD should be employed (0.2 to 0.3 mCi/Kg in pediatric patients). A delay after injection of no less than 60 minutes and preferably 90 minutes should be employed with Tc-99mHMPAO to allow for background clearance. No less than 10 minutes and preferably 60 minutes should elapse before Tc-99m-ECD imaging with SPECT. Quality control on the radiopharmaceutical should be performed prior to injection, according to the package insert. Careful patient monitoring is mandatory throughout the scan, since patients considered for the study often have dementia, neurologic dysfunction, stroke, psychiatric disease, or another condition that requires monitoring.
3828
If sedation is required, it should be given after injection of the radiopharmaceutical if at all possible. Attention to medications is critical, since both the presence of and withdrawal from (prescription and illicit) drugs can affect biodistribution of the tracer within the brain. ECD or stabilized HMPAO should be prepared in advance if ictal scanning is to be attempted. The patient must be monitored carefully and continuously with IV catheter in place and ready availability of agent nearby and injected very rapidly after seizure onset, since generalization of seizure focus can occur very rapidly. To deploy properly, the nursing staff need to perform the injection and thus must be trained in radiation safety and proper handling of radiopharmaceuticals. Acetazolamide (Diamox) is employed to assess vasodilatory reserve, as it is a cerebrovascular-specific vasodilator. It should be given as a slow IV push over 5 minutes at about 20 minutes prior to radiotracer injection. The typical dosage is 1,000 mg for adults and 14 mg/kg for children. This agent is contraindicated for patients with known sulfa drug allergy, history of complicated migraine headache, or within 3 days of acute stroke. It rarely may cause postural hypotension, increasing the need for monitoring when arising from the scanning table. Dipyridamole and adenosine have also been used to assess vascular reserve but are not in routine use. All patients should void immediately prior to imaging to improve comfort. This is especially important when using acetazolamide, which is a mild diuretic.
Processing
and
Interpretation
Filtering should be performed in all three dimensions. A low-pass filter is generally recommended, typically a Butterworth. Other filters can be used, but spatially varying filters may create artifacts, particularly in low-count studies. The whole brain should be reconstructed, taking care to include the vertex and cerebellum. Any summing of data should occur only after reconstruction at single pixel slices. Attenuation correction should always be performed in intact skulls and the data should be evaluated in three orthogonal planes. It is also important to evaluate the raw data for acquisition errors. These should be evaluated on rotating display to check for
3829
patient motion during acquisition, which can also create serious artifacts. Rigorous quality control is required for all these as well as other SPECT studies. Averaged slices should approach the full width at half maximum of the device. It is worthwhile to reconstruct the transaxial plane images along the anterior/posterior commissural line. A typical normal study is demonstrated in Fig. 63.3. In interpretation, it is important that background subtraction not be excessive and that a continuous color scale (or a continuous gray scale) be employed to avoid artifactual edges. The range of normal should be considered in rendering an opinion, and one should use a normal database that is specific to the radiopharmaceutical employed. Correlation with clinical data is required. Any regional perfusion defects should also be correlated with the locations of any CT or MR abnormalities. This is substantially easier if some type of fusion imaging (overlay of anatomic and functional images with or without the use of fiduciary markers) is employed, but it can be performed visually. The extent and severity of defects should be reported. The recommendations for image acquisition, processing, and interpretation conform generally and specifically to the recommendations of the Brain Imaging Council of the Society of Nuclear Medicine in place at the time this chapter was written. Readers are encouraged to check the current guidelines prior to initiating a brain imaging program (http://www.snm.org) .
Indications According to the American Academy of Neurology (AAN), the established indications for these techniques are confirmation of AD, presurgical ictal identification of seizure foci, and evaluation of acute brain ischemia. Most nuclear medicine physicians in clinical practice would consider brain death evaluation to be an established indication for planar brain scans. Areas considered promising by the AAN are determination of stroke subtypes, assessment of vasospasm following subarachnoid hemorrhage, and (nonictal) localization of seizure foci. Therapeutic options for these disorders, including medical or
3830
interventional, are available. The expansion of research into use of brain SPECT in psychiatric disorders, where therapeutic options are broad but objective P.1478 evaluation of biologic aspects of the diseases is extremely weak, is promising but as yet not validated. The remainder of the discussion will be limited primarily to the accepted indications, with brief discussion of other applications considered promising, applications to presurgical planning (balloon occlusion test and Wada test), and testing of vascular reserve.
FIGURE 63.3. (Color Plates) Normal Hexamethylpropyleneamine Oxime (HMPAO)
Study. Selected
images from a normal technetium-99m-HMPAO study. A . Axial images. B . Representative central images in three standard planes (axial, sagittal, and coronal to the long axis of the brain). The large arrow indicates the brainstem; the small arrow indicates the basal ganglia. With modern equipment, it should be possible to routinely obtain scans that resolve the gyri, basal ganglia, and brainstem. Interpreters should consult an atlas to familiarize themselves with the normal distribution of radiotracer in the brain structures.
3831
Applications for brain SPECT include the following.
Stroke
and
Ischemia
The extent of stroke can be determined a short time after its occurrence with functional brain scanning (Figs. 63.4, 63.5). Imaging with Tc-99m-ECD has demonstrated a specificity of 98% and sensitivity of 86% for localization of acute strokes. This contrasts with several hours for standard MR and CT, although MR molecular diffusion imaging has allowed improved assessment of strokes acutely. Therapeutic administration of tissue plasminogen activator is effective, but the time between initial onset and therapy is critical. The time of onset must be established and hemorrhagic stroke excluded prior to administration of tissue plasminogen activator. CT is employed to rapidly exclude hemorrhage, and administration is begun as soon thereafter as possible. Functional imaging has not been widely employed on the theoretical basis that it would delay therapy when stroke has already been confirmed clinically. Therapy of acute stroke has otherwise been limited to anticoagulation and supportive care. Anticoagulation requires only exclusion of hemorrhage, which is best accomplished with CT. The use of functional imaging to evaluate for ischemia caused by vasospasm after subarachnoid hemorrhage is in widespread use. Vasospasm of clinical impact occurs in 30% of patients with subarachnoid hemorrhage (Fig. 63.6). SPECT brain imaging in association with neurologic examination and transcranial Doppler artery narrowing assessment allows effective noninvasive monitoring and early intervention. Because there are multiple potential therapies (hyperdynamic hypertension, hemodilution and hypovolemia, calcium-channel blockers, micro-balloon cerebral angioplasty, intraarterial papaverine), which require definitive diagnosis, this technique has significant clinical use. In patients with ischemic stroke, a nearly normal brain SPECT study may be an indication of lacunar infarction. Using this criterion, SPECT
3832
was 69% sensitive and 100% specific in identifying a lacunar stroke. High-resolution SPECT has the potential to image lacunar strokes as well. Early differentiation of the mechanism is important, since embolic disease should prompt cardiac evaluation for a source and consideration of anticoagulant therapy. Prediction of prognosis after stroke or transient ischemic attack (TIA) is another area of likely application. Lesion volume correlates with early outcome in acute stroke, and large, P.1479 severe perfusion defects are predictive of nonnutritive perfusion. After TIA, a prolonged deficit on functional scanning with brain SPECT predicts a high likelihood of ischemic stroke in the period following the TIA.
FIGURE 63.4. (Color Plates) Cerebral Infarction. Transaxial images reformatted into the plane of the orbitomeatal line
3833
(standard CT format) from a iodine-123-iofetamine scan show a large region of absent perfusion (arrows) in the distribution of the left middle cerebral artery after cerebral infarction. Note the decreased cerebellar activity on the side opposite the infarct (arrow), an example of crossed cerebellar diaschisis. This phenomenon results from decreased right cerebellar metabolism caused by decreased neuronal communication between the right cerebellum and the infarcted portion of the left cerebral hemisphere. (All iofetamine images in this chapter were obtained during phase III trials of the agent under approved protocol. Modern equipment allows significant improvement in resolution.)
In the early phase of cerebrovascular compromise, blood flow is maintained through autoregulated vasodilatation, leading to an increase in blood volume. As the compensatory capacity of autoregulation is exceeded, blood flow falls, while oxygen metabolism is maintained, corresponding to an increase in oxygen extraction fraction (the beginning of misery perfusion). Once oxygen extraction fraction has increased maximally, continued decline in blood flow leads to a decline in oxygen delivery and metabolism (the onset of ischemia). Severe and prolonged compromise results in infarction of brain tissue, with decreased demand for oxygen metabolism, while vasodilatation persists, leading to a decline in oxygen extraction fraction (and onset of “luxury perfusion―). As revascularization occurs, blood flow to the region increases and the infarcted area typically remains in a state of luxury perfusion for days to weeks. By judicious use of acetazolamide to test vasodilatory reserve of the cerebral vessels, it is possible to perform the equivalent of pharmacologic coronary stress imaging for cerebral vessels. Studies with and without acetazolamide may provide information on the mechanism of ischemia. These studies may also be useful in presurgical planning when carotid surgery or intracranial/extracranial bypass surgery is contemplated, because they can indicate the physiologic significance of an anatomic vascular lesion. Interpretation
3834
of these studies depends on identification of a significant area of relatively decreased perfusion (actually indicating increased perfusion in the unaffected portions of the brain) after stimulation that was not present on the study without stimulation, which would indicate impaired vasodilatory reserve (Fig. 63.7). This is exactly analogous to evaluation of coronary artery reserve with dipyridamole or adenosine stress imaging, as discussed in Chapter 57. Ramsay et al used a combination of acetazolamide and cerebral Tc-99m-HMPAO to demonstrate reduction in cerebral blood flow reserve, which improved after carotid endarterectomy in 45% of selected patients with unilateral ICA stenosis. As with coronary studies, evaluation of clinical data, accurate registration of images, and comparison to a radiopharmaceuticalspecific normal database are important. Injection during vascular occlusion of a carotid artery can test crosscirculation across the circle of Willis, demonstrating the precise areas of decreased perfusion P.1480 during occlusion. This is the nuclear medicine version of the Matas test (Fig. 63.8). The distribution of amobarbital injected for localization of speech and memory functions (the Wada test) may also be assessed accurately using functional agents as tracers.
3835
FIGURE 63.5. (Color Plates) Subcortical Cerebral Infarct. Transverse axial images of technetium-99m ethyl cysteinate dimer in patient with acute stroke shows absence of uptake in the left lenticular nucleus.
Dementia The most important use for PET imaging in the workup of the dementia patient is to aid in making an accurate diagnosis as early in the course of AD as possible and, a review by Silverman concluded that F-18-FDG PET appears to supersede SPECT in this area. It can be diagnosed with an accuracy of approximately 80% with functional SPECT brain scans. In patients with AD of varying severity, the magnitude and extent of hypometabolism correlate with the severity of the dementia symptoms. Usually, there are only minor decreases
3836
in the parietal lobes in patients with early mild AD. Moderately affected patients show significantly decreased metabolism in the left midfrontal lobes, bilateral parietal lobes, and the superior temporal regions. In patients with severe AD, the same regions are affected, but the hypometabolism is much more pronounced, with sparing only of the sensorimotor, visual, and subcortical areas. The typical pattern is decreased activity in parietal and posterior temporal regions bilaterally but often asymmetrically (Fig. 63.9). The characteristic AD radiotracer distribution pattern is similar for SPECT and PET imaging. The classic bilateral temporoparietal defect of AD can also be seen in other conditions, including carbon monoxide poisoning, hypoglycemia, mitochondrial encephalomyelopathy, severe Parkinson disease, and diffuse Lewy body dementias (DLBD). The relatively recent development of several pharmaceuticals for AD provides an important area for PET imaging. For example, patients P.1481 treated with donepezil were found to have relatively stable cerebral metabolism at 24 weeks compared with a placebo group, which showed a 10% decline.
3837
FIGURE
63.6. (Color Plates) Subarachnoid Hemorrhage and
Cerebral Vasospasm. A . Transverse axial images of technetium-99m ethyl cysteinate dimer show baseline scan on rows 2 and 4 and scan during vasospasm on rows 1 and 3. New defect is seen in left hemisphere in the posterior frontoparietal cortex corresponding to ischemia from vasospasm. B . Angiogram of left internal carotid artery shows vasospasm in the midportion of the left middle cerebral artery. C . Angiogram of left internal carotid artery shows resolution of vasospasm after percutaneous
3838
transluminal
micro-balloon
angioplasty.
Frontotemporal lobe degeneration (including the subtype Pick disease, which is associated with cognitive and language dysfunction and behavioral changes) can generally be differentiated from AD. Frontotemporal lobe degeneration is classically associated with frontotemporal perfusion defects on brain SPECT. Again, the pattern is not specific to this group of diseases (depression, alcoholism, schizophrenia, Pick disease, severe AD, and progressive supranuclear palsy form the differential). Vascular dementias can result in defects in any portion of the brain and can coexist with AD in a proportion of the elderly. Further complicating the picture is an increasing tendency for nonspecific defects to occur in older patients, especially as a result of sulcal enlargement owing to brain atrophy. The evaluation of dementia is, therefore, not a simple task and P.1482 should require anatomic as well as functional imaging; it is recommended that quantitative analysis software be employed for corroboration. The use of quantitative software techniques (Fig. 63.10) has aided the visual interpretation and accuracy of dementing diseases and has helped to identify hypoperfusion or hypometabolism of the posterior cingulate gyrus and parietal precuneus as the earliest findings in AD.
3839
FIGURE 63.7. (Color Plates) Acetazolamide Vascular Reserve Testing in Occlusive Carotid Disease. Transverse axial images of technetium-99m ethyl cysteinate dimer with acetazolamide on rows 1 and 3 and at rest on rows 2 and 4. Images show decreased uptake in the left hemisphere with acetazolamide, which improves largely at rest. This finding indicates exhausted vasodilatory reserve in the left hemisphere corresponding to left internal carotid artery occlusion.
3840
A negative PET scan indicated that pathologic progression of cognitive impairment during the mean 3-year follow-up was unlikely to occur. There is substantial clinical overlap between elderly patients with AD, vascular dementia, and pseudodementia caused by depression. Thus, having scans that can help in the differential diagnosis is quite important. The therapeutic options (such as cholinesterase inhibitors in the recent past and now glutamate regulators) for treating AD have been slowly emerging and have had some clinical benefits in the last decade. DLBD, which is characterized by visual hallucination, fluctuating cognitive decline, and Parkinsonian symptoms, has been reported to be the second most common type of degenerative dementia, accounting for up to 20% of all dementia cases at autopsy. SPECT findings include temporoparietal hypoperfusion similar to that seen in AD patients and occipital lobe hypoperfusion. Occipital hypometabolism (or hypoperfusion) P.1483 measured by PET and SPECT yielded differential diagnostic accuracy of probable DLBD from probable AD with sensitivity and specificity of 92%. Similarly, hypometabolism in the primary visual cortex differentiated probable DLBD from probable AD with 86% sensitivity and 91% specificity. An MR study indicated that absence of medial temporal atrophy could separate DLBD from AD or vascular dementia with specificities of 100% and 88% respectively, but sensitivity was only 38%.
3841
FIGURE
63.8. (Color Plates) Positive Balloon Occlusion
Study. This patient was asymptomatic after inflation of a left carotid artery balloon, but images show decreased regional blood flow in the left middle cerebral artery distribution (A). Baseline study (B) shows normal tracer distribution. If occlusion study had been normal, the baseline images would be omitted. The radiopharmaceutical is technetium-99m hexamethylpropyleneamine oxime.
Vascular disease contributing to dementia, which is potentially reversible, can be unmasked by administration of acetazolamide with SPECT and compared against anatomic imaging, suggesting vascular disease. In multi-infarct dementia, patients typically present abruptly with a history of prior stroke and hypertension. Generally speaking, multiple perfusion deficits are evident that may be equivalent or more extensive than the findings on an MR scan. The likelihood of development of etiologically specific therapies for at least some
3842
dementias remains high, making continued investigation of diagnostic techniques of utmost importance. Follow-up scans improve sensitivity and specificity by demonstrating an appropriate pattern of progression in true dementias.
Seizure
Disorders
The seizure disorder most commonly referred for medical imaging is the patient with a temporal lobe syndrome, typically complex partial seizures. Interictally, seizure foci tend to have decreased activity, and increased activity is seen with ictus (Fig. 63.11). In simplistic terms, focal neuronal loss or damage results in a local phenomenon similar to diaschisis. Because of secondary activation and spread of the seizure foci, it is difficult to pinpoint the exact seizure focus unless injection can be made early during the seizure. This should be confirmed by video/electroencephalogram (EEG) review. However, ictal SPECT may identify seizure foci, which are not identified by MR, and is especially useful in epileptogenic cortical developmental disorders, where EEG often fails to find the epileptogenic focus. An ictal SPECT study showing an area of increased regional blood flow, which may correspond to an area of decreased regional blood flow on interictal SPECT, is strong evidence for an epileptogenic P.1484 P.1485 lesion. Ictal SPECT is reported to be more accurate than interictal SPECT and PET, with overall accuracy rates greater than 90%. In a group with structural lesions, the two techniques were shown to have comparable results. Both ictal SPECT and interictal PET have complementary roles where localization is difficult. If these findings also concur with the EEG findings and or CT/MR evidence of a lesion, the need for more invasive depth electrocorticography or intraoperative electrocorticography may be obviated or more accurately guided. For this indication, the technique will be applied mainly in larger referral centers that perform surgery for removal of the seizure foci.
3843
FIGURE
63.9. PET in Alzheimer Disease. Transaxial plane
from fluorine-18 fluorodeoxyglucose PET scan in Alzheimer disease. Note preservation of metabolism in sensorimotor cortex, visual cortex, basal ganglia, thalami, and cerebellum. Note deficit of metabolic activity in temporoparietal association cortex bilaterally. (Image courtesy of Satoshi Minoshima, MD, PhD.)
3844
FIGURE
63.10. (Color
Plates)
Three-dimensional
SSP
Display of SPECT in Alzheimer Disease. Stereotactic threesurface projection display of Z-scores in brain SPECT in Alzheimer disease showing statistically significant decreases in perfusion in bilateral temporoparietal association cortices and also in posterior cingulate. GLB, THL, CBL, and PNS indicate that each corresponding row is normalized to global, thalamic, cerebellar, and pons uptake, respectively.
3845
FIGURE 63.11. (Color Plates) Ictal and Interictal SPECT Scans in Complex Partial Epilepsy. Transverse axial images of technetium-99m ethyl cysteinate dimer injected with seizure ictus on rows 1 and 3 and also injected during interictal time period on the right temporal uptake in
rows 2 and 4. Ictal imaging shows increased uptake in hemisphere, which is predominantly in the right lobe. Interictal imaging shows relatively decreased the right temporal lobe.
3846
When newer methods for analyzing PET images such as statistical parametric mapping were used to detect temporal interhemispheric asymmetry, hypometabolism was identified on the side chosen for resection in most cases (sensitivity, 71%; specificity, 100%) and was predictive of favorable postsurgical outcome in 90% of the patients. The findings from some studies have suggested that metabolic dysfunction of the thalamus ipsilateral to the seizure focus becomes more severe with long-standing temporal or frontal lobe epilepsy and also with secondary generalization of seizures. FDG PET is also a sensitive and specific technique for investigating patients with seizures of probable frontal lobe origin, because many of these seizures begin in the medial or inferior aspects of the frontal lobe and scalp EEG readings do not provide adequate localization of foci. Performance of ictal PET studies is logistically impractical, primarily because of the relatively short half-life of the positronemitting isotopes, such as F-18 or oxygen-15. Receptor PET imaging with carbon-11-flumazenil, although still investigational, may have a useful clinical role in patients with partial epilepsy who have normal or nondiagnostic FDG PET, in patients with bilateral FDG findings but unifocal seizure activity on EEG, and in patients after surgical resection who continue to have seizures. Excision of well-localized foci can lead to elimination of seizures or significantly improve pharmacologic control in 80% of surgical patients. Viral encephalopathies can present a diagnostic imaging challenge. Launes et al reported abnormally increased accumulation of Tc-99mHMPAO in the affected temporal lobe, even at an early stage of herpes simplex encephalopathy, when CT was normal.
Brain
Tumors
PET can play an important role in the evaluation and management of patients with brain tumors, including the grading of tumors, determination of prognosis, and differentiation of recurrent tumor from radiation necrosis. The sensitivity for making the determination of radiation necrosis versus tumor recurrence may be as high as 86%, with a specificity as high as 56%. FDG studies have concluded
3847
that high-grade tumors are hypermetabolic, whereas low-grade tumors are hypometabolic. One distinction from this typology is P.1486 juvenile pilocytic astrocytomas, which typically have a high glucose metabolism despite their benign nature. It should be noted that PET does not differentiate between primary lymphomas of the CNS, secondary tumors of the brain, or malignant gliomas, because all of these may be hypermetabolic.
FIGURE 63.12. Brain Perfusion Planar Scans for Brain Death. A. Planar scan demonstrates uptake in brain parenchyma and therefore does not meet Society of Nuclear Medicine criteria for the corroboration of the clinical impression of brain death. B . Planar scan in another patient shows absence of uptake in brain and therefore does meet criteria for the corroboration of the clinical impression of brain death by total absence of cerebral blood
Brain
flow.
Trauma
3848
SPECT brain imaging has been proposed to confirm the presence of a focal or diffuse injury in patients with persistent symptoms after trauma but normal or nondiagnostic anatomic imaging studies. The increased sensitivity of functional SPECT relative to CT or MR favors this use and is supported by the procedure guidelines for SPECT using brain perfusion radiopharmaceuticals (http://www.snm.org) .
Brain
Death
Criteria for Tc-99m brain perfusion radiopharmaceutical planar scanning have been retrospectively validated. Brain perfusion agents have advantages over conventional agents such as Tc-99m-GH or Tc99m-DTPA, are less dependent on the quality of bolus injection, are easier to interpret, and allow evaluation of posterior fossa blood flow. Radionuclide cerebral angiography without brain perfusion radiopharmaceuticals requires rapid acquisition of dynamic images in technically challenging situations, cannot image flow in posterior fossae, and may result in difficult or equivocal interpretation. Radionuclide scintigraphy is not affected by drug intoxication, hypothermia, or hypovolemia. In the presence of brain death, the radioactive bolus stops at the base of the skull because of increased intracranial pressure. It is important to have a good bolus injection, and if distinct activity is not identified in the common carotid artery, the injection should be repeated. Absence of intracerebral arterial flow and no visualization of major venous sinuses on subsequent static images fit the nuclear medicine criteria for brain death (Fig. 63.12). The “hot nose sign,― caused by increased collateral blood flow in the nasal area, could be a secondary sign in brain death.
SUGGESTED
READINGS
Alexandrov AV, Black SE, Ehrlich LE, et al. Simple visual analysis of brain perfusion on HMPAO SPECT predicts early outcome in acute stroke. Stroke 1996;27:1537–1542.
3849
Assessment of brain SPECT: report of the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology. Neurology 1996;46:278–285. Bonte FJ, Harris TS, Roney CA, Hynan LS. Differential diagnosis between Alzheimer's and frontotemporal disease by posterior cingulate sign. J Nucl Med 2004;45(5):771–774. Conti PS. Introduction to imaging brain tumor metabolism with positron emission tomography (PET). Cancer Invest 1995;13:244–259. Chiro D, Ommaya AK, et al. Isotope cisternography in the diagnosis and follow-up of cerebrospinal fluid rhinorrhea. J Neurosurg
1968;28:522–529.
Drzezga A, Arnold S, Minoshima S, et al. F-18 FDG PET studies in patients with extratemporal and temporal epilepsy: evaluation of an observer-independent analysis. J Nucl Med 1999;40(5):737–746. Harbert JC. Radionuclide cisternography. Semin Nucl Med 1971;1:90–106. Hartshorne ME. Positron emission tomography. In Orrison WW, Lewine JD, Sanders JA, Hartshorne MF, eds. Functional Brain Imaging. St. Louis, MO: Mosby, 1995:187–212. Hustinx R, Pourdehnad M, Kaschten B, Alavi A. PET imaging for differentiating recurrent brain tumor from radiation necrosis. Radiol Clin North Am 2005;43(1):35–47. Idea RJ, Lewis DH. Timely diagnosis of brain death in an emergency trauma center. AJR Am J Roentgenol
3850
1994;163:927–928. Launes J, Nikkinen P, Lindroth L, et al. Diagnosis of acute herpes simplex encephalitis by brain perfusion single photon emission computed tomography. Lancet 1988;1(8596):1188–1191. Lewis DH, Hsu S, Eskridge J, et al. Brain SPECT and transcranial Doppler ultrasound in vasospasm-induced delayed cerebral ischemia after subarachnoid hemorrhage. J Stroke Cerebrovasc Dis 1992;2:12–21. Messa C, Fazio F, Costa DC, Ell PJ. Clinical brain radionuclide imaging studies. Semin Nucl Med 1995;15:111–143. Minoshima S, Foster N, Sima AA, Frey KA, Albin RL, Kuhl DE. Alzheimer's disease versus dementia with Lewy bodies: cerebral metabolic distinction with autopsy confirmation. Ann Neurol 2001;50(3):358–365. Moretti JL, Caglar M, Weinmann P. Cerebral perfusion imaging tracers for SPECT: which one to choose? J Nucl Med 1995;36:359–363. Mrhac L, Zakko S, Parikh Y. Brain death: the evaluation of semiquantitative parameters and other signs in HMPAO scintigraphy. Nucl Med Commun 1995;16:1016–1020. O’Connell RA. Psychiatric disorders. In Van Heertum RL, Tikofsky RS, eds. Cerebral SPECT Imaging. 2nd ed. New York: Raven Press, 1995. Silverman DH. Brain 18F-FDG PET in the diagnosis of neurodegenerative dementias: comparison with perfusion SPECT and with clinical evaluations lacking nuclear imaging. J Nucl Med
3851
2004;45(4):594–607. Silverman DH, Alavi A. PET imaging in the assessment of normal and impaired cognitive function. Radiol Clin North Am 2005;43(1):67–77. Silverman DH, Small GW, Chang CY, et al. Positron emission tomography in evaluation of dementia. JAMA 2001;86(17):2120–2127. Society of Nuclear Medicine Brain Imaging Council. Ethical clinical practice of functional brain imaging. J Nucl Med 1996;37:1256–1259. Vernet O, Farmer JP, Lambert R, Montes JL. Radionuclide shuntogram: adjunct to manage hydrocephalic patients. J Nucl Med 1996;37:406–410.
3852
Editors: Brant, William E.; Helms, Clyde A. Title: Fundamentals of Diagnostic Radiology, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins > Table of Contents > Section XII - Nuclear Radiology > Chapter 64 - Positron Emission Tomography
Chapter
64
Positron
Emission
Tomography
Bijan Bijan David K. Shelton William E. Brant PET imaging is the basis of molecular imaging in today's clinical practice. The new hybrid imaging instrument, PET-CT (PET plus CT), combines anatomic and physiologic imaging and opens a door to a new era in radiology and nuclear medicine. Currently the clinical applications of PET imaging are focused in four main areas: (1) oncology, comprising about 80% of the current practice of clinical PET; (2) neurologic applications (Alzheimer disease, epilepsy); (3) cardiac applications (coronary artery disease, myocardial viability); and (4) infection and inflammation imaging (fever of unknown origin, immunocompromised patients). With the expanding approval of reimbursement of PET for more indications, the number of PET scanners has increased dramatically throughout the United States.
PET
Instrumentation
PET imaging is based upon positron emitters, which are used as labelling tracers for metabolic molecules. Positron-emitting radionuclides include fluorine-18 (F-18), nitrogen-13 (N-13), oxygen15 (O-15), carbon-11 (C-11), and rubidium-82 (Rb-82). Current clinical PET imaging is based on F-18, an unstable radioisotope with a half-life of 109 minutes that is produced in a cyclotron. Its relatively
3853
short half-life requires that imaging be performed within relatively short transit proximity to a cyclotron. Other positron emitters have even shorter half-lives (75 seconds to 20 minutes) and must be imaged at the cyclotron site within minutes of their production. F-18 decay releases a positron, which has the same mass as an electron but is positively charged. Within milliseconds of emission, the positron annihilates with a nearby electron to release two highenergy (511 keV) γ-ray photons. These photons move apart in opposite directions (a near 180° angle). Because of their high energy, these photons are highly penetrative in soft tissue and therefore leave the body with limited absorption or deflection. The PET imaging system consists of a ring of scintillation detectors set to detect coincident photons that strike the detectors within a very narrow time window. Non-coincident, mostly scattered, photons are rejected from the data set. Simultaneous detection of two 511keV photons by any two detectors indicates that an annihilation event has occurred somewhere in the column of space between the two detectors. These raw data projections are reconstructed into cross-sectional images by algorithms similar to those used in CT and MR. PET scans viewed alone provide limited morphologic detail and can be difficult to interpret. Spatial resolution of current PET systems is 4 to 5 mm. Systems under development improve spatial resolution to 2 mm. Recently developed hybrid instruments combine PET scanners with CT scanners. CT provides excellent anatomic detail but lacks functional information. PET provides metabolic measurements and functional information but lacks precise morphology. The union of the two allows correlation of complementary findings into a single comprehensive examination. A typical PET-CT scanner consists of a PET scanner immediately adjacent to a multidetector CT scanner. A patient couch, accurately calibrated for position, runs through both scanning assemblies. Scans may be obtained from either instrument independently or from both simultaneously. Computer software is used to fuse the two sets of images into composite images. P.1488
3854
Attenuation correction must be applied to PET images. Attenuation of either of the paired photons by absorption or scatter would result in rejection of data and nondetection of annihilation events. Attenuation is increased when the origin of the photons is from deeper within the body because of the greater thickness of intervening soft tissues. Correspondingly, attenuation is less in the thorax because of the airfilled lungs than in the abdomen or pelvis. The CT transmission data is used to create an attenuation map that is applied to the PET images to compensate for attenuation defects. The attenuation correction process increases sensitivity for detection of positron activity, but it may introduce artifacts into the images. Therefore PET interpretation includes viewing of both attenuation-corrected and attenuation-uncorrected PET image. The F-18 radioisotope is the radiotracer in the molecule 2-[flourine18]fluoro-2 deoxy-D-glucose (FDG). FDG is an analog of glucose and tends to concentrate in areas of high metabolic activity everywhere in the human body. Tumor cells with a high rate of mitosis are highly metabolically active and have an increased number of glucose transporters, thus concentrating FDG in tumor cells in higher concentrations than in normal tissue. In addition, FDG becomes trapped within cells because, unlike glucose, it is not metabolized in most tissues. Performing PET-CT. Patients fast for 4 to 6 hours prior to the scan to limit metabolic activity with the GI tract. Blood glucose should be under good control (90%) when hyperactivity is present in regional nodes. PET can be employed for assessment of the response to radiation
3881
therapy as well as for restaging after radical treatment. Recurrences occur in the surgical bed (one third of patients), in regional nodes, or distally in liver, lung, or bone. PET is sensitive (~100%) but not specific (57%) for surgical bed recurrence. Distant disease is effectively demonstrated (95% sensitive, 80% specific). Inclusion of PET in the management of esophageal cancer has a P.1500 significant impact in deciding the best route of treatment and prevents unnecessary surgery by accurately upstaging the disease process.
Stomach Stomach
mucosa
Cancer concentrates
FDG
physiologically.
Therefore,
detection of a hypermetabolic lesion in a highly active background is of limited accuracy, especially if the lesion is small. PET should not be used as the primary mode of diagnosis. Involvement of regional nodes along the lesser curvature can be overlooked on PET imaging because of the marked nearby gastric activity. Regional disease, including celiac nodal involvement, and distant metastases are assessed with better accuracy. PET images should be correlated closely with corresponding CT images. Prominent nodes on CT need to be assessed on corresponding PET images. Peritoneal surfaces, the greater omentum, and the cul-de-sac need to be assessed meticulously in association with CT images to detect subtle hypermetabolic foci. PET detection of hepatic metastases is limited; therefore contrast-enhanced CT or MR should be added to the staging evaluation. Pancreatic cancer staging is best accomplished by other crosssectional imaging modalities, as PET lacks the resolution to assess for subtle features such as vascular invasion. However, distant metastases can be accurately detected using PET imaging if radical treatment of a small, well-localized lesion is considered. Gastric activity can overshadow the pancreatic bed. Intake of water can slightly improve the assessment of this region. PET can assess the response to treatment reliably. Persistent high activity on
3882
postradiation cases suggests very poor prognosis. PET is useful in the differentiation of chronic pancreatitis from pancreatic cancer.
Colorectal
Cancer
Most (95%) colon cancers are FDG-avid adenocarcinomas. Approximately 20% of patients have metastatic disease at the time of diagnosis. PET provides one-stop accurate assessment for wholebody staging at the time of initial diagnosis (Fig. 64.16). In about 40% of cases, clinical staging is modified by PET results. Accurate up-staging by PET prevents unnecessary surgery in patients with advanced,
albeit
clinically P.1501
occult, disease. Currently PET is widely utilized for staging and restaging of patients and for assessment of response to treatment. PET is currently not widely employed for early detection or colon cancer screening. Some groups are using hybrid PET-CT with virtual colonoscopy to offer functional and anatomic imaging for early detection of adenomatous polyps and small colon cancers. Nonspecific colonic activity can be a source of overcalling (Fig. 64.6) . Diffuse activity is usually physiologic, especially in the descending and sigmoid colon. Benign conditions, including sigmoid diverticulitis, inflammatory polyps, and fecal material, can accumulate FDG. Any focal activity in the colon should be further evaluated. The rectal region may be obscured by bladder activity. Therefore the bladder should be emptied immediately prior to imaging, and scanning should begin in the pelvis and proceed superiorly.
3883
FIGURE cancer
64.16. (Color Plates) Colon Cancer Staging. A colon (between cursors) in the splenic flexure is hypermetabolic
on PET-CT as is metastatic adenopathy in the porta hepatis (arrowhead). A . CT. B . Corrected PET. C . Fused PET-CT. D . Uncorrected
PET.
PET has low sensitivity (29%) for regional node metastases because involved nodes are often small, have a limited number of tumor cells, and are located close to and masked by bowel activity. When pericolic nodes are FDG positive, specificity for malignancy is high (96%). PET sensitivity is equal to or slightly better than CT alone in detection of hepatic metastases, and PET-CT specificity for hepatic metastases approaches 100%. Small lesions ( Back of Book > Resources
Resources
Color
FIGURE image. )
Plate
7.26B. Alzheimer
Disease
3930
(AD). (See black and white
FIGURE 9.1. Positron Emission black and white image. )
Tomography
(PET). (See
FIGURE 9.7. Non-Hodgkin Lymphoma of the Masticator Space. (See black and white image. )
3931
FIGURE
40.7. Normal Flow Reversal at Bifurcation. (See
black and white image. )
3932
FIGURE image. )
40.9. Color Doppler Image. (See black and white
FIGURE
40.10. Power Doppler (Color Doppler Energy
[CDE])
Image. (See black and white image. )
3933
FIGURE 40.11. Aliasing on Spectral Doppler. (See black and white image. )
3934
FIGURE 40.12. Aliasing on Color Doppler Imaging. (See black and white image. )
FIGURE image. )
40.16. External
Carotid
3935
Artery. (See black and white
FIGURE image. )
40.18. Ulcerated
Plaque. (See black and white
FIGURE 40.19. Internal Carotid Artery (ICA) Stenosis. (See black and white image. )
3936
FIGURE 40.20B, C. Common Carotid Artery (CCA) Stenosis. (See black and white image. )
3937
FIGURE 40.21. Common Carotid Artery (CCA) Origin Stenosis. (See black and white image. )
3938
FIGURE
40.22. Common Carotid Artery (CCA) Occlusion.
(See black and white image. )
3939
FIGURE 40.23B–D. Internal Carotid Artery (ICA) Occlusion. (See black and white image. )
3940
FIGURE 40.24. Internal Carotid Artery (ICA) Occlusion. (See black and white image. )
FIGURE
40.25. String
Sign. (See black and white image. )
3941
FIGURE
40.27. Tortuous
Vessel. (See black and white image. )
FIGURE image. )
40.28. Calcified
Plaques. (See black and white
3942
FIGURE image. )
40.32C. Aortic
Dissection. (See black and white
FIGURE
40.33E. Inferior Vena Cava (IVC) Tumor Thrombus.
(See black and white image. )
3943
FIGURE 40.34E. Normal Transjugular Intrahepatic Shunt (TIPS) US. (See black and white image. )
FIGURE 40.32. Transjugular Intrahepatic Portal (TIPS) Occlusion. (See black and white image. )
3944
Portal
Shunt
FIGURE
40.38B–D. Pseudoaneurysm. (See black and white
image. )
3945
FIGURE image. )
40.39. Pseudoaneurysm. (See black and white
3946
FIGURE 40.40B–D. Arteriovenous Fistula in the Left Wrist. (See black and white image. )
FIGURE 40.41. Synthetic black and white image. )
Graft
Anastomotic
3947
Stenosis. (See
FIGURE image. )
40.42. Graft
Anastomosis. (See black and white
FIGURE 40.43A, B, D. Reversed Saphenous Vein Graft. (See black and white image. )
3948
FIGURE
40.44. Popliteal to Plantar Artery Vein Graft
Stenosis. (See black and white image. )
3949
FIGURE white
40.47C. Deep
Venous
Thrombosis. (See black and
image. )
FIGURE image. )
40.49. Venous
Insufficiency. (See black and white
3950
FIGURE 40.51. Normal Upper Extremity Venous US. (See black and white image. )
FIGURE 40.48D. Chronic Deep Venous Thrombosis. (See black and white image. )
3951
FIGURE
40.52. Internal
Jugular
and white image. )
3952
Vein
Thrombosis. (See black
FIGURE 40.53. Subclavian white image. )
FIGURE
Vein
Thrombosis. (See black and
52.19D. Intussusception. (See black and white
image. )
3953
FIGURE 52.54B, C. Testicular
Torsion. (See black and white
image. )
3954
FIGURE
52.64B. Vascular Neoplasms of the Liver. (See black
and white image. )
FIGURE
52.40A. Renal
Abscess. (See black and white image. )
3955
FIGURE 62.11. SPECT-CT of Metastatic Prostate Carcinoma With ProstaScint. (See black and white image. )
FIGURE
63.2. Ethyl Cysteinate Dimer (ECD) Versus
Hexamethylpropyleneamine Oxime (HMPAO) Perfusion. (See black and white image. )
3956
in
Luxury
FIGURE 63.3. Normal Hexamethylpropyleneamine (HMPAO) Study. (See black and white image. )
3957
Oxime
FIGURE image. )
63.4. Cerebral
Infarction. (See black and white
3958
FIGURE 63.5. Subcortical white image. )
Cerebral
3959
Infarct. (See black and
FIGURE 63.7. Acetazolamide Vascular Reserve Testing in Occlusive Carotid Disease. (See black and white image. )
3960
FIGURE
63.6A. Subarachnoid
Hemorrhage
Vasospasm. (See black and white image. )
3961
and
Cerebral
FIGURE 63.8. Positive and white image. )
FIGURE
Balloon
Occlusion
Study. (See black
63.10. Three-dimensional SSP Display of SPECT in
Alzheimer
Disease. (See black and white image. )
3962
FIGURE Partial
63.11. Ictal and Interictal SPECT Scans in Complex Epilepsy. (See black and white image. )
3963
FIGURE
64.1C. Breast Cancer PET-CT. (See black and white
image. )
3964
FIGURE 64.2. Physiologic Fluorodeoxyglucose (FDG) Activity in Muscle. (See black and white image. )
3965
FIGURE 64.3. Physiologic Fluorodeoxyglucose (FDG) Activity in the Brain. (See black and white image. )
FIGURE 64.4. Physiologic Fluorodeoxyglucose (FDG) Activity in the Heart. (See black and white image. )
3966
FIGURE 64.6C. Massive black and white image. )
Physiologic
3967
Colon
Activity. (See
FIGURE 64.7. Normal Physiologic Activity in the Urinary Tract. (See black and white image. )
3968
FIGURE 64.9. Warm Pulmonary Nodule: Carcinoma. (See black and white image. )
FIGURE 64.8C. Hot Pulmonary Nodule: Carcinoma. (See black and white image. )
3969
Bronchoalveolar
Bronchogenic
FIGURE 64.10. Multiple white image. )
Pulmonary
3970
Nodules. (See black and
FIGURE
64.11. Lung Cancer Staging. (See black and white
image. )
3971
FIGURE 64.12. Postradiation white image. )
FIGURE
64.13. Lymphoma
Pneumonitis. (See black and
Staging. (See black and white
image. )
3972
FIGURE image. )
64.14C. Spleen
Lymphoma. (See black and white
FIGURE
64.15. Lymphoma: Early Response to Therapy. (See
black and white image. )
3973
FIGURE
64.16C. Colon Cancer Staging. (See black and white
image. )
3974
FIGURE 64.17C. Colon Cancer: black and white image. )
FIGURE 64.18. Hepatic and white image. )
Anastomosis
Metastasis:
3975
Colon
Activity. (See
Cancer. (See black
FIGURE white
64.19. Necrotic
Hepatic
image. )
3976
Metastasis. (See black and
FIGURE image. )
64.20. Chronic
Cholecystitis. (See black and white
3977
FIGURE 64.21. Breast Cancer: Axillary Node Metastasis. (See black and white image. )
3978
FIGURE 64.22. Breast Cancer: Internal Mammary Node Metastases. (See black and white image. )
FIGURE 64.23. Uterine white image. )
Cancer
Recurrence. (See black and
3979
FIGURE image. )
64.24. Ovarian Cancer, Stage 1. (See black and white
FIGURE 64.25. Transitional Cell Carcinoma of the Bladder. (See black and white image. )
3980
FIGURE image. )
64.26. Adrenal
Metastasis. (See black and white
3981
FIGURE 64.27. Benign white image. )
Adrenal
Adenoma. (See black and
FIGURE 64.28. Squamous Cell Carcinoma of the Tongue. (See black and white image. )
3982
FIGURE 64.29. Metastatic Squamous Cell Carcinoma of the Nasopharynx. (See black and white image. )
3983
FIGURE
64.30. Metastasis to Spine. (See black and white
image. )
3984
FIGURE 64.31. Facet Hypertrophy, black and white image. )
3985
Not
Metastasis. (See
FIGURE 64.32. “Hot― and “Cold― Metastases. (See black and white image. )
3986
Osseous
FIGURE
64.33. Benign
Hypermetabolic
black and white image. )
3987
Bone
Marrow. (See
FIGURE
64.34. Diffuse Metastatic Disease With
Hypermetabolic
Bone
Marrow. (See black and white image. )
3988
FIGURE
64.35B. Benign Intra-axial Brain Cyst. (See black
and white image. )
3989
FIGURE
64.36B. Absence of Residual Brain Tumor. (See
black and white image. )
3990
FIGURE image. )
64.37. Brain
Metastases. (See black and white
FIGURE image. )
64.38. Alzheimer
Dementia. (See black and white
3991
FIGURE image. )
64.40. Abdominal
Abscess. (See black and white
3992
FIGURE image. )
64.41. Shoulder
Arthritis. (See black and white
3993
FIGURE
64.42. Rib
Fracture. (See black and white image. )
3994
FIGURE
64.43. Thymic
Rebound. (See black and white image. )
3995