Diseases of the Abdomen and Pelvis 2010-2013: Diagnostic Imaging and Interventional Techniques

Diseases of the Abdomen and Pelvis 2010-2013 Diagnostic Imaging and Interventional Techniques

J. Hodler • G.K. von Schulthess • Ch.L. Zollikofer (Eds)

DISEASES OF THE ABDOMEN AND PELVIS 2010-2013 DIAGNOSTIC IMAGING AND INTERVENTIONAL TECHNIQUES 42nd International Diagnostic Course in Davos (IDKD) Davos, March 21-26, 2010 including the Nuclear Medicine Satellite Course “Diamond” Davos, March 20-21, 2010 Pediatric Satellite Course “Kangaroo” Davos, March 20-21, 2010

IDKD in Greece

presented by the Foundation for the Advancement of Education in Medical Radiology, Zurich

Editors J. HODLER Radiology, University Hospital, Zurich, Switzerland

G. K. VON SCHULTHESS Nuclear Medicine, University Hospital, Zurich, Switzerland

CH. L. ZOLLIKOFER Kilchberg/Zurich, Switzerland

ISBN 978-88-470-1636-1

e-ISBN 978-88-470-1637-8

DOI 10.1007/978-88-470-1637-8 Springer Dordrecht Heidelberg London Milan New York Library of Congress Control Number: 2010922809 © Springer Verlag Italia 2010 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in other ways, and storage in data banks. Duplication of this publication or parts thereof is only permitted under the provisions of the Italian Copyright Law in its current version, and permission for use must always be obtained from Springer. Violations are liable for prosecution under the Italian Copyright Law. The use of general descriptive names, registered names, trademarks, etc., in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Product liability: The publisher cannot guarantee the accuracy of any information about dosage and application contained in this book. In every individual case the user must check such information by consulting the relevant literature. Cover design: Simona Colombo, Milan, Italy Typesetting: C & G di Cerri e Galassi, Cremona, Italy Printing and binding: Grafiche Porpora, Segrate (MI), Italy Printed in Italy Springer-Verlag Italia S.r.l., Via Decembrio 28, 20137 Milan Springer is a part of Springer Science+Business Media (www.springer.com)

IDKD 2010-2013

Preface

The International Diagnostic Course in Davos (IDKD) offers a unique learning experience for imaging specialists in training as well as for experienced radiologists and clinicians wishing to be updated on the current state of the art and the latest developments in the fields of imaging and image-guided interventions. This annual course is focused on organ systems and diseases rather than on modalities. This year’s program deals with diseases of the abdomen and pelvis. During the course, the topics are discussed in group workshops and in plenary sessions with lectures by world-renowned experts and teachers. While the workshops present state-of-the-art summaries, the lectures are oriented towards future developments. Accordingly, this Syllabus represents a condensed version of the contents presented under the 20 topics dealing with imaging and interventional therapies in abdominal and pelvic diseases. The topics encompass all the relevant imaging modalities including conventional X-rays, computed tomography, nuclear medicine, ultrasound and magnetic resonance angiography, as well as image-guided interventional techniques. The Syllabus is designed to be an “aide-mémoire” for the course participants so that they can fully concentrate on the lecture and participate in the discussions without the need of taking notes. Additional information can be found on the IDKD website: www.idkd.org J. Hodler G.K. von Schulthess Ch.L. Zollikofer

IDKD 2010-2013

Table of Contents

Workshops Emergency Radiology of the Abdomen: The Acute Abdomen . . . . . . . . . . . . . . . . . . . . . . .

3

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

Trauma of the Abdomen and Pelvis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14

Philip J. Kenney, Stuart E. Mirvis

Diseases of the Esophagus and Stomach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

Marc S. Levine, Ahmed Ba-Ssalamah

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis

28

Marc J. Gollub

Diseases of the Small Bowel, Including the Duodenum – MRI . . . . . . . . . . . . . . . . . . . . . .

32

Karin A. Herrmann

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases . . . . . . . . . .

37

Jaap Stoker, Richard M. Gore

CT Colonography: Updated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

Daniel C. Johnson, Michael Macari

Imaging of Diffuse and Inflammatory Liver Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

Pablo R. Ros, Rendon C. Nelson

Focal Liver Lesions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

63

Wolfgang Schima, Richard Baron

Imaging Diseases of the Gallbladder and Bile Ducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75

Angela D. Levy, Celso Matos

Diseases of the Pancreas, I: Pancreatitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

Thomas Helmberger

Diseases of the Pancreas, II: Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

89

Ruedi F. Thoeni

Adrenal Imaging and Intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

96

William W. Mayo-Smith, Isaac R. Francis

Renal Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

99

Richard H. Cohan, Ronald J. Zagoria

Urinary Tract Obstruction and Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Parvati Ramchandani, Julia R. Fielding

VIII

Benign Diseases of the Female Genital Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Caroline Reinhold, Rahel A. Kubik-Huch

Malignant Diseases of the Female Genital Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Evis Sala, Susan Ascher

Magnetic Resonance Imaging of Prostate Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

Imaging of the Male Pelvis: The Scrotum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Brent J. Wagner

Spread of Metastatic Disease in the Abdomen and Pelvis . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 James A. Brink, Ali Shirkhoda

Abdominal Vascular Disease: Diagnosis and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Johannes Lammer

Non-vascular Abdominal Disease: Diagnosis and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

An Approach to Imaging the Acute Abdomen in the Pediatric Population . . . . . . . . . . . 167 Alan Daneman, Simon G. Robben

Imaging Uronephropathies in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 Jeanne S. Chow, Fred E. Avni

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging . . . . . . . . . . . . . . . . . . . . 183 Gerald Antoch

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging . . . . . . . . . . . . . . . . . 190 Thomas F. Hany

Nuclear Medicine Satellite Course “Diamond” Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals . . . . . 199 Angelika Bischof Delaloye

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 Ariane Boubaker

PET in Hepatobiliary-Pancreatic Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri

PET in Tumors of the Digestive Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 Thomas F. Hany

Tumors of the Adrenergic System: Imaging and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 Cornelis A. Hoefnagel

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy . . . . . . . . . . . . . . . . . . . . 231 Dik J. Kwekkeboom

Table of Contents

IX

Table of Contents

Pediatric Satellite Course “Kangaroo” Imaging Cystic Kidneys in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 Fred E. Avni

Understanding Duplication Anomalies of the Kidney . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Jeanne S. Chow

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management . . . . . 247 Alan Daneman

Pediatric Intestinal Ultrasonography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 Simon G. Robben

IDKD 2010-2013

List of Contributors

Allegri V., 215 Antoch G., 183 Ascher S., 119 Avni F.E., 174, 239 Ba-Ssalamah A., 22 Barentsz J.O., 125 Baron R., 63 Bartolozzi C., 162 Battaglia V., 162 Bischof Delaloye A., 199 Boubaker A., 205 Bozzi E., 162 Brink J.A., 146 Bruel J.-M., 3 Chow J.S., 174, 243 Cohan R.H., 99 Daneman A., 167, 247 Fanti S., 215 Fielding J.R., 104 Francis I.R., 96 Futterer J.J., 125 Gollub M.J., 28 Gore R.M., 37 Hany T.F., 190, 219 Heiken J.P., 3 Heijmink S.W.T.P.J., 125 Helmberger T., 81 Herrmann K.A., 32

Hoefnagel C.A., 226 Hoeks C., 125 Hulsbergen-van der Kaa C., 125 Johnson D.C., 48 Kenney P.J., 14 Kubik-Huch R.A., 110 Kwekkeboom D.J., 231 Lammer J., 154 Levine M.S., 22 Levy A.D., 75 Macari M., 48 Maffione A.M., 215 Marincek B., 3 Matos C., 75 Mayo-Smith W.W., 96 Mirvis S.E., 14 Nelson R.C., 50 Ramchandani P., 104 Reinhold C., 110 Robben S.G., 167, 252 Ros P.R., 50 Sala E., 119 Schima W., 63 Shirkhoda A., 146 Stoker J., 37 Thoeni R.F., 89 Wagner B.J., 142 Zagoria R.J., 99

WORKSHOPS

IDKD 2010-2013

Emergency Radiology of the Abdomen: The Acute Abdomen Jean-Michel Bruel1, Borut Marincek2, Jay P. Heiken3 1 Medical

Imaging Department, Hôpital Saint-Eloi, CHRU de Montpellier, Montpellier, France of Diagnostic Radiology, University Hospital, Zurich, Switzerland 3 Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO, USA 2 Institute

Introduction The term “acute abdomen” defines a clinical syndrome characterized by a history of hitherto undiagnosed abdominal pain lasting less than one week. A large range of disorders, from benign, self-limited diseases to conditions that require immediate surgery, can cause acute abdominal pain. Eight conditions account for over 90% of patients who are referred to the hospital and who are seen on surgical wards with acute abdominal pain: acute appendicitis, acute cholecystitis, small bowel obstruction (SBO), urinary colic, perforated peptic ulcer, acute pancreatitis, acute diverticular disease, and non-specific, non-surgical abdominal pain (dyspepsia, constipation).

Imaging Techniques Clinical assessment of the acute abdomen is often difficult because the findings of the physical examination and laboratory investigations are often non-specific. Traditionally, plain abdominal radiographs have served as the initial imaging approach; however, because of their diagnostic limitations, plain radiographs now play only a limited role in this clinical setting. Currently, the major indications for plain radiography are to determine the presence of bowel obstruction, perforated viscus, urinary tract calculi, or a foreign body. The conventional radiographic examination consists of supine and either upright or left lateral decubitus images. Computed tomography (CT) now serves as the imaging test of choice for most adult patients with acute abdominal pain. It has been shown to be superior to plain radiography for diagnosing nearly all causes of acute abdominal pain. Several studies have demonstrated that the use of CT to evaluate patients with acute abdominal pain increases the accuracy of clinical diagnosis by >20% and results in management changes in up to 60% of patients. The major obstacles to replacing plain abdominal radiography with CT are its higher cost, more limited availability, and higher radiation dose. The use of CT in patients with an acute abdomen requires careful attention to the CT protocol.

1. The multidetector CT (MDCT) image data volume should be obtained from the dome of the diaphragm to the inferior aspect of the pubic symphysis (with thin collimation and a short acquisition time). CT data are reconstructed with thin, overlapped slices for multiplanar (coronal and sagittal planes) reformation and 3- to 5-mm thick contiguous axial slices. The image series is sent to a dedicated workstation and/or PACS. 2. Vascular enhancement by iodinated contrast medium (CM) is mandatory in most cases, unless contraindicated. Special attention should be paid to older patients and those with metabolic disorders (dehydration in SBO) in assessing the renal impact of CM administration. In general, the most helpful scanning phase is the late portal phase (70 s), but other scanning phases are useful in selected circumstances: arterial phase in bowel ischemia, bleeding, or visceral infarction; delayed phase (3 min) for assessing the lack of enhancement in a patient with suspected acute mesenteric ischemia. A pre-contrast CT scan allows demonstration of calcifications, lithiasis, and acute or subacute hemorrhage; multiphasic scanning should be used only for specific indications in order to limit radiation dose. 3. Changes in the CT protocol should be decided according to the clinical conditions and/or the preliminary results of the CT examination. In selected cases, colorectal opacification and/or image acquisition with the patient in the prone position may be helpful to clarify equivocal findings. In most cases, the systematic use of ingested oral contrast is not recommended. 4. The method of image evaluation is critical to optimize interpretation. Additional window level (M) and width (W) settings are useful to identify tiny bubbles of extraluminal gas (CT lung windows) or hyperattenuation from recent hemorrhage (narrow CT windows). The systematic use of multiplanar reformation (MPR), particularly in the coronal plane, is recommended, and 3D imaging may be helpful. Ultrasonography (US) is the initial imaging technique of choice for patients with suspected acute cholecystitis or acute gynecological abnormalities. It also is the primary

4

method for evaluating pregnant women and pediatric patients. Although less sensitive and specific than CT, US is an excellent imaging test for diagnosing acute appendicitis, when employed by experienced individuals. It can also be used to evaluate the presence or absence of the layered structure of the digestive tract wall or to assess the structure of a lesion identified at CT. Until recently, magnetic resonance imaging (MRI) has played a very limited role in patients with acute abdominal pain; however, it is now established in the imaging of pregnant women with abdominal pain who have had a negative or equivocal US examination. Recent studies assessing the use of MRI to evaluate all patients with acute lower abdominal pain have shown promising results. MRI may also have a role in patients with biliary diseases and/or pancreatitis. The differential diagnosis in a patient with an acute abdomen is influenced greatly by the nature and location of the pain. Therefore, the imaging strategies for acute pain localized to an abdominal quadrant should be discussed separately from those for acute pain that is diffuse or localized to the flank or epigastric region.

Acute Pain in an Abdominal Quadrant In many cases, acute abdominal pain can be localized to one either the right upper, left upper, right lower, or left lower abdominal quadrant.

Right Upper Quadrant Acute cholecystitis is by far the most common disease to involve the right upper quadrant. Other important diseases that can have a clinical presentation similar to that of acute cholecystitis are pyogenic or amebic liver abscess, spontaneous rupture of a hepatic neoplasm (usually hepatocellular adenoma or carcinoma), hepatitis, and myocardial infarction. The preferred imaging method for evaluating patients with acute right upper abdominal pain is US. It is a reliable technique for establishing the diagnosis of acute calculous cholecystitis. The imaging criteria include the detection of gallstones, the sonographic Murphy sign, gallbladder wall thickening ≥3 mm, and pericholecystic fluid. The association of three of these signs is highly suggestive of acute cholecystitis. Isolated gallbladder wall thickening may be secondary to other conditions, such as gallbladder adenomyomatosis, gallbladder carcinoma, HIV cholangitis, sclerosing cholangitis, acute hepatitis, cirrhosis, ascites, portal hypertension, hypoproteinemia, pancreatitis, and cardiac failure. In acute calculous cholecystitis, typically a calculus obstructs the cystic duct. The trapped concentrated bile irritates the gallbladder wall, causing increased secretion, which in turn leads to distention and edema of the wall. The rising intraluminal pressure compresses the vessels, resulting in thrombosis, ischemia, and subsequent necrosis and perforation of the

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

wall. Gallbladder perforation and complicating pericholecystic abscess typically occur adjacent to the gallbladder fundus because of the sparse blood supply. CT may be useful for confirmation of the sonographic diagnosis, but usually is not necessary. Emphysematous cholecystitis is a rare complication of acute cholecystitis that generally is associated with diabetes mellitus. US or CT demonstrates gas in the wall and/or lumen of the gallbladder, which implies underlying gangrenous changes (Fig. 1). Acalculous acute cholecystitis accounts for only approximately 5% of cases of acute cholecystitis but is especially common in patients in the intensive care unit. Prolonged bile stasis results in increased viscosity of the bile that ultimately leads to functional cystic duct obstruction. Both US and CT are accurate techniques for diagnosing liver abscesses. US usually demonstrates a round or oval hypoechoic mass with low-level internal echoes. Although the lesion may mimic a solid hepatic mass, the presence of through transmission is a clue to its cystic nature. Pyogenic liver abscesses most commonly are the result of seeding from appendicitis or diverticulitis or direct extension from cholecystitis or cholangitis. Amebic abscesses result from primary colonic involvement, with seeding through the portal vein. In most cases, the US appearances of pyogenic and amebic abscess are indistinguishable. The CT appearances of pyogenic and amebic abscesses also overlap substantially. Amebic abscesses are cystic masses of low attenuation. An enhancing wall and a peripheral zone of edema surrounding the abscess are common but not universally present. Extrahepatic extension of the amebic abscess with involvement of the chest wall, pleura, or adjacent viscera is a frequent finding. Whereas amebic abscesses usually are solitary and unilocular, pyogenic abscesses may be multiple or multi-

Fig. 1. Diabetic patient with emphysematous cholecystitis and gangrene of the gallbladder. CT shows air-fluid level in the gallbladder lumen and air in the gallbladder wall (arrows)

Emergency Radiology of the Abdomen: The Acute Abdomen

loculated and may demonstrate an irregular contour. Some pyogenic abscesses have a mixed cystic and solid appearance on US, CT, or MRI; rarely, they appear completely solid. A small percentage of hepatic abscesses, particularly those secondary to Klebsiella infection, are associated with portal vein thrombosis. Spontaneous rupture of a hepatocellular carcinoma with subsequent hemoperitoneum is a frequent complication in countries with a high incidence of this tumor, but is less commonly seen in Western countries. Subcapsular location and tumor necrosis have been implicated in the pathogenesis. US, and especially CT, are the most useful techniques for diagnosing a ruptured hepatocellular carcinoma, which appears as a peripheral or subcapsular mass. Transcatheter embolization of either the tumor or the bleeding hepatic artery is the treatment of choice. Spontaneous hemorrhage within a hepatocellular adenoma occurs most commonly in women taking oral contraceptives. Capsular rupture with subsequent hemoperitoneum is an uncommon complication. On CT, highdensity intraperitoneal fluid confirms the diagnosis of hemoperitoneum. Extravasation of CM, when present, is indicative of active bleeding.

Left Upper Quadrant Although infrequent, acute left upper quadrant pain is most often seen in splenic infarction, splenic abscess, gastritis, and gastric or duodenal ulcer. US is most frequently used for screening, while CT enables accurate further evaluation. The diagnosis of gastric pathology is established by endoscopy, with imaging playing a minor role. Common causes of splenic infarction include bacterial endocarditis, portal hypertension, and marked splenomegaly. Pancreatitis or tumors that extend into the splenic hilum can also result in infarction. Splenic infarction may be focal or global. Typical focal splenic infarcts appear as peripheral wedge-shaped defects, hypoechoic or isoechoic at US and hypoattenuating at CT. Most splenic abscesses are secondary to hematogenous dissemination of infection, e.g., bacterial endocarditis or tuberculosis. Intravenous drug abusers and immunocompromised individuals are predominantly affected. US and CT are sensitive, but the specificity of either one is low. On US, most abscesses appear as hypo- or anechoic, poorly defined lesions; on CT, they typically appear as rounded lesions of low attenuation and with rim enhancement. Spontaneous splenic rupture can occur in patients with hematological malignancy or secondary to rapid splenic enlargement from viral infections such as mononucleosis.

Right Lower Quadrant Acute appendicitis is not only the most frequent cause of acute right lower quadrant pain, it is also the most commonly encountered cause of an acute abdomen. Other

5

diseases that can present with acute right lower quadrant pain include acute terminal ileitis (Crohn’s disease), typhlitis, right-sided colonic diverticulitis and, in women, pelvic inflammatory disease, complications of ovarian cyst (hemorrhage, torsion, and leakage), endometriosis, or ectopic pregnancy. Less common causes of right lower quadrant pain include segmental infarction of the greater omentum, mesenteric adenitis, epiploic appendagitis, perforated cancer, and ileal or Meckel’s diverticulitis. The diagnosis of acute appendicitis is uncertain in up to one-third of patients. Thus, pre-operative imaging plays an important role in confirming or excluding the diagnosis. With the increasing use of medical imaging to evaluate patients with suspected acute appendicitis, the rate of both false-positive (unnecessary appendectomy) and false-negative (leading to complications from perforated appendicitis) diagnoses has decreased. The standard surgical teaching is that patients with typical clinical findings should undergo immediate appendectomy without pre-operative imaging. Nevertheless, at most medical centers pre-operative imaging is obtained even when the clinical presentation is typical. The most specific CT finding of acute appendicitis is a thickwalled appendix that contains an appendicolith (Fig. 2). The inflamed appendix often is dilated and fluid-filled. Additional helpful findings are stranding of the periappendiceal fat and thickening of the cecal apex. Findings that indicate appendiceal perforation include periappendiceal abscess, extraluminal gas, a right lower quadrant inflammatory mass, a defect in the appendiceal wall, and SBO. In the evaluation of suspected acute appendicitis in children, pregnant women, and women of reproductive age, US is an important imaging option. Demonstration of a swollen, noncompressible appendix >7 mm in diameter with a target configuration is the primary sonographic criterion (Fig. 3). Additional helpful US findings are “MacBurney’s sign” (maximum tenderness found with graded compression of the inflamed appendix) and demonstration of an appendicolith. These US signs may also be demonstrated by transvaginal highresolution US. The advantages of US include the lack of ionizing radiation, relatively low cost, and widespread availability. However, US requires considerable skill and is difficult to perform in obese patients, patients with severe pain, and patients likely to have a complicating periappendiceal abscess. When the sonographic findings are unclear, CT can provide a rapid and definitive diagnosis. Due to its exceptional accuracy, CT has emerged in many centers as the primary imaging test for patients with suspected acute appendicitis. In a small percentage of patients, diverticulitis manifests itself as a right-sided condition. Right-sided colonic diverticula are often congenital, solitary, and true diverticula, unlike sigmoid diverticula. In right-sided diverticulitis the normal appendix should be visible.

6

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

c

a

b

Fig. 2 a-c. Acute appendicitis. Axial (a, b) and sagittal (c) views of multidetector CT demonstrate a dilated appendix in retrocecal position, a calcified appendicolith at the base of the appendix (arrowheads), and inflammatory changes of the mesenteric fat (arrow)

a

b

c

Fig. 3 a-c. Ultrasonography of acute appendicitis in a 12-year-old girl. Oblique (a, b) and transverse (c) views show swollen appendix (diameter 10 mm, arrowheads) with a target configuration

Left Lower Quadrant Diverticulitis is the most common cause of acute left lower quadrant abdominal pain. The condition occurs in 10–20% of patients with diverticulosis and most commonly involves the sigmoid colon. CT is very sensitive and approaches 100% specificity and accuracy in the diagnosis or exclusion of diverticulitis; it has therefore largely replaced barium enema examinations. CT is also very useful in establishing the presence of pericolic complications. The CT diagnosis of acute diverticulitis is based on the identification of segmental colonic wall thickening and pericolic in-

flammatory changes, such as fat stranding, inflammatory mass, gas bubbles, or free fluid (Figs. 4, 5). Complications of acute diverticulitis include abscess, fistula (most commonly colovesical), SBO, peritonitis, septic thrombophlebitis, colonic obstruction, and ureteral obstruction. In patients with left lower quadrant pain, alternative diagnoses that should be considered are colitis (infectious/ inflammatory or ischemic), colonic carcinoma, epiploic appendagitis, neutropenic colitis, functional colonic disorders, and extragastrointestinal disorders (pyelonephritis or gynecological diseases). Epiploic appendagitis is a clinical condition mimicking acute colonic diverticulitis, with focal

7

Emergency Radiology of the Abdomen: The Acute Abdomen

Fig. 4. Sigmoid diverticulitis with pericolic abscesses. CT shows fine linear strands within pericolic fat, diverticula filled with air, barium, or fecal material, circumferential bowel thickening, and frank abscesses (arrows)

exquisite lower abdominal pain. It is diagnosed with CT (or US) by the demonstration of an ovoid lesion within the pericolonic fat, surrounded by inflammatory changes and abutting the colonic wall. As this disease resolves spontaneously within a few days, its correct diagnosis on CT images is important to avoid unnecessary surgery. Distinguishing sigmoid diverticulitis from carcinoma is a major differential diagnostic consideration but is often difficult on CT images despite CT findings suggestive of colon carcinoma, such as marked thickening of the colonic wall, focal thickening (length 1000 IU/L) without an intrauterine gestational sac, associated with an abnormal adnexal pattern and/or heterogeneous pelvic fluid, indicates an ectopic pregnancy.

Acute Abdomen with Diffuse Pain Any disorder that irritates a large portion of the gastrointestinal (GI) tract and/or the peritoneum can cause diffuse abdominal pain. The most common disorder is gastroenterocolitis. Other important disorders are bowel obstruction, ischemic bowel disease, and GI tract perforation.

Bowel Obstruction Bowel obstruction is a frequent cause of abdominal pain and accounts for approximately 20% of surgical admissions for acute abdominal conditions. The small bowel is involved in 60-80% of cases. Frequent causes of SBO are postoperative adhesions, hernias, and neoplasms. Mechanical large bowel obstruction is most commonly due to colorectal carcinoma, but volvulus and diverticulitis are also important causes. Colonic volvulus most commonly involves the sigmoid region, followed by the cecum. The diagnosis of bowel obstruction is established on clinical grounds and usually confirmed with plain abdominal radiographs. Due to the diagnostic limitations of plain radiography, CT is increasingly used to establish the diagnosis, identify the site, level, and cause of obstruction, and determine the presence or absence of associated bowel ischemia. CT can be useful for differentiating between simple and closed-loop obstruction. Closed-loop obstruction is a form of mechanical bowel obstruction in which two points along the course of the bowel are obstructed at a single site. It is usually secondary to an adhesive band or a hernia. Since a closed loop tends to involve the mesentery and is prone to produce a volvulus, it represents the most common cause of strangulation. However, only colonic volvulus is associated with classic features on plain abdominal radiography. CT is particularly reliable in higher grades of bowel obstruction. It has proved useful in characterizing bowel obstruction from various causes, including adhesions, hernia, neoplasm, extrinsic compression, inflammatory bowel disease, radiation enteropathy, intussusception, gallstone ileus, or volvulus. The essential CT finding of

9

Emergency Radiology of the Abdomen: The Acute Abdomen

bowel obstruction is the delineation of a transition zone between the dilated and decompressed bowel. Careful inspection of the transition zone and luminal contents usually reveals the underlying cause of obstruction. However, the presumed point of transition from dilated to nondilated bowel can be difficult to determine in the transaxial plane. MDCT facilitates this task by providing the radiologist with a volumetric data set that can be viewed in the transaxial, sagittal, or coronal plane or any combination of the three. These MPR views centered on the anticipated transition point help to determine the site, level, and cause of obstruction. Mechanical obstruction of the small bowel (SBO) has to be differentiated from paralytic ileus, large bowel obstruction (LBO), and non-obstructive massive distention of the colon. Paralytic ileus is a common problem after abdominal surgery. It may be diffuse or localized and has numerous causes, e.g., secondary to ischemic conditions, inflammatory or infectious disease, abnormal electrolyte, metabolite, drug or hormonal levels, or innervation defects. In LBO, CT demonstrates distension of the large bowel to the point of obstruction, with collapse of the distal large bowel. The distal small bowel loops may also be distended if the ileocecal valve is incompetent. Luminal obstruction by a colonic carcinoma and colonic volvulus (Fig. 7) are the main causes of LBO. Perforation of the cecum, due to gross distention resulting in ischemia of the cecal wall, is the main complication of severe LBO. A massively dilated colon may be seen in toxic megacolon and Ogilvie’s syndrome. In toxic megacolon secondary to severe colitis, CT demonstrates a thickened wall with “thumbprinting” caused by wall edema and inflammation. Ogilvie’s syndrome is an acute pseudoobstruction with dilation of the colon in the absence of a colonic transition zone. It may occur in severely ill patients after surgery and/or with neurological disorders, serious infections, cardiorespiratory insufficiency, and metabolic disturbances. Drugs that disturb colonic motility (e.g., anticholinergics or opioid analgesics) contribute to the development of this condition.

a

Fig. 7 a, b. Cecal volvulus. Transaxial CT (a) and coronal volume-rendered (b) CT demonstrate a markedly dilated cecum (C) in the left side of the pelvis. The arrow points to the area of twist of the ascending colon. Note the dilated small bowel loops due to the proximal colonic obstruction

Ischemic Bowel Disease Predominant causes of bowel ischemia are arterial or venous occlusion and hypoperfusion. Occlusive disease involves the mesenteric arteries, most commonly the superior mesenteric artery, in the large majority of cases. Bowel ischemia secondary to venous thrombosis is much less common. The only direct sign of vascular impairment of the bowel is diminished bowel wall enhancement, which is due to inadequate arterial inflow to the bowel. Increased bowel wall enhancement may be seen in some cases secondary to reactive hyperemia or compromised venous outflow. Other CT findings are direct visualization of a thrombus in the superior mesenteric artery or vein. Bowel distention and bowel wall edema are nonspecific findings and may accompany inflammatory or infectious causes. Bowel distention reflects the interruption of peristaltic activity in ischemic segments. Nonocclusive acute mesenteric ischemia usually is due to hypoperfusion secondary to severe cardiac disease, but also occurs in patients with end-stage renal or hepatic disease. An important radiographic manifestation of nonocclusive acute mesenteric ischemia due to low arterial flow is mesenteric arterial vasoconstriction. A less common form of non-occlusive acute mesenteric ischemia is severe vasculitis, which often affects younger individuals. Mesenteric vasculitis usually results in bowel edema and mucosal hyperenhancement. The small and the large bowel often are both involved. The duodenum is involved in approximately one-quarter of patients. The most common mechanical cause of bowel ischemia is obstruction. A closed-loop SBO is more likely than other types of obstruction to result in vascular compromise (strangulation obstruction). Strangulation obstruction has a reported prevalence of 5-40% and is a predominantly venous disease. The most frequent abnormality seen on CT is bowel wall thickening. The thickened bowel wall sometimes is associated with a target sign, consisting of alternating layers of high and low attenuation within the thickened bowel wall, which results from submucosal edema and hemorrhage. The bowel segment proximal to an obstruction can become ischemic

b

10

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

due to severe distention. CT findings that suggest subsequent infarction are non-enhancement of the bowel wall, gas in the bowel wall, mesenteric or portal veins, edema/hemorrhage in the mesentery adjacent to thickened and/or dilated bowel loops, and ascites (Fig. 8).

Perforation of the Gastrointestinal Tract Gastrointestinal perforation usually causes localized pain initially, and culminates in diffuse pain if peritonitis develops. Gastroduodenal perforation associated with peptic ulcer disease or a necrotic neoplasm has become less frequent in recent decades due to earlier diagnosis and improved therapy. At the same time, the incidence of gastroduodenal perforation resulting from endoscopic instrumentation has increased. Perforation of the small bowel is relatively uncommon but may be secondary to a foreign body, small bowel diverticulitis, or trauma. Spontaneous rupture of the colon is more frequent and can occur when the colon becomes markedly dilated proximal to an obstructing lesion (tumor, volvulus) or when the bowel wall is friable (ischemic or ulcerative colitis, necrotic neoplasm). Fiberoptic colonoscopy with or without biopsy is another cause of colonic perforation. Pneumoperitoneum can be recognized by the presence of subdiaphragmatic gas on an upright chest radiograph or an upright or left lateral decubitus abdominal radiograph. A large pneumoperitoneum generally is indicative of colonic perforation, whereas moderate quantities of free gas are seen with gastric perforation. Small bowel perforation usually results in either a limited amount of peritoneal gas or none, because the small bowel usually does not contain gas. Detection of subtle pneumo-

a

peritoneum is often difficult. As CT is far more sensitive than conventional radiography in demonstrating a small pneumoperitoneum, it has become the imaging test of choice when the results of conventional radiography are equivocal. Viewing the CT images at “lung window” settings improves the demonstration of small amounts of extraluminal gas. Retroperitoneal perforations (duodenal loop beyond the bulbar segment, or involving the appendix; posterior aspect of the ascending and descending colon, or the rectum below the peritoneal reflection) tend to be contained locally and remain clinically silent for several hours or days. Retroperitoneal gas has a mottled appearance and may extend along the psoas muscles. In contrast to intraperitoneal gas, retroperitoneal gas does not move freely when the patient’s position is changed from supine to upright for plain abdominal radiographs.

Acute Abdomen with Flank or Epigastric Pain Acute flank or upper abdominal pain radiating to the back is commonly a manifestation of retroperitoneal pathology, especially urinary colic, acute pancreatitis, or leaking abdominal aortic aneurysm.

Urinary Colic For decades, intravenous urography was the primary imaging technique used to evaluate patients with suspected urinary colic. Plain abdominal radiography and US may be useful for patients with a contraindication to radiation or iodinated intravenous CM. However, because

c

b Fig. 8 a-c. Strangulating small bowel obstruction due to an adhesive band that developed after cholecystectomy and appendectomy. Axial (a, b) and sagittal (c) views of multidetector CT show bowel wall thickening, enhancing bowel wall with submucosal edema, and/or hemorrhage giving a target sign (long arrows), indicating ischemia. Non-enhancement of the bowel wall of the jejunum (short arrow) corresponds to segmental infarction. A Ascites

11

Emergency Radiology of the Abdomen: The Acute Abdomen

of the low sensitivity of abdominal radiographs and US in the detection of urinary tract calculi, the role of unenhanced CT has become well established over the past 15 years. On CT, virtually all ureteral stones are radiopaque, regardless of their chemical composition. Uric acid stones have attenuation values of 300-500 Hounsfield units (HU), and calcium-based stones >1000 HU. In addition to the direct demonstration of a ureteral stone, secondary signs of ureterolithiasis, including hydroureter, hydronephrosis, perinephric stranding, and renal enlargement, may be visible (Fig. 9). Perinephric stranding and edema result from reabsorbed urine infiltrating the perinephric space along the bridging septa of Kunin. The more extensive the perinephric edema shown on unenhanced CT, the higher the degree of urinary tract obstruction. Focal periureteral stranding resulting from a local inflammatory reaction or irritation and induced by the passage of a stone helps to localize subtle calculi. Occasionally, a repeat CT examination using intravenous CM may be required, particularly if infectious complications are suspected. For the diagnosis of such complications (pyelonephritis), CT is helpful as it reveals a “striated nephrogram” after CM administration, as well as global enlargement of the kidney, renal and/or perirenal abscesses, or emphysematous pyelonephritis.

When no stone is detected, an alternative diagnosis must be established. Non-calculus urinary tract abnormalities causing symptoms of colic include acute pyelonephritis, renal cell carcinoma, acute renal vein thrombosis, spontaneous dissection of the renal artery, and renal infarction. Extraurinary diseases, such as retrocecal appendicitis, diverticulitis, SBO, pancreatitis, gynecological disorders, and retroperitoneal hemorrhage, may also simulate acute urinary colic.

Acute Pancreatitis An important disease causing upper abdominal pain is acute pancreatitis. US may be helpful for the demonstration of choledocolithiasis as a cause of acute pancreatitis and for the follow-up of known fluid collections. Since the CT findings correlate well with the clinical severity of acute pancreatitis, CT has become the imaging test of choice to stage the extent of disease (CT severity index of Balthazar) and to detect complications. The initial CT should be performed 48-72 h after disease onset (a CT examination performed too early in the course of the disease may not demonstrate any abnormality). Pancreatic enlargement due to interstitial parenchymal edema may progress to pancreatic exudate collecting in

a

b

Fig. 9 a-c. Right-sided urinary colic. Axial (a), coronal (b), and oblique (c) views of multidetector CT show dilatation of the proximal urinary tract due to a ureteral stone (arrow). Note slight secondary periuretral and perinephric stranding

c

12

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

a

c

b Fig. 10 a-c. Severe acute pancreatitis. Axial (a, b) and coronal (c) views of MDCT demonstrate an enlarged pancreatic area. a Only small areas of enhanced pancreatic parenchyma (p) remain within b extensive pancreatic necrosis (n). Note the tiny hyperattenuating calculi within the gallbladder and the lower bile duct (arrows). c The large amount of pancreatic necrosis contrasts sharply with the minimal extrapancreatic changes at this phase

the abdominal ligaments and potential spaces surrounding the pancreas. The pancreatic parenchyma may undergo necrosis or hemorrhage (Fig. 10). Severe pancreatitis is often complicated by infection of the necrotic site and/or thrombosis of the splenic and portal vein. Identifying infectious complications in patients with pancreatitis may be difficult; the demonstration of gas bubbles within peripancreatic collections is neither sensitive nor specific. CT-guided fluid aspiration for bacteriological examination is often the best technique to establish the presence of infection. Acute pancreatic and peripancreatic fluid collections may evolve into pseudocysts, which exhibit defined walls. A pseudocyst can erode peripancreatic vessels, thus causing bleeding or the formation of a pseudoaneurysm (Fig. 11).

Leaking Abdominal Aortic Aneurysm One of the most life-threatening diagnoses in patients with acute flank pain is a leaking abdominal aortic or iliac artery aneurysm. When a patient with suspected rupture of an abdominal aortic aneurysm is hemodynamically unstable, US is the initial imaging technique. The examination can be performed rapidly with portable equipment in the emergency room. However, para-aortic hemorrhage is poorly diagnosed by US. Instead, in hemodynamically stable patients, non-contrast-enhanced CT is the initial imaging test of choice as it is almost always able to demonstrate a para-aortic hematoma, if present, and may show additional findings helpful in establishing

Fig. 11. Recurrent pancreatitis. CT shows a pseudocyst of the pancreatic tail (PC) and a pseudoaneurysm of the gastroduodenal artery (PA). The splenic vein is indicated by the arrowhead

the diagnosis, such as a high-attenuating crescent sign. If endoluminal stent graft repair of the aorta is planned, contrast-enhanced CT should be performed.

Conclusions The imaging evaluation of patients with an acute abdomen has changed dramatically in the past decade. Plain abdominal radiographs largely have been replaced with US and CT. MDCT permits a rapid examination with high

Emergency Radiology of the Abdomen: The Acute Abdomen

diagnostic accuracy. Close cooperation with the referring physician prior to imaging remains essential for rapid and accurate diagnosis, as the character and location of the patient’s abdominal pain strongly influences the differential diagnosis and the choice of initial imaging test.

Suggested Reading Ahn SH, Mayo-Smith WW, Murphy BL et al (2002) Acute nontraumatic abdominal pain in adult patients: abdominal radiography compared with CT evaluation. Radiology 225:159-64 Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH (1990) Acute pancreatitis: value of CT in establishing prognosis. Radiology 174:331-336 Freeman AH (2001) CT and bowel disease. Br J Radiol 74:4-14 Gore RM, Miller FH, Pereles FS et al (2000) Helical CT in the evaluation of the acute abdomen. AJR 174:901-913 Mindelzun RE, Jeffrey RB (1997) Unenhanced helical CT for evaluating acute abdominal pain: a little more cost, a lot more information. Radiology 205:43-47

13

Novelline RA, Rhea JT, Rao PM, Stuk JL (1999) Helical CT in emergency radiology. Radiology 213:321-339 Paulson EK, Jaffe TA, Thomas J et al (2004) MDCT of patients with acute abdominal pain: a new perspective using coronal reformations from submillimeter isotropic voxels. AJR 183:899-906 Singh AK, Gervais DA, Hahn PF et al (2005) Acute epiploic appendagitis and its mimics. Radiographics 25:1521-34 Smith RC, Varanelli M (2000) Diagnosis and management of acute ureterolithiasis. AJR 175:3-6 Stoker J, van Randen A, Lameris W, Boermeester MA (2009) Imaging Patients with acute abdominal pain. Radiology 253:31-46 Taourel P, Kessler N, Lesnik A et al (2003) Helical CT of large bowel obstruction. Abdom Imaging 28:267-275 Urban BA, Fishman EK (2000) Tailored helical CT evaluation of acute abdomen. Radiographics 20:725-749 Werner A, Diehl SJ, Farag-Soliman M, Düber C. (2003) Multi-slice spiral CT in routine diagnosis of suspected acute left-sided colonic diverticulitis: a prospective study of 120 patients. Eur Radiol 13:2596-2603 Wiesner W, Khurana B, Ji H, Ros PR (2003) CT of acute bowel ischemia. Radiology 226:635-650

IDKD 2010-2013

Trauma of the Abdomen and Pelvis Philip J. Kenney1, Stuart E. Mirvis2 1 University 2 University

of Arkansas for Medical Sciences, Little Rock, AR, USA of Maryland School of Medicine, Baltimore, MD, USA

Introduction Trauma is a major health problem in all age groups but it is especially true in the young, due to high-velocity transportation, altercations, including with weapons and resulting in penetrating injuries, as well as falls and sportsrelated injuries. In addition, both the elderly and pregnant women are vulnerable to trauma. Improvements in the management of trauma include more rapid rescue, better organization of trauma centers, and advances in treatment. Current trends include increased non-operative management of trauma-related injuries, accurate imagingbased diagnosis, and greater emphasis on the efficient but cost-effective use of imaging. One aspect of the trend to non-operative care is the desire to avoid non-therapeutic surgery; this is possible if imaging can identify those patients who require surgery. Another is the realization that non-operative care can result in better long-term outcome, such as splenic salvage.

Computed Tomography vs. Ultrasound Controversy exists about the appropriate use of computed tomography (CT) vs. ultrasound (US), although each modality has its advantages and disadvantages [1, 2]. In general, CT has the best statistical accuracy for detecting, characterizing, and excluding injuries. In modern high-volume trauma centers, the CT apparatus must be located in the trauma suite such that even unstable patients can be examined quickly without compromise. This allows for the efficient use of CT in rapid and accurate diagnostics and obviates the need for outmoded studies, such as diagnostic peritoneal lavage. CT is also more reliable at excluding injury, allowing the patient to be discharged home and avoiding the expense of observation in hospital. However CT may be overused; indeed, in one study only three of 100 patients had alterations of clinical management due to follow-up CT [3]. US can detect significant injury which can then be appropriately treated; conversely, low-risk patients with normal sonograms may be observed and possibly avoid CT [2]. However, patients

with abnormal US findings often require further evaluation with CT. In a large study, US had 86% sensitivity and 98% specificity but with 43 false-negative and 23 indeterminate studies, including six splenic, one liver, one renal, one pancreatic, and one bowel injury [2]. In traumatized pregnant women, US should be the first examination as it can evaluate the pregnancy, documenting fetal death or viability. US is nearly as accurate in detecting the abnormal presence of fluid in pregnant patient as in non-pregnant patients [4]. If US shows fluid or other injury, CT is justified for further evaluation (Fig. 1). The best outcome for the fetus is assured by best care of the mother. The radiation risk is reasonable if there is lifethreatening injury, such that prompt diagnosis and treatment are paramount [5].

Urinary Tract The nearly universal use of CT has altered the assessment of urinary tract trauma. While significant hematuria has been shown to be the best indicator of urinary tract injury, presently the decision to perform CT has little to do with the presence or absence of hematuria. CT is a primary investigation, after standard radiographs, in those with significant mechanisms of injury or any signs or symptoms of significant injury. Intravenous urography has been replaced by CT (Fig. 2). The ability of US to evaluate renal injury is limited [6] whereas CT has excellent negative predictive value for renal injury. CT also accurately indicates the presence and type of renal injury [7]. Renal contusion appears as an illdefined region of diminished enhancement. Segmental renal infarction is identified as a wedge-shaped, well-defined area of non-enhancement. Renal artery occlusion can be accurately diagnosed by its complete lack of either contrast enhancement or excretion by the kidney, usually with little to no associated hematoma. Angiography is thus not needed, and conservative therapy is most often used today. Most renal injuries are lacerations, with simple lacerations limited to the cortex and deep lacerations extending into the collecting system, which may show extravasation. Delay scans of 2-10 min aid in demonstrating or excluding

15

Trauma of the Abdomen and Pelvis

a

b

Fig. 1 a-c. A pregnant woman suffered a high-speed motor vehicle collision. a US demonstrates intrauterine fetus (heart motion documented) and free pelvic fluid. b US shows perisplenic fluid with hypoechoic defect. c CT confirms splenic laceration; note the higher density of the perisplenic hematoma compared to the rim of fluid about liver (sentinel clot sign)

c

a

b

Fig. 2 a, b. Hematuria and left upper quadrant pain after a footballrelated injury. a Intravenous urogram shows no abnormality. b Subsequent CT for persistent pain showed free fluid in the pelvis and extensive splenic laceration with extensive “blush”. Surgery confirmed a grade 4 splenic injury

extravasation (Fig. 3), although in most cases small amounts of extravasation will resolve with conservative therapy. Subcapsular hematoma is delimited by the renal cortex and may deform the renal surface; perinephric hematoma extends from the renal surface to fill Gerota’s space but does not deform the renal contour, although it may displace the kidney. CT is excellent at demonstrating the extent of hematoma and in evaluating enlargement on

follow-up scans [7]. Renal fracture indicates a single complete fracture plane, often extending through the collecting system; multiple planes of disruption are seen in a shattered kidney. CT can also diagnose avulsion of the ureteropelvic junction (UPJ) or ureteral injury, demonstrating lack of opacification of the ureter, retroperitoneal water attenuation collections adjacent to the pelvis or ureter, and possibly extravasation of contrast on delay scans (Fig. 4) [7].

16

a

Philip J. Kenney, Stuart E. Mirvis

a

b b

Fig. 3 a, b. Hematuria after fall from a power line. a Initial CT shows left renal laceration with perinephric hematoma. b Delay image shows no leak from the collecting system

Ureteral injuries, including UPJ avulsion, are uncommon. They can occur either with penetrating trauma or high-velocity blunt trauma and have no specific signs or symptoms, but can be detected with CT. Routine CT scans show subtle suggestive signs, such as perinephric and peripelvic stranding or fluid, that indicate the need for delay scans, if not routinely done, to demonstrate extravasation. In a study based on over 4,000 trauma patients, CT enabled the correct identification of seven of eight UPJ avulsions [8]. The AAST Organ Injury severity scale for the kidney includes lesions with different appearances in each category (1: contusion, small subcapsular hematoma; 2: 1 cm laceration without extravasation; 4: deep laceration with extravasation or main renal artery or vein injury; 5 shattered kidney or UPJ avulsion) and has been shown to correlate with need for surgery and outcome [9]. Urethral injuries are predominantly seen in males. Anterior urethral ruptures most commonly occur due to straddle injury. Posterior urethral ruptures most often are

Fig. 4 a, b. Routine trauma CT image shows fluid and stranding about the right ureter. a Note that contrast filling of the ureter has not yet occurred. Delay image shows extravasation medial to the kidney typical, of UPJ avulsion. b Note the intact parenchyma

due to compressive force and resultant pubic bone fractures, although both anterior and posterior urethral injury can result from penetrating injury. Retrograde urethrography is the only accurate diagnostic imaging procedure. If a urethral injury is strongly suspected, a urethrogram should be performed before passage of a catheter (Fig. 5). However, in patients with moderate risk, a urethral catheter may be gently passed so that the patient may go on to CT. A pericatheter urethrogram may then be performed after any other injuries have been stabilized. Five types of urethral injuries are recognized. In type 1, the posterior urethra is stretched but intact; in type 2, there is a tear of the membranous urethra above the urogenital diaphragm; in type 3, the posterior urethral tear is above and below the urogenital diaphragm; type 4 is defined as a bladder-neck injury and type 5 as an anterior urethral injury [10]. Bladder injuries consist of contusions and ruptures; classically, they have been detected with standard radiographic cystography (Fig. 6). They may be extraperitoneal, most commonly, intraperitoneal, less commonly, or combined in about 5%. While CT with only intravenous contrast may fail to identify extravasation from

17

Trauma of the Abdomen and Pelvis

a

b

Fig. 5. Blunt trauma resulted in pubic rami fractures. Retrograde urethrogram reveals type 3 posterior urethral rupture

Fig. 7 a, b. Gross hematuria following motor-vehicle collision resulting in extensive pelvic fractures. a Standard CT shows pelvic fluid but no extravasation. b CT cystogram documents extraperitoneal bladder rupture (note the clot in the bladder)

Fig. 6. Gross hematuria after gunshot wound to the pelvis. Standard cystogram shows extraperitoneal rupture (X marks entry site, O exit)

a ruptured bladder, several studies have shown very high accuracy for CT-cystography (Fig. 7), which is now the standard in our institutions. In patients with suspected bladder rupture (primarily those with gross hematuria, over 25 RBC/hpf, with pelvic fractures or unexplained pelvic fluid), a standard CT with the

bladder catheter clamped is performed. If there is no extravasation, the bladder is drained and then re-filled with 300-500 mL of dilute contrast and the pelvis rescanned. Bladder ruptures are virtually always associated with fluid or hematoma in the pelvis, but such blood or fluid may be due to splenic or other injuries or to pelvic fracture. Extravasation confined to the lower pelvis and not outlining bowel loops (and which may extend up the retroperitoneum) indicates extraperitoneal rupture, which most often is managed conservatively. Extravasation high near the dome and outlining bowel loops or extending to the gutters or higher indicates intraperitoneal rupture, which is more often managed surgically [11]. In a study of 495 patients with potential pelvic injuries, CT-cystography detected 98% of the bladder injuries while standard cystography detected 95%. Of the patients with bladder injury (65% extraperitoneal, 35% intraperitoneal, 5 combined), 89% had gross hematuria and pelvic fracture, 9% had gross hematuria with no pelvic fracture, and one patient had microscopic hematuria and pelvic fracture [12]. CTcystography carried out with multidetector CT allows for reformatting, which can more clearly demonstrate the point of leakage and allows more accurate characterization of the type of injury [13].

18

Bowel and Mesenteric Injuries Bowel and mesenteric injuries are found in about 5% of patients undergoing surgery for trauma and are seen in 0.7% of all traumatized patients [1, 14]. The mechanism of injury is direct compressive force, including from seatbelts, although deceleration may play a role. Morbidity and mortality can occur, with peritonitis and abscess resulting if the injury is missed. Clinical signs and symptoms are non-specific. Although diagnosis by CT is not as straightforward as is the case for other abdominal organ injuries, CT is the most accurate diagnostic modality, with >90% sensitivity and specificity reported [14, 15]. The use of orally administered contrast is now somewhat controversial; while extravasation of oral contrast can be a very specific sign of bowel injury, it is rarely seen. Contrast administration delays performance of the scan, and most bowel injuries will be evident based on other signs. There is no one CT sign that is both sensitive and specific for bowel or mesenteric injury. Focal bowel wall thickening, mesenteric stranding, interloop fluid, and hematoma are common but less specific, particularly for surgically important injuries (Fig. 8). Active bleeding, vessel beading, abrupt termination of mesenteric vessels, and bowel wall defect are more specific but less sensitive signs [16]. Active bleeding is seen as a focal extraluminal collection with attenuation similar to that of the aorta at the same level and different from the adjacent organs. Free air is considered a good sign of perforated bowel, but in fact it has limited value. It is infrequently seen in those with bowel injury and may represent air tracking into the peritoneum from thoracic injuries. In a study reported in 2008, free air had a sensitivity of 24% albeit a specificity of 95%. There were three false-positives with intraperitoneal air instead resulting from supradiaphragmatic or bladder injuries [16]. If a single finding

Fig. 8. Motor-vehicle collision. Focal hematoma and thickening of cecum; at surgery, cecal laceration found

Philip J. Kenney, Stuart E. Mirvis

is noted, the likelihood of injury is low; a combination of findings, particularly free fluid without obvious source in combination with focal bowel wall thickening and/or mesenteric stranding, is very suggestive of bowel injury and such patients should be explored or followed very carefully [14-16]. In our practice, we have found that performing a repeat abdominal-pelvic CT 4-6 h after the admission scan can be helpful in patients with suspicious but non-diagnostic findings for full-thickness bowel injury, by demonstrating injury progression such as the development of free air, increasing intraperitoneal fluid, or stability of findings. Of course, management decisions are made in conjunction with any evolution of the clinical findings.

Splenic Injuries The spleen is the most frequently injured abdominal organ in blunt trauma. There may be signs of blood loss or left upper quadrant pain, but the diagnosis largely rests on imaging or surgical exploration. A trend to nonoperative management is supported by evidence that long-term health is better in those who have had splenic function preserved. This necessitates accurate noninvasive diagnosis and is aided by signs predictive of the success or failure of conservative management. Splenic injuries can cause free fluid, perisplenic or elsewhere, which can readily be detected by sonography. Splenic injury may alter echo-texture: lacerations may be anechoic if there is rapid bleeding, but more commonly are more echogenic than normal spleen [2]. With such findings on sonography, the decision whether to further evaluate with CT or to proceed to surgery can be made on clinical grounds. Splenic injuries may be missed by sonography, particularly if they are not associated with free fluid. In one large study, there were 43 false-negative sonograms, including six splenic ruptures that required surgery [2]. CT is quite sensitive in the detection of splenic injuries [17]. Subcapsular hematoma is seen as a crescentic, lowattenuation, peripheral rim; intraparenchymal hematoma as a rounded area within the spleen with low attenuation and no enhancement. Lacerations are common, appearing as linear or branching low-attenuation lesions that often extend to the surface; if so, they are often associated with perisplenic or free fluid. Hemoperitoneum tends to be of higher attenuation close to the source of bleeding; thus, when the spleen is the source, the collection adjacent to the spleen may be higher in attenuation than elsewhere, a finding referred to as the sentinel clot sign. Lacerations may involve the vasculature. There can be devascularization of the spleen by hilar injury, or active extravasation into the peritoneal cavity, or a confined area of extravasation (pseudoaneurysm) (Fig. 9). Both types of extravasation indicate that non-operative management may not succeed, although angiographic embolization may control the bleeding and allow splenic salvage [18].

19

Trauma of the Abdomen and Pelvis

a

b

Fig. 9 a, b. Blunt trauma. a Routine trauma CT image shows area of hyperdensity (arrowhead). b Delay image shows that the hyperintensity is no longer visible; the lesion has become isodense with blood pooling, indicating confined pseudoaneurysm rather than free active bleeding

A number of schemes have been devised to grade splenic injury on CT in an attempt to predict outcome, with variable correlation with the need for surgery [1]. One of the commonest is the AAST scoring system. In a large study, failure of non-operative management correlated with splenic injury grade: the failure rate was 50% of the parenchyma; contrast blush >1 cm; or large hemoperitoneum; however, further study showed that this approach also had limited predictive value with poor sensitivity although fair specificity [20]. The additional finding of traumatic pseudoaneurysm or active extravasation (which does not confer a specific stage in the AAST scoring system) increased the likelihood of failure of nonoperative management, regardless of grade [21]. Delayed images can help distinguish between active bleeding, which persists as a hyperdense area, and confined vascular injury (pseudoaneurysm), which washes out [22]. Patients with active bleeding are more likely to require surgery or other forms of intervention.

of the liver and the difficulty in clearly imaging all portions of the organ ultrasonographically. Injuries to the liver include contusion, seen on CT as an ill-defined area of low attenuation; subcapsular hematoma, a crescentic collection limited by the capsule; and intraparenchymal hematoma, a collection of blood within a liver laceration. Laceration is commonest, seen as linear or branching low-attenuation regions, sometimes with jagged margins, that can extend to the hepatic surface or to vessels. Superficial lacerations are 80%, and clinical signs of hemodynamic instability dictate the need for intervention more than imaging features. However in one recent study of 214 patients with hepatic injury, all 14 who showed intraperitoneal contrast extravasation on CT required surgery [25]. Massive hemoperitoneum in six compartments also correlated independently with the need for surgical intervention. Gallbladder injuries occur in 20 eosinophils per highpowered field. These patients usually have a dramatic positive response to oral steroids or inhaled steroid preparations.

Erosive gastritis is usually manifested on double-contrast studies by varioliform erosions with punctate or slit-like collections of barium surrounded by radiolucent mounds of edema. Varioliform erosions tend to be located in the gastric antrum and are often aligned on the crests of the folds. Aspirin and other NSAIDs are by far the most common cause of erosive gastritis. Occasionally, NSAID-induced erosive gastritis may also manifest as distinctive linear or serpiginous erosions clustered together on or near the greater curvature of the gastric body [11]. It has been postulated that these erosions result from localized mucosal injury, as the dissolving NSAID tablets collect by gravity in the most dependent portion of the stomach.

Esophageal Carcinoma

Helicobacter Pylori Gastritis

Early esophageal cancers classically appear on doublecontrast studies as small, protruded tumors [10]. They may be plaque-like or small polypoid lesions. Other,

Gastritis caused by the bacterium Helicobacter pylori can be diagnosed on barium studies by the presence of thickened folds in the antrum, body, or, less commonly, the

24

Marc S. Levine, Ahmed Ba-Ssalamah

change in its shape. Usually, ulcer healing produces a visible scar manifested by a central pit or depression, radiating folds, and/or retraction of the adjacent gastric wall [13].

Gastric Carcinoma

Fig. 4. Helicobacter pylori gastritis. Left posterior-oblique spot image from double-contrast upper gastrointestinal examination shows thickened, irregular folds in the gastric body due to chronic H. pylori gastritis

fundus of the stomach (Fig. 4) [12]. Other patients with H. pylori infection may have a polypoid form of gastritis, with grossly thickened, lobulated folds such that the lesions resemble those of Menetrier’s disease, lymphoma, or a submucosally infiltrating carcinoma [12]; thus, endoscopy and biopsy are required for a definitive diagnosis.

Gastric Ulcers Benign gastric ulcers classically appear en face as round or ovoid collections of barium, often surrounded by a smooth mound of edema or thin, straight folds radiating to the edge of the ulcer crater [13]. When viewed in profile, benign ulcers project beyond the contour of the adjacent gastric wall and are often associated with an ulcer mound or collar. In contrast, malignant gastric ulcers appear en face as irregular ulcer craters within a discrete mass, sometimes associated with nodularity or clubbing of adjoining folds due to tumor infiltration of the folds [13]. When viewed in profile, malignant ulcers project inside the lumen within a mass that forms acute angles with the gastric wall rather than the obtuse, gently sloping angles expected for a benign mound of edema. Most benign ulcers are located on the lesser curvature or posterior wall of the gastric antrum or body [13]. Occasionally, benign gastric ulcers may occur on the greater curvature of the distal stomach, in which case the vast majority are caused by aspirin or other NSAIDs [13]. As these NSAID-induced greater-curvature ulcers enlarge, they can penetrate inferiorly into the transverse colon, producing a gastrocolic fistula [14]. Ulcer healing may be seen as a decrease in the size of the ulcer or a

Advanced gastric carcinomas may appear on barium studies as polypoid, ulcerated, or infiltrating lesions. Other primary scirrhous carcinomas can have a “linitis plastica” appearance, with luminal narrowing, irregularly thickened folds, and nodularity of the mucosa [15]. Scirrhous carcinomas classically involve the gastric antrum, but 40% of these lesions are confined to the gastric body or fundus (Fig. 5) [15]. Early gastric cancers may be manifested by small polypoid or ulcerated lesions. However, in the western world, the vast majority of patients with gastric carcinoma already have advanced lesions at presentation. As a result, early gastric cancer is unlikely to be detected as long as barium studies are performed predominantly on symptomatic patients [16].

Gastric Lymphoma Chronic H. pylori gastritis can lead to the development of mucosa-associated lymphoid tissue (MALT) in the stomach. This lymphoid tissue is the precursor of low-grade, B-cell gastric MALT lymphomas, which, if untreated, may undergo blastic transformation to more high-grade lymphomas. Gastric MALT lymphomas may sometimes be recognized on double-contrast studies by variably sized, rounded, confluent nodules in the stomach [17]. In contrast, advanced gastric lymphomas may be manifested by thickened folds, multiple submucosal masses, ulcerated bull’s-eye lesions, or giant, cavitated lesions.

Fig. 5. Scirrhous adenocarcinoma of the stomach. Front spot image from double-contrast upper gastrointestinal examination shows irregular narrowing of the gastric body and fundus due to infiltration of the wall by tumor. Note transition (arrows) to uninvolved gastric antrum distally. About 40% of scirrhous carcinomas are confined to the body or fundus of the stomach with antral sparing

Diseases of the Esophagus and Stomach

Gastrointestinal Stromal Tumors Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. Approximately 70% of all GISTs are found in the stomach and only 2-5% originate from the esophagus. GISTs have a wide clinical spectrum, ranging from benign, incidentally detected nodules to large malignant tumors, and must be distinguished from other mesenchymal tumors [18]. The most frequent symptom related to gastric or esophageal tumor is dysphagia or heartburn; thus, most patients undergo endoscopy of the esophagus, stomach and duodenum, with simultaneous biopsy if tumor is seen [19]. If the histopathological results reveal a tumor, accurate staging is required. Endoscopic ultrasound (EUS) can depict the normal gastric and esophageal wall with its five-layered internal structures, thus allowing detailed evaluation of the depth of tumor penetration even in early-stage disease. EUS is useful in the diagnostic work-up of early cancer, as it can distinguish between T1a tumors, in which only mucosectomy is needed, and T1b tumors, in which a complete resection is indicated [20]. However, because ultrasound penetration is not deep enough when transducers with higher frequencies are used to visualize fine structures, evaluation of deep tumor infiltration may be difficult and assessment of metastases may be limited by the finite depth of penetration [21]. EUS also is examiner-dependent, time-consuming, and unable to pass stenotic tumors. These limitations can be overcome using multidetector CT technology (MDCT), with its ability to cover a large volume in a very short scan time with a single breathhold. Thin collimation and isotropic voxels allow imaging of the entire esophagus and stomach with high-quality multiplanar reformation and 3D reconstruction (Fig. 6)

Fig. 6. Hydro-MDCT of the esophagus in coronal reformation to follow the course of the esophagus, demonstrating normal wall thickening of the esophagus (≤3 mm) and homogeneous enhancement

25

[22, 23]. Furthermore, MDCT with water filling (hydroMDCT, HMDCT) provides information about esophageal and gastric wall infiltration, extramural extent of disease, lymph node involvement, and distant metastases (Fig. 7) [23, 24]. Inadequately distended hollow viscera on CT may hide large lesions and may even mimic pseudolesions. Thus, optimal distention of the esophagus and stomach is a necessary prerequisite for achieving good diagnostic imaging. When water is used as the oral contrast agent, subtle pathology is easier to visualize. Gas or CO2 resulting from the administration of effervescent granules can be used for hollow-organ distention alone or in combination with water. The use of negative rather than positive contrast media is preferred, especially if CT angiography images are needed. Subtle pathology is easier to visualize, especially when an adequate intravenous contrast material bolus is administered [23, 24]. Threedimensional reconstructions of CT data sets with multiplanar reconstruction, curved planar reformations, or other protocols are mandatory to exploit the full potential of MDCT. Three-dimensional virtual gastroscopy provides an endoluminal image similar to conventional fiberoptic gastroscopy. Therefore, HMDCT is a valuable tool for the complete staging of gastric and esophageal tumors and serves as an adjunct to endoscopy. However, EUS and MDCT are anatomically based diagnostic techniques with certain drawbacks. These include limited sensitivity with false-negative findings due

Fig. 7. Hydro-MDCT of the esophagus in coronal reformation shows a huge mass with inhomogeneous enhancement in the mediastinum arising from the esophageal wall, with infiltration of the left main bronchus (arrow) and the left diaphragm (arrowhead), with regard to the T4 tumor. Note the pleural effusion

26

to non-enlarged, tumor-involved lymph nodes, and limited specificity with false-positive findings due to enlarged lymph nodes not involved by tumor. Furthermore, after neo-adjuvant chemotherapy and re-evaluation, it is not possible with these techniques to distinguish between fibrotic changes or remaining vital tumor due to their morphological similarities [25].Thus, there is an urgent need for additional functional examination techniques. Positron emission tomography (PET) yields physiological information that provides a means to diagnose cancer based on altered tissue metabolism. PET takes advantage of the principle that biochemical changes often precede or are more specific than the structural changes associated with any given disease process [26]. Therefore, PET offers the potential to show early esophageal cancer or small lymph node metastases before any structural abnormality is detectable or to exclude the presence of tumor in an anatomically altered structure. For example, FDG-PET can detect metastatic lymph nodes that are not enlarged on CT, and can help differentiate pathological from non-specifically enlarged lymph nodes, which usually show no uptake of FDG. Tumor uptake of FDG, measured as the maximal standardized uptake value (SUVmax) in FDG-PET, even provides a quantitative estimate of tumor aggressiveness [27]. Recent studies demonstrated that FDG-PET can be used not only for pre-treatment staging, but also for assessment of treatment response, detection of recurrence, and prediction of survival in patients with adenocarcinoma of the esophagus or stomach [27]. However, for mucinous and signet-ring cell adenocarcinoma of the stomach, it is less useful. This may be due to low or absent FDG activity in these tumors, which is the result of a high content of metabolically inert mucus, leading to a reduced FDG concentration. Another reason could be the lack of expression of the glucose transporter Glut-1 on the cell membrane of most signet-ring cell and mucinous adenocarcinomas [28]. The spatial resolution of FDG-PET is lower than that of CT scans. When metastatic nodes exist around the primary tumor, it can be difficult to distinguish uptake in these nodes from the intense activity of the primary tumor. However, the advent of PET/CT imaging, enabling co-registration of both anatomical and functional information, has overcome this disadvantage and improved the localization of increased FDG uptake (Fig. 8) [29]. FDGPET/CT shows the extent of disease more accurately than other imaging methods, and this frequently leads to a radical change in patient management. Combined PET/CT imaging is therefore a valuable diagnostic tool for the primary diagnosis of GISTs or assessment of the response to therapy [30, 31]. PET/CT scans can also be used in staging patients with primary gastric lymphoma, as well as for monitoring these tumors after therapy [30]. However, the availability of FDG-PET, and FDG-PET/CT in particular, is still limited and their use expensive. Thus, at present, HMDCT plays a major role as a triage tool to aid in choosing the appropriate treatment

Marc S. Levine, Ahmed Ba-Ssalamah

a

b

Fig. 8 a, b. Hydro-MDCT PET scan of an early cervical esophageal carcinoma and enlarged right-sided lymph node. a On the axial HMDCT scan of the upper thorax, the small esophageal cancer is not clearly seen and the enlarged lymph appears suspicious. b Corresponding axial fused HMDCT-PET image shows increased activity in the region of the small primary tumor (arrow) and an enlarged metastatic lymph node (arrowhead)

for patients with esophageal and gastric tumors. HMDCT may help distinguish surgical candidates with limited disease from patients in need of pre-operative chemoradiation for down-sizing a tumor or from patients who need palliative therapy for advanced, unresectable tumor. When HMDCT shows definite advanced disease with extensive tumor spread, pre-surgical chemotherapy or radiochemotherapy is used to improve the prognosis. After completion of neo-adjuvant treatment, the tumor can then be restaged to determine whether surgical resection is indicated. Thus, pre-operative staging of esophageal and gastric tumors is by far the most important indication for HMDCT. This technique also plays an important role in the evaluation of postoperative complications such as fistulas when the findings on barium studies are equivocal.

References 1. Graziani L, Bearzi I, Romagnoli A et al (1985) Significance of diffuse granularity and nodularity of the esophageal mucosa at double-contrast radiography. Gastrointest Radiol 10:1-6 2. Gupta S, Levine MS, Rubesin SE et al (2003) Usefulness of barium studies for differentiating benign and malignant strictures of the esophagus. AJR 180:737-744 3. Levine MS, Kressel HY, Caroline DF et al (1983) Barrett esophagus: reticular pattern of the mucosa. Radiology 147:663-667

Diseases of the Esophagus and Stomach

4. Levine MS, Macones AJ, Laufer I (1985) Candida esophagitis: accuracy of radiographic diagnosis. Radiology 154:581-587 5. Levine MS, Woldenberg R, Herlinger H, Laufer I (1987) Opportunistic esophagitis in AIDS: radiographic diagnosis. Radiology 165:815-820 6. Levine MS, Loevner LA, Saul SH et al (1988) Herpes esophagitis: sensitivity of double-contrast esophagography. AJR 151:57-62 7. Sor S, Levine MS, Kowalski TE et al (1995) Giant ulcers of the esophagus in patients with human immunodeficiency virus: clinical, radiographic, and pathologic findings. Radiology 194:447-451 8. Bova JG, Dutton NE, Goldstein HM, Hoberman LJ (1987) Medication-induced esophagitis diagnosed by double-contrast esophagography. AJR 148:731-732 9. Zimmerman SL, Levine MS, Rubesin SE et al (2005) Idiopathic eosinophilic esophagitis in adults: the ringed esophagus. Radiology 236:159-165 10. Levine MS, Dillon EC, Saul SH, Laufer I (1986) Early esophageal cancer. AJR 146:507-512 11. Levine MS, Verstandig A, Laufer I (1986) Serpiginous gastric erosions caused by aspirin and other nonsteroidal antiinflammatory drugs. AJR 146:31-34 12. Sohn J, Levine MS, Furth EE et al (1995) Helicobacter pylori gastritis: radiographic findings. Radiology 195:763-767 13. Levine MS, Creteur V, Kressel HY et al (1987) Benign gastric ulcers: diagnosis and follow-up with double-contrast radiography. Radiology 164:9-13 14. Levine MS, Kelly MR, Laufer I et al (1993) Gastrocolic fistulas: the increasing role of aspirin. Radiology 187:359-361 15. Levine MS, Kong V, Rubesin SE et al (1990) Scirrhous carcinoma of the stomach: radiographic and endoscopic diagnosis. Radiology 175:151-154 16. White RM, Levine MS, Enterline HT, Laufer I (1985) Early gastric cancer: recent experience. Radiology 155:25-27 17. Yoo CC, Levine MS, Furth EE et al (1998) Gastric mucosa-associated lymphoid tissue lymphoma: radiographic findings in six patients. Radiology 208:239-243 18. Miettinen M, Sarlomo-Rikala M, Sobin LH, Lasota J (2000) Esophageal stromal tumors: a clinicopathologic, immunohistochemical, and molecular genetic study of 17 cases and comparison with esophageal leiomyomas and leiomyosarcomas. Am J Surg Pathol 24:211-222

27

19. Galmiche JP, Clouse RE, Balint A et al (2006) Functional esophageal disorders. Gastroenterology 130:1459-1465 20. Jung M (2005) [Mucosectomy as sufficient therapy for early squamous cell]. Chirurg 76:1018-1024 21. Kutup A, Link BC, Schurr PG et al (2007) Quality control of endoscopic ultrasound in preoperative staging of esophageal cancer. Endoscopy 39:715-719 22. Prokop M. New challenges in MDCT (2005) Eur Radiol 15 Suppl 5:E35-45 23. Ba-Ssalamah A, Prokop M, Uffmann M et al (2003) Dedicated multidetector CT of the stomach: spectrum of diseases. Radiographics 23:625-644 24. Ba-Ssalamah A, Zacherl J, Noebauer-Huhmann IM et al (2009) Dedicated multi-detector CT of the esophagus: spectrum of diseases. Abdom Imaging 34:3-18 25. Westerterp M, van Westreenen HL, Reitsma JB et al (2005) Esophageal cancer: CT, endoscopic US and FDG PET for assessment of response to neoadjuvant therapy-systematic review. Radiology 236:841-851 26. Luketich JD, Schauer PR, Meltzer CC et al (1997) Role of positron emission tomography in staging esophageal cancer. Ann Thorac Surg 64:765-769 27. Cerfolio RJ, Bryant AS (2006) Maximum standardized uptake values on positron emission tomography of esophageal cancer predicts stage, tumor biology, and survival. Ann Thorac Surg 82:391-394; discussion 394-395 28. Chen J, Cheong JH, Yun MJ et al (2005) Improvement in preoperative staging of gastric adenocarcinoma with positron emission tomography. Cancer 103:2383-2390 29. Hsu WH, Hsu PK, Wang SJ et al (2009) Positron emission tomography-computed tomography in predicting locoregional invasion in esophageal squamous cell carcinoma. Ann Thorac Surg 87:1564-1568 30. Suga K, Yasuhiko K, Hiyama A et al (2009) F-18 FDG PET/CT findings in a patient with bilateral orbital and gastric mucosa-associated lymphoid tissue lymphomas. Clin Nucl Med 34:589-593 31. Antoch G, Kanja J, Bauer S et al (2004) Comparison of PET, CT, and dual-modality PET/CT imaging for monitoring of imatinib (STI571) therapy in patients with gastrointestinal stromal tumors. J Nucl Med 45:357-365

IDKD 2010-2013

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis Marc J. Gollub Memorial Sloan Kettering Cancer Center, New York, NY, USA

Computed Tomography Enterography Computed tomography enterography (CTEG) is a focused CT scan examination of the small intestine that combines the advantages of isotropic, thin-section multiplanar CT; the large volumes of neutral-density oral contrast; and rapid administration of intravenous contrast. Oral contrast agents such as 0.1% barium, polyethylene glycol (PEG) and methylcellulose contain additives that inhibit fluid reabsorption and allow maximal bowel distention. Intravenous contrast is injected at the “enteric phase” (about 45 s) to provide maximum wall enhancement against the neutral-density lumen (0-30 HU) [1]. Pharmacological manipulation to interrupt small-bowel spasm and encourage gastric emptying is commonly used, including glucagon and metoclopramide, respectively. The indications for CTEG include Crohn’s disease and other enteritides, obscure gastrointestinal bleeding (OGIB), detection of intestinal masses, and sprue. A prospective blinded comparison of CTEG with wireless capsule endoscopy (WCE) using clinical consensus as the gold-standard found similar sensitivities (82 vs. 83%) for active small-bowel Crohn’s disease, but CTEG was far more specific than WCE (89 vs. 53%). Although there is less experience in using CTEG for OGIB, one study found a bleeding source in 45% of 22 patients, in three of whom the source had been missed by initial WCE [2]. No comprehensive reports of mass detection have been published yet, but in our experience CTEG appears to represent a first-line test for suspected carcinoid. This chapter will discuss the pitfalls and limitations of CTEG and CT enteroclysis (CTEC).

Pitfalls of Computed Tomography Enterography General Even with the aid of neutral-density oral contrast to assist in mass conspicuity, mildly enhancing masses or mural inflammation may be subtle and overlooked without proper adjustment of the window and level. We recommend, in addition to standard abdominal settings, a

liver-type setting (W215, L135) and an “enterographictype” setting (W430, L155). Since CTEG does not create an enteral challenge, in contrast to CTEC, its use in the setting of low-grade small-bowel obstruction should be discouraged in favor of CTEC or magnetic resonance enteroclysis. However, even during a non-obstructive episode, CTEG may confer an advantage by delineating a subtle, underlying cause such as a mass or stricture. Fastidious technique and timing are basic requirements in CTEG. Inadequate oral intake, early or late scanning, poor intravenous enhancement (low rate, reduced dose), or underlying hypo- or hypermotility disorders can interfere with good distention and wall conspicuity. Some of these technical pitfalls can be overcome with low-kVp imaging, close monitoring of the drinking schedule, and pharmacological manipulation. Poor jejunal distention is a generally expected limitation compared with other examinations. Fortunately, many abnormalities are diagnosed at CTEG solely by virtue of mural hyperenhancement.

Crohn’s Disease Proof of disease, grading of severity, and assessment of penetrating disease are important tasks prior to surgical planning and even prior to medical treatment, since the therapeutic anti-tumor necrosis factor alpha (TNFα) antibodies (e.g., infliximab) are expensive and have side effects of infection and drug-induced immune disease [3]. The classic findings at CTEG in patients with active Crohn’s disease have been well described [2]. The American College of Radiology has deemed CTEG to be the most appropriate imaging test for Crohn’s disease [4]. Nonetheless, studies testing the accuracy of CTEG compared with other radiological and endoscopic tests have identified some clear limitations and pitfalls for CTEG in the setting of suspected Crohn’s disease, including: 1. non-specificity of certain findings, such as mural hyperenhancement without skip areas; 2. in known Crohn’s disease, signs of very early disease, such as aphthous ulcers, which will not be detected unless a mucosal study is done; 3. measurement of disease severity, including quantitation of ulcers or fistulas; and

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis

29

4. correlation with clinical findings. Wold et al. found CTEG to be superior to small-bowel follow-through for the detection of abscesses and fistulas, with a sensitivity and specificity of 78/83% vs. 62/90%, respectively [5]. Vogel et al. showed that, although CTEG is accurate for determining the presence or absence of strictures and fistulas (sensitivity 100, 92% respectively), it is less accurate in determining their number (sensitivity 67%, both). This may be clinically significant since uncorrected strictures or fistulas can result in postoperative symptoms [6]. Solem et al. found that the sensitivity of WCE and CTEG in the detection of active small-bowel Crohn’s disease was identical (both 83%), and there was no significant difference with ileocolonoscopy (74%) or small-bowel followthrough (65%) [7]. Hara et al., investigating disease activity, showed that findings at CTEG correlated with symptoms in 80% of patients, indicating CTEG’s excellence as a monitoring test [8]. In a study by Higgins et al., clinicians suspected only 84% of CTEG-identified strictures and CTEG excluded strictures in >50% of patients with a clinically suspected stricture [9]. Furthermore, a survey of clinicians showed that, following the use of CTEG, clinical management changes 50% of the time [10].

contrast, and 1800 mL of neutral-density contrast. With this approach, the majority of patients with a source of OGIB were detected compared with WCE, surgery, or follow up. Ten of 22 (43%) studies showed positive findings, including the detection of angioectasias [12]. Huprich et al. emphasized that, while three-phase CTEG can be done with the goal of detecting angioectasias and other arterialphase dominant lesions, the radiation incurred is substantial (effective dose 59 mSv per exam) such that the relative risks and benefits have to be weighed, including patient age, necessity of repeat examinations, and the known risk of radiation with multiple CT exams, especially in younger patients [12, 13]. A follow-up prospective study using three-reader comparisons, presented by Huprich at RSNA 2009, indicated that two out of three phases are often adequate, but conclusive results have not been published. In a similar study by Hara, using three-phase CTEG, 33% of lesions were detected (specificity 89%) and 52% were detectable in retrospect. Some of the missed lesions were in the stomach and colon, emphasizing the advantage of CTEG in depicting abnormalities throughout the GI tract. Currently, the conditions that appear to be undetectable by CTEG, as proven by other tests, include ulcers, vascular lesions, and non-bleeding lesions [14].

Obscure Gastrointestinal Bleeding

Neoplasms

In the search for sources of gastrointestinal (GI) bleeding, CTEG has found increasing use since it may be a more sensitive, non-invasive method – using intravenous contrast and CT angiography techniques – made possible by neutral-density oral contrast. However, its application is limited to being a diagnosis-only test, with no therapeutic capability. Abstracts from the 2009 Radiological Society of North America (RSNA) meeting in which the detection of bleeding rates using various techniques was compared suggested equal sensitivity with tagged RBC and catheter digital subtraction angiography, such that bleeds of as little as 0.3 mL/min were detected. The search for bleeding sources in the small bowel may be subsequent to an apparently negative upper- and lower-GI endoscopy. It should be kept in mind that missed gastric and colonic pathology may still be discovered at imaging if these areas are well-filled, since endoscopic examinations are performer-dependent, imperfect gold-standards. Conventional barium studies do not identify mucosal erosions or vascular ectasia and have yields of 6% for small-bowel follow-through and 10% for small-bowel enteroclysis [11]. Angioectasias are the most common cause of OGIB in patients over the age of 50. These lesions are typically small, may be multiple, and are sessile or slightly raised. They are typically only visible at endoscopy or potentially at arterial-phase catheter-based or CT angiography. A major limitation of CTEG is its relative insensitivity to these small, flat, vascular lesions. The most recent study used optimized thin-sections, three scanning phases (arterial, enteric, and delayed), rapid boluses of intravenous

The accuracy of CTEG in the detection of masses such as carcinoid, adenocarcinoma, lymphoma, gastrointestinal stromal tumor, and polyps is not known. Capsule endoscopy, an examination being used more frequently to examine the small bowel, has several documented limitations, such that CTEG plays an important ongoing role for small-bowel mass detection and is especially important for masses that may have a predominant extraluminal component [15]. A recent RSNA 2009 Abstract indicated that masses in the setting of OGIB were better detected at CTEG than at WCE. Hyperenhancing lesions (e.g., some melanoma metastases or carcinoid tumors) will be easily detected with careful observation and windowing; however, the detection of isoattenuated masses (some melanoma metastases, polyps, and even some primary adenocarcinomas) and smaller masses may require excellent luminal distention and perusal for secondary findings, such as increased wall thickening, increased luminal caliber, or complications (intussusception, obstruction). CTEG may underestimate the number of lesions due to their smaller size or the lower degree of vascularity of small tumors.

Computed Tomography Enteroclysis This fused test combines the advantages of enteral challenge from a catheter small-bowel examination (enteroclysis) with the isotropic, multiplanar, cross-sectional images obtained at helical CT. Indications for CTEC include small-bowel obstruction, small-bowel masses, OGIB, Crohn’s disease, and malabsorption.

30

Pitfalls of Computed Tomography Enteroclysis General Careful attention to technique, including the use of infusion rates that allow uniform bowel distention; catheter placement in the proper location, with correct balloon insufflation and tip placement; the appropriate administration of pharmacological agents; and proper use of multidetector CT with reformatted images, will obviate many sources of misinterpretation. Spasm and inadequate filling should be easily recognized and addressed by adjustment of the pump flow-rate and use of spasmolytics (glucagon or Buscopan). Variable window and level settings are advised to adjust for high-attenuation contrast or, in the interpretation of a neutral-density exam, to appreciate subtle differences in enhancement of the bowel wall.

Crohn’s Disease In the diagnosis of suspected or known Crohn’s disease, a comparison of WCE with neutral-density oral contrast CTEC (nCTEC) in 56 patients showed that 27 could not undergo WCE due to strictures (≤10 mm) for fear of capsule retention. In the other 41 patients, the limitations of nCTEC included its inability to detect very early, minimal inflammatory changes or small mucosal lesions such as villous denudation, aphthoid ulcerations, or erosions. These were far better detected by WCE than CTEC (p = 0.004) [16-18]. Since the detection of non-elevated or non-depressed lesions in Crohn’s disease and superficial erosions in NSAID enteropathy will be limited, these entities would probably be best investigated using others methods, such as push enteroscopy, single- or double-balloon endoscopy, WCE or “air” (C02) double-contrast fluoroscopic enteroclysis without CT [19, 20]. In addition, in late-stage Crohn’s disease, nCTEC may show fewer fistulae than positive oral contrast CTEC (pCTEC), as the higher-density contrast may fill the GI tract more conspicuously.

Masses In the workup of small-bowel masses or OGIB, several pitfalls may be encountered. False-positive masses may be seen in patients with Kerkring fold thickening or even transient intussusception [21]. Unfortunately, the findings may be so convincing as to necessitate surgery to prove the lack of a mass. In Pilleul’s study of 219 patients with possible small-bowel neoplasms, the overall accuracy was 84.7%. There were five false-positive masses (2.3%) ranging from 6 to 25 mm in size. In two of these, small-bowel fold thickening was found at surgery and in the others no mass could be detected [22]. CTEC may also fail to identify jejunal polyps 2.6 mm [31]. As a result of the inflammatory hyperemia, the wall of the diseased appendix shows hyperenhancement during the arterial phase of contrast administration. The wall is circumferentially and asymmetrically thickened (usually 1-3 mm). Periappendiceal inflammation, the hallmark of appendicitis, is characterized by increased hazy density or linear fat stranding in the adjacent mesoappendix, by fluid-containing abscesses, and by ill-defined, heterogeneous soft-tissue densities representing a phlegmon [26-28]. There may be secondary inflammatory and edematous changes – with thickening of the wall of the adjacent ileum and cecum – that may mimic primary ileocolic inflammatory disease.

42

Fig. 5 Acute appendicitis. Ultrasound shows a thickened, fluidfilled appendix, surrounding infiltration and free fluid

Jaap Stoker, Richard M. Gore

The combination of right lower quadrant inflammation, phlegmon, and an abscess adjacent to the cecum is suggestive but not diagnostic of appendicitis. Indeed, if an abnormal appendix or an appendicolith is not shown, the differential diagnosis also must include Crohn’s disease, cecal diverticulitis, ileal diverticulitis, perforated cecal or appendiceal carcinoma, and pelvic inflammatory disease. Abscesses may be found in locations distant from the cecum because of the length and position of the appendix and the patterns of fluid migration in the peritoneal cavity. It should be noted that (depending on referral patterns) the majority (60-70%) of patients with suspected appendicitis who are referred for cross-sectional imaging do not have this disease. Instead, most of these patients have benign, self-limited gastrointestinal disorders such as viral gastroenteritis. CT and US often can suggest a specific alternate diagnosis [26-28]. Adnexal cysts, masses, salpingitis, and tubo-ovarian abscesses are readily shown on US. Ureteral calculi and pyelonephritis can be detected on CT and US. Enlarged lymph nodes in the right lower quadrant suggest mesenteric adenitis or infectious ileitis; mural thickening of the terminal ileum can be seen in Crohn’s disease or infectious ileitis. Although CT has a higher accuracy than US, the latter technique can be used initially, to reduce radiation exposure. Only those patients with a negative or inconclusive examination should proceed to CT [34]. Another approach is to use MRI instead of CT, which combines the benefits of a lack of ionizing radiation exposure with high-contrast resolution cross-sectional imaging [35, 36]. Initial reports on MRI in diagnosing acute appendicitis are encouraging, particularly concerning pregnant women (Fig. 7) [37, 38].

Fig. 6. Acute appendicitis. Coronal reformatted CT image shows a thickened, fluid-filled appendix (arrow) along the lateral aspect of the right psoas muscle

CT after intravenous contrast medium administration is the preferred technique, as it facilitates identification of the inflamed appendix and suggests alternative diagnoses. In patients with (imminent) renal insufficiency, a non-contrast CT can be performed. The diagnosis of acute appendicitis on non-contrast CT scans requires the detection of a thickened appendix (diameter >6 mm) with associated inflammatory changes in the periappendiceal fat or abnormal thickening of the right lateroconal fascia, with or without a calcified appendicolith. The addition of coronal and sagittal reformatted images increases diagnostic confidence by virtue of the more reliable demonstration of the entire appendix, surrounding fat and lymph nodes, and periappendiceal infection and inflammation [32]. Confidence in image interpretation also improves with the use of thinner reconstruction sections [33].

Fig. 7. Acute appendicitis. Coronal HASTE (half-Fourier acquisition single-shot turbo spin-echo) in a 28-year-old woman at 18 weeks gestation, clinically suspected of having appendicitis but in whom ultrasound was non-diagnostic. MRI shows a thickened retrocecal appendix (arrow) with increased signal intensity and minimal infiltration of the surrounding fat. The MRI diagnosis of appendicitis was confirmed at surgery

43

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

Scan times are short, namely 10-15 min. At present, however, MRI is not widely used in the Emergency Department for the diagnostic work-up of patients with acute abdominal pain, due to a lack of availability and expertise and the uncertainly as to its cost-effectiveness. Further studies should be directed at better defining the role of MRI in acute abdominal pain, especially its role compared to US and CT. Nonetheless, in pregnant women with an inconclusive US, MRI is currently the preferred imaging technique.

Diverticulitis It is estimated that 10-25% of individuals with diverticulosis will suffer from episodes of peridiverticular inflammation during their lifetime. In the USA, this complication accounts for approximately 200,000 hospitalizations and a health-care expenditure of four billion dollars annually. Among the patients who are hospitalized, 10-20% require emergency surgery [39]. Clinical signs indicative of diverticulitis are inaccurate, although the combination of direct tenderness only in the left lower quadrant, the absence of vomiting, and an elevated C-reactive protein level is suggestive in 25% of these patients (Laméris et al., personal communication). Inflammatory change in the pericolic fat (Fig. 8) is the hallmark of diverticulitis on CT and is seen in 98% of patients with the disease. The extent of the inflammatory reaction is related to the size of the perforation, degree of bacterial contamination, and the host response. Mild cases may manifest as areas of slight in-

crease in the density of the fat adjacent to the involved colon or as fine linear stranding with small fluid collections or bubbles of extraluminal air. In sigmoid diverticulitis, the fluid is typically decompressed into the inferior interfascial plane. Due to the hypervascularity of the inflamed area, contrast-enhanced CT scans often reveal engorged mesenteric vessels in the involved pericolic fat. Pericolic heterogeneous soft-tissue densities representing phlegmons and partially loculated fluid collections indicating abscess are seen in more severe cases. The abscess cavities usually contain air bubbles or air-fluid levels. They develop within the sigmoid mesocolon or are sealed off by the sigmoid colon and adjacent small bowel loops. Less commonly, they may form in the groin, flank, thigh, psoas muscle, subphrenic space, or liver [40]. On CT, diverticula are seen at the site of perforation or adjacent to it in about 80% of cases. They appear as small outpouchings of air, contrast, or fecal material projecting through the colonic wall. Symmetrical mural thickening of the involved colon of approximately 4-10 mm is found in about 70% of cases; however, if there is marked muscular hypertrophy, the wall of the colon can measure up to 2-3 cm in thickness. CT can also demonstrate intramural abscesses and fistula, and is helpful in patients with suspected colovesical fistulas. In the latter case, a pericolic inflammatory mass involves the bladder wall; the presence of intraluminal gas confirms the diagnosis. CT has a reported sensitivity of up to 98% in the diagnosis of diverticulitis [40]. Additionally, it can demonstrate disease extent, such as abscess and peritonitis remote from the colon, and can guide percutaneous abscess drainage. The diagnosis of other pathological conditions that may clinically simulate diverticulitis can also be achieved with CT. Although the accuracy of US in most prospective studies is not inferior to that of CT [41], it has its limitations and a recent large cohort study demonstrated the superiority of CT. Moreover, CT is more accurate in indicating alternative diagnoses [41] and provides a better overview of disease extent, which is important for clinical management (Hinchey classification). An initial study demonstrated that MRI is also accurate in diagnosing diverticulitis [42], but further studies are needed to evaluate its role. The major advantage of MRI in middle-aged patients with diverticulitis is avoidance of intravenous contrast medium and thus of contrastinduced nephropathy.

Epiploic Appendagitis

Fig. 8. Acute diverticulitis. Coronal reformatted CT image shows mural thickening of the sigmoid colon, inflammatory changes, and a gas bubble (arrow) in the sigmoid mesocolon

Primary epiploic appendagitis is a relatively uncommon condition that results from acute ischemia and inflammation of the appendices epiploicae. This disorder is often associated with torsion and infarction of these appendices and can simulate diverticulitis if it occurs in the sigmoid

44

Fig. 9. Epiploic appendagitis at the junction of the descending colon and sigmoid colon. A fatty density mass is surrounded by increased attenuation. Thrombosed vessels (arrow) can be seen in the central portion of the epiploic appendage

and appendicitis if located in the proximal colon. A characteristic appearance of a small, round, or oval fatcontaining mass with an associated inflammatory reaction of the pericolic fat is seen on US and CT (Fig. 9) [43]. The thrombosed vessels within the affected appendage epiploica can sometimes be visualized. Epiploic appendagitis is a self-limited process with clinical resolution in a few days. Follow-up CT examination may show total resolution, shrinkage, and eventual calcification of the inflamed and infarcted epiploic appendix.

Acute Abdominal Pain Inflammatory conditions of the colon are the most frequent causes of acute abdominal pain requiring urgent treatment (e.g., appendicitis, diverticulitis). In the majority of patients with acute abdominal pain, the clinical diagnosis is unclear or incorrect such that imaging is mandatory [44]. In this setting, imaging tests give different results when used in a setting of a specific disease (e.g., acute appendicitis) than when referring to all patients with acute abdominal pain (a myriad of diagnoses). Thus, papers reporting on a specific diagnosis are less informative than those relatively scarce papers reporting on patients with acute abdominal pain in general. Plain abdominal radiographs have a low accuracy in this setting and result in management changes in only 4% of the patients [45, 46]. In the majority of patients, further imaging is warranted after plain radiography. US was shown to increase the

Jaap Stoker, Richard M. Gore

diagnostic accuracy of clinical evaluation (from 70 to 83%) and resulted in changes in management in 22% of patients [47-49]. CT has a significant impact on the diagnosis, as shown in one series in which the accuracy improved from 71% in the pre-CT clinical diagnosis to 93% in the post-CT diagnosis, with a change in management of 46% [50]. Also, the level of confidence increases significantly with CT [51]. In two randomized controlled trials from the same institution, CT was studied: (1) within 24 h versus routine workup by plain X-ray or (2) when considered necessary, US, CT, or fluoroscopy and CT within 1 h versus routine workup. The first study demonstrated significantly fewer deaths (0 vs. 7; p=0.014) [52]. Length of hospital stay was not significantly different but with CT there was a significant overstatement of serious diagnoses. In the second study, routine CT significantly improved diagnostic certainty, but there were no other significant differences to routine workup in patient management [53]. The overall inter-observer agreement of abdominal CT is good [54], and for specific urgent diagnoses it is very good (e.g., for appendicitis, diverticulitis, and bowel obstruction the values are 0.84, 0.90, and 0.81, respectively). A cohort study evaluated 11 diagnostic strategies of clinical diagnosis, plain radiography (supine abdominal X-ray and upright chest X-ray), US, and CT in 1021 patients with acute abdominal pain [34]. Clinical diagnosis (performed by surgical residents) had a sensitivity of 88% and a specificity of 41% for urgent diagnoses. Plain radiography did not contribute to a higher sensitivity or specificity for urgent cases of acute abdominal pain. Regarding the clinical diagnosis, US reduced the number of false-positive urgent diagnoses to 85%, but at the expense of a lower sensitivity (70%, thus missing 30% of the urgent conditions). CT had the highest sensitivity (89%) and specificity (77%) as a single imaging strategy. The best strategy for the detection of urgent diagnoses was an initial US, reserving CT for patients with negative or inconclusive US examinations [34]. This strategy led to a sensitivity for urgent diagnosis of 94%. Initial US reduced CT use and associated radiation burden by 51% compared to CT in all patients. Strategies driven by body mass index, age, or location of the pain had a lower sensitivity for urgent diagnoses than achieved with this conditional strategy. The use of MRI in acute abdominal pain primarily has been studied in acute appendicitis and diverticulitis [36].

Perianal Fistulas Perianal fistulas mostly occur either as the result of fistulous disease originating from the anal glands near the anal crypts (cryptoglandular hypothesis) or in patients with Crohn’s disease [55]. Infection of the anal glands may result in abscess formation. This is a relatively common condition, with a prevalence of approximately 0.01%, predominantly affecting young adults. The course

45

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

of the fistula track may be simple and superficial or complicated. The latter may be intersphincteric (through the internal anal sphincter and then downward through the intersphincteric space) or trans-sphincteric (transversing not only the internal sphincter but also the external sphincter or puborectis muscle) or have a supralevator extent or an extrasphincteric extension to the rectum, without involvement of the anal sphincter. These complicated tracks need detailed imaging for proper therapy, as inadequate treatment may lead to recurrent disease. The surgeon must be aware both of the presence and number of tracks and of their extent, the location of the internal opening, and the presence of abscesses. Pre-operative evaluation of perianal fistulas may include physical examination, examination under anesthesia (EUA), endoscopic ultrasound (EUS), or MRI. Physical examination has significant shortcomings, especially in patients with recurrent disease. While EUA can be used for determining disease extent, immediately followed by treatment, it has limitations and disadvantages, mostly related to probing. Firstly, not all fistulas have an external opening that can be probed, and probing may miss secondary tracks. It is well-recognized that missed extensions are the commonest cause of recurrence, which reaches 25% in some series. Forceful probing may lead to perforation of the levator plate, worsening the extent of the disease. Patients with recurrent disease are most likely to harbor missed disease but are also the most difficult

to assess. Digital palpation frequently cannot distinguish between scarring due to repeated surgery and induration due to an underlying extension. EUS, which can be enhanced by hydrogen peroxide instillation within the track, may be used to determine disease extent. Although initial reports were encouraging, later studies have been less sanguine especially when EUS was compared to MRI. This discrepancy probably relates to operator expertise since EUS is highly operator-dependent. Insufficient penetration beyond the external sphincter, especially with high-frequency transducers, limits the ability of EUS to resolve ischioanal and supralevator sepsis, with the result that it may miss extensions from the primary tract. This technique is impressive in demonstrating the internal opening whereas infection is distinguished only with difficulty from postoperative fibrosis; however, hydrogen peroxide instillation facilitates this differentiation. Low simple tracks presumably can be identified by EUS as accurately as by MRI, but the latter is definitely superior in cases involving complex or high tracks [56]. MRI has been proven to provide the most comprehensive assessment of patients with perianal fistulas, facilitating accurate identification of tracks and extensions as well as abscesses. MRI examination for perianal fistulas should include T2-weighted sequences in multiple planes, a fat-saturation sequence, and a contrastenhanced (fat-saturation) T1-weighted sequence [57]. Tracks (Figs. 10, 11) are identified on T2 as hyper-

a

b

a

b

Fig. 10 a, b. Crohn’s disease and perianal fistulas. a Axial T2-weighted turbo spin-echo demonstrates multiple tracks (arrowheads), intersphincteric and outside the anal sphincter. On the right, an abscess (A) is seen in the ischioanal space. b Coronal T2-weighted turbo spin-echo demonstrates a trans-sphincteric track in the ischioanal space (arrowhead) and extension of the abscess (A) in the levator ani at both sides and in the rectal wall (arrow)

Fig. 11 a, b. Cryptoglandular perianal fistula. a Coronal T2-weighted turbo spin-echo with endoanal coil shows a fibrous track (arrowhead) extending through the left external sphincter. Normal sphincter anatomy is seen on the right. E External sphincter, I internal sphincter, PR puborectal muscle, LA levator ani plate. b Axial T2-weighted turbo spinecho demonstrates the hyperintense track (arrow), surrounded by fibrous tissue (arrowhead), along the left external sphincter (E)

46

intense longitudinal structures, often with a hypointense, fibrous wall. Collateral inflammation is often appreciated on fat saturation sequences. After the administration of intravenous contrast medium, the lining of the wall will enhance. Non-enhancing fluid can be identified in the center of the track or the track can be completely obliterated by granulation tissue. In the latter case, there is complete enhancement of the part of the track that is hyperintense on T2 sequences (Fig. 11). Abscesses are readily appreciated on fat saturation sequences, although some small fluid collections may be more difficult to identify. The use of external phased array coils may result in limitations in detecting superficial extensions and difficulty in locating the precise level of the internal opening. In such cases, endoluminal MRI may provide more information (Fig. 11). A prospective triple-blinded comparison of the accuracy of anal endosonography (AES), pelvic MRI, and surgical EUA in perianal Crohn’s disease showed that AES correctly classified fistulas in 91% of the cases, compared with 87% for pelvic MRI and 91% for surgical evaluation [58]. A combination of any two of the three modalities increased the accuracy to 100%. Another study, in which MRI and AES were compared with surgical findings, showed MRI to be superior to AES in fistula classification, with sensitivities of 84 vs. 60% for the two modalities, and specificities of 68 and 21%, respectively. Several studies have indicated a positive effect of preoperative MRI on patient outcome. In one study, the therapeutic effect of MRI before EUA was 21.1% [59]. Additionally it was shown that disease recurrence after surgery could be reduced by about 75% if surgery was guided by the MRI findings. The differential diagnosis of perianal fistulas primarily concerns fistulas originating from skin appendages: acne conglobata, suppurative hidranitis, and pilonidal sinus. The first two disorders are easy to recognize clinically, but this is more difficult with pilonidal sinus. Imaging can be used to differentiate between perianal fistula and pilonidal sinus. In a study in seven patients with pilonidal sinus and 14 sex- and age-matched individuals with perianal fistulas, these conditions could be readily discriminated by the absence in the former of intersphincteric sepsis or an enteric opening [60]. Osteomyelitis of the pelvis or femur may give rise to abscesses and tracks that extend to the anorectal region, whereas osteomyelitis is a rare finding in perianal fistulas due to Crohn’s disease. Differentiating these two conditions is usually not difficult, as the predominant disease localization (either extensive bone marrow edema or extensive tracks with intersphincteric extension and internal opening) will establish the diagnosis.

References 1. Gore RM, Laufer I, Berlin JW (2008) Ulcerative and granulomatous colitis: idiopathic inflammatory bowel disease. In: Gore RM, Levine MS (eds) Textbook of gastrointestinal radiology, 3rd edn. Saunders, Philadelphia, pp 1071-1109

Jaap Stoker, Richard M. Gore

2. Markose G, Ng CS, Freeman AH (2003) The impact of helical computed tomography on the diagnosis of unsuspected inflammatory bowel disease in the large bowel. Eur Radiol 13:107-113 3. Furukawa A, Saotome T, Yamasaki M et al (2004) Cross-sectional imaging in Crohn disease. Radiographics 24:689-702 4. Horsthuis K, Bipat S, Bennink R, Stoker J (2008) Inflammatory bowel disease diagnosed with US, MR, scintigraphy, and CT: Meta analysis of prospective studies. Radiology 247:64-79 5. Johnson KT, Hara AK, Johnson CD (2009) Evaluation of colitis: usefulness of CT enterography technique. Emerg Radiol 16:277-282 6. Desmond AN, O’Regan K, Curran C et al (2008) Crohn’s disease: factors associated with exposure to high levels of diagnostic radiation. Gut 57:1524-1529 7. Ripolles T, Martinez MJ, Pardes JM et al (2009) Crohn disease: Correlation of findings at contrast-enhanced US with severity at endoscopy. Radiology 253:241-248 8. Migaleddu V, Scanu AM, Quaia E et al (2009) Contrast-enhanced ultrasonographic evaluation of inflammatory activity in Crohn’s disease. Gastroenterology 137:43-61 9. Ziech M, Stoker J (2010) MRI of the small bowel: enterography. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 117-134 10. Papanikolaou N, Gourtsoyianni S (2010) MRI of the small bowel: enteroclysis. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 135-148 11. Rimola J. Rodriguez S, Gracia-Bosch O et al (2009) Role of 3.0-T colonography in the evaluation of inflammatory bowel disease. Radiographics 29:701-719 12. Ajaj W, Lauenstein TC, Langhorst J et al (2005) Small bowel hydro-MR imaging for optimized ileocecal distension in Crohn’s disease: should an additional rectal enema filling be performed? J Magn Reson Imaging 22:92-100 13. Zijta F, Stoker J (2010) Magnetic resonance imaging of the colon (Colonography): Results. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 185-204 14. Ajaj WM, Lauenstein TC, Pelster G et al (2005) Magnetic resonance colonography for the detection of inflammatory diseases of the large bowel: quantifying the inflammatory activity. Gut 54:257-263 15. Horsthuis K, Bipat S, Stokkers PC, Stoker J (2009) Magnetic resonance imaging for evaluation of disease activity in Crohn’s disease: a systematic review. Eur Radiol 19:1450-1460 16. Röttgen R, Herzog H, Lopez-Häninnen E, Felix R (2006) Bowel wall enhancement in magnetic resonance colonography for assessing activity in Crohn’s disease. Clin Imaging 30:27-31 17. Langhorst J, Kühle CA, Ajaj W et al (2007) MR colonography without bowel purgation for the assessment of inflammatory bowel diseases: diagnostic accuracy and patient acceptance. Inflamm Bowel Dis 13:1001-1008 18. Rimola J, Rodríguez S, García Bosch O et al (2009) Magnetic resonance for assessment of disease activity and severity in Crohn disease. Gut 58:1113-1120 19. Kelly CP, LaMont JT (2008) Clostridium difficile – more difficult than ever. N Engl J Med 359:1932-1940 20. Horton KM, Corl FM, Fishman EK (2000) CT evaluation of the colon: inflammatory disease. Radiographics 20:399-418 21. Turner DR, Markose G, Arends MJ et al (2003) Unusual causes of colonic wall thickening on computed tomography. Clin Radiol 58:191-200 22. Gluecker TM, Williamson EE, Fletcher JG et al (2003) Diseases of the cecum: a CT pictorial review. Eur Radiol 13 Suppl 4:L51-61 23. Kirkpatrick ID, Greenberg HM (2003) Gastrointestinal complications in the neutropenic patient: characterization and differentiation with abdominal CT. Radiology 226:668-674 24. Thoeni RF, Cello JP (2006) CT imaging of colitis. Radiology 240:623-638

47

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

25. Raman SS, Osuagwu FC, Kadell B et al (2008) Effect of CT on false positive diagnosis of appendicitis and perforation. N Engl J Med 358:972-973 26. Pinto Leite N, Pereira JM, Cunha R et al (2005) CT evaluation of appendicitis and its complications: imaging techniques and key diagnostic findings. AJR Am J Roentgenol 185:406-417 27. Daly CP, Cohan RH, Francis IR et al (2005) Incidence of acute appendicitis in patients with equivocal CT findings. AJR Am J Roentgenol 184:1813-1820 28. Paulson EK, Harris JP, Jaffe TA et al (2005) Acute appendicitis: added diagnostic value of coronal reformations from isotropic voxels at multi-detector row CT. Radiology 235: 879-885 29. Puylaert JB, Rutgers PH, Lalisang RI et al (1987) A prospective study of ultrasonography in the diagnosis of appendicitis. N Engl J Med 317:666-669 30. van Randen A, Bipat S, Zwinderman AH et al (2008) Acute appendicitis: meta-analysis of diagnostic performance of CT and graded compression US related to prevalence of disease. Radiology 249:97-106 31. Moteki T, Ohya N, Horikoshi H (2009) Prospective examination of patients suspected of having appendicitis using new computed tomography criteria including “maximum depth of intraluminal appendiceal fluid greater than 2.6 mm. J Comput Assist Tomogr 33:383-389 32. Kim YJ, Kim J-E, Kim HS, Hwang HY (2009) MDCT with coronal reconstruction: clinical benefit in evaluation of suspected acute appendicitis in pediatric patients. AJR Am J Roentgenol 192:150-152 33. Johnson PT, Horton KM, Kawamoto S et al (2009) MDCT for suspected appendicitis: effect of reconstruction section thickness on diagnostic accuracy, rat of appendiceal visualization, and reader confidence using axial images. AJR Am J Roentgenol 192:893-901 34. Laméris W, van Randen A, van Es HW et al (2009) Imaging strategies for detection of urgent conditions in patients with acute abdominal pain: diagnostic accuracy study. BMJ 338:b2431 35. Lee KS, Pedrosa I (2010) Magnetic resonance imaging of acute conditions of the gastrointestinal tract. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 283-314 36. Stoker J (2008) Magnetic resonance imaging and the acute abdomen. Br J Surg 95:1193-1194 37. Oto A, Ernst RD, Ghulmiyyah LM et al (2009) MR imaging in the triage of pregnant patients with acute abdominal and pelvic pain. Abdom Imaging 34:243-250 38. Cobben L, Groot I, Kingma L (2009) A simple MRI protocol in patients with clinically suspected appendicitis: results in 138 patients and effect on outcome of appendectomy. Eur Radiol 19:1175-1183 39. Humes DJ, Solyamani-Dodaran M, Fleming KM et al (2009) A population-based study of perforated diverticular disease incidence and associated mortality. Gastroenterology 136:11981205 40. Kircher MF, Rhea JT, Kihiczak D, Novelline RA (2002) Frequency, sensitivity, and specificity of individual signs of diverticulitis on thin-section helical CT with colonic contrast material: experience with 312 cases. AJR Am J Roentgenol 178:1313-1318 41. Laméris W, van Randen A, Bipat S et al (2008) Graded compression ultrasonography and computed tomography in acute

42.

43. 44. 45.

46. 47. 48. 49. 50. 51.

52.

53.

54. 55. 56. 57. 58.

59. 60.

colonic diverticulitis: meta-analysis of test accuracy. Eur Radiol 18:2498-2511 Heverhagen JT, Sitter H, Zielke A, Klose KJ (2008) Prospective evaluation of the value of magnetic resonance imaging in suspected acute sigmoid diverticulitis. Dis Colon Rectum 51:1810-1815 Singh AK, Gervais DA, Hahn PF et al (2005) Acute epiploic appendagitis and its mimics. Radiographics 25:1521-1534 Stoker J, van Randen A, Laméris W, Boermeester MA (2009) Imaging in acute abdominal pain. Radiology 253:31-46 MacKersie AB, Lane MJ, Gerhardt RT et al (2005) Nontraumatic acute abdominal pain: unenhanced helical CT compared with three view acute abdominal series. Radiology 237:114-122 Kellow ZS, MacInnes M, Kurzencwyg D et al (2008) The role of abdominal radiography in the evaluation of the nontrauma emergency patient. Radiology 248:887-893 Walsh PF, Crawford D, Crossling FT et al (1990) The value of immediate ultrasound in acute abdominal conditions: a critical appraisal. Clin Radiol 42:47-49 Allemann F, Cassina P, Rothlin M, Largiader F (1999) Ultrasound scans done by surgeons for patients with acute abdominal pain: a prospective study. Eur J Surg 165:966-970 Dhillon S, Halligan S, Goh V et al (2002) The therapeutic impact of abdominal ultrasound in patients with acute abdominal symptoms. Clin Radiol 57:268-271 Tsushima Y, Yamada S, Aoki J et al (2002) Effect of contrastenhanced computed tomography on diagnosis and management of acute abdomen in adults. Clin Radiol 57:507-513 Rosen MP, Sands DZ, Longmaid HE, 3rd et al (2000) Impact of abdominal CT on the management of patients presenting to the emergency department with acute abdominal pain. AJR Am J Roentgenol 174:1391-1396 Ng CS, Watson CJ, Palmer CR et al (2002) Evaluation of early abdominopelvic computed tomography in patients with acute abdominal pain of unknown cause: prospective randomised study. BMJ 325:1387 Sala E, Watson CJ, Beadsmoore C et al (2007) A randomized, controlled trial of routine early abdominal computed tomography in patients presenting with nonspecific acute abdominal pain. Clin Radiol 62:961-969 van Randen A, Lameris W, Nio CY et al (2009) Inter-observer agreement for abdominal CT in unselected patients with acute abdominal pain. Eur Radiol 19:1394-1407 Halligan S, Stoker J (2006) Imaging fistula-in-ano. Radiology 239:18-33 Ziech M, Felt-Bersma R, Stoker J (2009) Imaging of perianal fistulas. Clin Gastroenterol Hepatol 7:1037-1045 Horsthuis K, Stoker J (2004) MRI of perianal Crohn’s disease. AJR Am J Roentgenol 183:1309-1315 Schwartz DA, Wiersema MJ, Dudiak KM et al (2001) A comparison of endoscopic ultrasound, magnetic resonance imaging, and exam under anesthesia for evaluation of Crohn’s perianal fistulas. Gastroenterology 121:1064-1072 Buchanan G, Halligan S, Williams A et al (2002) Effect of MRI on clinical outcome of recurrent fistula-in-ano. Lancet 360:1661-1662 Taylor SA, Halligan S, Bartram CI (2003) Pilonidal sinus disease: MR imaging distinction from fistula in ano. Radiology 226:662-667

IDKD 2010-2013

CT Colonography: Updated Daniel C. Johnson1, Michael Macari2 1 Department 2 Department

of Radiology, Mayo Clinic, Scottsdale, AZ, USA of Radiology, New York University Langone School of Medicine, New York, NY, USA

Computed tomography (CT) colonography has been in development for more than a decade, with hundreds of articles now published on its performance and technical capabilities. With the conclusion and publication of the National CT Colonography trial [1] and endorsement of the technique for screening by a multi-society task force (including the American Cancer Society, American College of Radiology, US Multi-society Task Force on Colorectal Cancer) [2], the clinical validation of CT colonography in the prepared colon has been completed. This chapter highlights the most important current issues for CT colonography. Patient acceptance of routine colorectal screening, including CT colonography, remains a major barrier. In 2000, only 43% of US adults age 50 or older had undergone a sigmoidoscopy or colonoscopy within the previous 10 years or had used a fecal occult blood home test kit within the preceding year [3]. The major disincentive for patients considering CT colonography as a screening option is the laxative purgation (the same as that required for colonoscopy) [4]. Advantages include the lack of required sedation and intravenous line placement for CT colonography, a quick return to work following the examination, and no need to inconvenience others for transportation to and from the exam. The risk of perforation at CT colonography is considerably less than at colonoscopy. Furthermore, the examination only requires two breath-holds on the CT scanner (in the supine and prone positions), with the completion of most average examinations in 10 min, which may help reassure hesitant patients. Still, the reality of a full bowel preparation, an enema tip, and full (although brief) colonic insufflation is likely to delay the decision for screening for some people. The performance of CT colonography has undergone exhaustive testing. The Pickhardt trial demonstrated a sensitivity similar to colonoscopy [5], but concerns were raised that community practices might not be able to achieve these results. The National CT Colonography trial (ACRIN 6664) studied 2531 individuals nationally across 15 centers, including academic and private practices. The findings of this trial were similar to those of the Pickhardt trial and have reassured many groups [1]. Radiologist’s training and testing were required for the

ACRIN trial. Although some participants required more training than others, all of them received a passing score of 90% for easy and moderately difficult to detect lesions [6]. The ACRIN trial also insisted on strict adherence to protocol requirements, including stool tagging regimens, mechanical insufflation of the colon, and thin-section and low-dose CT techniques [7]. It is clear that meticulous attention to all aspects of the examination is required to achieve optimal results. Extracolonic abnormalities are common in patients of screening age [8-11]. A pragmatic approach to these findings is needed; for example, radiologists should recommend follow-up studies for those patients with findings most likely to be of clinical significance. Patients (and clinicians) will be grateful if additional testing is minimized; for those that need addition studies, the recommended optimal follow-up should be included in the report. It is unfortunate that the risk associated with the low radiation dose required for CT colonography has been misunderstood. The standard dose at CT colonography is about half of the dose used for a standard body CT examination. This results in an average dose of approximately 5 mSv. The real risk of this exposure is unknown, but the Health Physics Society has stated that for doses in this range the risks for the development of radiationinduced cancer are too small to measure or are nonexistent [12]. Even if a very small risk is assumed from radiation exposure at CT, it must be balanced against the risk of developing colon cancer and of other alternative procedures. The risk of perforation (1:1000) and death (1:17,000) at colonoscopy are real and can be measured [13], but as a society there is a consensus that these risks are outweighed by the risk of developing colon cancer (about 1 in 13 without screening) [14]. Maintaining high-quality interpretations is a responsibility that each individual, each practice, and our specialty should assume. The American College of Radiology has established a national CT colonography database within the National Radiology Data Registry (NRDR) [15]. Selected process and outcome metrics can be quickly entered on-line and compared to national benchmarks. These measures include process metrics related to the CT

CT Colonography: Updated

technique and the adequacy of patient preparation, and outcome metrics related to colon perforation, true-positive and false-positive rates for large (≥1 cm) polyps, and the prevalence of significant extracolonic findings. Practices seriously interested in providing the best care should be encouraged to participate in this data registry and manage their practice such that benchmark metrics are achieved. A spirit of cooperation between radiologists and gastroenterologists is needed for optimal patient care. Guidelines will need to be jointly developed for the proper use of colonography and colonoscopy, and for processes to efficiently transfer patients with polyps to colonoscopy. Those practices that are able to do this effectively will offer patients a service of high value – and will likely find themselves very busy. In summary, CT colonography has completed its clinical validation and is now ready for widespread clinical application. Radiologists committed to performing the examination to the highest quality must obtain the education and equipment needed. We must focus our efforts on the best in patient care, and ignore the political distractions that will come. We have an obligation to educate referring physicians on the correct use of the technique. Collaborations with gastroenterologists to ensure sameday polypectomy for selected patient will enhance patient care. Extracolonic findings must be vigilantly and properly reported so that only those patients with highly significant lesions are recommended for additional followup testing. Lastly, we should be committed to ongoing quality measures to both improve and maintain the highest standards of care. Radiology has another exciting opportunity to serve the public, and to potentially help reduce the incidence of a common cancer killer.

References 1. Johnson CD, Chen MH, Toledano A et al (2008) Accuracy of CT colonography for detection of large adenomas and cancer. N Engl J Med 359:1207-1217

49

2. Levin B, Lieberman DA, McFarland B et al (2008) Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology 134:1570-1595 3. Colorectal (Colon) Cancer. http://cdcgov/cancer/colorectal/ statistics/screening_rateshtm 4. Beebe TJ, Johnson CD, Stoner S et al (2007) Assessing attitudes toward laxative preparation in colorectal cancer screening and effects on future testing: potential receptivity to computed tomographic colonography. Mayo Clinic Proceedings 82:666-671 5. Pickhardt PJ, Choi JR, Hwang I et al (2003) Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. New Eng J Med 349:2191-2200 6. Fletcher JG, Johnson CD, Toledano A et al (2005) ACRIN 6664: Lessions for CT colonography (CTC) training and certification. Radiological Society of North America Scientific Assembly and Annual Meeting Program, Chicago, IL 7. Johnson CD, Chen MH, Toledano A et al. The National CT Colonography Trial Protocol, ACRIN 6664. http://wwwacrinorg/ Portals/0/Protocols/6664/Protocol-ACRIN%206664%20 Amendment%201,%207706pdf 8. Gluecker TM, Johnson CD, Wilson LA et al (2003) Extracolonic findings at CT colonography: evaluation of prevalence and cost in a screening population. Gastroenterology 124:911916 9. Hara AK, Johnson CD, MacCarty RL, Welch TJ (2000) Incidental extracolonic findings at CT colonography. Radiology 215:353-357 10. Hellstrom M, Svensson MH, Lasson A (2004) Extracolonic and incidental findings on CT colonography (virtual colonoscopy). AJR Am J Roentgenol 182:631-638 11. Rajapaksa RC, Macari M, Bini EJ (2004) Prevalence and impact of extracolonic findings in patients undergoing CT colonography. Journal of Clin Gastroenterol 38:767-771 12. Radiation risk in perspective (2004) Position Statement of the Health Physics Society 13. Waye JD, Kahn O, Auerbach ME (1996) Complications of colonoscopy and flexible sigmoidoscopy. Gastrointest Endosc Clin N Am 6:342-377 14. Lifetime Probability of Developing or Dying From Cancer. http://wwwcancerorg/docroot/CRI/content/CRI_2_6x_Lifetime_Probability_of_Developing_or_Dying_From_ Cancerasp?sitearea=&level= 15. National Radiology Data Registry. https://nrdracrorg/portal/ Nrdr/Main/pageaspx

IDKD 2010-2013

Imaging of Diffuse and Inflammatory Liver Diseases Pablo R. Ros1, Rendon C. Nelson2 1 Department 2 Department

of Radiology, University Hospitals Health System, Case Western Reserve University, Cleveland, OH, USA of Radiology, Duke University, Durham, NC, USA

Introduction The category of diffuse liver diseases includes a variety of disorders that typically involve the liver in a non-focal fashion. It is important to note, however, that even within this group there may be focal abnormalities, representing unusual expression of a disease that typically has a diffuse manifestation. At the same time there may be focal neoplasms, some of which are benign and others malignant. With recent advances in cross sectional imaging, the detection, characterization, and follow-up of diffuse liver disease has been greatly facilitated. This chapter is divided along the traditional classification of cirrhosis, vascular disorders, congenital, metabolic and storage, and neoplastic diseases. In addition, diffuse and focal inflammatory/infectious diseases are discussed.

Fibrous Tissue Deposition (Cirrhosis) Background The preliminary events that lead to cirrhosis involve a mechanism of injury whereby there is repeated exposure of some noxious agent to the liver, resulting in hepatocyte injury and/or death. Etiological noxious agents include alcohol, viral infection (specifically hepatitis B and hepatitis C), non-alcoholic steatohepatitis, autoimmune disorders (such as primary sclerosing cholangitis and primary biliary cirrhosis), and toxic agents (such as aflatoxin and iron, specifically, primary hemochromatosis). Pathologically, the liver attempts to repair itself from the injury by regenerating hepatocytes and depositing collagen. With repeated exposure to the noxious agent, there is ongoing hepatocyte regeneration and collagen deposition, resulting in the formation of regenerative nodules and, eventually, the formation of fibrous scars, respectively. Note that fibrous tissue deposition, specifically in the form of collagen, can be treated and is reversible to a point. However, with more severe or profound degrees of collagen and fibrous tissue deposition, this process is no longer reversible and is henceforth referred to as cirrhosis. Con-

siderable research has been devoted to the non-invasive quantification of fibrosis using ultrasound (US) and magnetic resonance imaging (MRI). Some of these techniques use elastography, in which a mechanical impulse is used to push or distort the liver followed by measurement of the amount of liver movement. A stiffer, more fibrotic liver will move less than a soft, less fibrotic liver.

Regenerative Nodules Regenerative nodules are classified as micro-regenerative or macro-regenerative. Micro-regenerative nodules measure 12 cm or an anteroposterior dimension >9 cm is considered enlarged. In some patients with splenomegaly, focal iron deposits can be appreciated within the parenchyma of the spleen. These are seen on MRI as hypointense foci on either T1-weighted gradient

a

b

c

d

52

Pablo R. Ros, Rendon C. Nelson

echo or T2-weighted pulse sequences and are referred to as Gamna-Gandy bodies. Porto-systemic shunts can occur either intrahepatically or extrahepatically. Intrahepatic porto-systemic shunting is a result of high-pressure, low-volume arterial blood that mixes at the level of the hepatic sinusoids with lowpressure, high-volume venous blood. In the cirrhotic liver, arterial flow is increased and portal venous flow is decreased, thereby increasing the prevalence of these shunts. During the arterial phase of imaging, there is typically a parenchymal blush that is associated with early portal venous enhancement. Extrahepatic porto-systemic shunts are also seen in the setting of cirrhosis. The most common is a spontaneous spleno-renal shunt whereby venous blood from the spleen is shunted to the left renal vein via the left inferior adrenal vein. In this setting, a prominent collateral vein is usually seen draining directly into an enlarged left renal vein. Porto-systemic varices are also common in patients with cirrhosis and portal hypertension. These include esophageal varices, splenic hilar varices, perigastric varices, and peripancreatic varices. In some patients, profound shunting of blood from the splenic vein retrograde into the inferior mesenteric vein precludes the development of splenomegaly.

Ancillary Findings in Cirrhosis Other findings that often occur in the setting of cirrhosis include ascites, mesenteric edema, and bowel wall edema, particularly in the small bowel. Some of these changes are due to portal hypertension although they may be exacerbated by hypoproteinemia.

a

c

Vascular Disorders Arterial-Portal Shunts Although some arterial-portal shunts appear to arise spontaneously, many of them are due to prior liver intervention, such as a biopsy, or in the setting of cirrhosis, as noted above. Furthermore, they are often associated with other vascular malformations, such as focal nodular hyperplasia and cavernous hemangiomas. On contrast-enhanced CT or MRI, findings include a focal or wedge-shaped parenchymal blush, a large hepatic artery, early portal venous enhancement, and a possible increase in the luminal diameter of the portal vein (Fig. 2). Most of these arterial portal shunts are subclinical and do not require intervention.

Perfusion Abnormalities Perfusion abnormalities are common following the administration of contrast agents. They are seen as focal, often web-shaped, areas of hyperenhancement that are most pronounced during the late hepatic arterial phase, hence the term transient hyperattenuation defect (THAD) for CT and transient hyperintensity defect (THID) for MRI. Although the etiology of these defects is not always apparent, many of them are due to alterations in portal venous flow. For example, they are often visualized in the setting of liver tumors (either primary or metastatic) that block flow from a branch of the portal vein. They can also be seen when there are aberrant veins draining directly into the liver, leaving parenchyma that receives venous blood from the

b

d

Fig. 2 a-d. Arterial-portal shunt in a 49-year-old man with hepatitis C. Axial images through the liver during the late hepatic arterial phase demonstrate a wedgeshape area of focal parenchymal hyperenhancement in the left hepatic lobe (a), a large hepatic artery (b) as well as early and vivid enhancement of a large periumbilical collateral vein (c, d). These findings are consistent with an intrahepatic arterial-portal shunt

53

Imaging of Diffuse and Inflammatory Liver Diseases

aberrant source only and not from the portal vein. These perfusion abnormalities typically occur: a) in the subcapsular area; b) in the periligamentous area (about both the fissure for the falciform ligament and the ligamentus venosum), due to aberrant internal mammary or periumbilical veins; c) about the porta hepatis, due to aberrant gastric veins (right more common than left); d) about the gallbladder fossa, due to aberrant cholecystic veins. On contrast-enhanced CT or MRI, they appear as transient areas of hyperenhancement during the late hepatic arterial phase and are often web-shaped and subcapsular. There is usually not an appreciable abnormality in this area on the unenhanced images or during the portal venous or equilibrium phases.

of the portal vein, they are uncommonly associated with portal vein thrombosis, either bland or malignant. On CT and MRI, occlusive malignant portal vein thrombosis typically enlarges the portal vein, the luminal diameter of which may exceed 23 mm. Furthermore, the thrombus itself may enhance and this enhancement can be quite vivid during the late hepatic arterial phase. With Doppler US, sampling of the thrombus itself reveals low-resistance arterial wave forms that often flow in the hepatofugal direction. In patients with cirrhosis and HCC who are anticipating liver transplantation, it is important to determine whether portal vein thrombosis is bland or malignant. If the thrombosis is indeed malignant, the patient cannot be considered a candidate for transplantation. The best way to confirm the diagnosis is to percutaneously biopsy the intraluminal thrombus itself under direct real-time US guidance.

Portal Vein Thrombosis

Budd-Chiari Syndrome

In portal vein thrombosis, the thrombus can be either bland or malignant and either occlusive or non-occlusive. Bland thrombosis typically occurs in the setting of trauma, cirrhosis, following an orthotopic liver transplant at the end-to-end anastomosis, in certain hypercoagulable states, and in 25% of patients with BuddChiari syndrome. If the thrombus is not occlusive, an eccentric intraluminal filling defect is identified, which often resolves with anti-coagulant therapy. In this case, there is no cavernous transformation (Fig. 3). With occlusive portal vein thrombosis, however, the imaging findings are different. Early on, the portal vein demonstrates a non-enhancing intraluminal filling defect that may distend the vein and increase the luminal diameter. Furthermore, there is often enhancement of the wall of the vein via the vaso vasorum. Over time, typically in the range of 3-6 weeks, the thrombus undergoes retraction, with the development of cavernous transformation, mainly via collaterals that develop in the vaso vasorum itself. Malignant portal vein thrombosis typically occurs in patients that have HCC, in the setting of either cirrhosis or, less commonly, non-cirrhosis. Interestingly, although metastases to the liver commonly obstruct small branches

Budd-Chiari syndrome may occur when there is obstruction to the outflow of blood from the hepatic veins. In Western countries, such as the USA, the majority of these cases (70%) are idiopathic. In the Orient, however, they are commonly due to congenital webs that occur within the hepatic veins themselves. Other disorders that are associated with Budd-Chiari syndrome include trauma, pregnancy, certain hypercoagulable states, and malignant hepatic vein thrombosis. The latter is commonly associated with primary tumors of the liver, right adrenal gland, and right kidney, specifically HCC, adrenal cortical carcinoma, and renal cell carcinoma, respectively. Budd-Chiari syndrome can occur with occlusion of one, two, or all three hepatic veins or with occlusion of the suprahepatic inferior vena cava. Furthermore, the imaging findings may differ depending upon whether the disease is acute or chronic. With acute Budd-Chiari syndrome, there may be one or more intraluminal thrombi, most commonly identified by Doppler US. With chronic Budd-Chiari syndrome, however, the hepatic veins are small and often difficult to identify, although tortuous intrahepatic collateral veins or shunts may be apparent. In this setting, shunts may develop from one hepatic vein

a

Fig. 3 a, b. Non-occlusive portal vein thrombosis in a 39-year-old woman with a long history of taking birth control pills. a Axial T2weighted and b post-contrast (late hepatic arterial phase) images through the liver demonstrate an eccentric intraluminal defect in the main portal vein

b

54

a

c

Pablo R. Ros, Rendon C. Nelson

b

Fig. 4 a-d. Axial images through the liver during the portal venous phase demonstrate intense enhancement of the central portion of the liver and hypoenhancement of the peripheral portion (especially on a and b). The hepatic veins are not visualized. There is a cleft in the medial aspect of both the right and left hepatic lobes (c, d) consistent with chronic Budd-Chiari syndrome and peripheral parenchymal atrophy

d

that is obstructed to another hepatic vein that is not. Shunting can also occur from a hepatic vein that is obstructed to the hepatic vein in the caudate lobe. Finally, shunting may occur from a hepatic vein that is obstructed to the portal vein, one of the reasons why 25% of patients with Budd-Chiari syndrome develop portal vein thrombosis. Furthermore, chronic occlusion of the hepatic vein can result in significant enlargement of the caudate lobe and atrophy of the peripheral portion of the right and left hepatic lobes. At times, large intrahepatic collateral veins can be seen shunting blood to the hepatic vein in the caudate lobe. Furthermore, on T1-weighted MRI, the caudate lobe is often hyperintense. Following contrast administration on either CT or MRI, there is often differential enhancement of the central and peripheral portions of the liver (Fig. 4); that is, early on, the central portion of the liver hyperenhances but the periphery does not. Later on, there is a flip-flop phenomenon in which the central portion washes out and the peripheral portion accumulates contrast media. Over time, profound atrophy of the peripheral parenchyma can be seen. Patients commonly have and first clinically present with ascites, which develops shortly after hepatic vein occlusion. In suprahepatic inferior vena caval obstruction, the hepatic vein in the caudate lobe cannot be used as a conduit to shunt blood from the hepatic veins to the inferior vena cava. As a result, there is neither hypertrophy of the caudate lobe nor development of large intrahepatic collateral veins. It is important to note that, in the setting of Budd-Chiari syndrome, the liver is often swollen, which

may narrow the lumen of the intrahepatic inferior vena cava. Although this is not the cause of Budd-Chiari syndrome, it certainly exacerbates the condition.

Passive Hepatic Congestion Right-sided heart failure can result in the delayed drainage of blood from the liver into the inferior vena cava and right atrium. Images through the lower chest typically demonstrate either cardiac enlargement or a large pericardial effusion, although in the setting of restrictive pericarditis the heart may be normal in size. The key finding with right-sided congestive heart failure or passive hepatic congestion is enlarged and distended hepatic veins and inferior vena cava. Secondary signs include reflux of contrast material from the superior vena cava into the hepatic veins, although this is occasionally seen in normal patients as well. In addition, there may be a mottled enhancement pattern in the liver that is more pronounced peripherally and more apparent during the late hepatic arterial phase. This pattern, referred to as the “nutmeg liver”, may not be apparent during the portal venous and/or equilibrium phases. Over time, the liver can become fibrotic, at which point it may have all of the manifestations of cirrhosis and portal hypertension.

Macro-regenerative Nodules In a small percentage of patients with outflow obstruction of the hepatic veins, large regenerative nodules develop

55

Imaging of Diffuse and Inflammatory Liver Diseases

that tend to hyperenhance during the late hepatic arterial phase. These nodules have been associated with both Budd-Chiari syndrome and passive hepatic congestion. They typically are iso-enhancing during the portal venous and equilibrium phases and demonstrate only slow, periodic growth over time.

Hepatic Veno-occlusive Disease Hepatic veno-occlusive disease (VOD) typically occurs in bone marrow transplant recipients who have undergone total body irradiation. The definition of VOD is progressive non-thrombotic occlusion of the hepatic venules. Although some reports have noted sporadic reversal or hepatofugal flow in the portal vein with VOD, there are no reliable imaging findings for this diagnosis. As a result, VOD can only be reliably diagnosed by microscopic examination of biopsy tissue.

Peliosis Hepatis Peliosis hepatis is a rare entity in which the hepatic sinusoids throughout the liver dilate, resulting in numerous blood-filled lacunar spaces ranging in size from 1 to 3 mm. Similar lacunar spaces can also occur in the spleen, lymph nodes, bone marrow, and lungs. Although the cause of peliosis is poorly understood, it is believed to be due to outflow obstruction of the sinusoid. It typically occurs in patients who use anabolic steroids, corticosteroids, tamoxifen, or birth control pills; following cardiac or renal transplantation; with chronic debilitating diseases, such as tuberculosis, malignancy or AIDS; in association with diabetes, sprue, or Hodgkin’s disease; or in patients exposed to arsenic or polyvinyl chloride. Imaging shows numerous small cystic lesions that demonstrate an enhancement pattern similar to that of the blood pool.

Hepatic Infarction Parenchymal infarction in the liver is relatively uncommon for two reasons: 1. the liver has a dual blood supply; 2. the hepatocytes are relatively insensitive to hypoxia. It has been seen, however, in patients with shock, sepsis, eclampsia, sickle cell disease or trait, or arteritis; in those who have taken birth control pills; and in those who have suffered arterial embolic events such as those due to endocarditis, rheumatic heart disease, trauma, intra-arterial chemotherapy, or iatrogenic tumor embolization. With imaging, infarcted parenchyma may be hypoechoic on US: anechoic bile lakes may be visualized as necrosis progresses. On contrast-enhanced CT, there is a wedge-shaped subcapsular region of hypoenhancement that later on may contain bile lakes and gas bubbles. On MRI, the edema of infarction is typically hypointense on T1-weighted images and hyperintense on T2-weighted images; as with CT, hypoenhancement and bile lakes are seen as well.

Metabolic and Storage Diseases Steatosis Hepatic steatosis results from a variety of abnormal processes, including the increased production or mobilization of fatty acids (e.g., obesity, steroid use) or the decreased hepatic clearance of fatty acids due to hepatocellular injury (e.g., alcoholic liver disease, viral hepatitis). Histopathologically, the hallmark of all forms of fatty liver is the accumulation of fat globules within hepatocytes. The distribution of steatosis can be variable, ranging from focal, to regional, to diffuse. Diffuse steatosis is common and estimated to occur in approximately 30% of obese patients. Patients with steatosis are usually asymptomatic although some may present with right upper quadrant pain or abnormal liver function parameters. The histopathological findings of non-alcohol-related liver steatosis, also known as non-alcoholic fatty liver disease (NAFLD), vary from steatosis alone to steatosis with inflammation, necrosis, and fibrosis. Non-alcoholic steatohepatitis (NASH), with or without cirrhosis, is positioned at the most severe end of the NAFLD spectrum. The histopathological findings of NASH include steatosis (predominately macrovesicular), mixed lobular inflammation, and hepatocellular ballooning. Unlike steatosis alone, NASH may progress to cirrhosis. Diffuse fatty change is easily identified on CT. The attenuation value of normal liver is usually ~8 HU greater than that of spleen on non-contrast CT images. In patients with fatty change, however, an abnormally decreased density will be demonstrated, typically 10 and 25 HU less than the spleen on non-contrast and contrast-enhanced CT images, respectively. The diagnosis of hepatic steatosis is more reliably made on non-contrast images. Undoubtedly, the most sensitive technique to detect fatty change of the liver is the use of in-phase and out-phase gradient echo MRI pulse sequences (Fig. 5). Hepatic fatty change is, however, not always uniform but can instead present as a focal area of steatosis in an otherwise normal liver (focal steatosis) or as subtotal fatty change with sparing of certain areas (focal sparing). On imaging, several features allow the correct identification of focal fatty change or focal spared areas: 1. the typical periligamentous and periportal location; 2. lack of mass effect; 3. sharply angulated boundaries of the involved area; 4. non-spherical shape; 5. absence of vascular displacement or distortion; 6. lobar or segmental distribution.

Iron Overload Iron overload states may arise from hemochromatosis, with the preferential accumulation of iron within hepatocytes, or hemosiderosis, in which iron is deposited in Kupffer cells.

56

a

Pablo R. Ros, Rendon C. Nelson

b

c

Fig. 5 a-c. Diffuse fatty liver in a 41-year-old female presenting with epigastric pain. a Axial CECT image demonstrates diffuse low attenuation of the liver without displacement of the hepatic vessels. b In- and c out-of-phase T1-weighted images show significant signal drop in the liver on the out-of-phase images

Primary Hemochromatosis Hereditary or primary hemochromatosis is an autosomal recessive disorder of iron metabolism that is characterized by the abnormal absorption of iron from the gut, with subsequent excessive deposition of iron into hepatocytes, pancreatic acinar cells, myocardium, joints, endocrine glands, and skin. In addition, cells of the reticuloendothelial system (RES) in patients with primary hemochromatosis are abnormal and unable to store processed iron effectively. Consequently, these patients do not accumulate iron into the RES. The clinical findings of cirrhosis and its complications (portal hypertension, development of HCC) predominate in patients with longlasting disease. On CT, excessive iron storage within hepatocytes will result in an overall increased density. However, this CT appearance of a hyperdense liver is non-specific, since similar features can be seen with gold deposition and in Wilson’s disease, type IV glycogen storage disease, and following amiodarone administration. The use of noncontrast CT in patients with suspected hemochromatosis is important because excessive iron cannot be detected in the setting of enhancing parenchyma. MRI is far more specific than any other imaging modality for the characterization of iron overload, due to the magnetic susceptibility effect of iron. The superparamagnetic effect of accumulated iron in the hepatocytes results in a significant reduction of signal intensity on T2-weighted images. Comparison of the signal intensity of liver with that of paraspinal muscles provides a useful internal control. HCC, seen in 35% of patients with advanced hemochromatosis, is usually easily detected on T1- and T2-weighted images due to the background of decreased signal intensity of the liver.

thalassemia major, sideroblastic anemia, pyruvate kinase deficiency, chronic liver disease), the excess iron is processed and accumulates in organs containing reticuloendothelial cells, including liver, spleen, and bone marrow. Diffuse, increased attenuation of the liver and spleen is seen on CT (Fig. 6). On MRI, the extrahepatic signal intensity changes in the spleen and bone marrow allow primary hemochromatosis to be distinguished from hemosiderosis. Although, in general, the clinical significance of transfusional iron overload states is negligible, patients with chronic hemosiderosis can develop symptoms similar to those of the primary form as well as cirrhosis and HCC. Wilson’s Disease Wilson’s disease, also known as hepatolenticular degeneration, is a rare autosomal recessive abnormality of copper metabolism that is characterized by the accumulation

Hemosiderosis In patients with hemosiderosis or siderosis, due to transfusional iron overload states or dyserythropoiesis (e.g.,

Fig. 6. Hemosiderosis in a 45-year-old female with long history of sickle cell anemia requiring multiple transfusions. Axial nonenhanced CT image demonstrates increased attenuation of the liver

57

Imaging of Diffuse and Inflammatory Liver Diseases

of toxic levels of copper in the brain, cornea (KayserFleischer rings), and liver, the latter due to impaired biliary excretion. Hepatic deposition of copper, predominantly seen in periportal areas and along the hepatic sinusoids, evokes an inflammatory reaction resulting in acute hepatitis with fatty change. Subsequently, chronic hepatitis may result in liver fibrosis and, eventually, macronodular cirrhosis. Due to the high atomic number of copper, a hyperdense liver may be seen on unenhanced CT scans. However, this finding is not universally present; instead, usually only non-specific signs, such as hepatomegaly, fatty change, and in advanced cases, cirrhosis, are observed. During the early stage of the disease, and due to the paramagnetism of ionic copper, MRI can be valuable as it demonstrates focal copper depositions, such as multiple nodular lesions, typically appearing hyperintense and hypointense on T1- and T2-weighted images, respectively.

Fig. 7. Diffuse metastatic breast cancer (pseudocirrhosis pattern) in a 43-year-old woman treated for metastatic breast cancer. Axial contrast-enhanced CT image demonstrate several small low-density lesions in the liver and a nodular contour of the liver due to hepatic capsular retraction

Amyloidosis Lymphoma In amyloidosis, fibrils of protein-mucopolysaccharide complexes are deposited throughout the body. The disease is classified based on the biochemical composition of the amyloid fibrils. Primary amyloidosis is due to the deposition of immunoglobin light chains and is associated with multiple myeloma and monoclonal gammopathy. Secondary amyloidosis results from the deposition of amyloid A protein and is associated with chronic infection, rheumatoid arthritis, and malignancies. Exceeded only by the spleen and kidney, the liver is the third most common solid organ prone to amyloid deposition. Hepatic amyloidosis has a non-specific imaging appearance, with the most common finding being diffuse hepatomegaly. CT sporadically demonstrates focal areas of low attenuation within the liver, corresponding to sites of amyloid deposition (amyloid pseudotumor).

Neoplastic Diseases Metastatic Disease Neoplastic infiltration due to diffuse metastatic disease can occur with many primary tumors. Melanoma, malignant neuroendocrine tumors, pancreatic adenocarcinoma, breast carcinoma, and colonic adenocarcinoma are some of the more commonly encountered causes of diffuse hepatic metastatic disease. The CT appearances of hepatic metastases depend on the vascularity of the lesions compared with the normal liver parenchyma. Diffuse metastatic involvement may produce only subtle imaging findings and be detectable only through indirect features, such as diffuse parenchymal heterogeneity, vascular and architectural distortion, or alterations of the liver contour. The latter, particularly seen in patients with treated breast cancer metastases, has been described as the “pseudocirrhosis” sign (Fig. 7).

Lymphoma can infiltrate the liver both primarily and secondarily. Primary lymphoma of the liver is extremely rare. Conversely, the liver is often secondarily involved in Hodgkin’s and in non-Hodgkin’s lymphoma. Typically, the liver parenchyma is diffusely infiltrated with microscopic nests of neoplastic cells, without significant architectural distortion. Consequently, lymphomatous involvement is difficult to detect by imaging alone. Associated abnormalities, such as splenomegaly and lymphadenopathy, may narrow the differential diagnosis.

Diffuse Infectious and Inflammatory Diseases Fungal Infections Hepatosplenic fungal infection is a clinical manifestation of disseminated fungal disease in patients with hematological malignancies or compromise of the immune system. The reported prevalence of fungal dissemination ranges from 20 to 40%. Most hepatic fungal microabscesses occur in leukemia patients and are caused by Candida albicans.

Candidiasis Candida albicans in the liver may evoke little or no inflammatory reaction, cause a superlative response, or occasionally produce granulomas. The typical histological pattern of hepatic candidiasis is characterized by microabscesses, with the yeast or pseudohyphal forms of the fungus in the center of the lesion and a surrounding area of necrosis and polymorphonuclear infiltrate. At contrast-enhanced CT, fungal microabscesses usually appear as multiple, round, discrete areas of low attenuation, generally ranging in size from 2 to 20 mm (Fig. 8). These microabscesses usually enhance centrally

58

Pablo R. Ros, Rendon C. Nelson

Hepatic contrast-enhanced CT may typically reveal multiple, diffuse, small, low-density areas in both the liver and spleen. MRI features of hepatic sarcoidosis are also non-specific and include organomegaly, multiple lesions of low signal intensity relative to background parenchyma with all sequences, increased periportal signal, irregularity of the portal and hepatic vein branches, and patchy areas of heterogeneous signal. Tuberculosis

Fig. 8. Hepatic candidiasis in a 66-year-old female with leukemia who presented with abnormal liver function tests. Axial contrastenhanced CT image demonstrates multiple, small, low-attenuation lesions distributed throughout the liver and spleen. Splenomegaly and bilateral pleural effusions are also seen

after intravenous administration of contrast medium, although peripheral enhancement may occur as well. At MRI, the untreated nodules are rounded lesions 140 ms) will demonstrate the presence of a homogeneously “light-bulb-bright” lesion, which is characteristic of a benign lesion, either cyst or hemangioma. The exceptions to this would include cystic metastases, and gastrointestinal stromal tumor (GIST) and neuroendocrine tumor metastases. Hemangiomas show three distinctive patterns of enhancement at CT/MR imaging [16], with the common characteristic feature that areas of lesion enhancement closely follow the enhancement characteristics of blood pooling elsewhere [17]. Small lesions (up to ~2 cm) may show immediate and complete filling in the arterial phase, with sustained enhancement in the venous and delayed phases (type I, also termed flash filling) [18] (Fig. 1). The most common enhancement pattern is one of peripheral

c

Fig. 1 a-c. Hemangioma type 1. a Unenhanced CT shows a small hypodense lesion adjacent to the falciform ligament (arrow). b Contrastenhanced CT in the arterial phase shows rapid and complete enhancement of the hemangioma, which persists in the venous phase (c). The attenuation of the hemangioma in the enhanced phases is similar to that of the aorta

65

Focal Liver Lesions

a

b

c

d

Fig. 2 a-d. Hemangioma type III. a T2-weighted turbo spin-echo (TSE) MR image shows a very hyperintense lesion in the right lobe. b-d Dynamic gadolinium-enhanced T1-weighted gradient-recalled echo (GRE) images: b arterial, c venous, and d equilibrium phases show peripheral nodular enhancement with progressive centripetal fill-in. d In the equilibriumphase, after 5 min, there is no complete fill-in

nodular discontinuous enhancement that progresses with increased fill-in over time (type II). Larger lesions (>5 cm) or lesions with central thrombosis/fibrosis may lack central fill-in (type III) (Fig. 2). With SPIO agents, the blood pooling effect, with accumulation of contrast material in the sinuses, may lead to the prolonged enhancement of hemangiomas on T1 images and signal intensity loss on T2 images, despite the lack of Kupffer cells in these lesions. This imaging feature helps in the differentiation between hemangiomas and metastases [19]. Recent studies have shown that non-contrast diffusionweighted imaging may help to differentiate between hemangioma and solid lesions, as the apparent diffusion coefficient of hemangiomas is higher than that of solid lesions [20].

Focal Nodular Hyperplasia This benign lesion is usually of no clinical consequence other than the confusion it causes when incidentally detected during abdominal imaging examinations. The sonographic appearance of focal nodular hyperplasia (FNH) is non-specific; the lesion may be isoechoic, or slightly hypoechoic [21] to liver, while in patients with diffuse hepatic steatosis it is always hypoechoic. One characteristic feature is the presence of a central scar, seen in approximately two-thirds of large lesions but in only one-third of small lesions (10 mm ranges from 37 to 88%, patients with symptomatic polyps >10 mm are encouraged to undergo cholecystectomy while those with polyps 30-50% >50%

100 90 80 70 60 50 40 30 20 10 0

0 1 2 3 4

0 2 4 6

%

Mortality Complications

0-3

4-6

7 - 10

Fig. 4. CT severity index related to the degree of necrosis of the pancreatic parenchyma. Grade A = 0, B = 1, C = 2, D = 3, E = 4; no necrosis = 0, 30% necrosis = 2, 50% necrosis = 6 (from [26,27])

About a fifth of the patients without necrotic changes of the pancreatic gland will also develop local complications. Fluid collections are seen in up to 50% of patients with AP. In about half of these patients, the collections will resolve spontaneously within several weeks. In the rest, however, the fluid collections will persist, eventually followed by encapsulation, superinfection (abscess), or pseudocyst formation.

Modified CT severity index Pancreatic inflammation Normal pancreas Intrinsic pancreatic abnormalities with inflammatory changes in peripancreatic fat Pancreatic or peripancreatic fluid collection or peripancreatic fat necrosis Pancreatic necrosis None ≤30% >30% Extrapancreatic complications (one or more of pleural effusion, ascites, vascular complications, parenchymal complications, or gastrointestinal tract involvement)

Points 0 2 4

0 2 4 2

In a septic patient, not-water-like fluid collections and rim enhancement on contrast-enhanced CT or MRI studies should be considered as abscesses until proven otherwise. Gas, a characteristic sign of an infected fluid collection, is detected in only 20% of patients with pancreatic abscesses. Percutaneous aspiration or drain placement is the proper treatment. In contrast to abscess formations, superinfected necrotic areas of the pancreas are much more difficult to treat. Due to the more solid consistency of the infected necrosis, percutaneous drainage therapy is mostly frustrating; however, biopsy is often needed to prove the diagnosis. In most cases, either percutaneous, endoscopyguided necrosectomy, or surgical intervention has to be considered. Pseudoaneurysm formation and hemorrhage may result from the extravasated pancreatic enzymes that cause vascular injury. They are typically late complications that occur after several episodes of severe AP. While pseudoaneurysms are generally easily detected by any kind of imaging modality, retroperitoneal hemorrhage is best depicted by contrast-enhanced CT or unenhanced MRI. Angiography with arterial embolization is the treatment of choice and in general is superior to surgical therapy [29].

Groove Pancreatitis Complications Pseudocysts are fluid collections with a noticeable capsule that typically develop 4-5 weeks after the onset of AP. On US, CT, and MRI, they have a cyst-like appearance, usually without septations. Since large cysts are prone to complications (e.g., rupture, infection, hemorrhage, biliary obstruction, or fistulization to the gastrointestinal (GI) tract), cysts >5-7 cm in diameter should be treated by percutaneous drainage or operative marsupialization.

First described by Becker in 1973, groove pancreatitis is a rare, late complication occurring after several attacks of AP. It is defined as an inflammatory reaction and fluid collection located in the groove between the head of the pancreas, the duodenum, and the common bile duct. The anterior anlage of the pancreas seems to be mainly affected, with duodenal stenosis and/or strictures of the common duct in about 50% of the cases. Therefore, this disease may mimic cancer of the pancreatic head, necessitating surgi-

85

Diseases of the Pancreas, I: Pancreatitis

cal exploration. Dynamic CT and MRI with delayed enhancement of collagen fibrous tissue during the late postequilibrium phase may reveal a potential soft-tissue mass to be fibrosis and thus, in the absence of complications, obviate the need for surgical exploration [30-33].

Autoimmune Pancreatitis Autoimmune pancreatitis (AIP) is a relatively new syndrome of clinical and histological findings that was first described by Yoshida in 1995 [34]. The condition has also been described as lymphoplasmocytic sclerosing pancreatitis with cholangitis, non-alcoholic duct-destructive chronic pancreatitis, and chronic sclerosing pancreatitis. The features of AIP include hypergammaglobulinemia, elevation of serum IgG4, IgG4-containing immune complexes, and a number of other antibodies as antinuclear antibodies, as well as antibodies against lactoferrin, carbonic anhydrase type II, and rheumatoid factors. Histologically, there is fibrosis and a lymphoplasmacytic infiltration of the interlobular ducts. The majority of lymphocytes are CD8+ and CD4+, while B lymphocytes are less frequent. In general, the diagnosis of AIP is established by clinical signs, together with laboratory and morphological findings. An association with other autoimmune diseases, such as Sjögren-syndrome, primary biliary cirrhosis, primary sclerosing cholangitis, Crohn’s disease or ulcerating colitis, systemic lupus erythematosus, and retroperitoneal fibrosis is found in a third of the cases. At imaging, a focal (“mass-forming”) or diffuse (“sausage-like”) enlargement of the pancreas may be present. In contrast-enhanced studies, peripancreatic nodular or rim-like enhancement can be appreciated. Focal AIP of the pancreatic head that involves the pancreatic and distal common bile duct must be differentiated from pancreatic carcinoma, necessitating biopsy proof [35, 36]. In most patients, the symptoms as well as the laboratory and morphological abnormalities appear to respond to steroid treatment [34, 37-42].

Chronic Pancreatitis The hallmarks of chronic pancreatitis (CP) are a continuing (aseptic) inflammation of the gland accompanied by irreversible morphological and functional damage. The most common reasons are chronic alcohol abuse (70%) and cholelithiasis (20%), with rare cases arising from cystic fibrosis or idiopathically. Patients are typically in their 3rd to 4th decade of life and present with a history of epigastric pain (95%), weight loss (95%), and signs of endocrine/ exocrine deficiency (diabetes mellitus 58%, malabsorption syndrome and steatorrhea 80%). Acute exacerbations of CP are accompanied by episodes of pain attacks that may mimic an acute abdomen. With progressive destruction of the gland, CP may be painless after several years. In 1.5-12% of cases it is complicated by pancreatic cancer.

Tumor markers such as CA 19-9 and CA-50 may be elevated transiently and are non-specific. Laboratory tests of secretin-creozyme and secretin-caerulein have a high diagnostic accuracy except in early stages of the disease but are invasive and cumbersome for the patient. However, these tests are of particular importance in the diagnostically challenging, newly defined small-duct CP, in which chronic inflammation occurs without ductal abnormalities. In CP, the most characteristic findings are dilatation of the pancreatic main duct and of the ductal side branches (70-90%), small cystic changes, scattered glandular and ductal calcifications (40-50%), and ductal protein plugs. The grade and shape of the ductal dilatation may help to differentiate chronic (benign) obstructions from malignant occlusions: in CP, the contour of the pancreatic duct and its side branches is commonly irregular (73%) while this is true only in 15% of pancreatic malignancies. Additionally, the duct usually accounts for 55 years WBC >16,000

Hematocrit decrease >10% Blood urea nitrogen increase >5 mg/dL Ca (serum) 600 mL Mortality (%) 50

Blood glucose >200 mg/dL Serum LDH >350 IU/L SGOT (AST) >250 U/L

Score 0-2 3-5 >5

missed diagnosis of carcinoma. However, if local or regional lymph node enlargement, vascular encasement, or remote metastases is displayed, the differential is ruled by these secondary signs of malignancy, in which case the tumor must be staged correctly for further treatment stratification (Tables 3, 4). In ambiguous cases, biopsy or even surgical exploration may be necessary. CP can cause a focal pancreatic mass indicative of a neoplasm. Moreover, it represents a major risk factor for pancreatic cancer, with a 26-fold increased risk of developing cancer, according to an international, multicenter cohort study [45]. Therefore, the differential between CP and pancreatic cancer remains challenging and underlines the need for multiple diagnostic approaches. In one study, US, CT, MRI, and positron emission tomography (PET)/CT for pancreatic cancer were shown to have a sensitivity of 76-83% and a specificity of 91-93% [46]. Nevertheless, the rate of incorrect diagnoses is as high as 25%. The use of various differential criteria (Table 5) may help to improve the overall diagnostic accuracy beyond that achieved based solely on image interpretation.

87

Diseases of the Pancreas, I: Pancreatitis

Table 5. Differential criteria for chronic pancreatitis (CP) versus pancreatic cancer (PCa)

History Duct Duct/parenchyma Calcification Enhancement Cysts Lymph nodes Metastases

CP

PCa

+++ Irregular 0.5 – Focal (+) ++ +++

Recently, the use of new methods and techniques, such as oxygen insensitivity testing, have been described in conjunction with conventional pathology studies of brush cytology in facilitating discrimination between CP and pancreatic cancer [47, 48]. Nevertheless, to date, biopsy is the most reliable diagnostic tool in ambiguous cases of pancreatic masses, with no significant difference whether the biopsy was performed endoscopically or percutaneously, while laparoscopic procedures are compromised by an increased potential for tumor seeding and adverse events.

Pancreatitis in Children Fortunately, pancreatitis in childhood is rare. The best known causes for pediatric pancreatitis are traumas to the pancreas (typically, a bicycle accident), the genetically transmitted hereditary pancreatitis, and cystic fibrosis. Nevertheless, in most of the cases the reason is unknown (idiopathic pancreatitis), although microcalculi and protein plugs have been suggested. In general, the diagnosis can be established based on the history, clinical findings, and imaging findings, which are not different from those described in adults [49-51].

References 1. Mitchell RM, Byrne MF, Baillie J (2003) Pancreatitis. Lancet 361:1447-1455 2. Glasbrenner B, Kahl S, Malfertheiner P (2002) Modern diagnostics of chronic pancreatitis. Eur J Gastroenterol Hepatol 14:935-941 3. Etemad B, Whitcomb DC (2001) Chronic pancreatitis: diagnosis, classification, and new genetic developments. Gastroenterology 120:682-707 4. Halonen KI, Leppaniemi AK, Puolakkainen PA et al (2000) Severe acute pancreatitis: prognostic factors in 270 consecutive patients. Pancreas 21:266-271 5. Bank S, Indaram A (1999) Causes of acute and recurrent pancreatitis. Clinical considerations and clues to diagnosis. Gastroenterol Clin North Am 28:571-89, viii 6. Cavallini G, Frulloni L, Bassi C et al (2004) Prospective multicentre survey on acute pancreatitis in Italy (ProInf-AISP): results on 1005 patients. Dig Liver Dis 36:205-211 7. Losanoff JE, Asparouhov OK, Jones JW (2001) Multiple factor scoring system for risk assessment of acute pancreatitis. J Surg Res 101:73-78

8. Plock JA, Schmidt J, Anderson SE et al (2005) Contrast-enhanced computed tomography in acute pancreatitis: does contrast medium worsen its course due to impaired microcirculation? Langenbecks Arch Surg 390:156-163 9. Werner J, Schmidt J, Warshaw AL et al (1998) The relative safety of MRI contrast agent in acute necrotizing pancreatitis. Ann Surg 227:105-111 10. Robinson PJ, Sheridan MB (2000) Pancreatitis: computed tomography and magnetic resonance imaging. Eur Radiol 10:401-408 11. Pamuklar E, Semelka RC (2005) MR imaging of the pancreas. Magn Reson Imaging Clin N Am 13:313-330 12. Laurens B, Leroy C, André A et al (2005) [Imaging of acute pancreatitis]. J Radiol 86:733-46; quiz 746-747 13. Vaishali MD, Agarwal AK, Upadhyaya DN et al (2004) Magnetic resonance cholangiopancreatography in obstructive jaundice. J Clin Gastroenterol 38:887-890 14. Arvanitakis M, Delhaye M, De Maertelaere V et al (2004) Computed tomography and magnetic resonance imaging in the assessment of acute pancreatitis. Gastroenterology 126:715-723 15. Akahane T, Kuriyama S, Matsumoto M et al (2003) Pancreatic pleural effusion with a pancreaticopleural fistula diagnosed by magnetic resonance cholangiopancreatography and cured by somatostatin analogue treatment. Abdom Imaging 28:92-95 16. Sica GT, Miller FH, Rodriguez G et al (2002) Magnetic resonance imaging in patients with pancreatitis: evaluation of signal intensity and enhancement changes. J Magn Reson Imaging 15:275-284 17. Okai T, Fujii T, Ida M et al (2002) EUS and ERCP features of nonalcoholic duct-destructive, mass-forming pancreatitis before and after treatment with prednisolone. Abdom Imaging 27:74-76 18. Taylor SL, Morgan DL, Denson KD et al (2005) A comparison of the Ranson, Glasgow, and APACHE II scoring systems to a multiple organ system score in predicting patient outcome in pancreatitis. Am J Surg 189:219-222 19. Mentula P, Kylänpää ML, Kemppainen E et al (2005) Early prediction of organ failure by combined markers in patients with acute pancreatitis. Br J Surg 92:68-75 20. Leung TK, Lee CM, Lin SY et al (2005) Balthazar computed tomography severity index is superior to Ranson criteria and APACHE II scoring system in predicting acute pancreatitis outcome. World J Gastroenterol 11:6049-6052 21. Johnson CD, Toh SK, Campbell MJ (2004) Combination of APACHE-II score and an obesity score (APACHE-O) for the prediction of severe acute pancreatitis. Pancreatology 4:1-6 22. Gerlach H (2004) Risk management in patients with severe acute pancreatitis. Crit Care 8:430-432 23. King NK, Powell JJ, Redhead D, Siriwardena AK (2003) A simplified method for computed tomographic estimation of prognosis in acute pancreatitis. Scand J Gastroenterol 38:433-436 24. Triester SL, Kowdley KV (2002) Prognostic factors in acute pancreatitis. J Clin Gastroenterol 34:167-176 25. Sandberg AA, Borgstrom A (2002) Early prediction of severity in acute pancreatitis. Is this possible? Jop 3:116-125 26. Balthazar EJ (2002) Staging of acute pancreatitis. Radiol Clin North Am 40:1199-1209 27. Balthazar EJ (2002) Acute pancreatitis: assessment of severity with clinical and CT evaluation. Radiology 223:603-513 28. Mortele KJ, Wiesner W, Intriere L et al (2004) A modified CT severity index for evaluating acute pancreatitis: improved correlation with patient outcome. AJR Am J Roentgenol 183: 1261-1265 29. Mofidi R, Patil PV, Suttie SA, Parks RW (2009) Risk assessment in acute pancreatitis. Br J Surg 96:137-150 30. Irie H, Honda H, Kuroiwa T et al (1998) MRI of groove pancreatitis. J Comput Assist Tomogr 22:651-5 31. Gabata T, Kadoya M, Terayama N et al (2003) Groove pancreatic carcinomas: radiological and pathological findings. Eur Radiol 13:1679-1684

88

32. Shanbhogue AK, Fasih N, Surabhi VR et al (2009) A clinical and radiologic review of uncommon types and causes of pancreatitis. Radiographics 29:1003-1026 33. Triantopoulou C, Dervenis C, Giannakou N et al (2009) Groove pancreatitis: a diagnostic challenge. Eur Radiol 19:1736-1743 34. Okazaki K (2005) Autoimmune pancreatitis: etiology, pathogenesis, clinical findings and treatment. The Japanese experience. Jop 6(1 Suppl):89-96 35. Takahashi N, Fletcher JG, Hough DM et al (2009) Autoimmune pancreatitis: differentiation from pancreatic carcinoma and normal pancreas on the basis of enhancement characteristics at dual-phase CT. AJR Am J Roentgenol 193:479-484 36. Weili L, Jiaguo W (2009) Education and imaging: Hepatobiliary and pancreatic: autoimmune pancreatitis. J Gastroenterol Hepatol 24:1574 37. Sahani DV, Kalva SP, Farrell J et al (2004) Autoimmune pancreatitis: imaging features. Radiology 233:345-352 38. Farrell JJ, Garber J, Sahani D, Brugge WR (2004) EUS findings in patients with autoimmune pancreatitis. Gastrointest Endosc 60:927-936 39. Wakabayashi T, Kawaura Y, Satomura Y et al (2003) Clinical and imaging features of autoimmune pancreatitis with focal pancreatic swelling or mass formation: comparison with socalled tumor-forming pancreatitis and pancreatic carcinoma. Am J Gastroenterol 98:2679-2687 40. Ito K, Koike S, Matsunaga N (2001) MR imaging of pancreatic diseases. Eur J Radiol 38:78-93 41. Irie H, Honda H, Baba S et al (1998) Autoimmune pancreatitis: CT and MR characteristics. AJR Am J Roentgenol 170:1323-1327

Thomas Helmberger

42. Sahani DV, Sainani NI, Deshpande V et al (2009) Autoimmune pancreatitis: disease evolution, staging, response assessment, and CT features that predict response to corticosteroid therapy. Radiology 250:118-129 43. Sainani NI, Conwell DL (2009) Secretin-enhanced MRCP: proceed with cautious optimism. Am J Gastroenterol 104:1787-1789 44. Yamada Y, Mori H, Matsumoto S et al (2009) Pancreatic adenocarcinoma versus chronic pancreatitis: differentiation with triple-phase helical CT. Abdom Imaging 22 45. Lowenfels AB, Maisonneuve P, Cavallini G et al (1994) Prognosis of chronic pancreatitis: an international multicenter study. International Pancreatitis Study Group. Am J Gastroenterol 89:1467-1471 46. Tang S, Huang G, Liu J et al (2009) Usefulness of (18)F-FDG PET, combined FDG-PET/CT and EUS in diagnosing primary pancreatic carcinoma: A meta-analysis. Eur J Radiol (in press) 47. Cho SG, Lee DH, Lee KY et al (2005) Differentiation of chronic focal pancreatitis from pancreatic carcinoma by in vivo proton magnetic resonance spectroscopy. J Comput Assist Tomogr 29:163-9 48. van Kouwen MC, Jansen JB, van Goor H et al (2005) FDGPET is able to detect pancreatic carcinoma in chronic pancreatitis. Eur J Nucl Med Mol Imaging 32:399-404 49. Manfredi R, Lucidi V, Gui B et al (2002) Idiopathic chronic pancreatitis in children: MR cholangiopancreatography after secretin administration. Radiology 224:675-682 50. DeBanto JR, Goday PS, Pedroso MR et al (2002) Acute pancreatitis in children. Am J Gastroenterol 97:1726-1731 51. Levy MJ, Geenen JE (2001) Idiopathic acute recurrent pancreatitis. Am J Gastroenterol 96:2540-2555

IDKD 2010-2013

Diseases of the Pancreas, II: Tumors Ruedi F. Thoeni Department of Radiology and Biomedical Imaging, University of California, San Francisco, CA, USA

Introduction In the imaging of pancreatic disease and in assessment of the etiology of jaundice, abdominal ultrasound (US) and computed tomography (CT) traditionally have been employed [1]. These two methods are widely available and have the advantages of their familiarity to radiologists and clinicians and their non-invasiveness. With the introduction of magnetic resonance imaging (MRI), magnetic resonance cholangiopancreatography (MRCP), and endoscopic ultrasound (EUS), visualization of the pancreatic and biliary ducts improved, allowing tumors to be more accurately staged and safely sampled [2-12]. This led to a diminishing role for endoscopic retrograde cholangiography in the diagnostic arena but its therapeutic use has remained unchallenged. In recent years, technological advances with multidetector row CT (MDCT) imaging have improved the ability of CT to detect even small lesions in the pancreas and to stage pancreatic tumors more accurately [13]. Microbubble contrast enhancement and secretin stimulation have increased the diagnostic acumen of US and MRI, respectively, and may widen the utility of these techniques [14, 15]. Nevertheless, MDCT remains the primary tool in assessing patients with suspected pancreatic disease, while EUS and MRI are used as problem-solving modalities to confirm suspected lesions not identified with CT, to find additional lesions, and to obtain a definitive tissue diagnosis with EUS-guided tissue sampling. In recent years, position emission tomography/CT (PET/CT) has been increasingly employed in the assessment of patients with suspected pancreatic tumors but its ultimate role still needs further definition [16-21]. Also, somatostatin receptor scintigraphy has gained popularity in recent years for neuroendocrine tumors [22, 23]. This discussion will focus on diagnosing and staging the various pancreatic neoplasms with CT and MRI, mentioning EUS, PET/CT, and somatostatin receptor scintigraphy where appropriate.

Ductal Adenocarcinoma of the Pancreas About 90% of all neoplasms of the pancreas are ductal adenocarcinomas. Pancreatic adenocarcinoma is one of

the leading causes of cancer death in the western world, and the overall relative 5-year survival rate of only 5.5% is dismal [24]. Late clinical presentation with advanced disease and the aggressiveness of the tumor result in a low rate of surgical intervention and overall poor outcome. It is estimated that 42,470 men and women will be diagnosed with cancer of the pancreas in 2009 and that over 80% of these patients will die of the disease [25]. The tumor serum marker CA 19-9 is sensitive, although not specific for the diagnosis of adenocarcinoma of the pancreas. The treatment approach is based on whether the tumor can or cannot be resected at presentation. Therefore, imaging plays a crucial role in disease management. The initial diagnosis of pancreatic tumor, particularly if the patient presents with jaundice and the tumor is located in the head of the pancreas, may be made by US. The ultrasonographic signs of pancreatic carcinoma include a focal or diffuse pancreatic mass that is hypoechoic relative to normal gland parenchyma and dilatation of the pancreatic duct without or with biliary duct distention (double-duct sign). The accuracy of US for detecting the level of bile duct obstruction varies greatly, and ultrasonographic staging of pancreatic carcinoma is inferior to that of CT. Ultrasonography often fails to provide an adequate examination of the entire gland, resulting in an overall decrease in the sensitivity of this technique. Some of these limitations are overcome by endosonography, but tumors in the tail of the pancreas are not always accessible by EUS. For optimal evaluation of pancreatic neoplasms, MDCT is the modality of choice. A triple-phase protocol is recommended that includes thin sections (0.625 or 1.25 mm) through the abdomen, initially without intravenous contrast followed by a rapidly delivered bolus of contrast material (we use bolus tracking and 150 mL at 5 mL/s chased by 50 mL of saline). It is best to use a neutral oral contrast agent (water or VoLumen, Bracco Diagnostics, USA) because it permits optimal assessment of tumor extension to the stomach and/or duodenum and does not interfere with the determination of vascular invasion. We recommend a scan delay of 40-45 s (10-s delay from peak aortic enhancement) for the late arterial or pancreatic phase and a scan delay of 80 s for the hepatic phase. Rarely, an arterial phase

90

Ruedi F. Thoeni

at 20-25 s is performed if requested by surgery [26]. Arterial involvement and tumor mass are best detected in the pancreatic phase whereas the hepatic phase enables optimal visualization of the liver, veins, and the entire abdomen in the search for liver metastases and peritoneal seeding. One study demonstrated that a single-phase thinslice MDCT technique is sufficient for accurately assessing the resectability of pancreatic adenocarcinoma [27]. Pancreatic adenocarcinoma arises from the pancreatic duct. On MDCT, the tumor usually appears as a low-density mass, often associated with poorly defined margins (Fig. 1) and pancreatic and/or bile duct dilatation. The low-density central zone represents either hypovascular, scirrhous tumor surrounded by normal parenchyma or inflammatory tissue caused by obstructive pancreatitis. Occasionally, cystic degeneration is seen within the tumor [28]. Neoplastic pancreatic duct obstruction frequently produces a dilated duct as well as atrophy of the pancreatic parenchyma proximal to the neoplasm. Tumor obstruction of the main pancreatic duct can lead to rupture of the side branches, resulting in the formation of pseudocysts. Occasionally, a low-density mass cannot be identified because the tumor is isodense to the surrounding normal parenchyma. In these cases, a dilated duct with abrupt cut-off is often seen proximal to a small imperceptible tumor mass. Ancillary findings are local tumor extension, including direct invasion of neighboring organs such as the liver and the stomach, the arteries (loss of fat planes surrounding celiac axis, superior mesenteric artery, etc., vascular “cuffing”) and veins (tear-drop sign, flattening, irregularity of margins, etc., of the portal vein, superior mesenteric vein and its branches), and metastatic disease to local lymph nodes, as well as spread to the

liver, peritoneum (often associated with ascites), and more distant sites. The double-duct sign (dilatation of the biliary and pancreatic ducts) occurs in 90% for detecting pancreatic carcinoma can be achieved [1, 27]. Small metastatic implants on the liver and peritoneum are the lesions most likely to be missed by MDCT. MDCT generally provides accurate information on vascular involvement as long as a pancreatic protocol is observed; for resectability, sensitivities of >80% have been obtained [1]. The positive predictive values for predicting unresectability are much better than those for predicting resectability. Presently, most studies show a slight advantage of MDCT

a

b

Fig. 1 a, b. a Thin-section (1.25 mm) axial MDCT of a pancreas carcinoma in the pancreatic phase (~40 s). A low-attenuation mass is apparent in the head of pancreas near the uncinate process (white arrows), with encasement of the replaced right hepatic artery originating from the superior mesenteric artery (arrowhead). Retroperitoneal lymphadenopathy also is seen (black arrow). The mass is easily distinguished from adjacent normal pancreas. b Thin-section (1.25 mm) axial MDCT of a pancreas carcinoma in the hepatic phase (~80 s) in the same patient. Note the teardrop-shaped superior mesenteric vein (black arrowhead) as a sign of venous encasement. The low-attenuation pancreatic neoplasm (arrows) is less well seen

91

Diseases of the Pancreas, II: Tumors

over MR for detecting and staging pancreatic adenocarcinomas. A meta-analysis comparing CT, MRI, and US demonstrated a sensitivity and specificity of 91 and 85% for helical CT and a sensitivity and specificity of 84 and 82% for MRI whereas the results for resectability were similar [1]. For US, the sensitivity for diagnosing pancreas carcinoma and the specificity for determining resectability are much lower. The advantage of MRI is thought to be in the area of small tumors that do not alter the contour of the gland [10] and in detecting hepatic metastases. At present, MRI appears to be a problem-solving modality. It should be considered in patients with suspected pancreatic neoplasms in the presence of: (1) allergy to iodine contrast or other contraindications to iodine contrast administration, (2) a MDCT scan showing focal enlargement of the pancreas but no definable mass, (3) a clinical history suggesting malignancy and MDCT images that are equivocal or difficult to interpret, and (4) when distinction between chronic pancreatitis with focal enlargement and pancreatic cancer is needed. When choosing an imaging modality, one has to take into account that, today, MDCT of the pancreas requires a small fraction of the time needed for a complete MRI study of the pancreas. False-positive MDCT diagnoses of pancreatic cancer can occur, especially in patients with chronic pancreatitis; therefore, percutaneous aspiration biopsies are needed if non-operative treatment is planned. Fine-needle aspiration biopsy of pancreatic cancer under CT guidance is a frequently performed procedure and is associated with severe pancreatitis in 2 cm, a central enhancing scar is rarely seen, and calcifications are peripheral [28]. The margins usually are smooth and metastatic disease may be present at the time of diagnosis. Based on the above-mentioned criteria, a correct diagnosis of a serous cystic pancreatic tumor can be made in 62% of patients by CT, in 74% by sonography, and in 84% using both modalities [37]. Overall, the results for mucinous cystic tumors are inferior. Pancreatic pseudocysts and cystic forms of islet cell tumors, ductal carcinomas, solid and papillary tumors, and lymphangioma of the pancreas can be indistinguishable on CT from cystic neoplasms. Thus, EUS needle biopsies of the lesions often are necessary [38]. Better definition of the internal architecture of these cystic neoplasms is frequently obtained with MRI rather than CT. MRI also demonstrates the presence of mucin, seen as an area of increased signal intensity within the cysts on T1-weighted sequences. Septa and wall thickness of the lesions are well demonstrated by MRI but this is not always true for calcifications. MRI is of great help in distinguishing these cystic neoplasms from pseudo-

93

cysts of the pancreas, particularly if they are multiple. Both MRCP and MDCT with curved planar reconstruction can demonstrate the absence of a connection to the main pancreatic duct.

Intraductal Papillary Mucinous Neoplasm of the Pancreas A rare tumor that is considered a subtype of the mucinous cystic neoplasms of the pancreas is the intraductal papillary mucinous tumor of the pancreas (IPMN, formerly also called ductectatic cystadenoma or ductectatic cystadenocarcinoma). IPMN can be classified as main duct, branch duct (side-branch), or mixed type depending on the site and extent of involvement [39]. The cystic changes always demonstrate a connection to the pancreatic duct, which is a diagnostic feature that can be seen on MDCT and MRI. The branch duct tumor consists of cystic dilation of the side branches of the pancreatic duct, usually in the uncinate process. These ducts are lined with atypical, hyperplastic, or clearly malignant epithelium. In the late stages, the tumor nodules of the ducts produce copious mucinous secretions that fill the entire duct. Since extension into the parenchyma and beyond occurs relatively late in branch duct IPMN and overall malignant degeneration is rare, the overall prognosis is good. In 2544% of resected specimens of the other two types, malignancy is present. Resection is therefore the treatment of choice in these patients. CT shows markedly dilated ducts and cystic-appearing structures filled with mucinous material, which has a slightly higher attenuation than that of water. MRI seems to have a slight advantage over CT because it can visualize mucin within the cysts as well as the internal architecture of the lesion, including a solid mass and mural nodules (which are signs of malignancy) slightly better than CT. EUS also is well suited to detect mural nodules.

Solid Pseudopapillary Epithelial Neoplasm of the Pancreas

Fig. 3. Thin-section (1.25 mm) axial MDCT of a serous cystic neoplasm of the pancreas in the pancreatic phase. A lobulated and septated cystic mass is present in the tail of the pancreas (arrows). The individual cysts are small and the septa barely perceptible

Solid pseudopapillary epithelial neoplasms, previously called solid and cystic tumors of the pancreas, are rare tumors. They are seen almost exclusively in young women and are located mostly in the tail of the pancreas. This neoplasm is characterized by a solid peripheral area of tumor and a central zone of degeneration consisting of hemorrhage and cystic spaces filled with necrotic debris that can be visualized by CT (Fig. 4) and MRI. On imaging, these tumors appear as sharply defined, heterogeneous, large cystic pancreatic masses with solid components. They usually are benign but in older women they may be malignant [40]. Calcifications may be present in the capsule or in the inner portion of the mass. EUS also can be helpful in visualizing the nodules and the internal architecture of these masses.

94

Fig. 4. Thin-section (1.25 mm) axial MDCT of a solid pseudopapillary epithelial neoplasm of the pancreas in the pancreatic phase. A large enhancing mass (arrows) with partial necrosis is seen in the tail of the pancreas. The origin from the pancreas is clearly visualized on this image

Follow-Up for Cystic Neoplasms of the Pancreas Small lesions (≤3 cm) that are asymptomatic, show no signs of malignancy, and have negative fine-needle aspiration biopsy results can be followed every 6 months for one year and then yearly for a total of 4 years. If the cyst becomes symptomatic, increases in size during observation, shows malignant features including a thick wall, multiple irregular septation, and nodules, and/or has increased CEA or CA 19.9, positive cytology, or mucin in the aspirate, it should be surgically removed. Often, cysts detected in the elderly and fit patient are removed regardless of the features because of the increased incidence of malignancy in these lesions.

References 1. Bipat S, Phoa SS, van Delden OM et al (2005) Ultrasonography, computed tomography and magnetic resonance imaging for diagnosis and determining resectability of pancreatic adenocarcinoma: a meta-analysis. J Comput Assist Tomogr 29: 438-445 2. Midwinter MJ, Beveridge CJ, Wilsdon JB et al (1999) Correlation between spiral computed tomography, endoscopic ultrasonography and findings at operation in pancreatic and ampullary tumours. British J Surg 86:189-193 3. McLean AM, Fairclough PD (2005) Endoscopic ultrasound in the localisation of pancreatic islet cell tumours. Best Pract Res Clin Endocrinol Metab 19:177-193 4. Canto MI, Goggins M, Yeo CJ et al (2004) Screening for pancreatic neoplasia in high-risk individuals: an EUS-based approach. Clin Gastroenterol Hepatol 2:606-621 5. Thoeni RF, Mueller-Lisse UG, Chan R et al (2000) Detection of small, functional islet cell tumors in the pancreas: Selection of MRI Imaging sequences for optimal sensitivity. Radiology 214:483-490

Ruedi F. Thoeni

6. Semelka RC, Kroeker MA, Shoenut JP et al (1991) Pancreatic disease: prospective comparison of CT, ERCP, and 1.5 T MR imaging with dynamic Gadolinium enhancement and fat suppression. Radiology 181:785-791 7. Volmar KE, Vollmer RT, Jowell PS et al (2005) Pancreatic FNA in 1000 cases: a comparison of imaging modalities. Gastrointest Endosc 61:854-861 8. Ardengh JC, de Paulo GA, Ferrari AP (2004) EUS-guided FNA in the diagnosis of pancreatic neuroendocrine tumors before surgery. Gastrointest Endosc 60:378-384 9. Maguchi H (2004) The roles of endoscopic ultrasonography in the diagnosis of pancreatic tumors. J Hepatobiliary Pancreat Surg 11:1-3 10. Vachiranubhap B, Kim YH, Balci NC et al (2009) Magnetic resonance imaging of adenocarcinoma of the pancreas. Top Magn Reson Imaging 20:3-9 11. Manfredi R, Graziani R, Motton M et al (2009) Main pancreatic duct intraductal papillary mucinous neoplasms: accuracy of MR imaging in differentiation between benign and malignant tumors compared with histopathologic analysis. Radiology 253:106-115 12. Ku YM, Shin SS, Lee CH et al (2009) Magnetic resonance imaging of cystic and endocrine pancreatic neoplasms. Top Magn Reson Imaging 20:11-18 13. Brennan DD, Zamboni GA, Raptopoulos VD et al (2007) Comprehensive preoperative assessment of pancreatic adenocarcinoma with 64-section volumetric CT. Radiographics 27:1653-1666. 14. D’Onofrio M, Zamboni G, Faccioli N et al (2007) Ultrasonography of the pancreas. 4. Contrast-enhanced imaging. Abdom Imaging 32:171-181 15. Akisik MF, Sandrasegaran K, Aisen AA et al (2006) Dynamic secretin-enhanced MR cholangiopancreatography. Radiographics 26:665-677 16. Lytras D, Connor S, Bosonnet L et al (2005) Positron emission tomography does not add to computed tomography for the diagnosis and staging of pancreatic cancer. Dig Surg 22:55-61 17. Heinrich S, Goerres GW, Schafer M et al (2005) Positron emission tomography/computed tomography influences on the management of resectable pancreatic cancer and its cost-effectiveness. Ann Surg 242:235-243 18. Ruf J, Lopez Hanninen E, Oettle H et al (2005) Detection of recurrent pancreatic cancer: comparison of FDG-PET with CT/MRI. Pancreatology 5:266-272 19. Orlando LA, Kulasingam SL, Matchar DB (2004) Meta-analysis: the detection of pancreatic malignancy with positron emission tomography. Aliment Pharmacol Ther 20:1063-1070 20. Lee TY, Kim MH, Park do H et al (2009) Utility of 18F-FDG PET/CT for differentiation of autoimmune pancreatitis with atypical pancreatic imaging findings from pancreatic cancer. AJR Am J Roentgenol 193:343-348 21. Farma JM, Santillan AA, Melis M et al (2008) PET/CT fusion scan enhances CT staging in patients with pancreatic neoplasms. Ann Surg Oncol 15:2465-2471 22. Virgolini I, Traub-Weidinger T, Decristoforo C (2005) Nuclear medicine in the detection and management of pancreatic isletcell tumours. Best Pract Res Clin Endocrinol Metab 19:213-227 23. Buchmann I, Henze M, Engelbrecht S et al (2007) Comparison of 68Ga-DOTATOC PET and 111In-DTPAOC (Octreoscan) SPECT in patients with neuroendocrine tumours. Eur J Nucl Med Mol Imaging 34:1617-1726 24. Horner MJ, Ries LAG, Krapcho M et al (eds) posted to the SEER web site (2009) SEER Cancer Statistics Review, 1975-2006, National Cancer Institute. Bethesda, MD, based on November 2008 SEER data submission, http://seer.cancer.gov/csr/1975_2006 25. Cancer Facts & Figures (2009) American Cancer Society (ACS), Atlanta, Georgia 26. Fletcher JG, Wiersema MJ, Farrell MA et al (2003) Pancreatic malignancy: value of arterial, pancreatic, and hepatic phase imaging with multi-detector row CT. Radiology 229:81-90

Diseases of the Pancreas, II: Tumors

27. Imbriaco M, Megibow AJ, Ragozzino A et al (2005) Value of the single-phase technique in MDCT assessment of pancreatic tumors. AJR Am J Roentgenol 184:1111-1117 28. Sahani DV, Kadavigere R, Saokar A et al (2005) Cystic pancreatic lesions: a simple imaging-based classification system for guiding management. Radiographics 25:1471-1484 29. Morris-Stiff G, Webster P, Frost B et al (2009) Endoscopic ultrasound reliably identifies chronic pancreatitis when other imaging modalities have been non-diagnostic. JOP 10: 280-283 30. Aslanian H, Salem R, Lee J et al (2005) EUS diagnosis of vascular invasion in pancreatic cancer: surgical and histologic correlates. Am J Gastroenterol 100:1381-1385 31. Thoeni RF (2009) Imaging of endocrine tumors. In: Heiken JP (ed) Pancreatic cancer. Contemporary issues in cancer imaging. Cambridge University Press, Cambridge, pp 104-129 32. Gouya H, Vignaux O, Augui J et al (2003) CT, EUS combined protocol for preoperative evaluation of pancreatic insulinoma. AJR Am J Roentgenol 181:987-992 33. T. Zimmer, U. Stolzel, M. Bader et al (1996) Endoscopic ultrasonography and somatostatin receptor scintigraphy in the preoperative localisation of insulinomas and gastrinomas. Gut 39:562-568

95

34. Pisegna JR, Doppman JL, Norton JA et al (1993) Prospective comparative study of the ability of MR imaging and other imaging modalities to localize tumors in patients with Zollinger-Ellison syndrome. Dig Dis Sci 38:1318-1320 35. H. Liang H, Wang P, Wang XN et al (2004) Management of nonfunctioning islet cell tumors. World J Gastroenterol 10:1806-1809 36. Spinelli KS, Fromwiller TE, Daniel RA et al (2004) Cystic pancreatic neoplasms: Observe or operate. Ann Surg 239:651-659 37. Procacci C, Graziani R, Bicego E et al (1997) Serous cystadenoma of the pancreas: report of 30 cases with emphasis on the imaging findings. J Comput Assist Tomogr 21:373-382 38. Belsley NA, Pitman MB, Lauwers GY et al (2008) Serous cystadenoma of the pancreas: limitations and pitfalls of endoscopic ultrasound-guided fine-needle aspiration biopsy. Cancer 114:102-110 39. Ogawa H, Itoh S, Ikeda M et al (2008) Intraductal papillary mucinous neoplasm of the pancreas: assessment of the likelihood of invasiveness with multisection CT. Radiology 248:876-886 40. Lee JH, Yu JS, Kim H et al (2008) Solid pseudopapillary carcinoma of the pancreas: differentiation from benign solid pseudopapillary tumour using CT and MRI. Clin Radiol 63:1006-1014

IDKD 2010-2013

Adrenal Imaging and Intervention William W. Mayo-Smith1, Isaac R. Francis2 1 Department 2 Department

of Radiology, Warren Alpert School of Medicine Brown University, Rhode Island Hospital, Providence, RI, USA of Radiology, University of Michigan, Ann Arbor, MI, USA

Introduction The objectives of this chapter are: (1) to describe the different work-ups for adrenal masses, depending on clinical scenario, (2) define adrenal incidentaloma, (3) discuss the relative risk factors for benign and malignant adrenal masses, (4) describe the imaging techniques to differentiate benign from malignant adrenal masses, and (5) discuss the recommended medical work-up of an adrenal mass. The adrenal gland is a complex organ that is made up of the catecholamine-producing medulla and the steroidproducing cortex. It is a common site of primary tumors (functional and non-functional) and of metastases. The optimal work-up for an adrenal mass depends on the patient’s clinical scenario and whether detection or characterization is the primary concern. In general, it is useful to separate adrenal work-ups into one of three algorithms: 1. Detection of an adrenal tumor in a patient with a known biochemical abnormality. 2. Staging of a patient with a known primary neoplasm. 3. Characterization of an incidental adrenal mass detected on cross-sectional imaging. This is a relevant topic as there has been an over 20fold increase in medical literature reports on this subject in the past two decades.

Detection of Biochemically Active Adrenal Tumor Biochemically active adrenal neoplasms originate in the adrenal cortex, in which case they produce an excess of either glucocorticoids, aldosterone, or androgens, or in the adrenal medulla, thus producing an excess of catecholamines. Cushing’s syndrome results from an overproduction of cortisol by the adrenal cortex and approximately 80% of these cases are due to stimulation of the adrenal glands by a pituitary adenoma. A primary adrenal cortical tumor is seen in 20% of patients with Cushing’s syndrome and 95%) for the diagnosis of pheochromocytoma, its sensitivity is only 77-90%. Recent studies have suggested that MIBG scintigraphy should be used selectively and only in patients with familial or hereditary disorders, in the detection of metastatic

Adrenal Imaging and Intervention

disease, and in patients with biochemical evidence for pheochromocytoma and negative CT or MRI. These studies also concluded that MIBG scintigraphy does not offer any added advantage in patients with biochemical evidence for a pheochromocytoma, no hereditary or familial diseases, and a unilateral adrenal mass detected on CT or MRI [1, 2]. The standard treatment of a biochemically active adrenal tumor is open or laparoscopic resection. More recently, non-invasive techniques have been described, including selective arterial embolization, percutaneous injection of acetic acid, and radiofrequency ablation.

Staging Patients with Known Carcinoma Evaluation of the adrenal gland in the oncology patient is complicated because the gland is a frequent site of metastases, but benign adrenal adenomas are also common (detected in 2-5% of autopsy series). Thus, the presence of an adrenal mass does not necessarily implicate metastases. The role of cross-sectional imaging in the oncology patient is to detect enlargement of the adrenal gland and characterize the enlargement as either benign or malignant. More recently, PET imaging has facilitated the staging of neoplasms because adrenal metastases tend to demonstrate increased activity, having a greater uptake relative to the liver, while most benign adenomas do not. More recent studies have confirmed the high sensitivity of PET/CT in detecting malignant lesions but the specificity is lower (87-97%). This loss of specificity is attributable to a small number of adenomas and other benign lesions that mimic malignant lesions [3, 4]. Depending on the primary tumor, CT or PET/CT is a useful first-line exam to stage a known neoplasm. If the patient demonstrates multiple sites of metastatic disease, then evaluation of an adrenal mass is not important. If the adrenal mass is the only abnormality, further evaluation is required to differentiate an adenoma from a metastatic focus. Currently, there are two main criteria (anatomical and physiological) used to differentiate benign adenomas from malignant adrenal masses: (1) the intracellular lipid content of the adrenal mass, which represents the anatomical difference between adenomas and metastases, and (2) differences in vascular enhancement patterns, which represent the physiological difference. Approximately 80% of benign adenomas have abundant intracytoplasmic lipid in the adrenal cortex and thus are of low density on unenhanced CT or show signal drop-off on out-of-phase chemical shift MRI (CSMRI). Conversely, most metastases have little intracytoplasmic lipid and thus do not have a low density on non-contrast CT. At a threshold of 10 HU, CT has a 71% sensitivity and 98% specificity for characterizing adrenal adenomas. While a low HU is useful to characterize lipid-rich adenomas, it is estimated that up to 20% of adenomas do not contain sufficient lipid to be of low density on unenhanced CT [5-7]. More recently, Bae et al. showed that a histogram analysis of adrenal masses

97

(evaluating microscopic levels of lipid on a pixel by pixel basis) is useful to differentiate adenomas from metastases on non-contrast- and contrast-enhanced CT [8]. According to their findings, if an adrenal mass has >10% negative pixels, it is diagnostic of an adenoma (on either noncontrast- or contrast-enhanced CT). However, due to differing results in more recent studies, this approach is still considered as “research” and its use in clinical practice remains limited. The physiological difference in perfusion between adenomas and metastases can be used to differentiate these entities. Adenomas enhance rapidly with intravenous contrast (iodinated CT contrast or MR gadolinium chelates) and also have rapid washout. Metastases also enhance vigorously with dynamic contrast but the washout of contrast is more prolonged than in adenomas. This difference in contrast washout has been exploited to further differentiate benign from malignant adrenal lesions by comparing pre-contrast HU values with dynamic and 15-min delayed HU values [9, 10]. Absolute percent washout (APW) values are calculated by the formula: (HU at dynamic CT – HU at 15-min delayed CT)/(HU at dynamic CT – HU at non-contrast CT) × 100. A value ≥60% is diagnostic of an adenoma. Relative percent washout (RPW) is used when a non-contrast CT value is not available and the dynamic enhanced values are compared to 15-min delayed scans. RPW is calculated by the formula: (HU dynamic CT – HU 15-min delayed CT)/HU dynamic CT × 100, and a value >40% is diagnostic of adenoma. Adenomas can be differentiated from metastases using CSMRI if the patient has a non-diagnostic CT, is allergic to iodinated contrast, or in young patients, in whom radiation exposure is an issue [11, 12]. Most adrenal adenomas contain sufficient intracellular lipid and lose signal on the out-of-phase image compared to the spleen. Visual analysis is adequate in most cases to make this observation, but quantitative methods, such as the signal intensity index, may also be useful [13, 14]. If the CT, MRI, or PET findings are equivocal, adrenal biopsy using CT guidance should be performed, particularly to stage a lung carcinoma in patient who has no other sites of metastatic disease, as this may determine whether surgical resection is a therapeutic option. The role of adrenal biopsy has evolved in the last few years; in addition to the above indication of an indeterminate adrenal mass, adrenal biopsy can also be used to confirm metastatic disease to the adrenal glands in patients with suspected solitary adrenal metastasis. CT-guided biopsy has been shown to be safe, with a diagnostic yield of 8396% and a 3% complication rate [15].

Evaluation of an Incidentally Discovered Adrenal Mass As the indications for abdominal imaging (particularly CT) continue to increase, so does the detection of the incidental adrenal mass-given the high prevalence of

98

adenomas in the general population (3-7%) [16-18]. In general, the overwhelming majority of incidentally discovered adrenal masses (incidentalomas) are benign in a patient with no known malignancy [19]. An adrenal incidentaloma is defined as “an unsuspected and asymptomatic mass (≥1 cm) detected on imaging exams obtained for other purposes”. Risk factors for an incidental adrenal mass being malignant include lesion size, change in size, and their occurrence in a patient with a history of malignancy. For patients with no history of malignancy, most small (3 mm) from superficial disease, the accuracy of T2-weighted MRI, dynamic MRI, and contrast-enhanced T1-weighted imaging is 76, 98, and 63%, respectively [16]. Stage IB is defined as clinically visible lesions limited to the cervix uteri and is subdivided into stage IB1, in which lesions are 4 cm in their greatest dimension. The carcinoma appears as a mass of high signal intensity in contrast to the low signal intensity of the cervical stroma on T2-weighted images. Young women with stage IA or small IB (3 mm from pelvic side wall). When the tumor extends to the pelvic side wall (pelvic musculature or iliac vessels) or causes hydronephrosis, it is defined as stage IIIB. If the tumor invades the bladder or rectal mucosa it is stage IVA. There is segmental disruption of the low signal intensity of the bladder or rectal wall or segmental thickening of the rectal wall. Prominent strands between the tumor and the rectal wall may also indicate rectal invasion. MRI can confidently exclude bladder or rectal involvement, with a negative predictive value of 100% [17]. Distant metastases define stage IVB disease. Although pelvic lymph node metastases do not change the FIGO stage, para-aortic or inguinal lymph node metastases are also defined as stage IVB. CT has a limited role in staging cervical cancer due to its low accuracy in the detection of early parametrial extension [18]. However, CT has a diagnostic accuracy of approximately 90% in staging advanced cervical carcinoma and is very useful in evaluating the presence of distant metastases [19]. PET/CT allows the identification of involved nodes when CT findings are negative, resulting in a change in management in up to 25% of patients. In the detection of recurrent cervical cancer, it has a reported sensitivity, specificity, and accuracy of 90.3, 81.0, and 86.5, respectively. This is especially valuable to exclude the presence of distant disease prior to pelvic exenteresis [20].

Ovarian Carcinoma Ultrasound enables the detection and characterization of adnexal masses but has no role in staging. It can guide biopsy of adnexal or peritoneal masses in patients deemed unsuitable for primary surgery. CT is currently the modality of choice in staging ovarian cancer and can also be used to guide the biopsy of peritoneal or adnexal disease. CT provides information on the primary tumor,

the site and size of peritoneal deposits, and the presence of enlarged lymph nodes and ascites (Fig. 3). This information stratifies those patients with non-resectable disease, for whom neoadjuvant chemotherapy would be beneficial, from those patients who should undergo primary cytoreductive surgery. The primary ovarian tumor may be seen as mixed solid/cystic tumors, which are often bilateral, or as multilocular cystic lesions with thick internal septations and solid mural or septal components. Assessment can often be made as to whether the tumor invades the pelvic side wall or rectosigmoid colon or bladder, and associated complications, such as hydronephrosis and bowel obstruction, can be identified. Peritoneal deposits can be clearly identified; they are usually seen as discrete enhancing soft-tissue nodules. Liver, lung, and renal metastases and malignant pleural effusion indicate stage IV disease. PET/CT is of value in cases of suspected recurrence in which there is an increase in the level of the tumor marker CA-125 but indeterminate findings on CT or MRI [20]. Currently, the main role of MRI is in the characterization of ovarian masses rather than the staging of histologically proven ovarian cancer. MRI is very sensitive (95%) in the detection of peritoneal metastases, which show delayed enhancement on contrast-enhanced MRI [21]. Gadolinium-enhanced MRI is comparable to laparotomy but superior to serum CA-125 levels in the detection of residual or recurrent peritoneal and serosal implants in women who have been treated for ovarian cancer [21, 22]. MRI plays a crucial role in the detection of recurrent disease. It is important to realize that secondlook surgery is no longer routine and imaging diagnosis of recurrence may obviate a second-look laparotomy since secondary cytoreduction is only justified if resection is possible with no residual tumor. Imaging findings that indicate non-resectable recurrent tumor are invasion of the pelvic side wall, which should be suspected when the primary tumor lies within 3 mm of the pelvic side wall or when the iliac vessels are surrounded or distorted by tumor.

123

Malignant Diseases of the Female Genital Tract

a

b

c

d

Fig. 3 a-d. Advanced ovarian carcinoma. Contrast-enhanced CT images show (a) a left pleural effusion and a subcapsular liver deposit (arrow) and (b) peritoneal deposits along the surface of the liver and gastrosplenic ligament (arrows). There are also (c) a large omental cake and serosal deposits in the right paracolic gutter (arrows), (d) bilateral solid cystic ovarian masses and a large peritoneal deposit in the Pouch of Douglas (arrows)

References 1. Pecorelli S (2009) Revised FIGO staging for carcinoma of the vulva, cervix, and endometrium. Int J Gynaecol Obstet 105:103-104 2. Sala E, Wakely S, Senior E, Lomas D (2007) MRI of malignant neoplasms of the uterine corpus and cervix. AJR Am J Roentgenol 188:1577-1587 3. Shibutani O, Joja I, Shiraiwa M et al (1999) Endometrial carcinoma: efficacy of thin-section oblique axial MR images for evaluating cervical invasion. Abdom Imaging 24:520-526 4. Manfredi R, Mirk P, Maresca G et al (2004) Local-regional staging of endometrial carcinoma: role of MR imaging in surgical planning. Radiology 231:372-378 5. Sala E, Crawford R, Senior E et al (2009) Added value of dynamic contrast-enhanced magnetic resonance imaging in predicting advanced stage disease in patients with endometrial carcinoma. Int J Gynecol Cancer 19:141-146 6. Fujii S, Matsusue E, Kigawa J et al (2008) Diagnostic accuracy of the apparent diffusion coefficient in differentiating benign from malignant uterine endometrial cavity lesions: initial results. Eur Radiol 18:384-389 7. Shen SH, Chiou YY, Wang JH et al (2008) Diffusion-weighted single-shot echo-planar imaging with parallel technique in assessment of endometrial cancer. AJR Am J Roentgenol 190:481-488 8. Harry VN, Semple SI, Gilbert FJ, Parkin DE (2008) Diffusionweighted magnetic resonance imaging in the early detection of response to chemoradiation in cervical cancer. Gynecol Oncol 111:213-220

9. Fujii S, Matsusue E, Kanasaki Y et al (2008) Detection of peritoneal dissemination in gynecological malignancy: evaluation by diffusion-weighted MR imaging. Eur Radiol 18:18-23 10. Rockall AG, Sohaib SA, Harisinghani MG et al (2005) Diagnostic performance of nanoparticle-enhanced magnetic resonance imaging in the diagnosis of lymph node metastases in patients with endometrial and cervical cancer. J Clin Oncol 23:2813-2821 11. Ascher SM, Reinhold C (2002) Imaging of cancer of the endometrium. Radiol Clin North Am 40:563-576 12. Hricak H, Rubinstein LV, Gherman GM, Karstaedt N (1991) MR imaging evaluation of endometrial carcinoma: results of an NCI cooperative study. Radiology 179:829-832 13. Nicolet V, Carignan L, Bourdon F, Prosmanne O (2000) MR imaging of cervical carcinoma: a practical staging approach. Radiographics 20:1539-1549 14. Okamoto Y, Tanaka YO, Nishida M et al (2003) MR imaging of the uterine cervix: imaging-pathologic correlation. Radiographics 23:425-445 15. Scheidler J, Heuck AF (2002) Imaging of cancer of the cervix. Radiol Clin North Am 40:577-590,vii 16. Seki H, Takano T, Sakai K (2000) Value of dynamic MR imaging in assessing endometrial carcinoma involvement of the cervix. AJR Am J Roentgenol 175:171-176 17. Rockall AG, Ghosh S, Alexander-Sefre F et al (2006) Can MRI rule out bladder and rectal invasion in cervical cancer to help select patients for limited EUA? Gynecol Oncol 101:244-249 18. Hricak H, Gatsonis C, Chi DS et al (2005) Role of imaging in pretreatment evaluation of early invasive cervical cancer:

124

results of the intergroup study American College of Radiology Imaging Network 6651-Gynecologic Oncology Group 183. J Clin Oncol 23:9329-9337 19. Mitchell DG, Snyder B, Coakley F et al (2006) Early invasive cervical cancer: tumor delineation by magnetic resonance imaging, computed tomography, and clinical examination, verified by pathologic results, in the ACRIN 6651/GOG 183 Intergroup Study. J Clin Oncol 24:5687-5694 20. Subhas N, Patel PV, Pannu HK et al (2005) Imaging of pelvic malignancies with in-line FDG PET-CT: case exam-

Evis Sala, Susan Ascher

ples and common pitfalls of FDG PET. Radiographics 25:1031-1043 21. Ricke J, Sehouli J, Hach C et al (2003) Prospective evaluation of contrast-enhanced MRI in the depiction of peritoneal spread in primary or recurrent ovarian cancer. Eur Radiol 13:943-949 22. Low RN, Duggan B, Barone RM et al (2005) Treated ovarian cancer: MR imaging, laparotomy reassessment, and serum CA-125 values compared with clinical outcome at 1 year. Radiology 235:918-926

IDKD 2010-2013

Magnetic Resonance Imaging of Prostate Cancer Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer Department of Radiology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands

Introduction With a total of 192,280 new cases predicted for 2009, prostate cancer (PC) now accounts for 25% of all new male cancers diagnosed in the USA [1]. Furthermore, in their lifetime, one in six men will be clinically diagnosed with PC, although many more will be found to have histological evidence of PC at autopsy [2-4]. Presently, approximately one in ten men will die of PC [5, 6]. The ever-aging population and more widespread use of the blood prostate-specific antigen (PSA) test [7, 8], as well as the tendency to apply lower cut-off levels for this test [9], will further increase the diagnosis of PC [10]. An elevated PSA level, abnormal changes in PSA level and dynamics (such as PSA velocity or doubling time), or an abnormal digital rectal examination are biological indicators signaling an increased risk of PC. With the improvement and wider range of curative therapies, the detection and subsequent exact localization of PC have become increasingly important because of their influence on treatment strategy [11, 12], in particular, laparoscopic (robotic) radical prostatectomy and intensity-modulated radiation therapy (IMRT) [13]. The urologist’s inability to palpate the operating field during laparoscopic surgery makes it even more crucial to precisely localize the cancer. Moreover, the urologist must determine whether the cancer is near a neurovascular bundle since this affects the decision of whether or not to perform nerve-sparing prostatectomy [14]. IMRT also necessitates accurate PC localization. While the prostate receives a standard dose of radiation, a higher (boost) dose can be given to dominant intraprostatic lesion(s) since it is those lesions that regularly appear to be the sites of recurrent disease [15]. Furthermore, precision radiation dosimetry will decrease radiation complications, particularly rectal wall toxicity [16], thereby likely diminishing the development of postradiation rectal cancer [17]. In order to determine the optimal treatment for each patient, it is necessary to thoroughly evaluate him and to determine the cancer’s characteristics. In this regard, laboratory values (PSA level and dynamics), the results of the digital rectal examination (clinical staging), and histopathological prostatic biopsy findings (Gleason

score) are important aspects. Additionally, however, magnetic resonance imaging (MRI) can play an important role in detecting and localizing those areas most reflective of the actual aggressiveness of the cancer. This directly influences patient assessment and may lead to important changes in treatment strategy, which can mean the difference between treatment success and failure. In the mid-1980s, the first prostate MRI examinations were performed. Since that time MRI has evolved from a promising technique into a mature imaging modality for PC imaging [18, 19]. Beside anatomical information, MRI provides functional tissue-characteristic information. Multiparametric MRI consists of a combination of anatomical T2-weighted imaging and functional MRI techniques such as dynamic contrastenhanced MRI (DCE-MRI), diffusion-weighted imaging (DWI), and 1H MR-spectroscopic imaging (MRSI). Within a multiparametric MRI examination the relative value of its component techniques differ. In addition to T2-weighted MRI, which mainly assesses anatomy, MRSI [20] can add specificity for PC detection, while DCE-MRI [21] and DWI [22] are both very sensitive and very specific. The clinical challenges in the work-up of patients with either suspected or proven PC include detection, localization, TNM staging, determination of cancer aggressiveness, follow-up of patients in active surveillance protocols, and determination of the site and extent of cancer recurrence after therapy. This chapter reviews the MRI anatomy of the prostate and the basic MRI techniques that can be applied in PC, and discusses the clinical role of this imaging modality. At the end of this chapter, three clinically applicable protocols are provided.

Magnetic Resonance Imaging: Anatomy In order to effectively apply the various MRI techniques, it is important to first understand the prostate’s normal anatomy and its intrinsic age-related changes. The superior part of the prostate is called the base while its most caudal part is referred to as the apex, analogous to the

126

a

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

b c

d

e

Coronal

Sagittal Axial Fig. 1 a-e. Normal prostate with signs of benign prostatic hyperplasia (BPH) as seen on sagittal (a, d), coronal (b), and axial (c, e) highresolution T2-weighted images. The peripheral zone is white, BPH is blue, the urethra is yellow, and the seminal vesicles are green

anatomy of the heart. The prostate consists of three zones: (1) the peripheral zone, located posteriorly and caudally at its middle portion; (2) the transition zone, located interiorly, around the urethra; and (3) the central zone, which is posterior and superior to the transition zone. Ventral to the prostate is the anterior fibromuscular stroma. In aging, an important frequent change in prostate zonal anatomy occurs, namely, the transition zone becomes hypertrophic (as in benign prostatic hyperplasia), thus compressing the central gland. Consequently, most men who are imaged for prostate cancer have only two identifiable compartments in the prostate, the hyperplastic transition zone surrounded by the peripheral zone (Fig. 1). Up to 70-80% of PCs are located in the peripheral zone [23], with an overall analysis of these cancers showing that they are homogeneously distributed across the entire zone [24]. Additionally, over half of the prostates examined contained two or more distinct cancer foci [25]. Nevertheless, while up to 20-52% of all PCs originate in the transition zone, only a small percentage (3.6-25%) of these cancers occur solely in that zone [24, 26], and many such patients will have foci of concurrent peripheral-zone cancer [23, 27, 28]. Thus, a solitary transition zone cancer is rare in the general PC population.

Magnetic Resonance Techniques and Their Role in Detection and Localization For evaluation of the prostate, anatomical (high-resolution) MRI can be combined with functional and metabolic information. DWI, dynamic MRI, and MRSI provide information about the motion of free water molecules and thus about cellular density (neo-)vascularization and metabolism, respectively. These different types of information can be combined into a multiparametric MRI examination.

T2-Weighted Imaging Compared to CT computed tomography (CT) scanning, MRI has a high soft-tissue contrast resolution (Fig. 2). The use of a disposable endorectal coil combined with other external coils at 1.5 Tesla (T) increases the softtissue contrast significantly and is now the accepted clinical standard for MRI of the prostate, when information about submillimeter extracapsular penetration is of clinical importance [29]. A drawback is the extra time required for inserting and checking the position of the endorectal coil as well as the substantial expense and patient discomfort.

127

Magnetic Resonance Imaging of Prostate Cancer

a

b

Fig. 2 a, b. A 60-year-old patient with a urinary catheter, and stage 2b prostate cancer (PC) in the left peripheral zone. a Axial MDCT does not well-delineate the prostate and the tumor is not visible. b T2-weighted axial MRI shows good prostate delineation and tumor in the left peripheral zone (arrow) without capsular penetration

a

Fig. 3 a, b. a On axial T1-weighted MRI, the post-biopsy hematoma (H (H) is white. b On the T2-weighted image, both the hematoma (H (H), benign prostatic hyperplasia ((B), and tumor (*) are of low signal intensity

On MRI, PC typically appears as an area of low signal intensity within the brighter, healthy peripheral zone, as seen using a T2-dominated sequence [30-32] (Figs. 2b, 3). In the central gland, PC is not as clearly discernible because the transition zone generally has lower signal intensity than the peripheral zone and is more inhomogeneous due to the architectural changes induced by benign prostatic hyperplasia, which may mimic PC. A recent study showed that a homogeneously low T2 signal intensity and lenticular shape were significantly associated with the presence of transitionzone [4, 33]. It was reported that, relative to muscle, cancers with higher Gleason scores had lower signal intensities than cancers with low Gleason scores [32]. However, the number of patients in that study was limited. In a comparison of T2-weighted MRI with prostatectomy specimens, MRI attained 52-83% sensitivities in PC localization, while specificities were somewhat lower (46-88%). A study that directly compared endorectal MRI with digital rectal examination and transrectal ultrasound (TRUS)-guided biopsy localization revealed significant

incremental value from MRI [34]. In patients subjected to multiple prior negative TRUS-guided biopsies, anatomical MRI by means of T2-dominated acquisition plays an important role. In this patient population, a sensitivity of 83% and positive predictive value of 50% for MRI have been established [35]. Post-biopsy hemorrhage causes areas of low signal intensity on T2-dominated sequences, thereby making PC detection more difficult. However, it was shown recently that the amount of hemorrhage was significantly lower in areas of cancer than in healthy tissue [36].

Diffusion-Weighted Imaging This non-invasive technique measures the fractional anisotropy of water molecules within the prostate, expressed in apparent diffusion coefficient (ADC) mapping. Thereby, the movement of water molecules in cancer tissue has been shown to be more restricted, thus producing lower ADC values [37, 38]. In a recent study of 38 patients who underwent DWI at 1.5 T with an endorectal coil, the mean ADC values of regions of interest

128

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

Dynamic Contrast-Enhanced MRI

Fig. 4. Axial T2-weighted image obtained from a 70-year-old male with PSA = 35 and six negative TRUS biopsies shows a homogeneous area of low signal intensity in the right ventral prostate (arrows)

The use of intravenous contrast agents further enhances the localization accuracy of MRI. Endorectal DCEMRI, in which the contrast agent concentration is followed over time [46], discriminates between healthy prostate tissue and PC [47]. Early contrast enhancement and high (relative) peak enhancement are the most accurate predictors of PC of the peripheral zone, while fast washout of contrast agent and high permeability of the blood vessels (Fig. 6) are most sensitive for transition-zone PC [48, 49]. A recent study showed that the area under the receiver operating curve (AUC) for localizing PC increased significantly, from 0.68 with regular anatomical MRI to 0.91 with contrast-enhanced MRI. However, the limitations associated with the use of contrast agents are the lack of a uniform threshold, the low sensitivity for cancer, higher costs, and possible adverse reactions, of which the most serious is nephrogenic systemic fibrosis [50, 51]. 1H Magnetic Resonance Spectroscopic Imaging

Additionally, MRSI (Fig. 7) can be added to the imaging protocol to provide metabolic information based on the tissue citrate, choline, and creatine concentrations, and their relative ratios within the prostate. This is highly informative since the ratio between choline and citrate is markedly altered during the transformation from healthy to malignant prostate cells [52, 53] and an increasing choline+creatine/citrate ratio has been correlated with higher Gleason scores [54]. Presently, threedimensional (3D) MRSI of the entire prostate can be performed [55], thereby aiding in the diagnosis of

Fig. 5. Same patient as in Fig. 4. The ADC map (color coded) shows restriction in the area suspicious for tumor on this T2-weighted image (arrows)

within PC tissue were significantly lower than those within healthy prostate tissue [39]. In preliminary studies, the combination of this technique with MRSI [39] or T2weighted imaging [40] significantly improved localization accuracy (Figs. 4, 5). This was confirmed in a more recent study in 37 patients, in which sensitivity increased from 51% for T2-weighted imaging to 71% when the latter was combined with DWI [41]. In a recent multiparametric analysis, DWI was the best-performing parameter in localization [42]. Preliminary studies at 3 T have shown promising results [43-45].

Fig. 6. Same patient as in Fig. 4. Ktrans-map obtained with dynamic contrast-enhanced MRI shows pathological enhancement in area suspicious for tumor on this T2-weighted image (arrows)

129

Magnetic Resonance Imaging of Prostate Cancer

a

b

c

Fig. 7 a-c. Images obtained from a 65-year-old male with stage T3a PC in the left peripheral zone. The T2-weighted image (a) shows the tumor. MRSI of the right peripheral zone (b) shows low the choline and high citrate peak, whereas tumor in the left peripheral zone (c) shows high choline

central-gland PC (Fig. 7). The addition of 3D MRSI to MRI increases localization accuracy, by raising the specificity to as high as 91% [56]. However, a limitation of MRSI is its low spatial resolution and cumbersome post-processing. Compared to systematic biopsy, PC localization by means of MRI and MRSI was found to be more sensitive (67 and 76% vs. 50%) but less specific (69 and 57% vs. 82%) than systematic biopsy [57]. With whole-mount-section histopathology as the standard of reference, 3D MRSI had a significantly larger AUC (0.80) in localizing cancer than obtained with T2weighted MRI (0.68) [58]. The combination of T2weighted imaging and MRSI information to clinical data yielded the highest accuracy (AUC 0.85) in predicting the probability that a patient has insignificant PC [59], which was significantly higher than that obtained with clinical nomograms. A recent multi-institutional American College of Radiology Imaging Network study raised doubts on the additive value of MRSI over T2weighted imaging alone [60]. However, potential factors resulting in this conclusion were the selected prostatectomy population, the small size of the average cancer focus, and the inclusion of health centers without any previous MRSI experience.

sonable detection rates of 25-55% [62, 63]. Moreover, direct MRI-GB within the MRI scanner is technically feasible and can be performed on a routine basis. In patients with one previous negative TRUS biopsy, transrectal MRIGB performed at 1.5 T has produced promising cancer detection rates of 38-56% [64-66]. Lesions >10 mm can successfully be biopsied using this approach [66]. A multiparametric MRI approach consisting of T2weighted MRI, DWI, and DCE-MRI performed at 3 T has a median MRI-guided biopsy time of just 35 min and can generate an average of four biopsy cores per patient, as recently reported by Hambrock et al. (Fig. 8) [67]. Those

MRI-Guided Biopsies Random systematic TRUS-guided biopsy has relatively low detection rates and is prone to sampling error [61]. MRI-guided prostate biopsy (MRI-GB) has the potential to increase PC detection, as multiparametric MRI can target biopsies towards regions previously determined to be suspicious for cancer. Indeed, MRI findings have been used to direct biopsies under TRUS guidance, with rea-

Fig. 8. Same patients as in Fig. 4. MRI-guided biopsy of the tumorsuspicious region (red) showed PC with a Gleason score of 4+3. The biopsy needle is highlighted in white

130

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

authors showed that a cancer detection rate of 59% can be achieved in a large cohort of patients with more than two previous negative TRUS biopsies [68]. In addition, 93% of the cancers found were clinically significant, thus not contributing to the over-diagnosis of insignificant cancers. A limitation of MRI-GB is that multiparametric MRI for tumor localization and MRI-guided biopsy need to be performed in two different sessions, as image postprocessing and exact localization of the cancer demand time. Another disadvantage is movement of the prostate during the biopsy procedure [69]. A reduction of the MRI-GB intervention time remains an important challenge, perhaps solvable by robotics. In the future, MRIbased guidance might also be used in the focal treatment of PC, such as in the form of brachytherapy or cryotherapy.

Prediction of Prostate Cancer Aggressiveness Prostate cancer aggressiveness is pathologically graded by the Gleason score, which consists of a combination of the two most prevalent Gleason grades (range 1-5) based on the architectural characteristics of PC tissue [70]. Biopsy specimens obtained from random TRUS guidedbiopsy are subject to sampling error in approximately 64% of the cases [71]; this results in incorrect Gleason scores and thus incorrect patient risk stratification, which in turn leads to under- or overtreatment [72]. Apart from a relationship between muscle-normalized signal intensity on T2-weighted MRI and cancer Gleason scores [73], a correlation between cancer visibility on T2-weighted images and aggressiveness has been suggested, with low-grade cancers being detected in 43% and high-grade cancers in 79% of such cases [74]. Moreover, (choline+creatine)/citrate ratios, as determined by MRSI, have been shown to correlate with the Gleason

a

b

score [75-77]. Preliminary results in the evaluation of ADC as a marker of cancer aggressiveness are promising; ADC values were found to negatively correlate (ρ = –0.497, p10 mm for oval ones are generally considered to be malignant [101, 102]. Both MDCT [103] and MRI [104] have a low sensitivity (36 and 39%, respectively) for diagnosing PC lymph node metastases using these size and shape criteria. In studies that have employed thresholds as small as 6 mm [105], the specificity was very high (95-100%) but the sensitivity was too low (0-25%) to be useful in regular clinical practice for the evaluation of metastatic lymph node disease [106]. Some authors advocate restricting the application of these techniques to high-risk patients (e.g., with PSA levels >20 ng/mL) in order for them to be cost-effective [107, 108]. Thus, supplementary, invasive diagnostic examinations in the form of surgical pelvic lymph node dissection (PLND) are still commonly performed.

132

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

Magnetic resonance lymphography (MRL) using a lymph-node-specific contrast agent (Combidex/Sinerem) [109, 110] is an experimental technique that, compared to PLND, has a high negative predictive value (>96%) for the detection of lymph node metastasis in extended areas. Importantly, its use can render PLND unnecessary in negative cases [111]. Bone Most metastatic bone lesions are sclerotic [112]; a 50% change in bone mineral density is needed for metastatic bone lesions to be visible on X-ray images [113]. The most commonly used first-line diagnostic test to detect or exclude bone metastases is technetium-99m-diphosphonate bone scintigraphy. However, this approach lacks specificity, such that primary skeletal diseases may generate false-positive findings. Conventional Xray examinations can be used to exclude false-positive

a

findings on bone scintigraphy that have arisen due to conditions such as trauma, degenerative joint disease, and other chronic diseases. However, conventional Xray is too insensitive for the detection of metastatic bone lesions. Lecouvet et al. evaluated the accuracy of bone scintigraphy, targeted X-rays, and MRI in 66 patients with prostate cancer, 41 of whom had bone metastases (Fig. 11) [114]. Sensitivities were 46% for bone scintigraphy alone, 63% for bone scintigraphy and targeted X-rays, and 100% for MRI; the corresponding specificities were 32, 64, and 88%, respectively. Thus MRI was significantly more sensitive than any other approach ( 0.4 ng/mL following radical prostatectomy and by a PSA value 2 ng/mL above the nadir value (after the PSA bounce) following radiation therapy [119, 120]. Whenever such an elevation of PSA occurs, the main objective is to determine whether it is due exclusively to recurrent or residual disease in or outside the prostate, since locally recurrent disease might still be cured with adjuvant radiotherapy. The use of DRE and TRUS for recurrence detection is compromised because of the difficulty in distinguishing between fibrotic changes and recurrent disease. TRUSguided biopsy has a low sensitivity and specificity with respect to recurrence following prostatectomy or radiotherapy [121, 122] whereas MRI can help in the detection of recurrent disease. One of the major advantages of MRI is that it can be used to evaluate the recurrence of PC at either local or distant sites. After external-beam radiation therapy, prostatic tissue demonstrates diffusely low signal intensity on T2-weighted images, with indistinct zonal anatomy [123]. The contrast between benign, irradiated tissues and recurrent cancer is

therefore decreased, which makes the detection of recurrence more difficult. In one of the rare studies that used radical prostatectomy specimens after radiotherapy as the standard of reference, retrospectively assessed endorectal MRI had a low to fair accuracy for the detection of local post-radiotherapy recurrence (AUC 0.61-0.75), the prediction of extracapsular extension (0.76-0.87), and evidence of seminal vesicle invasion (0.70-0.76) [124]. Functional MRI techniques are of additional value in the detection of disease recurrence following radiation therapy (Fig. 12). On MRSI, the presence of three or more suspicious voxels in a prostate half had a sensitivity of 89% and a specificity of 82% in the detection of local post-radiotherapy recurrence (n=23) [125]. Furthermore, post-radiotherapy DCE-MRI was shown to have a sensitivity of 70-74% and a specificity of 73-85% for recurrence detection [126, 127]. The addition of DWI to T2-weighted imaging (AUC 0.61) also improved the detection of post-radiotherapy disease recurrence by 27% (AUC 0.88) [128]. In a retrospective study evaluating the value of anatomical T2-weighted MRI in the detection of post-prostatectomy disease recurrence, the sensitivity was reported to be 95% and the specificity 100% [129]. Thus, also in the evaluation of post-prostatectomy disease recurrence functional MRI techniques are of additional value (Fig. 13). In

a

b

a

b

Fig. 12 a, b. A 69-year-old patient 2.5 years after external radiation beam therapy; his PSA had risen to 2.1. a T2-weighted MRI shows entirely low signal gland; no tumor can be discriminated. b DCE-MRI shows enhancement of the right half of the prostate (arrow). Biopsy revealed Gleason 5+4 PC

Fig. 13 a, b. Patient with PSA recurrence after prostatectomy. a T1-weighted image shows no lesion whereas (b) DWI shows enhancement (red circle), which was due to recurrence of the tumor (Gleason 4+3)

134

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

this regard, DCE-MRI combined with anatomical T2weighted MRI was shown to improve sensitivity and specificity [130, 131]. According to Sciarra et al. [132], DCE-MRI is even better when used in combination with MRSI, resulting in AUC values of 0.94-0.96 for the detection of local recurrence compared to values of 0.810.94 for either DCE-MRI or MRSI alone. One of the limitations in the reported studies was the use of TRUSguided biopsies as the standard of reference [131].

Protocols The European Society of Urogenital Radiology and the Royal College of Surgeons (UK) are currently working on a set of guidelines, with the first version to be published at the end of 2011 in European Radiology. Currently, three protocols can be recommended: one for detection/localization and recurrence, one for staging, and one for the assessment of nodal size and bone marrow. Unfortunately, despite the enormous clinical potential of Combidex/Sinerem has, due to the inability of the pharmaceutical companies to provide convincing data to the FDA and EMEA, this contrast agent was not approved for nodal imaging and further development has been discontinued.

Detection, Localization, Recurrence This is a fast (15 mm in length; • infrarenal neck without thrombus or severe calcification; • angulation of the infrarenal neck 2.5 cm in diameter should be considered for treatment to prevent rupture. Meticulous imaging, including selective catheter angiography and 3D imaging with CTA or MRA, is necessary before surgery or endovascular treatment. The endovascular options are embolization and exclusion with a stent graft.

Renal Artery Aneurysm The causes are arteriosclerosis, systemic vasculitis (such as polyarteritis nodosa or lupus erythematosus), fibromuscular disease, soft-tissue disorders, and trauma. Arteriosclerotic and large aneurysms are usually calcified. The risk of rupture and chronic embolization are indications for treatment. Bypass surgery, coil embolization, and stent graft implantation are the therapeutic options.

159

Abdominal Vascular Disease: Diagnosis and Therapy

Therapeutic Embolization of Gastrointestinal Bleeding Bleedings in the upper gastrointestinal (GI) tract are usually treated by endoscopic coagulation therapy. However, if endoscopic therapy fails or the bleeding source is inaccessible to endoscopy, angiographic embolization is indicated. Patients after gastric resection involving bilioenteric anastomoses and patients bleeding from hepatobiliary and pancreatic pathologies (pseudoaneurysms) should be treated primarily by angiographic techniques. Due to the extensive collateral circulation in the upper abdomen, a detailed angiographic evaluation followed by embolization of all feeding arteries is required. The most common causes of upper GI bleeding are: • gastroduodenal ulcer; • proximal jejunal tumors (endocrine carcinomas, angiofibroma, melanoma metastasis, arteriovenous malformation etc.); • pseudoaneurysm of the gastroduodenal, pancreaticoduodenal, intrapancreatic, and splenic arteries after pancreatitis or trauma, bleeding through the Wirsung pancreatic duct; • pseudoaneurysm of the hepatic artery, bleeding through the common bile duct (triad of hemobilia: abdominal colic followed by jaundice and hematemesis or melena). In lower GI bleeding, diagnosis and treatment are preferentially done by colonoscopy. However, in selected cases embolization is required. For the primary diagnosis, a contrast-enhanced CT of the abdomen (400 mg iodine/mL, flow 4 mL, volume 100 mL) in the arterial and delayed phases usually demonstrates the area of bleeding nicely. This helps to find the bleeding source

a

b

during catheter angiography. Superselective catheterization and bowel paralysis with Buscopan (Boehringer Ingelheim, Germany) are mandatory for a bleeding angiogram. The causes of lower GI bleeding are: • small bowel tumours (endocrine carcinomas, angiofibroma, melanoma metastasis, arteriovenous malformation, etc.); • Meckels diverticulum; • large bowel diverticulum bleeding (most common cause in elderly patients); • hemangiomatosis and arteriovenous malformation of the colon; • colon cancer (rectal bleeding due to hemorrhoidal disease has to be ruled out primarily). GI bleedings are preferentially embolized by coils.

Treatment for Bleeding Complications after Surgery, Trauma, and Post-partum Bleeding in the abdomen may occur from iatrogenic causes, particularly in the kidneys and the liver after percutaneous interventions. It may also be due to trauma or tumor. Frequent and typical locations and causes are renal arteriovenous fistulas due to nephrostomy (Fig. 4) or biopsy, laceration of the hepatic arteries by percutaneous manipulations, psoas and pelvic bleeding due to traumatic arterial injury, and uterine artery bleeding post-partum. Temporary occlusion of the uterine artery can be a valid alternative to emergency hysterectomy in patients with intractable bleeding resulting from an atonic uterus. In other anatomical locations, the type, source, and location of the bleeding determine the method used to safely interrupt the extravasations.

c

Fig. 4 a-c. Hematuria and shock due to renal bleeding after nephrostomy. a CT showing large perirenal hematoma and active bleeding. b Selective angiography demonstrates the bleeding site. c Control angiography after selective coil embolization

160

Johannes Lammer

Therapeutic Embolization of Tumors Tumor embolization techniques range from palliative embolization in bleeding genitoureteral tumors, to chemoembolization techniques in hypervascularized hepatic neoplasms, in particular hepatocellular carcinoma (HCC), to the definitive treatment of benign lesions such as uterine fibroids. In the liver, tumor embolization is mainly used in patients with inoperable HCC but preserved liver function (Child-Pugh A and B) and in those with neuroendocrine tumor metastases. Classical treatment is chemoembolization with doxorubicin mixed with Lipiodol (Guerbet, France), sometimes in combination with a temporary blockade of the hepatic artery by Gelfoam or other embolization particles. In randomized trials, chemoembolization of unresectable HCC has shown to be superior to supportive treatment only. New techniques include doxorubicin-loaded particles as an alternative embolization agent for HCC (Fig. 5). In the Precision V trial, more than 200 patients were randomly assigned to either transarterial chemoembolization (TACE) with drug-eluting beads or to conventional chemoembolization with Lipiodol and doxorubicin. The former group had a higher rate of objective response and disease control within 6 months of follow-up. In addition, the use of drug-eluting beads resulted in a significantly

a

higher efficacy in high-risk patients (defined by ChildPugh B cirrhosis, ECOG 1 tumor symptoms, bilobar and recurrent disease), a lower liver toxicity, but also systemic doxorubicin-related toxicity. Regional chemotherapy with irinorecan drug-eluting beads in patients with colorectal liver metastases is under evaluation. Intrahepatic radiation by radioactive β-emitting particles directly injected into the hepatic arteries is another treatment modality for primary and metastatic liver tumors.

Venous Interventions TIPS Transjugular intrahepatic portocaval shunt (TIPS) to depressurize the portal venous system was introduced into clinical medicine at the end of the 1980s. Since then, a standardized technique has been developed that allows safe implantation of this artificial connection between the portal and hepatic veins. The introduction of stent grafts instead of bare stents has also led to improved patency of the shunt tract. TIPS is indicated in patients symptomatic from portal hypertension with acute or chronic bleeding of esophageal or gastric varices and in those with intractable ascites.

b

c d

Fig. 5 a-d. Hepatocellular carcinoma before and after transarterial chemoembolization (TACE) with drug-eluting beads. a Contrastenhanced CT shows HCC in right hepatic lobe. b Arteriography shows hypervascular HCC. c Control arteriography after TACE of HCC. d Contrast-enhanced CT shows complete necrosis of the HCC

Abdominal Vascular Disease: Diagnosis and Therapy

Endoscopic techniques to treat varices are competitive in bleeders whereas in some patients with ascites there are few alternatives. Randomized trials, meta-analyses, and Cochrane data review analyses have shown that TIPS is superior to endoscopic therapy in the prevention of rebleeding and is superior to paracenteses to remove ascites. In patients with acute or subacute Budd-Chiari syndrome, TIPS can be a life-saving procedure and help to overcome the acute phase, but the approach is burdened by a relatively high re-thrombosis rate. The risks after TIPS procedure are liver failure from shunted blood volume and encephalopathy.

Embolization of the Portal Veins An intervention of increasing importance is pre-operative embolization of the right or left portal vein in order to induce hypertrophy of the contralateral hepatic lobe prior to extended hemi-hepatectomy.

Suggested Reading ASTRAL Investigators, Wheatley K, Ives N, Gray R et al (2001) Revascularization versus medical therapy for renal artery stenosis. N Engl J Med 361:1953-1962 Bax L, Woittiez AJ, Kouwenberg HJ et al (2009) Stent placement in patients with atherosclerotic renal artery stenosis and impaired renal function: a randomized trial. Ann Intern Med 150:840-848 Blankensteijn JD, de Jong SE, Prinssen M et al (2005) Dutch Randomized Endovascular Aneurysm Management (DREAM) Trial Group. Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysms. N Engl J Med 352:2398-2405 Blum U, Voshage G, Lammer J et al (1997) Endoluminal stentgrafts for infrarenal abdominal aortic aneurysms. N Engl J Med 336:13-20 Covey AM, Tuorto S, Brody LA et al (2005) Safety and efficacy of preoperative portal vein embolization with polyvinyl alcohol in 58 patients with liver metastases. AJR Am J Roentgenol 185:1620-1626 D’Amico G, Luca A, Morabito A et al (2005) Uncovered transjugular intrahepatic portosystemic shunt for refractory ascites: a metaanalysis. Gastroenterology 129:1282-1293 EVAR trial participants (2005) Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet 365:2179-2186 Greenhalgh RM, Brown LC, Kwong GP et al, EVAR trial participants (2004) Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 364:843-848 Kaatee R, Beek FJ, de Lange EE et al (1997) Renal artery stenosis: detection and quantification with spiral CT angiography versus optimized digital subtraction angiography. Radiology 205:121-127

161

Kasirajan K, O’Hara PJ, Gray BH et al (2001) Chronic mesenteric ischemia: open surgery versus percutaneous angioplasty and stenting. J Vasc Surg 33:63-71 Khan S, Tudur Smith C, Williamson P, Sutton R (2005) Portosystemic shunts versus endoscopic therapy for variceal rebleeding in patients with cirrhosis. Cochrane Database Syst Rev 18:CD000553 Khuroo MS, Al-Suhabani H, Al-Sebayel M et al (2005) BuddChiari syndrome: long-term effect on outcome with transjugular intrahepatic portosystemic shunt. J Gastroenterol Hepatol 20:1494-1502 Lammer J, Malagari K, Vogl T et al, on behalf of the PRECISION V investigators (2009) Prospective randomized study of doxorubicin-eluting-bead embolization in the treatment of HCC: results of the PRECISION V study. Cardiovasc Intervent Radiol 12 [Epub ahead of print] Leertouwer TC, Gussenhoven EJ, Bosch JL et al (2000) Stent placement for renal artery stenosis: where do we stand? A metaanalysis. Radiology 216:78-85 Mann SJ, Pickering TG (1992) Detection of renovascular hypertension: state of the art 1992. Ann Intern Med 117:845-853 Perler AB, Becker GJ (1998) Vascular intervention – a clinical approach. Visceral vascular disease. Thieme, New York, Stuttgart, pp 517-637 Prinssen M, Verhoeven EL, Buth J et al (2004) Dutch Randomized Endovascular Aneurysm Management (DREAM) Trial Group. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 351:1607-1618 Rose SC, Quigley TM, Raker EJ (1995) Revascularization for chronic mesenteric ischemia: comparison of operative arterial bypass grafting and percutaneous transluminal angioplasty. J Vasc Interv Radiol 6:339-349 Saab S, Nieto JM, Lewis SK, Runyon BA (2006) TIPS versus paracentesis for cirrhotic patients with refractory ascites. Cochrane Database Syst Rev 18:CD004889 Salerno F, Cammà C, Enea M et al (2007) TIPS for refractory ascites: a meta-analysis of individual patient data. Gastroenterology 133:825-834 Soulez G, Oliva VL, Turpin S et al (2000) Imaging of renovascular hypertension: respective values of renal scintigraphy, renal Doppler US, and MR angiography. Radiographics 20:13551368 van de Ven PJ, Kaatee R, Beutler JJ et al (1999) Arterial stenting and balloon angioplasty in ostial arteriosclerotic renovascular disease: a randomized trial. Lancet 353:282-286 Van Jaarsveld BC, Krijen P, Pieterman H et al (2000) The effect of balloon angioplasty on hypertension in atherosclerotic renalartery stenosis. Dutch Renal Artery Stenosis Intervention Cooperative Study Group. N Engl J Med 342:1007-1014 Webster J, Marshall F, Abdalla M et al (1998) Randomized comparison of percutaneous angioplasty vs continued medical therapy for hypertensive patients with atheromatous renal artery stenosis. Scottish and Newcastle Renal Artery Stenosis Collaborative Group. J Hum Hypertens 12:329-335 Williams DM, Lee DY (1997) Dissected aorta, parts I-III. Radiology 203:23-44 Zheng M, Chen Y, Bai J et al (2008) TIPS vs. endoscopic therapy in the secondary prophylaxis of variceal rebleeding in cirrhotic patients: meta-analysis update. J Clin Gastroenterol 42:507-516

IDKD 2010-2013

Non-vascular Abdominal Disease: Diagnosis and Therapy Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi Division of Diagnostic and Interventional Radiology, University of Pisa, Pisa, Italy

Introduction Recent technological advances have given rise to a wide range of diagnostic and interventional imaging-guided procedures for an increasing number of abdominal parenchymal diseases, such as those involving the kidneys, adrenals, or pancreas. However, the liver, and the cirrhotic liver in particular, still represents the main field of application of abdominal imaging modalities. Imaging plays a key role in the diagnostic work-up of the nodular lesion, especially in its initial detection and characterization. In addition, the results of imaging assessment guide the therapeutic options (surgical, interventional, or palliative) and provide the standard of reference in the evaluation of tumor response.

Hepatocellular Carcinoma: Pathology The diagnosis of hepatocellular carcinoma (HCC) is based on imaging examinations in combination with clinical and laboratory findings. However, imaging of the cirrhotic liver remains challenging since regenerative nodules and pre-neoplastic hepatocellular lesions, such as dysplastic nodules, can mimic small HCCs. New imaging technologies have improved investigations into the multistep process that takes place during carcinogenesis, from regeneration towards dysplasia and full malignancy. The most important pathological alteration in the development of HCC is the derangement in the liver’s vascular supply due to progressive capillarization of the sinusoids, associated with the formation of an increasing number of unpaired arterioles. In addition, there are progressive, intracellular histological changes, including loss of the biliary polarization of hepatocytes, a derangement of their microscopic secretory structure, and the progressive depletion of Kupffer cells within nodular lesions [1]. Due to the wide spectrum of morphological changes seen in cirrhotic parenchyma, all cross-sectional imaging modalities should be applied in the detection and characterization of nodular lesions. Ultrasound (US) is the modality most commonly applied in the surveillance of cirrhotic patients and is thus considered as the first-line approach in their diagnostic workup. The high spatial

resolution of US enables the demonstration even of very small lesions (1 cm in maximum diameter, the vascular supply should be assessed as well. Moreover, the introduction of microbubble contrast agents and the development of contrast-specific scanning techniques have expanded the capabilities of US in examinations of the cirrhotic liver. The advent of second-generation contrast agents and lowmechanical-index real-time scanning techniques has been fundamental in improving the ease and reproducibility of US examinations [2]. Consequently, contrast-enhanced US has been introduced into the diagnostic flow chart as one of the imaging modalities able to demonstrate the vascular pattern of nodules >1 cm in diameter [3]. Dynamic multidetector computed tomography (MDCT) and magnetic resonance imaging (MRI) are further imaging modalities used to detect and characterize nodular vascular changes. They provide a detailed view of the hepatic parenchyma and allow a confident diagnosis of neoplasm. Importantly, these imaging evaluations are a fundamental prerequisite for guiding the therapeutic approach. The added value of MRI over other crosssectional imaging modalities is its ability to assess the components of a hepatic lesion in baseline image acquisitions. The use of tissue-specific MRI contrast media (hepatobiliary and reticuloendothelial agents) can reveal metabolic disorders that occur as a result of cellular dedifferentiation. Current cross-sectional techniques demonstrate the pathological changes associated with carcinogenesis, such that needle biopsy is no longer considered as the standard reference procedure for the definitive diagnosis of HCC [3].

Pre-neoplastic Lesions: Regeneration Regeneration is the result of the architectural rearrangement of the hepatic parenchyma in response to chronic injury. Generally, regenerative nodules are small (around 1 cm) but in some cases may have a maximum diameter v3 cm. At US examination, a typical regenerative nodule may appear as a hypoechoic or hyperechoic small nodule that enhances relative to background because of its thin hyperechoic rim, which represents perinodular fibrous

163

Non-vascular Abdominal Disease: Diagnosis and Therapy

Arterial

a

Late

b

Fig. 1 a-c. a Baseline US examination shows the presence of a large (up to 3 cm) hyperechoic nodule of the VI segment, partially exophytic, in a background of cirrhosis (arrow). b At dynamic MRI study, the nodule does not show signs of hypervascularization (appearing as hypointense on arterial phase); instead, the portal supply is still well evident, while the nodule appears slightly isointense during the late phase. c On this baseline T2-weighted image, the nodule appears as hypointense, due to its intranodular iron content, which increases signal intensity compared to the surrounding parenchyma (arrow)

tissue and which may mimic a pseudocapsule. Generally, no further examination is required. In cases of larger nodules, dynamic evaluation is required to exclude neoplastic dedifferentiation (Fig. 1 a). At dynamic contrast studies (contrast-enahnced US, MDCT, or MRI), there is no evidence of a vascular supply different than that of the surrounding parenchyma, due to the conspicuous and predominant feeding of these nodules by portal vessels (Fig. 1 b) [4]. MRI can depict some peculiarities that may differentiate regenerative nodules from other focal lesions. On baseline examination, regenerative nodules are isointense on T1- and T2weighted images due to the presence of normal liver cells. The frequent intranodular content of iron may decrease the relaxation time on T2, thus reducing the signal intensity of the nodules on these sequences (Fig. 1 c). Nodular signal intensity after the administration of tissue-specific contrast agent (hepatobiliary or reticuloendothelial) is not modified due to the preserved metabolic activity of hepatocytes and Kupffer cells.

Pre-neoplastic Lesions: Dysplasia Dysplastic nodules are classified as low-grade (LGDN) and high grade (HGDN), depending on the degree of cellular atypia and on changes in architectural structure.

c

HGDN are considered to be pre-malignant lesions, and the subsequent development of HCC from a HGDN within a period of years or even a few months has been documented [5]. Plain US and CT examinations are generally useless in the characterization of dysplastic nodules, due to the lack of specific imaging finding (Fig. 2 a). At dynamic study, the post-contrast enhancement patterns of HGDN may be highly variable. In fact, even if the main blood supply is still provided by branches of the portal vein, the presence of a vascular supply on arterial phase may be seen due to the development of sporadic unpaired arteries. However, this finding is not associated with wash-out, which instead represents the diagnostic finding for HCC (Fig. 2 b). Thus, in addition to displaying the degree of vascular supply, MRI demonstrates parenchymal alterations, thereby playing an important role in the differential diagnosis between pre-neoplastic and neoplastic lesions. In fact, at baseline MRI examination, dysplastic nodules typically show a characteristic hyperintensity on T1-weighted sequences due to the intranodular presence of glycogen or lipids, while on T2weighted sequences they may be slightly hyperintense, isointense, or even hypointense. In the hepatobiliary phase, after the administration of hepatospecific contrast agents, DNs are generally isointense or hyperintense compared to the surrounding liver

164

Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

Arterial

Late

a

b

c

Fig. 2 a-c. a At baseline US, it is possible to appreciate the presence of a suspected isoechoic nodule surrounded by a thin hypoechoic rim. b The US post-contrast examination does not show any arterial supply within the nodule, which remains hypoechoic in all post-contrast acquisitions. c On hepatobiliary phase at MRI examination, the nodule is isointense with the surrounding parenchyma, due to the preserved biliary function

parenchyma, reflecting the maintenance of biliary function but also cholestasis, which occurs in so-called green nodules (Fig. 2 c). Sometimes, the progression from dysplasia to HCC is detected in a very early phase, when it is possible to identify a focus of HCC within a pre-malignant lesion, known as “a nodule within a nodule”. On T2-weighted images, the typical appearance is a focus of high signal intensity located within a low-signal-intensity nodule and also showing the post-contrast signal behavior of HCC.

Hepatocellular Carcinoma: Imaging Findings As noted above, one of the key pathological features diagnostic of HCC is the vascular supply of the nodule. The progression from regeneration to overt HCC is characterized by neoangiogenesis, that is, the concomitant development of feeding arteries and efficient arteriovenous shunts. This pathological blood supply is well demonstrated on contrast-enhanced dynamic studies by the typical findings of wash-in during the arterial phase and subsequent wash-out (Fig. 3 a). However, a typical vascular behavior may not be present at dynamic imaging. In these cases, MRI both at baseline and following the administration of hepatospecific contrast media may lead to a definitive diagnosis of HCC. In addition, early or moderately differentiated HCC may show peculiar signal intensity on T1- and T2weighted baseline acquisitions (Fig. 3 b). On baseline T1-weighted images, HCC usually appears as a hypointense nodule because of its increased cellularity, and thus its higher amount of intracellular water; however, small, well-differentiated HCCs may show different

signal intensities, appearing as either hyperintense or isointense. Hyperintensity, as in the case of HGDN, may be related to the intracellular presence of glycogen and fat, which accumulate because of the loss of normal cellular metabolic activity. On baseline T2-weighted images, HCC usually shows mild signal hyperintensity, while small and well-differentiated tumors may be isointense to the surrounding parenchyma. In a comparison of histological data obtained from explanted cirrhotic livers with the MRI signal intensity of corresponding lesions, a relationship was found between lesion malignancy and nodular intensity on T2-weighted images [6]. It also has been shown that nodular signal intensity on T2-weighted images is significantly associated with the intranodular blood supply; in fact, signal intensity increases as the intranodular portal venous blood supply decreases [7]. Although their application has not yet been introduced into diagnostic guidelines, the use of tissue-specific contrast medium may give additional information, such as the atypical baseline or vascular pattern at dynamic study, as in the case of borderline lesions (dysplastic nodules) or well-differentiated HCCs. Regarding hepatobiliary contrast agents, the lack of contrast uptake is strongly related to overt HCC, due to the loss of normal metabolic function whereas the uptake is preserved in early HCCs, resembling that of HGDNs [8]. In daily practice, the advantages of the most recent generation of MRI contrast media can be exploited, as they illustrate the nodule’s characteristic vascular and hepatospecific phases, i.e., neoangiogenesis and lack of hepatobiliary function, respectively, which allow a highly confident diagnosis of HCC (Fig. 3 c). Specifically, reticuloendothelial system (RES) agents allow the carcinogenetic

165

Non-vascular Abdominal Disease: Diagnosis and Therapy

Arterial

Fig. 3 a-c. a MR dynamic examination reveals a nodule of the VIII segment that is hypervascular on arterial phase and shows clear cut washout in the late phase. b At baseline MRI examination, the nodules has a typical signal intensity, i.e., hypointense on T1weighted sequences and hyperintense on T2weighted sequences. c On hepatobiliary phase, the nodule is hypointense to the surrounding parenchyma, due to the complete loss of biliary function

Late

a

b

T1 w.i

pathway to be followed; for example, the progressive increase in sinusoid capillarization provides a hostile environment for reticuloendothelial cells, such that their progressive loss explains the very high signal intensity of HCC [9]. Nonetheless, due to the less consistent vascular phases, the application of RES agents is strongly limited and has largely been discontinued.

Therapy and Follow-up Nowadays, the therapeutic approach to HCC is based on surgical (transplantation and resection) and non-surgical approaches, mini-invasive modalities (percutaneous and intra-arterial therapies), and palliative approaches. The decision is based upon clinical and functional data as well as on the imaging findings, i.e., number of lesions and their size, location, degree of vascularization, and relationships with vascular and biliary structures. For example, a candidate for liver transplantation should fulfill imaging criteria, which include the presence of a single lesion measuring 6 mm in diameter and is non-compressible. These features should not be considered absolute and others should be taken into account, including edema of the mesentery, hyperemia of the wall of the appendix on color or power Doppler examination, the presence of an appendicolith and local fluid collections, or abscess formation (Figs. 2, 3). There are other conditions that may

b Fig. 2 a, b. Abortive appendicitis in a 9-yearold boy who had local peritoneal tenderness for one day, exactly at the site of the appendix, and normal temperature. a US shows borderline appendix diameter (6 mm) and slight edema of the apendiceal mesentery (echogenic triangle, arrow) but no surrounding edema. b Color Doppler US demonstrated marked hyperemia. Complaints subsided within 2 days without therapy

a

b Fig. 3 a, b. Rapidly progressing appendicitis in an 8-year-old boy who had local peritoneal tenderness exactly at the site of the appendix for 1 day and low grade fever. a US shows thickened appendix (8 mm) with edema of the appendicular mesentery (echogenic triangle, arrow) and extensive surrounding edema. b Color Doppler US demonstrated moderate hyperemia. At surgery, an inflamed appendix was removed

169

An Approach to Imaging the Acute Abdomen in the Pediatric Population

cause the appendix to become thick-walled and dilated; these include cystic fibrosis, Henoch-Schonlein purpura, and inflammatory bowel diseases.

Malrotation See the chapter “Malrotation: Techniques, Spectrum of Appearance, Pitfalls, and Management”, in the “Kangaroo” section of this syllabus for a review of this important cause of acute abdomen in the pediatric population.

Intussusception Intussusception is not an uncommon phenomenon in children and is one of the commoner causes of the acute abdomen in those between 6 months and 5 years of age. The vast majority of intussusceptions arise in the ileum and are either ileocolic or ileo-ileocolic. They are thought to occur because of hyperplasia of the lymphoid tissue in the ileum, possibly as a result of a viral infection. There are other types of intussusceptions that may relate to pathological lead points or gastro-jejunostomy tubes, or that are seen in the post-operative period. These are discussed separately at the end of this chapter. Another type of intussusception is the benign small bowel intussusception. This is often an incidental finding and does not present as an acute abdominal emergency. This entity is discussed separately in the chapter “Pediatric Intestinal Ultrasonography”, in the “Kangaroo” section of this syllabus.

intussusception promptly and accurately. The diagnosis can be made by US, AXR, or contrast studies of the colon. Ultrasonography has been shown in many series to be 100% accurate in depicting the presence or absence of the common types of ileocolic or ileo-ileocolic intussusceptions in children. These lesions have a characteristic sonographic appearance and are usually found just under the abdominal wall, most commonly on the right side of the abdomen (Fig. 4). Since it is a non-invasive procedure and because of its accuracy, sonography is the modality of choice for the evaluation of patients suspected of having an intussusception. The sonographic appearance of intussusception was excellently reviewed in a 1996 article by del-Pozo et al. (see “Suggested Reading”). Some of the characteristic signs of an intussusception can be seen on AXR, including the meniscus sign, target sign, and, less commonly, a soft-tissue mass (Fig. 5).

Ileocolic and Ileo-ileocolic Intussusceptions Diagnosis Children presenting with ileocolic or ileo-ileocolic intussusception commonly do not present with the classical clinical triad of abdominal pain, red currant jelly stool, and a palpable abdominal mass. Instead, the presentation may be non-specific. For this reason, the clinician often has to rely on imaging procedures to diagnose or exclude

a Fig. 4 a, b. Sonograms of children with ileocolic intussusception. a Transverse scan through an intussusception shows the typical target sign, which can be easily detected just deep to the abdominal wall anteriorly. The characteristic of this target sign is multiple concentric rings representing alternating mucosal and muscular layers. b Transverse scan through an intussusception shows the presence of a lymph node (arrow) within the mesentery which has been drawn into the intussusception between the layers of the intussusceptum

Fig. 5. Child with ileocolic intussusception that extended into the transverse colon. Abdominal radiograph shows a gasless right upper abdomen due to the presence of the intussusception on the right. The arrows indicate a soft-tissue mass that represents the intussusception, which is in the transverse colon

b

170

Alan Daneman, Simon G. Robben

However, in our institutions, we have relied on the plain radiograph only in those instances in which there is a clinical consideration of peritonitis. In this clinical setting, AXR is essential to exclude perforation, which is the major contraindication for attempted enema reduction. Contrast enema represents a more invasive diagnostic procedure, requiring radiation. Moreover, in a study by Daneman and Navarro, in 2003, in which patients were administered contrast enema for diagnosis, 50% of the enemas were negative for intussusceptions (see “Suggested Reading”). Using sonography as the diagnostic modality enables one to avoid performing unnecessary contrast enemas in those patients without intussusception. Reduction According to the many series in the recent literature, the reduction rate of intussusception ranges from 80% to as high as 95%. These series have used either fluoroscopic or sonographic guidance for reducing the intussusception and either hydrostatic (barium, water-soluble contrast, saline) or pneumatic reduction. The fact that different techniques have been used with similar success rates suggests that it is not important which technique is used. Non-operative reduction of an intussusception should only be attempted after the surgical team has evaluated the patient and the patient is clinically stable, well-hydrated, has no evidence of peritonitis, and has an intravenous line in place. The major contraindications to administering the enema are the clinical findings of peritonitis or shock or signs of perforation on an abdominal radiograph. In order to improve the reduction rate, delayed, repeated reduction attempts can be used as long as the intussusception moves in response to the initial attempted reduction and the child becomes asymptomatic and maintains stable vital signs. It has been shown that this approach is safe and effective, with a good success rate. Navarro et al., in a study published in 2004 (see “Suggested Reading”), used this approach in approximately 15% of patients with intussusceptions, achieving successful reduction in 50% of those intussusceptions not reduced on the first attempt. There does not appear to be a fixed optimal timing between attempts, and delayed second or third attempts can be made several hours after the first. Pathological Lead Points Pathological lead points are found in about 5-7% of all intussusceptions. The commonest are Meckel diverticulum, polyps, Henoch-Schonlein purpura, and cystic fibrosis. Less common causes are lymphoma, duplication cyst, and various inflammatory lesions of the bowel. Management of these patients remains a challenge. Contrast or air enema techniques are not always diagnostic in documenting the presence of a pathological lead point. Sonography is extremely useful in this regard as it may depict two-thirds of pathological lead points, providing a specific diagnosis in one-third of these cases (Fig. 6).

Fig. 6. A 3-month-old child with an intussusception due to a duplication cyst. In the transverse scan, the duplication cyst (C) is easily identified as a pathological lead point

However, it remains a diagnostic challenge as to how to search for pathological lead points in those patients in whom there is a high index of suspicion for this lesion and in whom the sonogram is negative. In such cases, the choice of which other imaging modalities to use will depend on the clinical situation in each particular patient. We recommend attempted enema reduction in all patients with a lead point if there is no contraindication to non-operative reduction. Postoperative Intussusception and Intussusception with Gastroenterostomy Tubes Intussusception may occur as a complication in 3 mm Enlarged bladder

175

Imaging Uronephropathies in Children

dilatation can also result from vesicoureteric reflux (VUR) and intravesical obstruction, especially in male fetuses. In most cases, the US evaluation will be able to differentiate between etiologies. In some patients, especially those with bilateral and complex uropathies, fetal MR imaging will provide additional information. Other organ malformations also can be associated with UT dilatation; therefore, the US survey should be as meticulous and complete as possible. Chromosomal analysis may be indicated in selected patients. The prognosis of a uropathy will depend upon the type and extent of the anomalies. Amniotic fluid volume is important to the prognosis as well; oligohydramnios, thought to be related to decreased urine production, is a poor prognostic indicator. It is of utmost importance that any relevant information is correctly transmitted to the postnatal team that will be in charge of caring for the newborn.

Postnatal Management of Fetal Pelvis Dilatation Certain conditions require immediate postnatal confirmation and therapeutic maneuvers, such as obstructive posterior urethral valves and prolapsed ectopic ureterocele into the urethra. In those cases, US and micturating cystourethrography (MCU) should be performed directly after birth. In all other cases, the work-up can be planned without urgency. Patients with ureterovesical junction obstruction (UVJO) and complex uropathies should be put on prophylactic chemotherapy at least until the final diagnosis is made. An algorithm based on US examination is presently applied by most teams (Fig. 1). Micturating voiding urethrography is only applied if US displays a significant anomaly (Table 1). MCU is used to detect high-grade VUR and urethral anomalies. If VUR is not present, complementary imaging is necessary to determine the precise

origin of the dilatation. Renal function is assessed through isotopic studies, while complicated UT malformations are best evaluated by MR imaging. The latter is particularly helpful for the assessment of a very dilated UT and complicated duplex-kidney systems. The type of treatment (conservative or surgical) will depend upon the diagnosis, renal function on follow-up, and complications. Today, the trend is increasingly toward a conservative approach. The length of follow-up must be adapted to the type of anomaly as well as to clinical and imaging follow-up. Key Points – Fetal renal dilatation is a common finding during obstetric US. – Thresholds of 4 and 7 mm during the 2nd and 3rd trimester, respectively, are widely accepted. – Postnatally, these patients must be further evaluated by US; voiding cystourethrography is performed only if an anomaly is found at birth or at the age of one month. – The trend is toward a more conservative approach to treatment, based on clinical and imaging follow-up.

Imaging Cystic Kidneys in Children Introduction Renal cystic diseases may be discovered or suspected at any stage during fetal life or at any age in childhood. They encompass a large number of conditions that can be separated into those with or without hereditary transmission. Imaging, mainly US, plays an important role in differentiating between the various types of cystic diseases as it shows the features of renal involvement as well as associated anomalies.

US : 1st US around day 5

abnormal: pelvis ≥7 mm + dilated calices, or other anomalies VCUG

normal

normal US at 1 mo

abnormal

US at 3 mo

abnormal pelvis ≥ 10 mm other malformation, “extended criteria” normal

pelvis ≥10 mm

Stop follow-up

pelvis >10 (15) mm

further morphological & functional evaluation: scintigraphy, IVU, MRU …

Stop follow-up

Fig. 1. Algorithm in fetal hydronephrosis (HN). Antenatal diagnosis of mild to moderate renal pelvis dilatation [6]. VCUG, Voiding Cystourethrography

176

Cystic Kidneys in the Fetus In the fetus (and during the perinatal period), cystic renal disease should be suspected whenever bilateral hyperechoic kidneys or cysts (uni- or bilateral) are discovered during an obstetric US examination. The imaging approach to the diagnosis should be based on a detailed sonographic analysis that includes measurement of renal length, the presence or absence of cortico-medullary differentiation (CMD), and the presence, number, size, and location of the cysts. This evaluation should be completed through an analysis of the entire fetus, looking for associated anomalies. The timing of detection and the amount of amniotic fluid are the most important prognostic factors. Furthermore, a detailed clinical and familial inquiry is essential in disease evaluation. Autosomal recessive polycystic kidney disease (ARPKD) is the main diagnosis to consider in case of bilateral, markedly enlarged, hyperechoic kidneys without CMD. The prognosis is usually poor if the amniotic fluid volume is markedly decreased. In case of moderately enlarged hyperechoic kidneys, three diagnoses have to be considered: 1. nephropathy due to a mutation in the TCF2 gene, 2. a milder form of ARPKD, and 3. autosomal dominant polycystic kidney disease (ADPKD). In TCF2-mediated nephropathy, CMD is absent or present and whenever cysts are visible they are typically subcortical. In the milder form of ARPKD, cysts may be observed in the medullary area. The main sonographic feature of ADPKD is a striking cortical hyperechogenicity associated with increased CMD. Whenever cysts are the main US finding in the fetal kidneys, their number and location are the main criteria for the differential diagnosis. The diagnosis of unilateral multiple cysts suggests a multicystic dysplastic kidney. Bilateral multiple cysts can be visualized in a large number of renal or syndromic diseases, the most common diagnoses being bilateral multicystic dysplastic kidney, ADPKD, bilateral obstructive dysplasia, and glomerulocystic kidneys.

Renal Cystic Diseases in Children In children, renal cystic diseases are usually discovered during US examination performed in the follow-up of a known perinatally diagnosed disease, during the work-up of syndromes diagnosed after birth, during screening in an at-risk family, or as an incidental finding. The US approach is the same as that described for the fetus. The role of imaging in the diagnosis and follow-up of renal involvement is to search for complications such as hemorrhage or urolithiasis. Specifically, an important role for US is the detection of hepatic-biliary complications. Key Points – Renal cystic diseases can be diagnosed in the perinatal period or in later childhood.

Jeanne S. Chow, Fred E. Avni

– Hyperechogenicity or cysts are cardinal findings. – A familial history, detailed clinical inquiry, and associated findings help in establishing the diagnosis.

Renal Ectopia and Duplications One of the most interesting areas of pediatric uroradiology is studying and understanding the multitude of congenital abnormalities of the urinary tract. During normal renal development, the kidneys ascend from the renal pelvis while rotating medially. If the kidneys do not ascend or ascend past their normal location in the renal fossae, they are ectopic. In some cases they are as low as the pelvis and in others as high as the thoracic cavity. If the kidneys fuse during ascent, pelvic cake kidneys, midline horseshoe kidneys, or left- or right-sided cross-fused ectopic kidneys form. Since the embryological origin of the kidneys (metanephros) is separate from that of the ureters (ureteric buds), the site of ureteral insertion is normal even if the kidney is ectopic. However, the renal blood supply from the aorta will vary depending on the level of ectopia. The ureteric bud must meet the metanephros in order for the kidney to form. Without this interaction, kidney formation is not induced. If two ureteric buds meet at the metanephric blastema, then the kidney becomes “duplex”. Ureteral duplication may be complete or, more commonly, incomplete. In incomplete ureteral duplication, a single ureteric bud bifurcates and meets the metanephros during approximately the 5th to 6th week of gestation. The two branches of the ureter may join at the level of the renal pelvis (bifid pelvis) or at the proximal, mid-, or distal ureter (bifid ureter) and terminate in a single distal ureter that inserts orthotopically into the bladder. Since the two moieties of the kidney share a common distal ureter, they behave similarly and usually appear normal. Rarely, one of the ureteral buds may be blind-ending and never appear to “reach” the kidney (blind-ending ureteral duplication). The associated kidney has a single collecting system. In complete ureteral duplication, two separate ureteric buds arise from the Wolffian duct. The lower-pole ureter is considered the analogue to the normal single-system ureter. Thus, the lower pole of the kidney has all of the same abnormalities that can affect a single-system kidney, including VUR, ureteropelvic junction obstruction (UPJO), and UVJO. The upper-pole ureter is “abnormal” and ectopic (Weigert-Meyer rule). The ectopic ureter inserts medially and inferiorly to the normal ureteral orifice, usually in the bladder. In girls, the ectopic ureter may insert below the bladder base, into the urethra or the vagina. A vaginal ectopic ureter can cause constant urinary dribbling in girls and incontinence [9]. In boys, ectopic ureters never terminate below the urinary sphincter and thus never result in incontinence; however, the ureter can terminate in Wolffian duct derivatives, including the

177

Imaging Uronephropathies in Children

seminal vesicals and vas deferens. Very rarely, three completely or incompletely separated ureters form, resulting in ureteral triplication [10]. Ectopic ureters are often obstructed but rarely reflux. If the ectopic ureter inserts into the urethra at the level of the urinary sphincter, urinary flow is obstructed or refluxes depending whether the sphincter is closed or open [11]. The more distal the ureteral insertion, the more dysplastic and dysfunctional the associated renal parenchyma. Ectopic ureters, and all the associated abnormalities, can also occur in single-system kidneys (single ectopic ureter) [12]. A ureterocele is the dilated submucosal terminal segment of the ureter. It is associated with varying degrees of ureteral obstruction and subsequent dilatation of the renal pelvis and calyces. In girls, ureteroceles are most commonly seen in association with ectopic upperpole ureters. In boys, they are most commonly associated with single-system kidneys and are orthotopic. Although ureteroceles protrude into the bladder, when the intravesical pressure equals that of the ureterocele, the ureterocele can flatten and become imperceptible (efface). When the intravesical pressure exceeds that of the ureterocele, the latter everts or intussuscepts into its ureter. Ectopic bladder-neck ureteroceles or large simple ureteroceles can prolapse into the urethra and cause bladder outlet obstruction. Key Points

sent with intermittent pain from intermittent obstruction, with hydronephrosis only evident during obstruction. To be correctly diagnosed, these children must be imaged at the time of their painful episodes [13]. The conundrum of UPJO is that we are still unable to predict whether the degree of obstruction and thus its eventual effects on renal function will improve or worsen over time. US is routinely used to describe the degree of obstruction and the appearance of the renal parenchyma. However, functional imaging studies, primarily MAG-3 studies with Lasix (MAG-3/Lasix renogram) and MR urography, are used to help quantify the degree of obstruction and the contributing function of each kidney. An obstruction of the distal ureter as it enters into the bladder results in UVJO. Most such cases are primary and due to a ureteral obstruction, although secondary UVJO can occur with an abnormally thickened bladder. The insertion of the obstructed ureter may be orthotopic (primary mega-ureter) or ectopic. An orthotopic or ectopic ureterocele may also be associated with obstruction. Primary mega-ureter accounts for the majority of the cases of UVJO. In most patients with this condition, the degree of dilatation improves over time [14] such that surgical repair is required only for a minority of affected patients. Surgery is indicated if the degree of dilatation worsens, renal function is impaired, or the obstruction is thought to be contributing to stasis and UT infections.

– Renal ectopia is due to abnormalities in the normal ascent of the kidney. – Ureteral duplication may be incomplete (more common) or complete. – The Weigert-Meyer rule states that the upper-pole ureter of a duplex kidney inserts ectopically, medially, and inferiorly to the orthotopic location. – The lower-pole ureter is the analogue of the single-system kidney.

Key Points

Urinary Tract Obstruction

Children with lower-spine abnormalities may have detrusorsphincter dysynergia, in which the bladder contracts but the urinary sphincter does not relax during voiding, resulting in chronic obstruction of the bladder outlet. The uncoordinated voiding causes urinary retention, increased bladder pressures, secondary VUR, and secondary UVJO. Similarly, children with voiding dysfunction without a neurogenic cause can also develop secondary reflux. Urethral obstruction can occur in the posterior or anterior urethra in boys whereas the urethra is rarely obstructed in girls. Posterior urethral valve obstruction is the most common congenital urethral obstruction and is caused by an obstructing membrane just below the level of the verumontanum. Anterior urethral obstruction is most commonly due to traumatic strictures and mostly located in the bulbar urethra. Anterior urethral valves or diverticula are rare. Depending on the severity of the ob-

Urinary tract obstructions occur at three main areas: the ureteropelvic junction, the ureterovesical junction, and the bladder outlet (i.e., the urethra). Rarely, the midureter can be obstructed by webs, fibrosis, or compression from the inferior vena cava, or there may be obstruction at the level of the infundibula in the kidney. On US, the normal hypoechoic medullary pyramids seen routinely in infancy and childhood should not be confused with dilated calyces or a sign of obstruction. The most common congenital obstruction of the kidney is UPJO, which is due to a stenosis at the junction of the renal pelvis and proximal ureter. Since most children are now diagnosed prenatally and followed postnatally, they rarely present with symptoms and signs of obstruction, such as infection, pain, or renal stones. Some children have UPJO due to a crossing renal artery and pre-

– Ureteropelvic junction obstruction is the most common cause of urinary tract obstruction. – Most cases of ureterovesical junction obstruction improve with time.

Voiding Abnormalities and Secondary Vesicoureteric Reflux

178

struction, the portion of the urethra proximal to the obstruction may be dilated, the bladder wall may be hypertrophied, and secondary reflux and UVJO may occur. In circumcised boys, meatal stenosis is another cause of urethral obstruction. Retrograde urethrogram under fluoroscopy or ultrasound is the best way of studying the anterior urethra whereas the posterior urethra can only be studied during voiding. Key Points – Voiding dysfunction can lead to vesicoureteral reflux. – Most obstructions of the bladder outlet in boys are congenital (e.g., posterior and anterior urethral valves) or post-traumatic (bulbar urethral stricture).

Renal Masses in Children Once a mass is established to be intrarenal, its histology can be predicted based on its appearance and on the patient’s age. Most intrarenal masses occurring in the newborn period are benign. Although rare, the most common solid intrarenal mass seen in newborns is a mesoblastic nephroma [15]. These large, solid, enhancing masses are benign, although the cellular subtype is the most aggressive and can cause paraneoplastic syndromes. These must be removed but the prognosis is excellent. In the newborn, the most frequently occurring cystic abnormality of the kidney is multicystic dysplasic kidney, which can involve the entire kidney or be segmental. The condition is due to a congenital abnormality of the kidney in which the collecting system forms as cysts, and the renal parenchyma is dysplastic and nonfunctional. Multicystic dysplasic kidney is now commonly diagnosed in utero. Over time, the cyst fluid resorbs and a tiny nub of tissue remains. These are typically treated non-surgically. The most common renal mass in toddlers is Wilms’ tumor. Children with aniridia, WAGR (Wilms’ tumor, aniridia, genitourinary abnormalities, and mental retardation), Deny-Drash syndrome, Beckwith-Weidemann syndrome, hemihypertrophy, or nephroblastomatosis are predisposed to developing this tumor. Wilms’ tumor is a solid, cystic, and often hemorrhagic mass and is far more common but radiographically indistinguishable from either clear cell sarcoma or malignant rhabdoid tumor. However, if a tumor has a large subcapsular hematoma, and if there are brain metastases, malignant rhabdoid tumor should be considered [16]. Centrally located multilocular masses of the kidney may be a multilocular cystic nephroma, which is more common in boys in childhood and in women in adult life [17]. In children over 11 years of age, renal cell carcinoma becomes more common than Wilms’ tumor, although the likelihood of either tumor is extremely rare [18]. It is crucial to confirm that the child has no clinical indicators of UT infection, because focal pyelonephritis or lobar

Jeanne S. Chow, Fred E. Avni

nephronia mimics a tumor in appearance and thus must always be considered in the differential diagnosis of a renal mass. If the renal mass is bilateral, the appearance and clinical presentation are extremely helpful in predicting the histology. If there are multiple large masses and the kidneys are also enlarged, bilateral nephrogenic rests due to nephroblastomatosis are most likely. Nephrogenic rests are remnant fetal renal tissue that never fully matured. As they have a high propensity to develop into Wilms’ tumors, these masses need frequent surveillance. If the masses are partially echogenic, angiomyolipoma should be considered, especially if the patient has tuberous sclerosis. Wilms’ tumors, lymphoma, and infections may also be bilateral. Solitary simple cysts are much less commonly seen in children than in adults. Calyceal diverticula may appear as simple cysts but they actually communicate with the adjacent calyx and can become superinfected. Delayed intravenous pyelogram, CT, or MR imaging, which show contrast within the cyst, is able to distinguish calyceal diverticula from simple cysts. If there are multiple simple cysts, especially in enlarged kidneys, ADPKD should be considered. Key Points – Most newborn renal masses are benign. – The most common renal malignancy in toddlers is Wilms’ tumor. – Focal pyelonephritis mimics renal tumors.

Imaging Renal Failure in Children Introduction Ultrasound plays a central role in pediatric imaging, particularly in pediatric nephrology, in which it helps to differentiate between the etiologies of renal failure. For some diseases, the US pattern will be specific, while for others there will be little or no parenchymal changes. The US evaluation should therefore be very meticulous and correlated to the biological and clinical data [18-20].

Sonographic Technique Renal US has to be carried out with the highest-resolution transducers, depending on the patient’s size. The use of both curved and linear transducers is essential. US studies include measurements of the kidneys and of any dilatation as well as the evaluation of renal echogenicity (cysts? calcifications?) and CMD. Doppler analysis also must be performed. In case of UT dilatation, the cause and level of obstruction must be determined, including the bladder, within the field of investigation. It might be of interest in some patients to evaluate the liver, spleen, and biliary tract, too.

179

Imaging Uronephropathies in Children

Acute Renal Failure Acute renal failure (ARF) is defined as urine production 1 cm after chemotherapy [15]. There were no false-positives reported. All cases of residual disease seen on CT as masses >3 cm and 95% of those cases in which the residual masses were 90% [4]; however, FDG-PET showed a poor performance in the detection of locoregional metastases. In the study by Fritscher-Ravens et al., there were several false-negatives, especially in patients with mucinous adenocarcinoma.

191

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging

Otherwise, the same results regarding locoregional and distant metastases were achieved with FDG-PET and PET/CT [5]. In the study by Petrowsky et al. [6], 61 patients with malignancies of the biliary tract, proven by histology or cytology, were evaluated by FDG-PET/CT and contrast-enhanced CT. PET/CT detected all gallbladder cancers (n = 14). Overall, 45 of 61 tumors were correctly identified with PET/CT (sensitivity 74%) and 40 by contrast-enhanced CT scan (sensitivity 66%). All 12 distant metastases were detected by PET/CT, but only 3 out of the 12 by CT (p 2 cm in diameter regardless of localization, with an intensity less than mediastinal blood pool structures, is considered negative. • With respect to partial volume effect, uptake above surrounding background in residual masses 1.5 cm in a patient with no evidence of lymphoma before therapy should be considered positive for lymphoma only if their uptake exceeds that of mediastinal blood pool, except in patients in whom a complete response is determined in all known lymphoma sites. In such patients, these “new lung lesions” most often correspond to infectious or inflammatory changes. • Residual hepatic or splenic lesions >1.5 cm should be considered positive if their uptake exceeds that of liver or spleen, respectively, and negative if their uptake is equal or lower than that of the surrounding organ. Diffusely increased splenic uptake (> normal liver) is compatible with splenic involvement except 10% activity has cleared from esophagus) is 50% decrease

230

in 5-hydroxyindoleacetic acid (5-HIAA) excretion in 24% (Table 1). Despite the absence of an objective response, palliation was achieved in 60% of patients and without significant side effects [24]. In view of the often indolent character of this disease, the value of a prolonged symptomatic response should not be underestimated. In a study at Duke University Medical Center, 98 patients with metastatic carcinoid were treated with 131I-MIBG. In this group, a subjective response was found to correlate with prolonged survival [25]. In patients with carcinoid tumors not qualifying for 131I-MIBG therapy because of no or insufficient uptake by the tumor, palliative treatment with high doses of unlabeled MIBG also proved beneficial in 60% of the cases, albeit with a shorter mean duration [26]. Improved biochemical and palliative effects of 131I-MIBG treatment due to enhanced tumor/non-tumor ratios by pre-dosing with non-labeled MIBG have also been reported [27]. A combination of higher doses of 131I-MIBG and unlabeled MIBG is used for therapy whenever comparative scintigraphy demonstrates a >20% increase of the tumor/nontumor-ratio following the addition of unlabeled MIBG.

References 1. Hoefnagel CA (1994) Metaiodobenzylguanidine and somatostatin in oncology: role in the management of neural crest tumours. Eur J Nucl Med 21:561-581 2. Khafagi FA, Shapiro B, Fig LM et al (1989) Labetalol reduces Iodine-131 MIBG uptake by pheochromocytoma and normal tissues. J Nucl Med 30:481-489 3. Reubi JC (1995) Neuropeptide receptors in health and disease: the molecular basis for in vivo imaging. J Nucl Med 36:1825-1835 4. Bardiès M, Bardet S, Faivre-Chauvet A et al (1996) Bispecific antibody and Iodine-131-labeled bivalent hapten dosimetry in patients with medullary thyroid or small-cell lung cancer. J Nucl Med 37:1853-1859 5. Goldsmith SJ (2009) Update on nuclear medicine imaging of neuroendocrine tumors. Future Oncol 5:75-84 6. Troncone L, Rufini V, Montemaggi P et al (1990) The diagnostic and therapeutic utility of radioiodinated metaiodobenzylguanidine (MIBG). 5 years experience. Eur J Nucl Med 16:325-335 7. Wiseman GA, Pacak K, O’Dorisio MS et al (2009) Usefulness of 123I-MIBG scintigraphy in the evaluation of patients with known or suspected primary or metastatic pheochromocytoma or paraganglioma: results from a prospective multicenter trial. J Nucl Med 50:1448-1454 8. Hoefnagel CA, De Kraker J (2004) Pediatric tumors. In: Ell PJ and Gambhir SS (eds) Nuclear medicine in clinical diagnosis and treatment, 3rd edition. Churchill Livingstone, Edinburgh, pp 195-206 9. Leung A, Shapiro B, Hattner R et al (1997) The specificity of radioiodinated MIBG for neural crest tumors in childhood. J Nucl Med 38:1352-1357 10. Vik TA, Pfluger T, Kadota R et al (2009) (123)I-mIBG scintigraphy in patients with known or suspected neuroblastoma: results from a prospective multicenter trial. Pediatr Blood Cancer 52:784-790

Cornelis A. Hoefnagel

11. Yeh SDJ, Larson SM, Burch L et al (1991) Radioimmunodetection of neuroblastoma with Iodine-131-3F8: correlation with biopsy, Iodine-131-Metaiodobenzylguanidine and standard diagnostic modalities. J Nucl Med 32:769-776 12. Hoefnagel CA, Rutgers M, Buitenhuis CKM et al (2001) A comparison of targetting neuroblastoma with mIBG and anti L1-CAM antibody mAB chCE7: therapeutic efficacy in a neuroblastoma xenograft model and imaging of neuroblastoma patients. Eur J Nucl Med 28:359-368 13. Hoefnagel CA, Delprat CC, Zanin D, van der Schoot JB (1988) New radionuclide tracers for the diagnosis and therapy of medullary thyroid carcinoma. Clin Nucl Med 13:159-165 14. Hoefnagel CA and Lewington VJ (2004) MIBG therapy. In: Ell PJ and Gambhir SS (eds) Nuclear medicine in clinical diagnosis and treatment, 3rd edn. Churchill Livingstone, Edinburgh, pp 445-457 15. Troncone L, Galli G (1991) Proceedings International Workshop on The Role of [131I]metaiodobenzylguanidine in the Treatment of Neural Crest Tumors. J Nucl Biol Med 35:177362 16. Baulieu J-L, Guilloteau D, Baulieu F et al (1988) Therapeutic effectiveness of Iodine-131 MIBG metastases of a nonsecreting paraganglioma. J Nucl Med 29:2008-2013 17. Gedik GK, Hoefnagel CA, Bais E, Olmos RA (2008) 131IMIBG therapy in metastatic phaeochromocytoma and paraganglioma. Eur J Nucl Med Mol Imaging 35:725-733 18. Safford SD, Coleman RE, Gockerman JP et al (2003) Iodine131 metaiodobenzylguanidine as an effective treatment for malignant pheochromocytoma and paraganglioma. Surgery 134:956-962 19. Hoefnagel CA (1999) Nuclear medicine therapy of neuroblastoma. Q J Nucl Med 43:336-343 20. Yanik GA, Levine JE, Matthay KK et al (2002) Pilot study of iodine-131 metaiodobenzylguanidine in combination with myeloablative chemotherapy and autologous stem-cell support for the treatment of neuroblastoma. J Clin Oncol 20:21422149 21. Voûte PA, van der Kleij AJ, de Kraker J et al (1995) Clinical experience with radiation enhancement by hyperbaric oxygen in children with recurrent neuroblastoma stage IV. Eur J Cancer 31A:596-600 22. Hoefnagel CA, de Kraker J, Valdés Olmos RA, Voûte PA (1994) 131I-MIBG as a first-line treatment in high-risk neuroblastoma patients. Nucl Med Commun 15:712-717 23. Kraeber-Bodéré F, Bardet S, Hoefnagel CA et al (1999) Radioimmunotherapy in medullary thyroid cancer using bispecific antibody and iodine-131-labeled bivalent hapten: Preliminary results of a phase I/II clinical trial. Clin Cancer Res 5:3190s-3198s 24. Zuetenhorst H, Taal BG, Boot H et al (1999) Longterm palliation in metastatic carcinoid tumours with various applications of meta-idobenzylguanidine: pharmacological MIBG, 131I-labeled MIBG and the combination. Eur J Gastroenterol Hepatol 11:1157-1164 25. Safford SD, Coleman RE, Gockerman JP et al (2004) Iodine131 metaiodobenzylguanidine treatment for metastatic carcinoid. Results in 98 patients. Cancer 101:1987-1993 26. Taal BG, Hoefnagel CA, Valdés Olmos RA et al (1996) Palliative effect of Metaiodobenzylguanidine in metastatic carcinoid tumors. J Clin Oncol 14:1829-1838 27. Taal BG, Hoefnagel CA, Boot H et al (2000) Improved effect of 131I-MIBG treatment by predosing with non-radiolabeled MIBG in carcinoid patients, and studies in xenografted mice. Ann Oncol 11:1437-1443

IDKD 2010-2013

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy Dik J. Kwekkeboom Department of Nuclear Medicine, Erasmus Medical Center, Rotterdam, The Netherlands

Introduction The development of peptide receptor scintigraphy in combination with radioiodinated somatostatin analogues allowed the in vivo demonstration of somatostatin-receptorpositive tumors in patients [1]. Later, other radiolabeled somatostatin analogues were developed, two of which subsequently became commercially available. With the advent, over the past decade, of positron emission tomography (PET) tracers for somatostatin receptor imaging, superior image quality and increased sensitivity in tumor site detection have become possible, as confirmed by several research groups. In the 1990’s, attempts at treatment with radiolabeled somatostatin analogues were undertaken in patients with inoperable and/or metastasized neuroendocrine tumors. Improvements in the peptides (higher receptor affinity) and the available radionuclides (β instead of γ emission), together with precautions to limit the radiation dose to the kidneys and bone marrow, led to better results and a virtually negligible percentage of serious adverse events.

between a successful localizing study and a disappointing one. For details of the scanning protocol, the reader is referred to the procedural guidelines for somatostatin receptor scintigraphy with [111In-DTPA0]octreotide, published by the Society of Nuclear Medicine [2].

[111In-DTPA0]Octreotide Scintigraphy: Normal Scintigraphic Findings and Artifacts Normal scintigraphic features include visualization of the thyroid, spleen, liver, and kidneys, and in some patients the pituitary gland (Fig. 1). In addition, the urinary bladder and

Somatostatin-Receptor-Based Radionuclide Imaging The many drawbacks of [123I,Tyr3]octreotide, the radioiodinated somatostatin analogue first used for imaging in patients, resulted in its replacement by the chelated and 111In-labeled somatostatin analogue [111In-DTPA0]octreotide (OctreoScan, Covidien, Petten, The Netherlands), which is commercially available and is now the most commonly used agent for somatostatin receptor imaging (SRI). The preferred dose of [111In-DTPA0]octreotide (containing at least 10 mg of the peptide) is about 200 MBq. This dose is appropriate for single photon emission computed tomography (SPECT), which shows increased sensitivity in the detection of somatostatinreceptor-positive tissues and provides better anatomical delineation than planar views. The acquisition of sufficient counts per view and the generation of spot images with a sufficient counting time – as opposed to the low count density of whole-body scans – are other important advantages of this approach and may make the difference

Fig. 1. Normal distribution in somatostatin receptor imaging (SRI). Variable visualization of the pituitary and thyroid (arrows, upper panels). Faint breast uptake can sometimes be seen in women (right middle panel, arrow). Normal uptake in the liver, spleen, and kidneys, and also some bowel activity is seen in the lower panel. Gallbladder visualization in the lower right panel (arrow). Anterior views

232

Table 1. Pitfalls and causes of potential misinterpretation of positive results with [111In-DTPA0]octreotide scintigraphy Radiation pneumonitis Accessory spleen Focal collection of stools Surgical scar tissue Gallbladder uptake Nodular goiter Ventral hernia Bacterial pneumonia Respiratory infections Common old (nasal uptake) Cerebrovascular accident Concomitant granulomatous disease Diffuse breast uptake Adrenal uptake Urine contamination Concomitant second primary tumor

Table 2. Causes of potential misinterpretation of negative results with [111In-DTPA0]octreotide scintigraphy The presence of unlabeled somatostatin, because of octreotide therapy or due to the production of somatostatin by the tumor itself, may lower tumor detectability Different somatostatin receptor subtypes have different affinities for the radioligand; variable tumor differentiation and receptor expression also influences tumor detectability. This may be important especially in patients with insulinomas and medullary thyroid carcinomas Liver metastases of neuroendocrine tumors may appear isointense because of a similar degree of tracer accumulation by the normal liver. Correlation with anatomical imaging and/or SPECT imaging may be helpful

the bowel are usually visualized to variable degrees. Visualization of the pituitary, thyroid, and spleen is due to receptor binding whereas uptake in the kidneys is for the most part due to re-absorption of the radiolabeled peptide by the renal tubular cells after glomerular filtration. While there is predominant renal clearance of the somatostatin analogue, hepatobiliary clearance via the bowel also occurs, thus necessitating the administration of laxatives in order to facilitate the interpretation of abdominal images. False-positive results of SRI with [111In-DTPA0]octreotide have been reported in virtually all cases; however, the term “false-positive” is a misnomer because it includes somatostatin-receptor-positive lesions unrelated to the pathology for which the investigation was performed (see the review by Gibril et al. [3]). The most common of these are listed in Table 1. The potential causes of a falsenegative study interpretation are given in Table 2.

Imaging Results of [111In-DTPA0]Octreotide Scintigraphy in Neuroendocrine and Other Tumors Somatostatin receptors have been identified in vitro in a large number of human neoplasias, in particular, neuroendocrine tumors (NETs) have a high incidence and

Dik J. Kwekkeboom

density of somatostatin receptors. NETs comprise a group of tumors that includes pituitary adenoma, pancreatic islet cell tumor, carcinoid, pheochromocytoma, paraganglioma, medullary thyroid cancer, and small cell lung carcinoma [4]. Tumors of the nervous system, including meningioma, neuroblastoma, and medulloblastoma, also very often express a high density of somatostatin receptors, as do tumors not classically originating from endocrine or neural cells, such as lymphoma, breast cancer, renal cell cancer, hepatocellular cancer, prostate cancer, sarcoma, and gastric cancer. In the majority of these tumors, somatostatin receptor (SR) subtype-2 is predominantly expressed, although low amounts of other SR subtypes may be concomitantly present [5]. It should also be emphasized that selected non-tumoral lesions may express SRs. For instance, SRs are expressed on the epithelioid cells of active granulomas in sarcoidosis and by inflamed joints in active rheumatoid arthritis, especially the proliferating synovial vessels [6]. Therefore, SR expression is not specific for tumoral pathologies. The most common indication for [111In-DTPA0]octreotide scintigraphy is the detection and localization of gastroenteropancreatic neuroendocrine tumors (GEPNETs) and their metastases, the staging of these patients, the follow-up of patients with known disease, and, lastly, the selection of patients with inoperable and/or metastatic tumors for peptide receptor radionuclide therapy (PRRT) [7-12].

Newer Ligands for Somatostatin Receptor Imaging 99mTc-depreotide

(Neotect) is a commercially available somatostatin analogue that has been approved specifically for use in the detection of lung cancer in patients with pulmonary nodules [13]. Due to the relatively high abdominal signal background and the impossibility to perform delayed imaging because of the tracer’s short half-life, it is less suited for the detection of abdominal NETs [14]. Analogues that are used for PET or hybrid PET/CT imaging are of particular interest because of two advantages that they have over γ-emitting analogues. First, many of them have a better affinity for SR subtype-2, the subtype most commonly expressed by NETs; likewise, there are some analogues that better target other SR subtypes and are therefore more appropriate for visualizing the respective tumors. Second, PET and the combined anatomical and functional information obtained with PET/CT provide images with high spatial resolution, which results in a higher sensitivity of this type of scanning. However, based on a review of the results obtained with these newer analogues, there are also causes for concern. Importantly, in many studies these newer analogues were compared to [111InDTPA0]octreotide scintigraphy using inadequate scanning protocols or comparisons were made between two or more new analogues such that a validated reference method was lacking. Also, the multitude of newly available PET analogues has led to a situation in which each center has to accumulate its own results on the normal findings and

233

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy

artifacts of their scanning methods. This hampers the exchange of data and their shared interpretation. It is, however, likely that one of the new PET analogues, [68Ga-DOTA0, Tyr3]octreotide or [68Ga-DOTA0, Tyr3]octreotate, will become the new standard for SRI using PET. This is due to the fact that these somatostatin analogues have a high affinity for SR subtype-2, and 68Ga is a generator-produced rather than a cyclotron-produced product, such that labeling of the compound is simplified [15]. An additional reason favoring the use of [68Ga-DOTA0,Tyr3]octreotide or [68Ga-DOTA0, Tyr3]octreotate as the standard analogue for PET imaging is that the 90Y or 177Lu-labeled counterparts of these compounds are administered for PRRT; thus, the peptide used in diagnostic imaging closely mimics the one that is used for therapy. Other radionuclide-coupled ligands that do not rely on the presence of SRs for tumor visualization have also been tested in patients with GEPNETs. The oldest of these, 123I-MIBG, performs poorly compared to [111In-DTPA0]octreotide scintigraphy [16]. PET scanning with (18F)2-fluoro-2-deoxy-D-glucose 18 ( F-FDG) has gained importance for tumor staging and the evaluation of treatment response for a number of tumor types. The method is based on glucose consumption by the tumors, such that fast-growing tumors usually show high tracer uptake. However, 18F-FDG PET is less suited for GEPNETs, because of the slow growing nature of these tumors. Therefore, the technique is recommended only in patients with negative SRI findings [17], a situation that usually correlates with more aggressive tumor behavior and faster tumor growth. Newer PET radioligands that have been clinically tested in patients with GEPNETs include 18F-DOPA and 11C-5-hydroxy-tryptophan [18]. PET with these ligands has been reported to be more sensitive than SRI with [111In-DTPA0]octreotide. These PET ligands, however, have a short half-life and therefore have to be synthesized in the close vicinity of or in the hospital where they are to be administered. Also, both 18F-DOPA and 11C-5hydroxy-tryptophan, unlike the radiolabeled somatostatin analogues used in PET, lack a sequel in PRRT.

Somatostatin-Receptor-Based Radionuclide Therapy As noted above, the functioning and non-functioning endocrine pancreatic tumors and carcinoids that make up

the GEPNETs are usually slow-growing. The treatment of tumor metastases with somatostatin analogues results in reduced hormonal overproduction and symptomatic relief in most cases. Treatment with somatostatin analogues is, however, seldom successful in terms of tumor size reduction [19]. A new treatment modality for patients with inoperable or metastasized endocrine GEPNETs is the use of radiolabeled somatostatin analogues. The majority of endocrine GEP tumors possess SRs and can therefore be visualized with SRI. A logical sequence to tumor visualization in vivo is to then treat these patients with radiolabeled somatostatin analogues. In the early phases of this approach, virtually all patients considered as candidates for PRRT had well-differentiated GEPNETs. At the time of these early studies, in the mid- to late 1990s, no other chelated somatostatin analogues labeled with β-emitting radionuclides were available, such that [111In-DTPA0]octreotide was used for PRRT. The results of these studies, in which high doses of the radionucleotide were administered to patients with metastasized NETs, were encouraging with regard to symptom relief but partial remissions (PRs) were exceptional [20, 21] (Table 3). The next generation of SR-mediated radionuclide therapy was based on the use of the modified somatostatin analogue [Tyr3]octreotide, which has a higher affinity for SR subtype-2, and a different chelator, DOTA instead of DTPA, in order to ensure a more stable binding of the intended β-emitting radionuclide, 90Yttrium (90Y). The resulting compound (90Y-DOTATOC; OctreoTher, Novartis, Switzerland), was used in several phase-1 and phase-2 PRRT trials [22-25] (Table 3) but renal insufficiency and myelodysplastic syndrome were reported as serious adverse events (SAEs). The incidence of these SAEs could, however, be dramatically reduced through adequate renal protection, achieved by the co-infusion of amino acids. Consequently, SAEs have become relatively rare, occurring in +4 SD), moderately increased (>+2 SD), or normal or small (+4SD) hyperechoic kidneys diagnosed during the late first and early second trimesters, Meckel-Gruber syndrome should be considered first, especially if the medulla appears enlarged and hypoechoic and if polydactyly and cerebral anomalies are associated. If the condition is detected during the second and third trimesters, the main diagnosis to be added to the differential would be autosomal recessive polycystic kidney disease (ARPKD) and Bardet-Biedl syndrome (BBS). In ARPKD, CMD may be partially absent, completely absent, or even reversed. A few visible cysts are rare but may be seen in utero. Oligohydramnios is a frequent finding and is associated with pulmonary hypoplasia, which confers a very poor prognosis [7-10]. In BBS, the kidneys are enlarged and hyperechoic and there is post-axial polydactyly. The other symptoms of the disease will develop after birth. Cysts can be observed already in utero or appear after birth [7-12]. In case of moderately enlarged hyperechoic kidneys (+2 SD), three diagnoses have to be considered first: 1. TCF2 mutation associated nephropathy; 2. ARPKD; 3. autosomal dominant polycystic kidney disease (ADPKD). An anomaly of TCF2 (leading to HNF-1β-related morphological anomalies) was recently shown to represent the main cause of fetal hyperechoic kidneys [11]. This mutation is associated with a wide spectrum of renal morphological and structural anomalies that histologically include glomerulocystic-type changes, cystic dysplasia, and renal agenesis. Hepatic ductular plate anomalies are commonly associated findings. In such kidneys, besides renal hyperechogenicity, CMD may or not be visible. Cysts may be detected already in utero or, more often, after birth; they are located in the subcortical area. A familial history of diabetes is a frequent finding. The involvement and extent of the kidney lesions related to ARPKD can vary from 10 to 90%, with the US appearances varying accordingly. In cases with mild involvement, the kidneys may be moderately enlarged, with a hyperechoic cortex and a few small cysts mainly within the pyramids. After birth, cysts may also develop, throughout the medulla first and within the cortex thereafter. Fetuses with mildly enlarged kidneys have a better prognosis for survival than those with massive enlargement [9]. Already in utero, ADPKD may be suspected based on a marked hyperechoic renal cortex that increases CMD. The kidneys are usually normal in size or slightly enlarged. Such finding should prompt familial inquiry. Cysts may be observed in utero but usually develop after birth [13].

Fred E. Avni

Table 2. Causes of moderately enlarged hyperechoic kidneys in the fetus TCF2 mutation Autosomal recessive polycystic kidney disease Autosomal dominant polycystic kidney disease Maternally related diseases Infection Ischemia Metabolic diseases Dysplasia Nephrotic syndromes “Transient”

Once these three diagnoses are excluded, there is a wide spectrum of other diseases that can lead to hyperechoic kidneys; clinical inquiry may suggest the diagnosis [14, 15]. Complementary examinations, such as chromosomal analysis, are directed at searching for infectious, toxic, maternally related, or ischemic causes and will help to reach a diagnosis (Table 2).

Renal Cyst(s) Discovered in the Perinatal Period A unilocular, single renal cyst occurring in otherwise normal-appearing kidneys can be detected in utero or after birth. It should be differentiated, especially if the cyst is septated, from a cystic tumor, segmental cystic dysplasia, a dysplastic upper-pole of a duplex kidney, or a urinoma. Associated urinary tract malformation and dilatation may help make the diagnosis. Noteworthy is the fact that ADPKD may start asymmetrically (with a single cyst) [16]. Whenever multiple cysts are detected, the first criterion for the differential diagnosis is uni- or bilateral involvement. Multiple cysts detected in one kidney only most often correspond to a multicystic dysplastic kidney (MCDK), which usually has a straightforward US appearance: multiple cysts of various sizes without interconnection, no recognizable normal renal parenchyma, and no central renal pelvis. MDCK should be differentiated from obstructive dysplasia (associated with urinary tract obstructive malformation), in which the dilated urinary tract is recognizable. The disease can also occur in the upper pole of a duplex kidney. MDCK evolves such that in most cases the kidney will eventually shrink. This can be followed by US [17, 18]. Bilateral multiple renal cysts can be visualized in a large number of isolated renal or syndromic diseases (Table 3) [3, 4] and may or may not be associated with global renal Table 3. Bilateral multiple cysts in the perinatal period Bilateral multicystic dysplastic kidney disease Bilateral obstructive dysplasia (+urinary tract dilatation) Autosomal dominant polycystic kidney disease Autosomal recessive polycystic kidney disease Glomerulocystic disease (subcortical cysts) Syndromes with cystic dysplasia, including (but not limited to) Ivemark syndrome, Zellweger syndrome, Meckel Gruber syndrome, Bardet-Biedl syndrome, tuberous sclerosis complex

241

Imaging Cystic Kidneys in Children

Table 4. Bilateral macrocysts Tuberous sclerosis complex Simpson-Golabi-Behmel TCF2 mutation syndrome Bilateral multicystic dysplastic kidney disease

hyperechogenicity (see above). A step-by-step approach, including detailed US analysis, familial history, and complementary examinations, will lead to the diagnosis. Amniotic fluid volume and associated morphological or chromosomal anomalies are mandatory for the prognosis. An interesting subgroup includes diseases with macro-cysts that are already obvious at birth [19, 20] (Table 4). Many renal cystic diseases present with some type of hepatic ductular anomaly. However, these are rarely detected by perinatal US. Furthermore, in some patients it may be of interest to perform fetal MRI in order to better characterize the renal involvement [21].

Cystic Renal Diseases in Childhood Renal cystic diseases in children are usually discovered during an US examination performed: (1) in the followup of a known, perinatally diagnosed disease, (2) during the workup of syndromes diagnosed after birth, or (3) during screening in an at-risk family. They may also be detected as an incidental finding in an US examination performed for other reasons. The main sonographic features of renal cystic diseases are obviously the presence of visible cysts uni-or bilaterally [1, 3]. The sonographic approach is similar to the one described in the fetus and is based on the characteristics of the cortex, medulla, and CMD as well as the size, location, and number of cysts. Certain features will orient the diagnosis [22, 23]: subcapsular cysts are indicative of glomerulocystic disease; medullary cysts suggest medullary cystic dysplasia (ARPKD); diffuse cortical cysts point to ADPKD, and macro-cysts to ADPKD or tuberous sclerosis complex. The role of US is in the follow-up of these patients once the disease has been diagnosed, including monitoring disease evolution and the potential development of complications (hemorrhage, lithiasis). Another important role for US is to verify the occurrence of hepatic complications of the diseases and the development of portal hypertension [24-26]. For this purpose, MRI may provide additional information important for patient management [27]. Nephronophthisis (NPS) deserves special mention. This genetically transmitted, autosomal recessive disease is being increasingly recognized due to improved genetic mapping. The symptoms of NPS include anemia, polyuria, and polydysia, along with chronic renal failure. The disease can be part of a syndrome (i.e., Joubert syndrome) or an isolated finding. At US examination, the kidneys are relatively small and cysts will develop at the corticomedullary junction [28, 29].

Conclusions The differential diagnosis of renal cystic disease is a difficult challenge. Cystic disease is suspected based on the discovery of hyperechoic kidney or/and cysts. It can be identified by approaching the diagnosis in a step-by-step manner that includes US analysis, familial history, and clinical evaluation.

References 1. De Bruyn R, Gordon R (2000) Imaging in cystic renal disease. Arch Dis Child 83:401-407 2. Avni EF, Garel L, Cassart M et al (2006) Perinatal assessment of hereditary cystic renal diseases: the contribution of sonography. Pediatr Radiol 35:405-414 3. Rizk D, Chapman AB (2003) Cystic and inherited kidney diseases. Am J Kidn Dis 42:1305-1317 4. Deshpande C, Hennekam RCM (2008) Genetic syndromes and prenatally detected renal anomalies. Semin Fetal Neonat Med 13:171-180 5. Winyard P, Chitty LS (2008) Dysplastic kidneys. Semin Fetal Neonat Med 13:142-151 6. Cohen HL, Cooper J, Eisenberg P et al (1991) Normal length of fetal kidneys. AJR Am J Roentgenol 157:545-548 7. De Bruyn R, Marks SD (2008) Post-natal investigation of fetal renal disease. Semin Fetal Neonat Med 12:133-141 8. Tsatsaris V, Gagnadoux MF, Aubry MC et al (2002) Prenatal diagnosis of bilateral isolated fetal hyperechogenic kidneys. It is possible to predict long-term outcome? BJOG 109:1388-1393 9. Chaumoitre K, Brun M, Cassart M et al (2006) Differential diagnosis of fetal hyperechogenic cystic kidneys unrelated to renal tract anomalies. Ultrasound Obstet Gynecol 28:911-917 10. Ickowicz V, Eurin D, Maugey-Laulom B et al (2006) MeckelGruber syndrome: sonography and pathology. Ultrasound Obstet Gynecol 27:296-300 11. Decramer S, Parant O, Beaufils S et al (2007) Anomalies of the TCF2-Gene are the main cause of fetal bilateral hyperechogenic kidneys. J Am Soc Nephrol 18:923-929 12. Cassart M, Eurin D, Didier F et al (2004) Antenatal renal sonographic anomalies and post-natal follow-up of renal involvement in Bardet-Biedl syndrome. Ultrasound Obstet Gynecol 24:51-54 13. Brun M, Maugey-Laulom B, Eurin D et al (2004) Prenatal sonographic patterns in autosomal dominant polycystic kidney disease. Ultrasound Obstet Gynecol 24:55-61 14. Slovis TL, Bernstein J, Gruskin A (1993) Hyperechoic kidneys in the newborn and young infant. Pediatr Nephrol 7:294-302 15. Nortier JL, Debiec H, Tournay Y et al (2006) Neonatal disease in NEP alloimmunization: lessons for immunological monitoring. Pediatr Nephrol 21:1399-1405 16. McHugh K, Stringer DA, Hebert D, Babiak GA (1991) Simple renal cysts in children: diagnosis and follow-up with US. Radiology 178:383-385 17. Kuwertz-Broeking E, Brinkmann OA, VanLengerke HJ et al (2004) Unilateral MDK: experience in children. BJU Internat 93:388-392 18. Aslam M, Watson AR (2006) Unilateral MDK: long-term outcomes. Arch Dis Child 91:820-823 19. Glazier DB, Fleisher MH, Cummings KB, Barone JG (1996) Cystic renal disease and TS in children. Urology 48:613-615 20. Newmann HPH, Schwarzkopf G, Henske EP (1998) Renal angiomyolipomas, cysts and cancer in TSC. Semin Pediatr Neurol 5:269-275 21. Cassart M, Massez A, Metens T et al (2004) Complementary role of MRI after US in assessing bilateral UT anomalies in the fetus. AJR Am J Roentgenol 182:684-695

242

22. Jaim M, Lequesne GW, Bourne AJ, Henning P (1997) High-resolution US in the differential diagnosis of cystic diseases of the kidney in infancy and childhood. J Ultrasound Med 16:235-240 23. Traubici J, Daneman A (2005) High-resolution renal sonography in children with ARPKD. AJR Am J Roentgenol 184:1630-1633 24. Lipschitz B, Berdon WE, Defelice AR, Levy J (1993) Association of congenital hepatic fibrosis with ADPKD. Pediatr Radiol 23:131-133 25. Premkumar A, Berdon WE, Levy J et al (1988) Emergence of hepatic fibrosis and portal HT in ARPKD. Pediatr Radiol 18:123-129

Fred E. Avni

26. Avni EF, Guissard G, Hall M et al (2002) Hereditary polycystic kidney diseases in children: changing sonographic patterns through childhood. Pediatr Radiol 32:169-174 27. Turkbey B, Ocak I, Daryanani K et al (2009) ARPKD and congenital hepatic fibrosi. Pediatr Radiol 39:100-111 28. Salomon R, Saunier S, Niaudet P (2009) Nephronophtisis. Pediatr Nephrol 24:2333-2344 29. Blowey DL, Querfeld U, Geary D et al (1996) US findings in juvenile nephronophtisis. Pediatr Nephrol 10:22-24

IDKD 2010-2013

Understanding Duplication Anomalies of the Kidney Jeanne S. Chow Departments of Urology and Radiology, Children’s Hospital, Boston, MA, USA

Introduction Learning the basic rules governing duplication anomalies of the kidney is particularly rewarding because the fundamental rules can be applied reliably to the great variety of cases and using all different imaging modalities. Although ultrasound is typically the starting point of imaging of the urinary tract in infants and children, renal duplex anomalies are also studied by micturating cystourethrography (MCU), also called voiding cystourethrography (VCUG), intravenous pyelography (IVP), magnetic resonance urography (MRU), computed tomography (CT), MAG-3 lasix renogram, and DMSA scan. While more and more duplex anomalies are being discovered in utero, most people with duplex anomalies are asymptomatic and thus may never be studied at all.

Embryology The ureteric bud must meet the metanephros in order for the kidney to form. Without this interaction, kidney formation is not induced. If two ureteric buds meet the metanephric blastema, then the kidney becomes duplex and the collecting system and ureter become duplicated (Fig. 1). Ureteral duplication may be complete or incomplete.

a

In incomplete ureteral duplication, a single ureteric bud, which is derived from the the mesonephric (Wolffian) duct, bifurcates and meets the metanephros during approximately week 5-6 of gestation. The two branches of the ureter may join at the level of the renal pelvis (bifid pelvis) or the proximal, middle or distal ureter (bifid ureter) and terminate in a single distal ureter that inserts orthotopically into the bladder. Rarely, one of the ureteral buds may be blind-ending and never appear to “reach” the kidney (blind ending-ureteric duplication). Since the upper and lower poles of the duplex kidney with a bifid ureter have a common distal ureter, the upper and lower poles typically appear similar (and normal). If there is reflux or obstruction at the end of the common ureter, both the upper and the lower pole will be affected. In complete ureteral duplication, two separate ureteric buds arise from the mesonephric duct to meet the metanephric blastema. The lower-pole ureter is the analogue of the normal single-system ureter and has a normally located ureteral orifice in the corner of the trigone of the bladder. The upper-pole ureter is the “accessory ureter” and inserts medially and inferiorly to the normal ureteral orifice (Weigert-Meyer rule). Since the lower pole of the duplex kidney is analogous to the normal single-system kidney, abnormalities of the lower pole are

Mesonephric Duct Accessory Ureter

Fig. 1 a, b. Duplex Kidney: embryology of the ectopic pathway. a Embryology of a complete ureteral duplication showing that the accessory ureter joins the upper pole of the metanephros. b As the kidney continues to develop, the bladder insertion of the lower-pole ureter ascends to the normal (orthotopic) position in the bladder trigone. The upper-pole ureter descends so that its orifice in the bladder is medial and inferior to the lowerpole ureter

Metanephros Ureter

b

244

Jeanne S. Chow

a

Fig. 2. Comparison of the calyceal axes

b

similar to those of a single (non-duplex) collecting system, such as vesicoureteral reflux and obstruction at the ureteropelvic or ureterovesical junction. Reflux into the lower-pole ureter and intrarenal collecting system can be easily distinguished from reflux into a single (non-duplex) system kidney by carefully noting the axis of the calyces (Fig. 2). Normally, the axis of the calyces (the line drawn from the lowest to the highest calyx) of a single-system kidney is toward the contralateral shoulder. Since only the lower calyces are opacified in lower-pole reflux, the axis of the visualized calyces is altered and lies toward the ipsilateral shoulder [1]. Ureteral obstruction of the lower pole ureter can occur at the level of either the renal pelvis or the insertion of the ureter into the bladder. Ureteropelvic junction obstruction (UPJO) of the lower pole of the kidney is visualized by ultrasound [2]. Occasionally, when the upper pole is dysplastic, lower-pole UPJO can be mistaken for obstruction of a single collecting system. Ureterovesical junction obstruction (UVJO) can occur due to primary mega-ureter, or it may be secondary to the effect of a dilated obstructed upper-pole ureter. In these cases, when the upper-pole ureter is decompressed, the lower pole obstruction also resolves. The upper pole ureter inserts ectopically, medially, and inferiorly, to the normal ureteral orifice into any mesonephric duct derivative. In addition to forming the ureter, the Wolffian (mesonephric) duct contributes to the formation of the trigone of the bladder, the urethra, and the vagina in females, and the posterior urethra and genital ducts in males. The “ectopic pathway” follows the pathway created by the Wolffian duct (Fig. 3). Girls with an ectopic ureter inserting into the vagina or perineum may present with constant urinary dribbling [3] (Fig. 4). In boys, the ureter can terminate in Wolffian duct derivatives, including the seminal vesicles and vas deferens. However, ectopic ureters in boys never terminate below the urinary sphincter and thus never cause incontinence.

Fig. 3 a, b. The ectopic pathway in boys (a) shows that the ectopic ureter may insert from just below the trigone of the bladder, to the posterior wall of the uretha as low as the veromontanum, and the ejaculatory duct and its branches. The ectopic pathway in the girl (b) shows that the ectopic ureter may insert from just below the trigone of the bladder down to the posterior wall of the urethra, to the vulva and vagina

Ectopic ureters are often obstructed, usually at the level of the ureterovesical junction, but rarely reflux. If the ectopic ureter inserts into the urethra at the level of the urinary sphincter, it is both obstructed and refluxes, depending on whether the sphincter is closed or open, the so-called sphinteric ectopic ureter (Fig. 5) [4]. The more distal the ureteral insertion, the more dysplastic and dysfunctional is the associated renal parenchyma that it drains. A ureterocele is the dilated submucosal terminal segment of the ureter. It is associated with the upper-pole ureter of a double collecting system in girls. In boys, ureteroceles are rare, but when they do occur they are

245

Understanding Duplication Anomalies of the Kidney

Prolapse Fig. 7. Prolapsing ectopic ureterocele

Fig. 4. Vaginal ectopic ureter

Sphincter closed

Sphincter open (Voiding)

b

Fig. 8 a, b. Ureterocele disproportion. a A typical ectopic, obstructed, left upper pole ureter ending in ureterocele. b Ureterocele disproportion

Fig. 5. Sphincteric ectopic ureter

a

a

b

Effacement of ureterocele “Intussusception” of ureterocele Fig. 6 a, b. Effacing (a) and everting (b) ureterocele

most commonly associated with single-system kidneys and their orifices are orthotopic. These are associated with varying degrees of ureteral obstruction and pelvicalyceal dilatation. Although ureteroceles protrude into the lumen of the bladder, when the intravesical pressure equals that of the ureterocele, the ureterocele can flatten and become

imperceptible (efface) (Fig. 6a). When the intravesical pressure exceeds that of the ureterocele, the ureterocele everts or intussuscepts into its ureter (Fig. 6b). Everting ureteroceles are often confused for periureteral diverticula. Ectopic ureteroceles involving the bladder neck can prolapse into the urethra and cause obstruction of the bladder outlet (Fig. 7). Typically, ureteroceles of upper pole ureters are associated with hydroureteronephrosis. Rarely, the upper pole is diminutive and dysplastic, and the ureterocele is relatively large. This combination is described as ureterocele disproportion (Fig. 8). The condition can be mistaken for single ectopic ureters [5]. Ureteroceles can also be associated with multicystic dysplastic kidneys. As the fluid in the multiple cysts resolves, these develop the appearance of ureterocele disproportion. Very rarely, three complete ureters or three incompletely separated ureters form, resulting in complete or incomplete in ureteral triplication [6].

246

Jeanne S. Chow

In paired organs, such as the kidney, when there is a congenital anomaly in one and there is something wrong with the other, it is almost always the same anomaly, but often different in degree. Thus, if one kidney is duplex, the other is more likely to be duplex as well.

8. Ureteroceles associated with duplex kidneys and ectopic upper-pole ureters are more common in girls than in boys. In boys, ureteroceles are usually associated with single (non-duplex) kidneys. 9. Ureteroceles are dynamic.

Conclusions

References

These are the fundamental principles to remember: 1. The ureters are duplicated but the kidney is called “duplex”. 2. Ureteral duplication may be incomplete or complete. 3. The Weigert-Meyer rule applies to complete ureteral duplication and states that the upper-pole ureteral orifice is ectopic. When the ectopia is slight, the upper pole is normal. When the ectopia is moderate or severe, the upper pole is abnormal. 4. The lower pole is the analogue of a single-system kidney. 5. Lower-pole reflux can be distinguished from reflux into a single-system kidney by the axis of the calyces. 6. Ectopic ureters terminate along the ectopic pathway and can cause incontinence in girls, but never in boys. 7. Ureteroceles are caused by obstruction; they do not cause obstruction.

1. Claudon M, Ben-Sira L, Lebowitz RL (1999) Lower pole reflux in children: uroradiologic appearance and pitfalls. AJR 172:795-801 2. Fernbach SK, Zawin JK, Lebowitz RL (1995) Complete duplication of the ureter with ureteropelvic junction obstruction of the lower pole of the kidney: imaging findings. AJR 164: 701-704 3. Carrico C, Lebowitz RL (1998) Incontinence due to an infrasphincteric ectopic ureter: why the delay in diagnosis and what the radiologist can do about it. Pediatric Radiology 28:942-949 4. Wyly JB, Lebowitz RL (1984) Refluxing urethral ectopic ureters: recognition by the cyclic voidng cystourethrogram. AJR 142:1263-1267 5. Share JC, Lebowitz RL (1989) Ectopic ureterocele without ureteral and calyceal dilatation (ureterocele disproportion): findings on urography and sonography. AJR 152:567-571 6. Gill RD (1952) Triplication of the ureter and renal pelvis. J Urol 68:140-147

IDKD 2010-2013

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management Alan Daneman Department of Radiology, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada

Introduction The midgut is that part of the bowel supplied by the superior mesenteric artery (SMA) and it extends from the mid-portion of the second part of the duodenum to the distal third of the transverse colon. The term “malrotation” is broadly used to describe the spectrum of developmental abnormalities of the midgut that are associated with abnormal rotation and/or fixation [1-3]. Malrotation may occur as an isolated entity. However, it is also associated with congenital defects in the development of the abdominal wall (e.g., omphalocele and gastroschisis) or diaphragm (e.g., congenital diaphragmatic hernia). Furthermore, it may also be associated with abnormalities of visceral situs and with certain syndromes.

Embryology During embryological development, the midgut undergoes a complicated process of growth and lengthening, herniation, rotation, reduction, and fixation [1]. Growth and lengthening of the midgut result in its herniation into the umbilical cord along the axis of the SMA, such that the apex of the herniated midgut is located at the level of the omphalomesenteric duct. This herniated bowel undergoes a 270° counterclockwise rotation, with subsequent reduction of the midgut back into the abdomen before the end of the first trimester. The proximal loop of herniated bowel (including the small bowel up to the level of the omphalomesenteric duct) is the first to reduce, followed by reduction of the distal loop (including the distal small bowel and colon to the level of the distal third of the transverse colon). The final phase involves fixation of the bowel into its final anatomical position. The cecum, however, is usually initially reduced into the upper abdomen on the right. This precedes elongation of the proximal colon until the cecum finally reaches the right lower quadrant. The latter process may only be finally achieved during early infancy. This complicated sequence is essential for the midgut to assume its normal position in the abdomen. One important aspect is the development of the normal duodenal

loop. Initially, the duodenum rotates to the right, then posteriorly, and finally to the left of the SMA, thereby completing the normal duodenal loop. The third part of the duodenum crosses the midline from right to left in the angle between the SMA (anterior to the duodenum) and the aorta (posterior). The fourth part of the duodenum ascends to the left of the spine to reach the duodeno-jejunal flexure (D-J flexure) at the ligament of Treitz. The position of the D-J flexure has been traditionally used as a landmark for documenting normal rotation of the midgut on contrast examinations of the upper gastrointestinal (GI) tract. The other landmark used is the position of the cecum in the right iliac fossa. When the D-J flexure and cecum are in their normal positions, the small bowel mesentery has a long base extending from the left upper quadrant (D-J flexure) obliquely down to the right lower quadrant (cecum). This long mesenteric base protects against the development of volvulus. Abnormalities due to the arrest of rotation and/or fixation can occur at any phase of the above-described process and may involve only a part or all of the midgut. This leads to many variations of malrotation and/or malfixation and accounts for the spectrum of clinical presentations and radiological appearances.

Causes of Symptoms The majority of these variations of malrotation and /or malfixation are associated with clinical symptoms that usually present within the first few months of life and can be life-threatening [1, 2]. Others may be associated with few or no symptoms, with the latter type often found only incidentally. Abnormalities of rotation and/or fixation usually lead to a situation in which the cecum and duodenum lie closer to each other than normal, such that the base of the small bowel mesentery is much shorter than normal. The obstruction occurs primarily in the duodenum (Figs. 1, 2) and is most often due to peritoneal (Ladd) bands that anchor the cecum to the retroperitoneum across the duodenum in the right upper quadrant. However, even more importantly, obstruction may also be due to volvulus as a re-

248

Alan Daneman

Fig. 1. Series of anteroposterior fluoroscopic images from a contrast examination of the upper GI tract in a young infant with malrotation. In the initial image (top left), before contrast was injected into the nasogastric tube, there is a non-specific bowel gas pattern with diminished gas in the mid and right abdomen. Contrast administration is followed by intermittent dilatation of the second and proximal third parts of the duodenum; initially, no dilatation is visible but subsequent images show varying degrees of dilatation. The dilated third part of the duodenum reaches only to the level of the right pedicles and does not cross the midline. The distal duodenum is on the right. The normal D-J flexure is absent. The findings are diagnostic of midgut malrotation with obstruction of the duodenum by Ladds bands and suggest a volvulus, which was confirmed at surgery. This case illustrates the potential non-specificity of the abdominal radiograph in the presence of duodenal obstruction and the intermittent nature of the duodenal dilatation

sult of the narrow base of the small bowel mesentery (Fig. 2). Much less commonly, there may be an associated internal hernia (Fig. 3). The clinical picture and imaging appearance depend on the nature and degree of the obstruction (which may be intermittent) as well as on the presence or absence of vascular compromise.

Imaging Modalities

Fig. 2. Lateral fluoroscopic view of contrast examination of the upper GI tract in a neonate with malrotation and volvulus. The proximal duodenum is dilated and ends in a beak that leads into a corkscrew pattern of non-dilated small bowel. The beak and corkscrew pattern are typical of obstruction due to volvulus, which was confirmed at surgery. The lateral view is often better than the anteroposterior view shown in Fig. 1 in demonstrating the volvulus but the latter view remains essential to depict the position of the of D-J flexure

As the clinical findings are often non-specific, pediatricians and surgeons rely on the radiologist to confirm or exclude the diagnosis [1-3]. The radiologist thus plays an exceptionally important role in the diagnosis of a malrotation and must be able to recognize the spectrum of appearances of its many variations, as depicted by any imaging modality (Figs. 1-5). Failure to do so may lead to a delay in treatment and thus potentially to bowel necrosis (which may require extensive resection) and even death. On a plain abdominal radiograph, malrotations may show a wide spectrum of appearances [3]. In contrast to what might be expected, the finding of duodenal distention with gas as a typical component of duodenal obstruction is often absent. If the duodenum is fluid-filled or

249

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management

a

b

Fig. 3. Anteroposterior abdominal radiograph from an upper GI series and follow-through contrast examination in a young teenager with midgut malrotation and an internal hernia. Contrast outlines the stomach and duodenal cap and then leads into a very well defined, rounded collection of small bowel loops on the right, representing a paraduodenal internal hernia. This appearance of contained, well defined small bowel is typical of an internal hernia

Fig. 5 a, b. Transverse sonograms of the upper abdomen in a neonate with malrotation and volvulus proven at surgery. The volvulus appears as a characteristic whirlpool in gray-scale (a) imaging and on color Doppler evaluation (b). The small bowel and mesenteric veins are coiled around the superior mesenteric artery

Fig. 4. Contrast enema in a young infant with malrotation proven at surgery. Note that the proximal colon turns back towards the left in the right upper quadrant, and the cecum and appendix lie in the upper abdomen close to the position of the duodenum. As a result of the proximity of the duodenum and cecum, the small bowel mesentery is much shorter than normal, thus predisposing to development of a volvulus

collapsed, it may not be visible and gaseous distension of the stomach alone may suggest gastric obstruction. A volvulus may lead to the presence of a soft-tissue mass due to the fluid-filled bowel, or the abdomen may be almost completely gasless. In children with severe vascular compromise due to volvulus, there is often gaseous

distention (with air fluid levels) of the entire small bowel, resembling a low bowel obstruction or ileus. The appearances of malrotation are indeed often non-specific (Fig. 1) and may be normal even in the presence of volvulus. Due to this wide variation of appearances on plain radiographs, the clinician should never rely on plain film findings to exclude malrotation. Any child in whom there is a clinical suspicion of malrotation should be studied with modalities that will directly depict the position of the bowel and/or the nature of the obstruction. These include following modalities: 1. Contrast examinations of the gastrointestinal (GI) tract: An upper GI series (alone or in combination with a follow-through series) and/or a contrast enema (Figs. 1-4) [4-8]. The most important features of the GI tract to define are the position of the D-J flexure (which requires a straight A-P view on the upper GI series) and/or the position of the cecum and proximal colon.

250

2. Cross-sectional imaging of the abdomen, particularly with sonography (but also with computed tomography and magnetic resonance) [9-15]. These modalities may depict a dilated duodenum, malposition of the D-J flexure, the whirlpool sign indicative of volvulus (Fig. 5), an internal hernia, or abnormalities of the relationship of the SMA and superior mestenteric vein. Meticulous attention to technique is critical with each of these modalities in order to delineate the relevant structures accurately. However, even when a technically perfect examination is performed, none of the above features are 100% accurate in allowing the confirmation or exclusion of malrotation. Accordingly, there has been much debate over the years as to which modality should be used first and what protocol of subsequent modalities may be required in order to enable the radiologist to rapidly make an accurate diagnosis. Independent of which modality is chosen first, the radiologist should never hesitate to ask for or perform other types of examinations aimed at obtaining more information that may increase confidence in the diagnosis – whether malrotation is present or not. As information from the various examinations is accumulated, the balance of evidence from all the examinations should be weighed together so that the correct diagnosis can be determined. Most institutions still use the upper GI series as the modality of choice in children in whom malrotation is suspected clinically [4-8]. This examination is relatively non-invasive, easy to perform, and the position of the D-J flexure is highly accurate in predicting malrotation (Fig. 1). Indeed, the presence or absence of malrotation can be readily made in most children using this modality alone. However, there will remain a group of children in whom the diagnosis of malrotation will be difficult based on the upper GI series alone, either because of technical difficulties in some patients or because of difficulties in differentiating normal variations in duodenal anatomy from true abnormalities of rotation. It is in these children that the radiologist must not be reluctant to extend the GI contrast examination or to perform another type of examination that will generate further diagnostic information. For this purpose, the next most commonly evaluated factor to determine on contrast examinations of the GI tract is the position of the cecum [4, 5]. In the acute clinical situation, particularly in neonates and young infants, the quickest way to achieve this is by performing an immediate contrast enema (Fig. 4), i.e., before too much contrast from the prior upper GI series fills the small bowel. The advantage of the contrast enema is that it relatively quickly provides information on the position of the entire large bowel. However, interpretation of the cecal position may be difficult. It must be remembered that the position of the cecum and proximal small bowel has a wide range of normality, particularly in neonates and young infants, and may well be normal even in patients with malrotation. Furthermore, the cecum and proximal

Alan Daneman

colon may be easily displaced into positions that simulate malrotation by adjacent markedly dilated loops of small bowel. This is seen particularly in neonates with congenital obstruction involving the distal small bowel. In such patients, the presence of a microcolon usually excludes malrotation, as this combination is rare. In older children, especially if the clinical setting is not acute, contrast from the original upper GI series can be followed with serial plain radiographs to determine the cecal position. This approach may take much longer than performing a contrast enema and is usually less fruitful in the neonate and young infant, since in this age group it is not always possible to clearly depict the position of the cecum. In recent years, signs of malrotation have been visualized using sonography and other cross-sectional imaging modalities, but they are still not used as the modality of choice by most radiologists [9-15]. It is true, however, that sonography is being used much more frequently than suggested in the literature. Direct visualization of a volvulus (whirlpool sign) (Fig. 5) may obviate the necessity for contrast examination of the GI tract; but this sign is not present in those children without volvulus but who are symptomatic because of obstruction due to bands. Furthermore, the sign may be difficult to appreciate in children with volvulus and a large amount of dilated, gas-filled bowel. Inversion of the superior mesenteric artery and vein relationship may be present in normal rotation, and a normal relationship may be present in children with malrotation. Therefore, a normal sonogram does not exclude malrotation and, to date, sonography has not been used as a screening procedure for this condition. Nevertheless, it is essential that radiologists be able to recognize these abnormal signs when they are detected as an incidental or unexpected finding on cross-sectional imaging. Yousefzadeh [15] has drawn attention to the sonographic depiction of the third part of the duodenum (D3) in the angle between the SMA and the aorta, and has suggested that documentation of the presence of the D3 in this normal position excludes the presence of malrotation. This is an extremely interesting approach because, if shown to be accurate, then the sonographic depiction of D3 in this position would obviate the necessity for doing a fluoroscopic, contrast examination of the upper GI tract in those children suspected of having this condition. However, to date, there are no data to substantiate this suggestion. Filling the duodenum with fluid administered orally or through a feeding tube may facilitate delineation of the position of the duodenum and D-J flexure on sonography. This approach has been advocated by some as the technique of choice, and some groups have used it quite extensively but have yet to publish their data. Gent and LeQuesne presented their experience with this technique, based on over 100 cases, at the World Federation of Ultrasound in Medicine and Biology Meeting in Sydney in 2009. These authors illustrated their technique, which

251

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management

delineates the anatomy of the duodenum exquisitely. They have been able to easily confirm the diagnosis of malrotation in the vast majority of patients, with no falsenegatives. In a few patients in whom the diagnosis was unsure, an upper GI series was required for confirmation. To the best of this author’s knowledge, there has, to date, been no large single series comparing the accuracy of this technique to the routine contrast upper GI series. Color Doppler sonographic assessment of bowel perfusion may be useful in patients in whom bowel perfusion is in doubt. However, its role is limited as poor perfusion is usually seen in sicker patients and surgeons are keen to get these children to the operating room as soon as possible without necessarily relying on the need for diagnostic imaging.

Conclusions Accurate diagnosis of malrotation requires (i) a high index of suspicion, (ii) an understanding of normal variations of the GI tract, (iii) an appreciation of the spectrum of appearances of malrotation on imaging modalities, and (iv) attention to meticulous technique when performing GI contrast examinations or cross-sectional imaging. Although the diagnosis of malrotation can be accurately made in many instances based on a single examination, the radiologist should never be reticent about using more than one examination (or repeating examinations) when attempting to confirm or exclude malrotation in difficult cases. As information is accumulated from the various diagnostic examinations, the balance of evidence derived from all of them should be weighed together to determine the correct diagnosis. If uncertainty remains after all imaging avenues have been exhausted, as does happen (though uncommonly), then the decision whether to perform laparoscopy or to operate should be left to the surgeon.

References 1. Strouse PJ (2000) Disorders of intestinal rotation and fixation (“malrotation”). Pediatr Radiol 34:837-851 2. Lampl B, Levin TL, Berdon WE, Cowles RA (2009) Malrotation and midgut volvulus: a historical review and current controversies in diagnosis and management. Pediatr Radiol 39:359-366 3. Daneman A (2009) Malrotation: the balance of evidence. Pediatr Radiol 39(2):S164-S166 4. Long FR, Kramer SS, Markowitz RI et al (1996) Intestinal malrotation in children: Tutorial on radiographic diagnosis in difficult cases. Radiology 198:775-780 5. Long FR, Kramer SS, Markowitz RI, Taylor GE (1996) Radiographic Patterns of Intestinal Malrotation in Children. RadioGraphics 16:547-556 6. Katz ME, Siegel MJ, Shackelford GD, McAlister WH (1987) The Position and Mobility of the Duodenum in Children. AJR 148:947-951 7. Donnolly LF, Rencken IO, de Lorimier AA, Gooding CA (1996) Left paraduodenal hernia leading to ileal obstruction. Pediatr Radiol 26:534-536 8. Manji R, Warnock GL (2000) Left paraduodenal hernia: an unusual cause of small-bowel obstruction. Can J Surg 44:455-457 9. Dufour D, Delaet MH, Dassonville M et al (1992) Midgut malrotation, the reliability of sonographic diagnosis. Pediatr Radiol 22:21-23 10. Chao HC, Kong MS, Chen JY et al (2000) Sonographic features related to volvulus in neonatal intestinal malrotation. J Ultrasound Med 19:371-376 11. Shimanuki Y, Aihara T, Takano H et al (1996) Clockwise whirlpool sign at color Doppler US: an objective and definite sign of midgut volvulus. Radiology 199:261-264 12. Yoo SJ, Park KW, Cho SY et al (1999) Definitive diagnosis of intestinal volvulus in utero. Ultrasound Obstet Gynecol 13:200-203 13. Loyer E, Eggli KD (1989) Sonographic evaluation of superior mesenteric vascular relationship in malrotation. Pediatr Radiol 19:173-175 14. Weinberger E, Winters WD, Liddell RM et al (1992) Sonographic diagnosis of intestinal malrotation in infants: importance of the relative positions of the superior mesenteric vein and artery. AJR Am J Roentgenol 159:825-828 15. Yousefzadeh DK (2009) The position of the duodenojejunal junction: the wrong horse to back on in diagnosing or excluding malrotation. Pediatr Radiol 39:S172-S177

IDKD 2010-2013

Pediatric Intestinal Ultrasonography Simon G. Robben Department of Radiology, Maastricht University Medical Centre, Maastricht, The Netherlands

Introduction Ultrasonography (US) is the imaging modality of choice for the initial evaluation of diseases in children for many reasons. First, it is relatively inexpensive and patient friendly. Second, it lacks radiation and motion artifacts. Third, the small size of the child compensates for the limited penetration of sound waves and facilitates the use of high-frequency transducers. Fourth, flow studies are possible using the Doppler mode, while realtime imaging allows the visualization of movements such as peristalsis. Moreover, US involves direct contact with the patient, thus offering a unique opportunity to ask specific questions and to perform additional physical examinations, emphasizing the role of the radiologist as a clinician. Accordingly, ultrasonography has become the most important imaging technique in children and can be considered as the workhorse of pediatric radiology. Initially, US of the stomach and intestines was not popular for obvious reasons: bowel gas has the annoying characteristic of reflecting all sound waves or creating artifacts because of its abnormally low acoustic impedance.

a

However, increased knowledge, improved technique (e.g., graded compression), improved hardware (highfrequency transducers), and improved software (adaptive imaging, compound imaging) have resolved the limitations to US use. Nowadays it is impossible to imagine US without intestinal US, especially in pediatrics! The hallmark of intestinal US is the “gut signature”, i.e., the characteristic appearance of the layers of the gut (Table 1 and Fig. 1).

Table 1. Gut signature, characterized by alternating hyper- and hypoechoic layers Layer

Echogenicity

Mucosal surface Mucosa Submucosa Muscularis propria Serosal surface

Hyperechoic Hypoechoic Hyperechoic Hypoechoic Hyperechoic

b

Fig. 1 a-c. Gut signature in a stomach, b ileum, and c appendix (between arrows)

c

253

Pediatric Intestinal Ultrasonography

Gastroesophageal Junction The gastroesophageal junction (Fig. 2) can be visualized with US in 87-95% of children with suspected gastroesophageal reflux disease (GERD), and reflux of gastric content into the esophagus can be demonstrated ultrasonographically [1, 2]. A threshold of three reflux periods within 10 min corresponds to a sensitivity of 82% and a specificity of 85% for GERD [1]. Farina et al. increased the sensitivity to 98% using color Doppler US [3]. However, in premature infants the results differ: the sensitivity is 38% and the specificity is 100% compared to 24-hour pH-metry [4]. US studies have also shown that there is a positive correlation between GERD and the length of the abdominal esophagus, protrusion of the gastric mucosa, and an increased gastroesophageal angle (angle of His). Therefore, US can be the initial imaging technique in children suspected of having GERD. Moreover, it can evaluate the post-operative situation after a fundoplication and gastric emptying, another important contributing factor to the pathogenesis of GERD.

palpation is accurate but not always successful as it depends on factors such as the experience of the examiner, the presence of gastric distension, and the cooperation of the infant. In virtually all patients, US is very accurate in facilitating the diagnosis and therefore plays a key role in the initial care of these infants. It is important that the radiologist understands the anatomical changes of the pyloric channel in affected infants, as demonstrated by US. Pyloric muscle hypertrophy is shown to a variable degree during the US examination. In addition, a certain amount of thickening of the mucosa is present (Fig. 3). A muscle thickness that is consistently v3 mm is considered to be diagnostic of IHPS, although some clinicians have stated that the overall morphological and dynamic impression, including length of the pyloric canal as well as relaxation and peristalsis, are just as important. The US examination can be performed in a very short time with an accuracy approaching 100%.

Stomach Gastric emptying in vomiting patients can be evaluated with US, which depicts dynamic and anatomical abnormalities. Infantile hypertrophic pyloric stenosis (IHPS) [5] is a condition of unknown etiology that affects young infants ages 2-8 weeks and with a male-to-female ratio of approximately 4:1. In IHPS, the antropyloric portion of the stomach becomes abnormally thickened and manifests as an obstruction to gastric emptying. Typically, infants with IHPS are clinically normal at birth but during the first few weeks of postnatal life, they develop non-bilious “projectile” vomiting that leads to weight loss, dehydration and hypochloremic alkalosis, and eventually death. Surgical treatment is curative. The clinical diagnosis relies on palpation of the thickened pylorus. Abdominal

a

Fig. 2 a, b. Sagittal slice of a normal gastroesophageal junction (arrows) in a 3month-old boy. b Same slice during transit of air. During real-time ultrasonography, the direction of transit can easily be appreciated. S Stomach, L liver

Fig. 3. Hypertrophic pyloric stenosis (between large arrows). Muscle thickening is indicated by small arrows. A Antrum of stomach with retained fluid

b

254

Simon G. Robben

Foveolar hyperplasia consists of a polypoid thickening of the mucosal layer. It may be seen after long-standing prostaglandin therapy, hypertrophic gastropathy, or cow’s milk allergy, but it may also be idiopathic. While the condition may simulate pyloric hypertrophy, on closer US examination the obstruction will appear to be caused by thickened mucosa instead of thickened muscle. Rare causes of gastric outlet obstruction include pylorospasm, (eosinophilic) gastritis, food allergy, chronic granulomatous disease, hyperlipidemia, duplication cysts, ectopic pancreas, benign and malignant tumors, and bezoars.

Small Bowel Conventional radiography and US are the initial imaging modalities in children with abdominal pain or obstruction. The most important additional value of US over conventional abdominal radiographs in these children is its capability to visualize peristalsis, vascularity, bowel wall characteristics, dilatation of fluid-filled loops, and extra-intestinal abnormalities, e.g., ascites and other fluids. The jejunum and ileum can be distinguished from the colon based on anatomical location, caliber, contents, folds, and peristalsis. Anatomical location: The colon has a peripheral location in which the ascending and descending colon lie dorsally in both flanks and the transverse colon is located ventrally in the upper abdomen. The sigmoid colon traverses the left psoas muscle and courses into the pelvis whereas the small bowel has a more central position. Caliber: The diameter of the small bowel is small while, as its name indicates, the diameter of the large bowel is relatively large.

a

Contents: The small bowel is either empty or filled with liquid contents but little air, whereas the colon is generally filled with gas-filled bulky stools. Folds: The folds in the jejunum are more numerous, longer, thinner, and closer together than the ileal folds. In the terminal ileum, the mucosa may be thickened due to hyperplasia of lymphoid tissue. The colon is recognized by its haustrations. Peristalsis: The small bowel moves continuously due to persistaltic waves whereas the colon shows sparse movements. Baud proposed a systematic US approach for identifying small bowel disease, based on wall thickening [6]. 1. Determine wall thickening: normal (f3 mm), mild (36 mm), moderate (6-9 mm), or severe (>9 mm). 2. Determine location (proximal or distal) and extent (focal 5 cm, segmental 6-40 cm, or diffuse >40 cm). 3. Determine stratification. The bowel wall is stratified when the hyperechogenicity of the submucosa is preserved and the mucosa, submucosa, and muscularis propria are visible as separate layers. Non-stratification implies the absence of distinction between mucosa and submucosa or between all three layers (Fig. 4). 4. Determine the valvular fold pattern: normal, thickened, thumb-printing, and hyperplastic valvular folds. In general, thickened small bowel loops show decreased peristalsis and contain little air. They are therefore easily visualized and measured. At least three patterns can be distinguished. Stratified thickening of the small bowel is found in infectious ileitis, advanced appendicitis, early Crohn’s disease, and graft versus host disease. Non-stratified thickening occurs in Henoch-Schönlein purpura, advanced Crohn’s disease, tuberculous ileitis,

b

Fig. 4 a, b. Difference between a stratified wall thickening in early Crohn’s disease and b non-stratified wall thickening in advanced Crohn’s disease

255

Pediatric Intestinal Ultrasonography

3. Determine stratification, which is best seen on transverse images. The bowel wall is stratified when the hyperechogenicity of the submucosa is preserved and the mucosa, submucosa, and muscularis propria are visible as separate layers. Non-stratification implies an absence of distinction between the mucosa and submucosa or between all three layers. 4. Determine the haustral pattern, which is best seen on longitudinal images; also, confirm the absence or presence of the haustral folds and their length (normal or shortened) and aspect. These criteria distinguish three patterns. Fig. 5. Two benign small bowel intussusceptions in a patient with a malabsorption syndrome

protein-losing enteropathy, hereditary angioedema, ischemia, celiac disease, Burkitt’s lymphoma, Kawasaki’s disease, and viral enteritis. Non-stratified thickening with hyperplastic valvular folds can be seen in viral (and sometimes bacterial) lymphoid hyperplasia and Yersinia ileitis. Malrotation and midgut volvulus are discussed in the chapter by Daneman. Benign small bowel intussusception (BSBI) is a recently described entity. It differs from the classical symptomatic ileocolic intussusception in that it occurs predominantly in the right lower quadrant or periumbilical region, has a smaller diameter (mean diameter 1.4 vs. 2.5 cm), a thinner outer rim, and does not contain mesenteric lymph nodes (Fig. 5) [7]. Moreover, peristalsis in the intussuscepted loop persists, in contrast with ileocolic intussusception. Often, BSBI is an incidental finding but it occurs with increased frequency and number in celiac disease. In general, BSBI does not need immediate reduction because of its spontaneously resolving nature. Appendicitis, Meckels diverticulum, necrotizing enterocolitis, and ileocolic intussusception are discussed elsewhere in this volume.

Large Bowel The differences between the normal small and large bowel were described above. In addition to the systematic approach to the small bowel, Baud proposed an analogous approach to the colon [8]. 1. Determine wall thickening: normal (f3 mm), mild (36 mm), moderate (6-9 mm), or severe (>9 mm). 2. Determine the extent and location of disease: diffuse, cecum, ascending colon, proximal and distal transverse colon, descending colon, or rectosigmoid.

Stratified thickening is found in infectious colitis, advanced appendicitis, and inflammatory bowel disease (ulcerative colitis and Crohn’s disease). Non-stratified thickening with loss of haustral folds is found in early hemolytic uremic syndrome (HUS) and advanced Crohn’s disease. Non-stratified thickening with preservation of normal haustral fold length is found in pseudomembranous colitis and neutropenic colitis (typhlitis).

Rectum In several studies, US was used to measure the transverse diameter of the rectum. This seems to be a reliable approach to identifying rectal impaction and may replace digital rectal examination. All children with rectal impaction on digital examination were ultrasonographically shown to have had a rectal diameter >30 mm [9]. Moreover, several studies reported that in children with constipation the mean diameter of the rectum is significantly larger than in normal children. In a series of 225 children, the mean rectal diameter in normal children was 32 mm (SD 8.2) and in children with constipation 43 mm (SD 9.7) [10]. To overcome the problem of agedependency of the rectal diameter in normal children, Bijos et al. proposed the rectopelvic ratio, defined as the ratio of the rectal diameter (as determined with US) to the distance between the anterior superior iliac spines [10]. A rectopelvic ratio >0.189 corresponds to a sensitivity for rectal impaction of 88% compared to proctoscopy. The rectal diameter can also be used to monitor therapy.

Anus Anal atresia is a relatively frequent congenital abnormality in which the anus is absent and a rectoperineal, rectovestibular, rectovaginal, rectourethral, or rectovesical fistula may be identified in almost all cases. Since the fistula demonstrates an internal sphincter, some clinicians instead prefer the term “ectopic anus” or “anorectal

256

malformation”. During pre-operative evaluation, it is important to assess the type of anal atresia, which may be high (distal rectal pouch above the puborectal sling), intermediate (at the sling), or low (through the sling). Transperineal US is a good diagnostic modality for defining the type of anal atresia as it can be used to measure the distance between the rectal pouch and the perineum (P-P distance). A P-P threshold value of 15 mm discriminates the low type of atresia from the intermediate and high types with a sensitivity of 100% and a specificity of 86% [11]. Moreover, internal fistula can be correctly identified in 82% of the patients with the high type of anal atresia [12]. Transperineal sonography is also a useful method for differentiating between an anteriorly displaced anus, which is a normal anatomical variant, and a low-type imperforate anus with perineal fistula, which is a pathological developmental abnormality requiring surgical repair [13]. In adults, transperineal ultrasonography is a simple, painless, cost-effective and real-time method to detect and classify perianal fluid collections, abscesses, fistulas, and sinus tracts [14-16]. These data can probably be extrapolated to the pediatric population, e.g., children with Crohn’s disease.

Cystic Intestinal Masses Cystic intra-abdominal masses originating from the alimentary canal are increasingly recognized because of the advent of routine prenatal US. These masses can be divided into cysts originating from solid organs (mesenchymal hamartoma, congenital splenic cyst, pancreatic pseudocyst, pancreatic cystadenoma, hydronephrosis, multicystic dysplastic kidney, multilocular cystic nephroma, adrenal hemorrhage, ovarian cysts and cystic neoplasms, hematocolpos, urachal cysts, abdominal and sacrococcygeal teratoma, and cerebrospinal fluid pseudocyst) and those originating from the alimentary canal and its appendages (hydrops of the gallbladder, choledochal cyst, mesenteric and omental cysts, gastrointestinal duplication cyst, meconium pseudocyst, and appendiceal abscess) [17] (Table 2).

Simon G. Robben

When a cystic mass is found on US examination, it should be evaluated for its size, shape, location, relation to organs, and contents and wall characteristics. In the majority of cases, US can provide a specific diagnosis or offer a narrow differential diagnosis [18]. Hydrops of the gallbladder is a rare cause of a right upper quadrant mass in children. It has been described in the absence of stones, infections, or congenital abnormalities, in which case it is probably caused by transient obstruction of the cystic duct or increased mucus secretion with ineffective emptying. It has also frequently been associated with Kawasaki’s disease. Rarely, it is associated with childhood infections, Henoch-Schönlein purpura, Cryptosporidium infection in immunocompromised children, Epstein-Barr virus infections, and typhoid fever. US examination reveals a dilated anechoic elliptical gallbladder without wall thickening. The bile ducts are not dilated and no stones are seen [17, 19]. Choledochal cysts are actually focal cystic dilatations of the biliary tree. Most patients present in the first decade of life with symptoms of episodic abdominal pain, mass, and jaundice. US is the best initial method of evaluating dilatation of the bile ducts. On US, the choledochal cyst is located in the porta hepatis, separate from the gallbladder, with bile duct(s) leading into or out of it. The entire biliary tree should be evaluated but intrahepatic bile duct dilatation may be absent. Surgical resection is necessary to prevent the development of ascending cholangitis, stones, or malignant degeneration [17, 18, 20]. Mesenteric cysts are cystic lymphangiomas that are most often found in the small bowel mesentery, especially the ileal mesentery but also in the greater and lesser omentum and occasionally in the mesenteric root and retroperitoneum. About one-third occur in children younger than 15 years. Newborns present with abdominal distention and a palpable mass whereas children are much more likely to present with pain, anorexia, vomiting, or fever. US can characterize the mass as a typical thin-walled unior multilocular cyst that displaces adjacent structures to the periphery of the abdomen (Fig. 6). Calcification of

Table 2. Cystic masses of the gut and gut-related structures Biliary system Choledochal cyst Hydrops of gallbladder Gastrointestinal tract Mesenteric cyst/lymphangioma Enteric/duplication cyst Omental cyst Meconium pseudocyst Miscellaneous Abscess Teratoma Sacrococcygeal teratoma

Fig. 6. Multilocular cystic lymphangioma

257

Pediatric Intestinal Ultrasonography

the wall is rare in children. The cyst content is anechoic but if hemorrhage has occurred then debris may be seen. Lymphangiomas may be so large that they are difficult to distinguish from severe ascites. However, ascites separates individual bowel loops and fills the perihepatic and perisplenic spaces [17, 18]. Gastrointestinal duplication cysts are spherical or tubular masses adherent to the gastrointestinal (GI) tract and sometimes communicating with it [17, 18, 21]. These cysts are lined with intestinal epithelium and contain smooth muscle within their walls. They may occur anywhere along the alimentary tract but the most common location is the ileum, followed by the stomach. Most patients present within the first year of life; their symptoms include GI obstruction and, less commonly, a palpable mass, intussusception, and abdominal distension. The cystic nature of duplication cysts can be easily appreciated on US. The content of the cyst is often anechoic but there may be debris after hemorrhage or due to mucoid material. Rarely, the cyst appears completely hyperechoic after hemorrhage. Two signs are virtually diagnostic of duplication cysts: 1. a double-layered wall consisting of echogenic mucosa and hypoechoic muscularis propria (Fig. 7); 2. peristalsis in the cyst. In 15-20% of cases, the cyst contains gastric mucosa, which accounts for the above-mentioned hemorrhages. Meconium pseudocyst is a manifestation of the cystic type of meconium peritonitis that results from in utero bowel perforation. Bowel perforation may be secondary to intestinal obstruction (meconium ileus, atresia, or volvulus) or idiopathic. The spilled meconium is encapsulated and forms a large meconium-filled (hyperechoic) cyst that is lined by a thick inflammatory and fibrotic membrane, often containing calcifications. The perforation may still communicate with the cyst postnatally. Dilated fluid-filled bowel loops may be seen next to the cyst if obstruction is still present. Distant peritoneal or scrotal calcifications are additional evidence for meconium peritonitis and may also be demonstrated with US or conventional radiographs [17, 18, 22]. Intra-abdominal abscesses often are complications of appendicitis, inflammatory bowel disease, or intra-abdominal surgery. However, they can also occur in patients on immunosuppressive therapy or in those with AIDS or chronic granulomatous disease. The typical presentation is intermittent fever, increased white blood cell count or C-reactive protein, and abdominal tenderness. US shows an ill-defined, fluid-filled mass with a thickened wall and irregular inner surface. The fluid (pus) demonstrates septations, debris, or debris-fluid levels. The presence of gas bubbles in the pus often makes differentiation from bowel loops difficult; the presence of peristalsis rules out an abscess. If feasible, ultrasonographically guided aspiration and/or drainage can be performed, if necessary in the ICU.

a

b

Fig. 7 a, b. Duplication cyst. a Large right-sided cyst on prenatal magnetic resonance imaging shows no typical features. b Postnatal ultrasound demonstrates gut signature in the wall of the cyst suggestive of duplication cyst

Abdominal teratomas usually arise from the retroperitoneum, presacral region, or the ovary, whereas mesenteric and gastric teratomas are rare. These tumors occur mainly in children, and 80% are benign. Mature, immature, and malignant types have been described. Most children present with a palpable abdominal mass that on US is seen to be either (multi)cystic or solid. Teratomas typically contain calcifications and fat. The latter is difficult to appreciate with US but highly echogenic tissue is suggestive of fatty tissue.

258

Simon G. Robben

Sacrococcygeal teratomas usually present as a large external cystic tumor located at the coccyx but in some cases consist of a large cystic intra-abdominal mass without any external mass. These pre-sacral tumors obstruct the rectum and bladder. A pre-sacral extension of an intra-abdominal cyst provides a clue to the diagnosis of sacrococcygeal teratoma. Cystic tumors carry a better prognosis than do solid, hypervascular tumors [17, 18].

ileum. The bowel wall thickening is extensive, asymmetrical, and poorly stratified whereas the mesenteric involvement is bulky and lobulated and appears to be in continuity with the bowel wall. Despite the extensive bowel wall thickening, the lumen may remain wide. Burkitt’s lymphoma can lead to intestinal obstruction and intussusception. It can also involve the liver, spleen, kidneys and pancreas. Extensive involvement of the omentum and peritoneum is rare [25, 26].

Cerebrospinal fluid pseudocyst (liquor cyst) is a complication resulting from ventriculoperitoneal shunt, with a frequency of approximately 3%. Risk factors for pseudocyst formation are related to inflammatory processes and CNS tumors. Pseudocysts tend to occur within 6 months of the last abdominal surgical procedure. Children present with abdominal pain, distention, or mass. US will demonstrate a sonolucent, well-defined mass in noninfected cysts, whereas infected cysts show septa, internal debris, and fluid-fluid levels. There is no statistically significant correlation between pseudocyst size and the presence of infection. It is important to identify the tip of the shunt within the cyst, producing the characteristic “railroad sign” [17, 23, 24].

Intra-abdominal lipomatous tumors (lipoma, lipoblastoma, and the rare liposarcoma) predominantly involve the mesentery and omentum [27-29]. They show hypo- or hyperechoic textures and are finely lobulated, homogeneous, or with fibrovascular septa. In most cases, magnetic resonance imaging is necessary for further evaluation.

Solid Intestinal Masses The many types of solid masses (often of neoplastic origin) that can be found during US examination are summarized in Table 3.

Inflammatory myofibroblastic tumor (inflammatory pseudotumor) most commonly occurs in the mesentery of children or young adults [29, 30]. The typical complaints are of fever, malaise, weight loss, or abdominal pain. The US characteristics of inflammatory pseudotumor are nonspecific: solid, well-defined (sometimes lobulated), with mixed echo-texture and frequent calcifications. Infiltration of the adjacent bowel may occur. Prominent vascularity may be shown with Doppler US.

Lymphoma Burkitt’s lymphoma Non-Hodgkin’s lymphoma

Fibromatosis or abdominal desmoid is part of the clinical-pathological spectrum of deep fibromatoses [29, 30]. The latter encompass a group of benign fibroproliferative processes that are locally aggressive and have the capacity to infiltrate or recur but not to metastasize. Mesenteric structures are the most common sites of origin of intraabdominal fibromatosis. Other locations are the abdominal wall, pelvis, and retroperitoneum. Thirteen percent of patients with mesenteric fibromatosis have familial adenomatous polyposis (FAP), specifically, the Gardner syndrome variant. In these patients, prior abdominal surgery is an important risk factor for the development of mesenteric fibromatosis. The US appearance is a solid, wellcircumscribed mass of variable echo-texture and homogeneity. Locally aggressive fibromatosis infiltrates the mesenteric fat.

Peritoneal, mesenteric, and omental Lipoma, lipoblastoma, and liposarcoma Inflammatory myofibroblastic tumor Fibromatosis (= desmoids) Neurofibromas (NF1-associated) Rhabdomyosarcoma Metastasis Mesenteric lymphadenitis

Neurofibromatous tumors are associated with neurofibromatosis type 1 (NF1). Abdominal involvement is found in 10-25% of patients with NF1, regardless of their age. Intra-abdominal neurofibromas present as hypoechoic heterogeneous masses or as multiple rounded, hypoechoic, well-circumscribed, variably sized mesenteric nodules.

Small and large bowel Juvenile colonic polyps Hamartomatous small bowel polyps Polyposis Carcinoid Vascular malformations

Rhabdomyosarcomas are rare intra-abdominal pediatric tumors involving the mesentery, peritoneum, or omentum, often associated with ascites. Leung et al. described an omental embryonal rhabdomyosarcoma consisting of lobulated round masses surrounded by tissue with a

Burkitt’s lymphoma is a fast-growing and aggressive malignant neoplasm predominantly affecting children and that may be associated with immunodeficiency. US will demonstrate bowel wall thickening combined with a mesenteric mass, predominantly of the cecum and distal

Table 3. Solid masses of the gut and gut-related structures

259

Pediatric Intestinal Ultrasonography

cerebriform pattern [31]. Rhabdomyosarcomas may also occur in the biliary tree, and are actually the most common pediatric tumor of the biliary tree. US findings are nonspecific but if a solid tumor in the biliary tree is detected, it is highly suggestive of rhabdomyosarcoma [32].

a

Intestinal polyps in the small bowel are usually hamartomatous polyps in patients with Peutz-Jeghers syndrome. These polyps occur more often in the jejunum than in the ileum. US is only able to detect polyps >15 mm and is therefore not a screening tool for the detection of uncomplicated polyps. However, it is valuable in diagnosing complications associated with polyps of the small bowel, e.g., intussusceptions. b

Juvenile polyps are the most common neoplasms of the large bowel in children. The sigmoid and rectum are preferential locations. The presenting symptom is rectal bleeding in over 90% of patients. Occasionally, a colo-colic intussusception is the first manifestation of a juvenile polyp. US demonstrates a pedunculated spherical nodule 10-25 mm in diameter and containing multiple 2- to 3-mm cysts. Administration of fluid within the bowel lumen will greatly improve the visualization of these polyps; however, the assessment of rectal polyps is unreliable. Mesenteric lymphadenitis. Abdominal lymph nodes varying from 0 to 10 mm in diameter are detected in almost all asymptomatic children (Fig. 8a). Approximately half of these lymph nodes (43-54%) are >5 mm in their short axis, with the number and shape of the nodes being age-independent [33, 34]. There is a gender dependency in that boys are more affected than girls [33, 35]. In children with recurrent or acute abdominal pain without known cause, the lymph nodes are significantly larger than in asymptomatic children, but there is considerable overlap between the two groups. Children with abdominal pain and abdominal lymph nodes >10 mm in their short axis may be considered as having primary mesenteric lymphadenitis, if no other acute inflammatory process is identified [34, 36]. There are few US criteria that are typical for a specific causative agent. The presence of calcifications and necrosis suggests tuberculous lymphadenitis (Fig. 8b).

Neonatal Bowel Obstruction Many neonatal bowel obstructions are caused by diseases discussed in other chapters, e.g., duplication cysts, malrotation or volvulus, meconium peritonitis, necrotizing enterocolitis, anal atresia, and pyloric hypertrophy. Additional causes of neonatal obstruction include various types of atresia, annular pancreas, meconium ileus, Hirschsprung’s disease, and meconium plug syndrome. These diseases are diagnosed by conventional abdominal radiographs and/or conventional contrast studies but US may be of additional value [37].

Fig. 8 a, b. Enlarged mesenteric lymph nodes. a Asymptomatic child shows lymph nodes with normal texture and ellipsoid shape; the largest node has a short axis of 8 mm. b Tuberculous mesenteric lymph node shows indistinct margins, round shape, and heterogeneous texture with hypoechoic areas of necrosis and a diameter of 20 mm

Atresias are congenital interruptions of the lumen of the alimentary canal caused by a failure of canalization during organogenesis. The result is dilatation proximal to the atresia and complete collapse distal to the atretic segment. While the clinical and radiographic presentation is typical, US may be used to demonstrate accompanying pathology: 1. exclusion of associated malrotation in cases of high obstruction; 2. evaluation of the colon and rectum in cases of high obstruction; single atresias are associated with a normalsized colon, multiple atresias, with microcolon; 3. evaluation of associated malformations of the heart, kidneys, and biliary tree; 4. demonstration of extrinsic duodenal compression in cases of high obstruction, e.g., duplication cysts, preduodenal portal vein, annular pancreas; and 5. demonstration of signs of prenatal meconium peritonitis as a cause of small bowel atresia (meconium cysts, subtle peritoneal or scrotal calcifications).

260

Meconium ileus is a small bowel occlusion caused by inspissated, abnormally tenacious, meconium in the distal ileum, invariably associated with cystic fibrosis. It occurs in 10-20% of neonates with cystic fibrosis. US may demonstrate the microcolon (mean diameter of 4 mm), small bowel dilatation with echogenic contents (in contrast to the hypoechoic contents in ileal atresia), pseudothickening of the bowel wall (caused by a circular layer of tenacious meconium adherent to the bowel wall), granular pattern (air bubbles in the tenacious meconium), and detection of meconium pellets in the distal ileum. Hirschsprung’s disease is caused by an absence of ganglion cells, which results in abnormal gut motility and a lack of gut relaxation. The length of the aganglionic segment is variable but always involves the distal end of the intestinal tract. In a small number of patients, the entire colon and even the ileum and jejunum are involved. US is of limited value because of the air-artifacts in dilated bowel loops. However in very early neonatal US examination, air is not yet present and distention of the colon in a neonate with distal occlusion suggests Hirschsprung’s disease. Meconium plug syndrome can be considered as neonatal constipation. US demonstrates moderate dilatation of the entire colon without a transition zone, echogenic colonic content, and dilatation of the proximal small bowel [37].

Conclusions Ultrasonography is a reliable initial imaging technique to evaluate a variety of gastrointestinal pediatric diseases and malformations. A solid knowledge of the pathophysiology of these conditions is necessary to understand their ultrasonographic manifestations.

References 1. Couture A, Baud C, Ferran J et al (2008) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg 2. Westra SJ, Derkx HH, Taminiau JA (1994) Symptomatic gastroesophageal reflux: diagnosis with ultrasound. J Pediatr Gastroenterol Nutr 19:58-64 3. Farina R, Pennisi F, La Rosa M et al (2008) Contrast-enhanced colour-Doppler sonography versus pH-metry in the diagnosis of gastro-oesophageal reflux in children. Radiol Med 113:591-598 4. Pezzati M, Filippi L, Psaraki M et al (2007) Diagnosis of gastro-oesophageal reflux in preterm infants: sonography vs. pHmonitoring. Neonatology 91:162-166 5. Hernanz-Schulman M (2003) Infantile hypertrophic pyloric stenosis. Radiology 227:319-331 6. Baud C (2008) Small bowel thickening. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg, pp 253-296 7. Park NH, Park SI, Park CS et al (2007) Ultrasonographic findings of small bowel intussusception, focusing on differentiation from ileocolic intussusception. Br J Radiol 80:798-802

Simon G. Robben

8. Baud C (2008) Infectious and inflammatory colitis. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg, pp 297-339 9. Joensson IM, Siggaard C, Rittig S et al (2008) Transabdominal ultrasound of rectum as a diagnostic tool in childhood constipation. J Urol 179:1997-2002 10. Bijos A, Czerwionka-Szaflarska M, Mazur A, Romanczuk W (2007) The usefulness of ultrasound examination of the bowel as a method of assessment of functional chronic constipation in children. Pediatr Radiol 37:1247-1252 11. Haber HP, Seitz G, Warmann SW, Fuchs J (2007) Transperineal sonography for determination of the type of imperforate anus. AJR Am J Roentgenol 189:1525-1529 12. Choi YH, Kim IO, Cheon JE et al (2009) Imperforate anus: determination of type using transperineal ultrasonography. Korean J Radiol 10:355-360 13. Haber HP, Warmann SW, Fuchs J (2008) Transperineal sonography of the anal sphincter complex in neonates and infants: differentiation of anteriorly displaced anus from low-type imperforate anus with perineal fistula. Ultraschall Med 29:383-387 14. Bonatti H, Lugger P, Hechenleitner P et al (2004) Transperineal sonography in anorectal disorders. Ultraschall Med 25:111-115 15. Maconi G, Ardizzone S, Greco S et al (2007) Transperineal ultrasound in the detection of perianal and rectovaginal fistulae in Crohn's disease. Am J Gastroenterol 102:2214-2219 16. Stewart LK, McGee J, Wilson SR (2001) Transperineal and transvaginal sonography of perianal inflammatory disease. AJR Am J Roentgenol 177:627-632 17. Wootton-Gorges SL, Thomas KB, Harned RK et al (2005) Giant cystic abdominal masses in children. Pediatr Radiol 35:1277-1288 18. Khong PL, Cheung SC, Leong LL, Ooi CG (2003) Ultrasonography of intra-abdominal cystic lesions in the newborn. Clin Radiol 58:449-454 19. Maurer K, Unsinn KM, Waltner-Romen M et al (2008) Segmental bowel-wall thickening on abdominal ultrasonography: an additional diagnostic sign in Kawasaki disease. Pediatr Radiol 38:1013-1016 20. Kim OH, Chung HJ, Choi BG (1995) Imaging of the choledochal cyst. Radiographics 15:69-88 21. Segal SR, Sherman NH, Rosenberg HK et al (1994) Ultrasonographic features of gastrointestinal duplications. J Ultrasound Med 13:863-870 22. Yang WT, Ho SS, Metreweli C (1997) Case report: antenatal sonographic diagnosis of meconium peritonitis and subsequent evolving meconium pseudocyst formation without peritoneal calcification. Clin Radiol 52:477-479 23. Pathi R, Sage M, Slavotinek J, Hanieh A (2004) Abdominal cerebrospinal fluid pseudocyst. Australas Radiol 48:61-63 24. Roitberg BZ, Tomita T, McLone DG (1998) Abdominal cerebrospinal fluid pseudocyst: A complication of ventriculoperitoneal shunt in children. Pediatr Neurosurg 29:267-273 25. Biko DM, Anupindi SA, Hernandez A et al (2009) Childhood Burkitt lymphoma: abdominal and pelvic imaging findings. AJR Am J Roentgenol 192:1304-1315 26. Wong S, Sanchez TR, Swischuk LE, Huang FS (2009) Diffuse peritoneal lymphomatosis: atypical presentation of Burkitt lymphoma. Pediatr Radiol 39:274-276 27. Moholkar S, Sebire NJ, Roebuck DJ (2006) Radiologicalpathological correlation in lipoblastoma and lipoblastomatosis. Pediatr Radiol 36:851-856 28. Prando A, Wallace S, Marins JL et al (1990) Sonographic features of benign intraperitoneal lipomatous tumors in childrenreport of 4 cases. Pediatr Radiol 20:571-574 29. Veyrac C (2008) Intraperitoneal masses. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg, pp 511-544

Pediatric Intestinal Ultrasonography

30. Levy AD, Rimola J, Mehrotra AK, Sobin LH (2006) From the archives of the AFIP: benign fibrous tumors and tumorlike lesions of the mesentery: radiologic-pathologic correlation. Radiographics 26:245-264 31. Leung RS, Calder A, Roebuck D (2009) Embryonal rhabdomyosarcoma of the omentum: two cases occurring in children. Pediatr Radiol 39:865-868 32. Roebuck DJ, Yang WT, Lam WW, Stanley P (1998) Hepatobiliary rhabdomyosarcoma in children: diagnostic radiology. Pediatr Radiol 28:101-108 33. Karmazyn B, Werner EA, Rejaie B, Applegate KE (2005) Mesenteric lymph nodes in children: what is normal? Pediatr Radiol 35:774-777

261

34. Watanabe M, Ishii E, Hirowatari Y et al (1997) Evaluation of abdominal lymphadenopathy in children by ultrasonography. Pediatr Radiol 27:860-864 35. Vayner N, Coret A, Polliack G et al (2003) Mesenteric lymphadenopathy in children examined by US for chronic and/or recurrent abdominal pain. Pediatr Radiol 33:864-867 36. Simanovsky N, Hiller N (2007) Importance of sonographic detection of enlarged abdominal lymph nodes in children. J Ultrasound Med 26:581-584 37. Couture A (2008) Bowel obstruction in neonates and children. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. SpringerVerlag, Berlin-Heidelberg, pp 131-251