Magnetic Resonance Angiography: Principles and Applications

Magnetic Resonance Angiography

wwwwwwwwwwww

James C. Carr, MD Timothy J. Carroll, PhD Editors

Magnetic Resonance Angiography Principles and Applications

Editors: James C. Carr, MD Director of Cardiovascular Imaging Associate Professor of Radiology and Medicine Northwestern University Feinberg School of Medicine Chicago, IL USA [email protected]

Timothy J. Carroll, PhD Associate Professor of Biomedical Engineering Director of MRI Research Department of Radiology Northwestern University Chicago, IL USA [email protected]

ISBN 978-1-4419-1685-3 e-ISBN 978-1-4419-1686-0 DOI 10.1007/978-1-4419-1686-0 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011940427 © Springer Science+Business Media, LLC 2012 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

“I would like to dedicate this to my wife, Jean and my son, Nate, who taught me to love books.” Timothy J. Carroll “I would like to dedicate this book to my wife, Maria, and children, Cristian and Sebastian, without whose constant support, dedication and inspiration, this work would not be possible.” James C. Carr

wwwwwwwwwwww

Preface

It is with tremendous enthusiasm and excitement that we introduce the book Magnetic Resonance Angiography: Principles and Applications to those who are interested in and involved with the ever-expanding field of magnetic resonance angiography (MRA). Vascular disease remains to this day the central cause for many serious clinical conditions, such as stroke, myocardial infarction, and peripheral arterial disease, each of which can precipitate further downstream complications, which can ultimately lead to serious morbidity and death in many individuals. While huge progress has been made to develop novel therapies for vascular pathologies, such as minimally invasive endovascular stents and stem cell therapies, much of these approaches still depend on an accurate depiction of the appearance and extent of abnormalities within the blood vessels. The original gold standard for diagnosing vascular disease was catheter-based digital subtraction angiography, which of course is invasive for the patient and can result in significant complications, albeit rare with current technology. The concept of inserting a needle directly into an artery so that iodinated contrast can be injected rapidly to opacify blood vessels under X-ray visualization will seem barbaric to readers of this textbook. In fact, it was not long ago that the accepted standard for diagnosing peripheral vascular disease was translumbar aortography, where the abdominal aorta was directly accessed percutaneously for diagnostic angiographic purposes. While such diagnostic techniques seem prehistoric and dated in the modern era, it is also a testament to how much progress has been made in the area of noninvasive vascular diagnosis that we have this attitude. Given the invasive alternatives, much effort has been spent on developing alternative noninvasive diagnostic techniques for assessing vascular disease. One of these tools, Doppler ultrasound, is used in routine clinical practice today to assess conditions, such as carotid artery stenosis and peripheral vascular disease. Ultrasound has the advantage of being noninvasive, cheap, and easily portable; however, it is most useful for superficial vessels, not having the penetration to image deeper structures, such as the intracranial vasculature or pulmonary circulation. Computed tomography (CT) is one of the most common diagnostic tools used today in medicine and can also be employed to image the vasculature. CT has achieved particular success with the recent development of multidetector scanners that allow rapid acquisition speeds at spatial resolutions that approximate digital subtraction angiography. While CT is quick and easy to operate, two major drawbacks include exposure to ionizing radiation and the requirement of injected potentially nephrotoxic-iodinated contrast for vessel opacification. Finally, magnetic resonance imaging (MRI), which is also used in all areas of medicine today, is ideally suited for imaging the vasculature. It is a noninvasive imaging tool too; however, in contrast to CT and catheter-based angiography, it does not use ionizing radiation and gadolinium, which is used as the contrast agent, is relatively nontoxic compared to iodinated agents. While MRI also has some disadvantages, such as not being able to scan patients with devices, such as pacemakers, it is the only modality that can combine anatomic depiction with functional assessment in the same study.

vii

viii

MRI emerged as a vascular imaging tool about 30 years ago with the discovery of time-offlight (TOF) imaging, where it became apparent that, as blood flowed through an external magnetic field, it emitted a signal that could be used for imaging purposes. TOF was the original noncontrast MRA technique; however, scan times were long and images were plagued by flow artifacts that could result in misdiagnosis. The technique has benefited from acceleration strategies developed over the years and is still used routinely in the intracranial circulation. The field of MRA really took off in the 1990s with the development of contrast-enhanced MRA. Contrast-enhanced MRA relied on an intravenous injection of a gadolinium-based extracellular contrast agent which, when imaged with a T1-weighted gradient echo pulse sequence, produced bright images of the blood vessels with suppression of the background tissues. The principal impact of contrast-enhanced MRA was that it produced images that were similar to conventional angiograms in very short periods of time. Much of the development in the field of MRA over the following decade focused on improving acquisition speed and spatial resolution for contrast-enhanced MRA. More recently, a condition known as nephrogenic systemic fibrosis (NSF) was described and its cause was linked to patients with renal dysfunction who were exposed to high doses of gadolinium. A lot more is known about NSF now and the condition has nearly been eliminated with better screening of patients and using lower doses of gadolinium. One less talked about consequence was that the field of MRA diverted its attention away from contrast-enhanced MRA to the development of noncontrast MRA techniques. The result was the development of a myriad of new techniques over the last few years that has become confusing to practicing clinicians and scientists in the field. Magnetic Resonance Angiography: Principles and Applications is designed to bring together into a single textbook all of the different MRA techniques, both contrast-enhanced and noncontrast, current contrast agents and implications for NSF, and strategies for applying these techniques in different clinical situations. The book is targeted to physicists, physicians (particularly those specializing in imaging), MRI technologists, residents, fellows, and students, both doctoral and postdoctoral. The book does not claim to have all of the answers and reflects a snapshot of current thinking in the field. Already there are new developments on the horizon. However, we hope to be able to provide a concrete basis for comprehending current MRA techniques and the clinical protocols in which they are applied so that new techniques can be more easily understood. With this objective in mind, we have divided the textbook into two parts. Part I (Chaps. 1–16) is focused on MRA techniques and part II (Chaps. 17–29) is focused on clinical applications. Part I begins with a review of MRI physics as it pertains to MRA. There are chapters devoted to all of the main MRA techniques, including contrast-enhanced MRA, time-of-flight, and phase contrast. There are several chapters devoted to newer noncontrast MRA techniques. A couple of specific areas, such as time-resolved angiography and coronary MRA are addressed independently. We have also attempted to describe in more detail-specific topics, such as highfield MRA, susceptibility-weighted imaging, acceleration strategies, such as parallel imaging, vessel wall imaging, targeted contrast agents, and low-dose contrast-enhanced MRA. Part II encompasses all of the clinical applications for MRA. Each chapter is divided into an initial “techniques” part, which describes the MRA techniques and protocols for that disease and vascular territory, and an “applications” part, which describes the pathology and imaging findings relevant to the disease state being discussed. There may be some repetition of techniques previously described in part I, although not with the same degree of detail. It is hoped that the techniques and protocols discussed will provide a foundation for the reader to develop his/her own protocols. We have deliberately avoided providing “canned” protocols, which may be somewhat restrictive given the numerous MRA techniques currently available. The “applications” part is designed to provide a comprehensive description of different pathologies together with MRA imaging findings. This should be particularly useful to physicians in practice, residents, and fellows in training. We have devoted specific chapters to NSF and its implications for contrast-enhanced MRA, MRI contrast agents, and newer topics, such as interventional MRI.

Preface

Preface

ix

Finally, we would like to thank all of the contributors to this book, without whom this text would not be possible. Each author is a highly respected expert in the field of MRA and this book would not have been feasible without their contribution and dedication. We would like to thank all our colleagues in the MRI community whose inspiration, support, and friendship have been invaluable. We would specifically like to thank our respective mentors, Dr. Paul Finn, MD, and Dr. Charles Mistretta, PhD, whose past and ongoing support has provided guidance and encouragement for many years. We would like to thank Springer for bringing this book to fruition, specifically Jennifer Donnelly and Frances Louie, whose patience and effort have been immeasurable. Chicago, IL Chicago, IL

James C. Carr, MD Timothy J. Carroll, PhD

wwwwwwwwwwww

Contents

Part I MRA Techniques 1

Basic Principles of MRI and MR Angiography .................................................. Frank R. Korosec

3

2

Time-of-Flight Angiography ................................................................................. Seong-Eun Kim and Dennis L. Parker

39

3

Phase-Contrast MRI and Flow Quantification ................................................... Bernd Jung and Michael Markl

51

4

Technical Aspect of Contrast-Enhanced MRA ................................................... Honglei Zhang, Wei Zhang, and Martin R. Prince

65

5

Time-Resolved, Contrast-Enhanced MR Angiography Using Cartesian Methods ...................................................................................... Stephen J. Riederer, Clifton R. Haider, Casey P. Johnson, and Petrice M. Mostardi

75

6

Flow-Dependent Noncontrast MR Angiography ................................................ Mitsue Miyazaki, Satoshi Sugiura, Yoshimori Kassai, Hitoshi Kanazawa, Robert Edelman, and Ioannis Koktzoglu

89

7

Low-Dose Contrast-Enhanced MR Angiography ............................................... Kambiz Nael, Roya Saleh, Gerhard Laub, and J. Paul Finn

107

8

Vessel Wall Imaging Techniques ........................................................................... Rui Li, Niranjan Balu, and Chun Yuan

113

9

Noncontrast Coronary Artery Imaging ............................................................... Allison Hays, Robert G. Weiss, and Matthias Stuber

129

10

Contrast-Enhanced MR Angiography of the Coronary Arteries ...................... Qi Yang and Debiao Li

141

11

MR Angiography and High Field Strength: 3.0 T and Higher.......................... Harald H. Quick and Mark E. Ladd

149

12

Susceptibility Weighted Imaging and MR Angiography.................................... Samuel Barnes and E. Mark Haacke

157

13

Non-Cartesian MR Angiography ......................................................................... Walter Block and Oliver Wieben

169

xi

xii

Contents

14

Parallel Imaging in Angiography ......................................................................... Nicole Seiberlich and Mark Griswold

185

15

Targeted Agents for Wall Imaging ....................................................................... Emily A. Waters and Thomas J. Meade

199

Part II

Clinical Applications

16

Intracranial Arterial and Venous Disease ........................................................... Dariusch R. Hadizadeh, Horst Urbach, and Winfried A. Willinek

213

17

Carotid and Vertebral Circulation: Clinical Applications ................................. Sugoto Mukherjee and Max Wintermark

225

18

Thoracic Aorta ....................................................................................................... Emily Ward and James C. Carr

239

19

Pulmonary MRA .................................................................................................... James F.M. Meaney and Peter Beddy

253

20

Abdominal Aorta and Mesenteric Vessels ........................................................... Klaus D. Hagspiel and Patrick T. Norton

269

21

Renal Vascular Diseases ........................................................................................ Tim Leiner and Henrik Michaely

283

22

MRA: Upper Extremity and Hand Vessels .......................................................... Ruth P. Lim and Vivian S. Lee

297

23

Lower Extremity Peripheral Arterial Disease..................................................... Jeremy D. Collins and Timothy Scanlon

319

24

Noninvasive Imaging for Coronary Artery Disease ............................................ Reza Nezafat, Susie N. Hong, Peng Hu, Mehdi Hedjazi Moghari, and Warren J. Manning

337

25

Venous Imaging: Techniques, Protocols, and Clinical Applications ...................................................................................... Amir H. Davarpanah, Philip Hodnett, Jeremy D. Collins, James C. Carr, and Tim Scanlon

351

26

Pediatric MR Angiography: Principles and Applications .................................. Bharathi D. Jagadeesan and David N. Loy

365

27

Contrast Agents for MR Angiography................................................................. Christoph U. Herborn

381

28

CE-MRA in the Age of Nephrogenic Systemic Fibrosis ..................................... Aditya Bharatha, Sean P. Symons, and Walter Kucharczyk

387

29

Emerging Interventional MR Applications ......................................................... Clifford R. Weiss, Aravindan Kolandaivelu, Jeff Bulte, and Aravind Arepally

395

Index ................................................................................................................................

403

Contributors

Aravind Arepally, MD, FSIR Division of Radiology, Piedmont Healthcare, Atlanta, GA, USA Niranjan Balu, PhD Department of Radiology, University of Washington, Seattle, WA, USA Samuel Barnes, MS Research Assistant, Department of Radiology, Loma Linda University Medical Center, Loma Linda, CA, USA Peter Beddy Consultant Radiologist, St. James’s Hospital, Dublin, Ireland Aditya Bharatha, MD, FRCP(C) Neuroradiology Fellow, Department of Medical Imaging, University Health Network, Toronto, ON, Canada Walter Block, PhD Associate Professor, Departments of Biomedical Engineering, Medical Physics, and Radiology, Wisconsin Institute for Medical Research, University of Wisconsin-Madison, Madison, WI, USA Jeff Bulte, PhD Professor and Director, Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, MD, USA Jeremy D. Collins, MD Assistant Professor of Radiology, Department of Radiology, Northwestern Memorial Hospital and Northwestern University Feinberg School of Medicine, Chicago, IL, USA Amir H. Davarpanah, MD Department of Radiology, Yale School of Medicine, New Haven, Connecticut Robert R. Edelman, William B. Graham Chairman, Northshore University Health System, Evanston IL, USA J. Paul Finn, MD Professor of Radiology, Medicine and Biomedical Physics, Department of Radiology, Ronald Reagan UCLA Medical Center, Los Angeles, CA, USA Mark Griswold, PhD Associate Professor, Department of Radiology, University Hospitals of Cleveland/Case Western Reserve University, Cleveland, OH, USA E. Mark Haacke, PhD Director, MR Research Facility, Department of Radiology, Harper Hospital/Wayne State University, Detroit, MI, USA Dariusch R. Hadizadeh, MD Department of Neuroradiology, University of Bonn, Bonn, Germany

xiii

xiv

Klaus D. Hagspiel, MD Professor of Radiology, Cardiology and Pediatrics, Director, Division of Noninvasive Cardiovascular Imaging, Department of Radiology, University of Virginia Health System, Charlottesville, VA, USA Clifton R. Haider, PhD Department of Radiology, Mayo Clinic, Rochester, MN, USA Allison Hays, MD Assistant Professor of Medicine, Division of Cardiology, Department of Medicine, Johns Hopkins Hospital, Baltimore, MD, USA Christoph U. Herborn, MD, MBA Associate Professor of Radiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Philip Hodnett, MD Department of Cardiovascular Imaging, Northwestern University Feinberg School of Medicine, Chicago, IL, USA Department of Radiology, New York University, NY Susie N. Hong, MD Advanced Cardiac Imaging Fellow, Cardiovascular Division, Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA Peng Hu, PhD Assistant Professor, Department of Radiology, Ronald Reagan Medical Center, Los Angeles, CA, USA Nobuyashu Ichinose, MS Senior Specialist, MRI Systems Development Department, Toshiba Medical Systems Corporation, Otawara, Japan Bharathi D. Jagadeesan, MD Endovascular Surgical Neuroradiology Fellow, Washington University School of Medicine, Barnes-Jewish Hospital, Mallinckrodt Institute of Radiology, Saint Louis, MO, USA Casey P. Johnson Department of Radiology, Mayo Clinic, Rochester, MN, USA Bernd Jung, PhD Department of Radiology, Medical Physics, University Hospital Freiburg, Freiburg, Germany Hitoshi Kanazawa, MS Senior Manager, MR Engineering, MRI Systems Development Department, Toshiba Medical Systems Corporation, Otawara, Japan Yoshimori Kassai, MS Group Manager, MRI Systems Development Department, Toshiba Medical Systems Corporation, Otawara, Japan Seong-Eun Kim, PhD Research Associate, Department of Radiology, Utah Center for Advanced Imaging Research, Salt Lake City, UT, USA Ioannis Koktzoglou, PhD Assistant Professor of Radiology, The University of Chicago, Northshore University Health System, Evanston IL, USA Aravindan Kolandaivelu, MD, BS Assistant Professor of Medicine, Division of Cardiology, Cardiac Arrythmia Service, Johns Hopkins Hospital, Baltimore, MD, USA Frank R. Korosec, PhD Professor of Radiology and Medical Physics, Department of Radiology, University of Wisconsin Hospital and Clinics, Madison, WI, USA Walter Kucharczyk, MD, FRCP(C), FIASTUM Professor, Departments of Medical Imaging and Surgery; Director, MRI and Spectroscopy, Joint Department of Medical Imaging, Senior Scientist; Hans Fischer Senior Fellow, Institute of Advanced Studies, University of Toronto, Toronto, ON, Canada Toronto General Research Institute, Toronto, ON, Canada Technical University of Munich, Munich, Germany Department of Medical Imaging, University Health Network, Toronto, ON, Canada

Contributors

Contributors

xv

Mark E. Ladd, PhD Director, Erhan L. Hahn Institute for MRI, University Duisburg-Essen, Essen, Germany Gerhard Laub, PhD Director, MR R&D West, MR Division, Siemens Healthcare USA, Pleasanton, CA, USA Vivian S. Lee, MD, PhD, MBA Vice Dean of Science, Professor of Radiology, Physiology and Neurosciences, Department of Radiology, New York University Langone Medical Center, New York, NY, USA Tim Leiner, MD, PhD Associate Professor of Radiology, Department of Radiology, Utrecht University Medical Center, Utrecht, The Netherlands Debiao Li, PhD Director, Cedars-Sinai Medical Center, Biomedical Imaging Research Institute, Los Angeles, CA, USA Rui Li, PhD Senior Fellow, Radiology Department, University of Washington, Seattle, WA, USA Ruth P. Lim, MBBS, MMed, FRANZCR Assistant Professor, Department of Radiology, New York University Langone Medical Center, New York, NY, USA David N. Loy, MD, PhD Instructor, Washington University School of Medicine, Barnes-Jewish Hospital, Mallinckrodt Institute of Radiology, Saint Louis, MO, USA Warren J. Manning, MD Cardiovascular Division, Department of Medicine, Department of Radiology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA Michael Markl, PhD Department of Radiology, Medical Physics, University Hospital Freiburg, Freiburg, Germany Director of Cardiovascular MR Research, Associate Professor of Radiology and Biomedical Engineering, Northwestern University Feinberg School of Medicine, Chicago, IL, USA Thomas J. Meade, PhD Eileen M. Foell Professor of Chemistry, Biochemistry and Molecular and Cell Biology, Neurobiology and Physiology, and Radiology, Department of Chemistry, Northwestern University, Evanston, IL, USA James F.M. Meaney, MB, FRCR, FFR Professor, Department of Radiology, Trinity College Dublin, St. James’s Hospital, Dublin, Ireland Henrik Michaely, MD Associate Professor of Radiology, Section Chief of Vascular and Abdominal Imaging, Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, Germany Mitsue Miyazaki, PhD Senior Fellow, MRI Department, Toshiba Medical Research Institute, Vernon Hills, IL, USA Mehdi Hedjazi Moghari, PhD Research Fellow, Department of Medicine, Cardiovascular Division, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA Petrice M. Mostardi Department of Radiology, Mayo Clinic, Rochester, MN, USA Sugoto Mukherjee, MD Assistant Professor, Section of Neuroradiology, Department of Radiology, University of Virginia Health System, Charlottesville, VA, USA Kambiz Nael, MD Radiology Resident, Department of Radiological Sciences, Ronald Reagan UCLA Medical Center, Los Angeles, CA, USA

xvi

Reza Nezafat, PhD Assistant Professor of Medicine, Director of Translational Cardiovascular Imaging Program, Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA Patrick T. Norton, MD Assistant Professor of Radiology, Cardiology and Pediatrics, Division of Noninvasive Cardiovascular Imaging, Department of Radiology, University of Virginia Health System, Charlottesville, VA, USA Dennis L. Parker, PhD Director, Department of Radiology, Utah Center for Advanced Imaging Research, Salt Lake City, UT, USA Martin R. Prince, MD, PhD Professor of Radiology, Department of Radiology, Weill Medical College of Cornell University, New York, NY, USA Harald H. Quick, PhD Director of MR Imaging, Institute of Medical Physics, University Erlangen-Nürnberg, Erlangen, Germany Stephen J. Riederer, PhD Professor, Department of Radiology, Mayo Clinic, Rochester, MN, USA Roya Saleh, MD Researcher, Department of Radiology, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA Timothy Scanlon, MD, MRCPI, FFR RCSI Department of Cardiovascular Imaging, Radiology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA Consultant Radiologist, Limerick Regional Hospital, Ireland Nicole Seiberlich, PhD Research Associate, Department of Radiology, University Hospitals of Cleveland/Case Western Reserve University, Cleveland, OH, USA Matthias Stuber, PhD Professor and Director, Department of Radiology, Center for Biomedical Research, University Hospital Lausanne, Lausanne, Switzerland Satoshi Sugiura Chief Specialist, MRI Systems Development Department, Toshiba Medical Systems Corporation, Otawara, Japan Sean P. Symons, MD, MPH, FRCP(C) Neuroradiologist, Department of Medical Imaging, Sunnybrook Health Sciences Centre, Toronto, ON, Canada Horst Urbach, MD Professor, Department of Radiology/Neuroradiology, University of Bonn, Bonn, Germany Emily Ward, MB, BCh, BAO Department of Radiology, Northwestern Memorial Hospital, Chicago, IL, USA Emily A. Waters, PhD Department of Chemistry, Northwestern University, Evanston, IL, USA Clifford R. Weiss, MD Assistant Professor of Radiology and Surgery, Division of Vascular and Interventional Radiology, Department of Radiology, Johns Hopkins Hospital, Baltimore, MD, USA Robert G. Weiss, MD Professor of Medicine and Radiology, Department of Medicine, Johns Hopkins Hospital, Baltimore, MD, USA Oliver Wieben, PhD Assistant Professor, Departments of Medical Physics and Radiology, University of Wisconsin-Madison, Wisconsin Institute for Medical Research, Madison, WI, USA Winfried A. Willinek, MD Associate Professor, Department of Radiology, University of Bonn, Bonn, Germany

Contributors

Contributors xvii

xvii

Max Wintermark, MD, MAS Associate Professor of Radiology, Neurology, Neurological Surgery and Biomedical Engineering, Department of Radiology, University of Virginia, Charlottesville, VA, USA Qi Yang, MD, PhD Attending Radiologist, Department of Radiology, Xuanwu Hospital, Capital Medical University, Beijing, China Chun Yuan, PhD Professor, Department of Radiology, University of Washington, Seattle, WA, USA Honglei Zhang, MD Assistant Professor, Department of Radiology, Weill Medical College of Cornell University, New York, NY, USA Wei Zhang, MD Research Fellow, Department of Radiology, Weill Medical College of Cornell University, New York, NY, USA

wwwwwwwwwwww

Part I MRA Techniques

wwwwwwwwwwww

1

Basic Principles of MRI and MR Angiography Frank R. Korosec

Introduction Magnetic resonance imaging (MRI) is a very versatile and useful imaging modality. It is capable of providing a wealth of diagnostic information, including information regarding blood flow. Magnetic resonance methods that provide images of the arteries are referred to as magnetic resonance angiography (MRA) methods. A variety of MRA methods exist. These methods can be categorized as noncontrast-enhanced and contrast-enhanced methods. The noncontrast-enhanced methods rely on the motion of blood (phase-contrast MRA, TOF MRA) or the magnetic properties of blood (steady-state free precession MRA) to differentiate signals from blood and stationary tissues, whereas contrast-enhanced methods require the intravenous injection of a contrast material to differentiate signals from blood and stationary tissues. In addition to providing information regarding the morphology of the blood vessels, some MRA methods can provide information regarding blood velocity or volume flow rates. If the MRA acquisitions are cardiac-gated, they can provide information regarding changes in blood velocity or volume flow rate throughout the cardiac cycle. Most MRA methods permit large volumes to be imaged, and the image sets may be retrospectively processed to provide observation of vessels from any perspective. Because MRI provides high-contrast images of soft tissues, MRI exams may be performed together with MRA exams to yield information regarding blood flow and target organ status in the same exam. In addition to providing information regarding vascular and tissue anatomy and blood flow, MRI can be used to obtain information regarding diffusion and perfusion (and a host of other qualities), making it a very

F.R. Korosec, PhD () Department of Radiology, University of Wisconsin Hospital and Clinics, E3/311, 600 Highland Avenue, Madison, WI 53792-3252, USA e-mail: [email protected]

effective modality for obtaining a comprehensive assessment of vascular disease. There are a great number of MRA methods, and each derives vascular signal by employing features and imaging parameters that take advantage of differences in physical properties between blood and stationary tissues. In order to most effectively utilize the MRA sequences and reap their greatest benefits, it is essential to have a good understanding of the principles of MRI; knowledge of the strengths, limitations, and capabilities of the different MRA methods; and comprehension of how the imaging parameters, and features of each of the MRA methods influence image quality. The physical principles of MRI and MRA will be briefly described in this chapter. Details of the MRA methods are described in more detail in later chapters.

MRI Overview MRI offers a number of benefits over some of the other diagnostic imaging modalities. First, MRI derives signals using a magnetic field and radiofrequency energy, not ionizing radiation or radioactive materials. Second, MRI is able to demonstrate, with striking contrast, differences in signal intensities among different soft tissues. A host of MR imaging parameters can be modified to exploit a variety of tissue-specific properties to manipulate image contrast. Third, MRI is capable of providing tomographic images of any plane without requiring movement of the patient or any of the equipment. The images may be acquired as a series of two-dimensional (2D) slices, or as a true three-dimensional (3D) volume. MRI may be used to image nuclei of atoms containing an odd number of protons and/or neutrons. The nucleus of the hydrogen atom, 1H, on the water molecule satisfies this criterion as it is composed of a single proton. Because of its abundance in the body, the hydrogen nucleus on the water molecule is routinely imaged in MRI. The protons on the fat molecule (as well as the protons on some other molecules) also appear in MR images.

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_1, © Springer Science+Business Media, LLC 2012

3

4

MRI systems use a strong magnetic field, B0, which is uniform throughout the imaging volume in the bore of the system, and is stable over time. The magnetic field strength is measured in tesla (T) or gauss (G), where 1 T = 104 G. The most commonly used clinical MRI scanners are whole-body imaging systems, with bore diameters of 60 cm or 70 cm,

F.R. Korosec

Fig. 1.1 A state-of-the-art 3 T MRI scanner

and magnetic field strengths of 1.5 T or 3 T. A state-of-the-art 3 T MRI system with a bore diameter of 60 cm is shown in Fig. 1.1. These systems use superconducting magnets, and require a reservoir surrounding the magnet to be filled with liquid helium periodically. The direction of the magnetic field is along the long axis (z-axis) of the bore. Resistive and permanent magnets also are used in MRI systems, but the field strengths of these systems are lower. The magnetic fields of some of these systems are aligned vertically. The hydrogen nucleus possesses a magnetic dipole moment and, from a classical physics perspective, will interact with a magnetic field as if it were a tiny bar magnet. The nuclear magnetic dipole moment of a hydrogen nucleus can either align or antialign with the strong magnetic field of the MRI scanner. In an ensemble of hydrogen nuclei, the majority of the nuclei will align with the magnetic field because it requires less energy to align than to antialign. The net sum of the nuclear magnetic dipole moments from an ensemble of nuclei will yield a bulk magnetization that is aligned with the applied magnetic field. This concept is demonstrated in Fig. 1.2. It is this bulk magnetization that is considered in discussions regarding MRI. In MRI, signal is generated when the bulk magnetization, M, is “tipped” out of alignment with the applied magnetic

Fig. 1.2 (a) Shown on a single xyz coordinate system for comparison with one another are vector representations of magnetic dipole moments aligned with (on the surface of the upward facing cone), and antialigned with (on the surface of the downward facing cone) the main magnetic field, B0. (b) Each vector has a longitudinal component, Mz, and a transverse component, Mxy. (c) Summing the transverse components yields a net sum of zero magnetization (since the orientation in the transverse plane is random, if enough dipole moments are considered, for each one

pointing in a given direction, statistically, there is likely one pointing in the opposite direction, resulting in total cancelation). (d) Summing the longitudinal components yields a net excess of magnetization aligned with the main magnetic field (since it requires less energy for the magnetic dipole moments to align with the magnetic field). (e) The result of summing the contributions from an ensemble of dipole moments is a net, or bulk, magnetization that aligns with the orientation of the main magnetic field, B0

1

Basic Principles of MRI and MR Angiography

5

Fig. 1.3 The bulk magnetization, M, rotates, or precesses, when it is in the transverse plane. The receiver coil detects an oscillating signal from the component of magnetization that points toward it (My in this case). The rate of precession and, therefore, the frequency of the detected signal, are proportional to the strength of the magnetic field, B0. In this diagram, the magnetic field is aligned along the z-axis

field, B0. The mechanism for tipping the magnetization is described below. The maximum signal is generated when the aligned (longitudinal) magnetization is tipped 90° so that it is perpendicular (transverse) to the direction of the applied magnetic field. The longitudinal magnetization is referred to as Mz, and the transverse magnetization is referred to as Mxy. The transverse component of magnetization rotates, or precesses, around the main magnetic field as shown in Fig. 1.3. The precessing transverse component of the magnetization can be detected using a receiver coil oriented perpendicular to the direction of the main magnetic field. The precessing magnetization induces a current in the receiver coil, proportional to the rate of change of the transverse component of the magnetization. Thus, it is the precessing transverse component of the magnetization that produces the MR signal. Because the bulk magnetization spins around the main magnetic field, and because the bulk magnetization is composed of nuclear magnetic dipole moments, the nuclear magnetic dipole moments are often referred to as “spins” in the context of describing the nuclear magnetic resonance phenomenon. The precessional rate of the bulk magnetization is proportional to the strength of the applied magnetic field, B0, and is given by the Larmor equation: ω 0 = g B0 ,

(1.1)

where ω0 is the precessional frequency, g is a constant referred to as the gyromagnetic ratio (its value depends on the nucleus under consideration, 42.58 MHz/T for hydrogen), and B0 is the strength of the main magnetic field. At a field

Fig. 1.4 When a magnetic field, B1, is applied along the x¢-axis in a frame of reference that rotates at the Larmor frequency, the bulk magnetization, M, will precess around B1 in the y¢z¢ plane. This method is used to tip the magnetization away from the longitudinal axis so that a component, Mxy, appears in the transverse plane

strength of 1.5 T, the precessional frequency of the bulk magnetization associated with the hydrogen atom is 63.87 million revolutions per second (63.87 MHz). The precessional frequency is commonly referred to as the Larmor frequency. When dealing with precessing magnetization, discussions are often made easier by introducing the rotating frame of reference. The rotating frame of reference is a frame of reference, such as a coordinate system, that rotates with the object that is being investigated. In this case, the object that is being investigated is the precessing bulk magnetization. Thus, if the xyz reference frame is made to rotate about the z-axis such that the x- and y-axes rotate at the precessional rate of the magnetization, it will appear as if the magnetization is not moving with respect to the rotating x- and y-axes, because the bulk magnetization will always stay in the same position relative to the x- and y-axes. Therefore, in the rotating frame, the precessional frequency of the bulk magnetization is zero. In the rotating frame, it is as if the main magnetic field, B0, is having no effect on the magnetization, so it can be ignored. With the introduction of the rotating frame, and the absence of B0 in this frame, it can be understood that the bulk magnetization can be tipped away from the longitudinal axis simply by applying a magnetic field perpendicular to the longitudinal axis. When this is done, the bulk magnetization will precess around this field. So, for example, if a magnetic field is applied along the x-axis in the rotating frame, the magnetization will precess in the yz plane as shown in Fig. 1.4. The magnetic field that is applied perpendicular to the bulk magnetization to tip it away from the longitudinal axis is referred to as the B1 magnetic field. Note that, in order to

6

Fig. 1.5 As viewed from the stationary frame of reference, the magnetization, M, precesses around the z-axis (at the Larmor frequency) as it is tipped away from this axis, tracing out a trajectory similar to the one shown here. The magnetic field, B1, is applied at the Larmor frequency so that it remains perpendicular to the precessing magnetization, M (ensuring that it is continuously acting to tip the magnetization away from the z-axis)

tip the magnetization away from the longitudinal axis, the B1 field must rotate at the Larmor frequency. If B1 does not precess at the Larmor frequency, the B0 field can no longer be ignored, and the magnetization will precess about some effective field determined by the combination of the B0 and B1 fields. Because the B1 magnetization precesses at 63.87 MHz, and this is in the radiofrequency range of the electromagnetic spectrum, the B1 pulse is often referred to as a radiofrequency (RF) pulse. The process of tipping the magnetization is often referred to as RF excitation. The strength of the RF excitation pulse is on the order of 50 mT, and the duration is typically on the order of just a few milliseconds. If the magnetization were observed from a stationary frame of reference as it were being tipped away from the longitudinal axis, it would be seen to precess about the longitudinal axis as it was being tipped, as shown in Fig. 1.5. The bulk magnetization can be brought from the z-axis into the transverse plane by applying a B1 pulse of the appropriate strength and duration. A B1 pulse that has such an effect is referred to as a 90° RF pulse, since this is the angle traversed by the tipping magnetization. As will be described later, it often is advantageous to tip the magnetization less than 90°. In this case, only a component of the magnetization will be brought into the transverse plane, and the signal will be proportional to sin(q), where q is the tip angle. MRI systems with higher field strengths produce more magnetization (more dipole moments align with a magnetic field of higher strength), which leads to generation of higher signals. Scanners with magnetic field strengths of 3 T are becoming more prevalent, and MRA methods performed on these systems yield excellent results [1–5].

F.R. Korosec

A host of RF coils are available for MRI and MRA applications at various magnetic field strengths. Coils that are cylindrical in shape may be used for transmission of the B1 pulse, and reception of the MR signal. These coils are typically referred to as transmit/receive RF coils. Coils that are not cylindrical in shape generally do not provide a uniform tip angle spatially, so these coils are used only for reception of the MR signal. These coils are referred to as receive only coils. When these coils are used, a cylindrical body coil that is built into the bore of the MRI system is usually used to transmit a spatially uniform B1 pulse. Some coils are composed of multiple smaller coils, or elements, each of which sends its signal to a separate receiver (amplifier, analogue-to-digital converter, etc.). Smaller elements are sensitive to noise from smaller regions of anatomy, and the noise detected by each element is uncorrelated. When the signals and noise from all elements are combined, the result is an image with improved signal-to-noise ratio relative to an image acquired with a single coil that is sensitive to noise in the entire volume covered by the multiple coil elements. These coils are referred to as phased array coils. Vendors currently are providing MRI systems with 32 or more receivers (channels) so phased array coils with 32 elements may be used. For some phased array coils, the number of elements exceeds the number of receivers in the MRI system. With these coils, signals from multiple elements may be combined and sent to a single receiver. Alternatively, the operator may specify that only certain elements be activated during a scan. For example, in a head/neck/spine coil, the operator may chose to activate only the elements required to image the head. In a subsequent scan, the operator may chose to activate only the elements required to image the neck. Some RF coils available for MR imaging are shown in Fig. 1.6.

MRI Contrast Mechanisms In MRI, varying degrees of contrast between different tissues can be achieved by modifying the MR imaging parameters. In this section, three magnetic properties of tissues, namely T2 relaxation, T1 relaxation, and proton density, are described. Later in this chapter, the imaging parameters that are modified to manipulate contrast based on these properties are discussed.

T2 (or Spin–Spin) Relaxation Time It already was shown that the bulk magnetization from a sample can be tipped into the transverse plane using a radiofrequency pulse. It also was shown that the transverse magnetization precesses, and as it does, it induces a signal in the

1

Basic Principles of MRI and MR Angiography

7

Fig. 1.6 Shown here are some of the coils used for MRI and MRA applications, including (a) an 8-channel head coil, (b), an 8-channel head and neck coil, (c) a 4-channel phased array peripheral vascular coil, and (d), an 8-channel phased array torso coil

receiver coil. What was not mentioned is that this transverse magnetization diminishes very rapidly. The rate at which the bulk transverse magnetization decreases is characterized by a constant referred as the T2 decay constant or the T2 relaxation time. Because it refers to relaxation of the transverse magnetization, it is often referred to as the transverse relaxation time. Also, because the effect is caused by spins interacting with neighboring spins, the decay constant is often referred to as the spin–spin relaxation time. The mechanism responsible for this diminution of transverse magnetization is described below. Recall that spins precess at a rate determined by the magnetic field that they experience (the Larmor equation). Recall also that, at 1.5 T, the spins from hydrogen precess at a rate of 63.87 MHz. In tissues, the magnetic field that each nucleus experiences is not exactly equal to the external magnetic field. This is because the magnetic field experienced by a spin is affected by the magnetic fields of the spins in its microscopic neighborhood. The microscopic neighborhood around a spin can cause the spin to experience a field slightly larger, or slightly smaller, than the applied external magnetic field. In tissues, the spins are constantly in motion, so the microscopic neighborhoods of spins are continually changing. The net result is that spins that are tipped into the transverse plane precess at slightly different frequencies (even if they are subjected to a uniform applied external magnetic field), and the precessional frequencies of the spins change over time. This is true even if the spins are in a macroscopically

homogeneous tissue, because of the presence of microscopic inhomogeneities. Due to the differences in the precessional frequencies of spins in different microscopic neighborhoods, the spins throughout the tissue eventually get out of synchronization with one another. That is, the spins that precess faster get ahead of those that precess slower. It is this dispersion, or dephasing, of spins that causes the net transverse magnetization to diminish over time. The behavior of the magnetic dipole moments during T2 relaxation is shown in Fig. 1.7. This figure shows that, in the rotating frame, spins that precess faster than 63.87 MHz, appear to precess clockwise (as viewed from the +z-axis), and spins that precess slower than 63.87 MHz appear to precess counterclockwise. As the spins get more out of phase with one another, components along the −x-axis cancel components along the +x-axis, and eventually components along the −y-axis cancel components along the +y-axis. The time it takes for the transverse magnetization to decrease to 37% of the initial magnetization is referred to as the T2 relaxation time. The T2 relaxation time is determined by the mobility of the spins in the tissue. Thus, different tissues with different spin mobility have different T2 relaxation times. The longer the T2 of a tissue, the longer its transverse magnetization persists. Differences in T2 times of different tissues can be exploited in MR imaging to achieve contrast among different tissues. Spin dephasing is exacerbated by inhomogeneities in the main magnetic field, B0. This is because the precessional rate

8

F.R. Korosec

Fig. 1.7 The rate of precession of spins is affected by neighboring nuclei. In the rotating frame of reference, spins that precess faster than 63.87 MHz appear to precess clockwise (as viewed from the +z-axis) relative to the reference frame, and spins that precess slower than 63.87 MHz appear to precess counterclockwise relative to the reference frame. This dephasing of the spins causes the magnitude of the transverse magnetization, Mxy, to decrease over time. The plot shows the

magnitude of the transverse magnetization as a function of time for two tissues. The time that it takes for the transverse magnetization to diminish to 37% of its initial magnitude is characterized by the T2 time constant. Different tissues have different compositions of nuclear neighbors, and therefore have different T2 time constants. This characteristic is exploited in MRI to obtain different signals from tissues with different T2 time constants

is dictated by the field strength (the Larmor equation). If the field varies spatially, then spins in different locations will have different precessional rates. This will lead to spin dephasing and loss of transverse magnetization in a manner similar to that responsible for T2 decay. Other factors contribute to spin dephasing as well (susceptibility differences, velocity distributions, acceleration distributions, etc.) The decay constant that characterizes how long it takes for the transverse magnetization to diminish to 37% of its initial strength when spin–spin interactions (T2) and all other effects are considered is referred to as T2* (T2 star). The formula for calculating T2* is

Recall that the net longitudinal magnetization is zero after the application of a 90° radiofrequency pulse. This is because after the application of the 90° RF pulse, the number of spins antialigned with the magnetic field is equal to the number of spins aligned with the magnetic field. In other words, the energy imparted to the spin system by the RF pulse causes half of spins to line up in a direction opposite to the magnetic field. After the application of the 90° RF pulse, antialigned spins give up energy to the lattice of the tissue, causing them to once again align with the B0 field. As spins continue to give up energy, aligned spins increasingly out number antialigned spins. The time constant that characterizes how long it takes for the longitudinal magnetization to return to 63% of its initial thermal equilibrium value (the value that it had before it was subjected to the RF pulse) is referred to as the T1 relaxation time. Because this process results from interactions of the spins with the structure of the tissue, or the tissue lattice, the T1 time constant is often referred to as the spin–lattice relaxation time; and because it characterizes the time it takes for the longitudinal magnetization to regrow, it often is referred to as the longitudinal relaxation time. The behavior of the magnetic dipole moments during T1 relaxation is shown in Fig. 1.8. As is the case with T2 relaxation, the T1 relaxation time is determined by the mobility of the spins in the tissue. Different tissues have different T1 relaxation times. The longer the T1 of a tissue, the longer it takes for its longitudinal magnetization to regrow. Differences in T1 times of different tissues can be exploited in MR imaging to achieve varying degrees of contrast among different tissues.

1 1 1 = + , * T2 T2 T2¢

(1.2)

where T2¢ is a time constant that characterizes the loss of transverse magnetization from all factors other than tissue specific, spin–spin interactions. Whereas dephasing due to T2 is advantageous because it permits tissues to be differentiated from one another, dephasing caused by other effects tends to unnecessarily diminish the transverse magnetization. Thus, it is desirable to minimize effects of signal loss from sources other than spin–spin interactions.

T1 (or Spin–Lattice) Relaxation Time At the same time that the transverse magnetization is diminishing, the longitudinal magnetization is increasing.

1

Basic Principles of MRI and MR Angiography

9

Fig. 1.8 After the magnetization is tipped into the transverse plane by the B1 pulse, the spins begin to collide with macromolecules that compose the lattice of the tissue, they give up energy, and they flip from the antialigned state to the aligned state. As more spins flip to the aligned state, the longitudinal magnetization, Mz, grows back to its initial thermal equilibrium value, M0. The plot shows the magnitude of the

longitudinal magnetization as a function of time for two tissues. The time that it takes for the longitudinal magnetization to regrow to 63% of its initial magnitude is characterized by the T1 time constant. Different tissues have different T1 time constants. This characteristic is exploited in MRI to obtain different signals from tissues with different T1 time constants

Table 1.1 Typical T1- and T2-relaxation times for selected tissues in a 1.0 T magnetic field

Table 1.2 Relative T1- and T2-relaxation times of different states of matter

White Matter Gray Matter Fat CSF

T1 (ms)

T2 (ms)

400 500 180 2,000

90 100 90 300

Solid Semisolid Liquid

T1

T2

Long Intermediate Long

Short Intermediate Long

CSF cerebrospinal fluid

Facts Regarding T1 and T2 Relaxation Rates As mentioned above, different tissues have different T1 and T2 values. This permits different tissues to be displayed with different signal intensities in MR images. The signal intensities of different tissues can be altered by manipulating the imaging parameters. Typical relaxation times for selected tissues (in a 1.0 T magnetic field) are listed in Table 1.1. It was stated above that both T2 and T1 relaxation times are dependent on the mobility of spins. The spins in liquids are fairly mobile. Therefore, liquids have long T2 values (see T2 of CSF in Table 1.1). At 1.5 T, the T2 of arterial (oxygenated) blood is about 250 ms, and the T2 of venous blood is about 220 ms. The spins in solids are relatively immobile, so, solids have short T2 values. In fact, the T2 values of solids are so short, that signals from solids do not appear in MRI images; the signals disappear before they can be detected. The T2 values of tissues are intermediate between the T2 values of liquids and solids.

T1 relaxation also is dependent on spin mobility, but in a fairly complicated manner. At 1.5 T, the T1 of arterial and venous blood is about 1,200 ms. The relative T1 and T2 times of solids, semisolids, and liquids are summarized in Table 1.2. These characteristics are schematically plotted in Fig. 1.9. For blood and most tissues, the T1 values increase with field strength (e.g., the T1 of arterial and venous blood is about 1,600 ms at 3 T), whereas the T2 values decrease slightly. The T1 and T2 values of blood and tissues may be decreased by intravenously injecting a gadolinium-based contrast material. Although T2 and T1 relaxation are described separately, they occur simultaneously. That is, as the spins are dephasing in the transverse plane, spins are also going from being antialigned with the B0 field to being aligned with it. Spin flips influence T1 and T2 (as spins flip, they lose coherence in the transverse plane). T2 is additionally influenced by spin– spin interactions. Thus, for most tissues imaged using MRI, T2 < T1. This means that the net magnetization in the transverse plane disappears before the longitudinal magnetization fully regrows. It is only for liquids that T2 may approach T1.

10

F.R. Korosec

Fig. 1.9 Schematic representation of (a) relative T1 relaxation times and (b) relative T2 relaxation times, of various states of matter

Bloch Equation The behavior of the net magnetization of protons due to excitation and relaxation can be summarized by the Bloch equation: M x i + M y j ( M z - M o )k dM , = M´g Bdt T2 T1

(1.3)

where M is the net magnetization, B represents the applied magnetic fields (including B0, B1, and gradient fields) and i, j, and k are unit vectors in the x, y, and z directions, respectively. Note that the behavior depends on both T1 and T2, which results in contrast between tissues with different T1 and T2 values. Solving this governing equation for M, with time-varying magnetic field gradients, yields the signal equations for different imaging sequences. In addition to being influenced by T2, T1, and proton density, the signal in MRI is influenced by diffusion, perfusion, magnetic susceptibility, chemical shift, temperature, magnetic field strength, motion (including that from flowing blood), presence of a contrast material, and many other factors. MRA techniques achieve contrast by harnessing the effects of motion (phase-contrast and TOF MRA), or the effects of injecting a gadolinium-based T1-shortening contrast material (contrast-enhanced MRA), or the inherently long T2 of blood (steady-state free precession MRA). The physical principles of the more common MRA techniques are introduced below.

Imaging Parameters and Their Effects on Image Contrast Signal intensity in MR depends on a number of factors, and there are a host of parameters in MR imaging sequences that can be modified to accentuate the influence of these factors. The signal intensity is proportional to the magnitude of the transverse magnetization at the time that the receiver is turned on. The rate at which transverse magnetization diminishes is characterized by the time constant T2. Different tissues have different T2 values. Tissues with short T2 values lose transverse magnetization quickly, whereas tissues with long T2 values lose it more slowly. Thus, allowing time to

elapse between tipping the magnetization transverse and sampling it allows achievement of signal differences from tissues with different T2 values. The time from when the magnetization is tipped transverse until the signal is detected is controlled by the MR imaging parameter TE (echo time). Because imaging sequences with longer echo times achieve signal differences based on variations in T2 decay times, they are referred to as T2-weighted imaging sequences, and the images produced with these sequences are referred to as T2-weighed images. The magnitude of the transverse magnetization is influenced by the magnitude of the longitudinal magnetization, since a component of longitudinal magnetization becomes transverse after the application of the RF pulse. After the longitudinal magnetization is tipped transverse, the longitudinal component begins to regrow. The rate of regrowth is characterized by the time constant called T1. Different tissues have different T1 regrowth times. In MR imaging, the longitudinal magnetization must be tipped many times in order to encode enough information to map the signals to the proper locations in the image (as described below). If only a short time elapses between applying the RF excitation pulses, the longitudinal magnetization will not fully regrow. The longitudinal magnetization from short T1 tissues will regrow more than the longitudinal magnetization from long T1 tissues. Thus, allowing only a short amount of time to elapse between applying RF excitation pulses will lead to different amounts of longitudinal magnetization being tipped transverse, which will result in different signal intensities from different tissues based on variations in T1 regrowth times. The time between applying RF excitation pulses is controlled by the MR imaging parameter TR (repetition time). Because imaging sequences with shorter repetition times achieve signal differences based on variations in T1 regrowth times, they are referred to as T1-weighed imaging sequences, and the images produced with these sequences are referred to as T1-weighed images. In order to minimize T2-weighting in these scans, the shortest possible echo time is used. In order to minimize T1-weighting in T2-weighted scans, a longer TR is used. When a long TR and a short TE are used, the signal intensities in the images are not dependent on T1 or T2 differences. In this case, the signal intensities depend on the density of hydrogen nuclei (on the water molecule) in each tissue. Since the hydrogen nucleus consists of a single proton, these imaging sequences are referred to as proton-density-weighted (also r-weighted) imaging sequences, and the images produced with these sequences are referred to as proton-densityweighted images. When is it desirable to achieve signal intensities in images that are strictly dependent on T1 values, T2 values, or proton densities of various tissues, a class of imaging sequences called spin-echo sequences is used. Typical TR and TE values used to achieve T1-, T2-, and

1

Basic Principles of MRI and MR Angiography

11

Table 1.3 TR and TE imaging parameters used to achieve different weighting in magnetic resonance images of the brain using a spin-echo imaging sequence at 1.5 T T1-weighting T2-weighting r-weighting

TR Short (400 ms) Long (3,000 ms) Short (3,000 ms)

TE Short (20 ms) Long (100 ms) Long (20 ms)

density-weighed images of the brain using a spin-echo imaging sequence are shown in Table 1.3. Spin-echo imaging sequences typically have scan times and properties that make them inappropriate for many MRA applications. Most MRA methods employ sequences called gradient-echo imaging sequences. In gradient-echo sequences, short TR and short TE values are used. This results in the signal for each tissue being dependent on combined effects of the T1, T2, and proton density of the tissue. There are several gradient-echo sequences used for MRA. Each yields signals that have different dependencies on the physical properties of the blood and tissues being imaged. For these sequences, additional imaging features may be employed to enhance signal differences between blood and stationary tissues. Also, imaging parameters, including tip angle, may be adjusted to enhance signal differences.

MR Image Formation To produce images in MRI, it is necessary to determine the origins of the signals so that they can be mapped to the appropriate positions in the image. This mapping of signal to position is accomplished by applying magnetic field gradients of predetermined amplitude and duration. To say that a magnetic field has a gradient means that it has a different strength at different locations in space. In MR imaging, the gradients are applied so that the magnetic field strength changes linearly with position. Magnetic field gradients can be applied along each of the three axes in the MR scanner. A magnetic field gradient has units of G/cm (or mT/m), and is represented as Gx, Gy, or Gz, depending on the axis along which it is applied. The gradient magnetic field adds to, or subtracts from, the main magnetic field B0. Thus, if a magnetic field gradient is applied along the x-axis, the total magnetic field at any position along the x-axis is given by B0 + Gx x, where x is the position along the x-axis. The value of x is 0 at the center of the magnet (isocenter of the gradient), and has greater positive values with increasing distance from the center in one direction, and greater negative values with increasing distance from the center in the other direction. If a 1 gauss/cm magnetic field gradient is applied in a 1.5 T scanner, the total magnetic field 10 cm from the isocenter of the gradient in one direction is 15,010 G

Fig. 1.10 The net magnetic field produced by applying a linear magnetic field gradient in the x direction. The lengths of the vectors represent the strength of the magnetic field. The gradient strength is denoted as Gx, and it is superimposed on the main magnetic field, B0. Note that when a gradient is applied along the x direction, the field varies linearly in the x direction, but is constant along the y and z directions

(15,000 G + 1 G/cm × 10 cm) and the total magnetic field 10 cm from the isocenter of the gradient in the other direction is 14,990 G (15,000 G – 1 G/cm × 10 cm), where the minus sign in the latter case is due to the negative direction with respect to the isocenter of the gradient. The vector representation of the magnetic field strength as a function of position resulting from the application of a linear magnetic field gradient is shown in Fig. 1.10. As was described previously, in the rotating frame of reference, the main magnetic field, B0, can be ignored. So, with this formalism, the total magnetic field that must be considered in the rotating frame becomes Gx × x for a magnetic field gradient applied along the x-axis, Gy × y for a magnetic field gradient applied along the y-axis, and Gz × z for a magnetic field gradient applied along the z-axis. For the remainder of this chapter, all discussions will pertain to the rotating frame of reference. Vector diagrams representing application of magnetic field gradients along the x-, y-, and z-axes in the rotating frame of reference are shown in Fig. 1.11. The strength of the magnetic field gradient can be plotted as a function of time. Such a plot is referred to as a pulse sequence timing diagram. Figure 1.12 demonstrates the characteristics of a pulse sequence timing diagram. In pulse sequence timing diagrams, the horizontal dimension represents time, and the vertical dimension represents the strength of the gradient at each point in time. The timing diagram shown in the top row of Fig. 1.12 represents a gradient that is applied along the x-axis (Gx) at a strength of 0.75 G/cm, for 4.0 ms. The gradient is turned on 1.0 ms after the clock starts, and it is turned off 5.0 ms after the clock starts. If the gradient were applied at a strength of 0.5 G/cm, and the timing parameters remained the same, the pulse sequence

12

Fig. 1.11 Vector representations of magnetic field gradients applied along the x-, y-, and z-axes. These representations ignore the contribution from B0. The length of the vector at each position represents the strength of the magnetic field at that position (less the contribution from B0). Note that all of the vectors point in the direction of the main magnetic field, meaning that they all add to, or subtract from, the main field, B0. In MRI, no magnetic fields are applied perpendicular to the main magnetic field (except the B1 pulses). In these diagrams, some of the vectors have

Fig. 1.12 (Top left) A timing diagram showing the strength of a magnetic field gradient (vertical axis) as a function of time (horizontal axis). Shown on the top right is a vector representation of the magnetic field gradient. The length of the vectors at each position (along the x dimension in this case) represents the strength of the magnetic field at that position. The bottom row shows a timing diagram and vector representation for a magnetic field gradient having a lower strength than the one shown in the top row

timing diagram would look like the one shown in the bottom row of Fig. 1.12. Typically, pulse sequence timing diagrams are used to give a rough indication of gradient strength and timing, and

F.R. Korosec

been intentionally omitted to reduce clutter. To complete Gx, the vectors at each x position should be reproduced along the y dimension (as shown in Fig. 1.10) and then each of those should also be reproduced along the z dimension. To complete Gy, the vectors at each y position should be reproduced along the x dimension, and then each of these should be reproduced along the z dimension. For Gz, the vectors at each z position are reproduced along the x dimension, but to complete the diagram, each of these vectors should be reproduced along the y dimension

so they do not include markings indicating actual times or gradient strengths. The remaining pulse sequence timing diagrams shown in this chapter will not include such markings. Also, the timing diagrams shown in Fig. 1.12 are drawn as if the gradients can be turned on and off instantaneously, which is not realistic. In reality, it takes time to ramp up the gradients to full strength, and an equivalent amount of time to ramp them down to zero strength. Magnetic field gradients are characterized by their maximum strength, and the time it takes to ramp them up to maximum strength. On state-of-the-art MRI scanners with high-performance gradients, the maximum gradient strength is about 5 G/cm (50 mT/m), and the amount of time it takes to ramp up the gradients on these systems is about 0.25 ms. Often times the gradient strength is divided by the ramp time to characterize the performance of the gradients in a single number, referred to as the gradient slew rate. For the numbers above, the gradient slew rate would be 200 mT/m/ ms. Stronger and faster gradients facilitate shorter TR and TE times, which is beneficial for MRA methods (as described later in this chapter). For the remaining pulse sequence timing diagrams shown in this chapter, the gradient ramp times will be represented by a sloped line before and after each gradient application. The remainder of this section describes how magnetic field gradients are used to select the magnetization that is going to be tipped into the transverse plane and then encode the signal from this magnetization so that it can be mapped to the proper position in the image.

1

Basic Principles of MRI and MR Angiography

13

Slice-Selection

Fig. 1.14 A spatially selective B1 pulse (right side of figure) is constructed by combining B1 fields that have different frequencies (left side of figure). The range of frequencies used to construct the B1 pulse is referred to as the bandwidth of the B1 pulse. The bandwidth of this pulse is −w3 to +w3, or ±w3. The spatially selective pulse shown here can be used in conjunction with a magnetic field gradient to generate transverse magnetization from spins having Larmor frequencies in the range −w3 to +w3

In order to form an image of a single slice, signal must be generated from only spins that are within that slice. This is accomplished by tipping only the magnetization within the slice of interest into the transverse plane so that it may be detected by the receiver coil. This can be achieved by applying a magnetic field gradient, and at the same time applying a specially tailored RF pulse. With the application of a magnetic field gradient along the z-axis, the Larmor equation (in the rotating frame) becomes w = g Gzz, where g is the gyromagntic ratio for protons, Gz is the strength of the magnetic field gradient, and z is the position along the z-axis. This means that, when the magnetic field gradient is being applied, spins at different positions along the z-axis have different Larmor frequencies. This is demonstrated in Fig. 1.13. By viewing Fig. 1.13, and by recalling that spins are only affected by an RF pulse that is transmitted at a frequency equal to the Larmor frequency of the spins, it can be realized that spins in a particular slice can be selectively tipped into the transverse plane by applying a magnetic field gradient, and at the same time applying an RF pulse that is made up of B1 fields that precess at the same frequencies as the spins that are within the slice that is to be imaged. An RF pulse made up of B1 fields having a range of frequencies will affect only spins that have a Larmor frequency that is within this range. Spins that have a Larmor frequency outside this range will be unaffected by the RF pulse. An RF pulse that is made up of B1 fields precessing within a range of frequencies is shown in Fig. 1.14. The range of

frequencies of the B1 fields making up the RF pulse is referred to as the bandwidth of the RF pulse. The RF pulse will affect only magnetization that has Larmor frequencies that are within the bandwidth of the RF pulse. Such a pulse is often referred to as a spatially selective, or slice-selective, RF pulse. The gradient that is applied to create a distribution in the Larmor frequencies of the spins so that magnetization from the spins within a specific slice can be selectively tipped is referred to as the slice-selection gradient. Changing the strength of the slice-selection gradient causes a change in the thickness of the slice that is selected (if the bandwidth of the RF pulse remains unchanged). Increasing the gradient strength causes spins in a thinner slice to be affected by the RF pulse, and decreasing the gradient strength causes spins in a thicker slice to be affected by the RF pulse. This is because increasing the gradient strength causes magnetization with Larmor frequencies that are within the bandwidth of the RF pulse to span a smaller distance, and decreasing the gradient strength causes magnetization with Larmor frequencies that are within the bandwidth of the RF pulse to span a larger distance. This phenomenon is demonstrated in Fig. 1.15. In Fig. 1.15, the bandwidth of the RF pulse includes frequencies from −625 Hz to +625 Hz (written as ±625 Hz), so it affects only spins that have Larmor frequencies within this range. If only a single slice is to be imaged, it is desirable to place that slice at the isocenter of the gradient. This is because the isocenter of the gradient is also the center of the magnet

Fig. 1.13 When a magnetic field gradient is applied along the z-axis, spins at different locations in the z dimension have different Larmor frequencies. The frequency as a function of position can be determined using the Larmor equation, w = g Gzz, where Gz denotes the strength of the magnetic field gradient, and z denotes the location along the z-axis

14

Fig. 1.15 The amplitude of the slice-selection gradient determines the thickness of the slice that is imaged. (a) A strong gradient causes the magnetic field strength to change rapidly with position, so when a strong gradient is applied, spins with Larmor frequencies in the range affected by the B1 pulse are clustered close together. (b) A weak gradient causes the magnetic field strength to change slowly with position, so when a weak gradient is applied, spins with Larmor frequencies in the range affected by the B1 pulse are spread out over a large distance

and the main magnetic field is most uniform at this point. So, if a single slice is to be imaged, the patient table is moved so that the anatomy that is to be imaged moves to the isocenter of the magnet. All of the discussion above pertained to imaging a slice at isocenter. If two slices are to be imaged, they cannot both be placed at the isocenter of the gradient so there is a mechanism to image slices off of isocenter. Imaging slices off of isocenter is accomplished by changing the frequencies of all of the B1 fields that make up the RF pulse. The bandwidth remains the same, but the frequencies of all the B1 fields that make up the RF pulse are offset to correspond to the position of the slice. A pulse sequence timing diagram demonstrating a sliceselective RF pulse and a slice-selection gradient is shown in Fig. 1.16. Here it is demonstrated that the gradient and RF pulse are applied simultaneously. The duration of the RF pulse is a function of its bandwidth. The duration of the gradient is equal to the duration of the RF pulse; the gradient must be on during the entire application of the RF pulse in order to maintain the distribution of spin frequencies. No other gradients may be on during application of the sliceselection gradient. The strength of the gradient is a function of the slice thickness, as explained above, and the strength of the RF pulse is a function of the tip angle. The greater the amplitude of the RF pulse, the farther the spins tip in the time that the RF pulse is applied. The negative lobe of the slice-selection gradient is used to refocus the spins. The spins along the slice-selection direction get out of phase with each other because the application of the slice-selection gradient causes the spins along this direction to precess at different frequencies while they are being tipped into the

F.R. Korosec

Fig. 1.16 A pulse sequence timing diagram showing a slice-selective RF (or B1) pulse, a slice-selection gradient, a phase-encoding gradient, a frequency-encoding (or readout) gradient, and the signal detected during activation of the receiver

transverse plane. Reversing the gradient for the appropriate amount of time causes the spins to come back into phase with each other.

Frequency-Encoding Now that a slice has been selected by tipping only the spins within that slice into the transverse plane, the signals from these spins must be encoded so that when they are detected, it can be determined from where they came, so that they can be mapped to the appropriate positions in the image. Differentiating the signal along one dimension of an object can be accomplished by applying a magnetic field gradient along that direction. Just as a slice-selection gradient applied along the z dimension causes spins at different locations along the z dimension to have different Larmor frequencies, a spatial-encoding gradient applied along the x dimension causes spins at different locations along the x dimension to have different Larmor frequencies. Thus, when the signals from the spins in the image slice are detected during the application of a magnetic field gradient, the frequencies of the signals can be used to map them to the proper location along this one dimension in the image. Because this gradient encodes position by giving spins at different positions different frequencies, it is referred to as the frequency-encoding gradient. It also often is referred to as the readout gradient because it is applied at the same time that the receiver coil is “reading out” the signal. An example of the effect that a frequency-encoding gradient has on the spins in an image slice is shown in Fig. 1.17. To understand more clearly how the frequency-encoding gradient allows the position of the spins to be determined,

1

Basic Principles of MRI and MR Angiography

consider the case of an object placed in the MR scanner as shown in Fig. 1.18. While the frequency-encoding gradient is being applied, the spins in the object precess at frequencies

15

Fig. 1.17 Application of a frequency-encoding gradient along the x dimension causes spins at different locations in the x dimension to have different Larmor frequencies. The signals can be mapped to the proper position in the image based on their precessional frequencies

that are dictated by where the spins are located along the magnetic field gradient. If the receiver coil is turned on while the frequency-encoding gradient is being applied, the detected signal will be a composite of all of the signals at all of the precessional frequencies of the spins in the object. When the signal is sufficiently sampled, the frequencyencoding gradient and the receiver coil are turned off and the signal is amplified, digitized, and sent to the signal processor. The signal processor performs a fast Fourier transform (FFT) on the sampled signal to determine the frequencies of all the signals composing the detected composite signal, and the relative number of spins precessing at each of the frequencies (signal intensity at each frequency). The output of the fast Fourier transform can be mapped to the columns of pixels (picture elements) in an image as shown in Fig. 1.18. So, the frequency-encoding gradient allows the relative amount of signal in each column of the image to be determined. To form an image, the spatial distribution of the signals in each column needs to be determined. This is the topic of the next section. Before this is addressed, a few aspects of frequency-encoding are discussed. The strength of the frequency-encoding gradient is chosen such that spins at one edge of the field-of-view (FOV) precess at a predetermined maximum frequency (added to the 63.87 MHz resulting from B0), and spins at the opposite edge of the FOV precess at the same predetermined maximum frequency but in the opposite direction (subtracted

Fig. 1.18 (Upper left) The sample, denoted by black squares in this example, extends over three of the six columns of the image. (Lower left) The detected signal contains contributions from all of the spins in the object, all precessing at frequencies determined by their location along the applied gradient. In this example, because the object extends over three of the six columns, the detected signal is composed of

signals having three different precessional frequencies, −w1, w1, and w2. (Lower right) Applying a fast Fourier transform to the detected signal reveals the frequencies that are contained in the signal, and the relative number of spins at each frequency. (Upper right) This information is used to map the signals from the spins to the appropriate column in the image

16

from the 63.87 MHz resulting from B0). The receiver is set by the operator to accept only frequencies within this predetermined range. This range of frequencies is referred to as the receiver bandwidth. A typical receiver bandwidth is ±16 kHz, which means that spins at one edge of the FOV precess at +16 kHz (faster than 63.87 MHz), and spins at the opposite edge of the FOV precess at −16 kHz (slower than 63.87 MHz). By keeping the receiver bandwidth the same and changing the strength of the gradient, the FOV in the frequencyencoding direction can be changed. Increasing the gradient strength decreases the FOV by causing spins that have precessional frequencies within the receiver bandwidth to be contained in a smaller region, and decreasing the gradient strength increases the FOV by causing spins that have precessional frequencies within the receiver bandwidth to extend over a larger region. The field of view is not restricted to the center of the magnet. Off-center FOVs are achieved by changing the frequencies that are accepted by the receiver. The receiver bandwidth is kept the same, but each of the frequencies that is accepted is changed by some amount depending on how far off of isocenter the FOV is shifted. A pulse sequence timing diagram, showing the timing of the frequency-encoding gradient and receiver activation relative to the timing of the slice-selection process is shown in Fig. 1.16. No other gradients may be applied during application of the frequencyencoding gradient.

Phase-Encoding The slice-selection gradient applied simultaneously with a tailored RF pulse allows the spins in a particular slice to be tipped into the transverse plane to generate signals, and the application of the frequency-encoding gradient permits determination of where the signals originate from along one dimension. To form an image, all that remains to be determined is from where along the second dimension in the image plane the signals originate. Encoding along the second image dimension is achieved by applying another gradient, this time along the y-axis. This gradient, because it gives signals at different positions and different phases, is referred to as the phase-encoding gradient. The phase of the magnetization is simply a measure of how far it precesses in the transverse plane during a given period of time. Phase is an angle and it is typically measured in degrees. If the magnetization in the transverse plane precesses one full revolution, it is said to have accumulated a phase of 360°. The phase-encoding gradient is applied after the magnetization is tipped into the transverse plane, and before it is read out; that is, it is applied after the slice-selection gradient is

F.R. Korosec

Fig. 1.19 (a) After the slice-selection process, the spins in the selected slice lie in the transverse plane, and are all oriented in the same direction (in this case, they all lie along the x-axis). (b) When a gradient, Gy, is applied along the y dimension, it causes spins at different locations along this dimension to precess at different frequencies. By the time the gradient is turned off, spins at different locations along the y dimension have accumulated different amounts of phase. This phase-encoding process is used to encode signals along this dimension with a unique phase so that when they are detected, they can be mapped to the proper position in the phase-encoding dimension in the image

applied and before the frequency-encoding gradient is applied. While the phase-encoding gradient is being applied, the spins at different locations along the y-axis precess at different rates. Thus, by the time the phase-encoding gradient is turned off, the spins at different locations along the y-axis will have precessed different amounts and thus will have accumulated different amounts of phase. This is shown schematically in Fig. 1.19. After the phase-encoding gradient has been turned off, the frequency-encoding gradient is turned on, and the signal from the selected slice is detected. In order to resolve all points along the y dimension, additional information is needed. The way this information is obtained is by applying additional phase-encoding gradients. After the application of each phase-encoding gradient, a frequency-encoding gradient is applied, and the signal is sampled again with the new phase-encoding value. The application of each phase-encoding gradient requires a separate TR. In each TR, a phase-encoding gradient of slightly greater amplitude is applied. This causes the spins to accumulate a slightly larger phase in each TR. The amplitude of the phaseencoding gradient is increased with each TR because, for any single application of the phase-encoding gradient, components of the magnetization from spins at different locations point in opposite directions causing the signals from some spins to partially cancel the signals from other spins. In order to appropriately determine the signal in each pixel along the phase-encoding dimension, the total number of phaseencoding gradients that must be applied is equal to the total number of rows of pixels that are to be resolved in the phaseencoding dimension (it is also true that the frequencyencoded signal needs to be sampled one time for each column

1

Basic Principles of MRI and MR Angiography

17

of pixels that is to be resolved in the frequency-encoding direction). An important difference between phase- and frequency-encoding is that increasing the number of rows of pixels in the phase-encoding direction results in an increase in scan time, whereas increasing the number of columns of pixels in the frequency-encoding direction does not affect the scan time (except in fast gradient-echo acquisitions, where increasing the number of frequency-encoding values necessitates an increase in the TR of the imaging sequence). It should be noted that for consistency of terminology in this chapter, frequency-encoding is being associated with rows of pixels in the image and phase-encoding is being associated with columns of pixels in the image, but the phaseand frequency-encoding dimensions can easily be swapped in practice. If it is desirable to have greater resolution along a particular dimension, the scanner operator would align the frequency-encoding along this dimension since aligning the phase-encoding along this dimension would necessitate a longer scan time. Just as the FOV in the frequency-encoding dimension can be adjusted by modifying the strength of the frequencyencoding gradient, the field of view in the phase-encoding dimension can be adjusted by modifying the strength of the phase-encoding gradients. Also, the center of the field of view can be offset in the phase-encoding dimension, just as it can be offset in the frequency-encoding dimension. Figure 1.16 shows the timing of the phase-encoding gradient with respect to the slice-selection gradient, the frequency-encoding gradient, the RF pulse, and the activation of the receiver. The phase-encoding gradient is shown with many amplitudes, representing the need to apply one phaseencoding amplitude for each row of pixels that is to be resolved in the phase-encoding dimension. It should be remembered that each phase-encoding value is applied in a separate TR. All other waveforms remain the same for all TR intervals. The scan time necessary to acquire the data to produce an image of a single slice in MRI is given by the following equation: Scan Time = TR ´ N y ´ Ave,

(1.4)

where TR is the repetition time of the imaging sequence, Ny is the number of phase-encoding values applied (one per TR), and Ave is the number of times each phase-encoding value is sampled for purposes of averaging the data. For certain imaging sequences used in MRI, the TR is chosen to achieve the desired image contrast (e.g., T1-weighting, T2-weighting, density-weighting). If the TR is much longer than the time necessary to acquire a single phase-encoded line of data from a given slice, then the data acquisition can be interleaved so that data from multiple slices may be obtained in a TR interval (i.e., excite slice 1, sample phase-encoding 1; excite

slice 2, sample phase-encoding 1; excite slice 3, sample phase-encoding 1; – continue until the TR interval has elapsed, then – excite slice 1, sample phase-encoding 2; excite slice 2, sample phase-encoding 2; excite slice 3, sample phaseencoding 2; – and repeat until all phase-encoding values have been sampled.) For this interleaved mode of acquisition, the scan time for imaging multiple slices is the same as that required for imaging a single slice, as long as the data for a given phase-encoding value for all slices can be acquired before the TR time elapses. If data are acquired from individual slices sequentially, such that all data for a given slice are acquired prior to collecting data for the next slice, then the scan time is increased by a factor equal to the number of slices imaged. This is true for the types of sequences most commonly used for MR angiography (i.e. fast gradient-echo sequences). The above methods, where individual slices are excited one at a time (in an interleaved fashion, or individually), and magnetization from only one slice is in the transverse plane at any given time, are referred to as 2D multislice imaging methods. If data are collected from contiguous slices, data are collected from a 3D volume, but the acquisition methods are still referred to as two-dimensional, multislice imaging methods. MR imaging may be performed using three-dimensional acquisition methods. With 3D acquisition methods, a component of the magnetization from a volume, or slab, containing all the slices to be imaged is tipped into the transverse plane in every TR, and the information is mapped to individual slices using a phase-encoding method. For 3D acquisitions, the pulse sequence timing diagram shown in Fig. 1.16 would be modified to include a phase-encoding gradient on the z-axis (sometimes referred to as slice-encoding, or slabencoding gradient). The phase-encoding gradient on the z-axis can be applied anytime after the application of the slice-selection gradient and before the application of the frequency-encoding gradient. With 3D acquisitions, data for every slice are sampled in every TR throughout the acquisition, so interleaved acquisition is not possible. For 3D acquisition methods, the scan time equation above must be multiplied by the number of slices imaged.

MR Data (k-Space) A representation of the matrix of data points acquired in an MRI scan is shown in Fig. 1.20. In MRI, analog data are sampled and converted to digital numbers. Each number (sampled point) is represented as a dot in this figure. Each dot along a row represents a different frequency-encoded sample (the number of columns of dots is equal to the number of columns of pixels that may be represented in the frequency-encoding dimension of the image). Each dot along a

18

column represents a different phase-encoded sample (the number of rows of dots is equal to the number of rows of pixels that may be represented in the phase-encoding dimension of the image). More frequency-encoding values (or more phase-encoding values) corresponds to more rows (or columns) of pixels across a FOV, leading to improved spatial resolution in an image. Actual data acquired during an MRI scan are shown in Fig. 1.21. In this figure, the dots have been replaced with a representation of the actual digital samples (the numeric values). Large numbers are represented with high-intensity signals, and small numbers are represented with low-intensity signals. In MRI, these numbers that are sampled correspond to Fourier weighting coefficients. A full discussion on Fourier

Fig. 1.20 In MRI, data are acquired at discrete points in what is referred to as k-space. The dots in this figure represent the locations of the acquired data points. The resolution in MR images is determined by how far from the origin data are sampled ( ± k x ,max, ± k y ,max ). The size of the unaliased field-of-view (FOV) in MR images is determined by the spacing between sampled points ( Dk x , Dk y )

Fig. 1.21 (a) MRI data. The signal intensity represents the magnitude of the Fourier weighting coefficients. The characteristics of the frequencyencoding gradient (in combination with the receiver bandwidth) determine which points are sampled along each row, and the characteristics of the phase-encoding gradient determine which points are sampled along each column. (b) An axial image of a human brain obtained after applying a two-dimensional fast Fourier transform (2D FFT) to the acquired data

F.R. Korosec

theory is beyond the scope of this chapter. However, a few concepts are introduced. According to Fourier theory, images can be formed by appropriately weighting and combing line-pair patterns (more specifically, cosine and sine function patterns). Fourier coefficients are the weighting values that must be applied to the line-pair patterns to produce the desired image. For twodimensional images, the Fourier coefficients can be arranged in a 2D matrix, and this matrix of coefficients may be displayed in what is referred to as k-space. The location of each point in k-space is associated with a line-pair pattern in image space. Locations along the kx-axis correspond to vertical line-pair patterns, locations along the ky-axis correspond to horizontal line-pair patterns, and locations along the diagonal axes correspond to diagonal line-pair patters. Locations near the center of k-space correspond to few line-pairs per image, whereas locations farther from the center of k-space correspond to more line-pairs per image. To fully represent an object in an image, an infinite number of line-pair patterns (and associated weighting coefficients) would be required. In MRI, only a subset of the Fourier coefficients are sampled, due to scan time considerations. The limited number of Fourier coefficients is confined between –kx,max and + kx,max in the frequency-encoding dimension, and –ky,max and +ky,max in the phase-encoding dimension. These Fourier coefficients, therefore, define the maximum spatial resolution in the image (highest density line-pair patterns in the x and y dimensions). The Fourier coefficients beyond these values are not sampled. The spacing between sampled Fourier coefficients defines the size of the FOV in MRI. The larger the spacing between the sampled k-space points, the smaller the FOV. This is demonstrated in Fig. 1.22. Another thing to note about k-space is that, in MRI, the imaging gradients are used to determine which k-space

1

Basic Principles of MRI and MR Angiography

19

Fig. 1.22 As the distance between k-space samples increases (in this case by acquiring fewer phase-encoded lines of data), the size of the field-of-view (FOV) decreases. When there is anatomy outside of the FOV, it “wraps around” and overlaps onto the anatomy within the FOV. This mismapping of the signal that originates from outside the FOV is referred to as spatial aliasing

points are sampled. The area (amplitude × duration) of the frequency-encoding gradient determines which point along the kx direction is sampled during data acquisition, and the area of the phase-encoding gradient determines which point along the ky direction is sampled. Negative gradient areas move through k-space along the negative kx and ky axes, and positive gradient areas move through k-space along the positive kx and ky axes. Once the FOV in the phase-encoding dimension is selected, Dky is determined, and the changes in the area of the phase-encoding gradient from one TR to the next can be calculated. Likewise, once the FOV in the frequency-encoding dimension is selected, Dk x is determined, and the changes in the area of the frequency-encoding gradient from one sample to the next can be calculated (in conjunction with the sampling interval of the receiver, which is dictated by the receiver bandwidth). Once the number of phase-encoding values ( N y ) and frequency-encoding values ( N x ) is determined, the number of k-space points can be determined, the kmax points can be determined Ny ö Nx æ çè ± k x ,max = ± 2 ´ D k x and ± ky ,max = ± 2 ´ Dky ÷ , and the ø duration of the frequency-encoding gradient, and maximum area of the phase-encoding gradient can be determined. By strategically applying combinations of magnetic field gradients, k-space may be sampled using a variety of trajectories. Instead of traversing k-space on a rectilinear grid [6] (typically referred to as Cartesian sampling), k-space can be sampled along 2D radial lines [7], where each radial line passes through the center of the k-space, and the radial lines are arranged with equal angles between them (like spokes on a wheel). Alternatively, k-space may be sampled along spiral trajectories [8], starting from the center and spiraling

outward. Other sampling trajectories exist as well, such as 3D radial [9], stack of spirals [10], 3D cones [11], as well as numerous others. The density of phase-encoding lines and readout samples required are determined by the Nyquist sampling criterion and can be described by the following equations, Dt £

2π , g Gx FOVx

(1.5)

2π , g TPE FOVy

(1.6)

DGy £

where Dt is the readout sampling interval (determined by the receiver bandwidth), Gx is the frequency-encoding gradient amplitude, FOVx and FOVy are field of view sizes in the x and y dimensions, DGy is the phase-encoding gradient amplitude step size, TPE is the duration of each phase-encoding gradient, and it is assumed that the imaging gradients can be turned on and off instantaneously (i.e., ramp time equals zero). If the requirements defined by these equations are not met, the reconstructed images will have aliasing artifacts, where copies of images will be superimposed as shown in Fig. 1.22. Finally, it has been described that points near the center of k-space represent low spatial frequency information (few line-pairs per image), and points at the periphery of k-space represent high spatial frequency information (many linepairs per image). This characteristic is often generalized by saying that the center of k-space contains the contrast information for the image, and the periphery of k-space contains the detail information for the image (information regarding

20

F.R. Korosec

Fig. 1.23 The images in the bottom row were produced by performing a fast Fourier transform (FFT) on the k-space data shown in the top row. Each image is shown directly below the k-space data used to produce it. These images demonstrate that the data at the center of k-space is responsible for representing the contrast in the image, whereas the data at the periphery of k-space is responsible for representing the detail in the image

the edges of objects). This concept is demonstrated in Fig. 1.23, which shows the image content derived from the center of k-space in one image, and the content derived from the periphery of k-space in a separate image.

Signal-to-Noise Ratio One parameter commonly used to characterize image quality is the signal-to-noise ratio (SNR). This is the ratio of the amplitude of the signal (desired information) to the amplitude of the noise (undesired information), and is expressed as Signal/Noise. Generally speaking, a higher SNR indicates better the image quality. In MRI, the SNR depends on a large number of factors, including the field strength of the MRI scanner, the imaging coil used, the temperature of the object being imaged, the imaging sequence selected (i.e., spin-echo, gradient-echo, fast spin-echo, etc.), the imaging parameters that are used to manipulate image contrast (i.e., TR, TE, TI, tip angle, etc.), and a host of other parameters. Once the exam is setup, and the imaging parameters are selected to produce the appropriate contrast in the image, there are still a number of parameters that may be changed for a variety of reasons (modify anatomic coverage, resolution, scan time, etc.). These parameters affect the SNR according to the following equation [12]: SNR μ voxel volume ´ acquisition time.

(1.7)

The variable acquisition time is the time that the receiver is turned on and data are being acquired, and a voxel (volume element) is the volume of the object represented by a pixel in the image. The proportionality symbol appears because the SNR is dependent on the scan setup, as well as the scan parameters responsible for image contrast as described above. The SNR is dependent on the voxel volume since larger voxel volumes result in more spins contributing to the signal from each voxel. The SNR also is dependent on the time spent acquiring data. When more data are acquired, the signal increases coherently, whereas the noise (because it is random) increases incoherently (sometimes adding constructively and sometimes destructively). The result is that the ratio of the signal and noise increase as the square root of the amount of data acquired. The voxel volume can be calculated as FOVx / N x ´ FOVy / N y ´ sl thick, where FOVx and FOVy are the fields of view in the frequency-encoding and phase-encoding directions, respectively, N x and N y are the number of frequencyencoding and phase-encoding values, respectively, and sl thick is the slice thickness. The amount of time spent acquiring data is equal to the time spent acquiring each frequency-encoding value (Dtx), multiplied by the number of frequency-encoding points sampled Nx (Dtx × Nx = duration of readout), multiplied by the number of phase-encoding values Ny (since the each frequency-encoding value is sampled at each phase-encoding value), multiplied by the number of times every data point is resampled for purposes of averaging,

1

Basic Principles of MRI and MR Angiography

21

Ave. The time spent acquiring each frequency-encoding point is equal to the inverse of the sampling rate (sampling rate = receiver bandwidth = RBW, so Dtx = 1/RBW). So, the total time spent acquiring data is equal to ((Nx × Ny × Ave)/ RBW). The equation for SNR becomes: SNR 2D μ

N x ´ N y ´ Ave (1.8) FOVx FOVy ´ ´ sl thick ´ , Nx Ny RBW

where, FOVx and FOVy are the fields of view in the frequency-encoding and phase-encoding directions, respectively, N x and N y are the number of frequency-encoding and phase-encoding values, respectively, slthick is the slice thickness, Ave is the number of signal averages, and RBW is the receiver bandwidth. For 3D acquisitions, data for every slice are sampled in every TR throughout the acquisition. This increases the SNR of every slice by the square root of the number of slices imaged. So, for 3D acquisitions, the SNR equation is as follows:

allow for high scan time reduction factors. Parallel imaging can be used to obtain higher spatial resolution or coverage in the same scan time, or to obtain the same resolution and coverage in less scan time, or various combinations of the above. Reducing the scan time is particularly useful when scans must be completed during a breath-hold interval or when temporal resolution is desired.

Magnetic Resonance Angiography The most widely used MRA techniques can be categorized as contrast-enhanced, TOF, phase-contrast, or steady-state MRA techniques. In the remainder of this chapter, a brief overview of the physical principles of these techniques is provided, examples of clinical applications are shown, and the benefits and limitations of each of the specific methods are briefly discussed.

Contrast-Enhanced MRA SNR 3D = SNR 2D ´ N z ,

(1.9)

where N z is the number of phase-encoding values in the slab thickness direction (the number of slices encoded).

Parallel Imaging When multielement receiver coils are used, scan times may be reduced using a class of methods referred to as parallel imaging methods. SMASH [13] (simultaneous acquisition of spatial harmonics), SENSE [14, 15] (sensitivity encoding), and GRAPPA [16] (generalized autocalibrating partially parallel acquisitions) are examples of parallel imaging methods. These methods rely on the sensitivity of the different coil elements to determine information about the spatial location of the detected signals, thereby reducing the number of phaseencoding values necessary for image formation. With these methods, the scan time reduction factor, R, is dependent on the number of elements contained in a coil. Coils with more elements permit greater R values. If coil elements are distributed in more than one dimension, then the amount of data in both phase-encoding dimensions in 3D acquisitions can be reduced. Reducing the amount of data acquired results in a reduced SNR [see (1.8)]. Parallel imaging methods suffer from an additional reduction in SNR-based on coil geometry. This additional reduction in SNR is characterized by what is called a g-factor. With current coil configurations and SNR limitations, the scan time reduction factors afforded by parallel imaging methods are on the order of 4–6. Parallel imaging methods are particularly effective when used with 3D CE-MRA where the SNR is sufficiently high to

Contrast-enhanced MRA (CE-MRA) methods have gained widespread acceptance due to their ease of use, and their ability to quickly, reliably, and robustly produce high-quality diagnostic images of large vascular territories. In many clinical practices, CE-MRA has replaced X-ray DSA as the method of choice for imaging certain vasculature, including the carotid and vertebral arteries, the aorta and renal arteries, and the vessels of the lower extremities. Contrast-enhanced MRA techniques [17] achieve signal differences between blood and stationary tissues by manipulating the magnitude of the magnetization, such that the magnitude of the magnetization from moving blood is larger than the magnitude of the magnetization from stationary tissues. Manipulating the magnetization to produce signal differences in CE-MRA techniques is achieved not only by employing the appropriate imaging sequence parameters, but also by injecting a contrast material intravenously to selectively shorten the T1 of the blood. By implementing a T1weighted imaging method, appropriately synchronized to acquire data during the first pass of the contrast material through the arteries of interest, images can be acquired that show arteries with striking contrast relative to surrounding stationary tissues and veins. A 3D CE-MRA image of the arteries in the lower extremities is shown in Fig. 1.24.

Contrast Material Currently, the contrast materials most widely used for CE-MRA are gadolinium-based. The gadolinium atom is highly paramagnetic. It has seven unpaired electrons in its

22

F.R. Korosec

Fig. 1.24 A coronal MIP image of the vessels of the thighs of a healthy volunteer acquired using a coronal 3D contrast-enhanced MRA acquisition method applied after intravenous injection of a gadolinium-based contrast material

outer shell that interact with the hydrogen nuclei on the water molecules. This interaction results in an increased regrowth rate of longitudinal magnetization from these hydrogen nuclei (reduced T1), so they appear bright in images acquired with T1-weighted MR acquisition methods. The T1-shortening induced by the gadolinium-based contrast material is described by the following equation: 1 1 = + R1 ´ [Gd], T1 T10

(1.10)

where T1 is the shortened longitudinal relaxation time of blood in the presence of gadolinium, T10 is the longitudinal relaxation time of blood in the absence of gadolinium, R1 is the relaxivity of the gadolinium-based contrast material, and [Gd] is the concentration of the gadolinium-based contrast material. Since gadolinium atoms are highly toxic, they must be chelated before they can be injected into the blood vessels. Currently there are several different chelates widely used in CE-MRA, which include gadobenate dimeglumine (Multihance), gadopentetate dimeglumine (Magnevist), gadodiamide (Omniscan), gadoversetamide (OptiMARK), and gadoteridol (ProHance). These materials are permeable through the blood vessel walls, resulting in enhancement of signals from tissues over time, as the contrast material makes its way through the arteries, into the veins, and eventually into the stationary tissues. These chelates are referred to as extracellular contrast materials. Contrast materials that are not permeable through vessel walls, and remain in the blood pool for more than an hour without leaking into the surrounding stationary tissues, have been developed (e.g., gadofosveset trisodium: Ablavar) [18, 19]. These so-called intravascular, or blood pool, contrast materials

can be used to increase the imaging time in order to achieve greater coverage and higher spatial resolution. The drawback of increased acquisition time is that it results in venous enhancement, which leads to difficulty in evaluating the arteries. Methods are being developed to separate arterial signal from venous signal, but are not yet available clinically. An additional benefit of some of the intravascular contrast materials is that they have a higher relaxivity. In other words, they provide greater vascular signal during the first pass of the contrast material by causing a more dramatic decrease in the T1 of blood (per unit volume of contrast material) compared to the currently available extravascular contrast materials. Exposure to gadolinium-based contrast materials has been associated with the development of nephrogenic systemic fibrosis (NSF) [20] in patients with compromised renal function who also have other confounding health issues. NSF is an irreversible, often fatal, disease. Screening measures are now in place to identify patients at risk of contracting this disease. Since awareness has been heightened, and screening measures have been implemented to identify high-risk patients, the incidence of this disease has been nearly eliminated. The gadolinium-based contrast material is typically injected intravenously through an 18–22-gauge angiocatheter. The angiocather is typically inserted into a large, easily accessible vein, such as one of the veins in the antecubital fossa. However, other injections sites may be selected as well, including the forearm, wrist, or the back of the hand. Injection into the right arm provides the most direct route to the heart, increasing the likelihood of distributing a tight, highly concentrated bolus of contrast material to the vessels of interest. The size of the angiocatheter is dependent on the size of the vessel and the desired injection rate. Injection rates typically range from 0.5 to 4.0 ml/s for CE-MRA. Injection volumes range from 0.1 to 0.3 mmol of contrast material per kg of patient body weight (mmol/kg), with typical values in the range of 20–40 ml of contrast material. Higher volumes usually provide images with higher SNR in which the small vessel detail is better delineated. The intent is to reduce the T1 of blood to as low as 50 ms during the first pass of the contrast material. The injection of the contrast material is immediately followed by a flush of normal saline. The injection may be performed manually, or using an MR-compatible, computer-controlled injector as shown in Fig. 1.25. With a computer-controlled injector, the volume and injection rate are precisely programmed to ensure reproducible results. With manual injection, the clinician controls the injection and can better monitor the status of the patient.

Imaging Sequence CE-MRA typically is performed using a three-dimensional, RF-spoiled, fast gradient-echo imaging sequence [21]. The

1

Basic Principles of MRI and MR Angiography

23

Fig. 1.25 Typical patient setup for a 3D CE-MRA acquisition of the renal arteries. An MRI-compatible computer-controlled injector is used to administer the gadolinium-based contrast material into a vein in the antecubital fossa. One syringe is used to inject the contrast material, and the second syringe is used to inject a saline flush. The anterior and posterior sections of an 8-channel torso coil (square items with Velcro straps) are placed on top of and underneath, respectively, the patient to image the region of interest

pulse sequence timing diagram looks like the one shown in Fig. 1.16, with the addition of a phase-encoding gradient on the slice-selection axis. Also RF spoiling is used to eliminate, or spoil, the transverse magnetization so it does not contribute to signal in subsequent TR intervals. With RF spoiling, the RF pulse is applied such that the longitudinal magnetization is tipped into a different location in the transverse plane (relative to the y-axis) for each TR. Eliminating the coherence of the transverse magnetization so it does not persist from one TR to the next makes RF-spoiled sequences less T2*-weighted, and therefore more strictly T1-weighted (as long as TE Ⰶ T2*). Imaging occurs after the longitudinal magnetization has reached its steady-state equilibrium value (a state where the signal remains constant from one TR to the next since the amount of longitudinal magnetization tipped away from the longitudinal axis by the RF pulse equals the amount of magnetization that regrows during each TR interval). RF spoiled gradient-echo imaging sequences are given different names by different MR vendors, including SPGR (spoiled gradient recalled echo), FLASH (fast low angle shot), and T1 FFE (T1 fast field echo). The steady-state equilibrium magnetization for an RF-spoiled gradient-echo sequence (normalized for protondensity and assuming negligible T2* decay and perfect spoiling) is given by the following equation [22]: M xy (q ,TE ) =

M 0 sin (q )(1 - E1 ) - TE/T2* e , 1 - E1 cos (q )

(1.11)

Fig. 1.26 Calculated steady-state signal as a function of tip angle for tissues with various T1 values using a spoiled gradient-echo imaging sequence with a TR of 5 ms. Note that the signal initially increases with increasing tip angle, but then decreases. See text for details. Note also that the signal increases with decreasing T1 value. For a given tip angle, the relative signal difference (contrast) between tissues can be determined. For 3D CE-MRA, it is desirable to maximize the signal different between blood and stationary tissues

where E1 is defined to be exp( - TR/T1 ) , M 0 is the thermal equilibrium magnetization value, q is the tip angle of the imaging sequence, TR is the repetition time of the imaging sequence, TE is the echo time of the imaging sequence, T2* is the time constant characterizing the decay rate of the transverse magnetization of the tissue under consideration, and T1 is the time constant characterizing the regrowth rate of the longitudinal magnetization of the tissue under consideration. The signal as a function of tip angle and T1 for an RF-spoiled gradient-echo sequence with a TR of 5 ms and TE  T2* is plotted in Fig. 1.26. It can be seen that the signal increases as the T1 value decreases. In CE-MRA, the injection of a gadolinium-based contrast material shortens the T1 of blood such that it gives rise to a large signal. All other tissues with longer T1 values give rise to lower signals. The graph also shows that the signal initially increases with tip angle as more magnetization is tipped into the transverse plane. However, as the tip angle is further increased, a point is reached where the tip angle is sufficiently large that it depletes the steady-state longitudinal magnetization that remains after tipping (and it is the longitudinal magnetization that contributes to the transverse magnetization in subsequent TR intervals), so the resulting signal gets smaller with increasing tip angles. The tip angle at which the largest signal is obtained is referred to at the Ernst angle. Rather than striving to achieve the largest signal from blood in CE-MRA, it is more desirable to maximize the image contrast (greatest signal difference between blood and other tissues).

24

Thus, tip angles greater than the Ernst angle are typically used when prescribing CE-MRA sequences. Prescribing an RF-spoiled gradient-echo sequence with a short TR and a large tip angle will yield signal differences based on differences in T1 relaxation times. As long as the scan is synchronized so that data are acquired when the contrast material is in the arteries of interest, the resulting T1weighted images will show the short-T1 arterial blood with a high signal, and the long-T1 veins and stationary tissues with diminished signal. In order to acquire the majority of the data for the contrast-enhanced angiogram during the first pass of the contrast material (when the concentration of the contrast material in the arteries is high, and the concentration in the veins and stationary tissues is low) it is necessary to use a short scan time. When imaging vessels in the abdomen, where it is necessary to acquire all the data during a breath-hold interval, it also is necessary to use a short scan time. To achieve short scan times in CE-MRA, a short TR is used, and a reduced data set is sampled. The TR may be reduced by using a specialized fast gradient-echo sequence, a high receiver bandwidth, the minimum echo time, a fractional-echo readout, a reduced number of frequency-encoding samples, and highperformance magnetic field gradients. A reduction in the data set may be achieved by employing a fractional FOV, a reduced number of phase-encoding values, and a reduced number of slices. Reducing the number of frequency-encoding values, phase-encoding values, or slices, for a specified volume of coverage, results in reduced spatial resolution in the MRA images. Designing CE-MRA imaging protocols requires compromises, and the best protocols are those that achieve the appropriate balance between scan time, anatomic coverage, and spatial resolution. The apparent spatial resolution may be increased by using zero filling [23, 24] to increase the number of pixels and/or slices in the reconstructed data set.

Scan Synchronization As has been described, in order to ensure the best possible image quality with CE-MRA, it is essential to properly synchronize data acquisition with the arrival and passage of the contrast material. If acquisition is completed before the arrival of the contrast material, the blood will not generate enough signal to appear in the angiogram. If the contrast material arrives during the scan – after data acquisition has begun, but before it is completed – the SNR of the arteries may be low, or artifacts may be present in the images [25]. The artifacts may manifest as a less intense area in the center of the vessels, ringing or replication of the vessel edges, or demonstration of only the edges of the vessels. The appearance of the artifacts depends on the size of the vessels

F.R. Korosec

affected, at what time during the acquisition the contrast material arrives in the vessels being imaged, the order in which the data are acquired, and the rate at which the concentration of the contrast material changes during the scan. If the data are acquired too late, the arterial signal will be diminished, and the veins and stationary tissues will be enhanced. Several methods have been developed to ensure proper timing of the acquisition relative to the passage of the contrast material. In one method, a timing scan is performed prior to acquiring the MR angiogram. For this timing scan, a small bolus of contrast material (1–2 ml) is injected, and then two-dimensional images are rapidly and repeatedly acquired [26]. From these images the arrival time of the contrast material can be determined and used to calculate when to start the acquisition of the three-dimensional angiogram after injecting the full bolus. In other methods, the signal in a volume [27] or an image [28] is monitored and acquisition of the angiogram begins when it has been determined that the contrast material has arrived in the vessels of interest. Finally, time-resolved methods have been developed that repeatedly acquire two-dimensional [29, 30] or three-dimensional [31–34] angiograms continuously during the passage of the contrast material, obviating the need to prospectively determine when the contrast has arrived. In addition to demonstrating the peak arterial frame, time-resolved methods provide information regarding the temporal characteristics of the passage of the bolus of contrast material. Images acquired using a commercially available time-resolved, CE-MRA technique known as TRICKs (Time-Resolved Imaging of Contrast Kinetics) are shown in Fig. 1.27.

Data Acquisition Order A current challenge of CE-MRA is obtaining high spatial resolution in the short amount of time available between arterial and venous enhancement. To address this issue, a method has been developed in which the data acquisition order is modified to acquire the low spatial frequency data early, during enhancement of the arteries, and the high spatial frequency data later, after enhancement of the veins and stationary tissues. This data acquisition order is referred to as elliptical centric phase-encoding order [35, 36]. It allows the acquisition time to be extended to increase the spatial resolution of the images without incurring substantial interference from enhancing veins and stationary tissues. In images, low spatial frequencies demonstrate the bulk of the contrast information, whereas high spatial frequencies demonstrate the detail information (see Fig. 1.23). Acquiring the low spatial frequency information (the contrast information) early in the scan, when only the arterial blood is enhanced, minimizes the amount of signal from veins and stationary tissues in the angiogram.

1

Basic Principles of MRI and MR Angiography

25

Fig. 1.27 Coronal MIP images of 8 of 13 time frames showing the vessels in the lower extremities of a patient acquired using the timeresolved 3D CE-MRA method called TRICKs. Note that the vessels in the

patient’s right leg fill much sooner and much more rapidly than the vessels in the patient’s left leg. The anatomical (nonvascular) MR images showed that this patient had a torn Achilles tendon in their right leg

Fig. 1.28 With 3D CE-MRA, it is common to subtract a “mask” image set, acquired prior to the injection of the contrast material, from an image set acquired during injection of the contrast material. This is an effective means of eliminating signals from stationary tissues that are common to both data sets, allowing the vessels to be much better

visualized. In this data set, the signal from the patient’s bladder is bright due to a prior injection of contrast material used to image vessels in a different anatomic location. Subtraction effectively eliminates this signal as well. Subtraction was performed on the source images. Shown here are the MIP images

If the contrast material reaches the veins and stationary tissues late in the scan, when the high spatial frequencies (the detail information) are being acquired, only the edges of these structures will appear in the images. This edge information from the veins and stationary tissues usually does not significantly interfere with observation of the arteries.

angiogram as shown in Fig. 1.28. The subtraction must be performed on the source slices prior to creation of the MIP images in order to fully retain the content of the three-dimensional data set. Mask subtraction in CE-MRA is analogous to subtraction in X-ray DSA in that the signals from stationary tissues (which remain constant between acquisitions) are eliminated by the subtraction, whereas the signals from vessels (which increased after the injection of the contrast material) persist after the subtraction. Disadvantages of subtraction are that it reduces the signal-to-noise ratio of the angiogram by a factor of 2 (because noise from two images, but signal from only one image, contribute to the angiogram), and it renders the angiograms susceptible to misregistration artifacts that result from patient motion.

Mask Subtraction The signal difference between vessels and stationary tissues can be increased in contrast-enhanced MRA methods by acquiring a precontrast mask image set and subtracting it from the postcontrast image set to produce a subtraction

26

F.R. Korosec

Multistation Exams Imaging more than one vascular territory, or station, during a single exam session presents challenges for CE-MRA. Two approaches are currently in use for performing multistation CE-MRA. In one approach, a separate injection of contrast material is administered just prior to scanning each territory [37]. There are benefits in this multi-injection approach, but the image quality can be adversely affected in two ways. First, the volume of contrast material available for each injection is only a fraction of the maximum allowed, because the volume must be split into multiple injections. The reduced volume leads to lower SNR in the angiograms. Second, residual contrast material remaining from the early injections enhances arteries in the mask acquisitions for the later injections. Because the arteries are slightly enhanced in the mask images, subtraction of the mask images leads to a signal reduction in the arteries in the subtracted angiogram. If the mask images from all stations are acquired before the first injection, they cannot be used to eliminate venous enhancement in the later images caused by residual contrast material remaining from the earlier injections. One benefit of a multi-injection method over a single-injection method (described below) is that more time can be spent imaging each station to achieve higher spatial resolution and extended coverage. This is possible because the scan times do not need to be limited in order to chase the bolus of contrast material. In a second approach utilized for multistation CE-MRA, multiple territories are imaged after administration of a single injection of contrast material [38–40]. After the administration of the contrast material is initiated in this single-injection approach, an image set is rapidly acquired from the first station, then the table is moved to bring the second station into the sensitive region of the MR scanner, and a scan is performed to image the second station. This table movement and data acquisition cycle is continued until all of the stations are imaged. With this method, it is necessary to scan quickly in order to image all the stations when the arterial signal is enhanced, and before the veins and stationary tissues enhance. Scanning quickly limits the spatial resolution and coverage that can be achieved at each station. Often the bolus of contrast material advances very quickly, and enhancement of veins and stationary tissues is inevitable in the later-acquired stations. When single-injection, multistation exams are performed, the mask images for all stations are acquired prior to administration of the contrast material, to prevent the contrast material from enhancing the vessels in any of the mask images. Finally, the rate of the injection must be reduced in

order to extend the duration of the contrast material administration to ensure that contrast material is present during imaging of all the stations.

Time-of-Flight MRA Just like CE-MRA techniques, TOF techniques [41, 42] derive contrast between flowing blood and stationary tissues by manipulating the magnitude of the magnetization, such that the magnetization is large for moving blood and small for stationary tissues. However, unlike CE-MRA techniques, TOF MRA techniques do not require the injection of a contrast material. Instead, TOF MRA techniques rely on the fact that blood is in motion, and stationary tissues are not. By appropriately adjusting the imaging parameters and scan prescription in TOF MRA techniques, a large signal may be obtained from moving blood, while a diminished signal is obtained from stationary tissues.

Imaging Sequence TOF methods can be implemented using two-dimensional or three-dimensional acquisition. For both acquisition methods, a spoiled, fast gradient-echo sequence is used. For twodimensional acquisition, thin slices are imaged, and the pulse sequence timing diagram is similar to the one shown in Fig. 1.16. For three-dimensional acquisition, thin slabs are imaged and the slabs are encoded into slices using a phase-encoding method. The pulse sequence timing diagram for three-dimensional TOF MRA is similar to the one shown in Fig. 1.16, with the addition of a phase-encoding gradient on the slice-selection axis. TOF sequences often employ additional gradients on the slice-selection and frequency-encoding axes to refocus unwanted phase accumulations accrued by spins that are in motion during the application of these imaging gradients. These additional gradients typically are referred to as flow-compensation gradients, velocity-compensation gradients, or first-momentnulling gradients [43, 44], and are discussed further in later chapters. They serve to reduce signal loss caused by intravoxel dephasing of spins traveling at different velocities. When a spoiled, fast gradient-echo sequence is used, the signal from tissues decreases with exposure to an increasing number of RF pulses. The longitudinal magnetization as a function of the number of RF pulses, n, is given by the following equation [45]:

(1 - (cosq E ) )+ M (cosq E ) M (n,q ) = M (1 - E ) n

z

1

0

1

1 - cos q E1

0

1

n

n ³ 1,

(1.12)

1

Basic Principles of MRI and MR Angiography

Fig. 1.29 Calculated signal for a spoiled gradient-echo imaging sequence with a flip angle of 30° and a TR of 5 ms shown as a function of the number of RF pulses experienced for tissues with various T1 values. Note that the signal decreases with increasing number of RF pulses experienced. Note also that the equilibrium signal is larger for tissues with shorter T1 values

where M z- (n,q ) is the longitudinal magnetization just prior to the nth RF pulse, E1def exp( -TR /T1 ) , M 0 is the thermal equilibrium magnetization, and q is the tip angle. given by (1.11). The steady-state equilibrium signal is often referred to as the saturation (or saturated) signal. The signal achieved using a spoiled gradient-echo sequence with a tip angle of 30°, a TR of 5 ms, and TE  T2* is plotted in the graph shown in Fig. 1.29. The signal is shown as a function of number of RF pulses experienced, and the signals are plotted for several T1 values. The graph shows that, independent of the T1 of the tissue, the signals decrease with exposure to an increasing number of RF pulses, until a steady-state equilibrium value is reached. In TOF MRA, the goal is to subject the flowing blood to a very few RF pulses, and to subject stationary tissues to a large number of RF pulses, thereby achieving a signal difference between blood and stationary tissues. The graph also shows that tissues with short T1 values have high steady-state equilibrium signal levels. Thus, tissues with short T1 values, like fat, show up bright in TOF MRA methods. Ensuring that flowing blood experiences only a few RF pulses can be accomplished by imaging slices, or thin slabs, oriented perpendicular to the primary direction of flow. When this is done, the moving blood enters the slice fully magnetized, experiences only a few excitation pulses, and then flows out of the slice before it becomes saturated. This ensures that the signal from blood will be relatively large. This phenomenon is referred to as inflow enhancement, or flow-related enhancement. The stationary tissues, however, remain in the slice, or slab, throughout image acquisition, and so they give rise to a diminished signal because the

27

magnetization from them is saturated due to the constant exposure to the excitation pulses. Data collection typically does not begin until the signals from stationary tissues are brought to their steady-state equilibrium level. The number of excitation pulses experienced by moving blood as it traverses the imaging slice (or slab) depends on a number of factors including the thickness of the slice (or slab), the velocity of the blood, the orientation of the vessel relative to the image slice, and the TR of the imaging sequence. In general, thinner slices (or slabs), faster-flowing blood, vessels oriented perpendicular to the slice (or slab), and longer TRs lead to increased vascular signal. A longer TR, however, also leads to increased signal from stationary tissues (less saturation). Therefore, an intermediate TR typically is selected with TOF MRA methods. Increasing the tip angle leads to diminished signal from stationary tissues, but can also lead to increased saturation of blood that experiences multiple excitation pulses. As a result, an intermediate tip angle typically is selected with TOF MRA methods. These factors and others must be carefully considered when prescribing a TOF MRA sequence.

Two-Dimensional TOF MRA For two-dimensional acquisition, data are acquired from multiple slices stacked contiguously along the long axes of the vessels of interest. Because the image quality is best if the slices remain perpendicular to the direction of flow, 2D TOF MRA methods are best suited for imaging vessels that are straight, such as the carotid arteries, or the vessels in the lower extremities. The data from the slices can be retrospectively reprojected or reformatted to demonstrate long segments of the vessels. With 2D TOF MRA, the spoiled, fast gradient-echo method is prescribed so that thin slices (1–3 mm), oriented perpendicular to the long axis of the vessels of interest are imaged. Prescribing the slices in this manner increases the likelihood that the blood will experience only a very few radiofrequency excitation pulses as it flows through the image slice. When thin slices are imaged, a moderately large tip angle (60°) can be used to suppress the signal from the stationary tissues without substantially suppressing the signal from blood that quickly moves through the image plane. Even when thin slices are imaged, the moderately large tip angle can cause saturation of the signal from slowly moving blood, such as that in the carotid bulb. The degree of saturation can be reduced by decreasing the tip angle and/or increasing the TR of the imaging sequence; however, it must be realized that this also will increase the amount of signal from stationary tissues. Increasing the TR also will lead to a longer scan time. When imaging vessels that contain pulsatile blood flow, it is necessary to synchronize the acquisition of data to the cardiac cycle. Cardiac leads are placed on the patient, and the

28

F.R. Korosec

Fig. 1.30 (a) To produce these cardiac-gated 2D TOF MRA coronal images of the lower legs, (b) a stack of contiguous axial slices was imaged. The data set was acquired in 4 scans (different shaded slices in (b)); the table was moved between scans so that the imaged slices were always in the most uniform magnetic field to yield the highest image quality. The images comprise a three-dimensional data set from which maximum intensity projections (MIPs) at any obliquity can be produced retrospectively

ECG signal is fed to the MRI scanner. Data are then acquired only during systole in order to take advantage of the maximum inflow of fresh, fully magnetized blood, thereby reducing the likelihood of saturating the blood. Acquiring segments of data always at the same point in each cardiac cycle also reduces ghosting artifacts that otherwise would occur due to the changing velocity of blood throughout the cardiac cycle. Images of the arteries in the lower extremities acquired using an ECG-gated, 2D TOF MRA method is shown in Fig. 1.30. With 2D TOF MRA, a spatial saturation pulse [46–48] can be applied parallel to the image slice at the beginning of each TR to reduce or eliminate unwanted signal from blood flowing into the image slice from a particular direction. For example, when imaging the vessels in the neck, applying a saturation pulse superior to an axial image slice is an effective means of eliminating signal from blood in the jugular veins, as shown in Fig. 1.31a. The signal from the blood in the jugular veins is diminished or eliminated as the blood flows through the region affected by the saturation pulse. The venous blood, therefore, has little signal to give when it flows into the image slice. If a saturation pulse was not used, and the blood in the jugular veins was left unsaturated, these vessels would interfere with observation of the carotid arteries as shown in Fig. 1.31b. When imaging the neck, if the saturation pulse is prescribed inferior to the image slice, the signal from the carotid arteries can be eliminated to produce an image of the jugular veins as shown in Fig. 1.31c. In general, inferior saturation pulses are used to suppress the signal from arteries above the heart and veins below the heart, whereas superior saturation pulses are used to suppress the signal from veins above the heart, and arteries below the heart.

Three-Dimensional TOF MRA For three-dimensional TOF acquisition [49, 50], a slab, oriented perpendicular to the long axis of the vessels of interest, is imaged and the slab is encoded into thin slices using a phase-encoding method. Because a slab is imaged, a small tip angle (30°) must be used so the signal from blood that remains in the slab does not become too saturated. The small tip angle necessary to preserve signal from blood also leads to an undesirable preservation of signal from stationary tissues. Therefore, when 3D TOF MRA methods are applied, other mechanisms must be implemented in order to improve contrast between blood and stationary tissues. Some of these mechanisms include magnetization transfer [51–53], fat-andwater out-of-phase imaging, and/or ramped tip angle [54] excitation. These methods are described in later chapters. Three-dimensional TOF MRA methods offer smaller voxels, shorter echo times, and inherently higher signal-tonoise ratios than 2D TOF MRA methods. These are features common to all three-dimensional acquisition methods as described previously. The use of phase-encoding, rather than slice-selection, leads to the ability to image thinner slices (better resolution in reformatted images, smaller voxels), and to use a shorter TE. The combination of the small voxels and the short echo time leads to a reduction in the amount of signal loss caused by intravoxel dephasing. Threedimensional TOF MRA images of the cerebral vessels are shown in Fig. 1.32. Even when using magnetization transfer, fat-and-water out-of-phase imaging, and ramped excitation pulses, contrast between blood and stationary tissues can be small in

1

Basic Principles of MRI and MR Angiography

29

Fig. 1.31 (a) In 2D TOF MRA of the neck, a spatial saturation pulse applied superior to, and very near, each axial imaging slice will eliminate signal from blood in the veins to yield an arteriogram. (b) When no spatial saturation pulse is applied, both arteries and veins appear in the image, making interpretation difficult. (c) Applying a spatial saturation pulse inferior to, and very near, each axial imaging slice will eliminate signal from blood in the carotid arteries to yield a venogram

Fig. 1.32 (a) Axial, (b) coronal, and (c) sagittal MIP images produced from axial images of the intracranial vessels of a healthy adult acquired using a 3D TOF MRA technique

3D TOF MRA when thick slabs are used to achieve extended coverage. To achieve greater coverage with reduced saturation effects, 3D TOF MRA data can be acquired from multiple thin slabs arranged perpendicular to the vessels of interest using a method referred to as multiple overlapping thin slab angiography (MOTSA) [42, 55]. The multislab method combines the thin-slice benefits of two-dimensional acquisition (reduced saturation of blood) with the benefits of three-dimensional acquisition (including an inherently high signal-to-noise ratio, small voxels, and a short echo time) in an effort to provide high quality images as shown in Fig. 1.33. With 3D TOF MRA, a spatial saturation pulse can be applied just outside the slab at the beginning of each TR to eliminate signal from blood that will flow into the imaging slab as described previously for 2D TOF MRA. However, with 3D TOF MRA, the saturation pulse is not as effective.

This is because, as the blood that was initially saturated traverses the slab, the longitudinal magnetization from this blood begins to regrow, and the once-saturated blood eventually gives rise to signal. In general, the farther the oncesaturated blood penetrates into the slab, the less effective the saturation of the signal due to the increasing amount of time that elapses between saturation and signal detection.

Phase-Contrast MRA Phase-contrast (PC) MRA [56, 57] methods differ from CE-MRA and TOF MRA methods in that PC MRA methods provide a direct quantitative measure of the velocity of the flowing blood. Like TOF MRA methods, PC MRA methods can be acquired using 2D or 3D acquisitions. The 2D acquisitions, since they are rapid, can be cardiac-gated to provide

30

F.R. Korosec

Fig. 1.33 (a) Coronal MIP of axial images of the carotid arteries acquired using a multiple overlapping thin slab angiography (MOTSA) technique with a ramped tip angle, and a TE such that fat and water are out of phase with each other. (b) Magnified and cropped oblique MIP of the left carotid bifurcation

Fig. 1.35 A timing diagram for a 2D phase-contrast MRA pulse sequence. The timing diagram is similar to the one shown in Fig. 1.16 with the addition of a bipolar flow-encoding gradient on the frequencyencoding axis (shown as a dotted grey line to distinguish it from the other magnetic field gradients) Fig. 1.34 The transverse magnetization, Mxy, precesses when a magnetic field is applied. The amount that the magnetization rotates is denoted by the phase angle, f

dynamic information regarding blood flow throughout the cardiac cycle. Phase-contrast MRA techniques derive contrast between flowing blood and stationary tissues by manipulating the phase of the magnetization, such that the phase of the magnetization is zero for stationary tissues and non-zero for moving tissues. Phase is a measure of how far the magnetization precesses, or rotates, from the time it is tipped into the transverse plane until the time it is detected. The phase of the precessing transverse magnetization is shown in Fig. 1.34.

Imaging Sequence The pulse sequence used for phase-contrast acquisitions is a modified, spoiled, fast gradient-echo sequence. The most

significant modification to the gradient-echo pulse sequence timing diagram is the addition of a bipolar flow-encoding gradient as shown for 2D acquisition in Fig. 1.35. For 3D acquisition, a phase-encoding gradient would be added to the slice-selection axis. As described below, in order to encode flow in all directions, a data set must be acquired with the bipolar flow-encoding gradient applied on each of the three gradient axes, and to determine and eliminate unwanted phase accumulations from sources other than the bipolar flowencoding gradient, a reference data set also must be acquired. The effect of applying the bipolar flow-encoding gradient in PC MRA is that spins will accumulate a phase in proportion to their velocity. For example, stationary spins will accumulate no net phase shift. This is because the first lobe of the gradient (positive polarity) will impart a positive phase to the stationary spins, and the second lobe of the gradient (negative polarity) will impart an equivalent negative phase to the stationary spins, resulting in no net phase accumulation for stationary spins at the end of the application of the flowencoding gradient. So the signal associated with voxels containing stationary spins will be zero.

1

Basic Principles of MRI and MR Angiography

31

Fig. 1.36 (a) Axial, (b) coronal, and (c) sagittal MIP images of the intracranial vessels of a patient acquired using an experimental 3D PC MRA technique that employs an very rapid undersampled 3D radial

acquisition method (Images courtesy of Kevin Johnson, University of Wisconsin-Madison)

Moving spins, on the other hand, will be left with a net phase accumulation at the end of the application of the bipolar flow-encoding gradient. Since the spins are in motion, they will experience different field strengths during application of the positive and negative lobes of the bipolar flowencoding gradient, and therefore the phase imparted by the positive lobe will not equal the phase imparted by the negative lobe. The resulting phase is proportional to the velocity of the spins since faster moving spins travel farther between (and during) application of the positive and negative lobes of the bipolar flow-encoding gradient. This leads to a bigger difference in the gradient strength experienced by the spins during application of the positive and negative lobes of the gradient, which translates to a greater difference in the phase imparted to the spins by the positive and negative lobes of the bipolar flow-encoding gradient. Note that phase will only be imparted to spins that move along the axis containing the bipolar flow-encoding gradient. The phase imparted to spins by the application of a bipolar flow-encoding gradient on the frequency-encoding axis is given by the following equation:

data set is subtracted from each of the three flow-encoded data sets to eliminate phase accumulations resulting from sources other than the bipolar flow-encoding gradient (magnetic susceptibility, eddy currents, measurement imperfections, etc.). Alternatively, the bipolar flow-encoding gradients may be strategically applied on more than one axis simultaneously, to improve the signal-to-noise ratio of the subtracted images [58, 59]. These alternative methods still require the acquisition of four scans to acquire velocity information in all three Cartesian directions. The scan time for encoding flow in all three Cartesian directions using two-dimensional acquisitions is 4 × TR × Ny × Ave, where TR is the repetition time of the imaging sequence, Ny is the number of phase-encoding values acquired, and “Ave” is the number of signal averages. If flow is encoded in only a single direction, the factor of four is reduced to a factor of two. With two-dimensional acquisitions, a single thick slab is typically imaged. If multiple slabs are imaged, the scan time is determined by multiplying the above equation by the number of slabs imaged. Phase-contrast methods may also be implemented using three-dimensional acquisition [60, 61] as shown in Fig. 1.36. Benefits of three-dimensional acquisition include an inherently high signal-to-noise ratio (due to the effective averaging of signal that is acquired throughout the entire scan) and small voxels (due to the thin slices made possible by the encoding method employed). Also, three-dimensional image sets can be retrospectively reprojected or reformatted to permit observation of the vessels from any orientation. When phase-difference processing is used, images reformatted perpendicular to a vessel can be used to determine volume flow rate through that vessel. Three-dimensional acquisition is more time consuming than two-dimensional acquisition and thus is used less frequently. Due to the long scan times associated with three-dimensional acquisitions, they are not commonly cardiac-gated.

f v = ±g Gbp vFEt 2 ,

(1.13)

where g is the gyromagnetic ratio of the spins, Gbp is the amplitude of the bipolar flow-encoding gradient, vFE is the velocity of the spins along the frequency-encoding axis, t is the duration of each lobe of the bipolar flow-encoding gradient, and the ± indicates that the sign of the phase depends on the sign of the velocity, and whether the first lobe of the bipolar flow-encoding gradient is positive or negative. This equation ignores the ramp times of the bipolar flow-encoding gradient. In order to encode flow in all directions, the bipolar flowencoding gradient must be applied on each of the three gradient axes in separate TR intervals. In addition, a fourth nonflowencoded acquisition must be acquired. The nonflow-encoded

32

F.R. Korosec

Fig. 1.37 Vector diagrams showing the values that are measured for each pixel when implementing (a) phase-difference, (b) complexdifference, and (c) magnitude processing of phase-contrast data. In phase-difference processing, the phase angle, f, through which the magnetization rotates during each of the two acquisitions is

measured. In complex-difference processing, the distance between the ends of the magnetization vectors from the two acquisitions is measured. In magnitude processing, the magnitude of the magnetization vectors from the two acquisitions is measured, and the average is calculated

The scan time required for encoding flow in all directions using three-dimensional acquisitions is 4 × TR × Ave × Ny × Nz, where Nz is the number of slice-encoding values acquired and all other abbreviations are as described previously. If flow is encoded in only a single direction, the factor of four is reduced to a factor of two. With phase-contrast acquisition methods, the subtraction provides high contrast between vessels and stationary tissues, permitting large fields-of-view to be imaged with minimal detrimental effects from saturation, as long as a relatively small tip angle (20°–30°) is used.

by each pixel. Additionally, spins moving in one direction are assigned a bright (white) signal, whereas spins moving in the opposite direction are assigned a dark (black) signal as shown in Fig. 1.38. So, a cursor or region of interest (ROI) can be placed on the phase-difference images, and a direct measure of the velocity (mm/s) of the blood can be determined. In addition, if the cross-section of a vessel is demonstrated in the image, information regarding the volume flow rate of the blood (ml/min) in that vessel can be derived [62]. Furthermore, if the acquisition is cardiac-gated [63, 64], and a series of images is produced demonstrating the cross-section of a vessel throughout the cardiac cycle, then the acquired series of images can be evaluated to determine volume flow rate in that vessel throughout the cardiac cycle. The physiologic effect of vascular pathology can be evaluated by comparing the volume flow rate on the contralateral side with that on the ipsilateral side, or by comparing the volume flow rate before and after a physical or drug-induced challenge. The accuracy of volume flow rate measurements obtained using phase-contrast MRA is dependent on imaging parameters and placement of the ROI [62]. Because phase-difference images show direction of flow in the form of black or white signal, these images may be used to effectively differentiate arteries from veins when they are aligned antiparallel to each other. More importantly, phase-difference images can be used to detect retrograde flow in cases such as in subclavian steal syndrome, where a stenosis or occlusion of the subclavian artery proximal to the vertebral artery causes reversed flow in the latter. In complex-difference processing, vector subtraction is performed to determine the signal in each pixel as shown in Fig. 1.37b. Complex-difference images are shown in Fig. 1.38. With complex-difference processing, the signal in the images is dependent on the motion of the blood (as it is in phase-difference images), but the dependence is not linear (as it is in phase-difference images). Furthermore, the direction of blood flow is not represented in complex-difference

Processing Methods The data acquired with phase-contrast techniques can be processed in three different ways to produce phase-difference, complex-difference, and magnitude images. In phase-difference processing, phase subtraction is performed to determine the signal in each pixel as shown in Fig. 1.37a. For phase-difference processing, the phase of the magnetization is measured at the point in time when the receiver is activated. Since the phase of the magnetization is influenced by factors in addition to the bipolar flow-encoding gradient, typically two measurements are made, with only the bipolar flow-encoding gradient changed between the measurements, so the phase imparted by other factors remains the same for the two measurements. When the data for the two measurements are subtracted, the phase accumulation imparted by the other factors is eliminated since it is the same for the two measurements. The phase imparted by the bipolar flow-encoding gradients is opposite in the two measurements, so subtraction yields an enhancement of the phase caused by application of the bipolar flow-encoding gradient. In phase-difference images, the signal value of each pixel is linearly proportional to the velocity of the spins represented

1

Basic Principles of MRI and MR Angiography

33

Fig. 1.38 Sagittal images of the midline intracranial vessels demonstrating that data acquired using phase-contrast MRA can be reconstructed to produce phase-difference images showing flow in the (a) right/left, (b) anterior/posterior, and (c) superior/inferior directions. In these images, the intensity of the signal is proportional to the velocity of the blood. The intensity in these images can be squared, summed, and the square root can be calculated to produce a “speed” image as

shown in (g). Data acquired using phase-contrast MRA can also be reconstructed to produce complex-difference images showing flow in the (d) right/left, (e) anterior/posterior, and (f) superior/inferior directions. The intensity in these images can be squared, summed, and the square root can be calculated to produce a “speed” image as shown in (h). Finally, the phase-contrast data can be reconstructed to produce a magnitude image as shown in (i)

images, so flow in opposite directions is not represented as black and white as it is in phase-difference images. Therefore, complex-difference images are not used to determine quantitative information regarding blood velocity or volume flow rate, but are used instead for demonstrating the morphology of the vessels. In magnitude processing, the length of the transverse magnetization vector is measured to determine the signal in each pixel as shown in Fig. 1.37c. This is the processing method traditionally used to produce images in MRI. With phase-contrast methods, the phase of the magnetization is manipulated, but the strength of the magnetization, or its magnitude, is unaltered. Thus, displaying the magnitude of the magnetization permits simultaneous demonstration of vessels and stationary tissues as shown in Fig. 1.38.

methods, only velocities within a certain range will be accurately represented in the images. The range of velocities that will be accurately represented is determined by the velocityencoding (Venc) parameter selected when the imaging sequence is prescribed. Blood traveling at velocities higher than the Venc value will be misrepresented in the image, so this value must be chosen carefully. This misrepresentation of velocities outside of the Venc value is referred to as velocity aliasing and appears differently in phase-difference and complex-difference images. In phase-difference images, aliased signal is easy to identify by an abrupt change from very dark to very bright signal as seen in the image shown in Fig. 1.40, or vice versa (see transitions in Fig. 1.39 for phase-difference processing). In complex-difference images, velocity aliasing manifests as a decreasing signal for blood traveling at velocities beyond the Venc value (see Fig. 1.39 for complex-difference processing), and blood traveling at some velocities above the Venc value gives rise to no signal at all. Thus, with complex-difference processing, selecting a Venc that is too low can yield very misleading results that are not easy to identify. With both phasedifference and complex-difference processing, choosing a Venc value that is too low should be avoided in order to

Velocity Encoding and Velocity Aliasing In phase-contrast images, the signals are proportional to the velocities of the imaged spins, as shown in the graphs in Fig. 1.39. The proportionalities are different for phase-difference and complex-difference processing. For both processing

34

Fig. 1.39 Graphs demonstrating the signal as a function of velocity for (top) phase-difference (PD) and (bottom) complex-difference (CD) processing. For phase-difference processing, the signal is linearly proportional to the velocity of the spins – for spins with velocities within the range defined by the velocity-encoding (Venc) value specified by

F.R. Korosec

operator. For complex-difference processing, the signal is sinusoidally proportional to the speed (the signal is the same for positive and negative velocities) of the spins – for spins with velocities within the range defined by the velocity-encoding (Venc) value specified by the operator

veins in an arteriovenous malformation (AVM) can be displayed in separate images that are sensitive to different velocity ranges as shown in Fig. 1.41.

Balanced Steady-State Free Precession MRA

Fig. 1.40 With phase-contrast MRA methods, a velocity-encoding, Venc, value must be selected. Velocities above this value will be misrepresented, or aliased. In phase-difference-processed images, such as the one shown here, velocity aliasing is easy to identify (arrows) as an abrupt change from very dark to very bright signal, or from very bright to very dark signal (not shown)

prevent velocity aliasing. Choosing a Venc value that is too high also should be avoided as it leads to magnetization from all blood accumulating only small phase shifts, which results in low SNR, and a small signal range. To avoid velocity aliasing, some a priori information regarding the anticipated pathology may be helpful in determining the appropriate velocity-encoding value. Alternatively, different velocity-encoding values may be used in different scans to highlight different vessels. For example, arteries and

MRA also can be performed using balanced steady-state free precession (bSSFP) imaging sequences [65]. In bSSFP imaging, both the longitudinal and transverse components of magnetization are maintained in a steady-state condition. In order to ensure that the transverse component of magnetization is not spoiled, all of the imaging gradients must be completely balanced so that the net phase accumulation imparted to the spins from one TR to the next is zero. A pulse sequence timing diagram for a bSSFP sequence is shown in Fig. 1.42. Note that, between RF pulses, the area under the positive gradient waveforms equals the area under the negative gradient waveforms, ensuring that the net phase imparted to any spin by the imaging gradients during this interval is zero. Depending on MRI vendor, this imaging sequence is called FIESTA (fast imaging employing stead-state acquisition), trueFISP (true fast imaging with steady-state precession), or bFFE (balanced fast field echo). Since the transverse magnetization is maintained from one TR to the next, those tissues with long T2 values give rise to large signals in bSSFP methods. Tissues with short T1 values also give rise to large signals. The signal intensity is proportional to T2/T1. The signal intensity achieved with signal

1

Basic Principles of MRI and MR Angiography

35

Fig. 1.41 Phase-contrast MRA images can be acquired with different velocity sensitivities (using different Venc values) to highlight vessels containing blood flowing in different velocity ranges. These sagittal complex-difference-processed phase-contrast images of an arteriovenous malformation (AVM) were acquired with velocity-encoding

Fig. 1.42 A timing diagram for a balanced steady-state free precession (bSSFP) gradient-echo pulse sequence. Note that the gradient waveforms on all three axes are balanced so that the positive area equals the negative area to ensure that no net phase is imparted to any spins during the TR interval. This allows the transverse magnetization to achieve a steady-state value

steady-state free precession methods (when steady-state has been achieved) is given by the following equation [66]: M xy+ = M 0 ´

(1 - E1 ) ´ sin(q ) , 1 - ( E1 - E2 ) ´ cos(q ) - E1 E2

(1.14)

where M xy+ is the transverse magnetization immediately following the RF pulse, E1,2 def exp( -TR /T1,2 ) , M 0 is the thermal equilibrium magnetization, and q is the tip angle. If TE = TR/2, then the transverse magnetization that is detected is given by the following equation [67]: M xy = M xy+ ´ E2 .

(1.15)

A cardiac image acquired with a bSSFP method is shown in Fig. 1.43. Owing to its relatively long T2, blood appears

values of (a) 20 cm/s and (b) 80 cm/s. Note that the slower flow in the nidus of the AVM is better demonstrated in the image shown in (a), whereas the faster flow in the feeding arteries is better demonstrated in the image shown in (b)

Fig. 1.43 One of 16 images of a patient’s heart acquired during a single breath-hold interval using a cardiac-gated, balanced steady-state free precession (bSSFP) method. With these methods, multiple images of the same anatomic slice are typically acquired. Each image shows the heart in a different phase of the cardiac cycle. When all of the images are displayed in rapid succession (as a movie loop), the cardiac dynamics are revealed

bright in this image, providing high contrast between it and the surrounding myocardium. Other tissues with large T2/T1 ratios also appear bright in images acquired with bSSFP methods, limiting the application of these methods for MRA applications. Fat and venous blood have large T2/T1 ratios. Fat suppression may be achieved to varying degrees of success by using water-selective excitation pulse, or repeated spectral fat-saturation pulses. Similarly, venous suppression may be achieved to varying degrees of success by using spatial saturation or inversion pulses. Balanced SSFP methods suffer from signal loss in regions of magnetic field nonuniformities, since, under these circumstances, the requirement for zero net phase accumulation

36

between TRs is not satisfied. Thus, shimming the magnetic field is very important when using bSSFP methods. In order to reduce phase accumulations caused by remaining magnetic field nonuniformities, short TRs are employed with bSSFP methods. Balanced SSFP methods have proven useful for cardiac imaging applications [68], but owing to the remaining challenges faced by these methods, they currently are not widely used for MRA applications.

Summary Under appropriate conditions, and when properly implemented, all MR angiographic methods can yield high-quality diagnostic images. Three-dimensional MRA methods offer small voxels, short echo times, inherently high signal-tonoise ratios, and the ability to retrospectively reformat image volumes to show tortuous vessels from any viewing angle. Three-dimensional CE-MRA methods can provide high quality vascular images with less sensitivity to artifacts (caused by intravoxel dephasing, signal saturation, cardiac pulsatility, and patient motion) and shorter scan times than noncontrast-enhanced MRA methods. Intravenous injection of a T1-shortening contrast material provides high SNR angiograms with high contrast between vessels and stationary tissues. Relatively high spatial resolution and large volume coverage may be achieved in reasonable scan times. In 3D CE-MRA images, vascular pathologies are well delineated. Owing to its many attributes, 3D CE-MRA is widely used for imaging vessels throughout the body, and in some situations is replacing X-ray DSA as the imaging method-of-choice. Three-dimensional TOF MRA methods are often used for imaging the intracranial arteries, as well as the extracranial carotid and vertebral arteries. Two-dimensional TOF methods are sensitive to signal loss from intravoxel dephasing and often are used as a highly sensitive and specific method for screening for stenoses of the carotid or lower extremity arteries. Two-dimensional TOF also may be used with an inferior spatial saturation pulse to image the intracranial veins. Twodimensional TOF MRA may be cardiac-gated to provide images in regions affected by cardiac pulsatility. Phase-contrast methods offer qualitative and quantitative information regarding blood velocity or volume flow rate. Two-dimensional phase-contrast methods offer short scan times, and often are used to provide multiple images, each sensitive to a different range of velocities, or are cardiacgated to provide velocity or flow information throughout the cardiac cycle. Three-dimensional phase-contrast images require long scan times, but the volume of information may be retrospectively analyzed to yield average velocity or flow information through any vessel, no matter what its orientation. The use of a bipolar flow-encoding gradient makes

F.R. Korosec

these sequences particularly susceptible to signal loss caused by intravoxel dephasing. bSSFP methods are widely used for cardiac applications, but owing to several remaining challenges, are not yet routinely used for MRA applications.

References 1. Bernstein MA, Huston J 3rd, Lin C, Gibbs GF, Felmlee JP. Highresolution intracranial and cervical MRA at 3.0T: technical considerations and initial experience. Magn Reson Med. 2001;46:955–962. 2. Thomas SD, Al-Kwifi O, Emery DJ, Wilman AH. Application of magnetization transfer at 3.0 T in three-dimensional time-of-flight magnetic resonance angiography of the intracranial arteries. J Magn Reson Imaging. 2002;15:479–483. 3. Willinek WA, Born M, Simon B, Tschampa HJ, Krautmacher C, Gieseke J, Urbach H, Textor HJ, Schild HH. Time-of-flight MR angiography: comparison of 3.0-T imaging and 1.5-T imaging-initial experience. Radiology. 2003;229:913–920. 4. Leiner T, de Vries M, Hoogeveen R, Vasbinder GB, Lemaire E, van Engelshoven JM. Contrast-enhanced peripheral MR angiography at 3.0 Tesla: initial experience with a whole-body scanner in healthy volunteers. J Magn Reson Imaging. 2003;17:609–614. 5. Gibbs GF, Huston J 3rd, Bernstein MA, Riederer SJ, Brown RD Jr. Improved image quality of intracranial aneurysms: 3.0-T versus 1.5-T time-of-flight MR angiography. Am J Neuroradiol. 2004;25:84–87. 6. Edelstein WA, Hutchison JM, Johnson G, Redpath T. Spin warp NMR imaging and applications to human whole-body imaging. Phys Med Biol. 1980;25:751–756. 7. Glover GH, Pauly JM. Projection reconstruction techniques for reduction of motion effects in MRI. Magn Reson Med. 1992;28: 275–289. 8. Meyer CH, Hu BS, Nishimura DG, Macovski A. Fast spiral coronary artery imaging. Magn Reson Med. 1992;28:202–213. 9. Barger AV, Block WF, Toropov Y, Grist TM, Mistretta CA.Timeresolved contrast-enhanced imaging with isotropic resolution and broad coverage using an undersampled 3D projection trajectory. Magn Reson Med. 2002;48:297–305. 10. Thedens DR, Irarrazaval P, Sachs TS, Meyer CH, Nishimura DG. Fast magnetic resonance coronary angiography with a three-dimensional stack of spirals trajectory. Magn Reson Med. 1999;41:1170–1179. 11. Gurney PT, Hargreaves BA, Nishimura DG. Design and analysis of a practical 3D cones trajectory. Magn Reson Med. 2006;55: 575–582. 12. Macovski A. Noise in MRI. Magn Reson Med. 1996;36:494–497. 13. Sodickson DK, Manning WJ. Simultaneous acquisition of spatial harmonics (SMASH): fast imaging with radiofrequency coil arrays. Magn Reson Med. 1997;38:591–603. 14. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: sensitivity encoding for fast MRI. Magn Reson Med. 1999;42:952–962. 15. Weiger M, Pruessmann KP, Kassner A, Roditi G, Lawton T, Reid A, Boesiger P. Contrast-enhanced 3D MRA using SENSE. J Magn Reson Imaging. 2000;12:671–77. 16. Griswold MA, Jakob PM, Heidemann RM, Nittka M, Jellus V, Wang J, Kiefer B, Haase A. Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson Med. 2002;47:1202–1210. 17. Prince MR, Yucel EK, Kaufman JA, et al.: Dynamic gadoliniumenhanced three-dimensional abdominal MR arteriography. J Magn Reson Imaging. 1993;3:877–881. 18. Kroft LJ, de Roos A. Blood pool contrast agents for cardiovascular MR imaging. J Magn Reson Imaging. 1999;10:395–403. 19. Perreault P, Edelman MA, Baum RA, Yucel EK, Weisskoff RM, Shamsi K, Mohler ER 3rd. MR angiography with gadofosveset

1

20.

21. 22.

23.

24.

25.

26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

Basic Principles of MRI and MR Angiography trisodium for peripheral vascular disease: phase II trial. Radiology. 2003;229:811–820. Thomsen HS, Morcos SK, Dawson P. Is there a causal relation between the administration of gadolinium based contrast media and the development of nephrogenic systemic fibrosis (NSF)? Clin Radiol. 2006;61:905–906. Zur Y, Wood ML, Neuringer LJ. Spoiling of transverse magnetization in steady-state sequences. Magn Reson Med. 1991;21:251–263. Haacke E, Brown R, Thompson M, Venkatesan R. Magnetic Resonance Imaging: Physical Principles and Sequence Design. New York: John Wiley and Sons, Inc.; 1999:455. Du YP. Parker DL. Davis WL, Cao G. Reduction of partial-volume artifacts with zero-filled interpolation in three-dimensional MR angiography. J Magn Reson Imaging. 1994;4:733–741. Bernstein MA, Fain SB, Riederer SJ. Effect of windowing and zero-filled reconstruction of MRI data on spatial resolution and acquisition strategy. J Magn Reson Imaging. 2001;14:270–280. Maki J, Prince M, Londy F, Chenevert T. The effects of time varying intravascular signal intensity and k-space acquisition order on three-dimensional MR angiography image quality. J Magn Reson Imaging. 1996;6:642–651. Earls J, Rofsky N, DeCorato D, et al. Breath-hold single-dose gadolinium-enhanced three-dimensional MR aortography: usefulness of a timing examination and MR power injector. Radiology. 1996;201:705–710. Foo T, Saranathan M, Prince M, Chenevert T. Automated detection of bolus arrival and initiation of data acquisition in fast, threedimensional, gadolinium enhanced MR angiography. Radiology. 1997;203:275–280. Huston J, Fain S, Riederer S, et al. Carotid arteries: Maximizing arterial to venous contrast in fluoroscopically triggered contrastenhanced MR angiography with elliptical centric view ordering. Radiology. 1999;211:265–273. Wang Y, Johnston D, Breen J, et al. Dynamic MR digital subtraction angiography using contrast enhancement, fast data acquisition, and complex subtraction. Magn Reson Med. 1996;36:551–556. Hennig J, Scheffler K, Laubenberger J, Strecker R. Time-resolved projection angiography after bolus injection of contrast agent. Magn Reson Med. 1997;37:341–345. Korosec F, Frayne R, Grist T, Mistretta C. Time-resolved contrastenhanced 3D MR angiography. Magn Reson Med. 1996;36:345–351. Willig DS, Turski PA, Frayne R, et al. Contrast-enhanced 3D MR DSA of the carotid artery bifurcation: Preliminary study of comparison with unenhanced 2D and 3D time-of-flight MR angiography. Radiology. 1998;208:447–451. van Vaals JJ, Brummer ME, Dixon WT, Tuithof HH, Engels H, Nelson RC, Gerety BM, Chezmar JL, den Boer JA. “Keyhole” method for accelerating imaging of contrast agent uptake. J Magn Reson Imaging. 1993;3:671–675. Willinek WA, Hadizadeh DR, von Falkenhausen M, Urbach H, Hoogeveen R, Schild HH, Gieseke J. 4D time-resolved MR angiography with keyhole (4D-TRAK): more than 60 times accelerated MRA using a combination of CENTRA, keyhole, and SENSE at 3.0T. J Magn Reson Imaging. 2008;27:1455–1460. Wilman A, Riederer S, Huston J, et al. Arterial phase carotid and vertebral artery imaging in 3D contrast-enhanced MR angiography by combining fluoroscopic triggering with an elliptical centric acquisition order. Magn Reson Med. 1998;40:24–35. Huston J, Fain SB, Wald JT, Luetmer PH, et al. Carotid artery: elliptic centric contrast-enhanced MR angiography compared with conventional angiography. Radiology. 2001;218:138–143. Earls JP, Patel NH, Smith PA, DeSena S, Meissner MH. Gadoliniumenhanced three-dimensional MR angiography of the aorta and peripheral arteries: evaluation of a multistation examination using two gadopentetate dimeglumine infusions. AJR Am J Roentgenol. 1998;171:599–604.

37 38. Ho KY, Leiner T, de Haan MW, Kessels AG, Kitslaar PJ, van Engelshoven JM. Peripheral vascular tree stenoses: evaluation with moving-bed infusion-tracking MR angiography. Radiology. 1998; 206:683–692. 39. Earls JP, DeSena S, Bluemke DA. Gadolinium-enhanced threedimensional MR angiography of the entire aorta and iliac arteries with dynamic manual table translation. Radiology. 1998;209: 844–849. 40. Meaney JF, Ridgway JP, Chakraverty S, et al. Stepping-table gadolinium-enhanced digital subtraction MR angiography of the aorta and lower extremity arteries: preliminary experience. Radiology. 1999;211:59–67. 41. Keller PJ, Drayer BP, Fram EK, et al. MR angiography with twodimensional acquisition and three-dimensional display: work in progress. Radiology. 1989;173:527–532. 42. Parker DL, Yuan C, Blatter DD. MR angiography by multiple thin slab 3D acquisition. Magn Reson Med. 1991;17:434–451. 43. Pattany PM, Phillips JJ, Chiu LC, Lipcamon JD, et al. Motion artifact suppression technique (MAST) for MR imaging. J Comput Assist Tomogr. 1987;11:369–377. 44. Haacke E, Lenz G. Improving MR image quality in the presence of motion by using rephasing gradients. AJR Am J Roentgenol. 1987;148:1251–1258. 45. Haacke E, Brown R, Thompson M, Venkatesan R. Magnetic Resonance Imaging: Physical Principles and Sequence Design. New York: John Wiley and Sons, Inc.; 1999:461. 46. Felmlee JP, Ehman RL. Spatial presaturation: a method for suppressing flow artifacts and improving depiction of vascular anatomy in MR imaging. Radiology. 1987;164:559–564. 47. Edelman RR, Atkinson DJ, Silver MS, et al. FRODO pulse sequences: a new means of eliminating motion flow, and wraparound artifacts. Radiology. 1988;166:231–236. 48. Ehman RL, Felmlee JP. Flow artifact reduction in MRI: a review of the roles of gradient moment nulling and spatial presaturation. Magn Reson Med. 1990;14:293–307. 49. Ruggieri PM, Laub GA, Marsaryk TJ, et al. Intracranial circulation: pulse-sequence considerations in three-dimensional (volume) MR angiography. Radiology. 1989;171:785–791. 50. Schmalbrock P, Yuan C, Chakeres DW, et al. Volume MR angiography: methods to achieve very short echo times. Radiology. 1990;175:861–865. 51. Wolff SD and Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med. 1989;10:135–144. 52. Pike GB, Hu BS, Glover GH, Enzmann DR. Magnetization transfer time-of-flight magnetic resonance angiography. Magn Reson Med. 1992;25:372–379. 53. Edelman RR, Ahn SS, Chien D, et al. Improved time-of-flight MR angiography of the brain with magnetization transfer contrast. Radiology. 1992;184:395–399. 54. Nagele T, Klose U, Grodd W, Petersen D, et al. The effects of linearly increasing flip angles on 3D inflow MR angiography. Magn Reson Med. 1994;31:561–566. 55. Blatter DD, Parker DL, Robinson R. Cerebral MR angiography with multiple overlapping thin slab acquisition. Radiology. 1991; 179:805–811. 56. Moran PR. A flow zeugmatographic interlace for NMR imaging in humans. Magn Reson Imaging. 1982;1:197–203. 57. Dumoulin CL and Hart HR. Magnetic resonance angiography. Radiology. 1986;161:717–720. 58. Pelc NJ, Bernstein MA, Shimakawa A, Glover GH. Encoding strategies for three-direction phase-contrast MR imaging of flow. Magn Reson Imaging. 1991;1:405–413. 59. Hausman R, Lewin JS, Laub G. Phase-contrast MR angiography with reduced acquisition time: new concepts in sequence design. J Magn Reson. 1991;1:415–422.

38 60. Dumoulin C, Souza S, Walker M, Wagle W. Three-dimensional phase contrast angiography. Magn Reson Med. 1989;9:139–149. 61. Pernicone JR, Siebert JE, Potchen EJ, et al. Three-dimensional phase-contrast MR angiography in the head and neck: preliminary report. AJR Am J Roentgenol. 1990;155:167–176. 62. Korosec F and Turski P. Velocity and volume flow rate measurements using phase contrast magnetic resonance imaging (review article). Int J Neuroradiology. 1997;3:293–318. 63. Nayler GL, Firmin DN, Longmore DB. Blood flow imaging by cine magnetic resonance. J Comput Assist Tomogr. 1986;10:715–722. 64. Pelc NJ, Herfkens RJ, Shimakawa A, Enzmann DR. Phase contrast cine magnetic resonance imaging. Magn Reson Q. 1991;7:229–254.

F.R. Korosec 65. Scheffler K, Lehnhardt S. Principles and applications of balanced SSFP techniques (review article). Eur Radiol. 2003;13: 2409–2418. 66. Haacke E, Brown R, Thompson M, Venkatesan R. Magnetic Resonance Imaging: Physical Principles and Sequence Design. New York: John Wiley and Sons, Inc.; 1999:480. 67. Scheffler K, Hennig J. Is true FISP a gradient-echo or a spin-echo sequence? Magn Reson Med. 2003;49:395–397. 68. Carr JC, Simonetti O, Bundy J, Li D, Pereles S, and Finn JP. Cine MR angiography of the heart with segmented true fast imaging with steady-state precession. Radiology. 2001;219: 828–834.

2

Time-of-Flight Angiography Seong-Eun Kim and Dennis L. Parker

Introduction Time-of-flight (TOF) angiography is a widely used means of producing angiographic images without the need for the injection of contrast agent. In this chapter, we discuss the MR imaging (MRI) physics which governs the creation of TOF angiograms. We also describe the MRI pulse sequence used for TOF imaging, sources of artifact, and compensatory mechanisms to reduce the deleterious effect of these artifacts.

MR Signal in Time of Flight The flow sensitivity of MRI methods is based on TOF effects, where the amplitude of the signal from flowing blood changes as it moves into the imaged volume, and phase effects, where motion of the blood during applied gradients results in a phase change due to motion. Both flow phenomena can be used to differentiate flowing spins from stationary spins by evaluating either the magnitude or the phase in the acquired MRI data. TOF effects influence the signal intensities of moving blood in MR image in nonangiographic applications. For example, flow voids resulting from the motion through a slice- and flow-related enhancement (FRE) have commonly been seen in both spin echo and gradient echo imaging, respectively. TOF image contrast is based on establishing a difference in the longitudinal magnetization of moving spins relative to stationary spins. The TOF effect was first reported in a nonimaging application by Suryan [1] and in the imaging of blood vessels by Hinshaw et al. [2]. The magnetization of a bolus of flowing blood is typically modified at one location and detected a short time later at another location. Since time elapses between the modification and detection S.-E. Kim, PhD () • D.L. Parker, PhD Department of Radiology, Utah Center for Advanced Imaging Research, Salt Lake City, UT, USA e-mail: [email protected]

of the flowing magnetization, this effect is referred to as the “TOF” effect. TOF techniques in MRA can be divided into those which yield a high signal from flowing spins and low signal from the background tissue (white blood) and those which yield strong signal from the background tissue and little signal from the flowing spins (black blood). White blood (sometimes called bright blood) techniques are the most commonly used for angiography and are the focus of the remainder of this chapter. In white blood TOF, static tissue is suppressed by using a spoiled gradient echo (SPGRE) sequence with relatively short TR. The spoiling suppresses (i.e., saturates) the signal from static tissue, and the TR is adjusted to be long enough so that a sufficient amount of blood can flow into the imaging plane creating contrast between flowing and static tissue. TOF images can be acquired as two-dimensional (2D) or three-dimensional (3D) images. The acquisition of a 3D volume receives signal simultaneously from the entire volume of interest while a 2D acquisition receives signal sequentially from a series of image slices, one slice at a time. In both cases, images are stored as a 3D image dataset. For physician review, a projection angiogram from the 3D dataset is created by projecting the reconstructed image values through the 3D image volume, most commonly using the maximum intensity projection (MIP) [3]. 3D techniques in general have the advantage of higher signal-to-noise ratio (SNR) and higher spatial resolution than 2D techniques at the expense of lower blood vessel signal because the blood remains in the slab for a significant fraction of the imaging time. 2D TOF images generally have higher contrast between blood and background tissue than 3D techniques but have lower spatial resolution due to slice thickness larger than 2 mm. To reduce blood signal saturation while maintaining the SNR of 3D acquisition, Parker et al. [4] developed the multiple overlapping thin 3D slab acquisition (MOTSA) technique. MOTSA acquisitions have the high SNR typical of 3D acquisition with improved vessel contrast common to 2D acquisitions.

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_2, © Springer Science+Business Media, LLC 2012

39

40

The goal of TOF MRA is to provide an accurate depiction of the blood within the vascular lumen (artery or vein) noninvasively and various TOF techniques have been successful in meeting this challenge. The best results are achieved when appropriate pulse sequence parameters are matched to each clinical application. In some cases, alterations to the standard SPGR pulse sequence, such as the use of magnetization transfer (MT) saturation, flow compensation, MOTSA [5], or TONE [6], can improve diagnostic accuracy. In general, TOF MRA techniques are clinically useful in the evaluation of small aneurysms, atherosclerosis, vasospasm, and inflammatory vasculitis.

S.-E. Kim and D.L. Parker

blood with plug flow, uniform velocity throughout the radius of the vessel, with speed v in a blood vessel with flow perpendicular to a slice to be imaged, as shown in Fig. 2.1. During the time, TR, between RF pulses, the fluid moves a distance, dz. Thus, in the time, TR, a length, dz, of nonexcited blood moves into the imaged slice. If dz is greater than slice thickness, z, the entire vessel segment within the slice is replaced by fresh inflowing blood shown as region I in Fig. 2.1a. If dz is less than z, then there will be sections of thickness dz that see one, then two, and three RF pulses as shown as region I, II, and II, respectively, in Fig. 2.1b. If the velocity of blood is exactly z/TR, the full slice thickness, z = vTR, will be completely replaced with fresh inflowing blood. The critical speed (2.1) is defined as:

Quantification of the Time-of-Flight Effect Vc ≡

z . TR

(2.1)

The TOF effect has been reviewed by Axel et al. [7], Gullberg et al. [8], and Nishimura [9] who modeled the signal on the basis of the Bloch equations for a variety of pulse sequence schemes for both plug and laminar flow. If the spins in stationary tissue experience a large number of RF pulses, the longitudinal magnetization of the stationary spins approaches a steady-state equilibrium value that is independent of position within the slice. However, when flowing spins, such as in blood, are flowing into and out of the slice, they may be subjected to fewer RF pulses resulting in a different steadystate magnetization. In general, the TOF effect leads to a diminished blood signal in spin echo imaging. However, in gradient echo images with short TR, the TOF effect results in inflow signal enhancement, which increases the signal from flowing blood relative to static tissue. As a simplistic example, consider

When the speed of blood is faster than Vc, the blood in the vessel that lies in the selected slice is completely refreshed by blood containing unsaturated spins. The fresh blood results in a higher signal relative to the stationary tissue signal since stationary tissue experiences many more RF pulses and is much more saturated. This effect is called wash-in, inflow enhancement or FRE. If v < Vc, partial saturation of the blood will begin to take place for distances into the slice greater than vTR. Fresh (fully magnetized) blood flowing into the imaging slice restores some of the signal intensity lost to partial saturation. In the simple example given above, the flow direction was perpendicular to the slice plane. Inflow enhancement of flow signal decreases for blood flow that transverses the blood vessel obliquely, and increases as

Fig. 2.1 TOF effects in the presence of plug flow with speed v in direction z. (a) When the velocity is higher than the critical velocity as given in (2.1), spins in the blood region I experience only one RF pulse and then exit the imaging slice before the next RF application. Spins in region II are not affected by the first excitation but by the next RF pulse.

(b) When the blood is moving slower than the critical velocity, the saturation effect can be understood by dividing the slice into multiple segments. Blood located in each segment experiences a different number of RF pulses, RF(n), resulting in different saturation effects depending on n

2

Time-of-Flight Angiography

41

the blood’s speed increases above Vc. Inflow enhancement also increases as slice thickness is reduced since Vc is proportional to the slice thickness, z. If the perpendicular component of velocity exceeds the critical velocity, there is complete inflow replacement in the imaging slice during each TR. A further increase in velocity results in no further increase in blood signal enhancement and might even begin to decrease the signal due to intravoxel phase dispersion. To extend this mathematical model to 3D TOF and MOTSA sequences, the slice thickness (z) can be replaced by the slab thickness, Nzdz. If the blood vessel is straight and perpendicular to the imaging slab, the critical velocity (2.2) is given by Vc =

N z dz TR

(2.2)

and for velocities greater than the critical velocity, the blood inside the slab is completely replaced by fresh blood flowing from the outside of the slab. For standard SPGR sequences, the TOF effect can be explained as the difference in signal saturation between flowing and static spins. Spins flowing into a slice may be less saturated than static spins resulting in angiographic contrast differences. The longitudinal magnetization of a stationary spin (i.e., after a series of RF pulses) with equilibrium magnetization, Mo, and longitudinal relaxation time, T1, subjected to the SPGR pulse sequence is given by [10]: M z ss =

M 0 (1 − e − TR/T1 ) , 1 − e − TR/T1 cos q

(2.3)

where q is the flip angle and TR is the repetition time of the sequence, T1 is the longitudinal relaxation time of blood which is almost same as T1 of tissue, and M0 is the equilibrium longitudinal magnetization. Before reaching the steady state, the longitudinal pulse (2.4) is given by: M z (n − ) = M z ss + (e − TR / T1 cos q )n −1 ( M 0 − M z ss ) n ≥ 1. (2.4) The transverse magnetization after the nth RF pulse (2.5) is given by M + (n) = M z (n − )sin qe − TE/T2 .

(2.5)

In cases when TR is much shorter than T1, as the number of RF pulses, n, increases, the contribution of the second term in (2.4) gets very small. Under these approximations, the transverse magnetization after the nth RF pulse (2.6) can be simplified as: M+ =

M 0 (1 − e − TR/T1 ) sin q e − TE/T2 . 1 − e − TR/T1 cosq

(2.6)

As the TR/T1 ratio decreases, the transverse magnetization given in (2.6) monotonically decreases. In other words, saturation of the MRI signal increases. The maximum signal occurs when the flip angle is the Ernst angle (qE = arcos(exp(−TR/T1)), which implies that saturation is dominant over the creation of transverse magnetization for q > qE. Inflowing fresh or fully magnetized blood entering into the imaging slice restores some of the signal intensity lost to partial saturation. Inflow enhancement increases with the blood velocity and as the imaging slice becomes perpendicular to the velocity direction. It also increases as slice thickness is reduced. As mentioned previously, if the perpendicular component of velocity exceeds the critical velocity, there is complete inflow replacement in the imaging slice during each TR. A further increase in velocity yields no further increase in the signal enhancement in blood and might even begin to decrease the signal due to intravoxel dephasing effects which are discussed in the next section. When the blood experiences a single RF excitation pulse in an SPRE sequence, the amount of inflow enhancement in the spoiled gradient sequence is given by the difference (2.7): M + (n) − M + = M0 sin q (e − TR/T1 cosq )n −1 ⎛ (1 − e − TR/T1 ) ⎞ − TE/T2* 1 − . ⎜⎝ (1 − e − TR/T1 cosq ) ⎟⎠ e

(2.7)

By using (2.7), the inflow enhancement effect can be simulated and quantified. Figure 2.2 shows computer simulations based on (2.7). In these simulations, we studied how the normalized signal of the flow enhancement effect (FRE) changes as the flip angle (q) increases for two cases: (a) TR/T1 = 0.02 and (b) when TR/T1 = 0.1. The normalized signal was obtained by dividing (2.6) by M0e−TE/T2*. The simulation demonstrates that FRE increases as n decreases when the TR and slice thickness are fixed. For a fixed TR and slice thickness, increasing n means faster flow velocity. This trend continues until v = Vmax, at which point the flow experiences only one RF pulse and the flow enhancement effect reaches a maximum. If n = 1, the maximum transverse magnetization (2.8) is given by: *

M + (n = 1) = M0 sin q e − TE/T2 .

(2.8)

Figure 2.2 shows that if there is no partial saturation (n = 1), the higher RF flip angle (90°) results in a maximum FRE. Comparing the curves, we see that the ratio of TR/T1 decreases and the flow enhancement intensity increases. When velocity v is smaller than the critical velocity, the flow enhancement effect can be modeled by subdivision of imaging slices into multiple compartments and summation of the geometric series shown in Fig. 2.1.

42

S.-E. Kim and D.L. Parker

Fig. 2.2 Plots of the flow enhancement effect (FRE) simulated from (2.8). The label n is the number of excitation RF pulses experienced by spins located in the imaging slices

In general, arterial velocity is an interesting function of the branching nature of the vascular bed. From the aorta to the capillaries, as the vascular branches, the vessel crosssection becomes progressively smaller. However, the number of branches increases in such a manner that the total cross-section increases as the branching increases. Thus, from geometry alone, arterial velocity decreases with distance from the heart. In general, the velocity remains sufficiently high for successful inflow enhancement until diameters smaller than 0.2 mm are reached. For vessels that can be seen with TOF MRA, the velocities become relatively constant (less pulsatile) for the more distal segments.

Phase Dispersion and Flow Compensation Signal intensity in MRI TOF imaging depends upon the imaging pulse sequence and the geometry and nature (velocity profile, pulsatility, etc.) of flow. Because of viscosity, flowing blood experiences frictional forces from the surrounding blood and vessel wall in addition to the force due to the pressure drop along the vessel. These forces generally result in a form of laminar flow. Laminar flow is characterized by a parabolic velocity profile, where the velocities in the center of the vessel are greater than those at the vessel wall. The magnetic field gradients used for spatial signal encoding in MRI impart a velocity-dependent phase at the time of signal acquisition. Because the velocity of blood varies considerably over dimensions that are much smaller than an image voxel, laminar flow can result in a range of phases for the signal generating spins within a voxel. If not properly compensated, this phase dispersion results in a loss in signal from flowing spins. High-velocity fluid motion through static vessels and arterial branches, such as the carotid bifurcation, produces complex flow patterns, including flow vortices (recirculation) and unsteady, nonrepetitive pulsatile, or turbulent flow.

Flow vortices and unsteady, turbulent flow increase phase dispersion of spin coherence due to the multiple directions of motion, acceleration, and higher order motions. Even simple pulsatile flow can result in magnitude and phase signal variations that are a complex function of time. These temporal changes can cause spatial misregistration of pulsatile flow in images. Venous flow is much less pulsatile than arterial flow. Flow velocities in the human body under normal conditions range from a few mm/s up to 180 cm/s. The phase dispersion among spins having the same constant velocities can be recovered with addition of the gradient waveform lobes known as “flow compensation” or “first-order gradient moment nulling.” If the spins are flowing through a gradient m(G), the time-dependent phase of flowing spins (f) at the location r (2.9) is given by f (t ) = ∫ g G (t ) × r (t )dt a f (t ) = r0 × ∫ g G (t )dt + v × ∫ g G (t )tdt + ∫ g G (t )t 2 dt +  2 a f (t ) = r0 × g × m 0 + v × g × m1 + × g × m2 +  (2.9) 2

where the expansion is obtained from a Taylor series expansion of r(t), r0 is the initial location of the spins, v is the velocity, a is an acceleration of the flow, and mj is the jth moment of the gradient. Equation (2.9) demonstrates that the behavior of the phase accumulated by the moving spins depends on the initial position, velocity, and gradient strength. If the velocity of each flowing spin is not constant, the phase accumulated after gradient application is not the same for each spin and varies with the velocity. This phase dispersion, if uncorrected, results in signal loss. To illustrate the effects of velocity on signal phase, an example of RF and gradient waveforms for a conventional SPGR sequence is shown in Fig. 2.3a. Slice selection (SS) is performed by the RF pulse in conjunction with the slice

2

Time-of-Flight Angiography

43

Fig. 2.3 3D SPGR pulse sequence for TOF acquisition without flow compensation (a) and with flow compensation in the slice selection and readout direction indicated by black arrows (b). The abbreviations

correspond to data acquisition (ADC), RF pulse (RF), readout (RO) gradient, phase-encoding (PE) gradient, and slice selection (SS) gradient

selection z gradient. The amplitude of the slice selection gradient is determined by the desired slice or slab thickness and the bandwidth of the RF pulse. A refocusing lobe is placed after the end of the RF pulse to compensate for the phase dispersion that occurs after the effective tipping of the magnetization into the transverse plane. The readout (RO) gradient, Gx, that is applied immediately after the RF pulse includes a dephasing lobe prior to signal acquisition so that the received signal is completely in phase at the time, TE, from the RF excitation. The amplitude of the readout gradient is determined by the desired resolution and the readout sampling bandwidth. Finally, a small gradient pulse is applied to the y gradient to create a variation in phase in the y direction during signal readout. This pulse sequence is repeated and the y gradient is stepped through values from a negative maximum to positive maximum. The SPGR works well for stationary tissues. The zeroth moment, m0, of both the slice selection and readout gradients is zero at the echo time, TE. The maximum signal occurs for the pulse sequence step, where the y phase-encoding gradient passes through the value, 0. At the time, TE, for this signal measurement, defined as the center of k-space, the stationary spins are completely in phase across the imaging volume. For the signal acquired when the y phase-encoding gradient has zero amplitude, all moments of the y gradient are zero. However, m1 and the higher gradient moments for the slice selection and readout gradients are not zero at time, TE. Thus, moving spins experience an additional phase shift due to motion during these gradients. Velocity dispersion due to laminar flow results in phase dispersion within a voxel and a net signal loss from flowing spins. Velocity-dependent phase dispersion can be corrected by a process known as velocity compensation or first-order gradient moment nulling (Fig. 2.3b). This is accomplished by adding additional area, positive and negative, to the gradient pulses used for to the slice selection and readout.

The magnitude of the lobes can then be adjusted to ensure that the moments of the waveform, when integrated over time, do not contribute to the velocity-dependent phase dispersion. In other words, flow compensation gradients “null” or zero the effects of phase dispersion. The amplitudes of the two compensation gradient lobes for each gradient are adjusted to null m0 and m1 at the times t = 0 and t = TE, respectively. In this manner, the phase of stationary and uniformly flowing spins is the same, independent of the velocity of flow velocity. An additional artifact can occur because of the difference in timing between the phase-encoding and readout gradients. If the time of the phase-encoding gradient is TP, there will be a position shift of v*(TE − TP) between the time the “y” coordinate is encoded and the “x” coordinate. Because the image is assumed to be recorded at time, TE, the blood may appear shifted from its actual position at that moment. When the flow is in the readout or phase-encoding direction, the shifted fluid appears to remain within the blood vessel. However, when the flow is diagonal within the readout and phase-encoding plane, the element of fluid appears to shift away from the vessel center with a distance of shift that is proportional to the flow velocity. If the flow is pulsatile, this shift will appear to change in size during the pulse sequence, resulting in blurring and ghosting artifacts. Such artifacts are often seen in the phase-encoding direction originating from the diagonal vessels in the circle of Willis. This last artifact can be eliminated by recognizing that the shift is due to a nonzero, velocity-dependent first moment of the phaseencoding gradient. By adding another lobe to the phaseencoding gradient, it is possible to step m0 through the values needed for imaging while at the same time nulling m1 at the time, TE. In 3D TOF, the rephrasing z gradient is usually used as a slice selection phase-encoding gradient. If first-order phase-encoding flow compensation is desired, both extra lobes on the z gradient can be stepped to achieve the desired m0 while maintaining a zero m1 at the time, TE.

44

Due to the additional lobes on the gradient waveforms, the minimum TE is increased. Spins experiencing higher orders of motion, such as constant acceleration, are not compensated by this technique. Acceleration terms, corresponding to the m2 moment, occur when flow is pulsatile or changes direction and in this case, some artifacts or signal loss can occur. Usually, this acceleration-caused signal loss is generally quite small and higher order flow compensation is usually not performed. However, it is possible to compensate for phase dispersion induced by constant acceleration by adding an additional lobe to the flow-compensating gradients. This is called second-order gradient motion rephasing or second-order gradient moment nulling, which compensates for constant acceleration. Changes in the rate of acceleration are described as jerk. Jerk can be compensated by third-order gradient moment nulling or third-order gradient motion rephasing with the addition of yet another lobe to the flow-compensating gradients. Signal loss due to these higher order terms is usually much smaller than the loss that would occur from the longer TE required to implement them such that most flow compensation techniques currently in use only compensate for first-order effects from constant velocity flow. If the TE of the pulse sequence can be made very short, all gradient moments will be small and there is no need for flow compensation. Thus, the phase dispersion in an ultrashort echo time sequence, such as a 3D radial acquisition technique, which requires no slice selection gradients and no readout prephasing, is minimal and there is no need for additional flow compensation [11].

S.-E. Kim and D.L. Parker

Flow-related enhancement can be reduced by the saturation of signal from the flowing spins, which depends on the number of pulses experienced by the blood, repetition rate, and blood T1 recovery time, and by intarvoxel phase dispersion.

Saturation effects are important in the setting of slow or in-plane flow, and can be minimized by using thinner slices and relatively long TR. Thus, 2D TOF [12] has the advantage in the setting of slow flow and is often the technique of choice for venous imaging. 2D TOF is also used in evaluation of cervical carotid stenosis to detect slow flow distal to a high-grade stenosis [13]. In 2D TOF imaging, a gradient echo sequence, usually 2D SPGR, is used to sequentially acquire a set of adjacent thin slices, generally 1–3-mm thick. TR, in a range of 20–30 ms, is used with a flip angle of 50–70°. If the velocity of flow is close to the critical velocity of 3–15 cm/s, as given in (2.1), the fluid in each slice experiences only a few RF pulses. The simulation shown in Fig. 2.2 demonstrates that the flow enhancement effect is maximized for flip angle in a range of 50–70°. The short TR and higher flip angle result in enhancement of the contrast between blood and background tissue, as there is insufficient time for longitudinal (T1) recovery of the static tissue magnetization and blood flowing into the slice is exposed to only one or two RF pulses for typical arterial velocities of 10–100 cm/s. The imaging plane is generally selected to be perpendicular to the flow direction, such as the axial plane for the carotid arteries. If arterial flow is principally along one axis and venous returns in the opposite direction (such as in the neck, with flow to the brain via the carotid arteries and return via the jugular veins), slice-selective saturation pulses can be used to eliminate the signal from flow in one direction. For example, to eliminate the signal from the jugular veins, a saturation RF pulse is applied superior to the axial slice (Fig. 2.4). To maintain saturation during the acquisition of all slices, this saturation pulse moves together with the axial slice as each subsequent slice is acquired. Blood flowing in the craniocaudal direction in the jugular vein is, thus, saturated before it flows into the imaging slice and produces no signal. Because the saturation pulse distinguishes arteries from veins only by the direction of blood flow, retrograde flow can result in the unwanted saturation of the desired vessel.

Fig. 2.4 Coronal MIP images of a 2D TOF acquisition of a healthy volunteer using the same acquisition parameters as Fig. 2.3. Coronal MIP image with no spatial saturation pulse (a) and with saturation pulse

superior to the imaging slices to suppress the venous signal (b) (images courtesy of Dr. Tae-Sub Chung, Gangnam Severance Hospital, Seoul, Korea)

2D TOF

2

Time-of-Flight Angiography

45

Fig. 2.5 2D TOF acquisition of a patient volunteer with atherosclerosis with TR = 28 ms, TE = 6.7 ms, flip angle = 50° , FOV = 16 cm, matrix = 256 × 256, slice thickness = 3 mm with 20% gap between slices.

(a) Coronal MIP; (b) axial source images; (c) 2D black blood T1w images

Transverse (axial) images acquired using 2D TOF in a patient with several severe stenoses just distal to the left and right bifurcation are shown in Fig. 2.5. Black blood 2D T1weighted images (Fig. 2.5c) confirm that atherosclerotic plaque caused the stenosis seen in 2D TOF. Figure 2.5a is obtained as the MIP through the “stack” of acquired image (Fig. 2.5b). Note that MIP images are not true angiograms but rather a projection produced from the source images. Any tissue with a short T1 relaxation time such as fat, depending on the exact scan technique used, may be hyperintense in the source images and thus on MIP images they may be potentially confused for flow. Further, the MIP image can be artifactual because only the brightest point along a projection line appears in the MIP image. It is possible that pathology or other important image details are masked behind brighter structures in the MIP image. Thus, when interpreting scans, it is important to review both the MIP and source images. The latter provide a means of assessing surrounding tissues and anatomy and often make identification of artifacts due to motion, signal loss, and the presence of fat easier. Pulsation and vessel motion induced by pulsation can cause artifacts in 2D TOF imaging. Flow compensation in the slice selection and frequency direction is helpful to reduce the inconsistent phase generated by pulsation [14]. Cardiac triggering can further reduce the artifact induced by pulsation. Fractional echo readout and a tailored excitation RF pulse are used to reduce the TE, since a shorter TE reduces the artifact induced by pulsation. Another drawback of 2D TOF is that complex flow, such as that seen distal to a stenosis, can have high-order motions (acceleration, jerk, etc.) and is not easily compensated for and can result in signal loss. Complex flow patterns specifically lead to signal loss in the region of a stenotic lesion and, thus, overestimation of the degree of stenosis. Although the slice thickness of

2D TOF sequences is usually less than 2 mm, 3D acquisitions can be used to produce images with thinner slices, which, together with the use of shorter TE, reduce the artifactual signal loss [15].

3D TOF Having intrinsically higher resolution and shorter echo times than 2D TOF, 3D TOF sequences suffer less from intravoxel phase dispersion. In 3D TOF imaging, a single slab of 3D slices is acquired. Like 2D TOF imaging, the slab is oriented to be perpendicular to the direction of flowing blood to ensure good inflow enhancement. SNR and the contrast between in-flowing blood and tissue in 3D TOF depend on the slab thickness and number of slices per slab. The transverse plane is often selected for a carotid or an intracranial 3D TOF acquisition to maximize the inflow enhancement effect; however, an oblique slab orientation is sometimes used, depending on the vessel geometry, to allow the desired imaging volume to be covered with fewer slices, which results in a shorter scan time. In 3D TOF acquisition, the blood inside the imaging slab generally experiences multiple RF pulses, so a smaller flip angle compared to that in 2D TOF is used to maximize the inflow enhancement effect. Further, spatial saturation is not usually applied in 3D TOF. Although spatial saturation could be applied in 3D TOF acquisition, it would increase the minimum TR and the venous signal in 3D TOF is already suppressed, except on a few entry slices, by the slow venous flow and thicker slab. Short TRs (~20 ms) in 3D TOF also result in reduced contrast between flowing and stationary tissue, including muscle and fat, which generally have a short T1 relaxation time compared with blood. Magnetization

46

S.-E. Kim and D.L. Parker

Fig. 2.6 (a) Schematic representation of TONE pulse. The flip angle of the TONE pulse increases as a function of distance in order to equalize the signal from flowing blood (indicated by arrow). (b) Coronal and sagittal MIP of 3D TOF acquisition of a volunteer at

1.5 T with and without TONE technique. The MIP image constructed with TONE shows the uniform vessel signal around edge (images courtesy of Dr. Tae-Sub Chung, Gangnam Severance Hospital, Seoul, Korea)

transfer is often used in 3D TOF to suppress the signal from the stationary tissue and to increase further the contrast between the flow and tissue. To minimize artifacts induced by turbulent or pulsatile flow, the echo time in 3D TOF is made as short as possible. Fractional echo readout and a short duration or asymmetric RF excitation pulse are used to achieve the shortest possible echo time. For situations where the signal from lipid can be a problem, an echo time can be selected, where the signal from lipid and water is out of phase, yielding better contrast between the two at the expense of a slight increase in problems due to the longer echo time. This out-of-phase echo time varies with the magnet field strength. When imaging with a thick 3D slab, the number of pulses experienced by the blood increases with distance into the slab. RF excitation that is uniform across the 3D slab results in greater inflow enhancement where the blood enters the slab, whereas distal slices in which blood has experienced

more RF pulses experience less flow enhancement signal. To compensate for the resulting variation in signal saturation, it has been found useful to make the RF excitation tip angle be a function of distance into the slab. For example, ramped RF excitation, where the tip angle is small at the slab entrance and increases with distance into the slab, is commonly used for 3D TOF to make the flowing blood signal across the imaging slab as uniform as possible [6, 16]. Ramped RF pulses, with spatially varying flip angle profiles parallel to the direction of flow, are often called tilted optimized nonsaturating excitation (TONE) pulses (Fig. 2.6). A variety of different flow compensation strategies are used in 3D TOF. Generally, flow compensation on the readout and slab selection gradient is used. In this case, the individual phase-encoding steps in the slice and phase-encoding gradient are not compensated. As described above, phaseencoding flow compensation can be used to compensate each step of both phase-encoding gradient waveforms. Although,

2

Time-of-Flight Angiography

47

Fig. 2.8 Axial MIP of 3D TOF acquisition of a healthy volunteer at 3.0 T (a) with linear view order and (b) with centric view order. Notice that the ghost artifact caused by pulsatile flow that propagates in the primary phase-encoding direction is reduced when centric view order is used

Fig. 2.7 Images obtained from four types of acquisition, from top to bottom. a–b, linear view ordering with (a) slice selection/frequencyencoding flow compensation and (b) three-directional flow compensation (3DFC). c–d, centric view ordering with (c) slice selection/ frequency-encoding flow compensation and (d) 3DFC. The cross-hatch artifacts arising from blood flow pulsations in the source images and other vessels near the circle of Willis are clearly evident in the two images in the top row. The artifacts are reduced by both 3DFC (second row of axial images) and centric view ordering acquisition order (third row), and the greatest reduction is observed when 3DFC is combined with centric view ordering acquisition (bottom row). More importantly, from the MIP images (right column), we note the change in apparent lumen diameter of the M1 segment (arrow) when using 3DFC. Without 3DFC, the obliquely oriented vessels have bright rims that are consistent with distortion or signal “pileup” (from Parker et al. [17], with permission)

compensation [20–22]. Ultrashort TE pulse sequences are also useful for scanning the brain in the presence of a metal clip or stent. In a 3D TOF acquisition, the order of k-space acquisition can be rearranged to reduce the prominence of artifacts induced by pulsatile flow (Fig. 2.8). A linear view order acquisition generally causes ghosts that are dominant along the phase-encoding direction [23]. An elliptical centric view order acquisition, which acquires the centric k-space first, spreads the ghosts evenly in the phase- and slice-encoding directions.

such phase-encoding flow compensation generally results in an increased echo time, it has the advantage of eliminating the misregistration artifact if blood is flowing obliquely to the frequency-encoding and one or both of the phase-encoding gradients (Fig. 2.7) [17]. To reduce the echo time in highresolution 3D TOF acquisition with three-direction flow compensation, a variable TE technique has been introduced [18, 19]. A variable TE technique in 3D TOF with threedirection flow compensation can minimize the TE at the center of k space. In this technique, k space of the 3D TOF is divided into several segments with different TE. Flowcompensation gradient lobes are calculated for each segment and the echo time at each segment is minimized. This results in a shorter echo time at the center of k space and a reduced flow-related signal void due to long echo times.

3D TOF with magnetization transfer has been employed for intracranial MRA to suppress the signal intensity of background brain tissue [24]. An MT pulse is a spectrally selective RF pulse that reduces the signal from tissue that has higher amounts of large molecules. Water molecules that are in contact with macromolecules generally move slowly and have a broad NMR resonance. Off-resonant excitation can saturate the magnetization on these water molecules. When these water molecules exchange with free water, the net magnetization in the tissue is reduced. Because blood has a lower concentration of macromolecules, off-resonance MT pulses can be used to selectively saturate the magnetization in station tissues with minimal effect on blood magnetization. In 3D TOF, MT is used to suppress the signal intensity of the brain parenchyma while leaving the signal from blood unaffected, thus improving smaller vessel visibility. Figure 2.9 shows a comparison of intracranial 3D TOF images acquired without and with the inclusion of MT pulses. An MT RF pulse generally requires a long pulse duration and higher RF power. Thus, MT pulse application causes increased TR and further increases the already long scan time of high-resolution 3D TOF acquisition. To increase the time

k-Space Sampling Strategies 3D MRA acquired using ultrashort TE sequences, such as projection or spiral acquisition, gives a minimal artifact from phase dispersion of flow across a voxel, even without flow

Magnetization Transfer

48

S.-E. Kim and D.L. Parker

Fig. 2.9 Axial MIP of 3D TOF acquisition of a volunteer at 1.5 T (a) without MT saturation and (b) with MT saturation. Notice that MIP with MT saturation demonstrates more small, distal, middle cerebral arteries’ detections (images courtesy of Dr. Tae-Sub Chung, Gangnam Severance Hospital, Seoul, Korea)

efficiency in 3D TOF, the majority of the MT effect is obtained by applying the MT pulses only around the center of k-space [25]. The use of MT saturation at high field strength becomes problematic because the SAR increases with the square of the RF transmission frequency, which is proportional to the main magnetic field strength. For this reason, MT saturation in TOF has been utilized only at 1.5 or 3.0 T. Because of the chemical shift between the fat and water, there is a small (3 ppm) difference in Larmor frequency between fat and water resonances, and a TE-dependent phase shift in their signal contributions to a given voxel. Choosing a TE at which fat and water are outphase can provide some fat suppression. However, the TE, where water and fat are out of phase, is greater than the minimum TE achieved by the pulse sequence and further increases the signal loss induced by intravoxel phase dispersion. As a general rule, the TE should be minimized at the cost of other considerations because of the rapid increase in intravoxel diphase dispersion that accompanies increased TE.

Multiple Overlapping Thin 3D Slab Acquisition 3D TOF MRA is susceptible to signal loss due to saturation of the moving spins in thick slabs. To reduce flow signal saturation while retaining the high spatial resolution, short echo times, and some of the SNR advantages of 3D techniques, a sequential acquisition of multiple 3D slabs has been developed. MOTSA has all the advantages of the single-volume 3D TOF techniques, and the use of overlapping thin slabs generally overcomes the problem of spin saturation. MOTSA is currently one of the most popular clinical 3D TOF applications [4]. This popularity is due in part to stronger in-flow enhancement and better vessel contrast-tonoise ratio properties of MOTSA compared to other techniques. Multiple thin slab acquisition reduces the signal saturation of slowly flowing blood compared to one thick single-slab acquisition, and overlapping slab acquisition

Fig. 2.10 Schematic representation of a MOTSA acquisition with three slabs and eight slices per slab. There are three overlap slices at each boundary between slabs

eliminates the signal void from each slab boundary region – often called the venetian blind artifact. The slices located in the overlapped region can be acquired twice with an entry slice from the first slab and an exit slice from the next slab [26]. This technique moderates the variation of the final MIP image intensity after taking the maximum of two images using a pixel-by-pixel comparison. Increasing the amount of slab overlap can obviously reduce the slab boundary artifact. Using a slab overlap of 50% can almost completely eliminate the artifact (Fig. 2.10). However, increasing overlap also increases the scan time per unit coverage in the slice direction. A sliding interleaved ky (SLINKY) sequence was proposed as an alternative technique to reduce the slab boundary artifact in multiple slab acquisition [27]. In a typical MOSTA data acquisition, the slab excitation is shifted

2

Time-of-Flight Angiography

after acquisition of all of ky and kz phase encodings. In SLINKY acquisition, the slab excitation is sliding every few views by one slab location. A designated partial set of the ky phase-encoding steps are collected and the acquisition of the partial set of the ky space is interleaved during continuous sliding of a slab along the slice selection direction [28]. SLINKY equalizes the flow enhancement effect across the entire slab dimension and eliminates the slab boundary artifacts while retaining good acquisition time efficiency compared with conventional MOTSA.

49

the extended scan time needed for phase encoding of the larger matrix is not clinically feasible. Parallel imaging (PI) techniques can accelerate scans at the cost of SNR and, thus, can achieve very-high-resolution MRA studies within a reasonable scan time [31]. For example, optimized coils coupled with PI techniques at 3.0 T can yield scan times similar to or shorter than those at 1.5 T. On this basis, intracranial TOF image quality improvements have previously been described in the transition from 1.5 to 3 T and 3 to even 7 T [32].

Summary TOF at High Magnetic Fields 3D TOF MRA is considered to be a safe, fully noninvasive imaging procedure with submillimeter spatial resolution. As such, it is routinely used to screen for cerebrovascular diseases, such as aneurysms and arteriovenous malformations. However, more subtle microvascular disease usually cannot be seen with the resolution capabilities of standard field strength MRA. Increased vessel contrast and spatial resolution are highly desirable for more sensitive detection of small aneurysms and vasculitis and for improved morphological characterization of larger aneurysms. The recent development of ultrahigh-field MRI scanners enables assessment of the cerebral arteries with a spatial resolution previously not achieved with standard MRI scanners. MRA at 7.0 T has shown superior contrast between blood and background tissue mainly because of the increased T1 recovery time of tissues and better suppression of the background signal relative to that of the blood vessels [29]. The increased SNR also yields improved visualization of the microvasculature of the human brain at high spatial resolution [30] (Fig. 2.11). However, at very high field strengths, the frequencyencoding dimension is limited by the TE-dependent artifacts of susceptibility-induced dephasing and pulsatile flow, and

In this discussion, we have presented a general overview of the basics of TOF MRA and provided discussion of the various trade-offs in image acquisition. Many of the concepts of TOF MRA were developed in the late 1980s and early 1990s, and hence are very mature. However, improvements in MRI techniques and technology have also led to progress in TOF MRA. These improvements include improved gradient capabilities, higher magnetic field strength, increased number and quality of RF channels and components, and new pulse sequences and acquisition techniques. In the past 20 years, gradient performance has increased from 1 mT/m and 20 T/m/s to over 40 mT/m and 200 T/m/s. Field strength in clinically used MRI scanners has now reached 3 T, and work is being performed at 7 T. The number of receiver channels has increased from 1 to as many as 128 on commercially available scanners, allowing the design of receiver coil arrays with more elements to allow parallel imaging to reduce image acquisition time and thereby reduce motion artifacts. Finally, new pulse sequences, including 3D radial acquisition, PROPELLER, and many others, have increased the flexibility available for novel TOF MRA techniques. Thus, as it is with many fields of MRI, TOF MRA continues to evolve in capability and ultimate utility.

Fig. 2.11 TOF MRA acquired at 7 T with a custom-built transmit/ receive head coil with eight stripline elements. TR/TE 21/3.4 ms; flip angle 40°; bandwidth 303 Hz/pixel; resolution 0.6 × 0.5 × 0.6 mm3 (noninterpolated); acquisition time 5 min 58 s; parallel imaging with

GRAPPA, reduction factor 2, 40 reference lines. The high spatial resolution enables nice depiction of fine, peripheral vessels (images courtesy of Mark Ladd of the Erwin L. Hahn Institute for MRI, Essen, Germany)

50

References 1. Suryan G. A time-of-flight method. Proc Indian Acad Sci Sect. 1959;A33:107. 2. Hinshaw WS, Bottomley PA, Holland GN. Radiographic thin section image of the human wrist by nuclear magnetic resonance. Nature. 1977;270:272–273. 3. Laub G. Displays for MR angiography. Magn Reson Med. 1990;14:222–229. 4. Parker DL, Yuan C, Blatter DD. MR angiography by multiple thin slab 3D acquisition. Magn Reson Med. 1991;17:434–451. 5. Davis WL, Warnock SH, Harnsberger HR, Parker DL, Chen CX. Intracranial MRA: single volume vs. multiple thin slab 3D time-offlight acquisition. J Comput Assist Tomogr. 1993;17:15–21. 6. Atkinson D, Brant-Zawadzki M, Gillan G, Purdy D, Laub G. Improved MR angiography: magnetization transfer suppression with variable flip angle excitation and increased resolution. Radiology. 1994;190:890–894. 7. Axel L, Shimakawa A, MacFall J. A time-of-flight method of measuring flow velocity by magnetic resonance imaging. Magn Reson Imaging. 1986;4:199–205. 8. Gullberg GT, Wehrli FW, Shimakawa A, Simons MA. MR vascular imaging with a fast gradient refocusing pulse sequence and reformatted images from transaxial sections. Radiology. 1987;165: 241–246. 9. Nishimura DG. Time-of-flight MR angiography. Magn Reson Med. 1990;14:194–201. 10. Vlaardingerbroek, MT, den Boer JA. Magnetic Resonance Imaging: Theory and Practice. Springer Verlag Telos; 1996:191–204. 11. Nielsen HT, Gold GE, Olcott EW, Pauly JM, Nishimura DG. Ultrashort echo-time 2D time-of-flight MR angiography using a halfpulse excitation. Magn Reson Med. 1999;41:591–599. 12. Keller PJ, Drayer BP, Fram EK, Williams KD, Dumoulin CL, Souza SP. MR angiography with two-dimensional acquisition and threedimensional display. Work in progress. Radiology. 1989;173: 527–532. 13. Heiserman JE, Drayer BP, Fram EK, et al. Carotid artery stenosis: clinical efficacy of two-dimensional time-of-flight MR angiography. Radiology. 1992;182:761–768. 14. Urchuk SN, Plewes DB. Mechanisms of flow-induced signal loss in MR angiography. J Magn Reson Imaging. 1992;2:453–462. 15. Keller PJ. Magnetic resonance angiography of the neck. Technical issues. Neuroimaging Clin N Am. 1996; 6:853–861. 16. Priatna A, Paschal CB. Variable-angle uniform signal excitation (VUSE) for three-dimensional time-of-flight MR angiography. J Magn Reson Imaging. 1995;5:421–427. 17. Parker DL, Goodrick KC, Roberts JA, et al. The need for phaseencoding flow compensation in high-resolution intracranial magnetic resonance angiography. J Magn Reson Imaging. 2003;18: 121–127.

S.-E. Kim and D.L. Parker 18. Song HK, Wehrli FW. Variable TE gradient and spin echo sequences for in vivo MR microscopy of short T2 species. Magn Reson Med. 1998;39:251–258. 19. Jeong EK, Parker DL, Tsuruda JS, Won JY. Reduction of flowrelated signal loss in flow-compensated 3D TOF MR angiography, using variable echo time (3D TOF-VTE). Magn Reson Med. 2002;48:667–676. 20. Schmalbrock P, Yuan C, Chakeres DW, Kohli J, Pelc NJ. Volume MR angiography: methods to achieve very short echo times. Radiology. 1990;175:861–865. 21. Glover GH, Lee AT. Motion artifacts in fMRI: comparison of 2DFT with PR and spiral scan methods. Magn Reson Med. 1995;33: 624–635. 22. Glover GH, Pauly MJ. Projection reconstruction techniques for reduction of motion effects in MRI. Magn Reson Med. 1992; 28:275–289. 23. Wilman AH, Riederer SJ, King BF, et al. Fluoroscopically triggered contrast-enhanced three-dimensional MR angiography with elliptical centric view order: application to the renal arteries. Radiology. 1997;205:137–146. 24. Dagirmanjian A, Ross JS, Obuchowski N, et al. High resolution, magnetization transfer saturation, variable flip angle, time-of-flight MRA in the detection of intracranial vascular stenoses. J Comput Assist Tomogr. 1995;19:700–706. 25. Parker DL, Buswell HR, Goodrich KC, Alexander AL, Keck N, Tsuruda JS. The application of magnetization transfer to MR angiography with reduced total power. Magn Reson Med. 1995;34: 283–286. 26. Blatter DD, Bahr AL, Parker DL, et al. Cervical carotid MR angiography with multiple overlapping thin-slab acquisition: comparison with conventional angiography. AJR Am J Roentgenol. 1993;161:1269–1277. 27. Liu K, Rutt BK. Sliding interleaved kY (SLINKY) acquisition: a novel 3D MRA technique with suppressed slab boundary artifact. J Magn Reson Imaging. 1998;8:903–911. 28. Liu K, Lee DH, Rutt BK. Systematic assessment and evaluation of sliding interleaved kY (SLINKY) acquisition for 3D MRA. J Magn Reson Imaging. 1998;8:912–923. 29. Kang CK, Park CW, Han JY, et al. Imaging and analysis of lenticulostriate arteries using 7.0-Tesla magnetic resonance angiography. Magn Reson Med. 2009;61:136–144. 30. Kang CK, Hong SM, Han JY, et al. Evaluation of MR angiography at 7.0 Tesla MRI using birdcage radio frequency coils with end caps. Magn Reson Med. 2008;60:330–338. 31. von Morze C, Purcell DD, Banerjee S, et al. High-resolution intracranial MRA at 7T using autocalibrating parallel imaging: initial experience in vascular disease patients. Magn Reson Imaging. 2008;26:1329–1333. 32. von Morze C, Xu D, Purcell DD, et al. Intracranial time-of-flight MR angiography at 7T with comparison to 3T. J Magn Reson Imaging. 2007;26:900–904.

3

Phase-Contrast MRI and Flow Quantification Bernd Jung and Michael Markl

Introduction Magnetic Resonance Imaging (MRI) techniques provide noninvasive methods for the accurate depiction of the vascular morphology and anatomy. In addition, the intrinsic motion sensitivity of MRI can be used to image vessels as in phasecontrast (PC) MR-angiography or to directly acquire and quantify blood flow [1–5]. Such techniques offer the unique possibility to acquire spatially registered functional information simultaneously with the morphological data within a single examination. Visualization and quantification of blood flow and tissue motion using PC MRI has been widely used in a number of applications. Characterization of the dynamic components of blood flow and cardiovascular function provide insight into normal and pathological cardiovascular physiology and are part of imaging protocols in daily clinical routine. In addition, the same underlying principles can be exploited to evaluate other dynamic processes in the human body such as CSF flow [6] or the cyclic wall motion of the heart [7].

ECG-triggered blood flow measurements as presented in 1986 by Nayler et al. [1]. The basic principle of PC MRI relies on the fact that the MR signal is inherently a vector quantity as illustrated in Fig. 3.1. The MR signal can thus be characterized not only by its magnitude but also by its phase f. As a consequence, a phase image can be reconstructed from any acquired MR data set in addition to the typical magnitude image reflecting the underlying anatomy. The MR-signal phase is affected by motion, which can be used to image vessels as in PC MR-angiography but also to quantify blood flow and motion of tissue. In the presence of a magnetic field gradient, the MR signal originating from a moving object presents with an additional signal phase, which is directly proportional to the velocity of the moving object. The measurement of MR signal phase, therefore, allows for motion quantification of the moving spins. The time-evolution of the MR-signal phase of the transversal magnetization of an object at the location r(t) can be derived from the spatial dependent Larmor frequency (3.1).     w L (r , t ) = g ( B0 + Δ B0 (r ) + r (t )G(t )).

(3.1)

Basic Principle of Phase-Contrast MRI The development of PC MRI initiated in the pre-MR imaging area with the first observation of coherent motion on the MR signal phase reported in 1954 by Carr and Purcell [8]. The basic concept was further developed for by Hahn who proposed to exploit the sensitivity of the MR signal phase to flow or motion for the measurement of sea water motion [9]. With the advent of MR imaging in the 1980s, the theory of phase velocity imaging was first described in 1982 [10] followed by first the presentation of MR velocity map images [11, 12] and clinical applications based on time-resolved B. Jung, PhD () • M. Markl, PhD Department of Radiology, Medical Physics, University Hospital Freiburg, Freiburg, Germany e-mail: [email protected]

In this equation, B0 denotes the static main magnetic field, DB0 reflects contributions by local field inhomogeneities, and G(t) is the time-dependent magnetic field gradient. The resulting signal phase f acquired at echo time TE can be derived by integrating the Larmor equation resulting in the following (3.2): TE     f (r ,TE) = f0 (r ) + g ∫ G(t )r (t )dt. t0

(3.2)

In this equation, the initial signal phase and the effects of field inhomogeneities are combined in the spatially dependent and typically unknown background phase f0(r). To evaluate the effect of flow or motion on the signal phase, the spatial location of a moving object or flowing spins can be    approximated in first order as r (t ) = r0 + v (t ) with constant

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_3, © Springer Science+Business Media, LLC 2012

51

52

B. Jung and M. Markl

Fig. 3.1 During an experiment, rf-excitation generates a precessing transverse magnetization vector, which induces the MR signal in a receiver coil surrounding the object under investigation. Following image acquisition anatomical (top right) and phase images (lower right) can be reconstructed from the raw data which reflect the local magnitude and phase of the transverse magnetization

velocity v, i.e., it is implied that tissue motion or blood flow does not change fast with respect to the temporal footprint (TE) of the data acquisition. Equation (3.2) then simplifies to the following (3.3):    TE   TE  f (r ,TE) = f 0(r ) + g r0 ∫ G(t )dt + g v ∫ G(t )tdt t0 t0     = f 0(r ) + g r0 M 0 + g vM1.

(3.3)

In addition to the background phase f0(r) the signal phase measured at TE is determined by the 0th and 1st order gradient moments M0 and M1. For PC-MRI, a bipolar gradient is typically used in order to quantify velocities, i.e., two gradients with identical amplitude G and duration Dt but with opposite polarity of the amplitude (see Fig. 3.2). The symmetry of this gradient scheme results in a vanishing 0th gradient moment M0 = 0. As a result, stationary spins no longer contribute to the signal phase and (3.3) reduces to the following (3.4):    f (r ,TE) = f0 (r ) + g vM1 .

(3.4)

The remaining non-zero first gradient moment M1 determines the velocity induced signal phase and moving spins experience an additional contribution in the signal phase which is proportional to the velocity of the moving spins as illustrated in Fig. 3.2. However, the measured signal phase f(r, TE) is still offset by the spatially dependent and unknown background phase f0(r). By subtracting the phase fref(r, TE) of an additionally performed reference scan the background phase can be eliminated. The resulting phase difference Df = f − fref finally yields the velocities of the moving spins (3.5):

Fig. 3.2 Bipolar velocity encoding gradient and temporal evolution of the MR signal phase for stationary spins and an object moving with constant velocity v

Δf   . v= g ΔM1

(3.5)

The calculated phase difference Df reflects changes in the MR signal phase associated with the motion component along the direction of the velocity encoding gradient. By adding a bipolar gradient to the MR pulse sequence along the read, phase, or slice direction and by subtracting a reference measurement without encoding gradient, flow or motion along these directions can be directly quantified. Such PC-MRI pulse sequences with single-direction velocity encoding are typically available on all commercial MR systems and permit the reconstruction of the following images (Fig. 3.3): • Magnitude image: the signal intensity represents the MR signal amplitude averaged over the two scans used for one-dimensional velocity encoding. • Complex difference image (flow encoded magnitude image): the signal intensity represents the absolute value of the phase shift and therefore contains no information about the flow direction (e.g., phase shifts of −170° and +170° shows the same pixel intensity). • Phase difference image: the signal intensity represents the phase angle DF between the reference and the motion encoded scan, i.e., the local velocities along the encoding direction. As is evident from (3.5), appropriate control of the first gradient moment (gradient strengths and duration) can be used to control flow or motion encoding. Owing to the 2p periodicity of the MR-signal phase, there exists an upper limit of the velocities that can be encoded using the PC principle. The velocity sensitivity “venc” refers to the maximum

3

Phase-Contrast MRI and Flow Quantification

53

Fig. 3.3 Images that are typically reconstructed from a 2D phase contrast acquisition with through-plane velocity encoding (bipolar encoding gradient along the slice direction). Note that the gray scale intensity for the phase difference image characterizes the motion

direction: the ascending aorta appears white due to the positive blood flow direction (flow foot to head) whereas the descending aorta appears black encoding to (negative) flow in the opposite direction (head to foot)

velocity that can be encoded by the bipolar gradient and is given by the following (3.6): venc =

±π  . γΔM1

(3.6)

Thus, the acquisition of moving objects requires the approximate knowledge of the blood flow or tissue velocities. If the first moment of the bipolar encoding gradient is too large, the signal phase can exceed values above ±p. Velocities positively exceeding the preselected venc of the PC-MRI acquisition result in velocity aliasing in the negative velocity range and vice versa. For example, a velocity of 120 cm/s measured with a venc of 100 cm/s will yield a velocity of −80 cm/s in the phase difference image. Figure 3.4 schematically illustrates the velocity aliasing effect for a parabolic flow profile. Figure 3.5 demonstrates the effect of different choices of venc with respect to the highest occurring velocities vmax in an in-vivo situation. No aliasing and a clear positive/negative blood flow velocities in the ascending/descending aorta can be observed with an optimal venc slightly exceeding the highest occurring velocities during peak systole (top left image). Aliasing in a few pixels can be seen in the ascending aorta if venc is chosen slightly smaller than vmax (lower left image). Aliasing in a considerably larger area of both ascending and descending aorta is evident if venc is far below the highest velocities (lower right image). Note also that if venc is chosen much higher than the systolic peak velocities in the aorta, no aliasing occurs but the velocity noise is substantially increased as indicated by the gray color within both ascending and descending aorta (top right image). It can be shown that the velocity noise s linearly increases with an increasing velocity sensitivity venc and is inversely related to the signal-to-noise ratio (SNR) in the corresponding magnitude images as follows (3.7) [13].

Fig. 3.4 Velocity aliasing for a parabolic flow profile. If the velocity sensitivity venc is smaller compared to the highest occurring flow velocities aliasing towards inverted velocity values occurs

Fig. 3.5 PC-MRI with through-plane velocity encoding in the ascending and descending aorta. The individual phase difference images show the encoded blood flow velocities during peak systole in the same subject for four different velocity sensitivities (venc)

54

B. Jung and M. Markl

Fig. 3.6 PC gradient echo pulse sequences for one-directional velocity encoding along the slice direction. The gray shaded bipolar velocity encoding gradient is added to a standard gradient echo sequence with

s=

2 venc . p SNR

(3.7)

To avoid velocity aliasing while maintaining a low velocity noise, it is thus important to select the velocity sensitivity higher than the expected vmax but as small as possible to achieve optimal velocity-to-noise performance for velocity quantification.

Phase-Contrast Pulse Sequences Typical applications of PC-MRI for measurement of blood flow in the aorta or cranial vessels are limited in scan time and therefore based on fast gradient echo pulse sequences. PC pulse sequences are formed by adding the bipolar velocity encoding gradient lobes to a gradient echo sequence as illustrated in Fig. 3.6 for velocity encoding along the slice (through-plane) direction. The addition of bipolar gradients for motion encoding prolongs the echo time TE which may cause flow artifacts, especially in the presence of pulsatile flow with rapid changes in velocity across a voxel. Therefore, a flow compensated acquisition (M1 = 0) is often used as the reference scan, i.e., an additional gradient is applied that refocus the MR-signal phase at TE independent from the velocity of moving spins (Fig. 3.6, left). For PC measurements a short echo time TE is desirable to minimize the signal loss due to T2* decay and reduce susceptibility and flow artifacts. However, the addition of bipolar gradients and use of flow compensated gradient waveforms results in relatively long echo times (in the order of 4–10 ms)

flow compensation along the slice direction. For improved performance, the bipolar gradient can be combined with the imaging gradient resulting in more time efficient gradient waveforms and reduced echo time TE

compared to a conventional gradient sequence as used for example used in contrast-enhanced measurements (1–3 ms). Therefore, several methods to reduce TE in PC-MRI have been introduced [14, 15]. One method that is implemented on most modern clinical MR systems utilizes the combination of the gradient waveforms of the reference scan with the bipolar gradients used for motion encoding, i.e., the flow encoding gradients are added onto the imaging gradients as depicted in Fig. 3.6 (right) resulting gradient waveforms with improved time efficiency. A further reduction of TE can be achieved by the implementation of two-sided flow encoding [14]. The difference between the two-sided flow encoding and conventional flow encoding with reference scan and flow encoded scan (also called one-sided flow encoding) is illustrated in Fig. 3.7. For one-sided flow encoding, the reference scan must be acquired with the same TE as the flow encoded scan to allow for a consistent subtraction of the background phase. For twosided flow encoding the moment DM1 of the flow encoding gradient is split in equal parts for both scans. The first scan (“up-case”) encodes the moment DM1/2, the second scan (“down-case”) −DM1/2 such that the difference yields the total encoding moment DM1. The corresponding gradients to encode DM1/2 require considerably a shorter duration and thus permit a noticeable reduction in echo time TE. Note that according to (3.5), the PC principle only permits the encoding and measurement of the motion component along the direction of the bipolar encoding gradient as shown in the pulse sequences described above for velocity encoding along the through-plane direction. To encode velocities along all three spatial dimensions at least four scans have to be acquired: one reference scan and three velocity encoded scans with different bipolar motion

3

Phase-Contrast MRI and Flow Quantification

55

Fig. 3.7 Different encoding strategies for one-directional velocity encoding along the slice direction

encoding gradients along read, phase, and slice direction but otherwise identical acquisition parameters [15]. This property results in relatively long scan times of PC-MRI which is one of the major drawbacks of PC measurements.

Implementation and Clinical Protocols To synchronize flow or motion sensitive measurements with periodic tissue motion or pulsatile flow, data acquisition is synchronized with the cardiac cycle and a k-space segmented pulse sequence scheme is used as illustrated in Fig. 3.8 [16]. The ECG signal is used to gate the MR measurement to consistently capture a series of time frame in the cardiac cycle for each heartbeat. Since the MR acquisition is not sufficiently fast to measure all required data during a single heartbeat, the periodic movements or flow are assessed by

Fig. 3.8 Time-resolved ECG synchronized CINE PC-MRI data acquisitions. One-directional through-plane velocity encoding requires the execution of two differently velocity encoded scan for each raw data line

56

reconstructing images from data acquired over several cardiac cycles. For each heartbeat and time frame, only a subset (NSeg) of all required (Ny) phase encoding steps are measured (k-space segmentation). The procedure is repeated for a different subset until the full raw data set is acquired for all time frames. As a result, time resolved (CINE) images can be derived depicting the dynamics of periodic physiological processes during the cardiac cycle. PC acquisitions can be combined with the CINE principle by successively acquiring the different velocity encoded scans for each raw data line (interleaved velocity encoding) [1, 3–5]. For an n-directionally encoded PC-MRI measurement n + 1 raw data lines have to be acquired for each phase encoding step. For a given repetition time TR and cardiac period of TECG, different imaging protocols can be constructed based on a trade-off between temporal resolution (Dt), spatial resolution (Ny phase encoding lines per slice) and total acquisition time Tacq. The selection of the number of phase encoding lines NSeg then determines the temporal resolution Dt = n TR NSeg and a total scan time Tacq = Ny/NSeg TECG of the PC CINE acquisition. From different PC protocols, 2D CINE PC with throughplane velocity encoding is mostly applied in the clinical routine. For a typical TR on the order of 5–10 ms measurements can be performed with temporal resolutions of up to 10–20 ms. Such PC-MRI protocols can be employed for volume flow measurements through a vessel. It should be noted that the CINE acquisition scheme is based on a periodic cardiac cycle. The regularity of the heart rate is important to ensure that data are consistently acquired for all time frames and heartbeats. An irregular heart rate would result in different entries in the data matrix acquired at different period of the heart cycle and thus induce artifacts. To avoid such inaccuracies, arrhythmia rejection algorithms have been introduced in clinical CINE PC-MRI protocols. Thoracic applications of PC-MRI are affected by breathing motions and should be performed at a constant breathing position to avoid respiration artifacts and errors in velocity encoding. For fast acquisitions, this can be realized by breath holding for the time of the acquisition [17]. However, the duration of the total acquisition is then limited by the breathhold capacity of the subject therewith limiting the spatial and/or temporal resolution of the acquisition. Breath hold is often not feasible for longer acquisitions such as high resolution or multidirectional PC-MRI. One possible remedy is to monitor the breathing pattern by a fast one dimensional MR scan without phase encoding (navigator) positioned at the interface between the liver and lung [18–20]. The large contrast between the liver tissue and air is easily tracked and can be used to detect the breathing position and to gate the data acquisition to the respiration cycle.

B. Jung and M. Markl

Applications of 2D CINE PC-MRI Most MR systems offer imaging protocols permitting both in-plane and through-plane velocity encoding. In-plane velocity mapping can be performed in one or two directions and is mainly used for visualizing flow patterns within the imaging plane such as jets through stenosed vessels or valves. The velocity encoding direction should be parallel to the direction of the flow in the vessel of interest. An in-plane acquisition of a high-velocity jet can be helpful for planning a subsequent through-plane scan on the jet location with maximum velocities, especially for the assessment of mitral or tricuspid valve regurgitation. The most common applications for CINE PC MRI focus on valvular, congenital, and other heart diseases. Flow quantification with through-plane velocity encoding can be used to estimate the fraction of regurgitant flow in case of valve defects or to quantify global cardiac function such a stroke volume and cardiac output [5, 21]. In the presence of pulmonary–systemic shunts, the pulmonary/systemic flow ratio can be determined and used to assess the severity of the disease [22–24]. For the quantification of left or right ventricular stroke volumes, cardiac output and regurgitant fractions throughplane velocity encoding is used and the temporal resolution is chosen such that cardiac cycle is covered by approximately 15–20 time frames. In combination with a spatial resolution of about 1–2 mm, these scans are typically performed during breath-hold. Typical velocity sensitivities are venc = 150 cm/s for aortic flow measurements and venc = 100 cm/s for flow in the pulmonary artery. However, it should be noted that in the case of a valve or vessel stenosis blood flow velocities can reach values of up to 800 cm/s. To avoid velocity aliasing the PC-MRI scan should be repeated using a higher velocity sensitivity venc until the phase aliasing has disappeared to allow for an appropriate flow analysis. Following successful data acquisition, the PC-MR data can be used for flow quantification as illustrated in Fig. 3.9 showing the systolic magnitude and phase difference image in the ascending and descending aorta. Vessel lumen segmentation of the ascending aorta for all time-frames throughout the cardiac cycle can be used to obtain the flow-time curve. In patients with aortic valve insufficiency, a slight backflow due to a mild aortic valve insufficiency can be observed that is not present for normal volunteers. The mild increase of flow after the closing of the aortic valve is caused by a contraction of the aortic wall, called the “Windkessel”-effect. Figure 3.10 shows systolic magnitude and the phase difference images in the ascending aorta at the level of the aortic valve for a patient with an aortic valve stenosis. At peak systole and maximum valve opening, blood flow is confined to a small area compared to the full diameter of the ascending

3

Phase-Contrast MRI and Flow Quantification

Fig. 3.9 Blood flow quantification in the aorta using 2D PC-MRI and through-plane velocity encoding. Left: Following data acquisition, the vessel contours of the ascending aorta (AAo) are identified by lumen segmentation to quantify time-resolved blood flow.

57

Right: Flow-time curves in the ascending aorta above the aortic valve in normal controls and two patients. In patients, the different extent of early diastolic regurgitant flow (arrows) indicates mild aortic valve insufficiency

Fig. 3.10 2D PC-MRI and through-plane velocity encoding in a patient with aortic valve stenosis and a dilated ascending aorta (Ao)

aorta as indicated by the dark color in the phase difference image (arrow). A further clinical application provided by CINE PC-MRI is provided by the evaluation of shunts reflecting defects in the septum of the heart that allow transfer of blood from the arterial to the venous system or vice versa (from the system with the higher to the lower pressure) [23]. To quantify the shunt volume two CINE flow measurements are performed. PC-MRI in the ascending aorta and in the pulmonary artery is used to quantify the left and right ventricular stroke volumes QS and QP, respectively. The ratio QP /QS represents the direction of shunt and the shunt volume. For QP /QS = 1 no shunt is present, if QP /QS > 1 a shunt from the aortic circulation to the pulmonary circulation is present and vice versa for QP /QS < 1.

Phase-Contrast Angiography While most clinical angiographic applications rely on the application of Gd contrast agent, 3D PC MR angiography (PC-MRA) based on velocity encoded 3D MRI with three-

directional encoding has proven to be a useful alternative. PC-MRA can provide detailed information on vascular geometry and may offer additional information on flow direction [25–28]. Several strategies exist for calculating a PC-MR angiogram from the data. A possible combination of velocity and magnitude data is shown in Fig. 3.11. The three-directional encoded PC-MRI data is used to calculate absolute velocities |v| for each image voxel which are additionally weighted by the magnitude images for suppression of background signal. Alternatively, complex difference images can be calculated directly from the raw data – as described above – for each individual velocity encoding direction and combined in a sum of squares sense [25, 26]. It is important to note that the depiction of the vessels is determined by the choice of the velocity sensitivity venc. The maximum intensity projections (MIP) of the cranial vessels in Fig. 3.12 demonstrate the effect of the velocity sensitivity on PC-MRA data. Small vessels with slow flow are more clearly visible using a smaller venc-factor (Fig. 3.12c, d). Larger vessels with higher blood flow velocities demonstrate a decreased signal due to (multiple)

58

B. Jung and M. Markl

Fig. 3.11 Derivation of 3D PC-MRA from PC-MRI data with three-directional velocity encoding and maximum intensity projects of PC-MRA data in the head and thorax

Flow-Sensitive 4D MRI

Fig. 3.12 Influence of the velocity sensitivity (venv) of PC MRA. PC MRA Maximum Intensity Projection (MIP) images acquired with venc’s of 80 (a), 40 (b), 20 (c), and 10 (d) cm/s. High venc images highlight the arterial system. As the venc is decreased the visualization of the high velocity arterial system is compromised, while, the visualization of the veins is improved. This is especially true for the slowflowing veins such as the internal cerebral vein (arrows) (Images courtesy of Kevin Johnson and Oliver Wieben, Departments of Medical Physics and Radiology, University of Wisconsin Madison, USA)

aliasing of the signal phase. For increased venc-factors, the visualization of larger arterial vessels is improved and venous signal is suppressed. The quality of the depicted vessels is best if the chosen flow sensitivity represents the physiological situation of the vessel segment of interest. Most PC-MRA implementations used nongated data acquisition which can result in artifacts for pulsatile blood flow. Further drawbacks of the PC-MRA method are long scan times and lack of respiration control, which limited most applications of 3D PC-MRA to static regions with low pulsatile flow such as the cranial vessels. In this context, new implementations based on time-resolved accelerated acquisitions or radial imaging strategies are promising and provide a considerable improved performance regarding flow artifacts, total scan time, and spatial resolution without the need for contrast agent administration [29, 30].

Traditionally, MRI imaging of flow is accomplished using methods that resolve two spatial dimensions (2D) in individual slices. Alternatively, 3D spatial encoding offers the possibility of isotropic high spatial resolution and thus the ability to measure and visualize the temporal evolution of complex flow and motion patterns in a 3D-volume. In this context, ECG synchronized flow-sensitive 3D MRI using three-directional velocity encoding (also termed “flow-sensitive 4D MRI” or “time-resolved 3D velocity mapping”) can be employed to detect and visualize global and local blood flow characteristics in entire targeted vascular regions (aorta, cranial arteries, carotid arteries, etc.) [31, 32]. For thoracic or abdominal applications, the data acquisition needs to be synchronized with the subject’s respiration. Owing to the acquisition of at least four data sets for three-directional velocity encoding, PC MRI inherits a trade-off between spatial/temporal resolution and total scan time. Nevertheless, a number of studies have reported methodological improvements (parallel imaging, adaptive navigator gating with increased efficiency, time-optimized velocity encoding gradients, etc.) permitting the acquisition of flow-sensitive 4D MRI data within the entire aorta or other arterial systems within reasonable scan times of the order of 10–15 min. For the subsequent analysis and visualization of complex, three-directional blood flow within a 3D volume, various visualization tools including 2D vector-fields, 3D streamlines, and time-resolved 3D particle traces have been proposed [33]. A representative data acquisition and data analysis strategy for the 3D visualization of blood flow characteristics in the thoracic aorta is illustrated in Fig. 3.13. Several groups have reported advances in the application of flow-sensitive 4D MRI including the analysis of blood flow through artificial valves [34], ventricular and atrial flow patterns [35, 36], blood flow characteristics in

3

Phase-Contrast MRI and Flow Quantification

59

Fig. 3.13 (a) Data acquisition and visualization of vascular geometry and 3D hemodynamics for flow-sensitive 4D MRI in the aorta using navigator gating and prospective ECG gating. (b) The resulting raw data comprises information along all 3 spatial dimensions, 3 velocity

directions, and time in the cardiac cycle. (c) 3D blood flow visualization permits the depiction of time-resolved 3D vascular hemodynamics within the entire thoracic aorta

the thoracic aorta [37–39], peripheral vessels [38], carotid arteries [40], large intracranial arteries [41], as well as flow in the pulmonary and venous systems [42]. Figure 3.14 illustrates the potential of the methods to assess, visualize and quantify flow characteristics in different vascular regions in the body. Recent studies indicate the potential of flow-sensitive 4D-MRI for the detailed visualization of complex flow patterns associated with vascular pathologies. The complete coverage of the vascular region of interest permits the assessment of spatial and temporal characteristics of 3D blood flow as shown in Fig. 3.15 for a patient with aortic valve stenosis and a dilatation of the ascending aorta. The disturbed complex flow patterns in the aorta illustrate the effect of the pathological valve function on aortic hemodynamics and may help to improve the understanding of the link between valve dysfunction and aortic dilatation often sees in such patients. Since flow-sensitive 4D MRI data reflects the true underlying time-resolved blood flow velocity vector field, it is possible to quantify the directly measured (e.g., flow rates) or derived parameters such as pressure difference maps [43],

wall sheer stress [44], pulse wave velocity [45], and others. Findings in recently reported studies combining the complete spatiotemporal coverage of flow-sensitive 4D MRI and advanced quantification strategies are promising and may help to define new clinical markers for the improved characterization of cardiovascular disease. Examples include relative pressure mapping within the heart and aorta [46] or renal arteries [27], wall shear stress analysis in the thoracic aorta [47], or assessment of onset and dynamics of regional turbulent kinetic energy in the aorta [48]. A disadvantage of PC MRI is related to the need for multiple acquisitions for encoding a single velocity direction, resulting in long scan times. New methods based on the combination of PC MRI and fast sampling strategies, e.g., radial imaging with 3D PC-VIPR as shown in Fig. 3.16, have been reported and are promising for further reduction in total scan time and/or increased spatial or temporal resolution [29, 49]. In this context, a number of studies have already demonstrated the potential of radial imaging techniques for the assessment of vascular function with increased efficiency compared to conventional methods [27, 50, 51].

60

B. Jung and M. Markl

In addition, the total acquisition time or temporal and spatial resolution associated with a specific MR technique may be further improved by using new spatiotemporal imaging acceleration [52, 53].

Sources of Errors Over the past three decades numerous studies have systematically validated the quantitative and qualitative analysis of blood flow using in-vitro and in-vivo experiments. It has to be noted, however, that there are a number of sources of inaccuracies in PC-MRI, which can results in errors in the measured velocities. The major sources of errors include eddy current effects [54], Maxwell terms [55], and gradient field distortions [56]. Appropriate correction strategies should, thus, be applied to ensure accurate flow quantification using PC-MRI data.

Maxwell Terms

Fig. 3.14 Flow-sensitive 4D MRI in different vascular territories. (a) 3D flow visualization in the large intracranial arteries in the circle of Willis using systolic streamlines. The tortuous routes of the different vascular segments and changes in regional velocities can clearly be appreciated (CCA/ICA/ECA: common/internal/external carotid artery). (b) Time-resolved 3D particle traces during peak flow in the carotid bifurcation. Note the typical helical flow pattern in the ICA bulb. (c) Blow flow characteristics in the distal abdominal aorta (Ao) and peripheral arteries in the left and right leg (IA iliac artery). Quantitative flowtime curves demonstrating typical triphasic pulsatile flow can be derived from the same data as shown for the distal abdominal aorta

It can be shown that, as a consequence of the Maxwell equations, a magnetic field gradient cannot be switched on, without generating additional unwanted nonlinear magnetic fields. These concomitant gradient terms or Maxwell’s terms arise whenever a gradient is activated. They are described by Maxwell’s equations and result in magnetic fields with nonlinear spatial dependences. Concomitant gradient crossterms arise when the longitudinal gradient Gz is activated with a transverse gradient (Gx or Gy). They result in phase errors and hence are affecting the velocity measurements using PC MRI. It is possible to reduce or correct these effects

Fig. 3.15 Flow-sensitive 3D MRI and visualization using systolic 3D streamlines in a patient with aortic stenosis and a dilated ascending aorta (AAo). Note the flow acceleration through the stenotic aor-

tic valve which results in a flow jet directed towards the outer wall of the AAo and considerable vortex formation in a large segment of the AAo

3

Phase-Contrast MRI and Flow Quantification

61

during postprocessing by fitting a polynomial onto the data as described by Bernstein et al. [55]. Corrections can be performed during image reconstruction, based on the knowledge of the gradient waveforms that can be used to derive the correction factors needed for Maxwell term compensation.

reflected in the strength or direction in the first order gradient moments and thus the velocity encoding. Therefore, gradient field distortions can lead to considerable deviations between the designed and the actual encoding in PC-MRI. The true gradient field demonstrates not only deviations from the nominal gradient strength but also from the intended gradient direction and thus affects not only the magnitude of encoded velocities but also velocity encoding direction. Dependent upon the spatial location of the velocity measurements, errors in velocity magnitude can be as high as 60%, while errors in the velocity encoding direction can be up to 45°. The true magnitude and direction of the underlying velocities can be recovered from the phase difference images by a generalized PC velocity reconstruction which requires the measurement of full three-directional velocity information. A generalized reconstruction of velocities can then applied using a matrix formalism that includes relative gradient field deviations derived from a theoretical model of local gradient field nonuniformity. In addition, approximate solutions for the correction of one-directional velocity encoding are available [56]. The gradient field model needed for the corrections is vendor specific and depends on the design on the gradient systems. Significant improvement in velocity quantification can be achieved by using the known field distortions to correct the measured phase shifts. Results of a phantom experiment with one-directional velocity encoding illustrating the effect of gradient field distortions are shown in Fig. 3.17. The relative velocity encoding errors predicted by the gradient field model are illustrated by the surface plot which demonstrates the relative deviations from nominal z-gradient strength a coronal plane transecting the flow phantom in longitudinal direction. Since steady flow was used for all experiments, the true mean through-plane velocities are expected to be constant as a function of spatial location (z) along the tube. The deviation of the measured velocities demonstrates the effect of gradient field nonuniformity and corresponded well to the predicted errors predicted by the gradient field model. Although errors associated with gradient field distortions have been known for some time, corrections for these inaccuracies are to date often not part of the standard image reconstruction process and mostly absent from commercial systems.

Gradient Field Nonlinearities

Eddy Currents

It is well known that nonuniformity in magnetic field gradients can cause significant image warping and require correction. In PC-MRI, these imperfections introduce errors in velocity measurements by affecting the first moments used to encode flow or motion [56, 57]. Any error in strength or direction of the local gradient from its ideal value is directly

Switching of imaging and encoding gradients in PC-MRI results in changes in magnetic flux, thereby inducing eddy currents in the conducting parts of the scanner system (coils) or in the patient. These eddy currents can cause alterations of the desired gradient strengths and duration and thus result in spatially varying phase errors in the MR images [58]. For

Fig. 3.16 PC VIPR is based on spoiled gradient echo sequence with bipolar flow encoding gradients (a) and a true 3D radial trajectory (b). Visualization of the flow conditions obtained with a PC VIPR scan of a 18 month-old male with pulmonary venolobar syndrome. Particle traces in various cardiac phases obtained from the PC VIPR scan from a posterior view (c). The vascular system is color coded in blue (veins) and red (arteries). (d) Atrial defect measured at 1.34 L/min. (e) Analomous Pulmonary Venous Return “Scimitar Vein” Flow – 0.42 L/min. (f) Abnormal Systemic Artery showing flow to the right lung (Images courtesy of Kevin Johnson and Oliver Wieben, Departments of Medical Physics and Radiology, University of Wisconsin Madison, USA)

62

Fig. 3.17 2D PC MRI with one-directional velocity encoding: Demonstration of the effect of gradient field inhomogeneities on velocity encoding. Left: PC-MRI velocity measurements in a tube phantom with constant laminar flow parallel to the direction of the main magnetic field (B0, z). Mid: Simulated and normalized z-gradient fields (left) depict the spatial variation of the relative deviation from the ideal

B. Jung and M. Markl

gradient strength in a coronal plane at the phantom locations. Right: Measured through plane velocities are compared to predicted measurement errors (thick black lines). Since the measured errors are in agreement with predictions, a correction can be performed based on the model of the encoding gradient fields

Fig. 3.19 Schematic illustration of a typical correction scheme for PC-MRI data. While Maxell and gradient field corrections can be automatically performed during image reconstruction, eddy current effects often require user interaction to identify regions for phase offset estimation (see also Fig. 3.18)

Fig. 3.18 Image-based eddy current correction of 2D PC-MRI data. The application of bipolar velocity encoding gradients can results in eddy current induced phase shifts as illustrated for a static phantom demonstrating which should ideally include constant and zero phase differences inside the entire object. The phase difference gradient is a result of spatially dependent eddy current induced phase offsets. For in-vivo PC-MRI (right) eddy current effects can be corrected by subtracting eddy current offsets estimated from regions of interest in static tissue

PC-MRI, the different gradient waveforms that are used for the subsequent velocity encodes lead to different eddy current induced phase changes in the phase images of each velocity encoded acquisition. As a result, subtraction of phase images does not eliminate errors related eddy currents and additional data processing is needed to restore the original velocity encoded signal phase. Several correction strategies have been proposed and are typically based on the subtraction of an estimation of the spatially varying eddy current induced phases changes. The spatial variation of the phase difference in regions containing static tissue is used to calculate the eddy current induced offset for the entire image. By fitting a plane or higher order polynomial to the phase difference data, the phase offset characteristics can be estimated and subsequently used to correct the entire image by subtraction (see Fig. 3.18) [54].

Compensation for Maxwell terms and for gradient field nonlinearities do typically not require user interaction. Maxwell corrections can be performed during image reconstruction, based on the knowledge of the gradient waveforms in the PC-MRI pulse sequence used for data acquisition. Similarly, gradient field models describing deviations between the designed and the actual velocity encoding gradients can be employed to automatically correct for gradient field nonlinearities during image reconstruction. For eddy current correction, automatic correction algorithms have been reported. However, for a reliable estimation of background offsets, user inter action may often be required to correctly identify static background signal (Fig. 3.19). It is of note that it is common to all sources of inaccuracies, that error in velocity encoding strength increase with increasing distance from the isocenter of the magnet. Velocity measurements with a single slice placed at or near the center of the magnet and with the vessel of interest close to the center of the FOV are therefore largely insensitive to encoding errors. The situation differs considerably, however, if multiple slices are acquired within a single acquisition or if blood flow is to be analyzed within a larger plane or a 3D imaging volume. For off-center flow quantification corrections for all of the above sources of inaccuracies should be performed to ensure reliable quantitative analysis of flow and derived parameters.

3

Phase-Contrast MRI and Flow Quantification

Acceleration Errors The MR signal phase is not only sensitive to the blood or tissue velocity, but it depends as well on the acceleration and higher-order terms. All those higher order terms are neglected when approximation blood flow or tissue motion in first order as (3.3). Although this assumption is reasonable in most situations, it may induce an error on the measured velocities. This effect may be reduced by optimizing the shape of the gradient waveforms or reducing TE (Oshinski et al. [59]). Although it is difficult to fully correct for this effect, it is possible to further reduce it at the cost of longer imaging times by integrating acceleration encoding [60] or Fourier encoding (Firmin et al. [61]).

References 1. Nayler GL, Firmin DN, Longmore DB. Blood flow imaging by cine magnetic resonance. J Comput Assist Tomogr. 1986;10:715–722. 2. Dumoulin CL, Souza SP, Darrow RD, Pelc NJ, Adams WJ, Ash SA. Simultaneous acquisition of phase-contrast angiograms and stationary-tissue images with Hadamard encoding of flow-induced phase shifts. J Magn Reson Imaging. 1991;1:399–404. 3. Pelc NJ, Herfkens RJ, Shimakawa A, Enzmann DR. Phase contrast cine magnetic resonance imaging. Magn Reson Q. 1991;7: 229–254. 4. Mohiaddin RH, Pennell DJ. MR blood flow measurement. Clinical application in the heart and circulation. Cardiol Clin. 1998;16: 161–187. 5. Chai P, Mohiaddin R. How we perform cardiovascular magnetic resonance flow assessment using phase-contrast velocity mapping. J Cardiovasc Magn Reson. 2005;7:705–716. 6. Naidich TP, Altman NR, Gonzalez-Arias SM. Phase contrast cine magnetic resonance imaging: normal cerebrospinal fluid oscillation and applications to hydrocephalus. Neurosurg Clin N Am. 1993; 4:677–705. 7. Drangova M, Zhu Y, Bowman B, Pelc NJ. In vitro verification of myocardial motion tracking from phase-contrast velocity data. Magn Reson Imaging. 1998;16:863–870. 8. Carr HY, Purcell EM. Effects of diffusion on free precession in nuclear magnetic Resonance Experiments. Phys Rev. 1954;94: 30–638. 9. Hahn EL. Detection of sea water motion by nuclear precession. J Geophys Res. 1960;65:776–777. 10. Moran PR. A flow velocity zeugmatographic interlace for NMR imaging in humans. Magn Reson Imaging. 1982;1:197–203. 11. Bryant DJ, Payne JA, Firmin DN, Longmore DB. Measurement of flow with NMR imaging using a gradient pulse and phase difference technique. J Comput Assist Tomogr. 1984;8:588–593. 12. Van Dijk P. Direct cardiac NMR imaging of heart wall and blood flow velocity. J Comput Assist Tomogr. 1984;171:429–436. 13. Andersen AH, Kirsch JE. Analysis of noise in phase contrast MR imaging. Med Phys. 1996;23:857–869. 14. Bernstein MA, Shimakawa A, Pelc NJ. Minimizing TE in momentnulled or flow-encoded two- and three-dimensional gradient-echo imaging. J Magn Reson Imaging. 1992;2:583–588. 15. Pelc NJ, Bernstein MA, Shimakawa A, Glover GH. Encoding strategies for three-direction phase-contrast MR imaging of flow. J Magn Reson Imaging. 1991;1:405–413.

63 16. Atkinson DJ, Edelman RR. Cineangiography of the heart in a single breath hold with a segmented turboFLASH sequence. Radiology. 1991;178:357–360. 17. Thomsen C, Cortsen M, Sondergaard L, Henriksen O, Stahlberg F. A segmented K-space velocity mapping protocol for quantification of renal artery blood flow during breath-holding. J Magn Reson Imaging. 1995;5:393–401. 18. Ehman RL, Felmlee JP. Adaptive technique for high-definition MR imaging of moving structures. Radiology. 1989;173:255–263. 19. Felmlee JP, Ehman RL, Riederer SJ, Korin HW. Adaptive motion compensation in MRI: accuracy of motion measurement. Magn Reson Med. 1991;18:207–213. 20. Taylor AM, Jhooti P, Wiesmann F, Keegan J, Firmin DN, Pennell DJ. MR navigator-echo monitoring of temporal changes in diaphragm position: implications for MR coronary angiography. J Magn Reson Imaging. 1997;7:629–636. 21. Didier D. Assessment of valve disease: qualitative and quantitative. Magn Reson Imaging Clin N Am.2003;11:115–134, vii. 22. Gatehouse PD, Keegan J, Crowe LA, Masood S, Mohiaddin RH, Kreitner KF, Firmin DN. Applications of phase-contrast flow and velocity imaging in cardiovascular MRI. Eur Radiol. 2005;15: 2172–2184. 23. Beerbaum P, Korperich H, Barth P, Esdorn H, Gieseke J, Meyer H. Noninvasive quantification of left-to-right shunt in pediatric patients: phase-contrast cine magnetic resonance imaging compared with invasive oximetry. Circulation. 2001;103: 2476–2482. 24. Underwood SR, Firmin DN, Rees RS, Longmore DB. Magnetic resonance velocity mapping. Clin Phys Physiol Meas. 1990;11 (Suppl A):37–43. 25. Bernstein MA, Grgic M, Brosnan TJ, Pelc NJ. Reconstructions of phase contrast, phased array multicoil data. Magn Reson Med. 1994;32:330–334. 26. Dumoulin CL, Souza SP, Walker MF, Wagle W. Three-dimensional phase contrast angiography. Magn Reson Med. 1989;9:139–149. 27. Lum DP, Johnson KM, Paul RK, Turk AS, Consigny DW, Grinde JR, Mistretta CA, Grist TM. Transstenotic pressure gradients: measurement in swine--retrospectively ECG-gated 3D phase-contrast MR angiography versus endovascular pressure-sensing guidewires. Radiology. 2007;245:751–760. 28. Dumoulin CL. Phase contrast MR angiography techniques. Magn Reson Imaging Clin N Am. 1995;3:399–411. 29. Johnson KM, Lum DP, Turski PA, Block WF, Mistretta CA, Wieben O. Improved 3D phase contrast MRI with off-resonance corrected dual echo VIPR. Magn Reson Med. 2008;60:1329–1336. 30. Bock J, Frydrychowicz A, Stalder AF, Bley TA, Burkhardt H, Hennig J, Markl M. 4D phase contrast MRI at 3 T: effect of standard and blood-pool contrast agents on SNR, PC-MRA, and blood flow visualization. Magn Reson Med. 2010;63:330–338. 31. Markl M, Harloff A, Bley TA, Zaitsev M, Jung B, Weigang E, Langer M, Hennig J, Frydrychowicz A. Time-resolved 3D MR velocity mapping at 3T: improved navigator-gated assessment of vascular anatomy and blood flow. J Magn Reson Imaging. 2007; 25:824–831. 32. Wigstrom L, Sjoqvist L, Wranne B. Temporally resolved 3D phasecontrast imaging. Magn Reson Med. 1996;36:800–803. 33. Buonocore MH. Visualizing blood flow patterns using streamlines, arrows, and particle paths. Magn Reson Med. 1998;40:210–226. 34. Kozerke S, Hasenkam JM, Pedersen EM, Boesiger P. Visualization of flow patterns distal to aortic valve prostheses in humans using a fast approach for cine 3D velocity mapping. J Magn Reson Imaging. 2001;13:690–698. 35. Fyrenius A, Wigstrom L, Ebbers T, Karlsson M, Engvall J, Bolger AF. Three dimensional flow in the human left atrium. Heart. 2001;86:448–455.

64 36. Kilner PJ, Yang GZ, Wilkes AJ, Mohiaddin RH, Firmin DN, Yacoub MH. Asymmetric redirection of flow through the heart. Nature. 2000;404:759–761. 37. Bogren HG, Buonocore MH. 4D magnetic resonance velocity mapping of blood flow patterns in the aorta in young vs. elderly normal subjects. J Magn Reson Imaging. 1999;10:861–869. 38. Frydrychowicz A, Harloff A, Jung B, Zaitsev M, Weigang E, Bley TA, Langer M, Hennig J, Markl M. Time-resolved, 3-dimensional magnetic resonance flow analysis at 3 T: visualization of normal and pathological aortic vascular hemodynamics. J Comput Assist Tomogr. 2007;31:9–15. 39. Kilner PJ, Yang GZ, Mohiaddin RH, Firmin DN, Longmore DB. Helical and retrograde secondary flow patterns in the aortic arch studied by three-directional magnetic resonance velocity mapping. Circulation. 1993;88(5 Pt 1):2235–2247. 40. Harloff A, Albrecht F, Spreer J, Stalder AF, Bock J, Frydrychowicz A, Schollhorn J, Hetzel A, Schumacher M, Hennig J, Markl M. 3D blood flow characteristics in the carotid artery bifurcation assessed by flow-sensitive 4D MRI at 3T. Magn Reson Med. 2009;61:65–74. 41. Wetzel S, Meckel S, Frydrychowicz A, Bonati L, Radue EW, Scheffler K, Hennig J, Markl M. In vivo assessment and visualization of intracranial arterial hemodynamics with flow-sensitized 4D MR imaging at 3T. AJNR Am J Neuroradiol. 2007;28:433–438. 42. Reiter G, Reiter U, Kovacs G, Kainz B, Schmidt K, Maier R, Olschewski H, Rienmueller R. Magnetic resonance-derived 3-dimensional blood flow patterns in the main pulmonary artery as a marker of pulmonary hypertension and a measure of elevated mean pulmonary arterial pressure. Circ Cardiovasc Imaging. 2008;1:23–30. 43. Tyszka JM, Laidlaw DH, Asa JW, Silverman JM. Three-dimensional, time-resolved (4D) relative pressure mapping using magnetic resonance imaging. J Magn Reson Imaging. 2000;12:321–329. 44. Stalder AF, Russe MF, Frydrychowicz A, Bock J, Hennig J, Markl M. Quantitative 2D and 3D phase contrast MRI: optimized analysis of blood flow and vessel wall parameters. Magn Reson Med. 2008;60:1218–1231. 45. Vulliemoz S, Stergiopulos N, Meuli R. Estimation of local aortic elastic properties with MRI. Magn Reson Med. 2002;47:649–654. 46. Ebbers T, Wigstrom L, Bolger AF, Engvall J, Karlsson M. Estimation of relative cardiovascular pressures using time-resolved three-dimensional phase contrast MRI. Magn Reson Med. 2001;45:872–879. 47. Frydrychowicz A, Stalder AF, Russe MF, Bock J, Bauer S, Harloff A, Berger A, Langer M, Hennig J, Markl M. Three-dimensional analysis of segmental wall shear stress in the aorta by flow-sensitive four-dimensional-MRI. J Magn Reson Imaging. 2009;30:77–84. 48. Dyverfeldt P, Sigfridsson A, Kvitting JP, Ebbers T. Quantification of intravoxel velocity standard deviation and turbulence intensity by generalizing phase-contrast MRI. Magn Reson Med. 2006;56: 850–858.

B. Jung and M. Markl 49. Gu T, Korosec FR, Block WF, Fain SB, Turk Q, Lum D, Zhou Y, Grist TM, Haughton V, Mistretta CA. PC VIPR: a high-speed 3D phase-contrast method for flow quantification and high-resolution angiography. AJNR Am J Neuroradiol. 2005;26:743–749. 50. Turk AS, Johnson KM, Lum D, Niemann D, Aagaard-Kienitz B, Consigny D, Grinde J, Turski P, Haughton V, Mistretta C. Physiologic and anatomic assessment of a canine carotid artery stenosis model utilizing phase contrast with vastly undersampled isotropic projection imaging. AJNR Am J Neuroradiol. 2007;28: 111–115. 51. Moftakhar R, Aagaard-Kienitz B, Johnson K, Turski PA, Turk AS, Niemann DB, Consigny D, Grinde J, Wieben O, Mistretta CA. Noninvasive measurement of intra-aneurysmal pressure and flow pattern using phase contrast with vastly undersampled isotropic projection imaging. AJNR Am J Neuroradiol. 2007;28:1710–1714. 52. Baltes C, Kozerke S, Hansen MS, Pruessmann KP, Tsao J, Boesiger P. Accelerating cine phase-contrast flow measurements using k-t BLAST and k-t SENSE. Magn Reson Med. 2005;54:1430–1438. 53. Jung B, Honal M, Ullmann P, Hennig J, Markl M. Highly k-t-spaceaccelerated phase-contrast MRI. Magn Reson Med. 2008;60: 1169–1177. 54. Walker PG, Cranney GB, Scheidegger MB, Waseleski G, Pohost GM, Yoganathan AP. Semiautomated method for noise reduction and background phase error correction in MR phase velocity data. J Magn Reson Imaging. 1993;3:521–530. 55. Bernstein MA, Zhou XJ, Polzin JA, King KF, Ganin A, Pelc NJ, Glover GH. Concomitant gradient terms in phase contrast MR: analysis and correction. Magn Reson Med. 1998;39:300–308. 56. Markl M, Bammer R, Alley MT, Elkins CJ, Draney MT, Barnett A, Moseley ME, Glover GH, Pelc NJ. Generalized reconstruction of phase contrast MRI: analysis and correction of the effect of gradient field distortions. Magn Reson Med. 2003;50:791–801. 57. Peeters JM, Bos C, Bakker CJ. Analysis and correction of gradient nonlinearity and B0 inhomogeneity related scaling errors in twodimensional phase contrast flow measurements. Magn Reson Med. 2005;53:126–133. 58. McCauley TR, Pena CS, Holland CK, Price TB, Gore JC. Validation of volume flow measurements with cine phase-contrast MR imaging for peripheral arterial waveforms. J Magn Reson Imaging. 1995;5:663–668. 59. Oshinski JN, Ku DN, Bohning DE, Pettigrew RI. Effects of acceleration on the accuracy of MR phase velocity measurements. J Magn Reson Imaging. 1992;2:665–670. 60. Forster J, Sieverding L, Breuer J, Lutz O, Schick F. Fast acceleration-encoded magnetic resonance imaging. Med Phys. 2001; 28:28–35. 61. Firmin DN, Nayler GL, Kilner PJ, Longmore DB. The application of phase shifts in NMR for flow measurement. Magn Reson Med. 1990;14:230–241.

4

Technical Aspect of Contrast-Enhanced MRA Honglei Zhang, Wei Zhang, and Martin R. Prince

Introduction Although MR angiography (MRA) has revolutionized imaging of vascular diseases, the complicated nature of blood/soft tissue contrast mechanisms and the unique artifacts associated with each of the many techniques have made it challenging for referring physicians to become comfortable interpreting noncontrast MRA studies. Fortunately, contrast-enhanced (CE) MRA provides the type of contrast arteriogram that clinicians and radiologists are comfortable interpreting while eliminating the risks of radiation, iodinated contrast, and arterial catheterization [1]. Just like DSA and CTA, CE-MRA provides reliable enhancement of the arterial lumen during the arterial phase of the Gd bolus injection. Although MR imaging is slower than DSA or CTA, advances in magnet technology, gradient performance, pulse sequence design, and MR contrast agents continue to improve CE-MRA image quality such that it rivals DSA in accuracy for diagnosing vascular anomalies and diseases. This chapter describes the basic principles underlying CE-MRA techniques, approaches to optimizing applications throughout the body, and methods of contrast agent bolus timing.

Theory Unlike conventional MRA techniques which rely on blood flow or intrinsic blood relaxation properties to distinguish vasculature from background tissue, CE-MRA uses a contrast agent (e.g., gadolinium) to shorten the T1 (spin lattice) relaxation time of blood so that the intraluminal signal is brighter than that of surrounding tissues [1, 2] on T1-weighted images. Blood can then be directly imaged using fast,

H. Zhang, MD () • W. Zhang, MD • M.R. Prince, MD, PhD Department of Radiology, Weill Medical College of Cornell University, New York, NY 10022, USA e-mail: [email protected]

three-dimensional, spoiled gradient echo or steady-state free precession pulse sequences. The short echo time of these 3D gradient echo sequences minimizes blood motion, e.g., flow artifacts which are problematic on noncontrast techniques. Spoiled gradient echo is preferred over steady-state free precession especially for large, field of view (FOV) applications to eliminate banding and susceptibility artifacts of steadystate free precession. T1 shortening effect of paramagnetic contrast agents, e.g., gadolinium, is independent of blood flow or scan plane. With CE-MRA, in-plane imaging of vessels allows a small number of slices, required in the plane of those vessels, to quickly image an extensive length of vessel at high spatial resolutions. CE-MRA data can be acquired in coronal, sagittal, or oblique planes to encompass the vascular anatomy of interest with a minimal number of slices. Imaging in the plane of the arteries also takes advantage of the tendency for MR to image at higher resolution in plane compared to through plane. Thus, CE-MRA is intrinsically fast, allowing high-resolution, breath-hold MR angiograms. The high resolution and quality of CE-MRA have yielded sensitivity and specificity for evaluating different vascular territories in the high 90% range using conventional X-ray angiography as gold standard of reference [3–5]. CE-MRA captures central k-space data during the peak arterial phase of a contrast agent bolus [2]. More peripheral k-space data may be collected before and/or after the Gd peak to enhance MRA image resolution. More k-space details are described under “Fourier consideration.” This 3D data can be reformatted into any obliquity to unfold tortuous arteries and to view lesions in multiple organs. Compared to computed tomographic angiography (CTA), volume rendering is more straightforward because arteries are the brightest structures; there is no need to cut away bones since calcium has low signal on MR. As long as the echo time (TE) is sufficiently short, less than 3 ms, vascular calcifications do not interfere with depiction of the artery lumen. The short echo time used with 3D Fourier transform (FT) gradient echo sequences also minimizes artifact from metal clips, bowel gas, and other susceptibility sources.

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_4, © Springer Science+Business Media, LLC 2012

65

66

Pulse Sequence CE-MRA primarily utilizes 3D spoiled gradient echo pulse sequences to take advantage of their high speed, short echo time, and T1 weighting with a single center of k-space for the entire volume of data [2]. CE-MRA can also be used with steady-state free precession with more signal-to-noise ratio (SNR) but also more artifacts, especially with larger FOV, where it is difficult to have good field homogeneity. PC MRA may also improve postinjection of paramagnetic contrast agent. Three-dimensional sequences have high spatial resolution with thin slices and intrinsically high SNR. This sequence is made fast by using short RF pulses or completely eliminating slice selection. Spoiling of the transverse magnetization is useful because it accentuates T1 contrast, thereby magnifying the effect of paramagnetic contrast agents (e.g., gadolinium). It also suppresses signal from background tissues to enhance image contrast and reduce visibility of background aliasing. Mask subtraction and fat suppression can also enhance image contrast, but can create artifacts. For example, motion between the mask and arterial phase creates misregistration artifacts. Poor field homogeneity can cause a fat saturation pulse to drift onto the water peak at the edge of the image or near-susceptibility sources giving the false appearance of occlusion. A recent advance is the acquisition of two echoes, in phase and out of phase with each TR and eliminating fat signal with the Dixon method. Repetition times (TR) used for CE-MRA are generally around 5 ms or less, and TE are typically 1–2 ms. Total scan times range from 10 to 30 s, although this can be lengthened further in regions not affected by respiratory motion for greater resolution, coverage, or SNR. It can also be shortened by sacrificing spatial resolution or coverage and repeated multiple times to provide time-resolved MRA showing the passage of contrast through the volume of interest. Using ultrashort TR, parallel imaging, partial Fourier, and/or sharing of peripheral k-space data, sliding window reconstruction with radial or spiral trajectories, time-resolved CE-MRA is possible at subsecond temporal resolution [1, 6–11]. Even faster temporal resolution can be acquired with Cartesian acquisition with projection reconstruction (CAPR), vastly undersampled isotropic projection reconstruction (VIPR), k–t broad-use linear acquisition speed-up technique (BLAST), or using 2D spoiled gradient echo sequences with complex mask subtraction [1, 12–15]. Gradient echo imaging without spoiling can also be useful when tissue perfusion is of interest.

Fourier (k-Space) Considerations MR imaging does not collect data voxel by voxel. Instead, MR scanners collect spatial frequency data, also known as Fourier or k-space data. With 3D MR imaging, the entire 3D

H. Zhang et al.

“Fourier” or “k-space” dataset is collected before reconstructing individual slices [2]. Because k-space maps spatial frequencies rather than spatial data, k-space data does not directly correspond to image space. Instead, different regions of k-space data determine different image features. For example, the center of k-space or “low” spatial frequencies dominates image contrast, whereas the periphery of k-space or “high” spatial frequencies contributes more to fine details, such as edges. Thus, the state of the contrast agent bolus and intravascular T1 in large vessels is captured at the moment corresponding to acquisition of central k-space [1, 2, 11]. If central k-space is collected when the arterial contrast agent concentration is peaking but has not yet reached capillaries and veins (arterial phase), only the arteries will be bright on MRA images. With optimal timing, contrast injection duration does not need to be as long as the acquisition time when the bolus peak in the artery of interest occurs during acquisition of central k-space data. Most protocols use contrast injection durations that are much shorter than the scan acquisition time. Shorter injection duration with the same total Gd dose results in a faster injection with higher blood Gd concentrations which increases SNR. On the other hand, shorter injections must be timed precisely to avoid artifacts which occur when the contrast bolus peak does not coincide perfectly with acquisition of the center of k-space. For a given Gd dose, the injection strategy is a trade-off between a fast injection for higher intravascular signal versus a slow injection for more uniform signal, easier bolus timing and fewer artifacts. As a general rule, injecting the total contrast dose over approximately half of the acquisition time is optimum if the Gd bolus is perfectly timed for the arterial phase to synchronize with the acquisition of central k-space data. Make sure to avoid a rapid increase or decrease in the concentration of the contrast agent during the acquisition of the central k-space as this causes edge-ringing artifacts (Fig. 4.1). Another factor that must be considered is phase-encoding order. Generally, phase encoding is performed “sequentially” such that central k-space is acquired at the midpoint of the scan. Alternatively, phase encoding can be performed in a “centric” fashion such that central k-space is acquired at the beginning of the scan. Although centric phase-encoding order has been shown to be more prone to artifacts [15], it greatly simplifies bolus timing and may be less susceptible to venous enhancement and artifacts from incomplete breath holding. Centric encoding is generally most effective when performed as an elliptical centric variant, as this concentrates the center of k-space into a shorter time at the beginning of the scan [16, 17]. Another choice is partial Fourier acquisition beginning near the k-space center which has some of the benefits of both centric and sequential ordering [18]. The best suppression of venous enhancement is with recessed elliptical centric encoding. Recessing the absolute center of k-space a few seconds from the beginning of the scan allows

4

Technical Aspect of Contrast-Enhanced MRA

Fig. 4.1 Time-resolved 2D MRA images show ringing artifact in the anterior tibial artery on early images, where there is enhancement of the periphery of k-space but not yet in the center of k-space. Ringing

67

disappears on the third image at 22 s indicating that the center of k-space is fully enhanced. Later images show enhancement of the veins and heel ulcer

Fig. 4.2 MR venogram of the legs with 9 mL gadofosvoset trisodium shows extensive acute deep venous thrombosis of the left upper soleus and gastrocnemius veins with surrounding inflammation

data acquisition to begin on the leading edge of the bolus while still synchronizing the center of k-space with the arterial Gd peak [19].

Contrast Agents At present, most CE-MRA examinations are conducted using gadolinium-based MR contrast agents. Gadolinium (Gd3+) is a paramagnetic metal ion that decreases both the spin–lattice (T1) and spin–spin (T2) relaxation times [3]. Because Gd3+ itself is biologically toxic, it is chelated with ligands, such as DTPA (gadopentetate dimeglumine), HP-DO3A (gadoteridol), gadoversetamide, or gadobenate dimeglumine to form low-molecular-weight contrast agents. These small, “extracellular” agents rapidly redistribute from the intravascular compartment into the interstitial space. Typically, 80% of gadolinium chelate leaks into the intravascular space within 5 min. Thus, imaging of arteries must be performed rapidly to exploit the “first pass” of the contrast agent. As compared with iodinated contrast agents, gadolinium chelates have a very low rate of adverse events and no nephrotoxicity, which is a significant advantage in patients with impaired renal

function. One exception is gadobenate dimeglumine which is ionic and has nearly as many reactions as iodinated contrast agents but very low risk of nephrogenic systemic fibrosis (NSF). Some new high-relaxivity contrast agents are large enough, e.g., USPIO, or bind to large, serum molecules, e.g., gadofosvoset trisodium, so that they stay within the intravascular compartment with minimal leaking out of the capillaries [20, 21]. These agents are referred to as blood pool or intravascular contrast agents. The high relaxivity of these blood pool agents makes them ideal for first pass, arterial phase imaging because a large, contrast effect is achieved with safe, low-injection rates. In addition, the longer intravascular half-life of these contrast agents may allow imaging longer for higher resolution or imaging additional vascular territories during the equilibrium, blood pool phase. Pulmonary and coronary 3D MRA may benefit from these new contrast agents. Other uses relate to venous imaging (Fig. 4.2) and “road map” imaging for monitoring vascular interventions. Blood pool agents may also be useful for identifying gastrointestinal (GI) bleeding using delays in a manner analogous to labeled red blood cells. Similarly, they can be used to detect slow or intermittent stent graft leaks [22].

68

Contrast Dose and Injection Rate In order to image bright blood with respect to background tissues during the arterial phase, sufficient contrast agent must be administered to transiently reduce the arterial blood T1 to substantially less than that of the brightest background tissue, fat, which has a T1 of approximately 270 ms [2]. During the “first-pass” arterial phase, blood T1 is more related to infusion rate and contrast agent relaxivity than total contrast dose, as arterial phase imaging occurs well before intravascular contrast equilibration. Total gadolinium dose and rate of delivery is a trade-off between maximizing intravascular signal (higher injection rate) and minimizing artifacts (longer injection duration). In the ideal situation, a large dose at a high rate would be optimal. This, however, must be weighed against safety, practicality, and cost. Gd chelates’ concentration in the arteries is estimated by the injection rate divided by the cardiac output. Higher injection rates give higher arterial Gd concentration and correspondingly higher SNR until the concentration is so high that T2* effects reduce signal. Diminishing return due to these T2* effects starts to occur for injection rates greater than 5 mL/s for standard, extracellular gadolinium chelates. It is also important to avoid Gd concentration changing too rapidly during acquisition of central k-space data as this creates ringing artifacts. But having a high injection rate for a long duration results in a large dose of Gd which can be expensive and may excessively exceed FDA limits. A good strategy for maximizing injection rate while minimizing the dose is to use a relatively rapid (2–3 mL/s) injection with a duration shorter than the scan duration. Typically, an injection duration that is half of the scan duration is optimum. For a 75 kg patient who received 15 mL of Gd (0.1 mmol/kg) at an injection rate of 3 mL/s for 5 s, the scan duration should be about 10-s long. Due to contrast agent dilution at the leading and trailing edges of the bolus, as well as varying transit times through different portions of the pulmonary circulation, the contrast bolus tends to lengthen as it travels from the antecubital vein (where it is typically injected), through the heart and lungs, to the arteries being imaged. This bolus dispersion depends on individual cardiovascular parameters which are difficult to predict, but at least 5–7 s of bolus prolongation occurs in most individuals [2]. Even greater bolus dispersion can be expected for peripheral arteries, whereas only minimal bolus dispersion can be expected for pulmonary arteries. For fast breath-held acquisitions of the systemic arterial system, the total dose of contrast should be administered at a rate that results in an injection time that is at least 5–7 s shorter than the data acquisition time. An injection duration which is about half the scan duration is recommended since it is not necessary to have peak Gd concentration while acquiring the periphery of k-space.

H. Zhang et al.

The optimal Gd dose also depends upon the vascular territory being imaged and the available software and hardware. Generally, the more contrast agent, the better the MRA image quality because larger contrast agent doses allow high injection rate for a longer time. However, the FDA-approved dose is only up to 0.3 mmol/kg body weight. Recently, concern over NSF has led to greater attention to the dose of Gd use for MR procedures because the risk scales with dose. It appears that for ionic and macrocyclic gadolinium chelates, there is negligible risk at 0.1 mmol/kg as the overwhelming majority of NSF cases have occurred at high doses (Fig. 4.3) [23]. For nonionic linear chelates which have the most cases of NSF, the risk of single dose appears to be less than 1 in 10,000 for patients with GFR 70 years old), 25–35 s in patients with cardiac disease or an aortic aneurysm, and up to 40–50 s when severe cardiac failure in conjunction with aortic aneurysm is present. Add 3 s if the IV is in the hand [1]. Best guess technique works well for scan time of at least 40 s or longer (with sequential k-space encoding) when the impact of 5 s of timing error is minimal.

H. Zhang et al.

background signals. Finally, the imaging bolus may not behave identical to the test bolus due to moment-to-moment patient variables, such as venous return and cardiac output, as well as the different total volume of injection. It is especially important to be aware that a test bolus with the arm at the patient’s side may not be accurate if the forearms are then repositioned above the patient’s head for the actual scan as overhead positioning the patient’s arms may stretch and pinch the subclavian veins.

Automatic Triggering Another approach to bolus timing uses a pulse sequence that can automatically detect contrast arrival in the aorta and synchronize 3D MRA data acquisition with the arterial phase of the contrast bolus. The operator selects a region of aorta to be sampled at 20-ms intervals. As contrast arrives in the aorta, the signal within the sampling region increases. A trigger threshold of typically 20% signal increase detects the leading edge of the contrast bolus. This gives the patient time to take in a deep breath and suspend respiration before the actual scan commences 5 s later. By beginning the scan near the center of k-space, i.e., centric or preferably recessed elliptical centric encoding, there is synchronization between the arterial phase of the bolus and elliptical centric acquisition of central k-space data. Commercial versions of this pulse sequence are available as “SmartPrep” (GE Medical Systems) or “Bolustrack” (Philips Medical Systems).

MR Fluoroscopy Test Bolus Technique A bolus-timing acquisition can be performed prior to acquisition of 3D MRA dataset using a small Gd contrast dose of ~2 mL followed by a 20-mL saline flush at the same rate as planned for the actual injection. Multiple, single-slice, fast gradient echo images in the vascular region of interest are acquired as rapidly as possible (every 1–2 s) for approximately 1 min. In order to minimize time-of-flight effects, the 2D test bolus images should either be oriented in the plane of imaged vessel (i.e., sagittal or coronal for the aorta) or alternatively be relatively thick (greater than 1 cm) with a superior saturation band or a blood-nulling inversion prepulse. The time of peak arterial enhancement (contrast travel time) is then determined visually or using ROI analysis. While quite effective, there are several drawbacks to the test bolus technique. Setting up, performing, and analyzing the test bolus lengthen the overall examination time. The test bolus rapidly redistributes into the interstitial space and is excreted by the kidneys into the ureters, thereby increasing

Perhaps, the most popular method of contrast bolus timing is to use extremely rapid MR fluoroscopy [1, 2, 16, 17]. With this technique, 2D gradient-refocused images are rapidly (less than 1 s/image) obtained through the vascular structure of interest, ideally using complex subtraction to improve contrast and decrease artifacts. Images are generated in near real time and updated at greater than once per second. The operator watches the contrast bolus arrive and then switches over to recessed elliptical centric 3D MRA when the desired enhancement is detected. This technique allows real-time, operator-dependent decision making. This may be particularly advantageous in cases with unusual or asymmetric flow patterns, such as unilateral stenoses and slow filling aneurysms. The ability to assess the vasculature “on the fly” under these circumstances is a great asset, as is the time and contrast savings associated with avoiding a test injection. It is useful to shift the MR fluoroscopic monitoring to a proximal region of vascular anatomy to get more advanced warning of contrast agent arrival. Using carotid MRA as an

4

Technical Aspect of Contrast-Enhanced MRA

71

example, MR fluoroscopic monitoring in the chest shows contrast agent arriving in the subclavian vein, then right heart, pulmonary arteries, left heart, arch, and finally proximal great vessels. Tracking the bolus over such a long period makes precise triggering easier. If the flow is very slow, the operator can compensate to allow for greater target vessel enhancement before triggering. Fluoroscopic triggering is improved by recessing the absolute center of k-space a few seconds from the beginning of the 3D acquisition, thus avoiding the previously described “leading-edge” artifact from early triggering.

More recent refinements traverse k-space using radial or spiral projections [1], which allow for sliding window reconstructions at very high temporal and spatial resolutions. VIPR and CAPR are particularly promising. VIPR can be combined with phase-contrast imaging as well. Another strategy that works particularly well with the sparse MRA datasets is to use training data acquired either pre or post contrast injection. Training data and post-Gd steady-state data help calculate high spatial resolution during high temporal resolution undersampling while contrast agent is injected, e.g., k–t BLAST or HYPR TRICKS.

Time-Resolved 3D Contrast MRA

Imaging Different Vascular Phases

Because proper bolus timing is difficult and a timing mistake can ruin the study, time-resolved techniques have been developed whereby multiple 3D datasets are acquired extremely rapidly (typically, 1–10 s per acquisition). Bolus timing is no longer a factor as multiple vascular phases are obtained without any predetermined timing (i.e., inject and begin scanning simultaneously). The operator simply selects the desired image set: pure arterial phase, maximum venous, etc. This is particularly useful in the carotid and pulmonary arteries, where the venous phase is extremely rapid, and in the calf, where variable rate filling may occur due to stenoses, occlusions, or rapid AV shunting due to soft tissue ulcers, cellulitis, and other inflammatory diseases [2]. The most straightforward way to accelerate acquisition time is using some combination of limiting the imaging volume, decreasing the resolution (e.g., fewer, thicker slices and fewer phase-encoding steps), decreasing TR, and partial Fourier parallel imaging. Recent developments in gradient systems allow TRs of 40%. Examples of matched cross-sectional images from three contrast weightings (TOF, T1WI, and CE-T1WI) for each category are provided. Arrowheads indicate the outer wall boundary; asterisk indicates the lumen and single arrows indicate LRNC. Reprinted with permission from Underhill et al. [36]

Fig. 8.11 Representative T1-weighted images of progression of atherosclerosis with intraplaque hemorrhage in right carotid artery. Each column presents matched cross-sectional locations in the carotid artery from baseline MRI (a) and MRI obtained 18 months later

(b). Lumen area was decreased, and wall area was increased in each section at second examination. CCA indicates common carotid artery; Bif bifurcation, ICA internal carotid artery, ECA external carotid artery (Reprinted with permission from Takaya et al. [37])

124

of subclinical CVD. There was a high prevalence of LRNC (71%) in subjects with wall thickness greater than 1.5 mm [38]. Vessel wall imaging is important to identify atherosclerosis considering the high prevalence of carotid lesions in patients with minimal stenosis. The feasibility of coronary vessel wall imaging was also demonstrated in this study. Coronary maximum wall thickness measured by black-blood MRI was greater for subjects with two or more cardiovascular risk factors compared to subjects with one or no risk factors [39]. Positive remodeling of the coronaries in response to atherosclerosis was also demonstrated [40].

Clinical Trials Safe noninvasive monitoring of change in atherosclerotic plaque size and composition is possible using MRI. This is useful for both patient follow-up and to assess the efficacy of drug treatment in reducing atherosclerotic disease burden. Currently, a number of clinical trials of statin therapy use MRI-based plaque measurements for assessing the efficacy of the treatment. Serial MRI studies monitoring a change in either plaque morphology or composition require high measurement reproducibility to ensure that the study endpoint can be assessed using a small number of subjects. At 1.5 T, Saam et al. [41] showed high measurement reproducibility denoted by low coefficients of variation (CV) of 5.8% for wall volume, 4.3% for lumen volume, and 3.2% for the percent atheroma volume. Compositional measurements, such as maximum percent LRNC had a larger CV (15%). Using MRI with this level of reproducibility will only require 14 subjects per treatment arm to identify a 5% treatment effect if the percent atheroma volume is used as an endpoint. Recent studies [42] established the corresponding CVs to be similar at 3 T. Based on these studies the lowest sample size can be achieved for studies using plaque burden change as the primary endpoint resulting in considerable time and cost savings. Accordingly, several studies using plaque burden as the primary endpoint have demonstrated significant reduction in plaque burden with treatment. With a 2-year simvastatin treatment, carotid and aortic vessel wall area (VWA) reduced by 14% and 10%, respectively, at 12 months and 18% and 15%, respectively, at 24 months [43]. Similarly, a high-dose atorvastatin (20 mg/day) treatment reduced vessel wall thickness and VWA of thoracic aorta plaques compared to lowdose atorvastatin (5 mg/day) [44]. In the carotid plaque composition (CPC) Study [45] of subjects with coronary artery disease or carotid disease treated with atorvastatin alone or combined with niacin and/or colesevelam, percent wall volume also decreased by 3.8% over the 3-year course of the study. The reductions in the first, second, and third years were 0.3%, 3.6%, and 0.1%, respectively, showing that

R. Li et al.

plaque burden can be used as the primary endpoint with a 2-year treatment. Plaque composition, such as the percent lipid composition, is increasingly used in clinical drug trials to monitor the efficacy of drug treatment. In a case control study of eight patients with coronary artery disease on intensive lipid-lowering therapy for 10 years, the main treatment effects were related to plaque composition with treated patients having a smaller core lipid area (0.7 vs. 10.2 mm2, respectively; P = 0.01) and percent lipid composition (1% vs. 17%, respectively) compared to untreated patients. In the Outcome of Rosuvastatin treatment on carotid artery atheroma: a magnetic resonance Imaging ObservatioN (ORION) trial of rosuvastatin treatment effects on carotid plaque, percent LRNC decreased by 41.4% in 24 months in all patients with LRNC at baseline [46]. A decrease in percent LRNC (8.4% to 5.2 over 3 years) was also noted in the CPC study. When normal wall was excluded percent LRNC decreased by 3.2%, 3.0%, and 0.91% in the first, second, and third years, respectively. Thus, using LRNC as the primary endpoint, a 1-year study can be used to test the efficacy of drug treatment. The ongoing carotid MRI substudy of the Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides and Impact on Global Health Outcomes (AIM-HIGH) comparing intensive LDLlowering plus HDL-raising therapy against LDL-lowering alone also monitors the effect on LRNC by serial MRI.

Future Directions Coronary Vessel Wall Imaging While vessel wall imaging of the carotids and aorta have advanced to the point where clinical studies are feasible, direct vessel wall imaging of the coronaries is still challenging because of the demands of high-resolution and high SNR due to the thin vessel walls and their position within the thorax in addition to the need for cardiac and navigator gating. Although coronary vessel walls have been visualized with DIR [47] and local pencil beam reinversion [48] image quality depends on effective gating and respiratory compensation. Imaging in the mid-diastolic quiescent period or using a subject-specific trigger delay measured from a cine scout can provide better image quality [49]. Breath-holding [47] or navigator monitoring of diaphragmatic motion [48] are generally used for respiratory compensation.

Molecular Imaging In addition to morphologic and compositional information MRI can provide functional information with the use of exogenous contrast agents. Ultrasmall superparamagnetic

8

Vessel Wall Imaging Techniques

125

particles of iron oxides (USPIO) have a large R2/R1 ratio and can be visualized based on their T2 or T2* effects. USPIO accumulation in macrophages in the carotid plaque has been validated by comparison against histology [50]. Plaques from symptomatic patients showed a decrease in signal intensity by 16.4% after USPIO infusion when compared to an 8.4% increase in plaques from asymptomatic patients [51]. Further studies have shown that asymptomatic arteries can also exhibit signal loss after USPIO infusion [52] suggestive of occult inflammation. Plaque activity monitoring by USPIO agents was used in the Atorvastatin Therapy: Effects on Reduction of Macrophage Activity (ATHEROMA) study [53] to monitor the effects of atorvastatin therapy. A significant reduction of USPIO-defined inflammation was observed in as early as 6 weeks. Gadolinium-based targeted molecular contrast agents have been demonstrated for thrombus imaging using a fibrintargeted peptide, inflammation targeting matrix metalloproteinases with specific contrast agents, gadofluorine or polyethylene glycol (PEG)-micelles incorporated with gadolinium DTPA for LRNC. Plaque neovessels can be targeted using alpha(v)beta3-integrin-targeted, paramagnetic nanoparticles. Details about these preclinical contrast agents can be found in the review by Briley-Saebo et al. [54].

carotid artery with black-blood MR images [59]. In-let and out-let flow rates were also acquired by phase contrast MRI. Finite element computational fluid dynamic simulation according to lumen boundary and flow rates was proposed to calculate the specific flow pattern and wall shear stress. Low and oscillating shear was found at the wall thickening place. Noninvasive vessel wall imaging is ideal to provide the boundary and flow information for hemodynamic studies.

Higher Resolution

1. Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases: part I: general considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation. 2001;104:2746–2753. 2. Naghavi M, Libby P, Falk E, et al. From vulnerable plaque to vulnerable patient: a call for new definitions and risk assessment strategies: Part I. Circulation. 2003;108:1664–1672. 3. Lu H, Clingman C, Golay X, van Zijl PC. Determining the longitudinal relaxation time (T1) of blood at 3.0 Tesla. Magn Reson Med. 2004;52:679–682. 4. Edelman RR, Chien D, Kim D. Fast selective black blood MR imaging. Radiology. 1991;181:655–660. 5. Parker DL, Goodrich KC, Masiker M, Tsuruda JS, Katzman GL. Improved efficiency in double-inversion fast spin-echo imaging. Magn Reson Med. 2002;47:1017–1021. 6. Yarnykh VL, Yuan C. Multislice double inversion-recovery blackblood imaging with simultaneous slice reinversion. J Magn Reson Imaging. 2003;17:478–483. 7. Yuan C, Kerwin WS, Ferguson MS, et al. Contrast-enhanced high resolution MRI for atherosclerotic carotid artery tissue characterization. J Magn Reson Imaging. 2002;15:62–67. 8. Yarnykh VL, Yuan C. T1-insensitive flow suppression using quadruple inversion-recovery. Magn Reson Med. 2002;48:899–905. 9. Yarnykh VL, Yuan C. Simultaneous outer volume and blood suppression by quadruple inversion-recovery. Magn Reson Med. 2006;55:1083–1092. 10. Wang J, Yarnykh VL, Hatsukami T, Chu B, Balu N, Yuan C. Improved suppression of plaque-mimicking artifacts in black-blood carotid atherosclerosis imaging using a multislice motion-sensitized driven-equilibrium (MSDE) turbo spin-echo (TSE) sequence. Magn Reson Med. 2007;58:973–981. 11. Clarke SE, Beletsky V, Hammond RR, Hegele RA, Rutt BK. Validation of automatically classified magnetic resonance images for carotid plaque compositional analysis. Stroke. 2006;37:93–97.

Current vessel wall MRI protocols have high spatial resolution in the image plane but low resolution (2–3 mm) in the slice direction which can limit detection of small plaque components [55]. 3D imaging can improve resolution in the partition encoding direction but blood suppression can be compromised if DIR-based techniques are used due to the large imaging slab thickness [55]. Diffusion preparation is well suited for 3D vessel wall imaging owing to its flow direction-independent blood suppression. Recently, new 3D vessel wall imaging approaches have been demonstrated using diffusion preparation [56]. Isotropic voxels can improve both accuracy and reproducibility of plaque component measurement. 3D isotropic black-blood imaging [56–58] can potentially improve the utility of vessel wall imaging in clinical studies.

Hemodynamic Study Previous sections are focused on the components of atherosclerotic plaque. However, the morphology of the lumen, which is clearly depicted by vessel wall imaging, is also used to evaluate the connection between mechanical forces and atherosclerotic disease. Steinman et al. reconstructed threedimensional models of the lumen and wall boundaries of

Conclusion Vessel wall imaging with black-blood prepulse provides not only morphological information about the lumen, but also the components of plaque, which are critical to evaluate plaque vulnerability, monitor plaque progression, and observe drug efficacy. Comprehensive understanding of atherosclerosis with the additional vessel wall imaging could improve the clinical management of vascular disease. Acknowledgment We would like to thank Marina S. Ferguson and Zach Miller for editing this chapter.

References

126 12. Yuan C, Kerwin WS, Yarnykh VL, et al. MRI of atherosclerosis in clinical trials. NMR Biomed. 2006;19:636–654. 13. Saam T, Ferguson MS, Yarnykh VL, et al. Quantitative evaluation of carotid plaque composition by in vivo MRI. Arterioscler Thromb Vasc Biol. 2005;25:234–239. 14. Chu B, Kampschulte A, Ferguson MS, et al. Hemorrhage in the atherosclerotic carotid plaque: a high-resolution MRI study. Stroke. 2004;35:1079–1084. 15. Moody AR. Magnetic resonance direct thrombus imaging. J Thromb Haemost. 2003;1:1403–1409. 16. Kim SE, Jeong EK, Shi XF, Morrell G, Treiman GS, Parker DL. Diffusion-weighted imaging of human carotid artery using 2D single-shot interleaved multislice inner volume diffusion-weighted echo planar imaging (2D ss-IMIV-DWEPI) at 3T: diffusion measurement in atherosclerotic plaque. J Magn Reson Imaging. 2009;30:1068–1077. 17. Cai J, Hatsukami TS, Ferguson MS, et al. In vivo quantitative measurement of intact fibrous cap and lipid-rich necrotic core size in atherosclerotic carotid plaque: comparison of high-resolution, contrast-enhanced magnetic resonance imaging and histology. Circulation. 2005;112:3437–3444. 18. Kerwin WS, O’Brien KD, Ferguson MS, Polissar N, Hatsukami TS, Yuan C. Inflammation in carotid atherosclerotic plaque: a dynamic contrast-enhanced MR imaging study. Radiology. 2006;241:459–468. 19. Kerwin W, Hooker A, Spilker M, et al. Quantitative magnetic resonance imaging analysis of neovasculature volume in carotid atherosclerotic plaque. Circulation. 2003;107:851–856. 20. Kerwin WS, Oikawa M, Yuan C, Jarvik GP, Hatsukami TS. MR imaging of adventitial vasa vasorum in carotid atherosclerosis. Magn Reson Med. 2008;59:507–514. 21. Balu N, Yarnykh VL, Scholnick J, Chu B, Yuan C, Hayes C. Improvements in carotid plaque imaging using a new eight-element phased array coil at 3T. J Magn Reson Imaging. 2009;30:1209–1214. 22. Yarnykh VL, Terashima M, Hayes CE, et al. Multicontrast blackblood MRI of carotid arteries: comparison between 1.5 and 3 tesla magnetic field strengths. J Magn Reson Imaging. 2006;23: 691–698. 23. Boussel L, Herigault G, de la Vega A, Nonent M, Douek PC, Serfaty JM. Swallowing, arterial pulsation, and breathing induce motion artifacts in carotid artery MRI. J Magn Reson Imaging. 2006;23:413–415. 24. Chan CF, Gatehouse PD, Hughes R, Roughton M, Pennell DJ, Firmin DN. Novel technique used to detect swallowing in volumeselective turbo spin-echo (TSE) for carotid artery wall imaging. J Magn Reson Imaging. 2009;29:211–216. 25. Kerwin W, Xu D, Liu F, et al. Magnetic resonance imaging of carotid atherosclerosis: plaque analysis. Top Magn Reson Imaging. 2007;18:371–378. 26. Babiarz L, Astor B, Mohamed M, Wasserman B. Comparison of gadolinium-enhanced cardiovascular magnetic resonance angiography with high-resolution black blood cardiovascular magnetic resonance for assessing carotid artery stenosis. J Cardiovasc Magn Reson. 2007;9:63–70. 27. Dong L, Underhill H, Yu W, et al. Geometric and compositional appearance of atheroma in an angiographically normal carotid artery in patients with atherosclerosis. AJNR Am J Neuroradiol. 2010;31:311–316. 28. Falk E. Why do plaques rupture? Circulation. 1992;86(6 Suppl): III30–42. 29. Hatsukami TS, Ross R, Polissar NL, Yuan C. Visualization of fibrous cap thickness and rupture in human atherosclerotic carotid plaque in vivo with high-resolution magnetic resonance imaging. Circulation. 2000;102:959–964. 30. Trivedi RA, U-King-Im JM, Graves MJ, et al. MRI-derived measurements of fibrous-cap and lipid-core thickness: the potential for identifying vulnerable carotid plaques in vivo. Neuroradiology. 2004;46(9):738–743.

R. Li et al. 31. Cai JM, Hatsukami TS, Ferguson MS, et al. In vivo quantitative measurement of intact fibrous cap and lipid-rich necrotic core size in atherosclerotic carotid plaque - Comparison of high-resolution, contrast-enhanced magnetic resonance imaging and histology. Circulation. 2005;112:3437–3444. 32. Yuan C, Zhang SX, Polissar NL, et al. Identification of fibrous cap rupture with magnetic resonance imaging is highly associated with recent transient ischemic attack or stroke. Circulation. 2002;105:181–185. 33. Underhill HR, Yuan C, Yarnykh VL, et al. Predictors of surface disruption with MR imaging in asymptomatic carotid artery stenosis. AJNR Am J Neuroradiol. 2010;31:487–493. 34. Singh N, Moody A, Gladstone D, et al. Moderate carotid artery stenosis: MR imaging-depicted intraplaque hemorrhage predicts risk of cerebrovascular ischemic events in asymptomatic men. Radiology. 2009;252:502–508. 35. Altaf N, MacSweeney ST, Gladman J, Auer DP. Carotid intraplaque hemorrhage predicts recurrent symptoms in patients with highgrade carotid stenosis. Stroke. 2007;38:1633–1635. 36. Underhill H, Hatsukami T, Cai J, et al. A noninvasive imaging approach to assess plaque severity: the Carotid Atherosclerosis Score. AJNR Am J Neuroradiol. 2010;31:1068–1075. 37. Takaya N, Yuan C, Chu B, et al. Presence of intraplaque hemorrhage stimulates progression of carotid atherosclerotic plaques - A high-resolution magnetic resonance Imaging study. Circulation. 2005 2005;111:2768–2775. 38. Wasserman BA, Sharrett R, Lai S, et al. Risk factor associations with the presence of a lipid core in carotid plaque of asymptomatic individuals using high-resolution MRI - The Multi-Ethnic Study of Atherosclerosis (MESA). Stroke. 2008;39:329–335. 39. Macedo R, Chen S, Lai S, et al. MRI Detects Increased Coronary Wall Thickness in Asymptomatic Individuals: The Multi-Ethnic Study of Atherosclerosis (MESA). J Magn Reson Imaging. 2008;28:1108–1115. 40. Miao CL, Chen SG, Macedo R, et al. Positive Remodeling of the Coronary Arteries Detected by Magnetic Resonance Imaging in an Asymptomatic Population MESA (Multi-Ethnic Study of Atherosclerosis). J Am Coll Cardiol. 2009;53:1708–1715. 41. Saam T, Kerwin WS, Chu B, et al. Sample size calculation for clinical trials using magnetic resonance imaging for the quantitative assessment of carotid atherosclerosis. J Cardiovasc Magn Reson. 2005;7: 799–808. 42. Li F, Yarnykh V, Hatsukami T, et al. Scan-rescan reproducibility of carotid atherosclerotic plaque morphology and tissue composition measurements using multicontrast MRI at 3T. J Magn Reson Imaging. 2010;31:168–176. 43. Corti R, Fayad ZA, Fuster V, et al. Effects of lipid-lowering by simvastatin on human atherosclerotic lesions - A longitudinal study by high-resolution, noninvasive magnetic resonance imaging. Circulation. 2001;104:249–252. 44. Yonemura A, Momiyama Y, Fayad ZA, et al. Effect of lipid-lowering therapy with atorvastatin on atherosclerotic aortic plaques detected by noninvasive magnetic resonance imaging. J Am Coll Cardiol. 2005;45:733–742. 45. Zhao X, Phan B, Chu B, et al. Testing the hypothesis of atherosclerotic plaque lipid depletion during lipid therapy by magnetic resonance imaging: study design of Carotid Plaque Composition Study. Am Heart J. 2007;154:239–246. 46. Underhill H, Yuan C, Zhao X, et al. Effect of rosuvastatin therapy on carotid plaque morphology and composition in moderately hypercholesterolemic patients: a high-resolution magnetic resonance imaging trial. Am Heart J. 2008;155:584.e581-588. 47. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human coronary artery lumen and wall imaging using black-blood magnetic resonance imaging. Circulation. 2000;102:506–510. 48. Botnar RM, Won YK, Peter Br, Matthias S, Elmar S, Warren JM. 3D coronary vessel wall imaging utilizing a local inversion

8

49.

50.

51.

52.

53.

Vessel Wall Imaging Techniques technique with spiral image acquisition. Magn Reson Med. 2001;46:848–854. Kim W, Stuber M, Kissinger K, Andersen N, Manning W, Botnar R. Impact of bulk cardiac motion on right coronary MR angiography and vessel wall Imaging. J Magn Reson Imaging. 2001 2001;14:383–390. Trivedi RA, Mallawarachi C, U-King-Im J-M, et al. Identifying inflamed carotid plaques using in vivo USPIO-enhanced MR imaging to label plaque macrophages. Arterioscler Thromb Vasc Biol. 2006;26:1601–1606. Tang T, Howarth S, Miller S, et al. Comparison of the inflammatory burden of truly asymptomatic carotid atheroma with atherosclerotic plaques in patients with asymptomatic carotid stenosis undergoing coronary artery bypass grafting: an ultrasmall superparamagnetic iron oxide enhanced magnetic resonance study. Eur J Vasc Endovasc Surg. 2008;35:392–398. Howarth S, Tang T, Trivedi R, et al. Utility of USPIO-enhanced MR imaging to identify inflammation and the fibrous cap: a comparison of symptomatic and asymptomatic individuals. Eur J Radiol. 2009;70:555–560. Tang T, Howarth S, Miller S, et al. The ATHEROMA (Atorvastatin Therapy: Effects on Reduction of Macrophage Activity) Study. Evaluation using ultrasmall superparamagnetic iron oxide-enhanced magnetic resonance imaging in carotid disease. J Am Coll Cardiol. 2009;53:2039–2050.

127 54. Briley-Saebo KC, Mulder WJM, Mani V, et al. Magnetic resonance imaging of vulnerable atherosclerotic plaques: current imaging strategies and molecular imaging probes. J Magn Reson Imaging. 2007;26:460–479. 55. Balu N, Chu B, Hatsukami T, Yuan C, Yarnykh V. Comparison between 2D and 3D high-resolution black-blood techniques for carotid artery wall imaging in clinically significant atherosclerosis. J Magn Reson Imaging. 2008;27:918–924. 56. Koktzoglou I, Li D. Submillimeter isotropic resolution carotid wall MRI with swallowing compensation: imaging results and semiautomated wall morphometry. J Magn Reson Imaging. 2007; 25:815–823. 57. Zhang Z, Fan Z, Carroll TJ, et al. Three-Dimensional T2-Weighted TSE MRI of the Human Femoral Arterial Vessel Wall at 3.0Tesla. Paper presented at: Proceedings of the 17th ISMRM Scientific Meeting and Exhibition2009; Honolulu. 58. Balu N YV, Chu B, Wang J, Hatsukami T, Yuan C. Carotid Plaque Assessment using Fast 3D Isotropic-Resolution Black-Blood MRI. Paper presented at: Proceedings 17th ISMRM Scientific Meeting and Exhibition2009; Honolulu. 59. Steinman DA, Thomas JB, Ladak HM, Milner JS, Rutt BK, Spence JD. Reconstruction of carotid bifurcation hemodynamics and wall thickness using computational fluid dynamics and MRI. Magn Reson Med. 2002;47:149–159.

wwwwwwwwwwww

9

Noncontrast Coronary Artery Imaging Allison Hays, Robert G. Weiss, and Matthias Stuber

Coronary Magnetic Resonance Angiography Introduction Despite advances in the diagnosis of coronary artery disease (CAD), there still exists a strong need for safer and costeffective techniques to improve visualization of the coronary lumen and vessel wall. Currently, the gold standard for the diagnosis of CAD is coronary X-ray angiography, however, this technique is invasive, costly, and not without risk to the patient. Furthermore, this test does not provide any information about early atherosclerotic disease progression that precedes luminal narrowing. Since its introduction, coronary magnetic resonance angiography (MRA) has been able to overcome some of the disadvantages of X-ray angiography and shows promise in the evaluation of CAD, particularly for proximal coronary disease. Coronary MRA provides a good alternative for patients because it is noninvasive, cost-effective, and without exposure to ionizing radiation. In addition, magnetic resonance techniques can detect abnormalities in the coronary vessel wall before luminal narrowing occurs. The ability to detect early atherosclerotic changes in the vessel wall, assess coronary function, and identify stenotic luminal disease are advantages of coronary MRA that make it a potentially powerful and comprehensive tool for the evaluation of CAD.

A. Hays, MD Division of Cardiology, Department of Medicine, Johns Hopkins Hospital, Baltimore, MD, USA R.G. Weiss, MD Department of Medicine, Johns Hopkins Hospital, Baltimore, MD, USA M. Stuber, PhD () Department of Radiology, Centre Hospitalier Universitaire Vaudois, Center for Biomedical Imaging and University of Lausanne, Lausanne, Switzerland e-mail: [email protected]

However, for successful coronary MRA and coronary vessel wall data acquisition, there are many challenges which must be overcome. The heart is subject to intrinsic and extrinsic motion due to its periodic contraction and relaxation as well as the effect of respiration. Both of these motion components exceed the coronary artery dimensions by a significant degree, making high-resolution data acquisition technically challenging. Therefore, efficient motion suppression strategies must be implemented. In addition, enhanced contrast between the coronary lumen and the surrounding tissue is essential for the visualization of both coronary lumen and the coronary vessel wall. This chapter reviews the technical developments and recent advances that have contributed to the current state of coronary MRA imaging. Furthermore, we discuss the clinical applications as well as the benefits and limitations of current MR approaches. Finally, we address future applications and recent technical developments to improve visualization of the coronary lumen and vessel wall.

Overcoming the Technical Challenges of Coronary MRA Cardiac Motion Suppression The main sources of image artifacts in coronary MRA include cardiac and respiratory motion. To address the issue of cardiac motion, data acquisition is cardiac triggered with the R-wave of the surface electrocardiogram. Data are collected over multiple cardiac cycles (k-space segmentation, Fig. 9.1), and in order to minimize intrinsic cardiac motion, the segmented data are typically acquired in a short acquisition window and during a period of minimal myocardial motion. The duration of the acquisition window must be long enough for an acceptable scan time and short enough to minimize cardiac motion during data acquisition. Although the time of least cardiac motion, or “trigger delay” may be estimated from a patient’s heart rate [1], a more accurate

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_9, © Springer Science+Business Media, LLC 2012

129

130

A. Hays et al.

Fig. 9.1 Suppression of intrinsic myocardial motion is obtained by ECG triggering, image data collection in late diastole during period of minimal myocardial motion and image data collection in a narrow

acquisition window. Segments of the k-space are filled in subsequent cardiac cycles (=k-space segmentation). FFT Fast Fourier Transform

means is to determine a patient-specific trigger delay with a high-resolution cine scout scan [2]. The following coronary MRA scans can then be tailored for data acquisition to occur only during this predefined quiescent period. Typically, a period of minimal myocardial motion occurs in late diastole [3]; however, in patients with high heart-rates, end-systolic imaging may also be beneficial [4]. An alternate approach is with the use of automated algorithms that predict the period of minimal myocardial motion [5–7].

patients often have comorbidities preventing them from complying with prolonged breath-holds, limiting the spatial resolution. Finally, the use of techniques, such as fold-over suppression or signal averaging for signal enhancement, is greatly limited by the achievable breath-hold duration. Therefore, the broad applicability of breath-holds is more restricted in sick or noncompliant patients. As a result, the vast majority of the recent coronary MRA studies did not include breath-holding as a mechanism to suppress respiratory motion artifacts [15–17].

Respiratory Motion Compensation Free-Breathing Coronary MRA Breath-Hold Coronary MRA The effect of respiratory motion on coronary MRA poses a major challenge for coronary imaging. Two-dimensional (2D) breath-hold techniques were implemented early to suppress respiratory motion artifacts for coronary imaging [8]. The goal of this 2D approach was to acquire contiguous images of the proximal segments of the coronary arteries during serial breath-holds. Recently, with the initiation of steady-state with free precession (SSFP) in combination with parallel imaging, 3D data collection during a single breathhold became technically feasible [9–13]. In general, the advantage of the breath-hold approach is that rapid imaging is possible, and it is relatively easy to perform in wellmotivated subjects. However, there are several limitations associated with the breath-hold strategy. The position of the diaphragm may vary significantly during repeated breathholds and during a sustained breath-hold, and there is frequently upward diaphragmatic drift of approximately 1 cm [14]. Data acquisition during serial breath-holds may also result in mis-registration gaps that may appear as signal voids in the coronaries and lead to misinterpretation. In addition,

Alternate methods to breath-hold approaches include freebreathing coronary MRA with the use of respiratory navigators [18–21]. Using this method, many ECG-gated data are acquired but only those where the navigator-identified lung–liver interface position falls within a prespecified endexpiratory gating window are included for image reconstruction. Navigator gating can be enhanced with prospective adaptive slice tracking. With this technology, the imaged slice position is adjusted in real-time to account for the residual diaphragmatic displacement within the gating window [22]. This helps to overcome some of drawbacks associated with prolonged scan times and narrow gating windows and has resulted in equivalent or improved image quality with submillimeter spatial resolution and abbreviated scan times [14, 22, 23]. Typically, respiratory data collection occurs at end expiration when the diaphragm position is most consistent within the respiratory cycle (Fig. 9.2). Therefore, respiratory gating lengthens the scan time because no data are collected in the remainder of the respiratory cycle and on average data are collected during 50% of the R-R intervals [2].

9

Noncontrast Coronary Artery Imaging

Fig. 9.2 The navigator signal is typically obtained from the lung–liver interface. A computer algorithm detects the position of that interface (lung–liver interface) in real-time. If the computer identified lung–liver interface position is found inside of the gating window (window width can be adjusted by the user), the k-space segment that is collected

131

immediately after the navigator is accepted for reconstruction. However, if the lung–liver interface position is found outside of the gating window, that k-space segment has to be remeasured in a later cardiac cycle. Adding navigator gating to an imaging sequence typically increases scanning time by a factor of ~2

One study directly compared the quality of 3D coronary MRA acquired during a single breath-hold versus a realtime navigator-gated free breathing technique in patients with suspected CAD [24]. In this study, it was found that navigator gated and corrected coronary MRA was improved with respect to diagnostic accuracy of stenosis quantification and image quality.

Contrast Enhancement Coronary MRA examinations are typically performed without the addition of intravenously administered contrast agents. The relative signal of the coronary artery lumen is augmented by taking advantage of the natural T2 differences between blood and the surrounding myocardium by using fat-saturation prepulses [8], magnetization transfer contrast prepulses [25] or T2 preparatory pulses (Fig. 9.3) [26, 27]. With these techniques, the coronary lumen appears bright while the surrounding myocardium appears dark with reduced signal intensity. Novel intravascular contrast agents offer the ability to improve spatial resolution and SNR in coronary imaging. The advantages of these blood pool contrast agents are that the plasma half-life is longer and that they do not extravasate as quickly into the extracellular space as Gadolinium does [28, 29]. These factors result in reduced myocardial signal, greater blood pool enhancement, and allow image acquisition over longer time periods after intravenous administration of the contrast agent. Therefore, intravascular contrast agents are well-suited to be combined with navigator technology and prolonged scanning times. However, the use of specific intravascular or extracellular contrast agents are discussed in another chapter.

Fig. 9.3 Baseline images (a, c) and T2prep-enhanced images (b, d) of 3D coronary data set. Application of T2prep (b, d) suppressed cardiac muscle as well as skeletal muscle in the chest and back. Due to enhanced blood-to-muscle contrast in T2prep images (b, d), visualization of LAD and LCx is improved compared with reference images (a, c). Suppression of venous blood additionally helps to differentiate between great cardiac vein (GCV), LAD, and LCx, which is difficult to distinguish in image (a) (Reprinted with permission from Botnar et al. [27])

Whole-Heart Coronary MRA One of the challenges of coronary MRA is that extensive planning and scout scanning is required prior to imaging the coronaries adding to the complexity of the scan and

132

reducing the ease-of-use. An alternative approach that circumvents this issue is whole-heart imaging [30] which utilizes a volumetric approach to improve coverage (Fig. 9.4). Arbitrary views can then be reconstructed during postprocessing without compromising image quality and can enable visualization of more distal coronary segments. Whole-heart coronary MRA is typically performed during free breathing with navigator gating technology, and has progressed largely due to advances in parallel imaging and SSFP, which has high endogenous contrast. Contrast enhancement is achieved using spatially selective fat saturation and T2 preparation [30, 31]. One of the limitations of whole-heart imaging is the relatively prolonged scan times (approximately 10–15 min) that may be susceptible to slow diaphragmatic drift during the relatively long scanning time. However, there are currently mechanisms in place to compensate for such diaphragmatic drift that can occur with time [32, 33]. The use of the whole-heart technique was investigated in a study of 20 CAD patients and reported a sensitivity of 82% and a specificity of 91% for the identification of significant CAD when compared to X-ray coronary angiography in a single-center setting [34]. The average scan time for the study was less than 15 min during free breathing. Further improvements in image quality and scan time are anticipated with the development of larger cardiac coil arrays for coronary imaging and with high field imaging [35, 36].

Fig. 9.4 Example of the right and left coronary arteries using whole heart bright blood SSFP imaging in a 71-year-old healthy volunteer. Imaging parameters used were: TR = 279 ms, TE = 1.6 ms, slice thickness 0.7 mm, in plane resolution of 1.1 × 1.1 mm with an imaging time of 10 min (Image courtesy of Kai Lin and Debiao Li, Northwestern University, Chicago, IL)

A. Hays et al.

Visualization of CAD Identification of Coronary Stenosis Current breath-hold coronary MRA techniques were shown to accurately identify proximal coronary stenoses in several clinical studies. Gradient-echo techniques depict focal stenoses as signal voids (Fig. 9.5). In an early patient study, a segmented k-space 2D breath-hold ECG-gated gradientecho sequence was used to compare coronary MRA to the gold standard of X-ray coronary angiography [37]. This single-center 2D coronary MRA technique yielded a sensitivity of 90% and a specificity of 92% for correctly classifying individual vessels with significant CAD defined as greater than 50% diameter stenosis on X-ray angiography. Other studies have subsequently reported variable sensitivity and specificity values for the detection of significant CAD [38–43]. The variability in these studies may be due to differences in the MR sequences employed, patient selection, or arrhythmias which may degrade image quality. Newer breath-hold [44] and nonbreath-hold approaches for 3D coronary MRA have also demonstrated the ability of this technique to detect coronary stenoses. An international multicenter trial prospectively compared X-ray coronary angiography with coronary MRA using common hardware, software, and methodology [45]. This trial showed that freebreathing submillimeter 3D coronary MRA can accurately identify significant proximal and mid coronary disease while nonsignificant coronary disease can be excluded with high certainty. Although the specificity of the technique remains to be improved, it is likely that advances in hardware and software in combination with a higher magnetic field strength may reduce the false positive rate and improve accuracy for

Fig. 9.5 (a) Radiograph coronary angiogram of a patient with a 50% luminal stenosis of the mid-LAD (arrow). (b) Transverse multiplanar reformatted free-breathing 3D Balanced FFE coronary MRA acquired in the same patient. The location of the lesion (dotted arrow) corresponds to that of the radiograph image (Reprinted with permission from Spuentrup et al. [88])

9

Noncontrast Coronary Artery Imaging

133

the identification of more distal stenoses. Clearly, the above mentioned whole-heart approach is ready to be tested in a multicenter setting.

Coronary MRA for Coronary Artery Bypass Graft Assessment The utility of coronary MRA for the assessment of coronary bypass graft patency has been investigated since the 1980s and has made significant improvement in recent years. In early studies using ECG-triggered spin-echo [46] and gradient echo techniques [46, 47], visualization of grafts and graft patency was limited by cardiac and respiratory motion artifacts. As a result, this restricted the ability to accurately quantify graft stenosis. More recently, data collection during a single breath-hold became possible [48]. However, while all occluded grafts were successfully identified using breathholding techniques [48], only 67% of the patent grafts were correctly diagnosed. Some of the limitations of these earlier techniques occurred because of problems with the patient’s ability to perform breath-holds and the inherent dependence of the maximum achievable spatial resolution on the breathhold duration, or because of diaphragmatic drift. More advanced approaches for respiratory motion compensation, such as the use of retrospective respiratory navigators [20] has enabled 3D acquisition with high in-plane spatial resolution and removed the constraints related to breath-holds [49]. This method increased both sensitivity and specificity for the assessment of graft patency (87% and 100%, respectively); however, the detection of luminal stenosis was not as reliable. More recently, navigator gated and corrected 3D coronary MRA has been shown to accurately graft identify occlusion or stenosis [50]. Using this method, a sensitivity of 83% and a specificity of 98% for the definition of graft occlusion were obtained, and a reasonable diagnostic accuracy for the assessment of vein graft stenosis severity was reported. A subsequent study of bypass graft patency with steady-state free-precession angiography found a comparable sensitivity although with reduced specificity for graft stenosis severity compared to other techniques [51]. A practical limitation of coronary MRA bypass graft assessment is related to local signal loss/artifacts due to nearby metallic objects, such as hemostatic clips, stainless steel graft markers, and sternal wires. The limited ability to consistently identify severely diseased yet patent grafts is also a hindrance to clinical utility and acceptance.

Visualization of Anomalous Coronary Arteries and Coronary Aneurysms Traditionally, X-ray angiography has been the imaging test of choice for the diagnosis of coronary artery anomalies, a rare cause of myocardial ischemia and sudden death among

Fig. 9.6 A young woman with a history of Kawasaki disease. Multiplanar reformatted coronary MRA images of the left (a) and right (c) coronary arteries are compared with selective X-ray angiograms of the left (b) and right (d) coronary arteries. Two coronary artery aneurysms (CAA) in the left coronary and one in the right coronary (a through d, black arrows), as well as a stenosis between the two left CAAs (a and b, white arrows) are shown (Reprinted with permission from Greil et al. [61])

young adults. However, this technique is limited in its ability to define the course of the anomalous coronaries particularly with regards to the great vessels. Several published series [52–55] have documented a good correlation between coronary MRA with X-ray angiography in the characterization of anomalous coronaries. Early coronary MRA studies often used a 2D breath-hold gradient echo approach [52–58]. These 2D coronary MRA studies uniformly reported high accuracy, including several studies in which coronary MRA was determined to be superior to X-ray angiography [53, 54]. Most centers currently use 3D coronary MRA because of superior reconstruction capabilities with similar results [59]. As a result, coronary MRA is now the preferred imaging modality for the evaluation of anomalous coronary arteries [60]. Though coronary artery aneurysms are relatively uncommon, recent studies indicate an important role for coronary MRA for the assessment of this condition [39] and in the evaluation of coronary ectasia. 3D coronary MRA has shown to be valuable in the characterization of coronary aneurysms in pediatric patients with Kawasaki disease (Fig. 9.6), a rare small vessel vasculitic disease [61]. A strong correlation between coronary MRA and X-ray coronary angiography has also been reported for ectatic coronary arteries [62].

134

Fig. 9.7 Black-blood 3D RCA vessel wall image showing the anterior and the posterior coronary walls (arrows). The dotted arrows point to the contrast between the tissue in the path of the cylindrical pulse and the surrounding tissue by use of the local inversion prepulse (Reprinted with permission, Desai et al. [69])

A. Hays et al.

window and therefore, less susceptibility to intrinsic cardiac motion. In one study using a free-breathing 3D approach, the coronary vessel wall could be visualized with an inplane resolution of 0.78–1.0 mm in both healthy adults and those with nonsevere CAD [64]. In the CAD patients, the mean vessel wall thickness was significantly increased when compared to that of normal volunteers, whereas as the lumen diameter was similar, illustrating positive arterial remodeling. In another study of patients with type I diabetes, coronary vessel wall imaging using a similar blackblood approach revealed significantly greater wall thickness in diabetic patients with nephropathy than those with normal kidney function [68]. Therefore, black-blood coronary MRA represents a powerful investigative tool to noninvasively quantify subclinical atherosclerosis in a variety of patient populations. Preliminary studies of coronary vessel wall imaging at 3 T are promising and show the potential to detect preclinical disease and monitor treatment effects over time [69, 70].

Coronary Flow Imaging Coronary Vessel Wall Imaging Using Black-Blood Coronary MRA Early changes in the development of atherosclerosis include outward “Glagov” arterial remodeling with relative preservation of the lumen [63]. Because thickening of the vessel wall precedes luminal narrowing, MRI has the ability to detect early coronary atherosclerosis (Fig. 9.7). Conventional imaging of the coronary lumen, such as with X-ray angiography, may significantly underestimate the degree of subclinical atherosclerosis. Using black-blood MRA imaging techniques, the coronary vessel wall can be visualized and vessel wall thickness and plaque volume quantified [64]. Most black-blood imaging techniques involve a dual inversion approach to suppress luminal blood based on both its T1 properties and flow, and an initial inversion pulse is immediately followed by a spatially selective reinversion pulse to restore magnetization along the vessel. Initially, single cross-sectional slices of the coronary vessel wall were obtained during breath-holds and vessel wall thickness was measured in a subset of cases [65]. Subsequently, this technique has been refined using respiratory navigators for free-breathing data acquisition [66]. Using a localinversion 3D spiral technique [67], a large anatomical coverage of the coronaries can be obtained with thinner reconstructed slices than previously possible with 2D approaches. Although scan times are prolonged using this free-breathing technique (approximately 10–15 min), the advantage includes data acquisition in a very short acquisition

MR flow mapping has been validated as a noninvasive means to assess coronary flow velocity and velocity reserve, and has a strong correlation to measurements obtained using the gold standard, Doppler guidewire [71]. In response to stress, prior invasive coronary artery studies documented similar increases in peak diastolic coronary flow velocity as that obtained using MRI [71, 72]. The clinical utility of MR flow studies of the coronaries has been examined in specific populations of CAD patients, including after bypass grafting or stent placement. In one study, MR flow measurements using phase contrast methods permitted the noninvasive evaluation of coronary flow in CAD patients after percutaneous coronary intervention and showed similar results to Doppler guidewire measurements [73]. In this study, the authors reported a diagnostic accuracy of 86% for the detection of significant in-stent restenosis defined as >50% arterial narrowing. In a study of 69 coronary bypass patients, velocity-encoded flow mapping by MR was performed at baseline and with vasodilator stress to assess flow in the bypass grafts [74]. Using X-ray angiography as the standard, a sensitivity of 96% and specificity of 92% for the identification of stenosis >70% was reported. One limitation, however, was that flow scans could only be obtained in 80% of grafts because of suboptimal image quality. The measurement of coronary flow using MRI may provide a valuable noninvasive alternative to monitor the hemodynamic significance of coronary stenoses in a variety of patients with CAD.

9

Noncontrast Coronary Artery Imaging

135

Newer Technical Developments Radial and Spiral Imaging Techniques Non-Cartesian methods of k-space sampling, such as radial and spiral trajectories, have gained more widespread use in coronary MRA because of advantages in reducing motion artifacts and improved sampling efficiency. One benefit with radial imaging is a low inherent sensitivity to motion. In addition, it is a robust technique with respect to undersampling, yielding adequate image quality during a single breathhold for a targeted volume approach [75]. 3D Radial sampling applied to whole-heart imaging may lead to a loss in SNR because image reconstruction entails a nonuniform weighting of data. However, with a 3D radial approach, isotropic spatial resolution for whole-heart imaging has been achieved [31]. With further developments in surface coils to generate the high signal required and advanced reconstruction algorithms, this may potentially improve radial sampling techniques for coronary applications. Spiral sampling techniques have also been applied to generate high-quality submillimeter resolution coronary images using 2D techniques (Fig. 9.8) [76]. Advantages of this approach are that flow artifacts can be minimized and SNR increased because k-space sampling is more efficient. Spiral sampling algorithms have been applied successfully for acquiring high-resolution coronary MRA using a freebreathing real-time navigator approach [77], 3D acquisitions [78], and a segmented interleaved strategy [79]. One study directly compared 3D spiral coronary MRA to 3D Cartesian acquisition and found superior results in image quality, SNR and CNR with the spiral approach [77]. Therefore, spiral sampling techniques show promise for high-resolution coronary imaging and warrant further study.

Arterial Spin Labeling To improve the visualization of the vessel lumen and suppress signal from surrounding structures, an MR subtraction technique called arterial spin labeling can be employed to exclusively image the coronary lumen [80], analogous to images obtained using conventional X-ray angiography. A 2D spatially selective inversion pulse is applied to tag blood in the aortic root [81] and after a short-time delay for wash-in, the labeled blood in the proximal coronaries may be imaged using a volume slab approach (Fig. 9.9). An advantage of arterial spin-labeling techniques is that they enable the depiction of the lumen of the coronary tree alone at varying angles. Furthermore, no postprocessing is required to facilitate analysis. However, a drawback to this technique includes the need for two separate acquisitions, thus prolonging scan time [82, 83]. Recently, an arterial

Fig. 9.8 Coronary MRA images. Reformatted images obtained from different subjects are shown. Examples for the mid RCA and proximal RCA and LM/LAD/LCX portions of the vessels are well visualized. Additionally, some anatomical structures are indicated (RV right ventricle, LV left ventricle, Ao aorta, and PA pulmonary artery). The left column represents data sets obtained using the Cartesian segmented k-space gradient-echo method. The middle column shows data obtained using the single-interleaf spiral imaging approach. The measuring times for the Cartesian and single-interleaf spiral approaches were similar. The right column shows data sets measured using the double-interleaf spiral approach obtained in half of the total measuring time used for a Cartesian scan (Reprinted with permission from Bornert et al. [77])

spin labeling with local reinversion was developed that takes advantage of inflow and natural differences in T1, and avoids subtraction, thus abbreviating scan time [84, 85]. In the future, arterial spin-labeling techniques may benefit from the improved SNR and reduced decay of labeled product (due to prolonged T1 values) inherent to higher field imaging.

136

Fig. 9.9 Local reinversion coronary MR angiography obtained in a healthy adult subject. A segment of the right coronary artery (RCA) is displayed with high visual signal intensity, contrast, and vessel delineation. Simultaneously, a cross-section of the left main (LM), a short proximal segment of the left anterior descending (LAD), and a proximal segment of the left circumflex coronary artery (LCX) are visualized with high signal intensity. AO aorta, RV right ventricle (Reprinted with permission from Katoh et al. [84])

Steady-State Free-Precession Coronary MRA The application of SSFP imaging of the heart enhances contrast between the ventricular blood pool and the myocardium without the requirement of exogenous contrast enhancement. SSFP imaging has been employed widely using either using a single breath-hold [86] or free-breathing with respiratory navigators [87]. Given the inherent properties of SSFP imaging with high SNR, this sequence may be potentially useful for contrast enhancement in 3D coronary MRA, in which in-flow effects are in general decreased due to thick slab excitations [86]. When combined with fat saturation and T2 preparation techniques, SSFP approaches lead to images with a high contrast and vessel sharpness (Fig. 9.10) [88]. An initial study performed using an animal model [87] observed that free-breathing SSFP imaging resulted in high-quality coronary MRA when compared with standard T2-prepared gradient-echo imaging with substantial improvements in SNR, CNR, and vessel sharpness. Although spiral imaging demonstrated the highest SNR, SSFP imaging yielded the highest vessel definition and image quality score. In another study of healthy volunteers, reliable fat suppression and a significant increase in the blood-myocardial CNR was achieved postcontrast using SSFP techniques [89]. At 3 T, both SSFP and segmented k-space gradient echo imaging have been investigated. In an early study, segmented k-space gradient echo imaging was found to be superior to SSFP imaging which was attributed in part to greater magnetic field susceptibility at 3 T as well as limitations in specific

A. Hays et al.

Fig. 9.10 Navigator gated and corrected free-breathing coronary MRA data acquired in a healthy subject (male, 40 years) using a segmented 3D Balanced FFE imaging sequence (TR = 4 ms, TE = 2 ms). (a) Image displays a double-oblique view of the right coronary artery (RCA) and a left circumflex (LCX) with a high visual vessel definition (dashed arrows). (b) Transverse imaging plane demonstrating a left coronary system, including the left main (LM), the left anterior descending (LAD), the left circumflex (LCX) and a great cardiac vein (GCV). A signal attenuated “dark rim” delineating the coronary arteries is readily apparent (dashed arrows) (Reprinted with permission from Spuentrup et al. [88])

absorption rate (SAR) [90]. However, it should be noted that the combination of SSFP and parallel imaging techniques has made it possible to acquire whole-heart data as noted above [30]. This major step forward has significantly improved ease-of-use over volume-targeted approaches and allowed access to more distal vessels.

Parallel Imaging for Coronary MRA Parallel imaging techniques decrease acquisition time and enable an entire data set to be collected in one cardiac cycle. The development of such rapid parallel imaging approaches, such as “SENSE” [91], “SMASH” [92], or GRAPPA [93, 94], have been shown to reduce scan time for cardiac MRI substantially [95] and offer great potential to enhance coronary imaging. However, the trade-off is reduced SNR, which may be an important consideration in some applications. Parallel imaging techniques have been employed at higher magnetic field strengths, and have resulted in excellent image quality with abbreviated scanning times [96].

High Field Coronary MR Imaging Imaging at higher field strengths has the potential of higher SNR, higher spatial resolution, and shorter scanning times compared to lower magnetic field strengths. Three Tesla systems have now been in wide clinical use, and several reports

9

Noncontrast Coronary Artery Imaging

137

design will likely yield continued progress and address many of the imaging challenges specific to 7 T imaging.

Conclusion

Fig. 9.11 Representative examples of MR images of the RCA obtained at 7 T. Proximal (a) and more distal (b) segments of the RCA are visualized. The 13-cm RF transmit and receive coil provides sufficient penetration depth to visualize a considerable RCA segment length. Sharply delineated contiguous coronary segments with good contrast between the coronary lumen blood pool and the epicardial fat are shown (Reprinted with permission from van Elderen et al. [104])

have shown promising results using similar techniques as at 1.5 T [96, 97]. However, scanning at 3 T and higher field strengths pose several technical challenges, including increased susceptibility artifacts particularly at tissue borders and an increased magnetohydrodynamic effect affecting ECG triggering. Despite these limitations, recent advances in hardware and software (i.e., vector ECG [98], higher order shimming, B1 shimming [99, 100]) have enabled high-resolution free-breathing scans of at least comparable image quality to that obtained at lower field strengths. Coronary MRA may be improved at higher field strengths and an initial study of coronary angiography at 3 T in healthy adults reported a higher spatial resolution compared to that achievable at 1.5 T [97]. Sommer and coworkers directly compared early 3 T to 1.5 T coronary MRA to assess the accuracy of diagnosing CAD compared to the standard of X-ray coronary angiography [101]. Using navigator-corrected, 3D segmented k-space gradient-echo techniques at both field strengths, they found comparable image quality with a 30% increase in SNR and a 22% increase in CNR at 3 T. Overall, the diagnostic accuracy at both field strengths was equivalent, with sensitivity for the detection of CAD of 82% for both, and a specificity of 89% and 88% for 3 T and 1.5 T, respectively. However, newer techniques that were not employed at the time, such as adiabatic T2 preparation pulses [102], parallel imaging [96], or advanced shimming algorithms [103], will likely contribute to superior results of coronary MRA at higher field strengths. More recently, commercial human 7 T MR systems have become available and represent a potentially powerful tool for coronary imaging. Initial studies demonstrate the feasibility of in vivo human coronary MRA at 7 T using custombuilt coils and vector ECG hardware to address some of the obstacles found with high field imaging (Fig. 9.11) [104]. Further developments in contrast enhancement and coil

Coronary MRA, because of its noninvasive nature and the capacity for soft tissue characterization, has emerged as a powerful modality to evaluate the coronary lumen and vessel wall. Current MRA techniques can reliably identify both anomalous coronary arteries and coronary aneurysms; however, there is still limited multicenter data to suggest that MRA is comparable to X-ray angiography for the detection of stenotic disease. Despite this, it has a high negative predictive value for the assessment of both proximal and multivessel coronary disease. A further increase in both spatial resolution and contrast to noise ratio is needed before coronary MRA can be used to screen asymptomatic patients or to precisely characterize focal coronary stenosis. Although computed tomography angiography is an alternative imaging modality with higher spatial resolution than MRI, it has several limitations including the exposure of patients to ionizing radiation and nephrotoxic contrast agents. Its application to studies of heavily calcified vessels and coronary function is limited, as coronary blood-flow velocity cannot be quantified. In addition, the radiation and contrast doses limit studies in low-risk subjects, repeated studies in patients over time or with stress, and evaluation of patients with renal disease, all of which are important for screening populations and monitoring responses to therapy. Because of its multipurpose nature, coronary MRA is a well-suited imaging modality for the comprehensive characterization of CAD. With further refinement of techniques, including improvements in hardware and high field imaging, coronary MRA will likely emerge as the premier modality to characterize tissue in the coronary vessel wall and to evaluate coronary function.

References 1. Kim WY, Stuber M, Kissinger KV, Andersen NT, Manning WJ, Botnar RM. Impact of bulk cardiac motion on right coronary MR angiography and vessel wall imaging. J Magn Reson Imaging. 2001;14:383–390. 2. Stuber M, Botnar RM, Danias PG, et al. Double-oblique freebreathing high resolution three-dimensional coronary magnetic resonance angiography. J Am Coll Cardiol. 1999;34:524–531. 3. Hofman MB, Wickline SA, Lorenz CH. Quantification of in-plane motion of the coronary arteries during the cardiac cycle: implications for acquisition window duration for MR flow quantification. J Magn Reson Imaging. 1998;8:568–576. 4. Duerinckx A, Atkinson DP. Coronary MR angiography during peak-systole: work in progress. J Magn Reson Imaging. 1997; 7:979–986.

138 5. Jahnke C, Paetsch I, Nehrke K, et al. A new approach for rapid assessment of the cardiac rest period for coronary MRA. J Cardiovasc Magn Reson. 2005;7:395–399. 6. Roes SD, Korosoglou G, Schar M, et al. Correction for heart rate variability during 3D whole heart MR coronary angiography. J Magn Reson Imaging. 2008;27:1046–1053. 7. Wang Y, Grimm RC, Felmlee JP, Riederer SJ, Ehman RL. Algorithms for extracting motion information from navigator echoes. Magn Reson Med. 1996;36:117–23. 8. Edelman RR, Manning WJ, Burstein D, Paulin S. Coronary arteries: breath-hold MR angiography. Radiology. 1991;181:641–643. 9. Bornert P, Jensen D. Coronary artery imaging at 0.5 T using segmented 3D echo planar imaging. Magn Reson Med. 1995;34:779–785. 10. Goldfarb JW, Edelman RR. Coronary arteries: breath-hold, gadolinium-enhanced, three-dimensional MR angiography. Radiology. 1998;206:830–834. 11. Stuber M, Botnar RM, Danias PG, Kissinger KV, Manning WJ. Breathhold three-dimensional coronary magnetic resonance angiography using real-time navigator technology. J Cardiovasc Magn Reson. 1999;1:233–238. 12. Wang Y, Grimm RC, Rossman PJ, Debbins JP, Riederer SJ, Ehman RL. 3D coronary MR angiography in multiple breath-holds using a respiratory feedback monitor. Magn Reson Med. 1995; 34:11–16. 13. Wielopolski PA, van Geuns RJ, de Feyter PJ, Oudkerk M. Breathhold coronary MR angiography with volume targeted imaging. Radiology. 1998;209:209–219. 14. Danias PG, McConnell MV, Khasgiwala VC, Chuang ML, Edelman RR, Manning WJ. Prospective navigator correction of image position for coronary MR angiography. Radiology. 1997;203:733–736. 15. Lai P, Bi X, Jerecic R, Li D. A respiratory self-gating technique with 3D-translation compensation for free-breathing whole-heart coronary MRA. Magn Reson Med. 2009;62:731–738. 16. Sakuma H, Ichikawa Y, Chino S, Hirano T, Makino K, Takeda K. Detection of coronary artery stenosis with whole-heart coronary magnetic resonance angiography. J Am Coll Cardiol. 2006; 48:1946–1950. 17. Yang Q, Li K, Liu X, et al. Contrast-enhanced whole-heart coronary magnetic resonance angiography at 3.0-T: a comparative study with X-ray angiography in a single center. J Am Coll Cardiol. 2009;54:69–76. 18. Ehman RL, Felmlee JP. Adaptive technique for high-definition MR imaging of moving structures. Radiology. 1989;173:255–263. 19. Ehman RL, Felmlee JP. Flow artifact reduction in MRI: a review of the roles of gradient moment nulling and spatial presaturation. Magn Reson Med. 1990;14:293–307. 20. Li D, Kaushikkar S, Haacke EM, et al. Coronary arteries: threedimensional MR imaging with retrospective respiratory gating. Radiology. 1996;201:857–863. 21. Oshinski JN, Hofland L, Mukundan S, Jr., Dixon WT, Parks WJ, Pettigrew RI. Two-dimensional coronary MR angiography without breath holding. Radiology. 1996;201:737–743. 22. McConnell MV, Khasgiwala VC, Savord BJ, et al. Prospective adaptive navigator correction for breath-hold MR coronary angiography. Magn Reson Med. 1997;37:148–152. 23. Stuber M, Botnar RM, Danias PG, Kissinger KV, Manning WJ. Submillimeter three-dimensional coronary MR angiography with real-time navigator correction: comparison of navigator locations. Radiology. 1999;212:579–587. 24. Jahnke C, Paetsch I, Nehrke K, et al. Rapid and complete coronary arterial tree visualization with magnetic resonance imaging: feasibility and diagnostic performance. Eur Heart J. 2005;26: 2313–2319.

A. Hays et al. 25. Li D, Paschal CB, Haacke EM, Adler LP. Coronary arteries: threedimensional MR imaging with fat saturation and magnetization transfer contrast. Radiology. 1993;187:401–406. 26. Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishimura DG. Coronary angiography with magnetizationprepared T2 contrast. Magn Reson Med. 1995;33:689–96. 27. Botnar RM, Stuber M, Danias PG, Kissinger KV, Manning WJ. Improved coronary artery definition with T2-weighted, freebreathing, three-dimensional coronary MRA. Circulation. 1999;99:3139–3148. 28. Huber ME, Paetsch I, Schnackenburg B, et al. Performance of a new gadolinium-based intravascular contrast agent in free-breathing inversion-recovery 3D coronary MRA. Magn Reson Med. 2003;49:115–121. 29. Li D, Zheng J, Weinmann HJ. Contrast-enhanced MR imaging of coronary arteries: comparison of intra- and extravascular contrast agents in swine. Radiology. 2001;218:670–678. 30. Weber OM, Martin AJ, Higgins CB. Whole-heart steady-state free precession coronary artery magnetic resonance angiography. Magn Reson Med. 2003;50:1223–1228. 31. Stehning C, Bornert P, Nehrke K, Eggers H, Dossel O. Fast isotropic volumetric coronary MR angiography using free-breathing 3D radial balanced FFE acquisition. Magn Reson Med. 2004; 52:197–203. 32. Huber ME, Hengesbach D, Botnar RM, et al. Motion artifact reduction and vessel enhancement for free-breathing navigatorgated coronary MRA using 3D k-space reordering. Magn Reson Med. 2001;45:645–652. 33. Jhooti P, Gatehouse PD, Keegan J, Bunce NH, Taylor AM, Firmin DN. Phase ordering with automatic window selection (PAWS): a novel motion-resistant technique for 3D coronary imaging. Magn Reson Med. 2000;43:470–480. 34. Sakuma H, Ichikawa Y, Suzawa N, et al. Assessment of coronary arteries with total study time of less than 30 minutes by using whole-heart coronary MR angiography. Radiology. 2005; 237:316–321. 35. Buehrer M, Huber ME, Wiesinger F, Boesiger P, Kozerke S. Coil setup optimization for 2D-SENSE whole-heart coronary imaging. Magn Reson Med. 2006;55:460–464. 36. Niendorf T, Sodickson DK, Hardy CJ, et al. Towards whole heart coverage in a single breath-hold: Coronary artery imaging using a true 32-channel phased array MRI system. In: International Society for Magnetic Resonance Imaging; 2004; Kyoto, Japan: ISMRM; 2004. p. 703. 37. Manning WJ, Li W, Edelman RR. A preliminary report comparing magnetic resonance coronary angiography with conventional angiography. N Engl J Med. 1993;328:828–832. 38. Cheng L, Gao Y, Guaricci AI, et al. Breath-hold 3D steady-state free precession coronary MRA compared with conventional X-ray coronary angiography. J Magn Reson Imaging. 2006;23: 669–673. 39. Duerinckx AJ, Urman MK. Two-dimensional coronary MR angiography: analysis of initial clinical results. Radiology. 1994;193:731–738. 40. Müller MF, Fleisch M, Kroeker R, Chatterjee T, Meier B, Vock P. Proximal coronary artery stenosis: three-dimensional MRI with fat saturation and navigator echo. J Magn Reson Imaging. 1997;7:644–651. 41. Pennell DJ, Bogren HG, Keegan J, Firmin DN, Underwood SR. Assessment of coronary artery stenosis by magnetic resonance imaging. Heart. 1996;75:127–133. 42. Post JC, van Rossum AC, Hofman MB, de Cock CC, Valk J, Visser CA. Clinical utility of two-dimensional magnetic resonance angiography in detecting coronary artery disease. Eur Heart J. 1997;18:426–433.

9

Noncontrast Coronary Artery Imaging

43. Post JC, van Rossum AC, Hofman MB, Valk J, Visser CA. Threedimensional respiratory-gated MR angiography of coronary arteries: comparison with conventional coronary angiography. AJR Am J Roentgenol. 1996;166:1399–1404. 44. van Geuns RJ, Wielopolski PA, de Bruin HG, et al. MR coronary angiography with breath-hold targeted volumes: preliminary clinical results. Radiology. 2000;217:270–277. 45. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance angiography for the detection of coronary stenoses. N Engl J Med. 2001;345:1863–1869. 46. White RD, Caputo GR, Mark AS, Modin GW, Higgins CB. Coronary artery bypass graft patency: noninvasive evaluation with MR imaging. Radiology. 1987;164:681–686. 47. Aurigemma GP, Reichek N, Axel L, Schiebler M, Harris C, Kressel HY. Noninvasive determination of coronary artery bypass graft patency by cine magnetic resonance imaging. Circulation. 1989;80:1595–1602. 48. von Smekal A, Knez A, Seelos KC, et al. A comparison of ultrafast computed tomography, magnetic resonance angiography and selective angiography for the detection of coronary bypass patency [in German]. Rofo. 1997;166:185–191. 49. Kessler W, Achenbach S, Moshage W, et al. Usefulness of respiratory gated magnetic resonance coronary angiography in assessing narrowings > or = 50% in diameter in native coronary arteries and in aortocoronary bypass conduits. Am J Cardiol. 1997;80: 989–993. 50. Langerak SE, Vliegen HW, de Roos A, et al. Detection of vein graft disease using high-resolution magnetic resonance angiography. Circulation. 2002;105:328–333. 51. Bunce NH, Lorenz CH, John AS, Lesser JR, Mohiaddin RH, Pennell DJ. Coronary artery bypass graft patency: assessment with true as imaging with steady-state precession versus gadoliniumenhanced MR angiography. Radiology. 2003;227:440–446. 52. McConnell MV, Ganz P, Selwyn AP, Li W, Edelman RR, Manning WJ. Identification of anomalous coronary arteries and their anatomic course by magnetic resonance coronary angiography. Circulation. 1995;92:3158–3162. 53. Taylor AM, Thorne SA, Rubens MB, et al. Coronary artery imaging in grown up congenital heart disease: complementary role of magnetic resonance and x-ray coronary angiography. Circulation. 2000;101:1670–1678. 54. Post JC, van Rossum AC, Bronzwaer JG, et al. Magnetic resonance angiography of anomalous coronary arteries. A new gold standard for delineating the proximal course? Circulation. 1995;92:3163–171. 55. Vliegen HW, Doornbos J, de Roos A, Jukema JW, Bekedam MA, van der Wall EE. Value of fast gradient echo magnetic resonance angiography as an adjunct to coronary arteriography in detecting and confirming the course of clinically significant coronary artery anomalies. Am J Cardiol. 1997;79:773–776. 56. Machado C, Bhasin S, Soulen RL. Confirmation of anomalous origin of the right coronary artery from the left sinus of Valsalva with magnetic resonance imaging. Chest. 1993;104:1284–1286. 57. Doorey AJ, Wills JS, Blasetto J, Goldenberg EM. Usefulness of magnetic resonance imaging for diagnosing an anomalous coronary artery coursing between aorta and pulmonary trunk. Am J Cardiol. 1994;74:198–199. 58. Razmi RM Meduri A, Chun W, Rathi VK, Pohost GM. Coronary magnetic resonance angiography (CMRA): The gold standard for determining the proximal course of anomalous coronary arteries [abstract]. J Am Coll Cardiol. 2001:37:380. 59. Bunce NH, Lorenz CH, Keegan J, et al. Coronary artery anomalies: assessment with free-breathing three-dimensional coronary MR angiography. Radiology. 2003;227:201–8. 60. Bluemke DA, Achenbach S, Budoff M, et al. Noninvasive coronary artery imaging: magnetic resonance angiography and multidetector

139

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76. 77.

computed tomography angiography: a scientific statement from the American Heart Association committee on cardiovascular imaging and intervention of the council on cardiovascular radiology and intervention, and the councils on clinical cardiology and cardiovascular disease in the young. Circulation. 2008;118:586–606. Greil GF, Stuber M, Botnar RM, et al. Coronary magnetic resonance angiography in adolescents and young adults with kawasaki disease. Circulation. 2002;105:908–911. Mavrogeni S MA, Papadakis E, Foussas S, Douskou M, Baras P, et al. Correlation between magnetic resonance angiography (MRA) and quantitative coronary angiography (QCA) in ectatic coronary vessels. J Cardiovasc Magn Reson. 2004;6:17–23. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis GJ. Compensatory enlargement of human atherosclerotic coronary arteries. N Engl J Med. 1987;316:1371–1275. Kim WY, Stuber M, Bornert P, Kissinger KV, Manning WJ, Botnar RM. Three-dimensional black-blood cardiac magnetic resonance coronary vessel wall imaging detects positive arterial remodeling in patients with nonsignificant coronary artery disease. Circulation. 2002;106:296–299. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human coronary artery lumen and wall imaging using black-blood magnetic resonance imaging. Circulation. 2000;102:506–510. Botnar RM, Stuber M, Kissinger KV, Kim WY, Spuentrup E, Manning WJ. Noninvasive coronary vessel wall and plaque imaging with magnetic resonance imaging. Circulation. 2000;102:2582–2587. Botnar RM, Kim WY, Bornert P, Stuber M, Spuentrup E, Manning WJ. 3D coronary vessel wall imaging utilizing a local inversion technique with spiral image acquisition. Magn Reson Med. 2001;46:848–854. Kim WY, Astrup AS, Stuber M, et al. Subclinical coronary and aortic atherosclerosis detected by magnetic resonance imaging in type 1 diabetes with and without diabetic nephropathy. Circulation. 2007;115:228–235. Desai MY, Lai S, Barmet C, Weiss RG, Stuber M. Reproducibility of 3D free-breathing magnetic resonance coronary vessel wall imaging. Eur Heart J. 2005;26:2320–2324. Priest AN, Bansmann PM, Mullerleile K, Adam G. Coronary vessel-wall and lumen imaging using radial k-space acquisition with MRI at 3 Tesla. Eur Radiol. 2007;17:339–346. Hundley WG, Lange RA, Clarke GD, et al. Assessment of coronary arterial flow and flow reserve in humans with magnetic resonance imaging. Circulation. 1996;93:1502–1508. Nagel E, Bornstedt A, Hug J, Schnackenburg B, Wellnhofer E, Fleck E. Noninvasive determination of coronary blood flow velocity with magnetic resonance imaging: comparison of breath-hold and navigator techniques with intravascular ultrasound. Magn Reson Med. 1999;41:544–549. Nagel E, Thouet T, Klein C, et al. Noninvasive determination of coronary blood flow velocity with cardiovascular magnetic resonance in patients after stent deployment. Circulation. 2003; 107:1738–1743. Langerak SE, Vliegen HW, Jukema JW, et al. Value of magnetic resonance imaging for the noninvasive detection of stenosis in coronary artery bypass grafts and recipient coronary arteries. Circulation. 2003;107:1502–1508. Wang Y, Rossman PJ, Grimm RC, Riederer SJ, Ehman RL. Navigator-echo-based real-time respiratory gating and triggering for reduction of respiration effects in three-dimensional coronary MR angiography. Radiology. 1996;198:55–60. Meyer CH, Hu BS, Nishimura DG, Macovski A. Fast spiral coronary artery imaging. Magn Reson Med. 1992;28:202–213. Bornert P, Stuber M, Botnar RM, et al. Direct comparison of 3D spiral vs. Cartesian gradient-echo coronary magnetic resonance angiography. Magn Reson Med.2001;46:789–794.

140 78. Thedens DR, Irarrazaval P, Sachs TS, Meyer CH, Nishimura DG. Fast magnetic resonance coronary angiography with a threedimensional stack of spirals trajectory. Magn Reson Med. 1999;41:1170–1179. 79. Bornert P, Aldefeld B, Nehrke K. Improved 3D spiral imaging for coronary MR angiography. Magn Reson Med. 2001;45:172–175. 80. Reeder SB, Atalay MK, McVeigh ER, Zerhouni EA, Forder JR. Quantitative cardiac perfusion: a noninvasive spin-labeling method that exploits coronary vessel geometry. Radiology. 1996;200:177–184. 81. Stuber M, Bornert P, Spuentrup E, Botnar RM, Manning WJ. Selective three-dimensional visualization of the coronary arterial lumen using arterial spin tagging. Magn Reson Med. 2002; 47:322–329. 82. Wang SJ, Hu BS, Macovski A, Nishimura DG. Coronary angiography using fast selective inversion recovery. Magn Reson Med. 1991;18:417–423. 83. Wang SJ, Nishimura DG, Macovski A. Multiple-readout selective inversion recovery angiography. Magn Reson Med. 1991;17: 244–2451. 84. Katoh M, Spuentrup E, Barmet C, Stuber M. Local re-inversion coronary MR angiography: arterial spin-labeling without the need for subtraction. J Magn Reson Imaging. 2008;27:913–917. 85. Katoh M, Spuentrup E, Stuber M, et al. Flow targeted 3D steadystate free-precession coronary MR angiography: comparison of three different imaging approaches. Invest Radiol. 2009;44: 757–762. 86. Deshpande VS, Shea SM, Laub G, Simonetti OP, Finn JP, Li D. 3D magnetization-prepared true-FISP: a new technique for imaging coronary arteries. Magn Reson Med. 2001;46:494–502. 87. Spuentrup E, Buecker A, Stuber M, et al. Navigator-gated coronary magnetic resonance angiography using steady-state-freeprecession: comparison to standard t2-prepared gradient-echo and spiral imaging. Invest Radiol. 2003;38:263–268. 88. Spuentrup E, Bornert P, Botnar RM, Groen JP, Manning WJ, Stuber M. Navigator-gated free-breathing three-dimensional balanced fast field echo (TrueFISP) coronary magnetic resonance angiography. Invest Radiol. 2002;37:637–642. 89. Deshpande VS, Li D. Contrast-enhanced coronary artery imaging using 3D trueFISP. Magn Reson Med. 2003;50:570–577. 90. Kaul MG SA, Bansmann PM, et al. Evaluation of balanced steadystate free precession (TrueFISP) and K-space segmented gradient echo sequences for 3D coronary MR angiography with navigator gating at 3 Tesla. Rofo. 2004;176:1560–1565.

A. Hays et al. 91. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: sensitivity encoding for fast MRI. Magn Reson Med. 1999; 42:952–962. 92. Sodickson DK, Griswold MA, Jakob PM. SMASH imaging. Magn Reson Imaging Clin N Am. 1999;7:237–254, vii–viii. 93. Huang F, Akao J, Vijayakumar S, Duensing GR, Limkeman M. k-t GRAPPA: a k-space implementation for dynamic MRI with high reduction factor. Magn Reson Med. 2005;54:1172–1184. 94. Lai P, Huang F, Larson AC, Li D. Fast four-dimensional coronary MR angiography with k-t GRAPPA. J Magn Reson Imaging. 2008;27:659–665. 95. Weiger M, Pruessmann KP, Boesiger P. Cardiac real-time imaging using SENSE. SENSitivity Encoding scheme. Magn Reson Med. 2000;43:177–184. 96. Huber ME, Kozerke S, Pruessmann KP, Smink J, Boesiger P. Sensitivity-encoded coronary MRA at 3 T. Magn Reson Med. 2004;52:221–227. 97. Stuber M, Botnar RM, Fischer SE, et al. Preliminary report on in vivo coronary MRA at 3 Tesla in humans. Magn Reson Med. 2002;48:425–429. 98. Fischer SE, Wickline SA, Lorenz CH. Novel real-time R-wave detection algorithm based on the vectorcardiogram for accurate gated magnetic resonance acquisitions. Magn Reson Med. 1999;42:361–370. 99. Schar M, Vonken EJ, Stuber M. Simultaneous B(0)− and B(1)+ − map acquisition for fast localized shim, frequency, and RF power determination in the heart at 3 T. Magn Reson Med. 2010;63: 419–426. 100. Kubach MR, Bornstedt A, Hombach V, et al. Cardiac phasespecific shimming (CPSS) for SSFP MR cine imaging at 3 T. Phys Med Biol. 2009;54:N467–478. 101. Sommer T, Hackenbroch M, Hofer U, et al. Coronary MR angiography at 3.0 T versus that at 1.5 T: initial results in patients suspected of having coronary artery disease. Radiology. 2005; 234:718–725. 102. Nezafat R, Stuber M, Ouwerkerk R, Gharib AM, Desai MY, Pettigrew RI. B1-insensitive T2 preparation for improved coronary magnetic resonance angiography at 3 T. Magn Reson Med. 2006;55:858–864. 103. Schar M, Kozerke S, Fischer SE, Boesiger P. Cardiac SSFP imaging at 3 Tesla. Magn Reson Med. 2004;51:799–806. 104. van Elderen SG, Versluis MJ, Webb AG, et al. Initial results on in vivo human coronary magnetic resonance angiography at 7 Tesla. Magn Reson Med. 2009;62:1379–1384.

Contrast-Enhanced MR Angiography of the Coronary Arteries

10

Qi Yang and Debiao Li

Introduction

Technical Considerations for CMRA

X-ray angiography is the current gold standard for the diagnosis of coronary artery disease (CAD). However, X-ray angiography is an invasive and costly procedure associated with a small but definite morbidity and mortality. There is a clear need for a noninvasive and more reliable method of directly detecting functionally significant CAD. Magnetic resonance imaging (MRI) overcomes a lot of the problems associated with X-ray angiography and has shown great advantages for the diagnosis of CAD. In addition to being noninvasive, radiation free, and cost-effective, it can provide functional, hemodynamic, and metabolic data as well as anatomical images in the same setting for a comprehensive exam of CAD. Over the past 15 years, substantial advances have been made in the field of coronary MR angiography (CMRA). However, for successful CMRA imaging, a series of technical challenges have to be overcome, including small vessel diameter, intrinsic and extrinsic motion of the heart, and complex geometry of the arterial trees. In addition, sufficient contrast between the coronary lumen and the surrounding tissue is crucial for the visualization of the coronary vessels. In this chapter, we briefly review the historical development of noncontrast CMRA, starting from the motion compensation techniques, with subsequent developments, including SSFP sequences, whole-heart acquisitions approach, and contrast-enhanced CMRA. Then, the current status of technological developments of contrast-enhanced CMRA at 3 T and its clinical role for the evaluation of CAD are also considered.

Motion Compensation

Q. Yang, MD, PhD () Department of Radiology, Xuanwu Hospital, Capital Medical University, Beijing, China D. Li, PhD Cedars-Sinai Medical Center, Biomedical Imaging Research Institute, Los Angeles, CA, USA

The major obstacles for obtaining CMRA images are the two types of motion: cardiac motion related to myocardial contraction/relaxation and respiratory motion attributable to diaphragm and chest wall movement. The extent of motion exceeds the diameter of the coronary artery, blurring artifacts will occur unless adequate motion compensation techniques are applied. To account for cardiac motion, ECG signal is commonly used to synchronize data acquisition to the quiescent period of each heart beat [1] and CMRA data were collected over multiple cardiac cycles. To deal with respiratory motion two possible solutions have been used: breath-hold and free-breathing imaging. Breath-hold is a straightforward approach and is easy to implement. However, the spatial coverage and resolution are limited by the patient’s ability to hold his/her breath. The implementation of navigator techniques for coronary artery imaging has enabled free breathing CMRA and allows higher resolution and larger coverage [2]. Navigator methods involve the detection of a signal at the interface of the dome of the diaphragm and lung tissue. The navigator signal can be produced by a slice-selective 90–180° radiofrequency (RF) pulse pair, where the rectangular slices excited by each pulse are oriented to produce a diamondshaped intersection. Imaging time can be extended from previous one breath-hold to several minutes with free breathing.

Spatial Resolution Insufficient spatial resolution prevents the consistent visualization of the distal and branch vessels and submillimeter spatial resolution is desirable in CMRA [3]. However, higher spatial resolution is associated with decreased signal-tonoise ratio (SNR), longer imaging time. In addition, precise motion compensation is required. Therefore, it is important

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_10, © Springer Science+Business Media, LLC 2012

141

142

Q. Yang and D. Li

Fig. 10.1 Three-dimensional free-breathing coronary MRI using (a) an inversion recovery prepulse and novel intravascular contrast agent (B-22,956, Bracco Spa, Milan) and comparison with (b) conventional

noncontrast T2 preparation coronary MRI. Note the improved contrast with the intravascular agent. (From Huber ME et al. [12].)

to get appropriate balance between spatial resolution, SNR and imaging time when determining imaging parameters of CMRA.

making it sensitive to phase error from complex flow in region with stenosis or local field inhomogeneities. Moreover, the sequence uses a combination of very short TR and large flip angles so that power deposition is high. Power deposition becomes a limiting factor at 3 T.

Contrast Between Blood and Surrounding Tissue The coronary arteries are surrounded by fat proximally and by myocardial tissue along their entire course. It is natural to generate contrast between coronary luminal blood and the surrounding tissue based on refreshing of the inflow blood with gradient echo techniques. Breath-hold two-dimensional (2D) segmented k-space gradient echo sequence was the first robust approach for CMRA using in-flow effect [4]. The breath-hold 2D approach was easy to implement and had been used in the visualization of the proximal coronary arteries. This, in conjunction with frequency-selective saturation pulse, fat suppression can be achieved [5]. Thereafter, the introduction of three-dimensional (3D) gradient echo CMRA eliminates some drawbacks of 2D CMRA related to long scan time and low SNR [5, 6]. These features of 3D CMRA permit 3D postprocessing of the entire coronary arteries. Nevertheless, the signal from the stationary or slow moving blood also gets saturated in 3D acquisitions due to excitation of a thick slab, and thus reduces the contrast. Therefore, magnetization preparation is necessary to suppress the fat and myocardial signal and enable clear delineation of the coronary arteries [6, 7]. With the improvement in gradient systems, steady-state free precession (SSFP) imaging techniques have been extensively used in cardiac MRI [8]. The signal intensity on SSFP sequence is primarily determined by the T2/T1 ratio which makes it intrinsically beneficial for cardiac imaging. Transverse magnetization is largely preserved between RF pulses, allowing for higher flip angle and SNR than conventional spoiled gradient echo sequences. With SSFP sequence, relatively high spatial resolution images with adequate SNR and CNR can be acquired reliably. However, the balanced gradient structure accumulates phases in each readout train,

Benefits of Contrast Agents for CMRA The use of T1-shortening contrast agents has revolutionized MRA of the entire body [9]. It dramatically improved blood SNR and permitted the use of short repetition times. Vessel contrast for CMRA can be further improved by administering T1-shortening contrast agents [10]. It is less important how the imaging plane is selected (i.e., parallel or perpendicular to the flow direction) since the contrast is generated by T1 differences in CMRA. In addition, the blood signal becomes largely flow independent, which is important for the depiction of slow-flowing blood and the reliable detection of coronary artery stenosis. Various paramagnetic T1-shortening contrast agents have been used to generate “bright blood” CMRA images. Based on the capability of diffusing to interstitial space, these agents are typically classified as intra- or extravascular agents. For CMRA, extravascular contrast agent is typically administered in a short time to assure adequate T1-shortening of blood pool. The spatial resolution and/or 3D coverage for each contrast-enhanced scan are limited by confining data acquisition within a short time-frame that coincides with the arterial phase of contrast media. In order to overcome the inherent limitations of extracellular contrast agents, MR blood pool agents have been developed due to the much higher T1 relaxivity and longer half-life in the blood pool [11]. Several researchers have demonstrated that combining with an inversion prepulse to suppress myocardial signal intravascular agent allows for improved arterial contrast on 3D gradient echo CMRA with thick 3D volume [12] (Fig. 10.1).

10

Contrast-Enhanced MR Angiography of the Coronary Arteries

Methods of Contrast Agent Administration for CMRA MR Smartprep and fluoroscopic triggering have been proposed for real-time triggering of contrast-enhanced MRA in pulmonary arteries, aorta, abdominal arteries, and peripheral arteries [13, 14]. For contrast-enhanced CMRA, currently there are two approaches. The first approach to coronary imaging with extracellular contrast agents is to image during breath-holding and the first pass of a contrast agent. One must image rapidly after injection to preserve good contrast between blood in the coronary arteries and the myocardium. The second is slow injection of the contrast agent in conjunction with a free breathing; respiratory-gated sequence. The major advantage of slow-injection, respiratory-gated coronary artery imaging is the relative flexibility in choosing TR, spatial resolution, and coverage. One major limitation is the reduced contrast agent concentration in the blood pool, resulting in longer T1 as compared with dynamic scans with faster injection. Gadobenate dimeglumine (Gd-BOPTA, Bracco Imaging SpA, Milan, Italy) has a roughly twofold higher T1 relaxivity in blood compared to other clinical extracellular contrast agents currently in widespread use. The higher in vivo relaxivity and prolonged half-life of Gd-BOPTA make it more suitable for whole-heart CMRA with a gradient echo imaging sequence.

Advanced Methods in CMRA k-Space Trajectories The MR sampling employed to acquire the actual MR data has an impact on the motion sensitivity and on the image quality. A variety of k-space trajectories have been tested in an effort to maximize spatial resolution and scan efficiency. Due to the limited system and reconstruction requirements, most coronary artery imaging protocols use Cartesian k-space sampling for image acquisition. Cartesian trajectories acquire one or a series of phase-encoding lines within each R–R interval efficiently. Non-Cartesian trajectories, which include multishot echo planar, spiral, and radial, each offer advantages and disadvantages for CMRA. Advantages of spiral acquisitions include a more efficient filling of k-space, a higher SNR, and favorable flow properties [15]. Drawbacks include marked sensitivity to off-resonance effects and the need for specialized image reconstruction algorithms. Radial methods acquire a series of phase-encoding lines of data that are oriented in a radial pattern. Each radial line of data passes through the center of k-space. Thus, radial imaging methods naturally are resistant to motion and flow-related artifacts for

143

CMRA and offer the benefit of more rapid acquisitions [16]. As in spirally sampled MRI, radial k-space sampling depends on specialized reconstruction algorithms to produce images. These reconstruction algorithms are generally not available on commercial MRI scanners and therefore restrict radial MRI to few specialized research groups.

SSFP Whole-Heart CMRA Due to the intrinsically tortuous course of coronary arteries, it has always been of interest to cover the entire heart to depict the long segments of all major coronary arteries. Li et al. [2] first reported using multiple overlapped 3D slabs to cover the whole-heart with retrospective respiratory gating. Advances in SSFP sequences make it feasible to perform whole-heart imaging with higher blood signal intensity as compared to 3D gradient echo sequences. SSFP sequences permit acquisition of large 3D axial volume that encompasses whole-heart without losing arterial contrast in 10–15 min at 1.5 T [17, 18].

Contrast-Enhanced EPI Whole-Heart Technique Imaging speed is important for whole-heart CMRA as reduced scan time leads to improved study success rate and higher image quality. Echo planar imaging (EPI) is one of the data acquisition strategies in which several lines of k-space are acquired following a single RF excitation to reduce the total imaging time. Thick slab coverage (40 slices with 2 mm thickness) within a single breath-hold is possible by using a segmented EPI image acquisition. Conventional EPI techniques suffer from low SNR and spatial resolution. T1 shortening contrast agent can be used to boost the blood signal intensity and imaging contrast [19]. A recently proposed interleaved GRE-EPI acquisition scheme has been used to speed up contrast-enhanced whole-heart CMRA [20]. The preliminary results showed excellent delineation of all the major coronary arteries with scan time reduced by a factor of 2 compared with the SSFP acquisition.

Benefits of Parallel Imaging Parallel imaging techniques, such as the k-space-based technique (GRAPPA: generalized autocalibrating partially parallel acquisitions) and the image space-based technique (SENSE: sensitivity encoding) enable substantial scan time reduction in cardiac imaging. The major drawback of using parallel imaging techniques is the loss of SNR. Therefore, it is important to have an appropriate balance between speed

144

Q. Yang and D. Li

Fig. 10.2 Sample reformatted right coronary MR angiography: (a) acquired with no T2 Prep; (b) with two MLEV-weighted composite T2 Prep; (c) with adiabatic T2 Prep. Arrows in (b) point to the artifacts

resulting from T2 Prep sequence; note the suppression of the banding artifacts in (c) and also the homogeneity of the signal. (From Nezafat R et al. [23].)

and image quality. Fortunately, for higher fields like 3.0 T, the natural SNR gain can be used to trade the acquisition time. The development of 32- or even 128-channel phased-array coils which are designed for 2D parallel imaging, higher acceleration factors will permit higher acceleration and scan time reduction for whole-heart CMRA.

Contrast-Enhanced Whole-Heart CMRA at 3 T

Advantages and Disadvantages of High-Field Coronary Imaging The signal strength in MRI increases in proportion to the strength of the main magnetic field. The higher signal produced by 3.0 T or even 7.0 T MRI scanners results in a proportionate increase in SNR. In other words, the SNR of an image acquired at 3.0 T is roughly twice that of the same image acquired on a 1.5 T scanner. This SNR gain allows images with reduced imaging time or improved spatial resolution, with the same clinically acceptable SNR as the analogous 1.5 T image. However, CMRA at high-field strength system has several major challenges, particularly with the SSFP technique, including more off-resonance image artifacts due to increased B0 field inhomogeneity, greater image intensity variations due to B1 field inhomogeneity, and limited flip angles due to the higher RF power deposition. Many measures have been developed to address the problems with SSFP imaging at 3 T, such as shifting the synthesizer frequency to reduce off-resonance-related image artifacts, or improving field homogeneity by applying localized linear or second-order shimming [21, 22]. The use of conventional T2-preparation is more challenging, thus, a new adiabatic refocusing T2 Prep sequence was developed which exploits the insensitivity of the adiabatic pulses to field inhomogeneity, without exceeding the SAR limitations [23] (Fig. 10.2). Recently, the first human coronary MR images were successfully obtained at 7 T, and other studies that take advantage of new high-field specific improvements are ongoing [24].

Noncontrast whole-heart CMRA approach at 1.5 T necessitates the use of SSFP sequences, which has superior SNR to gradient echo sequences. However, due to increased image artifacts at 3 T with SSFP technique, further SNR gain has not been directly translated into improved coronary artery delineation. Conventional spoiled gradient echo sequences are relatively insensitive to the increased field inhomogeneities at 3 T. Contrast-enhanced whole-heart CMRA at 3 T with slow infusion of contrast agent with high T1 relaxivity has recently been proposed [25]. An inversion recovery prepared spoiled gradient echo sequence, which is relatively insensitive to the increased field inhomogeneities at 3 T, was employed. Slow infusion of a high relaxivity agent allows prolonged blood enhancement time required for whole-heart MRA. Ultrashort TR and high acceleration factor allows significantly reduced acquisition time using contrast-enhanced CMRA at 3 T. Previous comparison study performed in the same volunteers has proved that contrast-enhanced CMRA at 3 T provides higher coronary artery contrast-to-noise ratio (CNR), better vessel depiction, and shorter imaging time than the SSFP approach at 1.5 T [26] (Fig. 10.3). In a later study, the blood pool contrast agent gadofosveset was shown to improve the overall CNR and the delineation of distal coronary segments for contrast-enhanced CMRA at 3 T in comparison to noncontrast SSFP CMRA at 1.5 T [27].

Practical Recommendations for Whole-Heart Coronary MRA at 3 T Patient Training Patient training and practice before data acquisition for maintaining regular breathing is useful to improve the gating efficiency and image quality of CMRA. Patients were trained

10

Contrast-Enhanced MR Angiography of the Coronary Arteries

145

Contrast Injection For 3 T CMRA, 0.2 mmol/kg body weight of Gadobenate dimeglumine (MultiHance; Bracco Imaging SpA, Milan, Italy) was slowly infused using a power injector at a rate of 0.3 ml/s, immediately followed by 20 ml saline at the same rate. Sixty seconds after the initiation of contrast administration, whole-heart CMRA data acquisition was started.

Survey Scanning

Fig. 10.3 Maximum intensity projection images of a 45-year-old healthy male volunteer demonstrate that contrast-enhanced coronary MRA at 3.0 T (b) has better CNR than SSFP coronary MRA at 1.5 T (a). Original axial images show that the mid LAD (c, arrowhead) and LCX (c, arrow) are buried in pericardial fluid at 1.5 T, whereas the arteries and D2 are clearly delineated at 3.0 T (d). LM, LAD, LCX, D2, AO, LV, and RV indicate left main artery, left anterior descending artery, left circumflex, the second diagonal branch, aorta, left ventricle, and right ventricle, respectively. (Reproduced with permission from Liu X et al. [26].)

to perform regular, shallow breathing and to avoid changes in depth of breathing during the data acquisition. Abdominal belt was rolled tightly around the upper abdomen during deep inspiration to suppress the motion of the diaphragm.

Vector Electrocardiogram at 3 T Regular rhythm and accurate ECG synchronization and R-wave detection are imperative for coronary CMRA. Use of a patient-specific quiescent period is recommended, this can be identified by the acquisition of high temporal resolution ECG-triggered cine images. However, under the influence of higher field strength, magneto-hydrodynamic effect led to considerable artificial augmentation of the T-wave of the ECG. The augmentation of the T-wave may mislead the R-wave detection so that triggering is performed on the T-wave instead of the R-wave. The vector ECG allowed reliable R-wave triggering and has been found to be very robust for R-wave detection at 3 T. Carefully positioning of the ECG leads may impact image quality.

For localization of the coronary arteries and for identification of the period of minimal myocardial motion, three scout scans are recommended: Scout 1: A low-resolution 2D scout images were first obtained in three orthogonal orientations to identify the position of the heart and diaphragm. The scan is performed during free breathing. Scout 2: Retrospective ECG-triggered cine images (50 cardiac phases reconstructed) were acquired in a four-chamber view using a fast low-angle shot (FLASH) sequence during free breathing. The global cardiac motion was visually assessed from cine images to determine the patient-specific triggerdelay time and duration of data acquisition window per heartbeat.

Navigator Efficiency A major challenge for CMRA remains to be respirationinduced motion artifacts. Adaptive navigator-gating and motion correction is an effective method for reducing respiratory motion artifacts. However, drift of the diaphragm and fluctuations of the breathing pattern during this period can decrease the respiratory gating efficiency, increase respiratory motion artifacts, or even lead to failure of the measurement. Patient training and practice before data acquisition for maintaining regular breathing should be useful to improve the gating efficiency and image quality of CMRA.

Recent Clinical Applications of 3 T ContrastEnhanced Whole-Heart Coronary MRA 3 T contrast-enhanced whole-heart CMRA now represents the current state-of-the-art technique. In our recent singlecenter study, we have prospectively examined the diagnostic value of contrast-enhanced whole-heart CMRA at 3 T on patients suspected of CAD [28] (Figs. 10.4 and 10.5). Our study demonstrated that acquisition of CMRA at 3 T was successful in 62 of 69 (90%) patients, with the averaged

146

Q. Yang and D. Li

Fig. 10.4 Reformatted images (a) of MRA in a 55-year-old female patient demonstrate significant coronary stenoses at proximal LCX (arrow) and normal RCA which were consistent with the findings (arrow) of conventional coronary angiography (b, c)

Fig. 10.5 3 T contrast-enhanced whole-heart CMRA images of a 75-year-old male patient with atypical chest pain. CMRA MIP images (a and b) show a significant stenosis in the proximal LCX and a nonsignificant stenosis in the middle RCA (arrows), respectively. VR images (syngo InSpace, Siemens AG Healthcare, Erlangen, Germany) (c and d) have the same findings in LCX and RCA, which were consistent with the conventional coronary angiography (e and f). (Reproduced with permission from Yang Q et al. [28].)

acquisition duration of 9.0 ± 1.9 min. 3 T contrast-enhanced whole-heart CMRA allows for ruling out significant CAD with high sensitivity and moderate specificity. The sensitivity, specificity, and accuracy of whole-heart CMRA for detecting significant stenoses were 91.6% (87/95), 83.1% (570/686), 84.1% (657/781), respectively, on a per-segment basis. The results obtained in our study compare favorably with other single-center studies of whole-heart coronary MRA at 1.5 T [29]. These results are also quite similar to the recent experience with 64-slice multidetector CT angiography in a multicenter study [30]. However, a recent published meta-analysis suggests that coronary CTA has better sensitivity and specificity than CMRA and is therefore advantageous for detecting and ruling out clinically relevant coronary stenoses [31]. Despite the excellent diagnostic accuracy, coronary CTA has several disadvantages of requiring rapid injection of iodinated contrast medium and of exposing patients to ionizing radiation. In addition, blooming artifact from calcification leads to false positive diagnosis in many cases. MR does not suffer from these artifacts caused by calcification; and CMRA can potentially depict the lumen of calcified coronary arteries (Fig. 10.6). Both coronary CTA and CMRA can provide lumenographic information to determine the presence and extent of CAD. However, the functional implications of the lesion are more important. The combination of CMRA with tissue perfusion and viability provides a comprehensive assessment of the patient with known or suspected CAD. Due to the use of contrast agent, high-resolution 3D delay enhancement MR images can be reconstructed from the CMRA images. This facilitates the 3D reformation in any slice orientation as well as precise quantification of the damaged tissue and the direct association of the infracted territory to the respective coronary artery lesion. The major advantage is that it allows for a fast and comprehensive assessment of both coronary artery stenosis and myocardial tissue damage in a single noninvasive and radiation free test.

10

Contrast-Enhanced MR Angiography of the Coronary Arteries

147

Fig. 10.6 Maximum intensity projection (MIP) image of CTA in a 59-year-old man (a) shows a mixed plaque with a severe stenosis at proximal LAD. MRA MIP image (b) detects the stenosis with good correlation with CTA and X-ray angiography (c)

Summary and Future Directions in ContrastEnhanced Coronary MRA 4D (3D cine, or cardiac-motion resolved) coronary artery imaging represents one of the future directions for coronary MRA. Such technique can potentially provide high temporal and spatial resolution coronary artery images. The combination of contrast agent and spoiled gradient-echo sequence (e.g., FLASH) at 3 T is the method of choice for coronary MRA. This will permit a major step forward for the clinical use of CMRA. The use of blood pool contrast agent might open the door to further improve the diagnostic accuracy of contrast-enhanced CMRA at 3 T. The Holy Grail for CMRA is to obtain entire whole-heart data set in one cardiac cycle.

References 1. Wang Y, Vidan E, Bergman GW. Cardiac motion of coronary arteries: variability in the rest period and implications for coronary MR angiography. Radiology. 1999;213:751–758. 2. Li D, Kaushikkar S, Haacke E, et al. Coronary arteries: threedimensional MR imaging with retrospective respiratory gating. Radiology. 1996;201:857–863. 3. Schar M, Kim WY, Stuber M, Boesiger P, Manning WJ, Botnar RM. The impact of spatial resolution and respiratory motion on MR imaging of atherosclerotic plaque. J Magn Reson Imaging. 2003;17:538–544. 4. Edelman R, Manning W, Burstein D, Paulin S. Coronary arteries: breath-hold MR angiography. Radiology. 1991;181:641–643. 5. Li D, Paschal CB, Haacke EM, Adler LP. Coronary arteries: threedimensional MR imaging with fat saturation and magnetization transfer contrast. Radiology. 1993;187:401–406. 6. Botnar RM, Stuber M, Danias PG, Kissinger KV, Manning WJ. Improved coronary artery definition with T2-weighted, free-breathing, three-dimensional coronary MRA. Circulation. 1999;99: 3139–3148. 7. Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishimura DG. Coronary angiography with magnetization-prepared T2 contrast. Magn Reson Med. 1995;33:689–696.

8. Carr JC, Simonetti O, Bundy J, Li D, Pereles S, Finn JP. Cine MR angiography of the heart with segmented true fast imaging with steady-state precession. Radiology. 2001;219:828–834. 9. Prince MR, Yucel EK, Kaufman JA, Harrison DC, Geller SC. Dynamic gadolinium-enhanced three-dimensional abdominal MR arteriography. J Magn Reson Imaging. 1993;3:877–881. 10. Zheng J, Li D, Bae KT, Woodard P, Haacke EM. Three-dimensional gadolinium-enhanced coronary magnetic resonance angiography: initial experience. J Cardiovasc Magn Reson. 1999;1:33–41. 11. Anzai Y, Prince MR, Chenevert TL, et al. MR angiography with an ultrasmall superparamagnetic iron oxide blood pool agent. J Magn Reson Imaging. 1997;7:209–214. 12. Huber ME, Paetsch I, Schnackenburg B, et al. Performance of a new gadolinium-based intravascular contrast agent in free-breathing inversion-recovery 3D coronary MRA. Magn Reson Med. 2003; 49:115–121. 13. Foo TK, Saranathan M, Prince MR, Chenevert TL. Automated detection of bolus arrival and initiation of data acquisition in fast, threedimensional, gadolinium-enhanced MR angiography. Radiology. 1997;203:275–280. 14. Wilman AH, Riederer SJ, King BF, Debbins JP, Rossman PJ, Ehman RL. Fluoroscopically triggered contrast-enhanced threedimensional MR angiography with elliptical centric view order: application to the renal arteries. Radiology. 1997;205:137–146. 15. Meyer CH, Hu BS, Nishimura DG, Macovski A. Fast spiral coronary artery imaging. Magn Reson Med. 1992;28:202–213. 16. Stehning C, Bornert P, Nehrke K, Eggers H, Dossel O. Fast isotropic volumetric coronary MR angiography using free-breathing 3D radial balanced FFE acquisition. Magn Reson Med. 2004;52:197–203. 17. Sakuma H, Ichikawa Y, Suzawa N, et al. Assessment of coronary arteries with total study time of less than 30 minutes by using wholeheart coronary MR angiography. Radiology. 2005;237:316–321. 18. Weber OM, Martin AJ, Higgins CB. Whole-heart steady-state free precession coronary artery magnetic resonance angiography. Magn Reson Med. 2003;50:1223–1228. 19. Deshpande VS, Wielopolski PA, Shea SM, Carr JC, Zheng J, Li D. Coronary artery imaging using contrast-enhanced 3D segmented EPI. J Magn Reson Imaging. 2001;13:676–681. 20. Bhat H, Zuehlsdorff S, Bi X, Li D. Whole-heart contrast-enhanced coronary magnetic resonance angiography using gradient echo interleaved EPI. Magn Reson Med. 2009;61:1388–1395. 21. Deshpande VS, Shea SM, Li D. Artifact reduction in true-FISP imaging of the coronary arteries by adjusting imaging frequency. Magn Reson Med. 2003;49:803–809. 22. Schar M, Kozerke S, Fischer SE, Boesiger P. Cardiac SSFP imaging at 3 Tesla. Magn Reson Med. 2004;51:799–806.

148 23. Nezafat R, Stuber M, Ouwerkerk R, Gharib AM, Desai MY, Pettigrew RI. B1-insensitive T2 preparation for improved coronary magnetic resonance angiography at 3 T. Magn Reson Med. 2006;55:858–864. 24. van Elderen SG, Versluis MJ, Webb AG, et al. Initial results on in vivo human coronary MR angiography at 7 T. Magn Reson Med. 2009;62:1379–1384. 25. Bi X, Carr JC, Li D. Whole-heart coronary magnetic resonance angiography at 3 Tesla in 5 minutes with slow infusion of Gd-BOPTA, a high-relaxivity clinical contrast agent. Magn Reson Med. 2007;58:1–7. 26. Liu X, Bi X, Huang J, Jerecic R, Carr J, Li D. Contrast-enhanced whole-heart coronary magnetic resonance angiography at 3.0 T: comparison with steady-state free precession technique at 1.5 T. Invest Radiol. 2008;43:663–668. 27. Prompona M, Cyran C, Nikolaou K, Bauner K, Reiser M, Huber A. Contrast-enhanced whole-heart MR coronary angiography at 3.0 T

Q. Yang and D. Li

28.

29.

30.

31.

using the intravascular contrast agent gadofosveset. Invest Radiol. 2009;44:369–374. Yang Q, Li K, Liu X, et al. Contrast-enhanced whole-heart coronary magnetic resonance angiography at 3.0-T: a comparative study with X-ray angiography in a single center. J Am Coll Cardiol. 2009;54:69–76. Sakuma H, Ichikawa Y, Chino S, Hirano T, Makino K, Takeda K. Detection of coronary artery stenosis with whole-heart coronary magnetic resonance angiography. J Am Coll Cardiol. 2006;48:1946–1950. Miller JM, Rochitte CE, Dewey M, et al. Diagnostic performance of coronary angiography by 64-row CT. N Engl J Med. 2008;359:2324–2336. Schuetz GM, Zacharopoulou NM, Schlattmann P, Dewey M. Metaanalysis: noninvasive coronary angiography using computed tomography versus magnetic resonance imaging. Ann Intern Med. 2010;152:167–177.

MR Angiography and High Field Strength: 3.0 T and Higher

11

Harald H. Quick and Mark E. Ladd

Introduction The motivation for performing MR angiography (MRA) at higher magnetic field strength can be appreciated by answering a few simple questions: Would you like to increase your spatial resolution in time-of-flight (TOF) or contrast-enhanced (CE) MRA, increase your temporal resolution in dynamic MRA applications, decrease your contrast agent dose in CE-MRA, or even use new imaging contrasts for MRA not available at lower field strength? Higher static magnetic field strengths open opportunities in all of these areas. Doubling the static magnetic field strength from the clinical standard 1.5 T to the clinically already established 3.0 T is associated with doubling the inherent signal-to-noise ratio (SNR) in the MR experiment, and thus is a starting point to realizing parts of the above-mentioned wish list. While the benefits of 3.0 T over 1.5 T have already been successfully demonstrated in the literature for certain clinical MRA applications, the potential further gain in SNR cannot simply be extrapolated to higher field strength, such as the still young and emerging 7.0 T technology. In this chapter, the benefits of high-field MRI (i.e., 7.0 T) for angiographic applications are described. Since nothing comes for free in MR, we first have to take into account the physical parameters that change with increasing field strength and that ultimately bolster or undermine specific MRA applications. We begin by comparing 1.5 vs. 3.0 T MRA examples, take a look at TOF and non-TOF MRA alternatives in intracranial MRA at 7.0-T field strength, and then slowly

H.H. Quick, PhD () Institute of Medical Physics (IMP), Friedrich-Alexander-University Erlangen-Nürnberg, Henkestr. 91, 91052 Erlangen, Germany e-mail: [email protected] M.E. Ladd, PhD Erwin L. Hahn Institute for MRI, University Duisburg-Essen, Arendahls Wiese 199, 45141 Essen, Germany

move down below the neckline to explore MRA applications in non-neuro body regions at 7.0 T.

Increasing the Field Strength Physics Background and Practical Consequences for MRA One of the major benefits when moving to high-field MR is the potentially higher achievable overall SNR, whereas the specific absorption rate (SAR) is one of the greatest limiting factors. While the SNR increases linearly with the field strength, the SAR increases quadratically (Fig. 11.1), often practically limiting the use of radiofrequency (RF)-intense imaging sequences. Consequently, MRI pulse sequences that run without limitations at 1.5 T might encounter SAR limits when ported to 3.0 T without modification. In many cases, simple modifications to the image acquisition parameters are required. It may be sufficient to alter sequence parameters, such as lowering the excitation flip angle, increasing the repetition time (TR), and/or reducing the number of slices. The use of RF-based flow saturation pulses as is common in TOF MRA for the suppression of either venous or arterial flow might be limited as well in high-field applications due to SAR constraints. The longitudinal relaxation times T1 of most stationary tissues are significantly longer at higher field strength [1, 2]. This leads to improved background suppression in CE-MRA applications. Associated with this is an improved vessel-totissue contrast in T1-weighted CE-MRA at high field, and in conjunction with the higher MR signal sensitivity, CE-MRA has shown superior results in images acquired on 3.0-T scanners when compared to images acquired on 1.5-T scanners [3–5] (Fig. 11.2). The increase in SNR that goes along with scanning on 3.0-T scanners used compensated for decrease in contrast agent, and reduced risk to the patient dose in CE-MRA (Fig. 11.3). A direct comparison of CE-MRA images acquired at 1.5 T to images acquired at 3.0 T with one half of the contrast agent dose showed no appreciable decease in image quality [6–8].

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_11, © Springer Science+Business Media, LLC 2012

149

150

Fig. 11.1 Signal-to-noise ratio (SNR) and specific absorption rate (SAR) in arbitrary units as function of the magnetic field strength. While the SNR increases approximately linearly with increasing field strength, SAR increases quadratically. Thus, doubling the field strength from 1.5 to 3.0 T would theoretically double the SNR and quadruple the SAR. The desirable increase in SNR at high field strength can be traded for increased spatial and/or temporal resolution and/or for reduced contrast dose in CE-MRA. The undesirable side effect of increasing SAR may practically restrict certain imaging parameters (e.g., limited flip angle or number of slices, prolonged TR, constrained RF saturation pulses, etc.)

Fig. 11.2 In this field strength comparison (1.5 T (a) vs. 3.0 T (b)) of 3D CE-MRA of the supraaortal arteries, the SNR gain was used to increase the spatial resolution. In direct comparison to the 1.5-T image (a), the 3.0-T CE-MRA (b) reveals increased vessel detail and much better visualization of fine arteries. (a) Imaging parameters for 1.5 T were: no parallel imaging, flip angle 30°, voxel size 1.0 × 0.8 × 0.8 mm3, acquisition time 19 s. (b) Imaging parameters for 3.0 T were: parallel imaging with GRAPPA acceleration factor 3, flip angle 16°, voxel size 0.8 × 0.7 × 0.9 mm3, acquisition time 20 s (courtesy of Marc Saake, MD, and Arnd Dörfler, MD, University of Erlangen, Germany)

H.H. Quick and M.E. Ladd

Fig. 11.3 In this comparison of 3D CE-MRA of the supraaortal arteries performed at 3.0 T (a, b), the SNR gain of high-field MRA was used to reduce the dose of the contrast agent while maintaining spatial resolution and acquisition time. The CE-MRA in image (a) was acquired by administering 8 mL of contrast agent (Gadobutrol) while the CE-MRA in image (b) was acquired using only half of the contrast volume (4 mL of Gadobutrol). This comparison study in the same individual shows that bisecting the dose in high-field MRA can qualitatively lead to comparable results in 3D CE-MRA with regards to vessel display and contrast to noise (CNR) while maintaining spatial resolution and acquisition time. (a, b) Imaging parameters for 3.0 T were: parallel imaging with GRAPPA acceleration factor 3, flip angle 16°, voxel size 0.8 × 0.7 × 0.9 mm3, acquisition time 20 s (courtesy of Marc Saake, MD, and Arnd Dörfler, MD, University of Erlangen, Germany)

Analogous to CE-MRA, TOF MRA of the intracranial arterial vasculature at 3.0 T also benefits from the prolonged T1 relaxation times, resulting in good background signal suppression of stationary brain tissue and excellent visualization of the vasculature. Conspicuity of distal vessel segments is greatly improved due to increased spatial resolution, as has been shown in initial studies comparing TOF MRA of the intracranial vessels at 1.5 and 3.0 T [9–11] (Fig. 11.4). The proton resonance frequency increases linearly with field strength. While the Larmor frequency at 1.5 T is approximately 64 MHz, it doubles to 128 MHz at 3.0 T. At field strength of 7.0 T, the excitation RF frequency is slightly less than 300 MHz. Associated with the increase in Larmor frequency is a linear reduction in RF wavelength from about 52 cm in humans at 1.5 T to about 26 cm at 3.0 T to a further reduced value of about 11 cm at 7.0 T. This reduction in RF wavelength can result in RF inhomogeneities during RF excitation at high field. Aside from resulting in signal voids and limited RF signal penetration depth into the tissue, the measured signal response is then a function of the inhomogeneous

11

MR Angiography and High Field Strength: 3.0 T and Higher

Fig. 11.4 In this field strength comparison (1.5 T (a) vs. 3.0 T (b)) of 3D TOF MRA of the intracranial vessels (Circle of Willis), the SNR gain and enhanced TOF signal were used to increase the spatial resolution. In direct comparison to the 1.5-T image (a), the 3.0-T TOF MRA (b) reveals increased vessel detail and much better visualization of fine arteries with simultaneous diminished background signal. (a) Imaging parameters for 1.5 T were: four overlapping slabs, no parallel imaging, flip angle 25°, voxel size 0.9 × 0.4 × 1.0 mm3, acquisition time 6:38 min. (b) Imaging parameters for 3.0 T were: three overlapping slabs, parallel imaging with GRAPPA acceleration factor 2, flip angle 15°, voxel size 0.7 × 0.4 × 0.7 mm3, acquisition time 5:37 min (courtesy of Marc Saake, MD, and Arnd Dörfler, MD, University of Erlangen, Germany)

RF signal distribution during excitation rather than due to inherent tissue contrast. Furthermore, inhomogeneous RF excitation might lead to locally increased SAR values with associated risk of RF tissue heating. Single-channel, whole-body RF transmit coils which are commonly used in 1.5- and 3.0-T MR scanners are no longer an option for signal excitation at 7.0 T. For 7.0-T MR imaging, new multichannel RF transmit technology is an active area of research and development. In multichannel RF transmit, rather than driving the RF body coil with only a single RF channel, the signal-exciting RF coil consists of multiple individual RF transmit elements that can be independently driven with different RF signal phases and amplitudes. This technique has been termed RF shimming, since the RF field within a certain region of interest can be homogenized by shifting and manipulating regions of signal cancellation to be outside of the body habitus of the patient through selection of the appropriate RF signal parameters. Thus, RF inhomogeneities induced by the shorter RF wavelengths can be counteracted and the problem of inhomogeneous RF signal distribution can be successfully alleviated [12–14]. A detailed discussion of various physical parameters in 7-T MRI can be found in Ladd [15], including the impact on T1, T2, and T2* relaxation times and enhancements in sensitivity to magnetic susceptibility differences.

Intracranial Time-of-Flight MRA at 7 T Intracranial TOF MRA is one of the first MRA applications that has demonstrated potential benefits of 7.0-T MRI scanners. As has been shown in intracranial TOF investigations

151

performed at 3.0-T field strength [9–11], prolonged T1 relaxation time leads to improved suppression of MRI signal associated with extravascular brain parenchyma. Better suppression of background is responsible for visualization of distal vascular segments. However, intracranial TOF at 7.0 T has not exhibited the same benefit. The additional factor of 2.3 in field strength imposes SAR limitations that limit the frequency at which flow saturation RF pulse can be applied. In a study exploring the potential for intracranial 7.0-T TOF MRA, the group of Cho et al. [16, 17] demonstrated the benefits of high field strength to provide excellent spatial resolution intracranial TOF MRA (Fig. 11.5). In these studies, the investigators were able to depict the lenticulostriate arteries, which cannot be reliably visualized at lower field strength with TOF MRI. Other groups have also demonstrated that intracranial TOF at 7.0 T is at this early stage, already an attractive technique for transferring the inherent gain in sensitivity of 7.0 T into increased spatial resolution as well as into exquisite vessel detail [18, 19] (Fig. 11.6). It has even been observed that MRA at 7 T can be used to directly visualize arterial dilation in response to specific neural activity associated with task performance, providing a new type of functional MRA [20]. Beyond these initial research endeavors demonstrating the feasibility and potential of 7.0-T TOF MRA in healthy subjects, a possible clinical application is the detection of aneurysms. It is known that small aneurysms pose a risk of bleeding and the devastating effects of subarachnoid hemorrhage. Even small intracranial aneurysms are clinically significant and must therefore be reliably detected with any diagnostic imaging modaility. So far, however, TOF MRA at lower field strength is limited in that aneurysms smaller than 3 mm may not always be detected. A higher resolution and improved contrast would be very helpful. It must be remembered, though, that TOF MRA at high field strength is subject to increased artifacts. Initial results indicate that there may be advantages of ultrahigh field in the detection of aneurysms, but further investigations are needed to evaluate the clinical relevance of these benefits [21] (Fig. 11.7). One advantage of high-field TOF in the detection of aneurysms may be higher sensitivity to slow flow, ensuring more uniform depiction of slow-moving blood within the aneurysmal dome.

To TOF or Not to TOF? High Field Alternatives for Intracranial MRA TOF imaging is an established imaging technique for noncontrast-enhanced 3D MRA of intracranial vessels at 1.5 and 3.0 T. Imaging protocols that are optimized for field strengths of 1.5 or 3.0 T, however, cannot be transferred directly for imaging at 7.0 T; SAR limitations, the altered tissue T1 and T2 times at 7.0 T, and new image artifacts, such as those

152

H.H. Quick and M.E. Ladd

Fig. 11.5 Left–right: Maximum intensity projections (MIPs) of 3D TOF MRA of the intracranial vessels. The SNR gain and TOF signal at 7.0 T were used to systematically increase the spatial resolution. (1) A 3D TOF MRA image obtained at 1.5-T field strength. (2–4) 3D TOF MRA images obtained at 7.0-T field strength. The nominal imaging resolutions are given below the respective images. The 7.0-T TOF MRA

in this example offers sufficient SNR and vessel-to-background contrast to allow for a twofold increase in spatial resolution in all three spatial dimensions, thus reducing the voxel size by a factor of 8. Note the differences (dotted boxes) between the high-resolution image of 7.0 T in (4) and the low-resolution image of 1.5 T in (1). Acquisition times in both examples (1 and 4) were about 10 min (from Kang CK et al. [17].)

Fig. 11.6 This 3D TOF MRA study of intracranial vessels performed on a 7.0-T system in a normal volunteer impressively shows the inherently good background signal suppression (a) due to prolonged T1 relaxation times of static brain tissues. In combination with the strong TOF in-flow signal and high SNR, this renders intracranial TOF MRA in high-field applications a powerful MRA method to display even the finest vessels with high resolution and contrast (b, c). This 3D MRA data set was acquired using a 24-channel radiofrequency (RF) transmit/

receive coil. Due to SAR limitations, no RF flow saturation pulses were used. Consequently, the images (a–c) show both arteries and veins. Imaging parameters were: flip angle 25°, ten overlapping slabs with 48 partitions each to cover the full volume of the brain, spatial resolution 0.3 × 0.3 × 0.4 mm3, acquisition time 19 min for the full volume (courtesy of Markus Thormann, MD, and Oliver Speck, PhD, University of Magdeburg, Germany)

Fig. 11.7 1.5-T (a) vs. 7.0-T TOF MRA source images (b) vs. DSA (c) showing a 11-mm aneurysm of the distal right ICA of a 43-year-old woman. The aneurysm dome and the right A1 branch of the ACA and the M1 segment of the right MCA (arrows) are better delineated at TOF MRA source images at 7 T (b) in comparison to 1.5 T (a). This

finding was attributed to the increased signal intensity and higher spatial resolution of ultrahigh-field MRA. Lateral DSA (c) depicts the parent vessel and neighboring branches of the MCA and ACA (courtesy of Christoph Mönninghoff, MD, University Hospital Essen, Germany)

11

MR Angiography and High Field Strength: 3.0 T and Higher

153

Fig. 11.8 Noncontrast-enhanced images of a healthy volunteer: axially oriented source images of (a) TOF; sagittally oriented source images of (b) VIBE, and (c) MPRAGE. Corresponding maximum intensity projections (MIPs) of the intracranial system: (d) TOF, (e) VIBE, (f) MPRAGE. Note that the vessel-to-background contrast in TOF images (a, d) is high due to good background signal suppression while VIBE (b, e) and MPRAGE images (c, f) show residual anatomic background signal.

The VIBE and the MPRAGE sequences at 7.0 T open up new possibilities for noncontrast-enhanced vascular imaging in the diagnosis of intracranial vascular lesions. For example, in addition to the vessels themselves, which are hyperintense without application of contrast agent as in the TOF MRA, the perivascular structures in the MPRAGE are depicted in the source images with good resolution and can be evaluated with perfect registration to the vasculature (from Maderwald M, et al. [22])

encountered through increased susceptibility, necessitate optimization of the existing sequences. Against this background, we and other groups [22, 23] were motivated to consider other fast gradient echo sequences that might be potential alternatives for MR angiographic imaging at 7.0 T. Magnetization-prepared rapid gradient echo (MPRAGE), for example, has been shown to provide high signal intensity of blood vessels on contrast-enhanced images at 1.5 T [24]. When imaging at 7.0 T, it has been observed that the MPRAGE sequence – even without the administration of contrast agent – provides high signal in the arterial vasculature while the background shows intermediate signal, resulting in high vessel-to-background contrast potential [22, 23, 25]. Moreover, volume-interpolated breath-hold examination (VIBE), a fast T1-weighted 3D spoiled gradient echo sequence that was designed for short acquisition times and high spatial resolution through the use of asymmetric k-space sampling and pixel interpolation [26], has been used for CE-MRA of the abdominal vasculature [26] and intracranial vessels [27]. Similar to MPRAGE imaging, VIBE images acquired at 7.0 T show hyperintense intracranial vascular signal even without administration of contrast agent [22] (Fig. 11.8).

Increased resolution (approximately 0.5 × 0.5 × 0.5 mm3) at 7 T enables the intracranial vessels to be traced far into the periphery (Fig. 11.8d–f), and the use of MPRAGE to depict the intracranial vessels at this resolution without contrast agent is only possible with ultrahigh field strength. This option opens up new possibilities in the diagnosis of intracranial vascular changes. In addition to the vessels themselves, which appear hyperintense as in TOF MRA, the perivascular structures are depicted with good resolution and can be assessed in the source images (Fig. 11.8b, c). Here, further evaluations are needed to show, for example, how this can be helpful in the workup of vascular stenosis. It is possible that this approach can make the additional acquisition of a CT scan to visualize wall calcifications superfluous. Such calcifications are important to depict when considering an endovascular therapy.

Contrast-Enhanced MRA at 7 T As attractive as the approach of using T1-weighted fast gradient echo sequences (such as FLASH, VIBE, and MPRAGE) for non-CE-MRA may seem, it exposes a potential problem

154

H.H. Quick and M.E. Ladd

Fig. 11.9 Magnitude (left column (a, b)) and corresponding phase images (right column (c, d)) of an SWI data set acquired at 7.0 T (top row (a, c) single slice; bottom row (b, d) minimum intensity projection (MIP) over five slices). Note the high contrast even for very small anatomical structures visible on the magnitude and phase images, such as intracortical veins and the layered structure of gray matter. Although generally the magnitude images show more details, some

structures are only visible on the phase images. (e) Venogram at 7.0 T. The SWI data (TR/TE 22/15 ms; acquisition time 10 min; matrix 512 × 384 × 72; resolution 0.4 × 0.4 × 1.5 mm3) were processed by multiplying the unwrapped and filtered phase images with the magnitude images. In addition, the 3D median-filtered SWI data set was subtracted and inverted ((a–d) from Koopmans et al. [34]) ((e) from Rauscher et al. [35])

for CE-MRA applications. Such T1-weighted fast gradient echo sequences are the basis for CE-MRA that are often used in conjunction with image mask subtraction to improve background signal suppression, whereby the precontrast images are subtracted from the postcontrast images. With arterial vessels already displaying brightly in noncontrastenhanced images at 7.0 T, however, mask subtraction ends up subtracting high nonenhanced arterial signal from high contrast-enhanced arterial signal, consequently reducing the arterial vessel-to-background signal. It remains to be seen whether appropriate contrast-enhanced protocols can be adapted for 7.0 T. The relaxivity parameters, R1 and R2, of contrast agents are highly field dependent. The relaxivity ratios of gadopentetate dimeglumine are, for example, 3.7/4.1 Ls−1 mmol−1 and 4.6/3.7 Ls−1 mmol−1 at 1.5 and 3.0 T, respectively [28]. Early results indicate further changes in R1 and R2 relaxivity at 7 T [29]. It is difficult to fully predict which properties will dominate at higher field strengths because many of the effects depend on the details of the injection protocol, including differentiation between first-pass bolus techniques and steadystate techniques. Numerous studies have shown that gadolinium CE-MRA is feasible at 3.0 T and, in many cases, delivers results superior to those at 1.5 T [3–5]. At higher field strengths, the advantages are not as clear because of the

increasing sensitivity to, for instance, R2*. Certainly, other contrast agents, such as those based on superparamagnetic iron oxide or ultrasmall superparamagnetic iron oxide, also need to be considered [30]. A review of contrast agent results at 3.0 T is presented by Trattnig et al. [7].

MRA Techniques Unique to High-Field MR Because of the susceptibility difference between deoxygenated venous blood and the surrounding tissue, venous imaging at high field strength can produce extraordinary results. The use of a gradient echo sequence with a relatively long echo time (TE) to achieve T2* weighting produces excellent contrast in the venous system. Venules with a diameter of 100 mm can be readily visualized. The susceptibility differences between tissue types also lead to differences in signal phase between the tissues. By optimizing the echo time to ensure that the signals are out of phase, contrast can be optimized in the phase image. This information can be combined with the magnitude image to enhance tissue contrast – a technique termed susceptibilityweighted imaging (SWI) [31–34] (Fig. 11.9a–d). Fig. 11.9e shows a high-field venogram acquired at 7 T after taking both magnitude and phase into account [35].

11

MR Angiography and High Field Strength: 3.0 T and Higher

155

to observe how further developments in RF technology and imaging applications can bring us closer to the assessment of body MRA at 7.0-T field strength.

Conclusion

Fig. 11.10 7T: Non-neuro MRA applications are a relatively new and emerging research field for 7.0-T MR imaging. This is due to the lack of commercially and scientifically available multichannel transmit/ receive RF coils for body applications. First research groups are exploring the inherent diagnostic potential of high-field MRA in the body. The example shows TOF MRA of the liver vasculature, acquired in a normal volunteer at 7.0-T field strength. (a) Twenty imaging slices with a spatial resolution of 1.6 × 0.8 × 2.5 mm3 acquired in a 30-s breath hold. (b) Selected MIP of an axial slab with 33-mm slab thickness. Note the residual B1 inhomogeneity leading to periaortal signal loss

Below the Neckline: 7 T Non-neuro MRA Techniques All the above-mentioned techniques and examples for MRA at 7.0 T are aimed toward neurological applications. This narrow focus is largely a result of the limited availability of whole-body RF coils and RF excitation architectures for 7.0-T scanners. Initially, human whole-body high-field systems have been available with head-only transmit coils only. Investigators have been compelled to build their own RF systems and RF transmit/receive coils to access the rest of the human body and to explore the benefits and limits of 7.0-T imaging in non-neurological applications. The development of multichannel RF technology with the ability to perform B1 shimming and associated RF signal homogenization inside the body tissue has been a precondition for realization of body MRI at 7.0 T. The first attempts at construction of 7.0-T whole-body MRI have recently been published, demonstrating the potential of ultrahigh-field imaging and the need for further coil and sequence optimization [36, 37]. With the development of a custom-built 8-channel RF B1 shimming system [38] in conjunction with a custom-built 8-channel transmit/receive RF body array coil [39], our group has begun to explore 7.0-T MRA applications in the human torso. Feasibility studies in normal volunteers have been performed to investigate the potential for non-CE-MRA of the renal arteries as well as of the vasculature of liver [40]. First results of TOF MRA in the human liver at 7.0 T are shown in Fig. 11.10 demonstrating relatively good B1 signal penetration and homogeneity as well as good vessel-to-background contrast. Of course, this can only be considered a very first step in this rather young but fastevolving research field. Seen in this light, it will be exciting

Increasing the field strength from the clinical standard 1.5 T to the already clinically established 3.0 T has brought more SNR to MRA. According to numerous studies, this doubling in SNR has successfully been transformed into improved display of vascular detail through finer spatial resolution in both TOF and CE-MRA; it has accelerated time-resolved 3D CE-MRA to gain additional dynamic information; and it has been used to reduce the contrast agent dose in selected CE-MRA applications. Beyond the clinically established field strength of 3.0 T, the still young but very fast-developing field of 7.0 T is gaining attention. Early research studies have demonstrated superb image quality for intracranial 3D TOF at 7.0 T owing to improved background tissue suppression and higher spatial resolution which is possible due to the increase in SNR. Furthermore, 7.0 T allows for MRA applications not previously available at lower field strength thanks to inherently changed tissue contrasts. Physical and technical challenges, such as RF inhomogeneities due to the short RF wavelength at 7.0 T, are still to be overcome before the full potential of 7.0-T MRA can be assessed in the remainder of the human body. However, technical solutions, such as multichannel RF transmit coils and associated RF shimming approaches, are under development, and great progress is being made in extending the range of 7-T body applications, including angiography.

References 1. Fischer HW, Rinck PA, Van Haverbeke Y, Muller RN. Nuclear relaxation of human brain gray and white matter: analysis of field dependence and implications for MRI. Magn Reson Med. 1990;16:317–334. 2. Rooney WD, Johnson G, Li X, Cohen ER, Kim SG, Ugurbil K, Springer CS. Magnetic field and tissue dependencies of human brain longitudinal 1H2O relaxation in vivo. Magn Reson Med. 2007;57:308–318. 3. von Falkenhausen MM, Lutterbey G, Morakkabati-Spitz N, Walter O, Gieseke J, Blömer R, Willinek WA, Schild HH, Kuhl CK. Highfield-strength MR imaging of the liver at 3.0 T: intraindividual comparative study with MR imaging at 1.5 T. Radiology. 2006; 241:156–166. 4. Ba-Ssalamah A, Nobauer-Huhmann IM, Pinker K, Schibany N, Prokesch R, Mehrain S, Mlynárik V, Fog A, Heimberger K, Trattnig S. Effect of contrast dose and field strength in the magnetic resonance detection of brain metastases. Invest Radiol. 2003;38:415–422. 5. Fenchel M, Nael K, Deshpande VS, Finn JP, Kramer U, Miller S, Ruehm S, Laub G. Renal magnetic resonance angiography at 3.0 Tesla using a 32-element phased-array coil system and parallel imaging in 2 directions. Invest Radiol. 2006;41:697–703.

156 6. Krautmacher C, Willinek WA, Tschampa HJ, Born M, Träber F, Gieseke J, Textor HJ, Schild HH, Kuhl CK. Brain tumors: full- and half-dose contrast-enhanced MR imaging at 3.0 T compared with 1.5 T: initial experience. Radiology. 2005;237:1014–1019. 7. Trattnig S, Pinker K, Ba-Ssalamah A, Nöbauer-Huhmann IM. The optimal use of contrast agents at high field MRI. Eur Radiol. 2006;16:1280–1287. Review. 8. Kramer U, Fenchel M, Laub G, Seeger A, Klumpp B, Bretschneider C, Finn JP, Claussen CD, Miller S. Low-dose, time-resolved, contrast-enhanced 3D MR angiography in the assessment of the abdominal aorta and its major branches at 3 Tesla. Acad Radiol. 2010;17:564–576. 9. Willinek WA, Born M, Simon B, Tschampa HJ, Krautmacher C, Gieseke J, Urbach H, Textor HJ, Schild H. Time-of-flight MR angiography: comparison of 3.0-T imaging and 1.5-T imaginginitial experience. Radiology. 2003;229:913–920. 10. Al-Kwifi O, Emery DJ, Wilman AH. Vessel contrast at three Tesla in time-of-flight magnetic resonance angiography of the intracranial and carotid arteries. Magn Reson Imaging. 2002;20:181–187. 11. Gibbs GF, Huston J 3rd, Bernstein MA, Riederer SJ, Brown RD Jr. Improved image quality of intracranial aneurysms: 3.0-T versus 1.5-T time-of-flight MR angiography. Am J Neuroradiol. 2004;25:84–87. 12. Abraham R, Ibrahim TS. Proposed radiofrequency phased-array excitation scheme for homogenous and localized 7-Tesla wholebody imaging based on full-wave numerical simulations. Magn Reson Med. 2007;57:235–242. 13. Mao W, Smith MB, Collins CM. Exploring the limits of RF shimming for high-field MRI of the human head. Magn Reson Med. 2006;56:918–922. 14. Van de Moortele PF, Akgun C, Adriany G, Moeller S, Ritter J, Collins CM, Smith MB, Vaughan JT, Uğurbil K. B1 destructive interferences and spatial phase patterns at 7 T with a head transceiver array coil. Magn Reson Med. 2005;54:1503–1518. 15. Ladd ME. High-field-strength magnetic resonance: potential and limits. Top Magn Reson Imaging. 2007;18:139–152. Review. 16. Cho ZH, Kang CK, Han JY, et al. Observation of the lenticulostriate arteries in the human brain in vivo using 7.0 T MR angiography. Stroke. 2008;39:1604–1606. 17. Kang CK, Park CW, Han JY, et al. Imaging and analysis of lenticulostriate arteries using 7.0-Tesla magnetic resonance angiography. Magn Reson Med. 2009;61:136–144. 18. von Morze C, Xu D, Purcell DD, et al. Intracranial time-of-flight MR angiography at 7 T with comparison to 3 T. J Magn Reson Imaging. 2007;26:900–904. 19. Heverhagen JT, Bourekas E, Sammet S, Knopp MV, Schmalbrock P. Time-of-flight magnetic resonance angiography at 7 Tesla. Invest Radiol. 2008;43:568–573. 20. Cho ZH, Kang CK, Han JY, et al. Functional MR angiography with 7.0 T Is direct observation of arterial response during neural activity possible? Neuroimage. 2008;42:70–75. 21. Monninghoff C, Maderwald S, Theysohn JM, et al. Evaluation of intracranial aneurysms with 7 T versus 1.5 T time-of-flight MR angiography - initial experience. Rofo. 2009;181:16–23. 22. Maderwald S, Ladd SC, Gizewski ER, et al. To TOF or not to TOF: strategies for non-contrast-enhanced intracranial MRA at 7 T. MAGMA. 2008;21:159–167. 23. Zwanenburg JJ, Hendrikse J, Takahara T, Visser F, Luijten PR. MR angiography of the cerebral perforating arteries with magnetization prepared anatomical reference at 7 T: comparison with time-offlight. J Magn Reson Imaging. 2008;28:1519–26. 24. Liang L, Korogi Y, Sugahara T, et al. Evaluation of the intracranial dural sinuses with a 3D contrast-enhanced MP-RAGE sequence:

H.H. Quick and M.E. Ladd

25.

26.

27.

28.

29.

30.

31.

32. 33.

34.

35.

36. 37. 38.

39.

40.

prospective comparison with 2D-TOF MR venography and digital subtraction angiography. AJNR Am J Neuroradiol. 2001;22: 481–492. Van de Moortele PF, Auerbach EJ, Olman C, Yacoub E, Uğurbil K, Moeller S. T1 weighted brain images at 7 Tesla unbiased for Proton Density, T2* contrast and RF coil receive B1 sensitivity with simultaneous vessel visualization. Neuroimage. 2009;46:432–46. Rofsky NM, Lee VS, Laub G, et al. Abdominal MR imaging with a volumetric interpolated breath-hold examination. Radiology. 1999;212:876–884. Wetzel SG, Law M, Lee VS, Cha S, Johnson G, Nelson K. Imaging of the intracranial venous system with a contrast- enhanced volumetric interpolated examination. Eur Radiol. 2003;13:1010–1018. Weinmann HJ, Bauer H, Griesinger C, et al. Early distribution dynamics of polymeric magnetic resonance imaging contrast agents in rats. Acad Radiol. 2002;Suppl 2:S412-S416. No abstract available. Noebauer-Huhmann IM, Kraff O, Juras V, et al. MR contrast media at 7 Tesla: preliminary study on relaxivities. Proc Intl Soc Mag Reson Med. 2008;16:1457. Allkemper T, Heindel W, Kooijman H, Ebert W, Tombach B. Effect of field strengths on magnetic resonance angiography: comparison of an ultrasmall superparamagnetic iron oxide blood-pool contrast agent and gadopentetate dimeglumine in rabbits at 1.5 and 3.0 Tesla. Invest Radiol. 2006;41:97–104. Reichenbach JR, Barth M, Haacke EM, Klarhöfer M, Kaiser WA, Moser E. High-resolution MR venography at 3.0 Tesla. J Comput Assist Tomogr. 2000;24:949–957. Haacke EM, Xu Y, Cheng YC, Reichenbach JR. Susceptibility weighted imaging (SWI). Magn Reson Med. 2004;52:612–618. Deistung A, Rauscher A, Sedlacik J, Stadler J, Witoszynskyj S, Reichenbach JR. Susceptibility weighted imaging at ultra high magnetic field strengths: theoretical considerations and experimental results. Magn Reson Med. 2008;60:1155–1168. Koopmans PJ, Manniesing R, Niessen WJ, Viergever MA, Barth M. MR venography of the human brain using susceptibility weighted imaging at very high field strength. MAGMA. 2008; 21:149–158. Rauscher A, Barth M, Herrmann KH, Witoszynskyj S, Deistung A, Reichenbach JR. Improved elimination of phase effects from background field inhomogeneities for susceptibility weighted imaging at high magnetic field strengths. Magn Reson Imaging. 2008;26:1145–1151. Snyder CJ, DelaBarre L, Metzger GJ, et al. Initial results of cardiac imaging at 7 Tesla. Magn Reson Med. 2009;61:517–524. Vaughan JT, Snyder CJ, DelaBarre LJ, et al. Whole-body imaging at 7 T: Preliminary results. Magn Reson Med. 2009;61:244–248. Bitz A, Brote I, Orzada S, et al. An 8-channel add-on RF shimming system for wholebody 7 Tesla MRI including real-time SAR monitoring. Proceedings of the 17th Annual Meeting of the International Society for Magnetic Resonance in Medicine, Honolulu, HI, 2009, Abstract 4767. Orzada S, Quick HH, Ladd ME, et al. A flexible 8-channel transmit/ receive body coil for 7 T human imaging. Proceedings of the 17th Annual Meeting of the International Society for Magnetic Resonance in Medicine, Honolulu, HI, 2009, Abstract 2999. Umutlu L, Lauenstein TC, Kraff O, Maderwald S, Orzada S, Kinner S, Heilmaier C, Antoch G, Ladd ME, Quick HH. Non-enhanced vs. contrast-enhanced MRA at 7Tesla: A feasibility and comparison trial. Proceedings of the 18th Annual Meeting of International Society for Magnetic Resonance in Medicine, Stockholm, Sweden 2010; Abstract 1421.

Susceptibility Weighted Imaging and MR Angiography

12

Samuel Barnes and E. Mark Haacke

Introduction Magnetic resonance angiography (MRA) has undergone tremendous development since its inception [1–3]. To image the vessels in the brain at high fields clinically, there is no longer the need to use contrast agents thanks to the increased signal-to-noise at 3 T and the rapid scanning that is possible. Conventional time-of-flight [4] with or without magnetization transfer contrast (MTC) can give excellent coverage with high-resolution images. Similarly, susceptibility weighted imaging (SWI) can be used to create venographic images of vessels as small as 200–300 mm [5, 6]. In this chapter, we discuss the potential to image both arteries and veins in an SWI single or multiecho time-of-flight (TOF)like sequence. For the last 100 years, the arterial system has played the major role in the study of the brain’s hemodynamics and, from a surgical point of view, in the study and treatment of atherosclerosis. However, the other half of the story is told by the venous system. SWI, like blood oxygen level dependent (BOLD) imaging, is sensitive to deoxyhemoglobin in the veins and is able to generate exquisite images of the veins in the brain. SWI is also sensitive to nonheme iron in the form of ferritin or hemosiderin as well and can be used to quantify iron [7] along with T2* maps [8]. These two features make SWI a powerful means by which to study neurovascular diseases such as cerebral amyloid angiopathy (CAA), multiple sclerosis (MS), stroke, traumatic brain injury (TBI) and tumors [9].

S. Barnes, MS () • E.M. Haacke, PhD Department of Radiology, Loma Linda University Medical Center, Loma Linda, CA, USA Department of Radiology, Harper Hospital/Wayne State University, Detroit, MI, USA e-mail: [email protected]

Technical Issues with SWI and MRA in a Single Sequence Basic SWI Concepts The actual SWI sequence itself is akin to the usual flow compensated gradient echo sequence [5, 6]. More recent efforts have focused on developing a multiecho version of SWI with the potential to simultaneously measure T2* as well as phase [10]. Gradient echo imaging is particularly sensitive to local changes in magnetic field. This variation of field changes the spin phase over time and so, at longer echoes, the signal rapidly disappears (thanks to the additional T2¢ dephasing). The total relaxation rate is given by R2* = R2 + R2¢ where R2* = 1/T2*, R2 = 1/T2, and R2¢ = 1/T2¢. In addition to the magnitude image, SWI uses the phase data where, for a righthanded system, phase is given by j(phi) = −g (gamma) × D(delta) B×TE with g the gyromagnetic ratio of hydrogen (42.6 MHz/T) and D(delta)B is the local change in magnetic field caused by iron or calcium or other structural effects in the tissue. The phase image is high pass filtered to remove low spatial frequency phase variations caused by a variety of background field inhomogeneities such as those from poorly shimmed fields or air–tissue interface field effects [6, 11, 12]. Although these SWI filtered phase images are of great interest in and of themselves [6, 13], the phase can also be manipulated to create a phase mask that highlights positive or negative phases or both on the magnitude image. This phase mask is then used to create a new type of image that accentuates both the T2* contrast, from the magnitude image, and phase information into a single SW image (Fig. 12.1). It is this form of SWI data that has become best known [9, 14]. SWI data are collected with a long-TE, fully flowcompensated gradient echo scan; this can be in the form of a single-echo, multiple-echo [10, 15, 16], or segmented echoplanar approach [17]. Imaging with long echoes with reasonable signal-to-noise became possible by using 3D gradient echo imaging [11]. This allowed for thinner slices (1–2 mm),

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_12, © Springer Science+Business Media, LLC 2012

157

158

S. Barnes and E.M. Haacke

Fig. 12.1 Single slice comparing SWI and other images at 4 T, (a) SWI magnitude image (b) SWI phase image (c) SW Image (d) proton density image. Note that the SW image (c) combines features from both the magnitude (a) and phase (b)

which reduced dephasing across the slice and improved image quality. Flow compensation in all directions is useful in SWI because of the long echo times required and because phase is being used as a measure of susceptibility. Velocity compensation is usually enough to reduce flow related signal loss in the magnitude image, and flow induced changes in the phase image. If one considers two tissues with different susceptibilities D(delta)c(chi), changes in the phase image are generated according to the formula (for a right-handed system): Δj = −g ⋅ ΔB ⋅ TE, (12.1) (delta)(phi) = − (gamma)×(delta)B×TE where Δj = −g ( Δc B0 G + ΔBCS + ΔBgeometry + ΔBmain field )TE. (12.2) (delta)(phi) = − (gamma)((delta)(chi)B 0 G + (delta) BCS + (delta)Bgeometry + (delta)Bmain field)TE Here, G represents a constant dependent on the geometry of the object, CS refers to chemical shift, and Bgeometry refers to the geometry of the brain and air–tissue interfaces. The last two terms represent unwanted field effects. The first two terms are of particular interest to us and are meant to represent the local changes in field, such as those that might be caused by iron in tissue. In 12.2, we have rewritten D(delta)

B = D(delta)c(chi)B0 where D(delta)c(chi) represents the local susceptibility change between tissues. The last two terms tend to be slowly varying spatial terms and can be mostly removed using a high-pass spatial frequency filter. Ideally, we can isolate the first two terms −g (gamma)GD(delta)c (chi) B0TE and −g (gamma)D(delta)BCSTE. Both of them lead to similar phase results inside the object of interest. A paramagnetic object (such as iron) causes a local increase in field and therefore a negative phase change relative to surrounding tissues, while a diamagnetic substance (such as calcium) causes a local decrease in field and therefore a positive phase change. The phase filter is designed as a homodyne filter whereby a low spatial frequency phase image is divided into the original phase to leave behind high spatial frequency phase. The size of the low pass filter Nf is usually quoted as something like 64 × 64 for example. However, it might make more sense to refer to the size of the smallest object df that is basically removed by the filter. This is given by df = FOV/Nf. As an example, consider the FOV = 256 voxels and Nf = 64. In this case, df = 4 voxels. The implications are that objects greater than or equal to 4 voxels will be suppressed from the filtered phase image. When aliasing is particularly severe, the phase image can be unwrapped, prior to high pass filtering, to improve the results of the filtering [18]. In some cases, it may be possible to directly remove background phase effects caused by the gross geometry (D(delta)Bgeometry) in 12.2 with minimal filtering. When the geometry and average susceptibility of the object itself (e.g., the brain) and its surroundings (e.g., the sinuses) are roughly known, their magnetic field effects can be calculated and removed using a forward modeling approach [19]. These calculated phase effects are then divided out from the measured phase image, prior to high-pass filtering, removing many artifacts and leaving the local changes in susceptibility and phase unaltered, improving the results of the filtering. Whichever filtering technique is used, the resultant filtered phase image is used as described below in all subsequent steps and will be referred to as the SWI filtered phase image. As mentioned above, a mask must be created from the filtered phase image and then applied to the magnitude image to generate the SWI data. This mask focuses on certain phase values that will enhance the contrast of the original magnitude image. For example, if areas with increased iron are the subject of interest, then the mask is designed to enhance information related to negative phase (in a right-handed system) as follows: ⎧ p + j ( x) ⎪ f ( x) = ⎨ p ⎪⎩ 1

for − p < j ( x ) < 0

(12.3)

otherwise

(pi) + (phi)(x)/(pi) for −(pi) < (phi)(x) < 0 where the phase values can range from −p(pi) to p(pi), j(phi)(x) is the phase at location x, and f(x) is the phase

12 Susceptibility Weighted Imaging and MR Angiography

159

mask. This phase mask can be multiplied by the original magnitude image an integer m number of times to create the SW image: r¢ ( x ) = f m ( x ) ⋅ r( x ). (rho)¢(x) = f m(x) (rho)(x)

(12.4)

The number of times the mask is applied will change the contrast in the SW image. It has been shown that four multiplications produces good CNR for a wide range of phase values [6]. If the aliasing in the phase image has not been fully removed, which is nearly inevitable in the regions near the frontal sinuses, the residual phase artifacts will destroy the existing magnitude contrast in the SW image rather than enhancing it. This can be overcome by improving the phase filtering, using one of the more advanced techniques discussed above, or alternatively not applying the phase mask in those areas. Problem areas can be identified by calculating the local field gradients, the phase mask can then be adjusted accordingly to remove problem areas with high local field gradients [20]. This preserves existing magnitude contrast in problem areas while still getting regular SWI contrast in the rest of brain. Once this SWI data has been created, it is possible to highlight veins by performing a minimum intensity projection (mIP) over a number of slices. This shows venous connectivity in the same way a maximum intensity projection (MIP) shows arterial connectivity in MRA. A disadvantage of using mIPs is that the dark background surrounding the brain will mask out the brain if it is included in the projection. The slice included in the mIP having the smallest visible brain area will dictate how much of the brain is visible in the final projection. This can be problematic at the top and bottom of the brain where the size changes very rapidly. This problem can be partly overcome by using a brain extraction algorithm, such as a complex threshold approach [21], to set the noise values outside the brain to a value much higher than the brain during the mIP processing and then back to zero afterward. Clinically, mIPs are usually limited to projections over 4–8 mm, although it is certainly possible to project over more slices near the center of the brain or if the background has been removed to get a better visualized of a larger portion of the venous vasculature.

Fig. 12.2 A single sagittal slice from 0.5 mm isotropic data showing (a) the high-pass filtered phase image and (b) the SWIM image. The phase image has been inverted to match contrast on the SWIM image. Notice the dark phase located outside of the veins that is removed in the SWIM image

susceptibility values in turn depend directly on the local tissue composition such as iron content, deoxyhemoglobin levels, and calcium content. While the phase image is directly influenced by the changes in tissue susceptibility, it is also influenced by many other factors, most noticeably the shape and distribution of those susceptibility changes (see 12.2 above). Susceptibility mapping seeks to remove all of these influences by taking the phase image and calculating what susceptibility distribution could create that phase image. Unfortunately, this inversion is ill-posed as the mathematical kernel used in the inversion process becomes zero at certain orientations (the so-called magic angle 54.7°). This has been overcome with various techniques such as restrictions to certain geometric shapes [22, 23], multiple scans at different orientations [24], and various regularization schemes ranging from very simple to quite complicated [25–27]. In our own work, we used a simple regularization approach. Although the term susceptibility mapping is well known, we prefer to call this approach SWIM for susceptibility weighted imaging and mapping as we used the highpass filtered phase image in the process. In this approach, the phase is converted into a magnetic field distribution by dividing the phase by g (gamma)TE. To create the susceptibility map, the usual inverse filter is applied and regularized by preventing the inverse from dividing by zero by setting a threshold [27]. We have found that this approach works best for high resolution 0.5 × 0.5 × 0.5 mm isotropic data (Figs. 12.2 and 12.3).

Susceptibility Mapping and Measuring Oxygen Saturation Simultaneous SWI and MRA Several groups have recently started exploring the possibility of creating susceptibility maps. Susceptibility maps are to SWI what T2 maps are to T2 weighted images, albeit much more difficult to calculate. They are a quantitative map of the actual magnetic susceptibility values of the tissue. These

Both SWI and time-of-flight MRA rely on a flow compensated gradient echo sequence. SWI uses a long echo time and MRA uses a very short echo time. This fact enables SWI and MRA sequences to be combined into a single sequence with

160

S. Barnes and E.M. Haacke

Fig. 12.3 A MIP over 8 mm for (a) the high-pass filtered phase image and (b) the SWIM image. The contrast on the phase image has been inverted to match the SWIM image. Notice both show good vein connectivity and gray matter–white matter contrast. The MIP hides many of the extravascular dipole effects seen in Fig. 12.2 as they are dark and so are excluded from the MIP. The hidden artifacts in the phase image cause the images to look very similar in a MIP, but the SWIM image is still quantifiable, while the phase is not. (From Haacke EM et al. [60]; with permission.)

Fig. 12.4 A multiecho SWI sequence with five echoes was used to get both MRA and traditional SWI contrast. (a) First echo with TE = 6.1 ms MIP over 32 mm clearly depicts the arteries. (b) Fifth echo with TE = 22.1 ms mIP over 12 mm of SW images shows the veins and traditional SWI contrast

two echoes. The first echo is optimized for MRA contrast and the second echo is optimized for SWI contrast. One of the problems in implementing a simultaneous MRA and SWI scan occurs in trying to flow compensate all echoes, which takes considerable time. One way around this is to discard the flow compensation in the phase encoding direction for the first echo (MRA contrast), where it is most important to keep the echo time as short as possible. The second or last echo (SWI contrast) should be flow compensated in all directions as one does not want phase from flow to obscure phase from susceptibility. It is fairly easy to accomplish this given the fact that the SWI sequence itself is usually run with a very long echo to enhance phase effects. The real problem lies in the choice of excitation flip angles so as not to oversuppress the cerebrospinal fluid signal (CSF) and maintain good SWI venous contrast inside the CSF in the gray matter, but still get good excitation of the quickly refreshed arterial blood. A contrast agent can be used to enhance results. In this case, the first echo will show excellent arteries and veins while the SWI long echo will still suppress the veins unlike conventional MRA (see contrast agents section for more details) [28]. The possibility of using multiple echoes in SWI to achieve different types of contrast was first explored by Du et al. [16]. They used a flow compensated short echo to achieve time-of-flight MRA contrast, and a second long echo to achieve SWI contrast. In their implementation, the second echo was not flow compensated in the phase encode directions. This was shown to cause some artifacts in the phase image near flowing arteries and Deistung et al. [29] was able to implement a second echo that was fully flow compensated to remove these artifacts. The flow compensation was

achieved by fully rewinding the phase encode gradients to zero before performing a second flow compensated phase encode for the next echo. Further modifications such as adjusting the k-space ordering to optimize flip angle choice for each type of contrast and using multiple thin slabs have also been proposed [15]. With these techniques it is possible to achieve both a high quality TOF MRA and a SWI in a single scan that takes approximately the same amount of time as either of those run separately. These two-echo, two-contrast techniques suggest the possibility of using a general flow compensated multiecho sequence to achieve a variety of contrasts in both magnitude and phase. As discussed above, the short echoes can be used as an MRA and the long echoes for SWI contrast (Fig. 12.4). The echoes in the middle or after the optimal SWI contrast also have some value. The multiple echo data can be used to calculate T2* maps especially in slow flowing vessels such as the veins and sagittal sinus. They can also be useful for various phase filtering techniques. Finally, having a variety of echo times allows the echo with the optimal susceptibility contrast to be used. Shorter echoes are more optimal for hemorrhages that contain lots or iron, while long echoes are more optimal for imaging tiny venules.

Single Echo Approach While the multiecho approaches can produce excellent angiographic and venographic images, the possibility remains open to achieve similar results with a single echo [30]. SWI provides a natural separation of the vasculature, with the arteries being bright from inflow enhancement due

12 Susceptibility Weighted Imaging and MR Angiography

to the short TR and the veins dark from the long TE. By using a higher bandwidth and high isotropic resolution to reduce flow dephasing, it is possible to obtain a decent angiogram without significantly degrading the SWI venogram. Generally, by using a long TE and short TR (typically 20 and 30 ms at 3 T, respectively) the veins can be suppressed and the arteries brightened, allowing a very good separation of arteries and veins in a single echo scan. The arteries can be visualized using a standard maximum intensity projection (MIP) of magnitude information, and the veins visualized with SWI processing and a minimum intensity projection (mIP). The choice of flip angle, resolution, and echo time is very important for determining image quality of both the MRA and MRV. Higher flip angles improve the angiogram due to increased background suppression and better TOF inflow effect, but in turn degrade the venogram by oversuppressing the CSF. Choosing a medium flip angle (15–20°) and a slightly longer TR appears to be optimal. The longer TR allows more inflow enhancement for the angiography and keeps the CSF from being oversuppressed. Shorter echo times improve the angiography by reducing uncompensated higher-order flow losses, but this decreases venous contrast. At 3 T, decreasing the echo time below 20 ms substantially degrades venous contrast and is not recommended; an echo time of 20 ms is preferred. For this reason, this technique shows promise at higher fields (>3 T) as shorter echo times can then be used without degrading venous contrast. It is possible to reduce uncompensated higher-order flow losses by increasing the read bandwidth and acquiring with high isotropic resolution. A high read bandwidth will reduce the time between spatial encoding and the echo readout; this is referred to as the field echo. Reducing the field echo improves the flow compensation by minimizing higher order flow effects which are proportional to higher powers of the field echo time (Fig. 12.5). Keeping the echo time long but reducing the field echo achieves good flow compensation while maintaining T2* contrast. This is only true, however, if you have a homogeneous field. If you have field inhomogeneities, from a poorly shimmed field, air–tissue interfaces, or air–bone interfaces, flow through these inhomogeneities will cause the blood to start collecting phase immediately after the excitation pulse, making the flow losses dependant on the echo time, not the field echo time. Unfortunately the carotid and parts of the MCA lie in regions of poor field homogeneity due to the sinuses causing both to experience nearly complete signal loss at long echo times despite very short field echoes. This can be seen in Fig. 12.6, where only the part of the MCA that is within the inhomogeneous field (as seen with the phase image) experiences bad signal loss. High isotropic resolution reduces dephasing across a voxel and thus can reduce flow losses. This does reduce the quality of the SWI phase image as an isotropic aspect ratio of 1:1 in-plane to through-plane resolution is not ideal for SWI

161

Fig. 12.5 MIP images from a long echo scan (20 ms) at different bandwidths (a) 60 Hz/pixel (b) 120 Hz/pixel (c) 235 Hz/pixel and (d) 465 Hz/pixel. Note that the flow compensation improves at higher bandwidths due to the shortened field echo time as noted by the decreased flow losses in the MCA (arrows)

[31, 32]. The isotropic aspect ratio causes the phase for veins of certain sizes and orientations to have opposite signs. This has the potential to confuse clinical interpretation and could reduce contrast in the SWI processed images. This lost contrast can be completely recovered, however, by postprocessing the images and applying a down-sampling filter to generate a more ideal aspect ratio of 1:4. A simple k-space crop can be used, which would be equivalent to a lower resolution acquisition with 1:4 aspect ratio. Alternatively, a sliding window complex average can be performed, which takes advantage of the fact that the higher resolution was collected. The sliding window filter takes the average of the complex signal (magnitude and phase) over four slices, advances a single slice, calculates the next average, and repeats until all slices are processed. In this way, the through-plane resolution is reduced but the same number of slices as in the original series is maintained (actually the slices will be reduced by a small amount, the collapsing factor – one or three slices in this case). By reconstructing the thick slabs in an overlapping pattern, optimal partial voluming of small structures is guaranteed, increasing their visibility. This reconstruction also offers a distinct advantage over the original k-space data in that it uses the high-pass filtered phase in the complex downsampling process, and not the original phase. This reduces the dephasing in areas of rapid phase change (air– tissue interfaces), improving the quality of the downsampled images. The single echo SWI contrast and venography, as shown in Fig. 12.7, are of good quality with the veins being well depicted. Likewise, nearly all of the arteries are well depicted; however, there is some signal loss in parts of the fast-flowing MCA due to flow dephasing from the long TE. The more distal arteries are well depicted with little to no signal loss.

162

Fig. 12.6 (a) 7 ms magnitude (b) 7 ms phase (c) 24 ms magnitude. Notice the more proximal part of the MCA (arrow) that is visible at 7 ms is significantly reduced in amplitude at 24 ms while the more dis-

Fig. 12.7 Single echo SWI dataset with TE = 20 ms, TR = 35 ms, FA = 15°, BW = 160 Hz/pixel (a) MIP over 64 mm, (b) mIP over 8 mm. (Images adapted or reprinted with permission of Haacke et al. [61].)

Blood Properties at Different Field Strengths For brain parenchyma, the T1 values have been found to generally increase with field strength [33] and the T2 values do not change much until you get above 3.0 T and then they start to fall precipitously [34]. The T2* values of all tissues behave differently than the T2 values. They also fall with field strength but they begin to decrease at lower fields and change significantly between 1.5 and 3.0 T [35]. Relaxation values for blood are difficult to measure accurately as blood is constantly flowing and ex vivo measurements can be challenging due to changes in oxygen saturation. However, T2* values for venous blood do decrease dramatically across commonly used field strengths and range from 97 ms at 1.5 T to 7.4 ms at 7.0 T [36]. The decreasing values of T2* for all tissues as field strength increases has important implications for SWI.

S. Barnes and E.M. Haacke

tal part is still very visible. The proximal part is exposed to field inhomogeneities from the sinuses (indicated with dashed line) causing them to dephase despite the very short field echo time

At 1.5 T relatively long echo times (40 ms) are demanded to give time for significant T2* contrast to develop. These long echo times are also needed to develop adequate contrast in the phase image according for 12.2. But as field strength increases the echo times can be shorted, approximately linearly with the increase in field strength, while still maintaining both T2* contrast and susceptibility contrast in the phase image. Phase contrast is maintained as 12.2 shows it is dependent on the product of echo time and field strength, and T2* contrast from the veins in the magnitude image is maintained due to the falling T2* values with field strength. The reduced echo time allows a significant time savings in SWI as one moves to higher field strengths. If the ratio of field strength remains the same as the ratio of the R2* relaxivities, then the SWI data can be produced identically between field strengths by simply keeping the product of B0TE constant. In practice, this is not the case as R2* values do not behave linearly, although going from 1.5 to 3 T, and dropping the echo from 40 to 20 ms, still generates excellent contrast SWI data.

The Role of MRI Contrast Agents T1 shortening contrast agents, such as gadolinium, can substantially improve the quality of SWI venography. The T1 shortening of the contrast agent causes an increase in the available signal for blood. The deoxyhemoglobin in the blood causes a slight frequency shift in the venous blood (this is what causes the phase contrast); this frequency shift allows the echo time to be chosen such that venous blood and brain parenchyma are out of phase and the signals will cancel. This cancelation allows the shortened T1 and increased signal of the blood to improve contrast in the

12 Susceptibility Weighted Imaging and MR Angiography

163

Fig. 12.8 SWI of a rat anesthetized with isoflurane and oxygen. Left images show precontrast and right images are 2.5 h post USPIO (P904, 300 mmol Fe/kg). Single slice SWI (top) and filtered phase images (bottom), notice the veins are much better visualized even 2.5 h post contrast

Fig. 12.9 mIP over 24 mm SW images (a) precaffeine and (b) approximately 1 h after a 200 mg caffeine tablet. The higher amounts of deoxyhemoglobin in the veins cause them to be much more visible postcaffeine

veins; this is referred to as T1–T2* coupling. The usual Gadolinium based contrast agents are paramagnetic and for a 1 mM concentration in the blood they increase the local susceptibility of the blood by 2/9ths of the usual BOLD effect. This in turn introduces an increase in the phase of the blood by 2/9ths, further improving SWI contrast. The possibility of using direct T2* shortening agents in the blood also exists with ultrasmall superparamagnetic iron oxides (USPIO) or superparamagnetic iron oxides (SPIO). While these have very large T2* shortening properties they act evenly on both arteries and veins, which causes the arteries to go dark along with the veins. While this does cause artery–vein ambiguity, the level of small vessel detail seen is quite impressive (Fig. 12.8). If this could be run in some sort of pre–post fashion to separate the arteries from veins it could find a use in imaging the small vasculature of the brain.

Role of Caffeine and Acetazolmide The cerebral vascular reserve of the brain can be tested in a number of ways. One is to stress the brain with caffeine which is a vasoconstrictive agent and a member of the methylxanthine family (which are adenosine antagonists [37]). Typically, two cups of coffee can elicit a strong vascular response [38]. The resulting reduced flow leads to an increase in the BOLD effect since the brain’s oxygen utilization remains constant. The slower flow then leads to a higher concentration of deoxyhemoglobin in the veins. This leads to an increase in the oxygen extraction fraction, OEF, to maintain the cerebral metabolic rate (CMRO2). The effect on SWI data is an improvement in visualization of the veins since deoxyhemoglobin increases, the local field increases, phase increases and local T2* effects increase [39]. The nice thing about using SWI as a high-resolution BOLD method is that the effects of caffeine can be seen throughout the entire brain (Fig. 12.9).

164

The effects of caffeine as seen with SWI have been investigated recently in a number of interesting papers [40]. Applying a region-of-interest based analysis, the authors reported an approximately exponential signal decay in venous vessels with time that was in agreement with a linear pharmacokinetic model of oral absorption of caffeine [40]. Venous response reached a maximum in the time interval of between 40 and 50 min, while allowing a significant differentiation between coffee drinkers (venous signal change: −16.5% ± 6.5%) and abstainers (venous signal change: −22.7% ± 8.3%). Only small signal changes of about −2% ± 1% were found in both gray and white matter and −1% ± 2% in the ventricles, in accordance with the earlier findings. The effects of caffeine can alter baseline cerebral blood flow and in this way modulate temporal dynamics of the BOLD response through alterations in the strength of neurovascular coupling [41–44]. Another agent that is used to study cerebral blood flow changes is the vasodilator acetazolamide (C4H6N4O3S2) also known as diamox. It has been used to study epilepsy, intracranial hypertension, glaucoma, and altitude sickness. It has been shown to increase cerebral blood flow from 30 [45] to about 50% [46, 47], and venous oxygen saturation by approximately 20% relative to the usual resting state levels [47]. As such it is used to assess hemodynamic reserve and vasomotor reactivity [48]. Acetazolamide inhibits carbonic anhydrase, which leads to the production of HCO3−. HCO3− in turn induces a local extracellular acidosis by increasing the concentrations of CO2 and H+ in the extracellular fluid in the brain, which is assumed to act as a stimulus for the increase in blood flow [49]. These increases in blood flow lead to a reduction in local oxygen saturation [50, 51].

S. Barnes and E.M. Haacke

it very useful in imaging hemorrhages from a variety of sources including trauma, stroke and aging, visualizing blood products and the vascularization of tumors, and high-resolution MR venography [9, 52–54]. It has also proven useful in other applications relating to iron such as measuring iron content in multiple sclerosis lesions, and iron accumulation in aging [55]. For a summary of clinical applications, please see the recent review by Mittal et al. [9]. Multiple sclerosis (MS) has long been thought to be an inflammatory demyelinating disease. However, both 70-yearold data [56] and recent imaging evidence [57] suggests that venous obstruction is closely related to and possibly a source of MS. Our own studies using SWI to evaluate the veins in the thalamostriate region and the iron content have shown that the iron content builds up at the confluence of the draining veins in structures such as the putamen, globus pallidus, and the caudate nucleus (Fig. 12.10). This fits with the theory that MS is a chronic cerebrospinal vascular insufficiency [57]. SWI is able to not only help visualize the small veins that appear to be at the center of the lesion [58] but also the iron content in the lesions and in the basal ganglia and thalamus.

Cerebral Microbleeds

SWI is particularly well suited for imaging venous blood as it is very sensitive to deoxyhemoglobin and other iron containing products such as ferritin and hemosiderin. This makes

Another area where SWI plays a key role is visualizing small hemorrhages thanks to the hemosiderin deposits that are eventually left behind (Figs. 12.11 and 12.12). Cerebral microbleeds (CMBs), while usually asymptomatic, are becoming an important biomarker for many other diseases such as hypertension, cerebral amyloid angiopathy, intracerebral hemorrhage, and hemorrhagic stroke [53]. SWI is more sensitive at detecting CMBs than standard T2* weighted imaging. A recent study showed that three times more CMBs could be detected with SWI compared to T2* weighted imaging [59]. SWI can also be useful in ruling out some common CMB mimics that are present in standard T2* weighted images. Calcium mimics, deposits of calcium that

Fig. 12.10 (a) SWI filtered phase showing increases in iron content directly related to the thalamostriate draining veins. (b) Minimum intensity projection processing SWI data showing the venous drainage

system for the basal ganglia and thalamus. (c) A radiographic image of the veins from a cadaver brain study (Courtesy of Georges Salamon, MD, David Geffen School of Medicine at UCLA, Los Angeles, CA)

Clinical Applications

12 Susceptibility Weighted Imaging and MR Angiography

165

much smaller than the voxel but still a source of spatially varying static fields, say for example for nanoparticles such as USPIOs, then T2¢ will be invariant at resolutions on the order of hundreds of microns or greater. However, the effects of diffusion are dependent on resolution via the gradients used to generate high-resolution images. The implications for SWI are that the phase remains invariant but T2* effects decrease at higher resolutions so SWI will maintain much of its BOLDlike contrast while T2* will become less and less sensitive. Therefore, we expect SWI to be an ideal method to for very high-resolution techniques such as detecting tiny CMBs. Fig. 12.11 Cerebral microbleeds caused by trauma in a motorcycle accident (a) SWI (b) SWI Phase

Fig. 12.12 Case with numerous cerebral microbleeds (CMBs) caused by severe CAA, imaged at 1.5 T (a) mIP over 8 mm of SW images (b) Corresponding FLAIR image

cause small hypointensities that look like CMBs, can be easily identified using the SWI phase image. Calcium will have an opposite phase shift compared to the iron in a true CMB as calcium is diamagnetic and iron is paramagnetic. If a true CMB has dark phase (as in a right-handed system) the calcium will have bright phase making it easily distinguishable. Flow voids, while not eliminated in SWI, are minimized as SWI uses a sequence that is flow compensated in all directions (read, phase, and partition). This minimizes signal loss due to flowing blood and should reduce the likelihood of confusing a CMB for a flow void.

Conclusion Importance of High-Resolution Imaging As one goes to higher and higher resolution the effects of local field variations diminishes. This is because the phase dispersion across a voxel is reduced, which reduces signal loss from dephasing. What this means is that T2¢, hence T2*, is not scale invariant. If the local fields are caused by objects

References 1. Nayler GL, Firmin DN, Longmore DB. Blood flow imaging by cine magnetic resonance. J Comput Assist Tomogr. 1986;10:715–722. 2. Nishimura DG, Macovski A, Pauly JM. Magnetic resonance angiography. IEEE Trans Med Imaging. 1986;5:140–151. 3. Haacke EM, Lenz GW. Improving MR image quality in the presence of motion by using rephasing gradients. AJR Am J Roentgenol. 1987;148:1251–1258. 4. Potchen EJ, Haacke EM, Siebert JE, Gottschalk A (eds). Magnetic Resonance Angiography: Concepts and Applications. St. Louis: Mosby; 1993:xiv, 650. 5. Reichenbach JR, Venkatesan R, Schillinger DJ, Kido DK, Haacke EM. Small vessels in the human brain: MR venography with deoxyhemoglobin as an intrinsic contrast agent. Radiology. 1997; 204:272–277. 6. Haacke EM, Xu Y, Cheng YC, Reichenbach JR. Susceptibility weighted imaging (SWI). Magn Reson Med. Sep 2004;52(3):612–618. 7. Haacke EM, Miao YW, Liu M, et al. Correlation of putative iron content as represented by changes in R2* and phase with age in deep gray matter of healthy adults. J Magn Reson Imaging. 2010;32:561–576. 8. Gelman N, Gorell JM, Barker PB, et al. MR imaging of human brain at 3.0 T: preliminary report on transverse relaxation rates and relation to estimated iron content. Radiology. 1999;210:759–767. 9. Mittal S, Wu Z, Neelavalli J, Haacke EM. Susceptibility-weighted imaging: technical aspects and clinical applications, part 2. AJNR Am J Neuroradiol. 2009;30:232–252. 10. Denk C, Rauscher A. Susceptibility weighted imaging with multiple echoes. J Magn Reson Imaging. 2010;31:185–191. 11. Reichenbach JR, Venkatesan R, Yablonskiy DA, Thompson MR, Lai S, Haacke EM. Theory and application of static field inhomogeneity effects in gradient-echo imaging. J Magn Reson Imaging. 1997;7:266–279. 12. Haacke EM, Dmitriy SL, Yablonskiy A, Lin W. In vivo validation of the bold mechanism: a review of signal changes in gradient echo functional MRI in the presence of flow. Int J Imaging Systems Technology. 1995;6:153–163. 13. Duyn JH, van Gelderen P, Li TQ, de Zwart JA, Koretsky AP, Fukunaga M. High-field MRI of brain cortical substructure based on signal phase. Proc Natl Acad Sci USA. 2007;104: 11796–11801. 14. Haacke EM, Mittal S, Wu Z, Neelavalli J, Cheng YC. Susceptibilityweighted imaging: technical aspects and clinical applications, part 1. AJNR Am J Neuroradiol. 2009;30:19–30. 15. Park SH, Moon CH, Bae KT. Compatible dual-echo arteriovenography (CODEA) using an echo-specific K-space reordering scheme. Magn Reson Med. 2009;61:767–774.

166 16. Du YP, Jin Z. Simultaneous acquisition of MR angiography and venography (MRAV). Magn Reson Med. 2008;59:954–958. 17. Xu Y, Haacke EM. An iterative reconstruction technique for geometric distortion-corrected segmented echo-planar imaging. Magn Reson Imaging. 2008;26:1406–1414. 18. Rauscher A, Barth M, Reichenbach JR, Stollberger R, Moser E. Automated unwrapping of MR phase images applied to BOLD MR-venography at 3 Tesla. J Magn Reson Imaging. 2003; 18:175–180. 19. Neelavalli J, Cheng YC, Jiang J, Haacke EM. Removing background phase variations in susceptibility-weighted imaging using a fast, forward-field calculation. J Magn Reson Imaging. 2009;29: 937–948. 20. Jin Z, Xia L, Du YP. Reduction of artifacts in susceptibilityweighted MR venography of the brain. J Magn Reson Imaging. Aug 2008;28(2):327–333. 21. Pandian DS, Ciulla C, Haacke EM, Jiang J, Ayaz M. Complex threshold method for identifying pixels that contain predominantly noise in magnetic resonance images. J Magn Reson Imaging. 2008;28:727–735. 22. McAuley G, Schrag M, Sipos P, et al. Quantification of punctate iron sources using magnetic resonance phase. Magn Reson Med. 2010;63:106–115. 23. Cheng YC, Hsieh CY, Neelavalli J, Haacke EM. Quantifying effective magnetic moments of narrow cylindrical objects in MRI. Phys Med Biol. 2009;54:7025–7044. 24. Liu T, Spincemaille P, de Rochefort L, Kressler B, Wang Y. Calculation of susceptibility through multiple orientation sampling (COSMOS): a method for conditioning the inverse problem from measured magnetic field map to susceptibility source image in MRI. Magn Reson Med. 2009;61:196–204. 25. de Rochefort L, Liu T, Kressler B, et al. Quantitative susceptibility map reconstruction from MR phase data using bayesian regularization: validation and application to brain imaging. Magn Reson Med. 2010;63:194–206. 26. Kressler B, de Rochefort L, Liu T, Spincemaille P, Jiang Q, Wang Y. Nonlinear regularization for per voxel estimation of magnetic susceptibility distributions from MRI field maps. IEEE Trans Med Imaging. 2010;29:273–281. 27. Shmueli K, de Zwart JA, van Gelderen P, Li TQ, Dodd SJ, Duyn JH. Magnetic susceptibility mapping of brain tissue in vivo using MRI phase data. Magn Reson Med. 2009;62:1510–1522. 28. Lin W, Mukherjee P, An H, et al. Improving high-resolution MR bold venographic imaging using a T1 reducing contrast agent. J Magn Reson Imaging.1999;10:118–123. 29. Deistung A, Dittrich E, Sedlacik J, Rauscher A, Reichenbach JR. ToF-SWI: simultaneous time of flight and fully flow compensated susceptibility weighted imaging. J Magn Reson Imaging. 2009; 29:1478–1484. 30. Barnes SR, Jin Z, Du YP, Deistung A. MR angiography and venography of the brain. In: Haacke EM, Reichenbach J, eds. Susceptibility Weighted Imaging in MRI: Basic Concepts and Clinical Applications: Wiley; 2011. 31. Xu Y, Haacke EM. The role of voxel aspect ratio in determining apparent vascular phase behavior in susceptibility weighted imaging. Magn Reson Imaging. 2006;24:155–160. 32. Deistung A, Rauscher A, Sedlacik J, Stadler J, Witoszynskyj S, Reichenbach JR. Susceptibility weighted imaging at ultra high magnetic field strengths: theoretical considerations and experimental results. Magn Reson Med. 2008;60:1155–1168. 33. Bottomley PA, Foster TH, Argersinger RE, Pfeifer LM. A review of normal tissue hydrogen NMR relaxation times and relaxation mechanisms from 1–100 MHz: dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med Phys. 1984;11:425–448.

S. Barnes and E.M. Haacke 34. Jezzard P, Duewell S, Balaban RS. MR relaxation times in human brain: measurement at 4 T. Radiology. 1996;199:773–779. 35. Peters AM, Brookes MJ, Hoogenraad FG, et al. T2* measurements in human brain at 1.5, 3 and 7 T. Magn Reson Imaging. 2007; 25:748–753. 36. Koopmans PJ, Manniesing R, Niessen WJ, Viergever MA, Barth M. MR venography of the human brain using susceptibility weighted imaging at very high field strength. MAGMA. 2008;21: 149–158. 37. Fredholm BB, Battig K, Holmen J, Nehlig A, Zvartau EE. Actions of caffeine in the brain with special reference to factors that contribute to its widespread use. Pharmacol Rev. 1999;51:83–133. 38. Haacke EM, Filleti CL, Gattu R, et al. New algorithm for quantifying vascular changes in dynamic contrast-enhanced MRI independent of absolute T1 values. Magn Reson Med. 2007;58:463–472. 39. Haacke EM, Hu C, Parrish T, Xu Y. Whole brain stress test using caffeine: effects on fMRI and SWI at 3 T. Proc Intl Soc Mag Reson Med. 2003;11:1731. 40. Sedlacik J, Helm K, Rauscher A, Stadler J, Mentzel HJ, Reichenbach JR. Investigations on the effect of caffeine on cerebral venous vessel contrast by using susceptibility-weighted imaging (SWI) at 1.5, 3 and 7 T. Neuroimage. 2008;40:11–18. 41. Behzadi Y, Liu TT. Caffeine reduces the initial dip in the visual BOLD response at 3 T. Neuroimage. 2006;32:9–15. 42. Chen Y, Parrish TB. Caffeine’s effects on cerebrovascular reactivity and coupling between cerebral blood flow and oxygen metabolism. Neuroimage. 2009;44:647–652. 43. Liau J, Perthen JE, Liu TT. Caffeine reduces the activation extent and contrast-to-noise ratio of the functional cerebral blood flow response but not the BOLD response. Neuroimage. 2008;42: 296–305. 44. Perthen JE, Lansing AE, Liau J, Liu TT, Buxton RB. Caffeineinduced uncoupling of cerebral blood flow and oxygen metabolism: a calibrated BOLD fMRI study. Neuroimage. 2008;40:237–247. 45. Okazawa H, Yamauchi H, Sugimoto K, Toyoda H, Kishibe Y, Takahashi M. Effects of acetazolamide on cerebral blood flow, blood volume, and oxygen metabolism: a positron emission tomography study with healthy volunteers. J Cereb Blood Flow Metab. 2001;21:1472–1479. 46. Schytz HW, Wienecke T, Jensen LT, Selb J, Boas DA, Ashina M. Changes in cerebral blood flow after acetazolamide: an experimental study comparing near-infrared spectroscopy and SPECT. Eur J Neurol. 2009;16:461–467. 47. Vorstrup S, Henriksen L, Paulson OB. Effect of acetazolamide on cerebral blood flow and cerebral metabolic rate for oxygen. J Clin Invest. 1984;74:1634–1639. 48. Griffiths PD, Gaines P, Cleveland T, Beard J, Venables G, Wilkinson ID. Assessment of cerebral haemodynamics and vascular reserve in patients with symptomatic carotid artery occlusion: an integrated MR method. Neuroradiology. 2005;47:175–182. 49. Lassen NA. Is central chemoreceptor sensitive to intracellular rather than extracellular pH? Clin Physiol. 1990;10:311–319. 50. Hedera P, Lai S, Lewin JS, et al. Assessment of cerebral blood flow reserve using functional magnetic resonance imaging. J Magn Reson Imaging. 1996;6:718–725. 51. Sedlacik J, Kutschbach C, Rauscher A, Deistung A, Reichenbach JR. Investigation of the influence of carbon dioxide concentrations on cerebral physiology by susceptibility-weighted magnetic resonance imaging (SWI). Neuroimage. 2008;43:36–43. 52. Sehgal V, Delproposto Z, Haacke EM, et al. Clinical applications of neuroimaging with susceptibility-weighted imaging. J Magn Reson Imaging. 2005;22:439–450. 53. Greenberg SM, Vernooij MW, Cordonnier C, et al. Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol. 2009;8:165–174.

12 Susceptibility Weighted Imaging and MR Angiography 54. Tong KA, Ashwal S, Obenaus A, Nickerson JP, Kido D, Haacke EM. Susceptibility-weighted MR imaging: a review of clinical applications in children. AJNR Am J Neuroradiol. 2008;29:9–17. 55. Haacke EM, Cheng NY, House MJ, et al. Imaging iron stores in the brain using magnetic resonance imaging. Magn Reson Imaging. 2005;23:1–25. 56. Putnam TJ. Evidences of vascular occlusion in multiple sclerosis and “encephalomyelitis”:. Archives of Neurology and Psychiatry. 1937;37:1298–1321. 57. Zamboni P, Galeotti R, Menegatti E, et al. A prospective open-label study of endovascular treatment of chronic cerebrospinal venous insufficiency. J Vasc Surg. 2009;50:1348–1358, e1341–1343.

167 58. Tan IL, van Schijndel RA, Pouwels PJ, et al. MR venography of multiple sclerosis. AJNR Am J Neuroradiol. 2000;21: 1039–1042. 59. Nandigam RN, Viswanathan A, Delgado P, et al. MR imaging detection of cerebral microbleeds: effect of susceptibility-weighted imaging, section thickness, and field strength. AJNR Am J Neuroradiol. 2009;30:338–343. 60. Haacke EM, Tang J, Neelavalli J, Cheng YC. Susceptibility mapping as a means to visualize veins and quantify oxygen saturation. J Magn Reson Imaging. 2010;32:663–676. 61. Haacke EM, Reichenbach J, eds. Susceptibility Weighted Imaging in MRI. Hoboken, NJ: Wiley; 2011.

wwwwwwwwwwww

Non-Cartesian MR Angiography

13

Walter Block and Oliver Wieben

Introduction Magnetic resonance angiography (MRA) has several compelling features, including volumetric data acquisition and the availability of several contrast mechanisms. These can be used for imaging with and without exogenous contrast agents and to obtain anatomical as well as functional information, for example in the form of perfusion or velocity measurements. However, such MRA acquisitions require longer imaging times, which challenge scan completion without artifacts from physiological or involuntary patient motion, particularly for dynamic acquisitions, such as bolus chasing, real-time imaging, or cardiac-gated acquisitions. The inherently 3D nature of MRI allows for fine depiction of vascular territories. However, as MR samples its data in an alternative Fourier domain, only a relatively few samples can be obtained at one time. If one desires to obtain a timeresolved depiction as injected contrast enhances the vasculature, a four-dimensional space must be sampled. In one of the most demanding angiographic tasks, phase-contrast (PC) angiography can obtain quantitative velocity data that is resolved to a spatial coordinate and a cardiac cycle. This imaging task requires four-dimensional sampling (three spatial dimensions, time) of a vector quantity, the flow vector with three directional velocity components constrained by physiological and clinical implementation limits. With such demanding acquisition requirements over so many dimensions, non-Cartesian trajectories have been developed to offer increased performance in MRA. Non-Cartesian trajectories can offer increased performance in several ways, although not always simultaneously. NonCartesian methods can better utilize limited gradient hardware

W. Block, PhD () • O. Wieben, PhD Departments of Biomedical Engineering, Medical Physics, and Radiology, University of Wisconsin–Madison, Madison, WI, USA e-mail: [email protected]

speed, improve the efficiency in which k-space is covered, decrease sensitivity to motion, and improve flow properties. Some non-Cartesian methods often offer a variable sampling trajectory, where the center of k-space is sampled more often than higher spatial frequencies. These sampling patterns support time-resolved imaging in reconstruction methods that vary in performance, speed, accuracy, and complexity. In general, accuracy and performance improve when complexity and time within the reconstruction task are accounted for. Non-Cartesian methods offer possibilities to exploit the sparse nature of the vascular imaging task and the correlation between temporal frames in a time-resolved study. Vascular imaging differs from static imaging of many regions of the body in several important ways. Often, vascular images are much more sparse than nonangiographic images of other parts of the body. The sparse nature of these images can be due to the high contrast provided by injected contrast agents. The ability to subtract out static signal, as in phase-contrast imaging or through use of a precontrast mask, further increases the sparse nature of vascular images. In timeresolved imaging, significant correlation may exist between imaging frames and thus each image volume may not need to be acquired completely separately. The chapter first describes non-Cartesian acquisition and reconstruction theory, loosely classified as spiral and radial trajectories. A brief summary of methods being utilized to provide consistent performance with non-Cartesian methods is provided as these trajectories are generally less robust to several system and patient-induced imperfections than Cartesian methods. Methods that utilize non-Cartesian trajectories to improve time-resolved angiography are then discussed. Similar concepts used for accelerating time-resolved imaging can be used for quantitative resolve flow and perfusion throughout the cardiac cycle. Finally, trajectories with variable sampling densities are useful when using image estimation methods to increase performance. Here, some form of a priori information is used to constrain the reconstruction process.

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_13, © Springer Science+Business Media, LLC 2012

169

170

Non-Cartesian Trajectory Design In MR imaging, data can be sampled in k-space on any 2D or 3D trajectory that the time-varying gradients and safety regulations regarding peripheral nerve and muscle stimulation and tissue heating can support. Although the first MR imaging method proposed the acquisition of projections [1], spin-warp imaging on a Cartesian sampling grid [2] became the predominantly used trajectory. The acquisition of data on a rectilinear grid is robust against inhomogeneities in the static magnetic field resulting from imperfections the scanner. In non-Cartesian acquisitions, these inhomogeneities introduce off-resonance effects which in turn cause blurring of the PSF. Advances in scanner hardware improved the field homogeneity and alternative sampling patterns with nonuniform sampling densities revisited. Projection imaging [3] is a CT-like acquisition, where each echo represents a radial line traversing through the center of k-space. This method offers good suppression of motion artifacts and allows for imaging with very short echo times when the projections start in the center because they do not require any prewinding gradients. A disadvantage is the prolongation total imaging time because of the redundant oversampling of the central k-space region. k-space can be sampled with fewer echoes using spiral trajectories [4]. Images can be acquired with as little as 30–40 echoes, but this scheme is very sensitive to off-resonance effects. The sampling grids for these acquisitions schemes are shown in Fig. 13.1. The trajectories can also be extended or combined into 3D acquisitions, for example for truly 3D radial, 3D spiral, cone, stack of spheres, shells trajectory, cones, and spiral PR (Fig. 13.2). Hybrid 3D sampling patterns with non-Cartesian in plane encoding and traditional Fourier slice encoding have also been implemented, predominantly for the sampling of imaging volumes of shorter dimensions in the through plane direction. More complete reviews of sampling patterns can be found elsewhere in the literature [5].

W. Block and O. Wieben

General Considerations Sampling Region Sampling a cylinder of k-space saves 21.5% of the sampled space relative to a cube while sampling a sphere saves 47.6% of the required samples. While non-Cartesian trajectories can easily be tuned to cylindrical and spherical k-space regions, selection of phase-encoding and slice-encoding locations can achieve cylindrical sampling spaces. Gradient Spoiling As the readout direction is changing throughout a nonCartesian scan, some attention has to be given to the method that spoils transverse signal in gradient-recalled sequences. Winding the magnetization to the same physical k-space location after each readout is generally a good way to remove variations in the transverse steady-state signal throughout the scan. Field of View In general, non-Cartesian trajectories are designed to sample along the readout direction at k-space intervals of 1/FOV as in Cartesian trajectories, where the FOV is the largest dimension of the acquired volume. Sampling along the readout dimension is constrained by the maximum slew rate achievable, and thus k-space sampling intervals often vary, especially at the beginning of the readout. Repetitions of the spiral or radial readout are then rotated in such a way to fill k-space. To provide full k-space sampling, enough repetitions of the model readout are needed such that the space of the interleaves is less than 1/FOV in all areas of k-space. Off Resonance The extent of phase accrued by off-resonance spins during each readout is directly proportional to the readout duration. The amount of needed effort to remedy off-resonance effects, thus, increases with readout duration. The sophistication of any needed off-resonance processing depends also of course on the amount of inhomogeneity present in the vascular territory of interest. While the appearance of off resonance

Fig. 13.1 Strategies for 2D k-space sampling. Shown are the spin-warp (a), radial sampling (b), and interleaved spiral imaging (c) trajectory as examples for sampling patterns used in MR angiography

13

Non-Cartesian MR Angiography

Fig. 13.2 Strategies for 3D k-space sampling include cylindrical stack of spirals (a), spherical stack of spirals (b), a spiral–radial hybrid trajectory (c), concentric shells (d), a true 3D radial trajectory (e), and a stack of cones (f) (figure adopted from Irarrazabal and Nishimura [5]; reprinted with permission)

Fig. 13.3 Asymmetric FOV imaging with a radial trajectory with increased sampling density along the near-vertical projections. This creates a larger, horizontal FOV than vertical FOV (b) (courtesy of Steve Kecskemeti, University of Wisconsin-Madison)

varies with trajectory, off-resonance effects are generally manifested by blurring and signal dropout in non-Cartesian trajectories.

Asymmetric FOVs The largest spacing between k-space samples in the ensemble of readouts collected in the acquisition is usually designed to be no smaller than 1/FOV. The largest spacing may be designed to be 1/FOV for full sampling or greater than 1/ FOV trajectories which one intentionally oversamples in regions of sparse and high-contrast vasculature. Asymmetric FOVs that are tuned to certain vascular territory are generally easier to achieve with Cartesian trajectories than nonCartesian methods; however, non-Cartesian asymmetric FOVs are possible as shown in Fig. 13.3 [6]. Flow Sensitivity In general, trajectories whose first moment is small near the center of k-space have better flow properties [7]. Trajectories for which the first moment changes smoothly as a function of k-space radius also are more robust in MRA. In general, trajectories which originate at the center of k-space without previous slice encoding have more advantageous flow properties.

171

Sampling Density The sampling density of radial trajectories varies significantly with k-space radius, falling off as 1 / k (r ) for 2D radial trajectories and 1 / k 2 (r ) for true 3D radial trajectories. Sampling some spatial frequencies more often at the expense of others has liabilities when trying to cover all of k-space rapidly and has deleterious effects on SNR compared to flat sampling trajectories [8]. Oversampling lower spatial frequencies had shown advantages for representing time-resolved imaging [9–11], motion artifact suppression, inherent field map and coil sensitivity generation, and constrained reconstruction methods [12–14]. In general, radial trajectories in MRA have emphasized the value of variable sampling density while working to deemphasize deleterious effects from an inefficient coverage of k-space. While their acquisition is less sophisticated than spiral trajectories, the short acquisition time required for each radial line or projection significantly limits offresonance effects. Radial waveforms are generally easier to program in pulse sequences as well. Spiral waveform design initially emphasized efficient, rapid k-space coverage with flat sampling density [5, 15], where only a minimum portion of the waveform was slew rate limited. Meyers presented an analytical expression to approximate the density of spirals [4] throughout their trajectory. Very simple methods that grid spiral data points to the nearest neighbor on an oversized Cartesian matrix have also been demonstrated [16] which simplify the density compensation computation. More recently, spiral design has incorporated variable sampling density to mitigate effects from aliasing from outside the FOV and to provide some of the advantages of oversampling in representing time-resolved images volumes [17]. The advantages of spiral acquisitions grow with longer readout duration, though these increase problems with off resonance. The trajectories shown in Fig. 13.2 were first proposed by Irarrazaval and have since been utilized in numerous examples of spiral MRA, including stack of stars in the coronaries [18], cones in the peripheral vasculature [19], whole heart 3D radial imaging [20, 21], and spiral projection imaging [22]. As an example, most 3D breath-hold coronary imaging can cover only a thin plane within the breath-hold requirement. In Fig. 13.4, a variable density spiral design is used to cover the entire heart with 0.8 × 0.8 × 1.6-mm resolution in just 17 heartbeats [23].

Spiral Trajectory Design Although numerous implementations are possible, most spiral trajectories have been based on an Archimedes spiral. These trajectories follow the basis equation: k (t ) = λq (t )e - iq ( t ) .

172

W. Block and O. Wieben

Fig. 13.5 Radial sampling schemes for projections from −kr,max to +kr,max (a) and from 0 to +kr,max (b). The starting point for each readout is shown as an unfilled circle and the end point with an arrow head. Both schemes are characterized by a constant radial sampling interval Dkr and a maximum angular sampling interval Dkj,max

Projection Imaging More specific aspects of the design of various radial trajectories and their effects on PSFs are next provided. Fig. 13.4 Detail of small coronary branches. Images (a–c) were acquired with a resolution 0.8 × 0.8 × 1.6 mm3. Image (d) was acquired with a resolution of 1 × 1 × 2 mm3 and reconstructed with the iterative algorithm. Image (a) shows a detail of the conus artery. The distal RCA, including the posterior left ventricular artery (PL) and posterior descending artery (PDA), is displayed in image (b). A lesser cardiac vein (CV) is also shown. Image c shows a detail of the mid to distal RCA with an acute marginal right ventricle (RV) branch. (d) The mid LAD and diagonal branches (courtesy of J. Santos et al. [23]; reprinted with permission)

The desired gradient waveforms are given by derivative of the k-space trajectory and scaled by the inverse of the gyromagnetic ratio. Linear functions of q(t) lead to inefficient spirals with constant angular speed. The intuitive choice for efficient coverage of k-space would use a constant velocity spiral, where q (t ) = t . As this choice is not realizable in regions of the spiral where the slew rate is limited, trade-offs in the formulation of q(t) between constant angular speed and constant velocity were formulated by Bornert et al. [24]. While a more accurate solution for optimal use of gradient slew rate was formulated by King et al. [25], this solution required significant computation. A closed-form expression, which produces images which are indiscernible from the optimal solution, is provided by Glover et al. [26]. The complexity of spiral trajectory has created numerous strong publications, where computational power is often used to create shorter trajectories, more accurate sampling density functions, and more powerful off-resonance correction methods. In many cases, simpler approximations can provide adequate performance for many vascular applications. A strong review of these trade-offs is provided by Block and Frahm [27].

2D Projection Imaging In 2D projection imaging, each readout traverses through the center of k-space. The sampling trajectory can be described in polar coordinates with a radial component kr and an angle j. A total of Np repetitions are acquired with Nr samples and a sampling interval Dkr along the readout direction. As the 1D Fourier transform of each repetition provides a projection of the object, the technique is also known as projection reconstruction (PR). The ensemble of transformed projections forms a sonogram, similar to computed tomography (CT) reconstruction, which can then be reconstructed with filtered backprojection. Projection imaging, also known as radial sampling, leads to a nonuniform sampling density with emphasis on the low spatial frequencies. Let us consider the case, where each projection starts at −kr,max, traverses through the origin, and ends at +kr,max as shown in Fig. 13.5a. The radial sampling interval Dkr supports an alias-free reconstruction of distance D = 1/Dkr along the readout. The largest angular sampling distance Dkf,max occurs between adjacent spokes at the maximum sampled spatial frequency: Dkj ,max =

pN r Dk . 2N p r

(13.1)

According to the Nyquist theorem, sampling with Dkj,max = Dkr produces isotropic resolution over a radial FOV with a diameter D. This optimal sampling requires the following projections: N p,opt =

p Nr . 2

(13.2)

13

Non-Cartesian MR Angiography

173

More projections do not provide better spatial resolution or a larger FOV while fewer projections reduce the artifactfree FOV. In comparison, spin-warp imaging requires only Nr readouts (2/p = 63.7% less) for a squared FOV with identical resolution. This decrease in sampling efficiency is due to the oversampling of central k-space. It is important to note that any signal from outside the circular FOV does cause data inconsistencies in the projections and result in streak artifacts in the image. The band-pass filter applied to the received signal does limit the signal contributions along the readout direction but not perpendicular to it. While the repeated sampling of low spatial frequencies in each readout decreases the scan efficiency, radial sampling has very desirable properties for certain applications. Projection imaging is more robust to bulk motion because of the averaging effects from repeated sampling of the low spatial frequencies and more tolerable streak artifacts [3, 28]. This is also advantageous in diffusion-weighted imaging, where large gradients amplify artifacts from motion [28–30]. The trajectory can be modified so that each projection starts in the k-space origin (kr = 0) as shown in Fig. 13.5b. A free induction decay (FID) can then be acquired for the imaging of tissues with very short transverse relaxation times T2, such as in the lungs [31, 32]. Projection imaging can also be advantageous to suppress displacement artifacts in flow imaging [7]. Continuous and interleaved radial acquisitions were proposed for dynamic imaging in studies of the joints [33, 34], catheter tracking in interventional MR [35, 36], swallowing exams [37], and cardiac imaging [37]. The properties of angular undersampling for faster imaging have been explored in various studies and are discussed below.

Undersampled 2D Projection Imaging If the number of projections is decreased below Np,opt, then the angular sampling interval Dkj,max exceeds the radial interval Dkr and the high spatial frequencies are not sampled adequately. This leads to a reduced artifact-free FOV (rFOV) with a diameter d given by the inverse of the largest sampling interval: d=

1 Dkj ,max

=

2N p pDkr N r

.

(13.3)

The ratio of the diameters of the reduced FOV and the full FOV is given by Dkr d 2 Np = = . D Dkj ,max p N r

(13.4)

Figure 13.6 shows the PSFs for a fully sampled radial trajectory (Np = p/2 Nr) and with a reduced number of projections (Np 75%) RAS were found. Patients with intermediate RAS showed nonsignificantly decreased perfusion parameters. In this study and in another study by Michoux and colleagues, significant correlations between raising serum creatinine levels and decreasing renal perfusion parameters were found as well indicating that renal first pass perfusion parameters may reflect – at least to a certain degree – renal function [54, 77]. A smaller study by Vallee and colleagues [54] included four renal transplants with RAS and seven renal transplants with renal failure and reported on significantly reduced blood flow in transplants with either RAS or with renal failure where the latter showed a smaller residual perfusion. In a further study, the same group investigated 30 patients with normal renal function or chronic renal failure the cortical and medullary perfusion [77]. Significant reductions of the cortical perfusion, the medullary perfusion, and of the accumulation of contrast media in the medulla were found in the presence of renal failure. Similarly, detection of segmental perfusion deficits in renal transplants with focal rejection was also reported with DCE-MRI [81]. A further application of DCE-MRI is the evaluation of kidneys after stent placement [11, 64] when the renal artery cannot be assessed with CE-MRA due to stent-induced susceptibility artifacts (Fig. 21.8). DCE-MRI allows demonstrating normalized perfusion parameters after successful stent placement and hence proves the patency of the stent.

Future Applications Several clinical trials aimed at elucidating the optimal treatment strategy in patients with RAS have recently been published or are in the final phase of follow-up. The conventional wisdom, i.e., to dilate and/or stent in the presence of 50% or greater RAS, was greatly challenged in 2000 with the appearance of the results of the study by van Jaarsveld et al. [8], who found no difference at 12-months follow-up in 106 patients with RAS and hypertension randomized between medical antihypertensive therapy and angioplasty. The Dutch benefit of stent placement and blood pressure and lipid-lowering for the prevention of progression of renal dysfunction caused by atherosclerotic ostial stenosis of the renal artery (STAR) trial also reported a lack of effect of stenting. In 140 patients with impaired renal function and RAS >50%, stent placement had no clear effect on progression of impaired renal function but was associated with significant procedurerelated complications [82]. Another large study that provides evidence against renal artery revascularization is the randomized but unblinded ASTRAL trial that was recently published in the New England Journal of Medicine [83]. The ASTRAL trial found no evidence of clinical benefit from renal artery revascularization over optimized medical treatment in 806 patients with suspected atherosclerotic RAS followed for a median of 34 months. On the other hand, substantial risks were reported in 23/403 (5.7%) patients randomized to the interventional arm of the study [83]. A final study that is currently still underway is the 1,000-patient cardiovascular outcomes with renal atherosclerotic lesions (CORAL) study [84], which aims to address a similar question to ASTRAL. Results of this trial are expected in 2011.

292

Fig. 21.8 MRA and perfusion measurements pre- and postintervention. In (a), a coronal thin-slab MIP of the MRA (1.5 T, 1 × 0.9 × 1 mm³, PAT factor 2) of a 55-year-old male patient with hypertension who presented with a proximal high-grade stenosis of the right renal artery is shown. After stenting (left image), the stent artifact disrupts the local magnetic field so that no vascular enhancement can be appreciated at this site. In order to assess the renal parenchymal blood flow and success of the intervention renal perfusion measurement were performed at the baseline visit (b, right column) as well as after the intervention (b, left column). Before the intervention delayed perfusion of the affected,

Common to all of these trials is selection of patients purely based on the presence of RAS, not taking into account the functional consequences of the stenosis nor the degree of renal impairment. Future trials using combined MRA/renal MRI protocols will have to be performed to investigate whether such a protocol leads to optimized selection of patients in whom revascularization therapy is indeed beneficial. Apart from the above, the detection and differentiation of renoparenchymal disease independent from the presence of RAS may be another suitable indication for functional MR imaging techniques (Fig. 21.9). Larger single-center studies on this topic are currently being undertaken. Finally, it is worthwhile to mention a new therapeutic concept in the treatment of therapy-resistant RVH. Recently, a very encouraging multicenter safety and proof-of-principle study employing renal sympathetic denervation in 50 patients was published in The Lancet [85]. This novel therapeutic concept seems to be much more efficacious compared to simple renal artery dilatation and/or stent placement. Krum et al. [85] found a mean reduction of 27 and 17 mmHg in systolic and diastolic office blood pressure at 12 months follow-up. If this effect were sustained in larger and randomized trials, it would certainly constitute a revolution in treatment of RVH.

T. Leiner and H. Michaely

right side could be appreciated with hypo-enhancement compared to the left side. After the intervention there was again a bilateral regular enhancement of both kidneys. The perfusion changes of the affected right kidney can be visualized semiquantitatively by using signal intensity versus time curves as done in this example (c). Comparing the preinterventional signal (red line) with the postinterventional signal (black line), a significant change can be appreciated. The impaired perfusion of the right kidney before intervention is reflected by the slower upslope and the delayed and lowered peak signal intensity in the semiquantitative assessment (Reprinted with permission from Michaely et al. [88].)

Conclusions Although IA-DSA is still regarded as the most accurate test for anatomical detection of RAS, MRA is an attractive noninvasive alternative in the diagnostic workup of patients suspected of having RVH. In addition to anatomical diagnosis of RAS, CE-MRA enables precise quantification of the degree of renal impairment using MR perfusion sequences. Additional studies are needed to establish reference values, to determine the optimal postprocessing protocols, and to investigate whether such a comprehensive MR imaging protocol improves selection of patients in whom revascularization is beneficial. Because the prevalence of RAS among patients with hypertension is low, the cost-effectiveness of any diagnostic strategy is sensitive to the pretest probability of RVH. Therefore, careful clinical evaluation in order to achieve a pretest probability of at least 20% is an essential component in the workup of patients suspected of having RVH [86, 87]. Because missing RVH may have serious consequences, the most important requirement for an alternative test is that it has high sensitivity. The combination of renal artery imaging and assessment of renal perfusion will undoubtedly lead to better selection of patients who will benefit from interventional therapy.

21

Renal Vascular Diseases

Fig. 21.9 Different time frames from the TurboFLASH perfusion study in a healthy volunteer who nicely demonstrates a good corticomedullary differentiation and a normal excretory function of the kidney (a). In contrast, different time frames of the same sequence in a patient with hypertension show very poor cortical enhancement after contrast agent administration, also no excretion can be demonstrated (b). In the absence of renal artery stenosis as can be seen on this thin MIP of the

References 1. Safian RD, Textor SC. Renal-artery stenosis. N Engl J Med. 2001;344:431–442. 2. Textor SC, Wilcox CS. Renal artery stenosis: a common, treatable cause of renal failure? Annu Rev Med. 2001;52:421–442. 3. Crook ED. The role of hypertension, obesity, and diabetes in causing renal vascular disease. Am J Med Sci. 1999;317:183–188. 4. Cheung CM, Hegarty J, Kalra PA. Dilemmas in the management of renal artery stenosis. Br Med Bull. 2005;73–74:35–55. 5. Ives NJ, Wheatley K, Stowe RL, et al. Continuing uncertainty about the value of percutaneous revascularization in atherosclerotic renovascular disease: a meta-analysis of randomized trials. Nephrol Dial Transplant. 2003;18:298–304. 6. Plouin PF, Chatellier G, Darne B, Raynaud A. Blood pressure outcome of angioplasty in atherosclerotic renal artery stenosis: a randomized trial. Essai Multicentrique Medicaments vs Angioplastie (EMMA) Study Group. Hypertension. 1998;31:823–829. 7. van de Ven PJ, Kaatee R, Beutler JJ, et al. Arterial stenting and balloon angioplasty in ostial atherosclerotic renovascular disease: a randomised trial. Lancet. 1999;353:282–286.

293

arterial phase MRA, the patient underwent biopsy which revealed extensive nephrosclerosis (c). Comparison of the cortical SIVT curve of the healthy volunteer (green, upper line) and of the patient (red lower line) shows a normal SIVT curve for the volunteer (d). The patient’s SIVT is grossly pathologic with a delayed upslope, a lacking first pass peak, and a diminished overall signal (Reprinted with permission from Michaely et al. [81].)

8. van Jaarsveld BC, Krijnen P, Pieterman H, et al. The effect of balloon angioplasty on hypertension in atherosclerotic renal-artery stenosis. Dutch Renal Artery Stenosis Intervention Cooperative Study Group. N Engl J Med. 2000;342:1007–1014. 9. Webster J, Marshall F, Abdalla M, et al. Randomised 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. 1998;12:329–335. 10. Leiner T. Magnetic resonance angiography of abdominal and lower extremity vasculature. Top Magn Reson Imaging. 2005;16:21–66. 11. Michaely HJ, Sourbron S, Dietrich O, Attenberger U, Reiser MF, Schoenberg SO. Functional renal MR imaging: an overview. Abdom Imaging. 2006;32:758–771. 12. Detection, evaluation, and treatment of renovascular hypertension. Final report. Working Group on Renovascular Hypertension. Arch Intern Med. 1987;147:820–829. 13. Slovut DP, Olin JW. Fibromuscular dysplasia. N Engl J Med. 2004; 350:1862–1871. 14. Schreij G, de Haan MW, Oei TK, Koster D, de Leeuw PW. Interpretation of renal angiography by radiologists. J Hypertens. 1999;17:1737–1741.

294 15. Kaufman JA. Renal Arteries. In: Kaufman JA, Lee MJ, eds. Vascular and Interventional Radiolgy. The Requisites. Philadelphia: Mosby; 2004:323–349. 16. Vasbinder GB, Nelemans PJ, Kessels AG, et al. Accuracy of computed tomographic angiography and magnetic resonance angiography for diagnosing renal artery stenosis. Ann Intern Med. 2004; 141:674–682; discussion 682. 17. Bude RO, Forauer AR, Caoili EM, Nghiem HV. Is it necessary to study accessory arteries when screening the renal arteries for renovascular hypertension? Radiology. 2003;226:411–416. 18. Vasbinder GB, Nelemans PJ, Kessels AG, Kroon AA, de Leeuw PW, van Engelshoven JM. Diagnostic tests for renal artery stenosis in patients suspected of having renovascular hypertension: a metaanalysis. Ann Intern Med. 2001;135:401–411. 19. Hoogeveen RM, Bakker CJ, Viergever MA. Limits to the accuracy of vessel diameter measurement in MR angiography. J Magn Reson Imaging. 1998;8:1228–1235. 20. Schoenberg SO, Prince MR, Knopp MV, Allenberg JR. Renal MR angiography. Magn Reson Imaging Clin N Am. 1998;6:351–370. 21. Michaely HJ, Kramer H, Oesingmann N, et al. Intraindividual Comparison of MR-Renal Perfusion Imaging at 1.5 T and 3.0 T. Invest Radiol. 2007;42:406–411. 22. Vasbinder GB, Maki JH, Nijenhuis RJ, et al. Motion of the distal renal artery during three-dimensional contrast-enhanced breathhold MRA. J Magn Reson Imaging. 2002;16:685–696. 23. Maki JH, Chenevert TL, Prince MR. Three-dimensional contrastenhanced MR angiography. Top Magn Reson Imaging. 1996;8: 322–344. 24. Weiger M, Pruessmann KP, Kassner A, et al. Contrast-enhanced 3D MRA using SENSE. J Magn Reson Imaging. 2000;12:671–677. 25. Bakker J, Beek FJ, Beutler JJ, et al. Renal artery stenosis and accessory renal arteries: accuracy of detection and visualization with gadolinium-enhanced breath-hold MR angiography. Radiology. 1998;207:497–504. 26. Shetty AN, Bis KG, Kirsch M, Weintraub J, Laub G. Contrastenhanced breath-hold three-dimensional magnetic resonance angiography in the evaluation of renal arteries: optimization of technique and pitfalls. J Magn Reson Imaging. 2000;12:912–923. 27. Willoteaux S, Faivre-Pierret M, Moranne O, et al. Fibromuscular dysplasia of the main renal arteries: comparison of contrastenhanced MR angiography with digital subtraction angiography. Radiology. 2006;241:922–929. 28. Soulez G, Pasowicz M, Benea G, et al. Renal artery stenosis evaluation: diagnostic performance of gadobenate dimeglumine-enhanced MR angiography – comparison with DSA. Radiology. 2008;247:273–285. 29. Garovic VD, Achauer MA, Kittner T, Horak D, Sheng R, Stanson AW. Comparison of gadodiamide-enhanced MR angiography to intraarterial digital subtraction angiography for evaluation of renal artery stenosis: results of a phase III multicenter trial. J Magn Reson Imaging. 2010;31:390–397. 30. Wilson GJ, Maki JH. Non-contrast-enhanced MR imaging of renal artery stenosis at 1.5 tesla. Magn Reson Imaging Clin N Am. 2009;17:13–27. 31. Spuentrup E, Manning WJ, Bornert P, Kissinger KV, Botnar RM, Stuber M. Renal arteries: navigator-gated balanced fast field-echo projection MR angiography with aortic spin labeling: initial experience. Radiology. 2002;225:589–596. 32. Fenchel M, Martirosian P, Langanke J, et al. Perfusion MR imaging with FAIR true FISP spin labeling in patients with and without renal artery stenosis: initial experience. Radiology. 2006;238:1013–1021. 33. Katoh M, Buecker A, Stuber M, Gunther RW, Spuentrup E. Freebreathing renal MR angiography with steady-state free-precession (SSFP) and slab-selective spin inversion: initial results. Kidney Int. 2004;66:1272–1278.

T. Leiner and H. Michaely 34. Katoh M, Spuentrup E, Stuber M, Hoogeveen R, Gunther RW, Buecker A. Free-breathing renal magnetic resonance angiography with steady-state free-precession and slab-selective spin inversion combined with radial k-space sampling and water-selective excitation. Magn Reson Med. 2005;53:1228–1233. 35. Wyttenbach R, Braghetti A, Wyss M, et al. Renal artery assessment with nonenhanced steady-state free precession versus contrastenhanced MR angiography. Radiology. 2007;245:186–195. 36. Coenegrachts KL, Hoogeveen RM, Vaninbroukx JA, et al. Highspatial-resolution 3D balanced turbo field-echo technique for MR angiography of the renal arteries: initial experience. Radiology. 2004;231:237–242. 37. Herborn CU, Watkins DM, Runge VM, Gendron JM, Montgomery ML, Naul LG. Renal arteries: comparison of steady-state free precession MR angiography and contrast-enhanced MR angiography. Radiology. 2006;239:263–268. 38. Maki JH, Wilson GJ, Eubank WB, Glickerman DJ, Millan JA, Hoogeveen RM. Navigator-gated MR angiography of the renal arteries: a potential screening tool for renal artery stenosis. AJR Am J Roentgenol. 2007;188:W540-546. 39. Maki JH, Wilson GJ, Eubank WB, Glickerman DJ, Pipavath S, Hoogeveen RM. Steady-state free precession MRA of the renal arteries: breath-hold and navigator-gated techniques vs. CE-MRA. J Magn Reson Imaging. 2007;26:966–973. 40. Glockner JF, Takahashi N, Kawashima A, et al. Non-contrast renal artery MRA using an inflow inversion recovery steady state free precession technique (Inhance): comparison with 3D contrastenhanced MRA. J Magn Reson Imaging. 2010;31:1411–1418. 41. Jun-ling X, Da-Peng S, Yong-Li L, Ji-liang Z, Shao-cheng Z, Hao S. Non-enhanced MR angiography of renal artery using inflowsensitive inversion recovery pulse sequence: A prospective comparison with enhanced CT angiography. Eur J Radiol. 80:e57–63. 42. Schoenberg SO, Rieger J, Weber CH, et al. High-spatial-resolution MR angiography of renal arteries with integrated parallel acquisitions: comparison with digital subtraction angiography and US. Radiology. 2005;235:687–698. 43. van Assen HC, Vasbinder GB, Stoel BC, Putter H, van Engelshoven JM, Reiber JH. Quantitative assessment of the morphology of renal arteries from X-ray images: quantitative vascular analysis. Invest Radiol. 2004;39:365–373. 44. Marckmann P. Nephrogenic systemic fibrosis: epidemiology update. Curr Opin Nephrol Hypertens. 2008;17:315–319. 45. Cowper SE, Rabach M, Girardi M. Clinical and histological findings in nephrogenic systemic fibrosis. Eur J Radiol. 2008;66: 191–199. 46. Idee JM, Port M, Medina C, et al. Possible involvement of gadolinium chelates in the pathophysiology of nephrogenic systemic fibrosis: A critical review. Toxicology. 2008;248:77–88. 47. Thomsen HS, Marckmann P. MRI contrast media are used to improve visualization of abnormal structures or lesions in various parts of the body. Introduction. Eur J Radiol. 2008;66:153–159. 48. European Medicines Agency. European Medicines Agency makes recommendations to minimise risk of nephrogenic systemic fibrosis with gadolinium-containing contrast agents. http://www.ema. europa.eu/docs/en_GB/document_library/Press_release/2009/11/ WC500015569.pdf. Accessed September 20, 2010. 49. United States Food and Drug Administration. FDA Drug Safety Communication: New warnings for using gadolinium-based contrast agents in patients with kidney dysfunction. http://www.fda.gov/ drugs/drugsafety/ucm223966.htm. Accessed September 20, 2010. 50. Schoenberg SO, Just A, Bock M, Knopp MV, Persson PB, Kirchheim HR. Noninvasive analysis of renal artery blood flow dynamics with MR cine phase-contrast flow measurements. Am J Physiol. 1997;272:H2477–2484.

21

Renal Vascular Diseases

51. Montet X, Ivancevic MK, Belenger J, et al. Noninvasive measurement of absolute renal perfusion by contrast medium-enhanced magnetic resonance imaging. Invest Radiol. 2003;38:584–592. 52. Schoenberg SO, Aumann S, Just A, et al. Quantification of renal perfusion abnormalities using an intravascular contrast agent (part 2): results in animals and humans with renal artery stenosis. Magn Reson Med. 2003;49:288–298. 53. Sourbron SP, Michaely HJ, Reiser MF, Schoenberg SO. MRImeasurement of perfusion and glomerular filtration in the human kidney with a separable compartment model. Invest Radiol. 2008; 43:40–48. 54. Vallee JP, Lazeyras F, Khan HG, Terrier F. Absolute renal blood flow quantification by dynamic MRI and Gd-DTPA. Eur Radiol. 2000;10:1245–1252. 55. Boss A, Martirosian P, Graf H, Claussen CD, Schlemmer HP, Schick F. High resolution MR perfusion imaging of the kidneys at 3 Tesla without administration of contrast media. Rofo. 2005;177:1625–1630. 56. Martirosian P, Klose U, Mader I, Schick F. FAIR true-FISP perfusion imaging of the kidneys. Magn Reson Med. 2004;51:353–361. 57. Lee VS, Rusinek H, Noz ME, Lee P, Raghavan M, Kramer EL. Dynamic three-dimensional MR renography for the measurement of single kidney function: initial experience. Radiology. 2003;227: 289–294. 58. Teh HS, Ang ES, Wong WC, et al. MR renography using a dynamic gradient-echo sequence and low-dose gadopentetate dimeglumine as an alternative to radionuclide renography. AJR Am J Roentgenol. 2003;181:441–450. 59. Baltes C, Kozerke S, Hansen MS, Pruessmann KP, Tsao J, Boesiger P. Accelerating cine phase-contrast flow measurements using k-t BLAST and k-t SENSE. Magn Reson Med. 2005;54:1430–1438. 60. Bock M, Schoenberg SO, Schad LR, Knopp MV, Essig M, van Kaick G. Interleaved gradient echo planar (IGEPI) and phase contrast CINE-PC flow measurements in the renal artery. J Magn Reson Imaging. 1998;8:889–895. 61. Michaely HJ, Schoenberg SO, Ittrich C, Dikow R, Bock M, Guenther M. Renal Disease: Value of Functional Magnetic Resonance Imaging With Flow and Perfusion Measurements. Invest Radiol. 2004;39:698–705. 62. Pettigrew RI, Avruch L, Dannels W, Coumans J, Bernardino ME. Fast-field-echo MR imaging with Gd-DTPA: physiologic evaluation of the kidney and liver. Radiology. 1986;160:561–563. 63. Michaely HJ, Sourbron SP, Buettner C, Lodemann KP, Reiser MF, Schoenberg SO. Temporal Constraints in Renal Perfusion Imaging With a 2-Compartment Model. Invest Radiol. 2008;43:120–128. 64. Michaely HJ, Schoenberg SO, Oesingmann N, et al. Renal artery stenosis: functional assessment with dynamic MR perfusion measurements – feasibility study. Radiology. 2006;238:586–596. 65. Gandy SJ, Sudarshan TA, Sheppard DG, Allan LC, McLeay TB, Houston JG. Dynamic MRI contrast enhancement of renal cortex: a functional assessment of renovascular disease in patients with renal artery stenosis. J Magn Reson Imaging. 2003;18:461–466. 66. Song T, Lee VS, Rusinek H, Wong S, Laine AF. Integrated four dimensional registration and segmentation of dynamic renal MR images. Med Image Comput Comput Assist Interv. 2006;9:758–765. 67. Huang AJ, Lee VS, Rusinek H. Functional renal MR imaging. Magn Reson Imaging Clin N Am. 2004;12:469–486, vi. 68. Carvlin MJ, Arger PH, Kundel HL, et al. Use of Gd-DTPA and fast gradient-echo and spin-echo MR imaging to demonstrate renal function in the rabbit. Radiology. 1989;170:705–711. 69. Grobner T. Gadolinium – a specific trigger for the development of nephrogenic fibrosing dermopathy and nephrogenic systemic fibrosis? Nephrol Dial Transplant. 2006;21:1104–1108.

295 70. Lee VS, Rusinek H, Johnson G, Rofsky NM, Krinsky GA, Weinreb JC. MR renography with low-dose gadopentetate dimeglumine: feasibility. Radiology. 2001;221:371–379. 71. Attenberger UI, Sourbron SP, Notohamiprodjo M, et al. MR-based semi-automated quantification of renal functional parameters with a two-compartment model – an interobserver analysis. Eur J Radiol. 2008;65:59–65. 72. Song T, Lee VS, Rusinek H, Kaur M, Laine AF. Automatic 4-D registration in dynamic MR renography based on over-complete dyadic wavelet and Fourier transforms. Med Image Comput Comput Assist Interv. 2005;8:205–213. 73. Rusinek H, Boykov Y, Kaur M, et al. Performance of an automated segmentation algorithm for 3D MR renography. Magn Reson Med. 2007;57:1159–1167. 74. Annet L, Hermoye L, Peeters F, Jamar F, Dehoux JP, Van Beers BE. Glomerular filtration rate: assessment with dynamic contrastenhanced MRI and a cortical-compartment model in the rabbit kidney. J Magn Reson Imaging. 2004;20:843–849. 75. Lee VS, Rusinek H, Bokacheva L, et al. Renal function measurements from MR renography and a simplified multicompartmental model. Am J Physiol Renal Physiol. 2007;292:F1548-1559. 76. Martin DR, Sharma P, Salman K, et al. Individual kidney blood flow measured with contrast-enhanced first-pass perfusion MR imaging. Radiology. 2008;246:241–248. 77. Michoux N, Montet X, Pechere A, et al. Parametric and quantitative analysis of MR renographic curves for assessing the functional behaviour of the kidney. Eur J Radiol. 2005;54:124–135. 78. Schoenberg SO, Knopp MV, Bock M, et al. Renal artery stenosis: grading of hemodynamic changes with cine phase-contrast MR blood flow measurements. Radiology. 1997;203:45–53. 79. Schoenberg SO, Knopp MV, Londy F, et al. Morphologic and functional magnetic resonance imaging of renal artery stenosis: a multireader tricenter study. J Am Soc Nephrol. 2002;13:158–169. 80. Ritt M, Janka R, Schneider MP, et al. Measurement of kidney perfusion by magnetic resonance imaging: comparison of MRI with arterial spin labeling to para-aminohippuric acid plasma clearance in male subjects with metabolic syndrome. Nephrol Dial Transplant. 2010;25:1126–1133. 81. Michaely HJ, Herrmann KA, Nael K, Oesingmann N, Reiser MF, Schoenberg SO. Functional renal imaging: nonvascular renal disease. Abdom Imaging. 2007;32:1–16. 82. Bax L, Woittiez AJ, Kouwenberg HJ, et al. Stent placement in patients with atherosclerotic renal artery stenosis and impaired renal function: a randomized trial. Ann Intern Med. 2009;150: 840–848. 83. Wheatley K, Ives N, Gray R, et al. Revascularization versus medical therapy for renal-artery stenosis. N Engl J Med. 2009;361: 1953–1962. 84. Murphy TP, Cooper CJ, Dworkin LD, et al. The Cardiovascular Outcomes with Renal Atherosclerotic Lesions (CORAL) study: rationale and methods. J Vasc Interv Radiol. 2005;16:1295–1300. 85. Krum H, Schlaich M, Whitbourn R, et al. Catheter-based renal sympathetic denervation for resistant hypertension: a multicentre safety and proof-of-principle cohort study. Lancet. 2009;373:1275–1281. 86. Mounier-Vehier C, Haulon S, Devos P, et al. Renal atrophy outcome after revascularization in fibromuscular dysplasia disease. J Endovasc Ther. 2002;9:605–613. 87. Mounier-Vehier C, Haulon S, Lions C, et al. Renal atrophy in atherosclerotic renovascular disease: gradual changes 6 months after successful angioplasty. J Endovasc Ther. 2002;9:863–872. 88. Michaely HJ, Dietrich O, Reiser MF, Schoenberg SO. Neue konzepte für die funktionelle MRT bei nierenerkrankungen. Der Nephrologe. 2006;1:40–49.

wwwwwwwwwwww

MRA: Upper Extremity and Hand Vessels Ruth P. Lim and Vivian S. Lee

Introduction

Vascular Anatomy

Catheter angiography is the acknowledged clinical gold standard for upper extremity arterial assessment. However, it is invasive and can be particularly challenging in the upper extremity, as iodinated contrast injection can initiate vasospasm and pain, particularly in young adults. MR evaluation of the upper extremity and hand vessels can provide a noninvasive comprehensive assessment, particularly since the advent of gadolinium-enhanced MRA (Gd MRA) in the early 1990s [1]. Indications for upper extremity MR angiography or venography are varied, including atherosclerosis, trauma, thromboembolic phenomena, and vasculitides. Benefits of MRA include lack of ionizing radiation, no requirement for iodinated contrast, and ability to acquire functional information including flow direction and velocity. Advances in magnet and coil technology have enabled continued improvements in spatial and/or temporal resolution. For successful imaging, a clear understanding of vascular anatomy, patient preparation, imaging protocols, and potential pathology is required. Some potential challenges that are faced when evaluating the arm are small caliber distal vessels, anatomic variants, relatively slow blood flow, and short arteriovenous transit times. This chapter begins with a review of vascular anatomy from the axilla to the fingertips. Scanning technique is discussed, including patient positioning and appropriate coil selection. Next, basic sequences used in MRA and MRV are reviewed in the context of upper extremity imaging, describing both contrast-enhanced and noncontrast-enhanced angiographic techniques, and strategies to optimize image quality. Pitfalls in image acquisition and interpretation are reviewed. Clinical indications for upper extremity and MRA and MRV are presented, and finally, areas under exploration are discussed.

Arterial Anatomy [2]

R.P. Lim, MBBS, MMed, FRANZCR () • V.S. Lee, MD, PhD, MBA Department of Radiology, New York University Langone Medical Center, New York, NY, USA e-mail: [email protected]

22

The axillary artery supplies the upper extremity, arising as the continuation of the subclavian artery at the lateral border of the first rib. The axillary artery gives off body wall and anterior and posterior circumflex humeral arteries, and then continues as the brachial artery at the inferior border of teres major. Branches in the arm include the profunda brachii artery, equivalent to the profunda femoris artery in the leg, continuous with the radial collateral artery, and ulnar collateral arteries that anastomose around the elbow with recurrent collateral vessels from the forearm. The brachial artery terminates by bifurcating into radial and ulnar arteries at the level of the neck of the radius. In the forearm, the ulnar artery is generally a larger caliber vessel than the radial artery, and gives rise to the common interosseous artery, which in turn divides into anterior and posterior interosseous arteries that supply the flexor and extensor compartments of the forearm. The ulnar and radial arteries give rise to palmar and dorsal carpal branches that anastomose around the wrist. Hand arterial anatomy can be arbitrarily divided by radial and ulnar artery supply. The radial artery typically runs volar to dorsal within the anatomical snuffbox at the level of the trapezium. It gives rise to the arteria radialis indicis, supplying the radial side of the index finger, and the princeps pollicis artery, which divides into two palmar digital branches to supply the thumb. The radial artery then terminates in the deep palmar arch where it anastomoses with the deep branch of the ulnar artery. The deep palmar arch gives rise to three palmar metacarpal arteries that anastomose with common palmar digital arteries from the superficial palmar arch. The ulnar artery continues in the hand as the superficial palmar arch, located approximately 1 cm distal to the deep palmar arch, at the same level as the distal border of the outstretched thumb. This is often incomplete, but if complete, anastomoses

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_22, © Springer Science+Business Media, LLC 2012

297

298

R.P. Lim and V.S. Lee

Fig. 22.1 Line drawing of the arterial anatomy of the upper limb, demonstrating major vessels of the arm and forearm (left), and hand arterial anatomy in detail (right). Arteries in the hand arising from the radial

artery and deep arch are depicted in burgundy, and those arising from the ulnar artery and superficial arch are shown in red

with the superficial palmar branch of the radial artery. Three common palmar digital arteries arise from the superficial arch, and one proper palmar digital artery that supplies the ulnar side of the little finger. The common palmar digital arteries subsequently divide into proper palmar digital arteries to the ulnar side of the index finger, the middle and ring fingers, and the radial side of the little finger. Conventional arterial anatomy of the upper extremity is depicted in Fig. 22.1. Arterial anatomic variants are not uncommon [3, 4]. These include a brachioradial artery, referring to a high origin of the radial artery from the axillary or proximal to mid brachial artery, found in up to 13.8% in one large cadaver series, and a much less common brachioulnar artery. Variants restricted to the forearm vessels are relatively rare, and include absence or duplications of the radial or ulnar arteries. Variations in the hand are very common, particularly of the superficial palmar arch, where nine variants have been described. A complete arch, where the superficial arch provides supply to all five digits, is seen in over two-thirds of

subjects. Less commonly, arterial supply to one or more digits is exclusively from the radial artery, or from a persistent median artery, located between the ulnar and radial arteries. The interested reader is referred to the work of Coleman and Anson [4].

Venous Anatomy [2] Venous drainage from the upper limb can be divided into superficial and deep venous systems, again paralleling lower limb vascular anatomy. The deep system predominantly drains the arm and forearm, and the deep veins usually run as dual venae comitantes with the major arteries. The superficial venous system provides most of the drainage of the hand in addition to the subcutaneous tissue of the upper limb. It begins distally as the palmar digital veins, draining into a dorsal venous plexus. In the forearm, the venous plexus drains into cephalic vein laterally and the basilic vein medially,

22

MRA: Upper Extremity and Hand Vessels

Fig. 22.2 Line drawing of the superficial venous anatomy of the upper limb. Deep veins (not shown) follow the arteries, generally as paired venae comitantes. The cephalic vein is the main superficial vein laterally, and the basilic vein provides superficial drainage medially. Anatomic variations in superficial venous drainage are common, particularly in the antecubital region

and variably, median forearm and median cubital veins runs between the two along the volar forearm. While the cephalic vein remains superficial until its termination in the axillary vein beyond the clavipectoral fascia, the basilic vein passes through the deep fascia in the mid arm, becoming the axillary vein at the lower margin of teres major. Superficial venous drainage is summarized in Fig. 22.2.

Scanning Patient Positioning Patient positioning and comfort are integral components of achieving high-quality MR images. For above-elbow imaging, patients can be imaged supine. Below the elbow, in order to minimize wrap artifact from adjacent body structures, it is desirable to isolate the arm or arms if a bilateral examination is required. If feasible for the patient, prone positioning with one or both arms outstretched above the head (Fig. 22.3) will minimize wrap [5]. In order to maintain patient comfort, the arms can be raised to shoulder level by propping them up with supporting padding. In elderly patients or patients with limited arm extension, a less rigorous alternative is to remain

299

in a supine position, with arms by their sides or resting in neutral position on the thigh. In either position, the arm of interest should be centered as close to the bore of the magnet as possible, where the magnetic field is most homogeneous. This might involve positioning the patient’s torso laterally within the bore when feasible. A final alternative, for smaller patients unable to lie prone is to lie in a lateral decubitus position with the arm above the head. One other consideration with positioning is to minimize rotation of the upper extremity, to minimize slice or partition number, and also to simplify image interpretation. For example, the forearm should ideally be positioned in full pronation or supination, and for the hand, digits should be extended and immobilized as much as possible. This may be achieved by taping the hand to a rigid support, such as an arm board as is used with peripheral intravenous cannulae. The patient should be consulted and made comfortable prior to commencement of scanning, which may include additional supportive padding, blankets, and oxygen for dyspnea. Keeping the patient warm is important to prevent vasoconstriction, particularly if the digital arteries are of interest. Patients should be warned against motion, particularly with respect to the target area.

Coil Selection Coil selection will depend on which area is of clinical interest and the patient position selected. Surface receive coils are desirable in order to maximize signal to noise ratio, as there is no requirement for signal reception from deep tissues with upper extremity imaging. For the hand, there are commercially available multichannel hand coils, however knee, head, or a surface-phased array coil placed over or wrapped around the hand(s) can be substituted in practice and yield diagnostic quality results. If large field of view coverage is desired, for example, from the axillary level to the fingertips, overlapping surface-phased array coils as are used for body and cardiac applications can be positioned over the area of interest and combined with spinal coils and the inherent body coil of the magnet as required for signal reception.

Key Sequences Localizer Bright blood localizer images are desirable, for example, balanced steady-state free precession (SSFP) images, for adequate image planning. More than one set of multiplanar localizing images may be required for upper extremity imaging, because of variability of patient positioning within the magnet and desired coverage.

300

R.P. Lim and V.S. Lee

Fig. 22.3 Examples of patient positioning and coils. (a) Patient lying prone with arm extended above head and hand imaged within a knee coil. Dedicated hand, elbow, or head coils may be substituted depending on availability and patient hand size. Smaller patients can also lie in a lateral decubitus position with contralateral arm down, if prone positioning is not possible. (b) Patient lying prone with both arms extended above the head, using a 6-element surface-phased array coil, helpful for

bilateral forearm and hand imaging, for example, in suspected emboli or vasculitis. (c) Patient lying supine with arm of interest placed as close to the center of the bore as possible using two 6-element surface phased array coils. This is useful for assessment of the arm, but can also be employed for forearm and hand assessment in frail patients where prone imaging is not possible

2D Fast Spin Echo Imaging

surrounding structures without the distraction of bright blood signal. T2-weighted FSE imaging is therefore a useful sequence when it is important to evaluate the vessel wall, as in vasculitis or atherosclerosis. Use of nonselective and selective 180° inversion prepulses is often used to ensure that blood is nulled in the slice of interest. This technique is often used in cardiac imaging where imaging during diastole is desirable to minimize cardiac motion but slow blood flow

Since spin echo imaging involves an initial 90° excitation pulse, followed by one (spin echo) or more [fast spin echo (FSE)] 180° refocusing pulses, blood that has moved through a slice will appear black, as it does not experience both of these radiofrequency pulses. This effect can be exploited in vascular imaging to enable assessment of the vessel wall and

22

MRA: Upper Extremity and Hand Vessels

makes blood nulling more difficult. Nulling of blood signal, or black-blood imaging, is dependent on blood flow exceeding a threshold velocity proportional to slice thickness and inversely proportional to half the echo time [6]. Blood moving slower than this velocity will not be nulled, as it will be exposed to both RF pulses, and therefore will appear bright. This phenomenon can be used to differentiate highflow from slow-flow vascular malformations or lymphatic malformations, as high-flow vascular malformations appear dark on T2-weighted FSE imaging, unlike the other two pathologies. However, a potential pitfall is that blood moving within the imaging plane will similarly appear bright, and this can lead to incorrect diagnosis of slow flow. As such, bright blood imaging is preferable for luminal assessment. As vessels are relatively small caliber in the upper extremity, multishot FSE is preferable to single-shot FSE imaging, providing higher signal to noise ratio which can be used to maximize spatial resolution. Fat suppression is desirable to increase conspicuity of pathologic findings.

Noncontrast-Enhanced MRA Sequences: Phase-Contrast and Time-of-Flight MRA Although Gd MRA has largely superseded traditional noncontrast-enhanced bright blood MRA techniques for the upper extremity, they can be considered when gadolinium is contra-indicated and can also provide limited functional information. Phase-contrast MRA is rarely used in clinical practice because of the inherent requirement for multiple gradients and because extremity MRA usually demands a large field of view, both of which contribute to impractical acquisition times. Time-of-flight MRA (TOF MRA) is briefly discussed in the context of the upper extremity. TOF MRA for upper extremity imaging should be considered when only relatively small coverage is desired, as the need to acquire images perpendicular to flow directly impacts imaging times. When used to image the hand, 2D TOF MRA is preferred, due to slow arterial flow, but oblique rather than perpendicular positioning will enable visualization of all vessels, as the palmar arches are perpendicular in orientation to the radial, ulnar, and digital vessels [7]. Strategies to maximize flow-related enhancement include: (a) lengthening repetition time, at the cost of increased imaging time and poorer background suppression; (b) using a lower flip angle, which worsens background suppression; (c) minimizing voxel size and slice thickness to combat intravoxel dephasing and through-plane saturation effects, at the cost of decreased SNR and longer acquisition; (d) minimizing TE; and (e) use of flow compensation gradients, which also minimize intravoxel dephasing but increase achievable echo times. These strategies can equally be applied to MR venography, where

301

flow is also slow. It can clearly be seen that the potential benefits and disadvantages must be carefully balanced and tailored to the target vascular bed. Newly described noncontrast techniques under development are discussed at the end of this chapter.

Gadolinium-Enhanced MRA Unlike the flow-related techniques described above, intravascular contrast is generated with a gadolinium chelate, an exogenous contrast agent that shortens the T1 relaxivity of blood when injected, thereby producing “bright blood” images on the first pass of the contrast agent. Crucial components to consider when performing Gd MRA include contrast dosage, timing and rate, and tailoring sequence parameters for desired spatial resolution and acquisition times.

Contrast Dosage Contrast dosage is weight-based, and single to double dose (0.1–0.2 mmol/kg) for standard extracellular gadolinium chelates is suitable for upper extremity MRA. If the entire upper extremity is desired, two-station MRA can be performed, in which case a two injection approach is generally preferable to a bolus chase approach, as this allows for differences in positioning of the arm versus the arm and forearm. Using the two injection approach, it is advisable to complete one station in its entirety before acquiring the “precontrast” mask images for the second station to allow for more accurate image subtraction, and elimination of venous contamination and bright background signal from the first injection. Using a smaller first and larger second contrast dose, for example, 0.08 mmol/kg followed by 0.12 mmol/kg, will minimize the effect of the first injection when reviewing source images from the second station. Gd MRA Timing Accurate timing of the postcontrast acquisition is essential for pure arterial phase imaging. The digital arteries are relatively slow to fill, and sufficient delay between injection and acquisition is required to ensure their opacification. Conversely, too delayed an acquisition will result in venous contamination, as arteriovenous transit times decrease moving proximal to distal within the upper extremity. For forearm and hand MRA, a blood-pressure cuff can be applied proximally and inflated to subsystolic pressure to retard venous filling. Appropriate timing can be approached a number of ways. An empiric time delay of 15 s has been reported to yield diagnostic results for the hand [8]. However, with differences in individuals’ circulation times, contrast dosages and injection rates, a bolus tracking or test bolus

302

R.P. Lim and V.S. Lee

approach is preferable. This can be centered on the brachial artery at the antecubital level. Using the test bolus approach, an accurate scan delay can be calculated by determining the

Scan delay = Time to peak enhancement +

For sustained arterial enhancement during image acquisition, it is desirable for duration of enhancement to approximate imaging time. This can be achieved with slower injection rates, of the order of 2 ml/s, followed by a saline chaser at the same injection rate. A final alternative is to use a time-resolved approach, with multiple measures. This obviates the need for estimation of appropriate timing, provided image acquisition commences before contrast has reached the imaging field of view. With sufficient temporal resolution, time-resolved imaging has the additional benefit of hemodynamic information, including speed of vascular enhancement.

Optimizing Gd MRA Parameters Sequence parameters need to be tailored to the upper extremity. There are competing demands for high spatial resolution, as digital arteries are of the order of 1 mm in diameter, and for relatively short acquisitions to minimize venous contamination. Ideally, voxel sizes in the 1.5 mm range for the arm and forearm, and 1 mm or less for the hand, are desirable. However, for the hand in particular, spatial resolution may be limited by signal to noise ratio, and time saving techniques described below will also affect SNR. Use of high-relaxivity contrast agents, such as gadobenate dimeglumine or gadofosvoset, and imaging at 3 T can mitigate this problem, to be discussed at the end of this chapter. One strategy that can shorten acquisitions is parallel imaging when multiple receiver coils are available in the phase encoding direction. Receiver coil data is used to provide spatial localization information, decreasing the number of required phase encoding lines, and facilitating large field of view rapid coverage. Image reconstruction techniques for parallel imaging include sensitivity encoding (SENSE) [10], simultaneous acquisition of spatial harmonics (SMASH) [11], and generalized autocalibrating partially parallel acquisitions (GRAPPA) [12]. With 3D Gd MRA, the slice or partition direction is essentially a second-phase encoding direction. The frequency encoding direction should be placed along the longest plane of the desired field of view, as this will not affect imaging time. For upper extremity MRA, an acquisition that is coronal to the outstretched hand and parallel to the plane of the radius and ulna will ensure the two shorter dimensions (transaxial to the upper extremity), fall along the slice and

time at which the midpoint of arterial enhancement will coincide with time at which the center of k-space (determining image contrast) is acquired [9]: 1 Injection time - Time tok - space center. 2

phase encode directions. Ideally, the slice direction should be along the shortest dimension, to maximize the use of rectangular field of view in the phase encode direction, at no cost to spatial resolution. Partial Fourier and/or zero interpolation can be employed in phase encode or slice directions, although this will decrease true spatial resolution of the acquisition. Centric encoding, where central lines of k-space are acquired early, before peripheral lines, can be employed to good effect by collecting central lines of k-space before contrast reaches the venous system. This allows more time for peripheral k-space to be filled after image contrast data has been collected. In this way, relatively long acquisition times of 30–40 s can provide high-resolution arterial phase hand MR angiograms [8, 13].

Time-Resolved Gd MRA Echo-sharing or keyhole imaging strategies can be used to further decrease imaging times and provide some hemodynamic information. The basic premise is that with multiple measures, the center of k-space is fully sampled with every acquisition; however, the periphery of k-space is undersampled, with uncollected data interpolated between consecutive measures. Various trajectories have been described in order to undersample k-space, including time-resolved imaging of contrast kinetics (TRICKS) [14] and time-resolved imaging with stochastic trajectories (TWIST), but all allow sub-10 s temporal resolution, which can be particularly important in the assessment of vascular malformations. As image acquisitions are faster, no more than single dose of contrast is required, and early arterial through to venous phase images can be acquired without a test bolus. A potential disadvantage of this approach is that the acquisition of multiple measures can generate a large amount of data, and reconstruction time can slow down the examination overall [5]. Also, Gibbs ringing artifact may be problematic with such techniques, where there are sharp transitions in MR signal between center and periphery of k-space acquisition, exacerbated by high contrast injection rates and aggressive k-space undersampling. As there is inevitably some loss of signal to noise ratio and subsequently spatial resolution with time-resolved MRA, it is most effective when hemodynamic information is desired, for example, in the evaluation of vascular malformations or fistulas.

22

MRA: Upper Extremity and Hand Vessels

303

Table 22.1 Overview of key sequences used in clinical vascular imaging Sequence 2D FSE

Description Blood moving through the imaging slice not experiencing both 90° (excitation) and 180° (refocusing) pulses or double inversion recovery prepulses null blood signal for black-blood images

TOF MRA

Multiple excitation pulses lead to saturation of stationary tissue, while fresh inflowing blood appears relatively bright

PC MRA

Dephasing and rephasing gradients cause phase shifts in moving protons in blood, resulting in a bright blood angiographic image Injection of gadolinium chelate causes rapid T1 relaxation of arterial blood on first-pass T1-weighted gradient echo imaging

Gd MRA

Time-resolved Gd MRA

Multiple Gd MRA measures are acquired utilizing image acceleration techniques including parallel imaging and keyhole imaging

Strengths Relatively rapid if a single shot (SSFSE) approach is employed Enables evaluation of vessel wall and surrounding structures Helpful in evaluating vascular malformations; flow voids will be seen with high-flow malformations Using saturation prepulses, direction of blood flow can be determined Bright blood luminal assessment without exogenous contrast Bright blood luminal assessment without exogenous contrast Excellent background suppression Independent of flow Rapid 3D volumetric coverage with good spatial resolution Multiple phase dynamic images can be obtained, including venous phase MRV Timing run not required Faster acquisitions allow for lower doses of contrast Provides hemodynamic information

Gadolinium-Enhanced MRV Although noncontrast methods of MR venography are available, particularly TOF MRA and, more recently, balanced SSFP imaging, gadolinium-enhanced 3D MRV provides a large field of view, rapid assessment. Gadolinium-enhanced MRV (Gd MRV) can be performed using a direct approach, where dilute gadolinium contrast is injected into the venous system of the extremity of interest distal to the area of concern. More commonly, indirect Gd MRV is performed, where image acquisition occurs after intravenously injected gadolinium has recirculated to opacify the entire venous system, usually approximately 3 min following injection. With the indirect approach, there are no restrictions to the intravenous access site, and any part of the venous system can be rapidly imaged. Image subtraction can highlight the venous system, particularly if arterial phase imaging is performed first. Similar to Gd MRA, a T1-weighted spoiled gradient echo sequence is used. Recirculation within the venous system necessarily dilutes the contrast, with less T1 relaxation of venous blood compared with a first-pass arterial examination. For this reason, double dose (0.2 mmol/kg) of contrast is also desirable for Gd MRV, provided renal function is normal.

Weaknesses Slow or in-plane blood flow may not be nulled, which may be incorrectly interpreted as intravascular thrombus 2D approach limits assessment to the acquisition plane

Must be acquired perpendicular to blood flow, impacting acquisition time Slow or in-plane flow causes saturation of blood which may lead to stenosis overestimation Relatively long TE increases intravoxel dephasing and susceptibility artifact Poor coverage: multidirectional gradients increase imaging time

Low risk of adverse reaction to gadolinium Risk of Nephrogenic Systemic Fibrosis in renal failure patients Susceptibility artifact can be problematic when metallic implants including vascular stents are present, particularly at higher field strength Decreasing acquisition may compromise spatial resolution Multiple measures and acceleration techniques require longer reconstruction times Large amount of data generated

Use of frequency selective fat saturation, and a lower flip angle can improve soft tissue contrast, valuable for Gd MRV. The strengths and potential disadvantages of T2-weighted dark blood FSE images, 2D TOF MRA, and Gd MRA are presented in Table 22.1, followed by suggested imaging parameters for key sequences in Table 22.2.

Pitfalls There are a number of potential pitfalls that may be encountered in upper extremity MRA. These include inaccurate timing, motion artifact, flow-related artifacts, pseudostenosis, vascular mimics, and nonvisualization of extraluminal pathology [15].

Inaccurate Timing As discussed above, accurate timing for upper extremity MRA can be challenging. Distal arteries will not be opacified with too early an acquisition, and Maki artifact may also arise, where larger structures such as the center of large

TR (ms) 4,000

20

3.3

3.3

Sequence FST2w FSE

2D TOF MRA (for neck)

Gd MRA

FST1wGRE

1.2

1.3

5

TE (ms) 70

12

25

40

Flip angle (º) 180

20 s

20–30 s Breath hold (60 mL/min/1.73 m2, a total of 0.20 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer Healthcare Pharmaceuticals, Wayne, NJ) is administered. For patients with eGFR 30–60 mL/min/1.73 m2, a total of 0.15 mmol/kg gadobenate dimeglumine (Multihance, Bracco Diagnostics, Princeton, NJ). Gadobenate dimeglumine has a highly stable cyclic ring structure greatly reducing the amount of free gadolinium at equilibrium, as discussed subsequently. The gadolinium dose is divided approximately 1:3 between the time-resolved calf acquisition and the bolus-chase LE-MRA runoff. Noncontrast LE-MRA sequences are performed in patients with eGFR > Feces Renal Renal

complex anatomy of a tortuous aorto-iliac system, tortuous collateral vessels, and the horizontally oriented proximal anterior tibial artery could result in artifactual signal loss mimicking disease. Time-of-flight imaging has limited sensitivity to slow flow due to challenges differentiating an occluded segment from proton saturation secondary to slower than expected flow. Venous inflow is suppressed through the application of a saturation pulse below the imaging plane. Three-dimensional slab thickness is limited by arterial saturation deep into the imaging volume. Iterations of twodimensional imaging start at the exit of arterial blood from the field of view, moving opposite to the direction of arterial blood flow. This prevents saturation band associated signal loss in the arteries while increasing venous signal suppression. Specific iterations have been developed to overcome the deep imaging volume arterial blood saturation, including multiple overlapping thin slab acquisition (MOTSA) and titled optimized nonsaturated excitation (TONE). MOTSA covers a large z-axis dimension by splitting the volume into several thinner slabs. TONE applies larger flip angles deep into a slab to increase the signal from saturated blood protons. Flow-related dephasing is an important consideration in time-of-flight imaging. Both intravoxel and intervoxel dephasing can occur. The accumulated phase shift is dependent on the blood velocity, the amplitude of the applied gradient, and the square of the gradient duration. Gradient moment nulling has been applied to reduce flow-related dephasing. Rather than a single bipolar gradient, gradient moment nulling utilizes a pair of bipolar gradients of opposite polarity. Stationary tissues experience no net phase accumulation under both scenarios; however, moving protons accumulate a net phase shift with a single bipolar gradient. The application of a pair of bipolar gradients compensates for constant flow velocities with no net phase accumulation. Gradient moment nulling cannot compensate for higherorder flow patterns with accelerating protons or turbulent flow. In the presence of higher-order flows, minimizing the TE reduces the phase shift accumulation by shortening the period of time between the applied gradients and the echo. Higher-order flows are more sensitive to longer TEs than

23

Lower Extremity Peripheral Arterial Disease

constant velocity flow patterns. In addition, local field inhomogeneities experienced by the moving protons cannot be corrected for gradient moment nulling. Shorter TEs reduce the amount of phase shift accumulated; however, as in the case of higher-order flows, cannot completely compensate and artifacts result. Time-of-flight imaging is poorly suited to image the entire z-axis field of view required in LE-MRA due to the excessive imaging times required for three-dimensional imaging and proton inflow saturation effects. Selective application of twodimensional time of flight may be useful in the assessment of flow directionality. However, an appropriate suspicion is required, as the traveling saturation band suppresses retrograde arterial flow. When combined with contrast-enhanced LE-MRA, the technique is able to readily differentiate antegrade and retrograde flow patterns, as vessel patency is already known. Two-dimensional time of flight has demonstrated accuracy in assessment of pedal vessel stenosis and occlusion [15], although it has largely been replaced by novel noncontrast techniques and contrast-enhanced methods for this application.

Phase-Contrast Angiography Phase-contrast MR angiography (PC-MRA) generates images of the vasculature based on the absolute net phase shift acquired by protons moving through a time-varying magnetic field. PC-MRA is performed by the application of both flow compensated and flow-encoding bipolar gradients. As discussed previously, gradient moment nulling utilizes paired bipolar gradients of opposite polarity to correct for phase shifts accumulated by protons moving in the direction of the applied gradient. By also acquiring a velocity-encoded, uncompensated gradient acquisition, data regarding the net phase shift of the protons is obtained. Both acquisitions experience phase shifts secondary to field inhomogeneities; however, only the velocity-encoded uncompensated acquisition also contains velocity-related phase shift information. Phase shifts accumulated by moving protons after correction for field inhomogeneities are linearly related to the proton velocity. Net phase shift is converted to proton velocity by the estimation of the peak velocity experienced in a region of tissue. For accurate application, the user must accurately prescribe a velocity corresponding to 180° of accumulated phase shift. Voxels with an average net phase accumulation in one gradient direction of 180° would be assigned the peak velocity; however, if any voxels experienced greater than 180 or −180° of phase shift, aliasing would occur. For example, at a velocityencoding setting of 150 cm/s, a phase shift of 210, corresponding to an actual velocity of 210/180 × 150 = 175 cm/s would generate an aliased velocity corresponding to a phase shift

325

of −150° or −150/180 × 150 = −125 cm/s. Hence appropriate choice of velocity encoding is necessary. Although problematic at flow quantification, small amounts of aliasing at phase-contrast angiography do not introduce significant errors secondary to the vector sum calculation for average voxel velocity. Voxel signal intensity is equal to the square root of the sum of the squares of the corresponding velocity in each encoded direction; hence aliased negative velocities close to the peak forward velocity in magnitude introduce much smaller errors than seen with single direction flow quantification. The choice of the velocity-encoding gradient is important to consider in the context of an ECG-gated or ungated acquisition, reflecting the variation between diastolic and systolic peak versus average flow, respectively. PC-MRA can be performed with both two-dimensional and three-dimensional acquisitions. Although acquisition of a PC-MRA sequence at peak systole results in greater signal to noise ratio, imaging times may be prohibitive with threedimensional techniques. Consequently, applications of PC-MRA to the lower extremities have relied on ECG-gated two-dimensional acquisitions in the coronal plane, adjusting the velocity-encoding gradient to the expected average velocity across the cardiac cycle. Velocity encoding is much lower to reflect the high-resistance vascular bed supplied by the lower extremity vasculature at rest. Steffens et al. reported their experience imaging from aortic bifurcation to the tibioperoneal trunk with two-dimensional ECG-gated PC-MRA in 115 patients with atherosclerosis [16]. The authors employed three-dimensional flow encoding with velocityencoding gradients of 30 and 20 cm/s in the pelvis and thighs, respectively. The imaging volume was divided into three stations with an acquisition time of 4–7 min per station; two averages were performed with completion of imaging in 30 min. The lesion analysis demonstrated a sensitivity and specificity of 95% and 90% with positive and negative predictive values of 90% and 96%, respectively. PC-MRA overcomes the limitations associated with flowrelated enhancement at time-of-flight imaging by generating signal intensity proportional to proton velocity. Technical improvements with undersampling techniques have resulted in shorter imaging times [17]. However, clinically available PC-MRA iterations have imaging times impractical for routine LE-MRA applications. A related technique, two-dimensional flow quantification, discussed in detail elsewhere in this textbook, has been described for selective assessment of stenosis grading in the lower extremities. Similar to Doppler sonography, the velocity increase at a stenosis can be calculated based on the corrected phase shift; using the modified Bernoulli equation, 4 × Vmax2, the pressure gradient across the lesion can be estimated. Mohajer et al. showed success grading lesion severity in the superficial femoral arteries by assessing the peak velocity at and measuring the delay in systolic peak velocity beyond lesions, using MRA as standard

326

of reference. Although accurate for assessing a problematic individual lesion, this technique is impractical for assessing serial lesions in patients with multifocal disease.

3D MRA with Balanced Steady-State Free Precession Bright-blood balanced steady-state free precession (bSSFP) imaging generates robust images of the soft tissues and vessels, with signal intensity dependent on the ratio of T2/T1 [18]. Signal intensity is independent of inflow. Both arteries and the veins demonstrate high signal intensity; consequently most angiographic adaptations utilize additional preparatory pulses to suppress venous signal. Similar to other noncontrast techniques, signal loss may in result in regions of flow acceleration. Short TR times combined with high flip angles generate high signal to noise ratios, making these techniques well suited to parallel imaging. One technique to suppress fat tissue at bSSFP is the Dixon method. This utilizes the frequency shift of protons associated with lipids relative to those in water. At 1.5 T, there is a frequency offset of 220 Hz, with lipid protons processing at a slightly lower rate. Acquiring an echo with the fat and water protons in phase generates increased signal from voxels containing both lipid and water; acquisitions at an effective TE with out-of-phase lipid and water protons generate low signal intensity voxels. Tuning the center frequency of the magnet can also produce in- and out-of-phase images. Complex addition of these two datasets can generate wateror fat-only images. bSSFP sequences are highly susceptible to field inhomogeneities, with consequent off-resonance artifacts. These become increasingly problematic with longer TRs. Application in LE-MRA is particularly challenging due to the complex surface anatomy of the lower extremities. Venous signal notwithstanding, however, diagnostic images of the arteries can be obtained with this technique.

Arterial Spin Labeling Arterial spin labeling (ASL) as the name implies relies on the inflow of labeled or tagged spins into the imaging volume. A nonselective inversion pulse suppresses signal from stationary tissues. Earlier versions of technique were implemented with segmented turbo fast low angle shot gradient echo sequences and required two separate acquisitions, one with a nonselective inversion pulse and the other with selective inversion of tissue upstream from the region of interest [19]. The positioning of the upstream tagged slab is determined by the desired delay time between tagging and imaging. Subtracting the two acquisitions yields an angiographic image with suppressed background tissues.

J.D. Collins and T. Scanlon

Subsequent ASL iterations combined a spatially nonselective inversion pulse with a spatially selective upstream inversion pulse into a single acquisition. The delay time between the application of the nonselective inversion pulse and the initiation of imaging allows selective suppression of different tissues based on T1 relaxation. Acquiring an untagged imaging volume for subtraction enables better suppression of background tissue. Further improvements on the technique have incorporated bSSFP and partial Fourier FSE sequences. Combining ASL with bSSFP yields bright-blood images with excellent venous suppression and high signal to noise ratio [20]. Partial Fourier FSE sequences (discussed in detail subsequently) can be used with ASL and benefit from less susceptibility artifacts secondary to the spin echo readout. ECG gating is required for partial Fourier FSE readout to avoid spin-dephasing artifacts from systolic acquisition. Although an eloquent noncontrast MRA technique, ASL has limited application for imaging the lower extremities. This is primarily secondary to long transit times in the extremities. The inflow time for a peripheral imaging volume can approach the blood T1, negating differences between the tagged blood and stationary tissues in the imaging volume. Applying smaller imaging volumes is impractical with consequent increases in imaging time. In addition, ASL does not detect retrograde arterial flows. The technique assumes antegrade brisk arterial flow to generate adequate signal. Patent vessels with slow or retrograde flow are not well assessed with this technique and may appear occluded.

Emerging Noncontrast Techniques Several novel noncontrast techniques for lower extremity angiography deserve mention. The goal of this section is to introduce the reader to the concepts underpinning generation of arterial signal and venous suppression in each, while highlighting challenges and known pitfalls. Of these ECG-gated 3D partial Fourier FSE and quiescent interval single shot (QISS) are available commercially. The clinical utility and accuracy of these techniques is the subject of ongoing investigation.

ECG-Gated 3D Partial Fourier FSE This technique makes use of systolic arterial flow voids on T2-weighted fast spin echo (FSE) images to generate arterial images [21]. Systolic-triggered acquisitions demonstrate bright signal in the veins. Diastolic-triggered acquisitions demonstrate bright signal in both arteries and veins. Subtracting the two acquisitions yields arterial only images. Each partition in the three-dimensional dataset is acquired with a single-shot acquisition, lengthening the TR to equal two to three heartbeats to allow sufficient time for recovery

23

Lower Extremity Peripheral Arterial Disease

of longitudinal relaxation. The application of partial fast Fourier TSE readout yields acceptable acquisition times for clinical imaging, with 3–6 min reported for each station. Optimization of the systolic and diastolic trigger times can be achieved by obtaining preparatory ECG-triggered images at different trigger delays to choose the frames with greatest systolic suppression and adequate diastolic signal. Also, a preparation scan varying the spoiling gradients in the frequency direction can facilitate choice of dephasing gradient amplitude. A challenge applying this technique to LE-MRA is unwanted T2 blurring, which occurs in the phase encoding directions. In LE-MRA applications, the frequency encoding direction is oriented along the length of the vessel to facilitate spoiling of the systolic acquisition arterial signal. Consequently, the phase encoding directions are oriented orthogonal to the vessels of interest; blurring in these directions reduce the sharpness of the vessel walls. Single-shot FSE readouts acquire data over a range of approximately 350 ms. Solutions to reduce T2 blurring include a dual shot acquisition with consequent doubling of image acquisition time. Alternatively, centric ordering with rectilinear k-space sampling has been reported to reduce vessel blurring through less phase spread in the phase encoding directions. Although a promising technique, clinical application is challenging in several scenarios. Patients with irregular arrhythmias and tachyarrhythmias are not well suited to this technique due to differences in the length of diastole, precluding adequate sampling of the diastolic acquisition. Applying a dual shot acquisition may be helpful, but prolonged imaging times increases the probability of macroscopic patient motion. In addition, patients with significant peripheral vascular disease may demonstrate limb-to-limb and in-station segmental differences in the optimum delay time for the systolic arterial flow void. Optimizing flow to visualize collateral vessels may exaggerate the severity and extent of disease in adjacent native arteries. Despite these limitations, several studies have demonstrated promising results in clinical patients with atherosclerosis, particularly when the analysis is limited to those patients with good quality studies [22].

ECG-Gated Flow-Sensitive Dephasing bSSFP ECG-gated flow-sensitive dephasing (FSD) bSSFP acquires systolic- and diastolic-triggered datasets with a bSSFP bright-blood readout. Similar in principle to ECG-gated partial Fourier FSE, this technique reduces systolic arterial signal by applying a weak gradient moment in the frequency encoding direction, dephasing flowing signal [23]. Fast, laminar arterial flow patterns and weak gradient moments facilitate suppression of arterial signal in systole, while maintaining venous signal. Constant venous diastolic flow and slower

327

arterial diastolic flow facilitates bright signal in both vessels for the diastolic acquisition. Subtracting the two datasets results in pure arterial images. Acquisition for a single station is approximately 3 min. This technique has primarily been applied for imaging the calves alone. An ECG-gated phase-contrast imaging preparatory scan facilitates appropriate selection of the systolic and diastolic trigger delays. Similarly, use of a scout sequence applying different gradient moment strengths in the systolic phase optimizes arterial signal suppression [24]. A challenge to the technique is bright signal at systole in both arteries and veins with bSSFP. In clinical practice, this technique encounters similar difficulties to the ECG-gated partial Fourier FSE sequence, described above. Advantages of FSD bSSFP are shorter overall imaging time, greater diastolic arterial signal intensity free from short TR-associated saturation effects, and interleaved acquisition of systolic and diastolic phases.

Quiescent Interval Single-Shot MRA Quiescent interval single-shot (QISS) MRA is a novel ECGgated noncontrast technique that employs a two-dimensional bSSFP pulse sequence [25]. QISS MRA relies on combination of stationary tissue and fat signal suppression, venous signal saturation, and bright-blood bSSFP gated for acquisition during diastole. The technique utilizes a slice-selective radiofrequency pulse after a user specified delay time from the R wave to suppress stationary tissues within the imaging slice. A saturation pulse applied inferior to the imaging slice suppresses antegrade venous signal. Following a delay termed the quiescent interval, a chemical shift-selective fat suppression pulse is applied with a subsequent radiofrequency pulse forcing in-plane protons into the steady state. A two-dimensional bSSFP pulse sequence is subsequently acquired, triggered in diastole. The sequence is performed as stacks of two-dimensional slices starting caudally. Each station consists of sixty 3 mm slices with 0.6 mm overlap and takes approximately 45–70 s to acquire. A total of 10–13 stations are needed to cover the entire lower extremity field of view with adequate overlap between stations. Thinner slices can be acquired through diseased segments to better assess vascular stenosis. The technique benefits from parallel imaging with a factor of 2 at 1.5 T. Spatial resolution is limited in the slice direction by the slice thickness of 3 mm; however, 1 × 1 mm resolution is achieved in-plane. QISS MRA performed well in a feasibility study of eight volunteers. Preliminary analysis in a patient cohort referred for contrast-enhanced LE-MRA demonstrated high accuracy for QISS MRA using contrast-enhanced LE-MRA as the reference standard [26]. QISS MRA demonstrated 95% sensitivity and 92% specificity for detecting a stenosis ³50% using bolus-chase gadolinium-enhanced MRA as the reference

328

standard. The negative predictive value of QISS MRA was 96%; the positive predictive value was 90%. QISS MRA image quality assessed by Likert scores was comparable to contrast-enhanced LE-MRA. QISS MRA has several advantages compared to other noncontrast ECG-gated angiographic techniques. The sequence requires little patient-specific adjustments to generate robust images of the arteries with excellent soft tissue and venous signal suppression. Independence from systolic phase imaging improves the reliability in patients with peripheral vascular disease. As with other noncontrast ECGgated techniques, QISS MRA image quality may be reduced in patients with irregular arrhythmias or tachyarrhythmias. Finally, the two-dimensional acquisition technique reduces the effects of macroscopic patient motion.

J.D. Collins and T. Scanlon

Plaque Imaging As performed clinically, LE-MRA imaging protocols are optimized to grade stenoses, but lack the soft tissue contrast and resolution to characterize plaque components. Dedicated acquisitions with small fields of view, surface coils, and spinecho T1, T2, and proton density weighted imaging are necessary to achieve the requisite contrast and spatial resolution for plaque characterization. Even with optimized acquisitions, plaque characterization is limited to macroscopic plaque components in the femoral through popliteal arteries. The appearance of specific plaque components has been validated through histological correlation studies of carotid plaque specimens. MRI is currently used to assess changes in plaque burden with treatment in ongoing clinical research. Due to long imaging times and uncertain clinical utility, plaque imaging at MRI remains a research tool.

LE-MRA at 3 T The increasing availability of 3 T scanners has fostered adaptations of 1.5 T LE-MRA sequences for the higher field strength system. Despite the benefits from imaging at 3 T, several pitfalls deserve mention. Higher parallel imaging factors are achievable secondary to increases in signal. Specific absorption rate limits can be problematic at 3 T when 180° refocusing pulses or higher flip angles are utilized. Contrast-enhanced LE-MRA benefits from higher signal to noise ratio. Intravascular gadolinium contrast agents easily overcome the prolongation of T1 relaxation at 3 T. Greater signal to noise ratio enables higher parallel imaging factors, translating into gains in both spatial and temporal resolution at a fixed acquisition time. The development of total imaging matrix technology enables robust continuous moving table LE-MRA with a single data acquisition covering up to 128 cm with isotropic resolution [14]. Several noncontrast techniques deserve specific discussion at 3 T. Increased susceptibility artifacts adversely impact steady-state sequences. At 3 T, venous signal is less than arterial signal on bSSFP sequences due to the effect of reduced oxygenation in venous blood. In addition, longer TRs generate further venous signal suppression on bSSFP sequences. Stafford et al., who applied the Dixon technique at 3 T using a three-dimensional bSSFP sequence and a TR of 3.4 ms for noncontrast LE-MRA, demonstrated good quality arterial images with less venous signal [27]. The investigators employed a 50% overlap between coronal 3D bSSFP acquisitions to overcome inhomogeneity artifacts. The reduction in imaging time afforded by parallel imaging is particularly useful in ECG-gated partial Fourier FSE, with improvement in T2 blurring described above at 1.5 T. Finally, ASL benefits from the longer T1 relaxation times at 3 T. Preliminary work with QISS MRA suggests that the increased signal to noise ratio improves image quality in the tibial vessels and lower extremity branch vessels compared to 1.5 T.

Imaging Processing Image postprocessing is an important trouble-shooting tool in the assessment of vessel pathology. Through more efficient display of three-dimensional data, postprocessing may reduce the amount of time required to review an imaging study. Both contrast-enhanced and noncontrast LE-MRA datasets are well suited to MPR postprocessing to better assess stenosis in the plane of a diseased vessel. MPR review is best performed on unsubtracted partition data. The higher signal to noise ratio and contrast to noise achieved with mask-subtracted contrast-enhanced datasets and noncontrast acquisitions with static tissue signal suppression are well suited to MIP algorithms. MIP postprocessing projects the voxel with the greatest signal intensity in a line through the entire dataset to generate an image in a given orientation. MIPs collapse an entire three-dimensional dataset into either a single frame or a smaller number of frames (sliding MIPS). The three-dimensional nature of the dataset can be partially preserved by generating a series of rotating images through small changes in the orientation of the dataset. A large dataset can be reviewed efficiently using a MIP algorithm. Artifacts that increase signal intensity of background tissues or cause misregistration between contrast-enhanced and mask datasets reduce the quality of MIP reconstructions. Reviewing postprocessed and subtracted datasets without considering the source data may lead to pitfalls in interpretation (Fig. 23.3); abnormalities on postprocessed images should be verified by reviewing the source data. Postprocessing greatly facilitates efficient review of timeresolved MRA datasets. As described above, time-resolved acquisitions generate multiple three-dimensional datasets with a single contrast injection. Reviewing a series of subtracted MIPs oriented in the coronal or sagittal planes organized by the time of acquisition enables efficient review of the

23

Lower Extremity Peripheral Arterial Disease

329

Atherosclerotic Disease

Fig. 23.3 MIP postprocessing artifacts. Gadolinium-enhanced LE-MRA performed in a 67-year-old male with buttock claudication. The patient has a right common iliac artery stent and a short segment nonflow limited dissection of the left common iliac artery. The thickslab MIP exaggerates the signal loss in the stent. The dissection flap is not well-seen secondary to bright contrast-enhanced blood on both sides of the intimal flap. Review of partition data clearly demonstrates flow through the stent and the intimal flap

entire time-resolved dataset. Correlation with partition data is recommended to confirm abnormal findings on MIP images.

Clinical Applications Contrast-enhanced LE-MRA is routinely performed to evaluate the vessels of the lower extremities. The lower extremity arterial system is divided by station into the inflow vessels (aorto-iliac), outflow vessels (femoral and popliteal arteries), runoff vessels (tibial), and pedal vessels (dorsalis pedis and medial plantar arteries).

Peripheral vascular disease is the third most common manifestation of atherosclerosis. The prevalence in the general population is estimated at 10%, increasing to 15% in individuals over the age of 70. Presenting symptoms are often indolent, with pain or muscle cramping with activity subsiding with rest. The effected muscle groups may indicate the level of arterial occlusive disease. Activity-associated claudication may progress to rest pain, nonhealing ulcers, and tissue gangrene. Risk factors include family history, longstanding diabetes mellitus, history of hypercholesterolemia, and smoking. Aorto-iliac occlusive disease manifests with intermittent buttock and thigh claudication. Impotence may also be a presenting symptom. Leriche’s syndrome is a constellation of symptoms and signs associated with infrarenal occlusion of the abdominal aorta (Fig. 23.4). Initially described in middle-aged males, patients complain of impotence and buttock claudication; on physical examination the common femoral arterial pulses are nonpalpable. Treatment is indicated for symptomatic claudication. The mainstay of therapy is surgical bypass, although endovascular treatment may successfully recanalize the native infrarenal aorta and iliac vessels in a minority of patients. External iliac and femoral arterial disease manifests with intermittent thigh claudication. Superficial femoral, popliteal, tibial, or peroneal artery disease presents with calf claudication (Fig. 23.5). Endovascular treatment with angioplasty and stenting is preferred in patients with suitable anatomy; surgical bypass may be performed in patients with distal targets. Acute occlusion may result from distal thromboembolization of plaque, as shown in Fig. 23.6. Without an upstream dissection, plaque donor site, or aneurysm with mural thrombus, a cardiac source should be considered and further assessment with echocardiography or cardiac MRI is indicated. Cardiac MRI can be performed at the same time as a thoracoabdominal angiogram. The presence of preexisting collateral vessels determines the acuity of intervention. Most patients with acute thromboembolism undergo operative thrombectomy with possible fasciotomy. Patients without limb threatening ischemia may be referred for catheter directed lysis. Atherosclerosis and aneurysm formation often coexists. The natural history of iliac artery aneurysms mirrors that of aneurysms in other locations. Iliac aneurysms are discovered incidentally in up to 65% of patients (Fig. 23.7). Symptoms are nonspecific in the remainder; the most common symptom is abdominal pain. The common iliac artery is considered aneurysmal if it is greater than 2.5 cm in diameter. Treatment is recommended if the aneurysm exceeds 3 cm in diameter. Aneurysm rupture is associated with a mortality of up to 80%. Other complications include thrombosis and thromboembolization. Treatment strategies depend on the aneurysm location and its association with an abdominal aor-

330

J.D. Collins and T. Scanlon

Fig. 23.4 Leriche’s Syndrome. 62-Year-old male presents with bilateral buttock claudication. (a) Pelvis, (b) thigh, and (c) calf MIPs from a subtracted contrast-enhanced runoff demonstrate infrarenal aortic

occlusion, with reconstitution of the common femoral arteries through mesenteric and abdominal wall collaterals

Fig. 23.5 Peripheral occlusive vascular disease. 59-Year-old male presents with left calf claudication. (a) Pelvis, (b) thigh, and (c) calf MIPS from a subtracted contrast-enhanced runoff demonstrate segmental

occlusion of the distal left superficial femoral artery in the region of the adductor canal, with two vessel runoff via the peroneal and posterior tibial arteries

tic aneurysm. Endovascular aneurysm repair is the preferred therapy. The popliteal artery is an uncommon location for aneurysm formation (Fig. 23.8). Popliteal aneurysms are commonly associated with aortic aneurysms (Fig. 23.8d). As with aneurysms in other vascular territories, complications are related to rupture, progressive thrombus accumulation with stenosis or occlusion, or thromboembolism. Preferred treatment methods repair the native vessel; saphenous vein interposition grafts are considered in those patients whose native vessels cannot be repaired.

artery [28]. This entity should be considered in a young male patient presenting with calf claudication and in older patients without risk factors for atherosclerotic disease. The average age at presentation is less than 30 years, although this entity has been diagnosed in the seventh decade of life. The male:female ratio is approximately 15:1. A rare congenital abnormality with a prevalence of 50% diameter stenosis on a per-vessel basis (and 94, 82, and 89% on a per-patient basis) [132]. Coronary MRI with a bolus infusion of Gd-BOPTA has been recently reported [133]. In the example shown in Fig. 24.11, whole-heart coronary images are acquired after a bolus injection of Gd-BOPTA and the three major coronary vessels are clearly depicted in the reformatted and 3D volume rendered images. A potential advantage of this approach is that its contrast injection method is compatible with LGE, making it possible to assess coronary artery stenosis and myocardium viability using a single bolus contrast injection. The bolus injection approach also simplifies the initiation time of coronary MRI acquisition compared to slow infusion. Despite the tremendous technical improvements in the last two decades, the sensitivity and specificity of coronary MRI for detection of CAD remain moderate based on singlecenter [5, 77, 78, 132, 134–137] (Table 24.1) and multicenter [63] studies. Coronary motion, SNR and CNR will remain major impediments to coronary MRI in the foreseeable future and these issues need to be addressed before clinical prime time for coronary MRI.

Coronary Vein MRI Knowledge of coronary vein anatomy is becoming increasingly important for diagnostic and interventional cardiac procedures including epicardial radiofrequency ablation [138, 139], retrograde perfusion therapy in high-risk or complicated coronary angioplasty [140], arrhythmia assessment [141, 142], stem cell delivery [143], coronary artery bypass surgery [144], and cardiac resynchronization therapy (CRT) [145, 146]. In patients with severe congestive heart failure, CRT has been proven as an adjuvant therapy to pharmacological treatment [147, 148]. In CRT, simultaneous pacing of the right ventricle and left ventricle (LV), or pacing the LV

344

R. Nezafat et al.

Fig. 24.12 The variations in the coronary venous anatomy can be observed in images from four healthy adult subjects acquired using magnetization transfer GRE during the systolic rest period. There are clear variations in the branching point, angle, and diameter of

different tributaries of coronary sinus, highlighting the potential for noninvasive assessment of the coronary venous anatomy in cardiac resynchronization therapy. For example, subject (a) has no visible lateral vein (LatV)

alone, results in hemodynamic improvement and restoration of a more physiological contraction pattern [149]. One of the technical difficulties of CRTs is achieving effective, safe and permanent pacing of the LV. Transvenous coronary sinus pacing is the most common technique as it has the least procedural risk, but it is associated with long procedure times, extensive radiation exposure from fluoroscopy, implantation failure and LV lead dislodgment. Two of the major difficulties of the transvenous approach are the small number of coronary vein branches adjacent to an appropriate LV wall and the great variability in coronary vein anatomy [146]. Ideally, coronary venous morphology should be assessed

noninvasively prior to CRT procedure, to determine whether epicardial or transvenous lead placement would be more appropriate. Coronary artery MRI techniques developed over the past two decades are applicable to imaging coronary veins except for magnetization preparation methods and optimal time window for imaging within the cardiac cycle. Magnetization transfer preparation sequences have been demonstrated as an alternative to T2 magnetization preparation, commonly used in coronary MRI, for both targeted [150] and whole-heart [151] approaches. Figure 24.12 shows example coronary vein MRI using a targeted approach with a magnetization

24

Noninvasive Imaging for Coronary Artery Disease

preparation sequence. Use of intravascular contrast agents such as gadocoletic acid trisodium salt [152] has also been reported to improve contrast in coronary vein MRI. With high relaxivity extracellular contrast agents such as Gd-BOPTA, coronary vein can also be easily visualized, similar to coronary arteries. While coronary artery MRI is commonly performed during mid-diastolic quiescent period, coronary vein MRI should be in the end-systolic quiescent period, as it coincides with the maximum size of the coronary veins [117].

References 1. Lloyd-Jones D, Adams RJ, Brown TM, et al. Heart disease and stroke statistics--2010 update: a report from the American Heart Association. Circulation. 2010;121:e46–e215. 2. Omran H, Schmidt H, Hackenbroch M, et al. Silent and apparent cerebral embolism after retrograde catheterisation of the aortic valve in valvular stenosis: a prospective, randomised study. Lancet. 2003;361:1241–1246. 3. Gerber BL, Coche E, Pasquet A, et al. Coronary artery stenosis: direct comparison of four-section multi-detector row CT and 3D navigator MR imaging for detection – initial results. Radiology. 2005;234:98. 4. Kefer J, Coche E, Legros G, et al. Head-to-head comparison of three-dimensional navigator-gated magnetic resonance imaging and 16-slice computed tomography to detect coronary artery stenosis in patients. J Am Coll Cardiol. 2005;46:92–100. 5. Pouleur AC, le Polain de Waroux JB, Kefer J, Pasquet A, Vanoverschelde JL, Gerber BL. Direct comparison of whole-heart navigator-gated magnetic resonance coronary angiography and 40and 64-slice multidetector row computed tomography to detect the coronary artery stenosis in patients scheduled for conventional coronary angiography. Circ Cardiovasc Imaging. 2008;1:114–121. 6. Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnostic accuracy of noninvasive coronary angiography using 64-slice spiral computed tomography. J Am Coll Cardiol. 2005;46:552–557. 7. Maintz D, Ozgun M, Hoffmeier A, et al. Whole-heart coronary magnetic resonance angiography: value for the detection of coronary artery stenoses in comparison to multislice computed tomography angiography. Acta Radiol. 2007;48:967–973. 8. Liu X, Zhao X, Huang J, et al. Comparison of 3D free-breathing coronary MR angiography and 64-MDCT angiography for detection of coronary stenosis in patients with high calcium scores. AJR Am J Roentgenol. 2007;189:1326–1332. 9. Choi SI, George RT, Schuleri KH, Chun EJ, Lima JA, Lardo AC. Recent developments in wide-detector cardiac computed tomography. Int J Cardiovasc Imaging. 2009;25(Suppl 1):23–29. 10. Hein PA, Romano VC, Lembcke A, May J, Rogalla P. Initial experience with a chest pain protocol using 320-slice volume MDCT. Eur Radiol. 2009;19:1148–1155. 11. Hoe J, Toh KH. First experience with 320-row multidetector CT coronary angiography scanning with prospective electrocardiogram gating to reduce radiation dose. J Cardiovasc Comput Tomogr. 2009;3:257–261. 12. Pasricha SS, Nandurkar D, Seneviratne SK, et al. Image quality of coronary 320-MDCT in patients with atrial fibrillation: initial experience. AJR Am J Roentgenol. 2009;193:1514–1521. 13. Rixe J, Conradi G, Rolf A, et al. Radiation dose exposure of computed tomography coronary angiography: comparison of dualsource, 16-slice and 64-slice CT. Heart. 2009;95:1337–1342.

345 14. Achenbach S, Anders K, Kalender WA. Dual-source cardiac computed tomography: image quality and dose considerations. Eur Radiol. 2008;18:1188–1198. 15. Achenbach S, Marwan M, Ropers D, et al. Coronary computed tomography angiography with a consistent dose below 1 mSv using prospectively electrocardiogram-triggered high-pitch spiral acquisition. Eur Heart J. 2010;31:340–346. 16. Achenbach S, Ropers D, Kuettner A, et al. Contrast-enhanced coronary artery visualization by dual-source computed tomography-initial experience. Eur J Radiol. 2006;57:331–335. 17. Achenbach S, Ropers U, Kuettner A, et al. Randomized comparison of 64-slice single- and dual-source computed tomography coronary angiography for the detection of coronary artery disease. JACC Cardiovasc Imaging. 2008;1:177–186. 18. Kuettner A, Gehann B, Spolnik J, et al. Strategies for dose-optimized imaging in pediatric cardiac dual source CT. Rofo. 2009; 181:339–348. 19. Lell M, Marwan M, Schepis T, et al. Prospectively ECG-triggered high-pitch spiral acquisition for coronary CT angiography using dual source CT: technique and initial experience. Eur Radiol. 2009;19:2576–2583. 20. Marwan M, Ropers D, Pflederer T, Daniel WG, Achenbach S. Clinical characteristics of patients with obstructive coronary lesions in the absence of coronary calcification: an evaluation by coronary CT angiography. Heart. 2009;95:1056–1060. 21. Pflederer T, Jakstat J, Marwan M, et al. Radiation exposure and image quality in staged low-dose protocols for coronary dualsource CT angiography: a randomized comparison. Eur Radiol. 2010;20:1197–1206. 22. Schepis T, Marwan M, Pflederer T, et al. Quantification of noncalcified coronary atherosclerotic plaques with dual source computed tomography: comparison to intravascular ultrasound. Heart. 2010;96: 610–615. 23. Blankstein R, Shah A, Pale R, et al. Radiation dose and image quality of prospective triggering with dual-source cardiac computed tomography. Am J Cardiol. 2009;103:1168–1173. 24. Kalra MK, Brady TJ. Current status and future directions in technical developments of cardiac computed tomography. J Cardiovasc Comput Tomogr. 2008;2:71–80. 25. Kimbiris D, Iskandrian AS, Segal BL, Bemis CE. Anomalous aortic origin of coronary arteries. Circulation. 1978;58:606–615. 26. Davis JA, Cecchin F, Jones TK, Portman MA. Major coronary artery anomalies in a pediatric population: incidence and clinical importance. J Am Coll Cardiol. 2001;37:593–597. 27. Maron BJ, Thompson PD, Ackerman MJ, et al. Recommendations and considerations related to preparticipation screening for cardiovascular abnormalities in competitive athletes: 2007 update: a scientific statement from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism: endorsed by the American College of Cardiology Foundation. Circulation. 2007;115:1643–1455. 28. Maron BJ, Carney KP, Lever HM, et al. Relationship of race to sudden cardiac death in competitive athletes with hypertrophic cardiomyopathy. J Am Coll Cardiol. 2003;41:974–980. 29. Corrado D, Basso C, Schiavon M, Thiene G. Screening for hypertrophic cardiomyopathy in young athletes. N Engl J Med. 1998; 339:364–369. 30. Eckart RE, Scoville SL, Campbell CL, et al. Sudden death in young adults: a 25-year review of autopsies in military recruits. Ann Int Med. 2004;141:829–834. 31. Taylor AJ, Rogan KM, Virmani R. Sudden cardiac death associated with isolated congenital coronary artery anomalies. J Am Coll Cardiol. 1992;20:640–647. 32. Taylor AJ, Byers JP, Cheitlin MD, Virmani R. Anomalous right or left coronary artery from the contralateral coronary sinus: “highrisk” abnormalities in the initial coronary artery course and heterogeneous clinical outcomes. Am Heart J. 1997;133:428–435.

346 33. Cheitlin MD, De Castro CM, McAllister HA. Sudden death as a complication of anomalous left coronary origin from the anterior sinus of Valsalva, a not-so-minor congenital anomaly. Circulation. 1974;50:780–787. 34. McConnell MV, Ganz P, Selwyn AP, Li W, Edelman RR, Manning WJ. Identification of anomalous coronary arteries and their anatomic course by magnetic resonance coronary angiography. Circulation. 1995;92:3158–3162. 35. Post JC, van Rossum AC, Bronzwaer JG, et al. Magnetic resonance angiography of anomalous coronary arteries. A new gold standard for delineating the proximal course? Circulation. 1995;92:3163–3171. 36. Taylor AM, Thorne SA, Rubens MB, et al. Coronary artery imaging in grown up congenital heart disease: complementary role of magnetic resonance and x-ray coronary angiography. Circulation. 2000;101:1670–1678. 37. Bunce NH, Lorenz CH, Keegan J, et al. Coronary artery anomalies: assessment with free-breathing three-dimensional coronary MR angiography. Radiology. 2003;227:201–208. 38. Hendel RC, Patel MR, Kramer CM, et al. ACCF/ACR/SCCT/ SCMR/ASNC/NASCI/SCAI/SIR 2006 appropriateness criteria for cardiac computed tomography and cardiac magnetic resonance imaging: a report of the American College of Cardiology Foundation Quality Strategic Directions Committee Appropriateness Criteria Working Group, American College of Radiology, Society of Cardiovascular Computed Tomography, Society for Cardiovascular Magnetic Resonance, American Society of Nuclear Cardiology, North American Society for Cardiac Imaging, Society for Cardiovascular Angiography and Interventions, and Society of Interventional Radiology. J Am Coll Cardiol. 2006;48:1475–1497. 39. Hundley WG, Bluemke DA, Finn JP, et al. ACCF/ACR/AHA/ NASCI/SCMR 2010 expert consensus document on cardiovascular magnetic resonance: a report of the American College of Cardiology Foundation Task Force on Expert Consensus Documents. Circulation. 2010;55:2614–2662. 40. Burns JC, Glode MP. Kawasaki syndrome. Lancet. 2004;364: 533–544. 41. Holman RC, Curns AT, Belay ED, Steiner CA, Schonberger LB. Kawasaki syndrome hospitalizations in the United States, 1997 and 2000. Pediatrics. 2003;112(3 Pt 1):495–501. 42. Du ZD, Zhao D, Du J, et al. Epidemiologic study on Kawasaki disease in Beijing from 2000 through 2004. Pediatr Infect Dis J. 2007;26:449–451. 43. Burns JC, Cayan DR, Tong G, et al. Seasonality and temporal clustering of Kawasaki syndrome. Epidemiology 2005;16:220–225. 44. Akagi T, Rose V, Benson LN, Newman A, Freedom RM. Outcome of coronary artery aneurysms after Kawasaki disease. J Pediatr. 1992;121(5 Pt 1):689–694. 45. Greil GF, Stuber M, Botnar RM, et al. Coronary magnetic resonance angiography in adolescents and young adults with Kawasaki disease. Circulation. 2002;105:908–911. 46. Mavrogeni S, Papadopoulos G, Douskou M, et al. Magnetic resonance angiography, function and viability evaluation in patients with Kawasaki disease. J Cardiovasc Magn Reson. 2006;8:493–498. 47. Mavrogeni S, Papadopoulos G, Douskou M, et al. Magnetic resonance angiography is equivalent to X-ray coronary angiography for the evaluation of coronary arteries in Kawasaki disease. J Am Coll Cardiol. 2004;43(4):649–652. 48. Mavrogeni SI, Manginas A, Papadakis E, et al. Correlation between magnetic resonance angiography (MRA) and quantitative coronary angiography (QCA) in ectatic coronary vessels. J Cardiovasc Magn Reson. 2004;6:17–23. 49. Harrison A, Abolhoda A, Ahsan C. Cardiovascular complications in Behcet syndrome: acute myocardial infarction with late stent thrombosis and coronary, ventricular, and femoral pseudoaneurysms. Tex Heart Inst J. 2009;36:498–500.

R. Nezafat et al. 50. Yakut ZI, Odev K. Pulmonary and cardiac involvement in Behcet disease: 3 case reports. Clin Appl Thromb Hemost. 2007;13(3): 318–322. 51. Kanaan S, Baker C, Starnes V. Resection of giant coronary artery aneurysms in a Takayasu’s arteritis patient. Ann Thorac Surg. 2008;85:1795–1796. 52. Malik IS, Harare O, AL-Nahhas A, Beatt K, Mason J. Takayasu’s arteritis: management of left main stem stenosis. Heart. 2003;89:e9. 53. Khalaf HH, Arafah MR, Refaat AA, Ibrahim MF. Coronary artery bypass grafting for Takayasu arteritis with severe coronary, carotid, subclavian, and renal artery involvement and subsequent pregnancy. Interact Cardiovasc Thorac Surg. 2006;5:153–155. 54. White R, Caputo G, Mark A, Modin G, Higgins C. Coronary artery bypass graft patency: noninvasive evaluation with MR imaging. Radiology. 1987;164:681–686. 55. White R, Pflugfelder P, Lipton M, Higgins C. Coronary artery bypass grafts: evaluation of patency with cine MR imaging. AJR Am J Roentgenol. 1988;150:1271–1274. 56. Rubinstein R, Askenase A, Thickman D, Feldman M, Agarwal J, Helfant R. Magnetic resonance imaging to evaluate patency of aortocoronary bypass grafts. Circulation. 1987;76:786–791. 57. Wintersperger B, Engelmann M, Von Smekal A, et al. Patency of coronary bypass grafts: assessment with breath-hold contrastenhanced MR angiography--value of a non-electrocardiographically triggered technique. Radiology. 1998;208:345–351. 58. Engelmann M, Knez A, von Smekal A, et al. Non-invasive coronary bypass graft imaging after multivessel revascularisation. Int J Cardiol. 2000;76:65–74. 59. Vrachliotis T, Bis K, Aliabadi D, Shetty A, Safian R, Simonetti O. Contrast-enhanced breath-hold MR angiography for evaluating patency of coronary artery bypass grafts. AJR Am J Roentgenol. 1997;168:1073–1080. 60. Klein C, Nagel E, Gebker R, et al. Magnetic resonance adenosine perfusion imaging in patients after coronary artery bypass graft surgery. J Am Coll Cardiol Img. 2009;2:437–445. 61. Manning WJ, Edelman RR. Magnetic resonance coronary angiography. Magn Reson Q. 1993;9:131–151. 62. Manning WJ, Li W, Edelman RR. A preliminary report comparing magnetic resonance coronary angiography with conventional angiography. N Engl J Med. 1993;328:828–832. 63. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance angiography for the detection of coronary stenoses. N Engl J Med. 2001;345:1863–1869. 64. Stuber M, Botnar RM, Danias PG, Kissinger KV, Manning WJ. Breathhold three-dimensional coronary magnetic resonance angiography using real-time navigator technology. J Cardiovasc Magn Reson. 1999;1:233–238. 65. Stuber M, Botnar RM, Danias PG, Kissinger KV, Manning WJ. Submillimeter three-dimensional coronary MR angiography with real-time navigator correction: comparison of navigator locations. Radiology. 1999;212:579–587. 66. Stuber M, Botnar RM, Danias PG, et al. Double-oblique freebreathing high resolution three-dimensional coronary magnetic resonance angiography. J Am Coll Cardiol. 1999;34:524–531. 67. Bhat H, Zuehlsdorff S, Bi X, Li D. Whole-heart contrast-enhanced coronary magnetic resonance angiography using gradient echo interleaved EPI. Magn Reson Med. 2009;61:1388–1395. 68. Bi X, Carr JC, Li D. Whole-heart coronary magnetic resonance angiography at 3 Tesla in 5 minutes with slow infusion of Gd-BOPTA, a high-relaxivity clinical contrast agent. Magn Reson Med. 2007;58:1–7. 69. Bi X, Deshpande V, Carr J, Li D. Coronary artery magnetic resonance angiography (MRA): a comparison between the whole-heart and volume-targeted methods using a T2-prepared SSFP sequence. J Cardiovasc Magn Reson. 2006;8:703–707.

24

Noninvasive Imaging for Coronary Artery Disease

70. Kim YJ, Seo JS, Choi BW, Choe KO, Jang Y, Ko YG. Feasibility and diagnostic accuracy of whole heart coronary MR angiography using free-breathing 3D balanced turbo-field-echo with SENSE and the half-fourier acquisition technique. Korean J Radiol. 2006;7:235–242. 71. Lai P, Bi X, Jerecic R, Li D. A respiratory self-gating technique with 3D-translation compensation for free-breathing whole-heart coronary MRA. Magn Reson Med. 2009;62:731–738. 72. Lai P, Larson AC, Bi X, Jerecic R, Li D. A dual-projection respiratory self-gating technique for whole-heart coronary MRA. J Magn Reson Imaging. 2008;28:612–620. 73. Liu X, Bi X, Huang J, Jerecic R, Carr J, Li D. Contrast-enhanced whole-heart coronary magnetic resonance angiography at 3.0 T: comparison with steady-state free precession technique at 1.5 T. Invest Radiol. 2008;43:663–668. 74. Nehrke K, Bornert P, Mazurkewitz P, Winkelmann R, Grasslin I. Free-breathing whole-heart coronary MR angiography on a clinical scanner in four minutes. J Magn Reson Imaging. 2006;23:752–756. 75. Niendorf T, Hardy CJ, Giaquinto RO, et al. Toward single breathhold whole-heart coverage coronary MRA using highly accelerated parallel imaging with a 32-channel MR system. Magn Reson Med. 2006;56:167–176. 76. Okada T, Kanao S, Ninomiya A, et al. Whole-heart coronary magnetic resonance angiography with parallel imaging: comparison of acceleration in one-dimension vs. two-dimensions. Eur J Radiol. 2009;71:486–491. 77. Sakuma H, Ichikawa Y, Chino S, Hirano T, Makino K, Takeda K. Detection of coronary artery stenosis with whole-heart coronary magnetic resonance angiography. J Am Coll Cardiol. 2006;48: 1946–1950. 78. Sakuma H, Ichikawa Y, Suzawa N, et al. Assessment of coronary arteries with total study time of less than 30 minutes by using wholeheart coronary MR angiography. Radiology. 2005;237:316–321. 79. Stehning C, Bornert P, Nehrke K, Eggers H, Stuber M. Free-breathing whole-heart coronary MRA with 3D radial SSFP and self-navigated image reconstruction. Magn Reson Med. 2005;54:476–480. 80. Weber OM, Martin AJ, Higgins CB. Whole-heart steady-state free precession coronary artery magnetic resonance angiography. Magn Reson Med. 2003;50:1223–1228. 81. Deshpande VS, Shea SM, Laub G, Simonetti OP, Finn JP, Li D. 3D magnetization-prepared true-FISP: a new technique for imaging coronary arteries. Magn Reson Med. 2001;46:494–502. 82. Ozgun M, Hoffmeier A, Kouwenhoven M, et al. (2005) Comparison of 3D segmented gradient-echo and steady-state free precession coronary MRI sequences in patients with coronary artery disease. AJR Am. J Roentgenol. 185:103–109. 83. Maintz D, Aepfelbacher FC, Kissinger KV, et al. Coronary MR angiography: comparison of quantitative and qualitative data from four techniques. AJR Am J Roentgenol. 2004;182:515–521. 84. Nezafat R, Herzka D, Stehning C, Peters DC, Nehrke K, Manning WJ. Inflow quantification in three-dimensional cardiovascular MR imaging. J Magn Reson Imaging. 2008;28:1273–1279. 85. Scott A, Keegan J, Firmin D. Motion in cardiovascular MR imaging. Radiology. 2009;250:331–351. 86. Lu B, Mao S, Zhuang N, et al. Coronary artery motion during the cardiac cycle and optimal ECG triggering for coronary artery imaging. Invest Radiol. 2001;36:250–256. 87. Hofman M, Wickline S, Lorenz C. Quantification of in-plane motion of the coronary arteries during the cardiac cycle: implications for acquisition window duration for MR flow quantification. J Magn Res Imaging. 1998;8:568–576.

347 88. Wang Y, Vidan E, Bergman G. Cardiac motion of coronary arteries: variability in the rest period and implications for coronary MR angiography. Radiology. 1999;213:751–758. 89. Ehman R, Felmlee J. Adaptive technique for high-definition MR imaging of moving structures. Radiology. 1989;173:255–263. 90. Wang Y, Rossman P, Grimm R, Riederer S, Ehman R. Navigatorecho-based real-time respiratory gating and triggering for reduction of respiration effects in three-dimensional coronary MR angiography. Radiology. 1996;198:55–60. 91. McConnell M, Khasgiwala V, Savord B, et al. Comparison of respiratory suppression methods and navigator locations for MR coronary angiography. AJR Am J Roentgenol. 1997;168:1369–1375. 92. Taylor A, Jhooti P, Wiesmann F, Keegan J, Firmin D, Pennell D. MR navigator-echo monitoring of temporal changes in diaphragm position: implications for MR coronary angiography. J Magn Res Imaging. 1997;7:629–636. 93. Huber A, Nikolaou K, Gonschior P, Knez A, Stehling M, Reiser M. Navigator echo-based respiratory gating for three-dimensional MR coronary angiography: results from healthy volunteers and patients with proximal coronary artery stenoses. AJR Am J Roentgenol. 1999;173:95–101. 94. Wang Y, Riederer S, Ehman R. Respiratory motion of the heart: kinematics and the implications for the spatial resolution in coronary imaging. Magn Reson Med. 1995;33:713–719. 95. Danias PG, McConnell MV, Khasgiwala VC, Chuang ML, Edelman RR, Manning WJ. Prospective navigator correction of image position for coronary MR angiography. Radiology. 1997;203:733–736. 96. Danias PG, Stuber M, Botnar RM, Kissinger KV, Edelman RR, Manning WJ. Relationship between motion of coronary arteries and diaphragm during free breathing: lessons from real-time MR imaging. AJR Am J Roentgenol. 1999;172:1061–1065. 97. Nezafat R, Stehning C, Gharib AM, et al. Improved spatial-temporal resolution MR coronary blood flow imaging at 3 T. J Cardiovasc Magn Res. 2005;7:199. 98. Nehrke K, Bornert P. Prospective correction of affine motion for arbitrary MR sequences on a clinical scanner. Magn Reson Med. 2005;54(5). 99. Lai P, Larson AC, Park J, Carr JC, Li D. Respiratory self-gated four-dimensional coronary MR angiography: a feasibility study. Magn Reson Med. 2008;59:1378–1385. 100. Larson AC, Kellman P, Arai A, et al. Preliminary investigation of respiratory self-gating for free-breathing segmented cine MRI. Magn Reson Med. 2005;53:159–168. 101. Park J, Larson AC, Zhang Q, Simonetti O, Li D. 4D radial coronary artery imaging within a single breath-hold: cine angiography with phase-sensitive fat suppression (CAPS). Magn Reson Med. 2005;54:833–840. 102. Keegan J, Gatehouse P, Yang G, Firmin D. Non-model based correction of respiratory motion using beat-to-beat 3d spiral fat-selective imaging. J Magn Reson Imaging. 2007;26:624–629. 103. Nguyen T, Spincemaille P, Cham M, Weinsaft J, Prince M, Wang Y. Free-breathing 3D steady-state free precession coronary magnetic resonance angiography: comparison of diaphragm and cardiac fat navigators. J Magn Reson Imaging. 2008;28:509–514. 104. Nguyen T, Spincemaille P, Prince M, Wang Y. 2128 Free-breathing steady-state free precession 3D coronary MRA: comparison of diaphragm and cardiac fat navigator techniques. J Cardiovasc Magn Res. 2008;10(Suppl 1):A397. 105. Nguyen T, Spincemaille P, Cham M, Weinsaft J, Prince M, Wang Y. Free-breathing 3-dimensional steady-state free precession coronary magnetic resonance angiography: comparison of four navigator gating techniques. Magn Res Imaging. 2009;27:807–814.

348 106. Uribe S, Muthurangu V, Boubertakh R, et al. Whole-heart cine MRI using real-time respiratory self-gating. Magn Reson Med. 2007;57:606–613. 107. Brau AC, Brittain JH. Generalized self-navigated motion detection technique: Preliminary investigation in abdominal imaging. Magn Reson Med. 2006;55:263–270. 108. Nehrke K, Bornert P, Manke D, Bock JC. Free-breathing cardiac MR imaging: study of implications of respiratory motion--initial results. Radiology. 2001;220:810–815. 109. Gurney PT, Hargreaves BA, Nishimura DG. Design and analysis of a practical 3D cones trajectory. Magn Reson Med. 2006;55: 575–582. 110. Santos JM, Cunningham CH, Lustig M, et al. Single breath-hold whole-heart MRA using variable-density spirals at 3 T. Magn Reson Med. 2006;55:371–379. 111. Sussman M, Stainsby J, Robert N, Merchant N, Wright G. Variable-density adaptive imaging for high-resolution coronary artery MRI. Magn Reson Med. 2002;48:753–764. 112. Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishimura DG. Coronary angiography with magnetization-prepared T2 contrast. Magn Reson Med. 1995;33:689–696. 113. Botnar RM, Stuber M, Danias PG, Kissinger KV, Manning WJ. Improved coronary artery definition with T2-weighted, freebreathing, three-dimensional coronary MRA. Circulation. 1999;99: 3139–3148. 114. Nezafat R, Derbyshire J, Ouwerkerk R, Stuber M, McVeigh E. Spectrally selective B1 insensitive T2 preparation sequence for 3 T imaging. Proceedings of 14th Scientific Meeting, International Society for Magnetic Resonance in Medicine. 2006:596. 115. Nezafat R, Ouwerkerk R, Derbyshire AJ, Stuber M, McVeigh ER. Spectrally selective B1-insensitive T2 magnetization preparation sequence. Magn Reson Med. 2009;61(6):1326–1335. 116. Li D, Paschal CB, Haacke EM, Adler LP. Coronary arteries: threedimensional MR imaging with fat saturation and magnetization transfer contrast. Radiology. 1993;187:401–406. 117. Nezafat R, Han Y, Peters DC, et al. Coronary magnetic resonance vein imaging: imaging contrast, sequence, and timing. Magn Reson Med. 2007;58:1196–1206. 118. Stuber M, Botnar RM, Fischer SE, et al. Preliminary report on in vivo coronary MRA at 3 Tesla in humans. Magn Reson Med. 2002;48:425–429. 119. Schar M, Kozerke S, Fischer Se Boesiger P. Cardiac SSFP imaging at 3 Tesla. Magn Reson Med. 2004;51:799–806. 120. Nezafat R, Stuber M, Ouwerkerk R, Gharib AM, Desai MY, Pettigrew RI. B1-insensitive T2 preparation for improved coronary magnetic resonance angiography at 3 T. Magn Reson Med. 2006;55:858–864. 121. van Elderen SG, Versluis MJ, Webb AG, et al. Initial results on in vivo human coronary MR angiography at 7 T. Magn Reson Med. 2009;62:1379–1384. 122. Terashima M, Meyer CH, Keeffe BG, et al. Noninvasive assessment of coronary vasodilation using magnetic resonance angiography. J Am Coll Cardiol. 2005;45:104–110. 123. Hu P, Chuang ML, Ngo LH, et al. Coronary MR imaging: effect of timing and dose of isosorbide dinitrate administration. Radiology. 2010;254:401–409. 124. Zheng J, Bae KT, Woodard PK, Haacke EM, Li D. Efficacy of slow infusion of gadolinium contrast agent in three-dimensional MR coronary artery imaging. J Magn Reson Imaging. Nov 1999; 10:800–805. 125. Goldfarb JW, Edelman RR. Coronary arteries: breath-hold, gadolinium-enhanced, three-dimensional MR angiography. Radiology. 1998;206:830–834.

R. Nezafat et al. 126. Knuesel PR, Nanz D, Wolfensberger U, et al. Multislice breathhold spiral magnetic resonance coronary angiography in patients with coronary artery disease: effect of intravascular contrast medium. J Magn Reson Imaging. 2002;16:660–667. 127. Herborn CU, Barkhausen J, Paetsch I, et al. Coronary arteries: contrast-enhanced MR imaging with SH L 643A--experience in 12 volunteers. Radiology. 2003;229:217–223. 128. Prompona M, Cyran C, Nikolaou K, Bauner K, Reiser M, Huber A. Contrast-enhanced whole-heart MR coronary angiography at 3.0 T using the intravascular contrast agent gadofosveset. Invest Radiol. 2009;44:369–374. 129. Kelle S, Thouet T, Tangcharoen T, et al. Whole-heart coronary magnetic resonance angiography with MS-325 (Gadofosveset). Med Sci Monit. 2007;13:CR469-CR474. 130. de Haen C, Anelli PL, Lorusso V, et al. Gadocoletic acid trisodium salt (b22956/1): a new blood pool magnetic resonance contrast agent with application in coronary angiography. Invest Radiol. 2006;41:279–291. 131. Tang L, Merkle N, Schar M, et al. Volume-targeted and whole-heart coronary magnetic resonance angiography using an intravascular contrast agent. J Magn Reson Imaging. 2009;30:1191–1196. 132. Yang Q, Li K, Liu X, et al. Contrast-enhanced whole-heart coronary magnetic resonance angiography at 3.0-T: a comparative study with X-ray angiography in a single center. J Am Coll Cardiol. 2009;54:69–76. 133. Hu P, Chan J, Smink J, et al. Contrast-enhanced whole-heart coronary MRI with a bolus infusion of gadobenate dimeglumine at 1.5 T. Magn Reson Imaging. 2011;66:392–398. 134. Bunce NH, Jhooti P, Keegan J, et al. Evaluation of free-breathing three-dimensional magnetic resonance coronary angiography with hybrid ordered phase encoding (HOPE) for the detection of proximal coronary artery stenosis. J Magn Reson Imaging. 2001; 14:677–684. 135. Sommer T, Hofer U, Hackenbroch M, et al. [Submillimeter 3D coronary MR angiography with real-time navigator correction in 107 patients with suspected coronary artery disease]. Rofo. 2002;174:459–466. 136. Bogaert J, Kuzo R, Dymarkowski S, Beckers R, Piessens J, Rademakers FE. Coronary artery imaging with real-time navigator three-dimensional turbo-field-echo MR coronary angiography: initial experience. Radiology. 2003;226:707–716. 137. Jahnke C, Paetsch I, Schnackenburg B, et al. Coronary MR angiography with steady-state free precession: individually adapted breath-hold technique versus free-breathing technique. Radiology. 2004;232:669–676. 138. Kusano KF, Morita H, Fujimoto Y, Hirose E, Ohe T. Catheter ablation of an epicardial accessory pathway via the middle cardiac vein guided by monophasic action potential recordings. Europace. 2001;3:164–167. 139. Haissaguerre M, Gaita F, Fischer B, Egloff P, Lemetayer P, Warin JF. Radiofrequency catheter ablation of left lateral accessory pathways via the coronary sinus. Circulation. 1992;86:1464–1468. 140. Kar S, Nordlander R. Coronary veins: an alternate route to ischemic myocardium. Heart Lung. 1992;21:148–157. 141. Stellbrink C, Diem B, Schauerte P, Ziegert K, Hanrath P. Transcoronary venous radiofrequency catheter ablation of ventricular tachycardia. J Cardiovasc Electrophysiol. 1997;8: 916–921. 142. Butter C, Meisel E, Engelmann L, et al. Human experience with transvenous biventricular defibrillation using an electrode in a left ventricular vein. Pacing Clin Electrophysiol. 2002;25:324–331. 143. Thompson CA, Nasseri BA, Makower J, et al. Percutaneous transvenous cellular cardiomyoplasty. A novel nonsurgical approach

24

Noninvasive Imaging for Coronary Artery Disease

for myocardial cell transplantation. J Am Coll Cardiol. 2003;41: 1964–1971. 144. Oesterle SN, Reifart N, Hauptmann E, Hayase M, Yeung AC. Percutaneous in situ coronary venous arterialization: report of the first human catheter-based coronary artery bypass. Circulation. 2001;103:2539–2543. 145. de Paola AA, Melo WD, Tavora MZ, Martinez EE. Angiographic and electrophysiological substrates for ventricular tachycardia mapping through the coronary veins. Heart. 1998;79:59–63. 146. Singh JP, Houser S, Heist EK, Ruskin JN. The coronary venous anatomy: a segmental approach to aid cardiac resynchronization therapy. J Am Coll Cardiol. 2005;46:68–74. 147. Bax JJ, Abraham T, Barold SS, et al. Cardiac resynchronization therapy: Part 1--issues before device implantation. J Am Coll Cardiol. 2005;46:2153–2167.

349 148. Bax JJ, Abraham T, Barold SS, et al. Cardiac resynchronization therapy: Part 2--issues during and after device implantation and unresolved questions. J Am Coll Cardiol. 2005;46:2168–2182. 149. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic heart failure. N Engl J Med. 2002;346:1845–1853. 150. Nezafat R, Han Y, Peters DC, et al. Magnetic resonance coronary vein imaging: sequence, contrast and timing. Magn Reson Med. 2007;58:1196–1206. 151. Stoeck CT, Han Y, Peters DC, et al. Whole heart magnetizationprepared steady-state free precession coronary vein MRI. J Magn Reson Imaging. 2009;29:1293–1299. 152. Rasche V, Binner L, Cavagna F, et al. Whole-heart coronary vein imaging: a comparison between non-contrast-agent-and contrastagent-enhanced visualization of the coronary venous system. Magn Reson Med. 2007;57:1019–1026.

wwwwwwwwwwww

Venous Imaging: Techniques, Protocols, and Clinical Applications

25

Amir H. Davarpanah, Philip Hodnett, Jeremy D. Collins, James C. Carr, and Tim Scanlon

Introduction Magnetic resonance imaging (MRI) is intrinsically sensitive to flowing blood. The earliest investigators of MRI recognized that blood flow altered the intraluminal magnetic resonance (MR) signal. MRI techniques in imaging blood flow have been principally directed toward interrogation of the arterial system [1]. However, the same techniques that have made magnetic resonance angiography (MRA) such a useful clinical tool have also been applied to the venous system [2]. MR angiographic techniques can be divided into two categories: contrast material-enhanced and nonenhanced MRA [3]. Since its introduction in 1994 by Prince [1], first-pass contrast-enhanced (CE) MRA with gadolinium-based contrast material has seen widespread acceptance. The approach has benefited from many technical advances, including strategies to synchronize arrival of the bolus of contrast material with MR acquisition, moving bed technology for multistation studies, such as peripheral MRA, shortened acquisition times with parallel imaging, and k-space sharing methods, such as time-resolved imaging of contrast kinetics or TRICKS, for “time-resolved” MR angiographic acquisitions [1–3].

A.H. Davarpanah, MD () Department of Radiology, Yale School of Medicine, Yale University, New Haven, Connecticut, USA e-mail: [email protected] P. Hodnett, MD Department of Radiology, New York University, NY J.D. Collins, MD Department of Radiology, Northwestern Memorial Hospital and Northwestern University Feinberg School of Medicine, Chicago, IL, USA J.C. Carr, MD Northwestern University, Feinberg School of Medicine, Chicago, IL, USA T. Scanlon, MD Consultant radiologist, Limerick regional Hospital, Ireland

Several factors contribute to a renewed interest in nonenhanced MR angiographic methods [1]. Improvements in MR hardware and software, including the widespread availability of parallel imaging, have helped reduce acquisition times and made some methods clinically practical. Moreover, the recent association between high doses of gadolinium-based contrast material for MRA and the debilitating and occasionally life-threatening entity, nephrogenic systemic fibrosis, originally known as nephrogenic fibrosing dermopathy, has made it imperative that patients with moderate to severe renal insufficiency and vascular disease have nonenhanced alternatives for angiography/venography [4–6]. The purpose of this chapter is to review the techniques, clinical applications, and useful protocols of magnetic resonance venography (MRV), both noncontrast and contrastenhanced, in relation to the chest, abdomen, pelvis, and lower limbs, to aid in everyday practice.

MRV Techniques Multiple techniques, both noncontrast and contrast-enhanced pulse sequences, have been used to assess venous anatomy and pathology [2, 3]. Both categories of techniques include different strategies, including time-of-flight (TOF) imaging, phase-contrast imaging, MR direct thrombus imaging (MRDTI), gadolinium-enhanced MRV, and black blood techniques [2]. Contrast-enhanced MR venography (CEMRV) remains the mainstay technique for imaging the venous system. Noncontrast (NC) MRV is employed when there are contraindications to gadolinium contrast administration.

Noncontrast-Enhanced MR Venography Given the frequency of renal functional impairment in patients with peripheral vascular disease (PVD) and concerns about nephrogenic systemic fibrosis, over the last few years, several unenhanced MRA techniques have been introduced [7]. Three-dimensional fast spin echo (FSE) sequences with ECG

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_25, © Springer Science+Business Media, LLC 2012

351

352

A.H. Davarpanah et al.

Table 25.1 Protocol for noncontrast MR venography (NCMRV) Technique 1. 2D single-shot SSFP (axial and coronal) 2. T1 GRE FS (axial and coronal) 3. 3D SSFP (axial/LAO)

Comment Free breathing; ECG gated to diastole Noncontrast Free breathing; respiratory gated with navigator

gating to generate systolic and diastolic images can be modified to obtain venous information. The ability to generate 3-dimensional reconstruction of venous anatomy without contrast and without arterial contamination is very appealing; however, the efficacy of these techniques in body MRV has only begun to be evaluated. In a recent study, it was demonstrated that portal vein and its branches can be well-delineated with unenhanced MRA performed with a single-breath-hold, coronal, in-plane, ECG-synchronized, 3D half-Fourier, FSE sequence [8]. Similarly, in another study, a combination of 3D half-Fourier, FSE, flow-refocused fresh-blood imaging (FR-FBI) and the swap phase-encode arterial double-subtraction elimination (SPADE) technique images demonstrated to be highly accurate and reproducible for detecting deep vein thrombosis (DVT) without interference from implant susceptibility artifacts, especially in patients with nonmagnetizing metal implants during orthopedic surgery [9]. NCMRV protocol is detailed in Table 25.1.

Time of Flight TOF imaging employs a gradient-recalled echo pulse sequence [10]. In contradistinction to spin echo imaging, the signal from flowing blood is consistently replenished by the inflow of unsaturated (excited) spins producing increased intraluminal signal, resulting in “bright blood” images. The acquisitions may be either two dimensional (slice by slice) or three dimensional (a volume is acquired and then partitioned). A major drawback of TOF is the long acquisition time, which may result in image degradation due to motion artifact from patient movement. Additionally, in regions of slow flow, turbulence, or significant in-plane coursing of the vessel, there is loss of the intraluminal signal. Portions of the vessel lumen then appear darker and may mimic intraluminal pathology [10]. TOF MRV has generally been superseded by other NCMRV techniques due to the reasons mentioned above; however, the technique may be of use at higher field strength due to the increased signal-to-noise ratio and resultant improved image quality. TOF imaging has been used for imaging the intracranial veins [11, 12] and portal venous system [13]. Steady-State Free Precession Steady-state free precession (SSFP) is a bright blood imaging technique that is used extensively for cine MRI of the heart [14] and has been employed successfully as a noncon-

trast MRA technique for imaging the arterial system in several different vascular territories [3, 15]. Since, SSFP produces bright signal from both arteries and veins, it can be used as a method for NCMRV. The technique is implemented in 2D, as a stack of images, or in 3D, where respiratory gating needs to be utilized if used in the thorax or abdomen. ECG gating, where image acquisition occurs during diastole, improves image quality, particularly in the thorax. A major advantage of SSFP is its short scan time resulting in acquisition times significantly shorter than TOF imaging. SSFP has been used as an NCMRV technique in the lower extremity venous system [16] and also in the thorax.

Signal Targeting Alternative Radiofrequency and Flow-Independent Relaxation Enhancement A novel noncontrast MRV technique, designed by Koktzoglou and Edelman, is Signal Targeting Alternative Radiofrequency and Flow-Independent Relaxation Enhancement (STARFIRE) [17]. STARFIRE uses a bSSFP pulse sequence for readout of the MR signal to make blood vessels appear bright. Although bSSFP can be effective for imaging large vessels, such as the abdominal aorta and inferior vena cava (IVC), it can be problematic to use this technique for making projection angiograms of smaller vessels, such as the peripheral arteries and veins. With bSSFP, the smaller vessels tend to be obscured in MIP images by using the signal intensities of fat and muscle. Although chemical shift-selective methods for fat suppression can be applied with bSSFP, a substantial drawback is that they produce nonuniform signal suppression over large fields of view owing to static magnetic field inhomogeneity, a problem which is exacerbated by the sensitivity of peripheral array coils to the high signal intensity of subcutaneous fat. Moreover, the effect of the fat suppression radiofrequency (RF) pulse may be attenuated during the course of the bSSFP echo train. The goal of the STARFIRE technique is to suppress fat and muscle signals while maintaining the high signal intensity of the vasculature. Different contrast mechanisms are involved for fat and muscle signal suppression. With STARFIRE, fat suppression is performed predominantly on the basis of a T1-dependent mechanism, relying on the fact that fat has a much shorter T1 relaxation time than blood. The main rationale for using a T1-based, rather than chemical shift-based, mechanism for fat suppression is that the T1 relaxation times of the tissues are only negligibly dependent on static magnetic field inhomogeneity [17]. STARFIRE produces flow-independent images of all blood vessels (i.e., both arteries and veins). The presaturation pulses equalize arterial signal on the tagged and untagged image sets and suppress arterial signal on subtraction images. STARFIRE enables imaging of the vasculature over large fields of view as a result of the uniform suppression of both fat and muscle signals [17].

25

Venous Imaging: Techniques, Protocols, and Clinical Applications

3D Fast Spin Echo Recently, unenhanced MRA, using 3D half-Fourier FSE has been shown effective for visualization of the coronary arteries, renal arteries, and peripheral vessels [2, 3]. The 3D half-Fourier FSE method is one of the new MRA techniques that allows selective visualization of both arteries and veins. In general, an FSE sequence uses a long TE value and shows rapid blood flow as signal void; however, reduced echo train spacing enables compact echo sampling and reduces this phenomenon, especially when the flow velocity is low, as in the vein [3]. FSE images can also be acquired during systole, where arterial signal is suppressed due to rapid flow, and during systole, where both arteries and veins are visible together. Subtraction of both acquisitions has been used for arterial imaging, such as in the thoracic aorta [18]. However, images from either systole or diastole depict veins and can therefore be used for NCMRV imaging. MR-Directed Thrombus Imaging MRDTI relies on a gradient echo sequence which visualizes thrombus directly against a suppressed background [19]. In this instance, the sequence is a T1-weighted, magnetizationprepared, three dimensional gradient echo volume acquisition. The sequence employs a water-only excitation radiofrequency pulse that effectively eliminates any signal from fat. Unlike simple TOF GRE sequences which produce high signal from flowing blood, an inversion pulse is incorporated into this acquisition and is timed to effectively eliminate blood signal. It is known that acute venous thrombus contains significant methemoglobin [2, 3]. Methemoglobin causes T1 shortening within the thrombus causing it to appear bright. By suppressing the surrounding tissue and blood signal, the conspicuity of the thrombus becomes more apparent. MRDTI has been used to detect lower limb DVT, avoiding some of the flow related artifacts associated with other NCMRV techniques. Susceptibility-Weighted Imaging Susceptibility-weighted imaging (SWI) is a means of enhancing contrast in MRI. It is complementary to conventional spindensity, T1- and T2-weighted imaging methods. SWI is particularly suited for imaging venous blood, as it is very sensitive to deoxyhemoglobin, making it useful for imaging hemorrhages from trauma, where blood products can be visualized. The iron in deoxyhemoglobin in venous blood acts as an intrinsic contrast agent, causing T2*-related losses in the magnitude image and a shift in the phase image caused by susceptibility differences. The oxygen in oxyhemoglobin shields the iron so that T2* and susceptibility effects are only seen in venous blood. This provides a natural separation of venous and arterial blood, and allows for venographic images without any arterial contamination. Veins are dark because of T2* losses, whereas the arteries are bright from TOF inflow enhancement. In addition, it is possible to increase the contrast in the arteries without overly degrading the venographic images by using a slightly higher flip angle, short TR, and a

353

Fig. 25.1 Schematic drawing illustrating the range of each sequence. The parenchyma sequence is more sensitive to lower concentrations while the angiographic phase is more sensitive to higher concentrations. An overlap exists, making both sequences usable in a given range

thin slab [2, 20]. This is particularly effective at high and ultrahigh field because shorter echoes can be used to reduce flow dephasing in the arteries without affecting the venous contrast. SWI has been used to image the intracranial venous system.

Contrast-Enhanced MR Venography CEMRV is similar to CEMRA of the arterial system, except that the contrast injection is timed to the venous phase (Fig. 25.1). A spoiled gradient echo pulse sequence is used, as with CEMRA, and gadolinium contrast agent is injected intravenously. Minimum TR and parallel imaging are sued to shorten the acquisition time. The short scan time allows rapidly flowing blood to become saturated, thereby losing its bright signal. The intravenous administration of gadolinium shortens the T1 relaxation time of blood to such an extent that recovery occurs after each RF pulse and thus blood appears white. The surrounding tissues appear dark as they have become saturated and produce little signal. CEMRV can be implemented as a conventional “timed” MRV or as time-resolved MRV. Table 25.2 shows sequence protocol for CEMRV.

Conventional CEMRV Conventional CEMRV is a timed examination of the vessel of interest, similar to CEMRA. CEMRV may be performed using a direct or indirect approach. With the indirect approach (Fig. 25.2), nondiluted contrast is injected in a nontargeted peripheral vein and imaging takes place during the venous phase of the vessel of interest. Because considerable dilution of contrast occurs by the time it arrives to the area of interest, images are acquired in the early equilibrium phase to avoid redistribution. In general, it is recommended that larger doses of contrast agent are used, i.e., 0.2 mmol/kg, in order to fully fill the large capacitance venous system. The transit time can be estimated by adding 10–20 s to the arterial transit time, measured using a timing

354

A.H. Davarpanah et al.

bolus technique. Bolus tracking is less successful for venous imaging due to contrast dilution during the venous phase. Three to four postcontrast 3D images of the venous system are acquired to ensure adequate venous enhancement. Alternatively, the direct approach uses significantly less contrast because of targeted administration proximal to the area of interest. The direct technique involves injection of dilute contrast in a peripheral vein and imaging of the draining venous system. Thus, it is analogous to conventional X-ray venography. Dilution of gadolinium with saline prior to injection is necessary to avoid T2* effect, resulting in signal loss. Contrast

is injected into a peripheral vein and images are acquired before gadolinium reaches the heart. To ensure optimal opacification, bilateral injection may be of benefit for imaging the central veins in the chest. Bolus triggering is optional but not necessary, since the injection rate is relatively low and the contrast reaches the region of interest quickly. A set of 3D images is obtained, with mask-mode subtraction, at different time points during venous filling in the same manner as conventional X-ray venography. Postprocessing includes MIPs with and without subtraction. First-pass venography can be used for visualization of the subclavian veins and the superior vena cava, using bilateral injection (Fig. 25.3), and of the forearm veins, using unilateral injection.

Table 25.2 Protocol for contrast-enhanced MR Venography (CEMRV) Technique 1. 2D single-shot SSFP (axial and coronal) 2. T1 GRE FS (axial and coronal) 3. VIBE (axial and coronal) Inject contrast 4. TRMRV (LAO/ coronal)

Comment Free breathing; ECG gated to diastole Pre and post contrast Pre and post contrast

4 cc GAD at 4 cc/s; this can be used for timing; alternatively, this can be repeated several times at different temporal resolution, if required 5. CEMRV (coronal) Breath-hold, if needed; 3–4 postcontrast 3D measurements 6. High-spatial-resolution If blood pool agent used, increase MRV spatial resolution to 0.8 mm3 7. High-spatial-resolution If blood pool agent used, increase VIBE spatial resolution to 0.8 mm3

Fig. 25.2 Contrast-enhanced MR venography at equilibrium. Preparation of pure contrast flushed with saline. Masks are obtained before injection for further subtraction. Contrast is injected at a high rate. MRA sequence is peformed once, followed by the parenchyma sequence which is repeated twice

Time-Resolved MRV A variant of conventional CEMRV is to perform timeresolved MRV (TRMRV). With this approach, the acquisition speed is considerably increased using acceleration strategies, such as echo sharing (i.e., TWIST, TRICKS) or parallel imaging. Multiple 3D images are acquired in rapid succession at a time resolution of 3–6 s per frame so that filling of the venous system can be observed in real time. In-line mask-mode subtraction and MIP calculation can be carried out so that venous images are displayed as a cine series showing filling of arteries and veins over time. Since speed of acquisition is the priority with TRMRV, spatial resolution needs to be sacrificed resulting in lower image quality compared to conventional CEMRV. TRMRV is particularly advantageous in regions, where arteriovenous transit is rapid, such as for pulmonary venous imaging (Fig. 25.4). Another advantage of TRMRV is the ability to use smaller doses of

25

Venous Imaging: Techniques, Protocols, and Clinical Applications

355

Fig. 25.3 Contrast-enhanced MR venography at first pass. Preparation of dilute contrast media. A mask is obtained before injection for further subtraction. Contrast is injected at a slow rate. MRA sequence is repeated multiple times during injection

Fig. 25.4 Coronal dynamic TWIST MIP showing left inferior pulmonary vein stenosis post ablation

contrast. Although a conventional CEMRV acquisition may require 20–40 ml of contrast, a temporally resolved acquisition may only use as little as 3–4 ml.

Contrast Agents Extracellular Gadolinium Contrast Agents Standard extracellular contrast media are currently used for almost all CEMRA and CEMRV applications. Angiography during the first pass provides strong and selective enhancement of the vessel of interest. Extracellular agents are welltolerated with low incidence of side effects. Extracellular contrast agents are characterized by rapid extravasation of the agent from the vascular space into the interstitium, resulting in decreasing vascular enhancement over time and increasing background signal. Thus, precise bolus timing is mandatory so that contrast arrival is timed to the venous system. There is a narrow window for imaging with extracellular agents. Blood Pool Contrast Agents Blood pool contrast agents are characterized by prolonged intravascular half-life and low rate of extravasation of contrast from the vascular to the extravascular space. This feature makes blood pool agents ideally suited for venous imaging

since they reside within the venous system for prolonged periods of time, allowing high spatial resolution images to be acquired. Gadofosveset trisodium is currently the only contrast media with predominant blood pool intravascular distribution that is approved for use in patients [21]. Gadofosveset is noncovalently bound to albumin in human plasma and is primarily excreted renally. In addition to its prolonged intravascular half-life, the agent has significantly higher relaxivity compared to conventional extracellular agents, allowing it to be administered at lower doses. In plasma, gadofosveset exhibits a relaxivity at 0.5 T that is approximately six to ten times that of gadopentetate dimeglumine [22].

Iron-Based Contrast Agents Iron-based blood pool agents are coated ultrasmall superparamagnetic iron oxide (USPIO) particles with a strong T1 and T2 shortening effect. Such compounds are retained within the intravascular space in a prolonged fashion. Because of the T1 and T2 shortening effects, a dual-contrast approach can be used for imaging. First-pass and equilibrium images can be acquired using T1 gradient echo approach, similar to conventional CEMRV, and T2-weighted TSE can be employed to exploit the T2 shortening effect. Promising results from USPIO at first-pass and steady-state angiography have been published, but no USPIO is approved yet [21]. Iron-based agents have also been used to assess deep venous thrombosis [23].

Clinical Applications MRV has applications in many clinical settings. In most instances, it complements other venous imaging modalities. In some cases, it is the primary diagnostic modality. If renal

356

A.H. Davarpanah et al.

Fig. 25.5 Axial T2 proton density sequence showing DVT with hematocrit effect

function or a history of contrast allergy precludes iodinated contrast administration and body habitus prevents adequate assessment by ultrasound, MRI can be performed.

Abdominal and Lower Extremity Deep Vein Thrombosis Assessment Although considered an underestimation, the reported incidence of acute DVT is approximately 5 cases per 10,000 per year. Complications include postphlebitic syndrome, pulmonary embolism, and possibly death. As many as 300,000 deaths per year in the USA are caused by pulmonary embolism [24–26]. For diagnosing DVT, physicians have heavily relied on TOF imaging. [25–27]. This basic protocol is supplemented with fat-suppressed T2-weighted FSE or fast IR images and, where appropriate, phase-contrast images. When using these sequences, nonturbulent flowing blood is bright (white). Thrombus appears dark if not black, especially when acute. When present, acute thrombus typically appears to occlude the entire lumen, although not infrequently the lumen is only partially occluded by thrombus (Fig. 25.5). Using TOF imaging, Evans et al. [28] reported 100% sensitivity and specificity for the detection of DVT in the thigh in 61 patients when compared with venography. In the same study, the authors reported a sensitivity of 87% with a specificity of 97% for the assessment of the calves when compared with venography. Fraser et al., in a study of 101 patients with suspected DVT, compared MRDTI with venography and reported sensitivities of 97 and 100% for femoropopliteal and iliofemoral DVT, respectively [19]. Preliminary investigations are ongoing using new contrast agents in the hope of facilitating the assessment of thrombus.

Fig. 25.6 Contrast-enhanced (CE) MRA pelvis thick MIP showing right persistent sciatic vein and IVC thrombus

Li et al. [23] demonstrated the potential utility of ferumoxytol for detecting lower extremity thrombus. This iron-based compound was able to detect thrombus using both black blood (FSE) and white blood (3-D GRE) techniques. The authors suggest that although the anatomy is better seen using the white blood technique the precise extent of thrombus was more readily appreciated on the black blood images [25, 26, 29]. Spuentrup et al. [30] have demonstrated the feasibility of using a fibrin-specific contrast agent (EP-2104R) to detect thrombus within the arterial system. Such approaches clearly have applicability in the venous system as well. The traditional reference standard for assessing venous thrombus is ascending venography. The reported sensitivity and specificity for venography have been as high as 89 and 97%, respectively [30]. However, X-ray venography is associated with complications, including pain, inflammation, extravasation, induced DVT, allergy, and renal failure [24, 26]. Accordingly, there are some situations, where MRI should be primarily considered. These include the following: (a) clinical suspicion of central (pelvic) venous thrombosis (Fig. 25.6), (b) pelvic trauma, (c) ovarian vein thrombosis, (d) cryptogenic stroke, (e) pelvic congestion syndrome (Fig. 25.7), (f) hepatic and portal venous thrombosis (Fig. 25.8), (g) venous anomalies. In addition, if ultrasound is unclear or if there is potentially a large differential diagnosis, MRI is the diagnostic tool of choice. Septic thrombophlebitis occurs primarily in women who have recently undergone a cesarean section [31, 32]. Although

25

Venous Imaging: Techniques, Protocols, and Clinical Applications

Fig. 25.7 3D coronal CEMRA/V showing enlarged left gonadal vein and pelvic varicosities, suggesting pelvic congestion syndrome

Fig. 25.8 3D CE FAT-SAT MRA/V showing right gonadal (ovarian) vein thrombus and main portal vein thrombus

the left ovarian vein may be thrombosed, it is the right ovarian vein (Fig. 25.8) that is principally affected. MRI is considered more sensitive than either computed tomography (CT) or ultrasound for making this diagnosis. Cryptogenic stroke is defined as brain ischemia with no apparent etiology [33]. Studies suggest that cryptogenic stroke may occur in up to 40% of stroke patients. This number appears to be higher in younger patients. A patent foramen

357

ovale is more common in patients with cryptogenic stroke when compared with stroke of determined origin or normal controls, suggesting that right-to-left shunting of emboli is a potential cause of stroke in this population. In a small study of 16 patients with cryptogenic stroke, all had intra-atrial communication: a patent foramen ovale was present in 14, whereas 2 subjects had atrial septal defects. In four cases, MRI showed pelvic thrombus. In an additional seven cases, MRI was suspicious for prior DVT. In each instance, ultrasound was normal [33]. MRI combined with MRV efficiently clarifies cases, where multiple diagnoses are simultaneously being entertained. For example, for a patient presenting with acute calf pain with swelling and possible mass, a typical differential might include DVT, ruptured Baker’s cyst, compartment syndrome, infection (myositis), and muscle injury. Whereas ultrasound may diagnose or exclude several of the aforementioned entities, MRI/MRV can readily distinguish and diagnose each entity. MRI is potentially able to distinguish between acute and chronic thrombus. Erdman et al. [34] demonstrated that perivascular edema seen on T2-weighted images is highly predictive of acute DVT. As discussed above, acute thrombus appears bulky filling much if not all of the involved lumen. In addition, the thrombus appears dark in the TOF images. Over time, the thrombus becomes less bulky with higher signal intensity due to recanalization and lysis of the clot. In MRDTI images, with age, methemoglobin is lost resulting in decreased signal over time [35]. The sequelae of prior DVT include partial filling defects and/or Webs, vessel narrowing, thickened vessel walls, and the development of collateral circulation. MRI is capable of identifying these changes, in addition to venous anomalies. In a fashion similar to the pelvic veins, MRI/MRV is useful in the assessment of the IVC, hepatic, portal, mesenteric, and renal veins. A detailed map of the portal and hepatic venous anatomy is an important tool for surgeons to plan hepatic resection and living-related donor liver transplantation or to aid an interventional radiologist plan a TIPS procedure. Other clinical applications include arterio-portal fistula, cavernous transformation of the portal vein, Budd-Chiari, hepatic and/or portal vein thrombosis, and congenital anomalies. In the past, conventional angiography was the standard method for visualization of the portal venous anatomy. 3D contrast-enhanced MR portography has been shown to be as effective as digital subtraction angiography for assessing the portal vein [36]. Half-Fourier FSE is an unenhanced MRA technique [2, 3], which allows coronal acquisition, which is not possible with TOF imaging, especially for body MRA, and thus enables shorter 3D acquisition time [3]. In addition, this sequence is T2 weighted and the liver parenchyma becomes relatively low

358

signal intensity, resulting in good contrast between the portal vein and liver parenchyma; however, the long TE decreases the signal at the portal confluence, where higher flow velocity might cause flow void in comparison with images acquired with true SSFP [3]. Recently, Ono et al. demonstrated high accuracy and reproducibility of noncontrast- enhanced MRV using both the FR-FBI and the SPADE techniques for detecting DVT, as compared using conventional X-ray venography as the reference standard. Forty-one legs of 32 consecutive patients in 32 consecutive patients [9].

Klippel-Trénaunay Syndrome and Parkes Weber Syndrome Klippel-Trénaunay syndrome is a congenital disorder classically characterized by three findings: a port-wine stain (nevus flammeus), abnormal venous structures (such as varicosities and venous malformations), and osseous and soft-tissue hypertrophy. In 1907, Frederick Parkes Weber noted similar findings in association with arteriovenous malformations. This entity is referred to as Parkes Weber or KlippelTrénaunay–Weber syndrome. The diagnosis of KlippelTrénaunay syndrome can be made when any two of the three features are present [37, 38]. Complications associated with Klippel-Trénaunay syndrome are most often related to the vascular system. Such complications include stasis dermatitis, thrombophlebitis, and cellulitis. More serious complications include deep venous thrombosis, pulmonary embolism, coagulopathy, and congestive heart failure (in patients with associated arteriovenous malformations). Various imaging techniques can be used in the diagnosis of suspected Klippel-Trénaunay syndrome. Both plain radiography and CT have been used to depict phleboliths that suggest the presence of a venous malformation and disease-related complications, such as deep venous thrombosis. Sonography with Doppler capabilities can be used to assess the condition of the venous system within an affected extremity. MRI can be used to evaluate extremity hypertrophy and vascular malformations in these patients. MR arteriography and MRV can be used to define both the type and extent of vascular malformations in Klippel-Trénaunay syndrome [39]. Specifically, MRI allows differentiation of low-flow (venous) from high-flow (arteriovenous) malformations. The venous malformations typically associated with Klippel-Trénaunay syndrome are hyperintense on T2-weighted images and lack flow voids. The arteriovenous malformations associated with Parkes Weber syndrome are high flow because they are fed by the arterial system and therefore typically have flow voids. Occasionally, however, conventional angiography or venography is needed to define the vascular lesions associated with these conditions [39, 40].

A.H. Davarpanah et al.

Pelvic Congestion Syndrome Chronic pelvic pain with associated ovarian vein varicosities is called pelvic congestion syndrome (PCS). It can present with atypical and intractable pelvic pain usually in women of childbearing age. The etiology of PCS, while poorly understood, may be from incompetent valves in ovarian or pelvic veins, with associated hormonal factors. Ovarian vein dilatation is seen in 10% of women, up to 60% of whom may develop PCS [41]. Noninvasive methods, such as MRI, contrast-enhanced CT, and duplex ultrasound, may be used to depict dilated veins. Catheter venography, however, remains the gold standard in accurately depicting flow dynamics, but is an invasive procedure associated with radiation exposure. Conventional MRI and computed tomography yield good cross-sectional information; however, they fail to demonstrate the flow dynamics. MRV has been used to study ovarian vein incompetence, but lacks sufficient temporal resolution to demonstrate retrograde ovarian vein filling. In a study of 23 female patients with signs and symptoms of pelvic venous congestion comparing MRV with phlebography as gold standard, Asciutto et al. found that MRV was highly sensitive (88%), however demonstrated lower specificity (67%) in detecting congested ovarian veins [42]. In this study, the presence of retrograde ovarian vein filling was demonstrated in only 66% cases by MRV when compared to gold standard phlebography. Time-resolved MRA (TR-MRA) has been proven to be a quick and noninvasive technique that allows visualization of the physiologic blood flow dynamics and is predicted to be helpful for the detection of PCS because of its presumed ability to accurately determine whether anterograde or retrograde flow in the ovarian vein is present. In two recent studies, TR-MRA has shown utility in demonstrating ovarian vein reflux and diagnosing PCS [43, 44]. Pandey et al. used 4D time-resolved angiography with central keyhole (TRAK) acquisition MR technique as a noninvasive alternate modality for diagnosing pelvic congestion syndrome. The technique achieves a high temporal resolution by sharing k-space and acquiring only the central lines (keyhole) repeatedly over a short period of time. Faster image acquisition allows for improved visualization of flow dynamics and vascularity. This technique is extremely useful for depicting early reversal of venous flow and incompetence of the ovarian vein. It also has the capability to better resolve the arterial phase helping in the detection of arterial feeders to the lesion and any underlying arteriovenous malformation [43]. Nutcracker Syndrome Nutcracker syndrome refers to compression of the left renal vein (LRV) by the superior mesenteric artery and aorta. Patients typically present with left flank pain and associated symptoms of pelvic congestion. Hematuria is frequently

25

Venous Imaging: Techniques, Protocols, and Clinical Applications

present, and vulvar or lower extremity varices are seen in a subset of patients. Diagnosing nutcracker phenomenon can be done by anatomic assessment, with finding LRV compression between the aorta and SMA, a ratio between the distended and narrowed portions of the LRV of >1.5 for the anteroposterior diameter and a sharp rather than 90° branching angle of the SMA from the aorta, all indicating the phenomenon. The standard (although invasive) method of physiologic assessment is measuring the reno-caval pressure gradient during phlebography. If the pressure gradient is >3 mmHg, the patient is considered to have venous hypertension. A noninvasive alternative is to measure the peak velocities in the distended and narrowed segments of the LRV using Doppler ultrasonography, and then calculating a velocity ratio. However, quantitative analysis of the venous flow by Doppler sonography has potential pitfalls. It is difficult to obtain a spectral Doppler sampling from an entrapped segment of the LRV in the aortomesenteric space. TR-MRA findings compatible with nutcracker syndrome include the angle of the superior mesenteric artery (SMA) to the aorta (normal > 60°), LRV diameter on the basis of coronal maximum-intensity-projection images, and continuity of flow from the LRV into the IVC [44]. Wong et al. showed the usefulness of FSE T2-weighted MRI in diagnosing nutcracker syndrome. Hyperintense LRV on MR FSE T2WI indicated marked flow stagnation, and this inferred the presence of venous hypertension and suggested a diagnosis of nutcracker syndrome [45].

May-Thurner Syndrome Iliac vein compression syndrome (IVCS), also termed as May-Thurner syndrome or Cockett syndrome, is caused by compression of the left common iliac vein between the right common iliac artery and overlying vertebrae. Pulsatile wall compression of the vein induces replacement of normal intima and media of vein is largely replaced by well-organized connective tissue covered with endothelium that could cause DVT or venous hypertension without thrombosis in the left lower extremity. The syndrome most commonly presents as DVT; however, patients also can present with left-sided leg pain, swelling, and venous insufficiency without a thrombosis, but these occur less frequently [46, 47]. Conventional venography is the gold standard for IVCS diagnosis; however, different modalities have been shown to demonstrate the compression successfully as well. Different MRV types are available for evaluation of lower extremity venous system with their limitations. TOF technique is susceptible to flow artifacts and saturation and long acquisition times are needed. Basically, two different contrast-enhanced techniques using subtraction of arterial phase from the venous-arterial equilibrium phase and direct

359

contrast-enhanced venous phase have been investigated. Capillary passage of extracellular contrast during the arterial phase reduces venous return resulting in a reduced venous signal. Furthermore, a selective venous enhancement without concomitant arterial display is often difficult to obtain in peripheral vascular segments in this technique. Direct contrast-injection 3D MRV seems to overcome these problems [48]. Ruehm et al. reported that direct CEMRV has sensitivity and specificity values over 90% compared to conventional venography in the evaluation of varicosities, postthrombotic changes of the lower extremity venous system [49]. Direct CEMRV performed with diluted and long-lasting automated injection of the extracellular contrast agent from the pedal veins seems to be an improved imaging technique of lower extremity deep venous system. In terms of equipment and application, this technique is simple and feasible at any center where lower extremity contrast-enhanced 3D MRA is being performed in routine clinical practice while providing a practical approach with better image quality.

Renal Vein Thrombosis The term renal vein thrombosis (RVT) is used to describe the presence of thrombus in the major renal veins or their tributaries. This condition may either present with acute symptoms or go unnoticed because of lack of symptoms until a complication like pulmonary embolism or worsening renal function draws attention to it. The etiology of RVT is variable, but can be extrinsic or intrinsic. The intrinsic form is triggered by an intrarenal thrombotic event precipitated by acidosis, hemoconcentration, or arteriolar constriction. In adults, this process is typically the result of an underlying renal neoplasm. Additional intrinsic causes include membranous glomerulonephritis, pyelonephritis, amyloidosis, polyarteritis nodosa, sickle cell anemia, cardiac disease/low flow states, trauma, diabetic nephropathy, lupus nephropathy, coagulopathy, dehydration, or trauma. Extrinsic processes include umbilical vein catheterization, extension of IVC thrombosis, pancreatitis, retroperitoneal fibrosis, metastasis, and pancreatic tail carcinoma. MRI characterizes renal vein and IVC involvement by RCC with a higher accuracy for staging than computed tomography [50]. Standard sequences supplemented with 3D gadolinium-enhanced images provide high contrast and spatial resolution. In one study, MRA could delineate the entire course of the renal vessels in 88% of cases compared to 58% with Color Doppler ultrasound and 43% on spin-echo MRI. Similarly, the anatomic variants, vessel displacement, collateral circulation, and neoplastic vessel infiltration were demonstrated more accurately by MRA [51]. Laissy et al. study the performance of gadolinium-enhanced TOF MRV in 26 patients with RCC and tumor thrombus. For detection of venous thrombus, the sensitivity and specificity are

360

A.H. Davarpanah et al.

100 and 96%, respectively [52]. Using gadolinium-based contrast for evaluation of bland versus neoplastic thrombus, sensitivity and specificity of 89 and 96%, respectively, are achieved [52].

Portal Vein Obstruction Portal vein occlusion may occur with tumor or bland thrombus or it may occur secondary to extrinsic compression. Infectious, inflammatory, and malignant conditions of the abdomen are the most commonly implicated local risk factors for the development of portal vein thrombosis. In adults, acute appendicitis, cholecystitis, acute necrotizing pancreatitis, cholangitis, diverticulitis, and perforated peptic ulcers all have been reported to cause portal vein thrombosis. Extrinsic compression most commonly is due to metastatic liver disease but also may occur with benign masses, such as hemangiomas. Lobar or segmental portal vein obstruction by tumor may cause discrete wedge-shaped regions of increased intensity on T2-weighted and Gd-enhanced images [53]. Tumor and bland thrombus can be distinguished from each other by the observation that tumor thrombus is higher in signal on T2-weighted images, is intermediate in signal intensity on TOF, and enhances with Gd. In comparison, bland thrombus is low in signal intensity on both T2-weighted and TOF images and does not enhance with Gd. Portal vein thrombosis can be demonstrated using bright blood techniques (e.g., TOF and three-dimensional GRE sequences and black blood techniques, e.g., spin echo with superior and inferior saturation pulses). After administration of Gd, parenchymal manifestations of portal vein compromise are seen on MRI as transient areas of increased enhancement during the arterial phase that correspond to areas of decreased portal perfusion during the portal venous phase. This phenomenon results from the compensatory autoregulatory mechanism geared toward maintaining the net parenchymal perfusion by augmenting hepatic arterial blood supply [54]. Several authors have demonstrated that contrast-enhanced threedimensional MRA is as accurate as digital subtraction angiography (DSA) in assessing the portal venous system and determining surgical respectability in patients with pancreatobiliary tumors [53, 55].

Upper Extremity and Central Vein Evaluation Primary indications for MRV/MRI of the chest and upper extremity include defining SVC and central venous obstruction and invasion, venous thrombosis, stenosis, occlusion, and venous access planning. In patients presenting with abrupt swelling of the arms or facial swelling suggesting SVC syndrome, both noncontrast- and contrast-enhanced MRV are essential diagnostic tools.

Fig. 25.9 Coronal SHARP post contrast showing right internal jugular vein thrombus

Central venous occlusion often results in congestion, edema, and venous hypertension. The underlying cause varies, but previous radiation therapy, extrinsic mass/compression, or inflammation frequently is present. Common central venous occlusive conditions include superior vena cava (SVC) and IVC syndrome. SVC syndrome may be the result of a complete or partial occlusion of the SVC or its tributaries. Eighty to ninety percent of secondary obstructions are neoplastic in origin. Common offending neoplasms are bronchogenic carcinoma (greater than 50%), lymphoma, and mediastinal tumors. Granulomatous diseases, aneurysms, constrictive pericarditis, and substernal goiter are among the non-neoplastic causes. Catheter-induced occlusion or stenosis has become a more frequent cause of SVC syndrome. IVC syndrome can have intrinsic or extrinsic etiologies. Intrinsic caval occlusion typically has a neoplastic etiology (leiomyoma, leiomyosarcoma, and endothelioma), but may be non-neoplastic (congenital membrane). Extrinsic obstruction often occurs at the mid-IVC as a result of enlarged lymph nodes or an adjacent retroperitoneal, renal, pancreatic, or hepatic mass. Functional obstruction can result from a pregnant uterus, valsalva maneuver, or supine positioning with a large abdominal mass. Whereas ultrasound readily detects acute thrombus within the arms, MRV more readily visualizes more central thrombus (Figs. 25.9 and 25.10). Many patients with end-stage renal disease undergo both short-term and long-term dialysis via indwelling catheters. Initially, a multitude of suitable access sites are available for catheter placement. However, over time, the number and quality of accessible sites diminish due to multiple stenoses and thrombosis associated with long-standing catheterization. In such patients, MRV readily evaluates potential access sites.

25

Venous Imaging: Techniques, Protocols, and Clinical Applications

Fig. 25.10 Partition CEMRA/V showing SVC occlusion secondary to thrombus

Fig. 25.11 STARFIRE sequence showing DVT

Finn et al. [56], using only TOF imaging [10] in three orthogonal planes, reported excellent correlation between MRV and contrast venography. The authors acknowledged that collateral vessels were better appreciated with contrast venography; otherwise, there was essentially complete agreement between the two imaging modalities. The major disadvantage of TOF acquisitions is the relatively long acquisition time. In addition, these techniques suffer from numerous artifacts, including decreased signal intensity in the vessel due to slow flow, loss of signal secondary to turbulent flow, and pulsation artifacts, especially in the SVC. Koktzoglou’s noncontrast STARFIRE technique [17] has been utilized for both lower limb DVT and central venous mapping. Initial results look promising (Fig. 25.11). Temporally resolved MRA/MRV may prove to be a useful alternative to standard gadolinium-enhanced MRV in patients with significant renal impairment as significant dose reductions of gadolinium are possible [4–6]. Studies concerning

361

NSF suggest that in addition to using a more stable gadoliniumcontaining compound, dose reduction also reduces the probability of acquiring NSF, suggesting that it may be possible to safely image patients with poor renal function [4]. Kim et al., in a retrospective analysis of 27 consecutive patients undergoing both conventional gadolinium-enhanced MRV and temporally resolved MRV, demonstrated that the addition of the time-resolved images improved specificity in the detection of venous occlusions and improved reader confidence while actually reducing image interpretation time [56– 58]. Unfortunately, the time-resolved imaging, while sensitive for occlusion (95%), was only moderately specific (56%). More recently, several investigators have reported findings with SSFP imaging as a means to assess the central veins. SSFP has several potential advantages over traditional TOF imaging [10], including (a) faster acquisition times due to the shortened repetition time, (b) less flow artifacts due to the incorporation of balanced gradients in the pulse sequence, and (c) better contrast to noise between flowing blood and the adjacent stationary tissue. Like TOF imaging, SSFP imaging does not require intravenous contrast. A primary cause for the reduced sensitivity of the SSFP pulse sequence appeared to be the variable signal intensity of thrombus over time [10]. Acute thrombus displayed increased signal making it relatively isointense to blood. Older thrombus was more readily detected presumably due to its longer T1 value. If gadoliniumbased contrast agents are unable to be administered, we currently use a combination of axial and sagittal, true FISP, black blood (e.g., Haste or double-inversion IR), and STARFIRE images [17]. In addition, cardiosynchronous acquisitions are usually obtained in the axial, saggital, and coronal planes through the SVC. Although these studies require longer acquisition and interpretation times, clinically useful information is routinely obtained noninvasively.

Pulmonary Vein Evaluation MRA has become an established modality for pulmonary vein (PV) evaluation prior to and following RF ablation in patients with atrial fibrillation. Contrast material-enhanced MR angiographic techniques are primarily used for PV evaluation because of their high spatial resolution and reliability [59]. In patients who are pregnant or at risk of NSF, the use of gadolinium-based contrast material is contraindicated. For these patients, unenhanced MR angiographic techniques must be of equal quality and reliability as the contrastenhanced MR angiographic techniques. Balanced SSFP imaging techniques have been used to evaluate the thoracic vasculature, including the coronary arteries and thoracic aorta, because of their inherent high signal-to-noise and contrast-to-noise ratios. While singleshot techniques are rapid and useful for the urgent evaluation

362

of patients with suspected acute aortic syndromes or patients incapable of extended breath holding, 3D techniques with near-isotropic resolution are preferred for evaluation of smaller vessels, such as PVs. Free-breathing, navigator-gated 3D SSFP techniques have been developed to evaluate the coronary and renal arteries. Francois et al. [59] demonstrated that unenhanced freebreathing, T2-segmented 3D SSFP imaging sequence for PV assessment is comparable with time-resolved 3D CEMRA in patients in the assessment of patients prior to and following RF ablation for atrial fibrillation. Qualitatively, the images obtained were not different between the techniques. PV variants detected by using CEMRA were identified at 3D SSFP imaging as well. No significant difference in PV measurements was found between the 3D SSFP and contrastenhanced MR angiograms, including assessment of PV stenosis (Fig. 25.4). While both CT and MR angiography are commonly used for PV imaging [59], CT angiography has substantial drawbacks, including the use of ionizing radiation and nephrotoxic [4–6] contrast agents. Because patients with atrial fibrillation return for follow-up examinations at frequent short-term intervals prior to and following RF ablation, MRA is frequently preferred. Advances in MR sequences have made MRA an excellent noninvasive method of evaluating the vasculature. MRA is particularly suited for evaluating the PVs, especially with the use of parallel imaging and time-resolved techniques to offset the effects of cardiac motion. MRA, when combined with other nonangiographic MRI techniques, can be used as the primary or sole imaging modality in the evaluation of patients with atrial fibrillation prior to and following RF ablation. In particular, TRMRV can produce purely pulmonary venous-phase images, free from overlap of other vascular structures, such as the pulmonary arteries or aorta. An important point is that pulmonary vein ostia have to be measured accurately using multiplanar reformatting to generate orthogonal dimensions. In addition to providing morphologic information on the PVs, cardiac MR can be used to assess cardiac function and help detect evidence of myocardial scar. Because of the risk of causing atrioesophageal fistulas during RF ablation, it is also important to define the relationship between the atrium and esophagus with cross-sectional imaging prior to therapy. Because the contrast-enhanced technique used in this study is a subtraction technique, to optimize visualization of the vasculature, extravascular structures are rarely noted. However, the 3D SSFP technique permits evaluation of the vasculature in addition to extravascular anatomy, including the detection of extracardiac pathologic anomalies, such as cardiac masses, foregut cysts, lung nodules, lymphadenopathy, and hiatal hernias.

A.H. Davarpanah et al.

Importantly, secondary complications of the ablation procedure can be readily assessed with both noncontrast and contrast-enhanced MR PV mapping, including left atrial appendage thrombus.

Future Directions Imaging at field strengths higher than the conventional 1.5 T, such as with the increasingly available 3-T systems, has different benefits and drawbacks for the various nonenhanced MR techniques [3]. The intrinsically higher signal-to-noise ratio at 3 T does lend itself well to the use of parallel imaging. Implementation of parallel imaging can be particularly beneficial to nonenhanced MR angiographic methods by reducing acquisition times and consequently decreasing undesirable blurring and motion artifacts. Most applications of parallel imaging suffer a trade-off of reduced signal for shorter scan times; however, partial Fourier FSE with parallel imaging allows for compensating benefits that include a reduction in T2 blurring. With shorter acquisition times, multistation imaging with nonenhanced methods becomes feasible.

Conclusion Using a myriad of techniques, MRV has proven to be a useful tool in the assessment of venous abnormalities. Developments on the horizon include techniques that provide time-resolved imaging for assessment of flow dynamics by using nonenhanced approaches.

References 1. Prince MR. Gadolinium-enhanced MR aortography. Radiology. 1994;191(1):155–64. 2. Spritzer CE. Progress in MR imaging of the venous system. Perspect Vasc Surg Endovasc Ther. 2009;21(2):105–16. 3. Miyazaki M, Lee VS. Nonenhanced MR angiography. Radiology. 2008;248(1):20–43. 4. Prince MR, Arnoldus C, Frisoli JK. Nephrotoxicity of high-dose gadolinium compared with iodinated contrast. J Magn Reson Imaging. 1996;6(1):162–6. 5. Marckmann P, Skov L, Rossen K, Heaf JG, Thomsen HS. Casecontrol study of gadodiamide-related nephrogenic systemic fibrosis. Nephrol Dial Transplant. 2007;22(11):3174–8. 6. Cowper SE, Robin HS, Steinberg SM, Su LD, Gupta S, LeBoit PE. Scleromyxoedema-like cutaneous diseases in renal-dialysis patients. Lancet. 2000;356(9234):1000–1. 7. Guerrero A, Montes R, Munoz-Terol J, et al. Peripheral arterial disease in patients with stages IV and V chronic renal failure. Nephrol Dial Transplant. 2006;21(12):3525–31. 8. Ito K, Koike S, Shimizu A, et al. Portal venous system: evaluation with unenhanced MR angiography with a single-breath-hold

25

9.

10.

11.

12. 13. 14.

15.

16.

17.

18.

19.

20. 21. 22.

23.

24.

25.

26.

27.

28.

Venous Imaging: Techniques, Protocols, and Clinical Applications ECG-synchronized 3D half-Fourier fast spin-echo sequence. Ajr. 2008;191(2):550–4. Ono A, Murase K, Taniguchi T, et al. Deep venous thrombosis: diagnostic value of non-contrast-enhanced MR venography using electrocardiography-triggered three-dimensional half-Fourier FSE. Magn Reson Med. 2010;64(1):88–97. Wehrli FW, Shimakawa A, Gullberg GT, MacFall JR. Time-offlight MR flow imaging: selective saturation recovery with gradient refocusing. Radiology. 1986;160(3):781–5. Kirchhof K, Welzel T, Jansen O, Sartor K. More reliable noninvasive visualization of the cerebral veins and dural sinuses: comparison of three MR angiographic techniques. Radiology. 2002;224(3): 804–10. Liauw L, van Buchem MA, Spilt A, et al. MR angiography of the intracranial venous system. Radiology. 2000;214(3):678–82. Laissy JP, Trillaud H, Douek P. MR angiography: noninvasive vascular imaging of the abdomen. Abdom Imaging. 2002;27(5):488–506. Carr JC, Simonetti O, Bundy J, Li D, Pereles S, Finn JP. Cine MR angiography of the heart with segmented true fast imaging with steady-state precession. Radiology. 2001;219(3):828–34. Liu X, Berg N, Sheehan J, et al. Renal transplant: nonenhanced renal MR angiography with magnetization-prepared steady-state free precession. Radiology. 2009;251(2):535–42. Stafford RB, Sabati M, Mahallati H, Frayne R. 3D non-contrastenhanced MR angiography with balanced steady-state free precession Dixon method. Magn Reson Med. 2008;59(2):430–3. Edelman RR, Koktzoglou I. Unenhanced flow-independent MR venography by using signal targeting alternative radiofrequency and flow-independent relaxation enhancement. Radiology. 2009;250(1):236–45. Miyazaki M, Sugiura S, Tateishi F, Wada H, Kassai Y, Abe H. Noncontrast-enhanced MR angiography using 3D ECG-synchronized half-Fourier fast spin echo. J Magn Reson Imaging. 2000;12(5): 776–83. Fraser DG, Moody AR, Morgan PS, Martel AL, Davidson I. Diagnosis of lower-limb deep venous thrombosis: a prospective blinded study of magnetic resonance direct thrombus imaging. Annals of internal medicine. 2002;136(2):89–98. Spritzer CE, Blinder RA. Practical aspects of vascular imaging using MRI. Crit Rev Diagn Imaging. 1990;31(2):145–85. Bremerich J, Bilecen D, Reimer P. MR angiography with blood pool contrast agents. European radiology. 2007;17(12):3017–24. Baxter AB, Melnikoff S, Stites DP, Brasch RC. AUR Memorial Award 1991. Immunogenicity of gadolinium-based contrast agents for magnetic resonance imaging. Induction and characterization of antibodies in animals. Investigative radiology. 1991;26(12):1035–40. Li W, Salanitri J, Tutton S, et al. Lower extremity deep venous thrombosis: evaluation with ferumoxytol-enhanced MR imaging and dual-contrast mechanism--preliminary experience. Radiology. 2007;242(3):873–81. Fowkes FJ, Price JF, Fowkes FG. Incidence of diagnosed deep vein thrombosis in the general population: systematic review. Eur J Vasc Endovasc Surg. 2003;25(1):1–5. Spritzer CE, Sostman HD, Wilkes DC, Coleman RE. Deep venous thrombosis: experience with gradient-echo MR imaging in 66 patients. Radiology. 1990;177(1):235–41. Spritzer CE, Norconk JJ, Jr., Sostman HD, Coleman RE. Detection of deep venous thrombosis by magnetic resonance imaging. Chest. 1993;104(1):54–60. Sampson FC, Goodacre SW, Thomas SM, van Beek EJ. The accuracy of MRI in diagnosis of suspected deep vein thrombosis: systematic review and meta-analysis. European radiology. 2007; 17(1):175–81. Evans AJ, Sostman HD, Knelson MH, et al. 1992 ARRS Executive Council Award. Detection of deep venous thrombosis: prospective

363

29.

30.

31.

32.

33.

34. 35.

36.

37.

38. 39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

comparison of MR imaging with contrast venography. Ajr. 1993; 161(1):131–9. Erdman WA, Jayson HT, Redman HC, Miller GL, Parkey RW, Peshock RW. Deep venous thrombosis of extremities: role of MR imaging in the diagnosis. Radiology. 1990;174(2):425–31. Spuentrup E, Botnar RM, Wiethoff AJ, et al. MR imaging of thrombi using EP-2104R, a fibrin-specific contrast agent: initial results in patients. European radiology. 2008;18(9):1995–2005. Bertsch NM, Mastrobattista JM, Kawashima A, Kramer LA. Antepartum bilateral ovarian vein thrombosis: magnetic resonance imaging diagnosis. Am J Perinatol. 1997;14(10):597–9. Twickler DM, Setiawan AT, Evans RS, et al. Imaging of puerperal septic thrombophlebitis: prospective comparison of MR imaging, CT, and sonography. Ajr. 1997;169(4):1039–43. Rordorf G, Koroshetz WJ, Copen WA, et al. Regional ischemia and ischemic injury in patients with acute middle cerebral artery stroke as defined by early diffusion-weighted and perfusion-weighted MRI. Stroke. 1998;29(5):939–43. Erdman WA, Parkey RW. MR imaging of deep venous thrombosis. Ajr. 1990;155(4):897. Kubik-Huch RA, Hebisch G, Huch R, Hilfiker P, Debatin JF, Krestin GP. Role of duplex color Doppler ultrasound, computed tomography, and MR angiography in the diagnosis of septic puerperal ovarian vein thrombosis. Abdom Imaging. 1999;24(1):85–91. Shimada K, Isoda H, Okada T, et al. Non-contrast-enhanced MR portography with time-spatial labeling inversion pulses: comparison of imaging with three-dimensional half-fourier fast spin-echo and true steady-state free-precession sequences. J Magn Reson Imaging. 2009;29(5):1140–6. Jacob AG, Driscoll DJ, Shaughnessy WJ, Stanson AW, Clay RP, Gloviczki P. Klippel-Trenaunay syndrome: spectrum and management. Mayo Clin Proc. 1998;73(1):28–36. Cha SH, Romeo MA, Neutze JA. Visceral manifestations of Klippel-Trenaunay syndrome. Radiographics. 2005;25(6):1694–7. Vilanova JC, Barcelo J, Smirniotopoulos JG, et al. Hemangioma from head to toe: MR imaging with pathologic correlation. Radiographics. 2004;24(2):367–85. Elsayes KM, Menias CO, Dillman JR, Platt JF, Willatt JM, Heiken JP. Vascular malformation and hemangiomatosis syndromes: spectrum of imaging manifestations. Ajr. 2008;190(5):1291–9. Ganeshan A, Upponi S, Hon LQ, Uthappa MC, Warakaulle DR, Uberoi R. Chronic pelvic pain due to pelvic congestion syndrome: the role of diagnostic and interventional radiology. Cardiovasc Intervent Radiol. 2007;30(6):1105–11. Asciutto G, Mumme A, Marpe B, Koster O, Asciutto KC, Geier B. MR venography in the detection of pelvic venous congestion. Eur J Vasc Endovasc Surg. 2008;36(4):491–6. Pandey T, Shaikh R, Viswamitra S, Jambhekar K. Use of time resolved magnetic resonance imaging in the diagnosis of pelvic congestion syndrome. J Magn Reson Imaging. 2010;32(3):700–4. Kim CY, Miller MJ, Jr., Merkle EM. Time-resolved MR angiography as a useful sequence for assessment of ovarian vein reflux. Ajr. 2009;193(5):W458-63. Wong HI, Chen MC, Wu CS, et al. The usefulness of fast-spin-echo T2-weighted MR imaging in Nutcracker syndrome: a case report. Korean J Radiol. 2010;11(3):373–7. Oguzkurt L, Ozkan U, Tercan F, Koc Z. Ultrasonographic diagnosis of iliac vein compression (May-Thurner) syndrome. Diagnostic and interventional radiology (Ankara, Turkey). 2007;13(3):152–5. Wolpert LM, Rahmani O, Stein B, Gallagher JJ, Drezner AD. Magnetic resonance venography in the diagnosis and management of May-Thurner syndrome. Vasc Endovascular Surg. 2002;36(1):51–7. Gurel K, Gurel S, Karavas E, Buharalioglu Y, Daglar B. Direct contrast-enhanced MR venography in the diagnosis of May-Thurner Syndrome. Eur J Radiol. 2010.

364 49. Ruehm SG, Wiesner W, Debatin JF. Pelvic and lower extremity veins: contrast-enhanced three-dimensional MR venography with a dedicated vascular coil-initial experience. Radiology. 2000;215(2): 421–7. 50. Heiss SG, Shifrin RY, Sommer FG. Contrast-enhanced three-dimensional fast spoiled gradient-echo renal MR imaging: evaluation of vascular and nonvascular disease. Radiographics. 2000;20(5):1341–52; discussion 53–4. 51. Pfluger T, Czekalla R, Hundt C, et al. MR angiography versus color Doppler sonography in the evaluation of renal vessels and the inferior vena cava in abdominal masses of pediatric patients. Ajr. 1999;173(1):103–8. 52. Laissy JP, Menegazzo D, Debray MP, et al. Renal carcinoma: diagnosis of venous invasion with Gd-enhanced MR venography. European radiology. 2000;10(7):1138–43. 53. McFarland EG, Kaufman JA, Saini S, et al. Preoperative staging of cancer of the pancreas: value of MR angiography versus conventional angiography in detecting portal venous invasion. Ajr. 1996; 166(1):37–43.

A.H. Davarpanah et al. 54. Morrin MM, Pedrosa I, Rofsky NM. Magnetic resonance imaging for disorders of liver vasculature. Top Magn Reson Imaging. 2002;13(3):177–90. 55. Smedby O, Riesenfeld V, Karlson B, et al. Magnetic resonance angiography in the resectability assessment of suspected pancreatic tumours. European radiology. 1997;7(5):649–53. 56. Finn JP, Zisk JH, Edelman RR, et al. Central venous occlusion: MR angiography. Radiology. 1993;187(1):245–51. 57. Tanju S, Sancak T, Dusunceli E, Yagmurlu B, Erden I, Sanlidilek U. Direct contrast-enhanced 3D MR venography evaluation of upper extremity deep venous system. Diagnostic and interventional radiology (Ankara, Turkey). 2006;12(2):74–9. 58. Thornton MJ, Ryan R, Varghese JC, Farrell MA, Lucey B, Lee MJ. A three-dimensional gadolinium-enhanced MR venography technique for imaging central veins. Ajr. 1999;173(4):999–1003. 59. Francois CJ, Tuite D, Deshpande V, Jerecic R, Weale P, Carr JC. Pulmonary vein imaging with unenhanced three-dimensional balanced steady-state free precession MR angiography: initial clinical evaluation. Radiology. 2009;250(3):932–9.

Pediatric MR Angiography: Principles and Applications

26

Bharathi D. Jagadeesan and David N. Loy

Introduction Magnetic resonance angiography (MRA) is perhaps the most valuable in children when compared to all other patient groups. While CT angiography has made vast strides in imaging the neurovascular system in adults, these advances do not lend themselves to easy application in the pediatric patient population primarily due to the risk from the considerable radiation exposure that CT angiographic procedures entail. Likewise, conventional catheter angiography procedures also result in radiation exposure. While many CTA and conventional digital subtraction angiography (DSA) procedures have been shown to result in radiation exposure which is less than the threshold for deterministic effects such as epilation or erythema [1–3], they nevertheless result in an increased risk for malignancy secondary to stochastic effects. Although the risk from any single procedure may be low, children with neurovascular disorders are often subject to multiple CTAs or angiograms in a year, resulting in large cumulative radiation doses. The estimated incidence of new cancers resulting from these procedures was 890.6/100,000 exposed males and 1,222.5/100,000 exposed females [4]. Swoboda et al. [4] found that the average 9-year-old patient with a neurovascular accumulates radiation exposure equivalent to 235 frontal and 177 lateral skull radiographs over the year from CT and DSA procedures. Radiation risks are higher for the youngest children and for females. Radiation risks for newborn males and females per unit dose of radiation are approximately three and six times higher than for adult patients, respectively [5, 6]. Most CT angiography studies are now performed on multidetector CT scanners. Likewise, multiphase CT imaging also results in increased radiation doses [7]. Improvements in

B.D. Jagadeesan, MD () • D.N. Loy, MD, PhD Washington University School of Medicine, Barnes-Jewish Hospital, Mallinckrodt Institute of Radiology, Saint Louis, MO, USA e-mail: [email protected]

image quality with these techniques come at a cost of increased radiation dose. Recent efforts by vendors to develop better radiation dose estimates in children have been an important step in support of the “Image Gently” campaign [8]. MRA should play an increasingly crucial role in the diagnosis, classification, prognostication, and follow-up of pediatric neurovascular disorders to reduce radiation exposure in children. Numerous unique challenges confront MRA in this developing patient population. These include the small size of intracranial vascular structures, the hyper-dynamic circulation in children with various multisystem disorders, and the immature or compromised renal function that is encountered in certain pediatric patient populations. Pediatric MRA techniques should possess high spatial and temporal resolution and should also be feasible without the administration of intravenous contrast when possible. In this chapter, we briefly review the basic principles of MRA as they specifically pertain to pediatric neuroimaging. We can then proceed to classifying and reviewing the common pediatric neurovascular disorders and the role of MRA techniques in each.

Principles of Magnetic Resonance Angiography in Children All MRA techniques strive to accentuate the contrast ratio between the intraluminal blood and the surrounding background tissue. They aim to achieve this by accentuating the signal originating from the flowing blood and by attenuating the signal originating from the stationary tissues. Some techniques rely on a combination of both these endeavors. Hence, most neurovascular imaging can be considered to be “white blood imaging” where the flowing blood is hyperintense. Black blood imaging which is popular elsewhere in the body for vascular imaging plays almost no role in routine neurovascular imaging with the notable exception of MR bold venography, which is useful in select scenarios.

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_26, © Springer Science+Business Media, LLC 2012

365

366

Techniques in Pediatric Magnetic Resonance Imaging Noncontrast Magnetic Resonance Angiography of the Head and Neck The basic noncontrast MRA techniques are time-of-flight MRA, phase contrast MRA, and BOLD MRI venography or susceptibility weighted imaging (SWI).

Time-of-Flight MRA Time-of-flight (TOF) angiography relies on the enhancement produced by imaging the inflow of fresh unsaturated protons in the blood stream into a slice or slab of tissue that has been presaturated with repetitive RF pulses with short TRs [9]. When compared to 2D TOF MRA, 3D TOF MRA offers much higher spatial resolution but it is relatively insensitive to slow flow. Modifications to 3D TOF imaging such as Multiple Overlapping Thin Slice Acquisition (MOTSA), addition of magnetization transfer, and the use of ramped flip angles through the imaging volume can considerably improve the signal in slow flow lesions [10–12]. Additionally, a few authors have also advocated the intravenous administration of a small dose (0.5 mL) of gadolinium, a T1 shortening agent, in order to improve contrast in distal intracranial branches [13]. Nevertheless, TOF MRA suffers from inherently poor sensitivity to turbulent blood flow in vascular beds with complex anatomy. Therefore, time-of-flight imaging is of limited use in complex head and neck vascular malformations or for imaging the carotid arteries. TOF MRA can also be misleading in patients with the evaluation of thrombosed vessels since thrombi with short T1 signal will appear hyperintense similar to flowing blood, making it impossible to distinguish between the two [14]. Despite these caveats, the lack of a need for intravenous contrast administration and the excellent spatial resolution with 3D techniques makes TOF angiography an attractive MRA technique for imaging the circle of Willis in children. Unlike techniques dependent on intravenous contrast administration, TOF can also be safely performed in children with renal dysfunction. TOF MRA studies can also be repeated multiple times in the same session or the same day (i.e., preand postoperatively) when indicated. This is invaluable for the study of children, especially those who cannot be sedated or those with contraindications to intravenous contrast including sickle cell disease. The more widespread availability of 3 Tesla scanners and the increase in signal-to-noise that is inherent to higher field strengths has significantly improved the quality of TOF MRA in children. Recently, several investigators have also described techniques for the simultaneous acquisition of TOF MR arteriographic images and susceptibility weighted MR venographic

B.D. Jagadeesan and D.N. Loy

imaging data with a multi-echo sequence [15, 16]. This exciting development allows for the acquisition of instantly co-registered distinct arterial and venous maps of the cerebral vascular system without additional imaging time when compared to routine TOF imaging. These techniques are likely to have a major role in the evaluation of complex brain vascular malformations such as pial or dural arteriovenous malformations in the near future. Another recent development is hybrid opposite contrast MR angiography, which combines TOF and flow sensitive black blood contrasts [17]. This technique combines the excellent sensitivity of TOF techniques for high flow vessels with the excellent sensitivity of black blood techniques for slow flow vessels without increasing the imaging time. This development may increase the ability of TOF MRA to diagnose cerebral vasculopathies involving medium and small vessels but DSA comparison studies will be essential for validation.

Phase Contrast MRA Phase contrast imaging technique utilizes the difference in phase accumulation between stationary and moving protons, which are subjected to magnetic gradients. In order to exploit this difference, two opposing magnetic gradients, which are similar in magnitude and duration, are applied to a selected imaging slice in rapid succession. In the case of stationary protons within the excited tissue, position dependent phase differences, which accumulate along the axis of the first gradient are promptly neutralized after the application of the second gradient. However, in the case of protons, which are not stationary along the axis of these gradients, e.g. protons within a blood vessel oriented along the axis of the gradient, the phase accumulation is not reversed by application of the second gradient. This leads to the accumulation of a net positive or negative phase in mobile protons in flowing blood. The magnitude of this net phase accumulation depends upon the velocity of blood flow, whereas the sign of the net phase accumulation depends on the direction of blood flow. Needless to say phase differences are also similarly produced by higher order motion such as acceleration and jerk. The magnitude and the duration of the gradients can be adjusted to determine the range of velocities that can be encoded between the phase angles of −P and +P degrees. This step is called velocity encoding (VENC). In order to obtain technically adequate phase contrast images, it is essential to set an appropriate VENC. A VENC that is too high may result in poor signal from slow flowing vessels, whereas a VENC that is too low will result in signal aliasing in vessels with rapid flow. In the end, this technique results in the generation of a phase image, which shows the direction of blood flow and a magnitude image, the signal in which is proportional to the overall flow rate within the interrogated blood vessel [18]. Additionally, phase contrast imaging can also be used to generate flow velocity curves by continuously interrogating a

26 Pediatric MR Angiography: Principles and Applications

blood vessel during the entire cardiac cycle. Unlike time of flight imaging, in phase contrast imaging, it is very easy to distinguish between clotted blood with short T1 signal and flowing blood. Therefore phase contrast imaging is especially useful in imaging for cerebral venous sinus thrombosis in children, a not uncommon situation in neonates. Additionally, it can be used in other populations of children such as those with sickle cell disease in whom intravenous contrast administration is risky or contra-indicated. The sensitivity of phase contrast MRA (PCMRA) to the direction of blood flow is also very useful in evaluating arteriovenous malformations and collateral vessels. Both 3-D and 2-D PCMRA techniques are available. The relatively long acquisition times constitute a major drawback with PC-MRA techniques [19]. Recently, accelerated 4D methods for phase contrast imaging have been described which also enable in vivo demonstration hemodynamics as vascular flow streamlines [20]. Such depictions are likely to increase fundamental understanding of the evolution of pediatric neurovascular diseases including vessel wall shear stresses and collateral recruitment after large vessel occlusions. Phase contrast imaging has also been used in highly constrained projection reconstruction techniques (HYPR) to increase the fidelity of time-resolved MRA images [21]. The relatively long acquisition times constitute a major pitfall for PC-MRA techniques. However, parallel imaging techniques, improved coil performance, and 3-T field strengths are rendering these techniques faster and more clinically feasible.

Time-Resolved Contrast-Enhanced MRA in Children Contrast-enhanced MR angiography methods have grown rapidly in the past decade. In order to gain the maximum information from a contrast-enhanced MRA study, it is essential to distinctly image the arterial and venous phases during the first pass of the intravenously injected T1 shortening agent through the cerebral vasculature [22]. These techniques are often also referred to magnetic resonance digital subtraction angiography or MRDSA since they rely on subtraction of an initial non-contrast mask image of the area in question from the sequential phases of the contrast-enhanced study or with continuous subtractions of stagnate signal in sliding window techniques. Image acquisition during arterial and venous transit of contrast is traditionally timed using bolus tracking methods. A test bolus of contrast agent is injected prior to image acquisition. The MRA acquisition is triggered using a fixed delay after contrast injection (calculated from the test bolus injection) or the bolus is monitored and triggered in real time using MR fluoroscopy. Bolus timing methods are often successful in adults. However, in children, the test bolus dose (approximately 1 mL of contrast agent) is likely to constitute a larger fraction of the total amount of contrast that can be

367

used for imaging based on the child’s weight. Furthermore, greater variability in the heart rates and shorter arteriovenous circulation times in children make test bolus injections MRA techniques more challenging and often unfeasible [23]. Therefore, time-resolved MRA techniques with rapid continuous dynamic image acquisition are more commonly used in pediatric MRA studies. Broadly speaking, these techniques rely on centric or elliptical reordering of k space acquisition, partial acquisition and refreshment of k space, parallel imaging methods, radial sliding window reconstruction techniques and sliding mask subtraction algorithms [24–29]. These approaches have been used in various different combinations to increase the temporal resolution of these techniques while maintaining spatial resolution. Currently, sub-second acquisition of individual MRA frames is possible. Chooi et al. [30] studied 15 children with cerebrovascular malformations using MRDSA after hand injection of 1 or more doses of gadolinium contrast agent and found that MRDSA contributed significantly to avoiding catheter angiography (in 33% of patients), and justified catheter angiography in 27% of their patients. They also reported that MRDSA could be safely performed in neonates. However, MRDSA was limited in the evaluation of high flow vascular malformations of the brain and it could not be reliably used for treatment planning. Continuous improvement in acquisition techniques is ongoing. Most recently, time of arrival mapping techniques have been applied to three-dimensional time-resolved MRDSA with promising results [31]. However, MRDSA studies have not to date replaced conventional catheter-based DSA techniques in the evaluation of complex cerebral and cervical arteriovenous vascular malformations, especially for the evaluation of small residual shunts after treatment. In the following sections, we shall briefly review the role of these MRA techniques in the evaluation of specific neurovascular disorders in children. Brain vascular malformations in children: Pediatric cerebrovascular malformations consist of both diseases that are unique to the pediatric patient population such as vein of Galen malformations and Sturge–Weber syndrome as well as malformations that occur both in children and adults such as pial arteriovenous malformations (AVMs), dural arteriovenous malformations, dural arteriovenous fistulas (DAVFs), and developmental venous anomalies with or without associated cavernomas. The following sections consist of a brief review of the latter type of malformations followed by a more detailed review of the appropriate roles for MRA.

Arteriovenous Malformations Pediatric AVMs differ from their adult counterparts demonstrating more immature architecture [32, 33], a more diffuse pattern of AV shunting [34], and perhaps a higher cumulative

368

B.D. Jagadeesan and D.N. Loy

Fig. 26.1 Pial AVM in a 6-year-old child with seizures. Saggital (a), coronal (b) and axial (c) maximum intensity projections from a precontrast time of flight MR angiography showed a small periventricular arteriovenous malformation nidus (white arrow) supplied by the right pericallosal arteries and draining into the deep venous system. Postcontrast time-of-flight MR angiography in the same patient with identical saggital (d), coronal (e), and axial (f) maximum intensity projections shows improved contrast in the images but also significantly

increased venous contamination. Axial SWI image (g) shows hyperintense nidus (white arrow). Note the hyperintense blood products within the atrium of the right lateral ventricle (Asterisks, G, see A also). Anteroposterior (h) and oblique (i) DSA obtained during injection of the right vertebral artery demonstrates the nidus of the arteriovenous malformation, which is also supplied by posterior pericallosal branches of the right posterior cerebral artery

lifetime risk for hemorrhage [35]. Intraparenchymal hemorrhage constitutes the commonest presenting symptom in children with AVMs [35]. In a study of 26 children with nontraumatic intracerebral hemorrhage as the presenting event, Papadias et al. [36] found that AVMs represented the underlying etiology in seven of the patients. MRA was utilized to study these malformations and the authors opined that the overall performance of MRA was satisfactory in the detection of AVMs but did suffer from lack of temporal information and poor spatial resolution. Fasulakis et al. [37] directly compared the performance of 3D TOF MRA and DSA in nine children with brain AVMs. They reported agreement between these two modalities in four patients and discordant results in one patient. DSA demonstrated an additional supplying vessel not visible on MRA in one case. On the other

hand, MRA demonstrated an additional supplying vessel not described on DSA in two cases. Although there are few studies which compare MRA with DSA specifically in the pediatric population, it is likely that the overall performance of MRA in the detection of AVMs in children is similar to MRA in adults (Fig. 26.1). DSA remains the gold standard because of the superior temporal and spatial resolution. Aneurysms in Children: Aneurysms are much less common in children than in adults accounting for less than 5% of all detected aneurysms. In most cases, the etiology of these aneurysms is unknown. In the remaining patients, trauma, infectious disease, or connective tissue disorders may constitute the underlying etiology. These children usually present with subarachnoid hemorrhage or symptoms related to mass effect.

26 Pediatric MR Angiography: Principles and Applications

369

Fig. 26.2 Cerebral aneurysm. Nine-month-old child who presented to the hospital with nausea, vomiting, and poor appetite, her clinical examination had revealed a left VIth nerve palsy. Subsequent MRI examination of the head with time-of-flight MR angiography (a), susceptibility weighted imaging (b), and gadolinium-enhanced T1-weighted imaging (c) of the brain showed a large saccular aneurysm arising from the left internal carotid artery. There is thrombus layering in a dependent

manner within the aneurysm. Note: The thrombus is hyperintense on the time of flight maximum intensity projection image (a) and maybe confused for blood flow but is clearly identifiable as thrombus based on its hypointensity on the SWI image and the postcontrast MR angiogram. Digital subtraction angiography confirmed the large saccular aneurysm arising from the left internal carotid artery (d, e)

Saccular aneurysms represent the commonest morphologic subtype but giant aneurysms are considerably more common in children when compared to adults. With respect to infants, Buis et al. [38] conducted a retrospective review of 110 published articles describing a total of 131 aneurysms in children less than 1 year of age and found that most aneurysms occurred in the anterior circulation with the MCA being the most common vessel involved. They also reported a mean size of 1.8 ± 1.4 cm. Patients who presented with hemorrhage tended to be younger and had smaller aneurysms. They reported no gender bias. Traumatic pseudoaneurysms account for 14–39% of intracranial aneurysms in children [39] and they can occur with closed head injury, especially in the pericallosal or callosomarginal branches of the anterior cerebral arteries where these branches impact against the falx cerebri [40]. Most literature comparing MRA with DSA in adults generally suggest that MRA is excellent at the detection of aneurysms larger than 5 mm. It is likely that the performance of MRA in the detection of aneurysms in children is similar. Additionally, given that giant aneurysms constitute 16–29% of all aneurysms in children as compared to just 2% of

aneurysms in adults, MRA may be expected to a higher overall percentage of aneurysms in children compared to adults (Fig. 26.2). Allison et al. [41] reported that false-negative MRA studies occurred in the presence of spasm, slow flow or small aneurysm size in a study comparing MRA with DSA in children.

Vein of Galen Malformation In normal adults, the vein of Galen is a deep venous structure that drains the internal cerebral veins and the basal veins of Rosenthal. It empties into the straight sinus. The term, vein of Galen malformation, can be broadly used to describe any condition in which the vein is abnormally large or has abnormal morphology. Perhaps the best way to approach these malformations is to consider them as either high flow or low flow malformations. High flow malformations are the result of abnormal arteriovenous shunting, whereas low flow malformations are either the result of venous egress failure of the vein of Galen or from excessive venous inflow from developmental venous anomalies.

370

B.D. Jagadeesan and D.N. Loy

High Flow Vein of Galen Malformations Vein of Galen Aneurysmal Malformation By the 11th–12th week of gestation, a transient venous structure develops in the roof of the diencephalon, which receives drainage from the choroid plexus of the telencephalon. This structure is called the median vein of the prosencephalon or the vein of Markowski. With the development of the basal ganglia, the choroidal venous drainage shifts to the paired internal cerebral veins and the median vein of the prosencephalon regresses, except for its caudal extreme, which is joined by the internal cerebral veins and thalamostriate to form the vein of Galen. When this pattern is not established and there is arteriovenous shunting to the persistent embryonic median vein of the prosencephalon, the resultant malformation is called a vein of Galen aneurysmal malformation (VGAM) [42]. The underlying arteriovenous communications take place in the cistern of the velum interpositum and the quadrigeminal cistern. There is resultant ballooning of the median prosencephalic vein, which is thin walled and unsupported [42–44]. These VGAMs were subdivided into choroidal and mural subtypes [45, 46]. In the choroidal type (Fig. 26.3), there is a nidus supplied by the choroidal arteries, which drains into the primitive vein of the mesencephalon. In the mural type, arteriovenous fistula(s) are located in the inferolateral wall of the median vein of the prosencephalon. The communications occur in the anterior end of the median prosencephalic vein. Arterial supply to the malformation arises from those arteries that normally feed the structures in the vicinity, namely the arteries that feed the tela choroidea and the quadrigeminal plate. They may belong to either the anterior or prosencephalic group (the anterior cerebral, anterior choroidal, middle cerebral, and the posterolateral choroidal arteries), or the posterior or mesencephalic group (the posteromedial choroidal, posterior thalamoperforating, quadrigeminal, and superior cerebellar arteries) [42, 43]. The arterial supply that is recruited from the subependymal and thalamo-perforator arteries arises secondarily from a sump effect. These secondary arterial-feeders usually tend to disappear following treatment of the main feeders to the shunt. Persistence of a primitive limbic arterial arch, which connects the anterior and posterior choroidal vessels, may also be noted [47]. The presence of a high-flow arteriovenous shunt also results in the persistence of fetal patterns of venous drainage from the deeper structures of the brain. The thalamostriate veins in a VGAM no longer drain afferent into the median vein of the prosencephalon but may drain into the posterior and inferior thalamic veins and eventually drain in to a subtemporal vein, or the lateral mesencephalic vein and into the superior petrosal sinuses. This pattern results in the typical

Fig. 26.3 Vein of Galen malformation in a 22-month-old child. Saggital (a) and axial (b) maximum intensity projection images from a noncontrast time-of-flight MRA study show an enlarged median prosencephalic vein fed by arterial branches from both the anterior and posterior cerebral arteries. The arterial feeders can be seen to complete a primitive limbic circle (a). The posterior choroidal feeders are depicted (b). Postcontrast saggital T1-weighted image (c) and sagittal reconstruction of a postcontrast susceptibility weighted image (d) from the same study show multiple enlarged draining vessels around the midbrain with the formation of multiple large venous aneurysms. Interestingly, the axial pre- (e) and postcontrast SWI (f) images show not only bright signal in the enlarged veins with arterialized flow from arteriovenous shunting but also dark passively congested cortical veins, suggesting venous hypertension. Venous hypertension is commonly associated with ventriculomegaly in these patients

epsilon shape (“epsilon vein”) on the lateral views in conventional catheter angiograms [48, 49]. Efferent venous drainage from a VGAM is usually into the straight sinus. However, the VGAM nidus may also drain to some extent into the pontomesencephalic vein via the choroidal veins. The pontomesencephalic vein in turn usually drains into the falcine sinus with or without a stenosis at the sinovenous junction. This results in persistence of the falcine sinus [45]. The presence of a connection between the VGAM nidus and the pontomesencephalic vein may also result in pial venous drainage due to communications between the choroidal and striate venous systems. Documentation of

26 Pediatric MR Angiography: Principles and Applications

the drainage pattern is crucial since the presence of pial venous drainage excludes a patient from transvenous endovascular embolization [47].

Vein of Galen Aneurysmal Dilatation In contrast to vein of Galen arteriovenous malformations, the vein of Galen aneurysmal dilatation (VGAD) consists of a pial AVM, which drains into a normally formed vein of Galen. The pial AVM may be supratentorial or infratentorial. The clinical manifestations depend on the location of the AVM and the extent of outflow obstruction at the drainage of the Vein of Galen into the straight sinus. The VGAMs and VGAD are also classified according to the Yasargil system [50]. In the Yasargil classification, VGADs constitute the type IV malformations, whereas the VGAMs are classified as types I–III according to the arterial feeders. Dural Vein of Galen Dilatations: Vein of Galen dilatation can also occur due to development of dural arteriovenous shunting after straight sinus thrombosis in adults or thrombosis of the falcine vein or vein of Galen in children [51, 52].

Low Flow Vein of Galen Dilatations There can also be dilatation of a normally formed vein of Galen secondary to congestive heart failure or from drainage of a venous angioma known as a vein of Galen Varix [47].

Role of MRA in Vein of Galen Malformations In the light of the above discussion, it becomes clear that in order for any angiographic technique to be useful in the evaluation of vein of Galen malformations, the technique must be capable of differentiating between high flow and low flow vein of Galen malformations, and it should also be able to provide exquisite information regarding the arterial supply and the venous drainage pattern associated with vein of Galen malformations. Time-resolved dynamic contrastenhanced MRA technique is likely to offer the greatest detail if performed with high temporal and spatial resolution. Physiological immaturity in renal function often precludes intravenous administration of gadolinium agents in neonates. Hyperdynamic circulations and congestive cardiac failure associated with VGAMs also makes time-resolved MRA challenging. Recently, Chooi et al. [30] studied a set of 15 pediatric patients with MRDSA and found that the extremely high flow rates associated with VGAMs resulted in poor visualization of the surrounding feeding cerebral vessels.

371

They also opined that the MRDSA results could not be used for treatment planning in these children. In this setting, a high quality time-of-flight MRA may be the study of choice. Newer rapid phase contrast imaging techniques may also be useful. Time-of-flight imaging is excellent in the depiction of the arterial supply but may or may not reveal the venous drainage pattern in its entirety. In this context, we have also found SWI to be a useful tool, since the BOLD technique offers exquisite representation of the deep venous system of the cerebral hemispheres [53]. By reviewing both the TOF and SWI images, it may be possible to adequately evaluate vein of Galen malformations but the technique has not been applied enough in VGAMs to make conclusions at this time. Segmental neurovascular malformations in children: The segmental neurovascular malformations are a group of rare disorders that share several common features. According to Krings et al. [54], these features consist of their segmental distribution pattern, their link to the concept of neural crest development, the metameric origin and cephalic migration of the cells concerned, their evolution with growth of the child, and their varying expressivity. These malformations include several eponymous syndromes such as Sturge–Weber syndrome, Wyburn-Mason syndrome, Klippel–Trenaunay syndrome, Cobbs syndrome, and PHACES [55–58]. Of these, Wybrun-Mason or Dechaume–Blanc syndromes represent the occurrence of brain arteriovenous malformations, orbital arteriovenous malformations and maxillofacial vascular abnormalities in the same patient. Extracranial vascular malformations of the head and neck are also a part of these segmental malformations. Midline brain vascular malformations seem to be associated with midline facial vascular malformations and lateral malformations are associated with maxillary malformations. The role of MRA in the detection of the orbital components of these malformations is clearly evident but in the presence of cerebral and facial malformations on MRA, the orbital anomalies can usually be studied more thoroughly with ophthalmologic evaluations. The cerebral AVMs found in these malformations differ from isolated AVMs in their multifocality, recruitment of dural vessels, and their low shunt volumes [54, 59]. Sturge–Weber (SWS) syndrome: This syndrome also known as encephalo-trigeminal angiomatosis consists of a large unilateral leptomeningeal venous malformation which undergoes progressive calcification with accompanying atrophy of the involved cerebral hemisphere and an associated dermal venous malformation or facial port wine stain over the ipsilateral V1 division territory. Children with this syndrome typically present with the cosmetic deformity and epilepsy. There is also enlargement of the choroid plexus in the ipsilateral cerebral hemisphere and accompanying vascular malformations in the orbit. In the majority of affected children, the

372

Fig. 26.4 Sturge–Weber Syndrome in a 2-year-old child with seizures. Noncontrast axial head CT (a) shows cortical calcification involving the left frontoparietal convexity. There is corresponding signal loss on an axial SWI image (b) and leptomeningeal enhancement on a postcontrast T1-weighted image (c) from the same patient suggesting a diagnosis

diagnosis can be made on the basis of dermatological and neurological findings. However, occasionally, the neurologic findings are mild and the dermatologic findings may be absent. In such cases, neuroimaging can play an important role. Recently, MRI/MRA studies have become the neuroimaging study of choice. On these studies, the diagnostic imaging features are gyral contrast enhancement, leptomeningeal vessel enhancement, enlarged deep medullary veins, and an enlarged ipsilateral choroid plexus (Fig. 26.4). Among these features, enhancement of the leptomeningeal vessels is usually considered to be the most important feature [60, 61]. MR contrast venography plays a useful role in this setting by demonstrating thrombosed cortical veins and enlarged deep draining medullary veins. Recently, Hu et al. [62] evaluated the role of SWI or MR bold venography in imaging patient with SWS. In their series of 12 patients, they found that SWI provide complimentary information to contrast-enhanced MRI. Specifically, SWI showed superior depiction of abnormal transmedullary veins, periventricular veins, cortical gyriform hypointensities, and gray–white junction abnormalities. However, conventional contrast-enhanced MRI was reportedly better at demonstrating abnormal leptomeningeal enhancement. Extracranial vascular malformations in children: These malformations can also be broadly classified as low-flow and high-flow lesions. Among the low-flow lesions, hemangiomas form a distinct class when compared to other venous, lymphatic or venolymphatic malformations of the head and neck (Fig. 26.5). The high-flow vascular malformations are essentially arteriovenous malformations. The goal of imaging studies in head and neck vascular malformations is to classify the malformations as high- or low-flow, evaluate their anatomy with respect to feeding and draining vessels, and to evaluate their effect on other tissues in the head and neck. Routine MR imaging by itself is often enough to

B.D. Jagadeesan and D.N. Loy

of Sturge–Weber Syndrome. Note: The extent of leptomeningeal enhancement in image C exceeds the boundaries of the calcified region in image A suggesting increased sensitivity of MRI in the detection of meningeal vascular abnormalities in these children

Fig. 26.5 Cystic hygroma. Two-year-old child with neck swelling. Axial T2-weighted image (a) shows a multilobulated mass which is predominantly T2 hyperintense but also shows a few locules with fluid levels and T2 hypointensity. Axial source image from a time of flight angiogram of the neck vessels (b) in the same patient shows no evidence for arterial supply to this mass. Coronal dynamic contrastenhanced MRA image obtained during the venous phase after injection of intravenous gadolinium (c) also shows no evidence for abnormal blood vessels associated with this mass. These imaging findings are characteristic of a cystic hygroma or a low flow veno-lymphatic malformation of the neck

address the questions regarding the anatomy of the vascular malformations with high spatial resolution and to document the relationship of the vascular malformations to the surrounding structures. The major role of MRA is simply to distinguish between high-flow and low-flow vascular malformations.

26 Pediatric MR Angiography: Principles and Applications

Thus, the role of MRA in the extracranial malformations is somewhat different from its role in intracranial malformations, where both high spatial and temporal resolution are required, this is simply due to the fact that extracranial vascular malformations tend to be much larger than their intracranial counterparts and are often distinctly situated away from normal vascular structures. Therefore, MRA techniques with high temporal resolution and a large field of view are preferred to evaluate the extracranial vascular malformations at the expense of spatial resolution. Ziyeh et al. [63] performed MR projectional angiography in eight patients with head and neck vascular malformations and compared the result with DSA. They found that MRPA was adequate in the differentiation of high-flow from low-flow vascular malformations. Cerebral venous thrombosis and magnetic resonance venography: The incidence of cerebral venous thrombosis in the neonatal population may be as high as 40.7 per 100,000 per year. In the Canadian Pediatric Ischemic Stroke Registry, the incidence cerebral venous sinus thrombosis was 0.6 per 100,000 children per year aged term birth to 18 years. Some of the many predisposing causes include disorders such as leukemia, trauma, dehydration, infection, prothrombotic disorders, congenital heart disease, and prematurity. In the majority of children, the superior saggital sinus and the transverse sinus are involved. Unlike adults, the presenting symptoms in children can be atypical or subtle. Hence neuroimaging and particularly magnetic resonance venography (MRV) plays a major role in the detection of pediatric cerebral sinovenous thrombosis. Time of flight MRV is unreliable in this setting because of the T1 shortening effect of thrombus which can mimic flow in a thrombosed vascular segment and the high incidence of false-positive findings. Rollins et al. [64] directly compared the performance of 2D time-of-fight techniques with 3D contrast-enhanced MR venograms of the cerebral venous system in 37 children and found that the results of the techniques were comparable in only 17 of the patients. In 19 children, various vascular anomalies such as flow gaps and segmental atresia were suggested by 2D TOF which were found to be absent on the postgadolinium studies. In the neonatal population where the accurate diagnosis of cerebral venous thrombosis is most crucial, Widjaja et al. [65] found that TOF MRV showed spurious flow gaps in 31 of 51 neonates who had normal venous anatomy on CT venography. When postcontrast imaging is performed without bolus timing or subtraction techniques, especially with a longer acquisition time, the organized thrombus itself may enhance resulting in a false-negative study. MRV studies with bolus timing methods and MR fluoroscopic monitoring of contrast bolus arrival in the superior sagittal sinus followed by centric

373

re-ordering of echo acquisition may be required. However, these studies may suffer from poor spatial resolution, which may be a significant drawback in very young children. Recently, Meckel et al. [66] evaluated a combination of these two techniques for the evaluation of cerebral venous thrombosis, and to differentiate between acute thrombus and chronic enhancing organized thrombus. They found that this combination of 4D MR venography outperformed other modalities such as TOF MRV and gradient echo imaging in the detection of dural venous sinus thrombosis. This method had the highest sensitivity in the detection of subacute venous sinus thrombosis. However, in common with other studies, their study too showed that the sensitivity of MRV techniques for the detection of cortical venous thrombosis was lower than that of traditional gradient recalled echo (GRE) MR imaging. Isolated cortical venous thrombosis, in the absence of dural venous sinus thrombosis is a rare phenomenon and it is possible that SWI bold venography will play an increasing role in its evaluation in the future. Phase contrast imaging is also commonly used in the evaluation of the cerebral venous system; its main advantages are the lack of need for intravenous contrast, the ability to differentiate between thrombus and flow, and it ability to provide quantitative hemodynamic information.

MRA in the Evaluation of Acute Ischemic Stroke in Children Acute ischemic stroke is a rare disorder in children thought to affect 1.2–3.3/100,000 children/year [67–71]. The majority of the cases of childhood stroke occur in the absence of any known risk factors. However, in many cases, a specific etiology can still be identified. The most common etiologies include vasculopathies, cardiac disease, and pro-thrombotic disorders. The vasculopathies, including transient cerebral arteriopathy (TCA), fibromuscular dysplasia, childhood primary angiitis of the central nervous system, and Moyamoya disease, can be identified in 18–80% of children with acute ischemic strokes depending on the population studied. Arterial dissection can also be found in 7.5–20% of cases of acute ischemic stroke in children [70–72]. Transient cerebral vasculopathy: TCA is the most common among these conditions. In one series, TCA accounted for approximately 26% of sample of children suffering from stroke. TCA is an inflammatory vasculopathy, which has been reported to be triggered by prior varicella zoster infection in approximately 44% of cases and by other bacterial or viral infections in the rest. It usually manifests as unilateral, often multiple, focal or elongated segmental narrowings of the terminal internal carotid artery and the proximal middle

374

B.D. Jagadeesan and D.N. Loy

Fig. 26.6 Moyamoya disease. Maximum intensity projection images (a) and volume rendered images (b) of the circle of Willis from a time-of-flight MR angiography study in an 8-year-old child demonstrate severe stenoses of the terminal internal carotid arteries and the proximal anterior and middle cerebral arteries with prominent deep lenticulostriate collateral vessels. The M2 and M3 segments of the middle cerebral arteries fill (retrograde on DSA) from dural vessels which are the result of a prior encephaloduromyosynangiosis procedure (a–d). Serial maximal intensity projection images of the vertebrobasilar circulation from time of flight studies obtained at initial presentation (e) and 2 years later (f, g) in the same child with progressive stenosis of the P1 segment of the right posterior cerebral artery

and anterior cerebral arteries with resolution on follow-up imaging [73]. Moyamoya disease and Moyamoya syndrome: In contrast to transient cerebral vasculopathy, bilateral progressive narrowing and eventual occlusion of the terminal ICA and the proximal MCAs and ACAs characterize Moyamoya disease and syndrome. Among the common underlying etiologies, sickle cell disease, hemolytic uremic syndrome, neurofibromatosis, Down’s syndrome, and radiation are the most widespread [74–76]. There is development of an abnormal network of collaterals including enlarged lenticulostriate arteries, which results in the puff of smoke appearance on conventional angiograms as well as development of extensive, dural to pial collaterals. Although ischemic stroke is more common in children with Moyamoya disease, hemorrhagic stroke may also occur and it is believed that abnormal dilatation of the anterior choroidal artery may predict an increased risk for hemorrhagic stroke [77]. It has also been shown that increasing

prominence of collateral vessels predicts an increased risk of stroke. Several studies have shown that MRA using time-offlight technique compares favorably with DSA in evaluating the cranial circulation in children with untreated sickle cell disease. It has also been shown that imaging at 3 T provides better information on the extent of collateralization compared to 1.5 T. This is simply because of the improved spatial resolution and signal-to-noise ratio of higher field strengths [78]. MRA can also be used in the follow up of patients with sickle cell disease who are treated with encephalo duro arterio synangiosis or encephalo duro arterio myosynagiosis (EDAS or EDAMS, respectively). Following EDAS, there is an increase in size of the distal MCA branches (underlying the graft) and a decrease in collaterals at the base of the brain, which imparts a gradual but significant decrease in the risk for stroke (Fig. 26.6). Yoon et al. [79] suggested that these findings could be adequately visualized with routine TOF-MRA.

26 Pediatric MR Angiography: Principles and Applications

It is unlikely that contrast-enhanced MRA will have any appreciable role in the evaluation of sickle cell disease in the near future given the decreased renal function in most children with sickle cell disease. However, new dynamic phase contrast studies may have a role in understanding the hemodynamics of treated and untreated sickle cell disease. Neff et al. [80] used 2D phase contrast MRA for flow quantification and found that there was 50% or more reduction in blood flow volumes in the internal carotid arteries of subjects with Moyamoya disease when compared with controls whereas flow in the basilar arteries increased substantially. It remains to be seen if such flow quantification studies can predict which subjects are at an increased risk for intracranial hemorrhage. For now, O15 PET studies remain the gold standard for quantification of oxygen extraction fractions and stroke risk in Moyamoya patients. Vascular dissection and stroke in children: Vascular dissections are often under diagnosed and may account for a significant proportion of strokes in children. Vascular dissections may represent the underlying etiology for arterial ischemic stroke in 7.5–20% of children with acute ischemic stroke. Dissections in children differ from those in adults in both clinical presentation and distribution. Unlike adults, most dissections in children present with infarcts and dissections are more common among male children. Likewise a greater proportion of dissections tend to be intracranial when they involve the anterior circulation although extracranial dissections of the vertebral arteries at the C1–C2 level are still the most common. Dissections can result in luminal narrowing of the affected vessel (when the dissection plane is subintimal) or pseudoaneurysm formation (when the dissection plane is subadventitial). Subarachnoid hemorrhages can also be the presenting manifestation of a dissection especially when there is pseudoaneurysm formation. Dissections can occur in the absence of any known trigger (i.e., spontaneous dissections) or may occur in the setting of trauma or a known vessel wall disorder such as fibromuscular dysplasia. When imaging the vessels in a child with suspected dissection of the craniocervical arteries, it is not only important to image the vessel lumen but also to image the vessel wall in order to detect the intramural hematoma. Axial high-resolution fatsaturated T1- and T2-weighted images are crucial in the detection of intramural hematomas. Childhood vasculitis: Aviv et al. [81] compared the performance of catheter angiography and MR angiography in the selection of childhood vasculitis. Children with a clinical diagnosis of primary angiitis of childhood were included in the study and those with other causes for childhood vasculopathy such as varicella or sickle cell disease were excluded. They studied a total of 25 patients and found that 3D time-offlight MRA identified only 45 of the 64 lesions detected on

375

catheter angiography and there was only a modest agreement between the two modalities (k = 0.4). However, when the performance of MRA for making the diagnosis of vasculitis was evaluated rather than the identification of individual lesions, MRA had a sensitivity of 95% and a positive predictive value of 83%. They reported that unilateral disease was common and that proximal MCA and ACA lesions accounted for most stenoses. These lesions lead most commonly to basal ganglia infarctions. The authors also concluded that MRA performs well in quantifying the degree of stenosis in individual affected segments when the stenosis is greater than 75% of the luminal diameter but overestimates the degree of stenosis in patients with 50–75% stenosis. Eleftheriou et al. [82] also compared the performance of MRA and DSA in childhood central nervous system vasculitis. They reported that MRA correctly detected only 28 of 44 lesions identified by catheter angiography. MRA was 63% sensitive and 89% specific with a PPV of 62% and an NPV of 88%. They found that MRA performed particularly poorly in the posterior circulation. The 3D TOF technique used in both these studies also did not offer any dynamic flow information. In conclusion, at present a positive MRA study may prove to be reliable indicator for the presence of cerebral vasculitis but a negative study does not exclude the presence of childhood vasculitis. Unlike adults where 20–40% of patient with cerebral vasculitis do not have angiographically detectable changes, most children with vasculitis do show changes on DSA. Therefore, a catheter angiogram is still necessary for diagnosis in most cases of suspected pediatric vasculopathy.

Spinal Vascular Malformations The most widely used classification for spinal vascular lesions was devised by Anson and Spetlzler and classifies them into type 1, spinal dural arteriovenous fistulas (dorsal intradural fistulas); type II, intradural glomus arteriovenous malformations (these may be compact or diffuse with an extramedullary component); type III, juvenile or metameric AVMs; and type IV, intradural perimedullary AVFs (ventral intradural fistulas). The incidence of pediatric spinal vascular malformations is not entirely known although it is thought that they constitute approximately 10–20% of all central nervous system vascular shunts. In younger children, spinal vascular malformations with arteriovenous shunting can be associated with hereditary hemorrhagic telengiectasia or Osler–Rendu–Weber syndrome. They may also be found in association with Klippel–Trenaunay syndrome, Parkes–Weber syndrome, or neurofibromatosis. Alternatively, they may form part of a spinal arteriovenous metameric syndrome. Most pediatric

376

spinal cord vascular lesions are isolated, spontaneous anomalies. The commonest of the spinal vascular malformations in adults, namely the type 1 malformations (dural AV fistula) are thought to be acquired in nature and rarely occur in children. Pediatric patients with spinal vascular malformations most commonly present with complications from hemorrhage or venous congestion. Hence, most undergo an initial MR imaging study where enlarged flow voids are usually evident on T2-weighted images. MRA is reported to be most useful for the diagnosis of type 1 spinal vascular malformations. MRA may play a role in identifying the approximate levels of arterial feeders in other types of spinal AVMs but is not helpful for characterization of the types of AV shunting. Identification of hypertrophied segmental feeders with MRA can be useful to direct spinal catheter angiography, thus reducing iodinated contrast doses in small children. MRA is insensitive for the detection of rare type IV malformations. The spinal vasculature presents many unique challenges for MRA. These arise from the unique architecture of the spinal vasculature, with paired ventral arteries and veins, which can only be distinguished from each other by dynamic imaging; the extremely small caliber of the vessels concerned; the extreme variability in the origin and course of crucial vessels such as the artery of Admakewicz, and the affliction of nervous tissue remote from the neurovascular abnormality. Therefore, any MRA technique for imaging the spinal vasculature has must exhibit high spatial resolution, high temporal resolution, and a relatively large field of view. TOF and PC MRA are not useful in the spine given the slow flow in spinal vessels and their small size. However, some authors have successfully used these techniques for the diagnosis and follow up of spinal vascular lesions in adult patients [83, 84]. Therefore, contrast-enhanced MRA methods are most commonly utilized for spinal MR angiography. The usual arteriovenous circulation time of blood in the spinal cord ranges between 9 and 12 s in adults and is likely shorter in children, especially those with hyperdynamic circulation from other vascular anomalies. In order to be useful, dynamic MRA techniques should ideally be able to acquire images of the entire spinal vasculature within the order of this time frame, which will allow the reader to differentiate between the arterial and venous systems. Hence unlike in the brain where subsecond temporal resolutions present the major

B.D. Jagadeesan and D.N. Loy

technical challenge for MRA, in the spine the challenge is to maximize the field of view and achieve high spatial resolution while maintaining a temporal resolution of several seconds. In small children, the field of view that is required may be smaller, but the vessels that have to be imaged are also smaller, making spinal vascular imaging even more challenging. Even with acquisition times that are longer than the arteriovenous circulation time, adequate differentiation between arteries and veins can be achieved by bolus timing the arrival of contrast in the aorta and acquiring the center of k space during the time window when the arteriovenous contrast difference is greatest. Typically dual-phase imaging is performed. The first phase is acquired during the maximal contrast difference between arteries and veins and the second phase is obtained immediately thereafter to study the venous system. Backes et al. [85] speculate that the second phase may be more focused with the use of a smaller field of view and higher resolution. They also suggest the use of intravascular contrast agents which stay in the blood pool for an extended time. Using dynamic bolus timed MRA, Mull et al. [86] identified the main feeding artery in 10 out of 11 patients with spinal intradural arteriovenous malformations and were also able to differentiate between fistulous and nidal malformations in four out of six patients. Jaspers et al. [87] have successfully used a key hole imaging technique to image the spinal vasculature with a temporal resolution of 6–8 s. However, as elsewhere, bolus-timing methods are not optimal for the study of spinal vessels in children. Some authors have suggested using a higher gadolinium contrast dose in the range of 0.2 mmol/kg to improve sensitivity [85], but this may not be feasible in children with immature or poor renal function. Therefore, dynamic subtraction imaging methods, which do not rely on timing may be required to optimally image the spinal vessels (Fig. 26.7). Hyodoh et al. [88] performed a dynamic study of the spinal vessels using a 3D fast spoiled gradient recalled acquisition in steady state (GRASS) method, and imaged five consecutive phases of the transit of contrast bolus through the spinal vasculature. They then performed a double subtraction of the maximum intensity projection (MIP) images to delineate the artery of Adamkiewicz. A similar technique can be used to improve the hemodynamic evaluation of spinal arteriovenous malformations in children.

26 Pediatric MR Angiography: Principles and Applications

Fig. 26.7 Type III metameric spinal AVM in a 6-year-old child who presented with neck pain and swelling. MRI and MRA examinations after conventional radiographic images showed lytic lesions involving her C1 to C4 vertebrae. T2-weighted images in the saggital (a) and (b) axial planes showed multiple hypointense vascular flow voids involving the C1 to C4 veretbrae, with intradural extension. There was also abnormal T2 hyperintensity within the cervical spinal cord parenchyma.

References 1. Swoboda NA, Armstrong DG, Smith J, Charkot E, Connolly BL. Pediatric patient surface doses in neuroangiography. Pediatr Radiol. 2005;35:859–866. 2. Glennie D, Connolly BL, Gordon C. Entrance skin dose measured with MOSFETs in children undergoing interventional radiology procedures. Pediatr Radiol. 2008;38:1180–1187. 3. Hollingsworth CL, Yoshizumi TT, Frush DP, et al. Pediatric cardiac-gated CT angiography: assessment of radiation dose. AJR Am J Roentgenol. 2007;189:12–18. 4. Raelson CA, Kanal KM, Vavilala MS, et al. Radiation dose and excess risk of cancer in children undergoing neuroangiography. AJR Am J Roentgenol. 2009;193:1621–1628. 5. Pierce DA, Shimizu Y, Preston DL, Vaeth M, Mabuchi K. Studies of the mortality of atomic bomb survivors. Report 12, Part I. Cancer: 1950–1990. Radiat Res. 1996;146:1–27. 6. Preston DL, Shimizu Y, Pierce DA, Suyama A, Mabuchi K. Studies of mortality of atomic bomb survivors Report 13: Solid cancer and noncancer disease mortality: 1950–1997. Radiat. Res. 2003;160: 381–407. 7. Yang CY, Chen YF, Lee CW, Huang A, Shen Y, Wei C, Liu HM. Multiphase CT angiography versus single-phase CT angiography: comparison of image quality and radiation dose. Am J Neuroradiol. 2008;29:1288–1295.

377

Maximum intensity projection images obtained during the arterial phase of a dynamic contrast-enhanced MR angiography examination of the neck show a large arteriovenous malformation in the upper cervical spine region supplied by both vertebral arteries (c), large venous channels can be seen draining the arteriovenous malformation to the internal jugular veins (d). Spinal DSA (e)

8. Strauss KJ, Goske MJ, Fursh DP, Butler PF, Morrison G. Image Gently Vendor Summit: working together for better estimates of pediatric radiation dose from CT. AJR Am J Roentgenol. 2009; 192:1169–1175. 9. Bosmans H, Wilms G, Dymarkowski S, Marchal G. Basic principles of MRA. Eur J Radiol. 2001;38:2–9. 10. Haacke EM, Brown RW, Thompson MR, Venkatesan R. Magnetic Resonance Imaging—Physical Principles and Sequence Design. New York: Wiley-Liss; 1999:10. 11. Blatter DD, Parker DL, Robison RO (1991) Cerebral MR angiography with multiple overlapping thin slab acquisition. Part I. Quantitative analysis of vessel visibility. Radiology. 179:805–811. 12. Atkinson D, Brant-Zawadzki M, Purdy D, Laub G. Improved MR angiography: magnetization transfer suppression with variable flip angle excitation and increased resolution. Radiology. 1994;190:890–894. 13. Ozsarlak O, Van Goethem JW, Parizel PM. 3D time-of-flight MR angiography of the intracranial vessels: optimization of the technique with water excitation, parallel acquisition, eight-channel phased-array head coil and low-dose contrast administration. Eur Radiol. 2004;14:2411–2444. 14. Ozsarlak O, Goethem JWV, Maes M, Parizel PM. MR angiography of the intracranial vessels: technical aspects and clinical applications. Neuroradiology. 2004;46:955–972. 15. Du YP, Jin Z, Hu Y, Tanabe J. Multi-echo acquisition of MR angiography and venography of the brain at 3 Tesla. J Magn Reson Imaging. 2009;30:449–454.

378 16. Deitsung A, Dittrich E, Sedalick J, Rauscher A, Reichenbach JR. ToF-SWI: simultaneous time of flight and fully flow compensated susceptibility weighted imaging. J Magn Reson Imaging. 2009; 29:1478–1484. 17. Tsuchiya K, Yoshida M, Imai M, Nitatori T, Kimura T, Ikedo M, Takemoto S. Hybrid of opposite-contrast magnetic resonance angiography of the brain by combining time-of-flight and black blood sequences: its value in Moyamoya disease. J Comput Assist Tomogr. 2010;34:242–246. 18. Marks MP, Pelc NJ, Ross MR, Enzmann DR. Determination of cerebral blood flow with a phase-contrast cine MR imaging technique: evaluation of normal subjects and patients with arteriovenous malformations. Radiology. 1992:182:467–476. 19. Markl M, Chan FP, Alley MT, et al. Time-resolved three-dimensional phase-contrast MRI. J Magn Reson Imaging. 2003:17:499–506. 20. Yamashita S, Isoda H, Hirano M, et al. Visualization of hemodynamics in intracranial arteries using time-resolved three-dimensional phase-contrast MRI. J Magn Reson Imaging. 2007;25:473–478. 21. Velikina JV, Johnson KM, Wu Y, Samsonov AA, Turski P, MIstretta CA. PC HYPR flow: a technique for rapid imaging of contrast dynamics. J Magn Reson Imaging. 2010;31:447–456. 22. Korosec FR, Frayne R, Grist TM, et al. Time-resolved contrastenhanced 3D MR angiography. Magn Reson Med. 1996;36: 345–351. 23. Muthupillai R, Vick GW, Flamm SD and Chung T. Time-resolved contrast-enhanced magnetic resonance angiography in pediatric patients using sensitivity encoding. J Magn Reson Imaging. 2003; 17:55–564. 24. Pinto C, Hickey R, Carroll TJ, et al. Time-resolved MR angiography with generalized autocalibrating partially parallel acquisition and time-resolved echosharing angiographic technique for hemodialysis arteriovenous fistulas and grafts. J Vasc Interv Radiol. 2006;17:1003–1009. 25. Lim RP, Shapiro M, Wang EY. 3D time-resolved MR angiography (MRA) of the carotid arteries with time-resolved imaging with stochastic trajectories: comparison with 3D contrast-enhanced boluschase MRA and 3D time-of-flight MRA. AJNR Am J Neuroradiol. 2008;29:1847–1854. 26. Jeong HY, Cashen TA, Hurley MC, Eddleman C, Getch C, Batjer H, Carroll TJ. Radial sliding-window magnetic resonance angiography (MRA) with highly-constrained projection reconstruction (HYPR). Magn Reson Med. 2009;61:1103–1113. 27. Eddleman CS, Jeong HY, Hurley MC, Zuehlsdorff S, Dabus G, Getch H, Batjer H, Bendok BR, Carroll TJ. 4D radial acquisition contrast-enhanced MR angiography and intracranial arteriovenous malformations: quickly approaching digital subtraction angiography. Stroke. 2009;40;2749–2753. 28. Willinek WA, Hadizadeh DR, von Falkenhausen M, et al. 4D timeresolved MR angiography with keyhole (4D-TRAK): more than 60 times accelerated MRA using a combination of CENTRA, keyhole, and SENSE at 3.0T. J Magn Reson Imaging. 2008;27:1455–1460. 29. van Vaals JJ, Brummer ME, Dixon WT, et al. Keyhole method for accelerating imaging of contrast agent uptake. J Magn Reson Imaging. 1993;3:671–675. 30. Chooi WK, Connolly DJ, Coley SC, Griffiths PD. Assessment of blood supply to intracranial pathologies in children using MR digital subtraction angiography. Pediatr Radiol. 2006;36:1057–1062. 31. Riederer SJ, HAider CR and Borisch E. Time-of-arrival mapping at three-dimensional time-resolved contrast-enhanced MR angiography. Radiology. 2009;253:532–542. 32. Klimo P, Jr, Rao G, Brockmeyer D. Pediatric arteriovenous malformations: a 15-year experience with an emphasis on residual and recurrent lesions. Childs Nerv Syst. 2007;23:31–37. 33. Gabriel EM, Sampson JH, Wilkins RH. Recurrence of a cerebral arteriovenous malformation after surgical excision. Case report. J Neurosurg. 1996;84:879–882.

B.D. Jagadeesan and D.N. Loy 34. Luessenhop AJ. Natural history of cerebral arteriovenous malformations. In: Wilson CB, Stein MB, eds. Intracranial Arteriovenous Malformations. Baltimore: Williams & Wilkins; 1984:12–23. 35. Humphreys RP. Brain vascular malformations. In: Choux M, Di Rocco C, Hockley A, Walker M, eds. Paediatric Neurosurgery. Edinburgh: Churchill Livingstone; 1999:665–677. 36. Papadias A, Taha A, Sgouros S, Walsh AR, Hockley AD. Incidence of vascular malformations in spontaneous intra-cerebral haemorrhage in children. Childs Nerv Syst. 2007;23:881–886. 37. Fasulakis S, Andronikou S. Comparison of MR angiography and conventional angiography in the investigation of intracranial arteriovenous malformations and aneurysms in children. Pediatr Radiol. 2003;33:378–378. 38. Buis DR, van Ouwerkerk WJ, Takahata H, Vandertop WP. Intracranial aneurysms in children under 1 year of age: a systematic review of the literature. Childs Nerv Syst. 2006;22:1395–1409. 39. Ventureyra EC, Higgins MJ. Traumatic intracranial aneurysms in childhood and adolescence. Case reports and review of literature. Childs Nerv Syst. 1994;10:361–369. 40. Nakstad P, Nornes H, Hauge HN. Traumatic aneurysms of the pericallosal arteries. Neuroradiology. 1986;28:335–338. 41. Allison JW, Davis PC, Sato Y, Jamees CA, HAque SS, Angtuaco EJ, Glasier CM. Intracranial aneurysms in infants and children. Pediatr Radiol. 1998;28:223–229. 42. Raybaud CA, Strother CM, Hald JK. Aneurysms of the vein of Galen: embryonic considerations and anatomical features relating to the pathogenesis of the malformation. Neuroradiology. 1989;31:19–28. 43. Casasco A, Lylyk P, Hodes JE, Kohan G, Aymard A, Merland JJ. Percutaneous transvenous catheterization and embolization of vein of Galen aneurysms. Neurosurgery. 1991;28:260–266. 44. Gupta AK, Varma DR. Vein of Galen malformations: review. Neurol India. 2004;52:43–53. 45. Berenstein A, Lasjaunias P. Arteriovenous fistulas of the brain. In: Surgical Neuroangiography: Endovascular Treatment of Cerebral Lesions. Berlin: Springer-Verlag; 1992:267–317. 46. Garcia-Monaco R, Lasjaunias P, Berenstein A. Therapeutic management of vein of Galen aneurysmal malformations. In: Vinuela F, Halbach VV, Dion JE, eds. Interventional Neuroradiology: Endovascular Therapy of the Central Nervous System. New York: Raven Press; 1992:113–127. 47. Alvarez H, Garcia Monaco R, Rodesch G, Sachet M, Krings T, Lasjaunias P. Vein of Galen aneurysmal malformations. Neuroimaging Clin N Am. 2007;17:189–206. 48. Lasjaunias P. Vascular Diseases in Neonates, Infants and Children: Interventional Neuroradiology Management. Berlin: Springer; 1997. 49. Lasjaunias P. Vascular Diseases in Neonates, Infants and Children: Interventional Neuroradiology Management. 2nd edition. Berlin: Springer; in press. 50. Yasargil MG. Microneurosurgery IIIB. New York: Thieme Medical Publishers; 1988:323–357. 51. Fournier D, Rodesch G, Ter Brugge K, et al. Acquired mural (dural) arteriovenous shunt of the vein of Galen: report of 4 cases. Neuroradiology. 1991;33:185–192. 52. Lasjaunias P, Ter Brugge K, Lopez Ibor, et al. The role of dural anomalies in vein of Galen aneurysms: report of six cases and review of the literature. AJNR Am J Neuroradiol. 1987;8:185–192. 53. Jagadeesan BD, Delgado Almandoz JE, Moran CJ, Benzinger TL. Accuracy of susceptibility-weighted imaging for the detection oa arteriovenous shunting in vascular malformations of the brain. Stroke. 2011;42:87–92. 54. Krings T. Geibprasert S, Luo CB, Bhattacharya JJ, Alvarez H, Lasjaunias P. Segmental neurovascular syndromes in children. Neuroimaging Clin N Am. 2007;17;245–258. 55. Bergstrand H, Olivecrona H, Tonnis W. Gefa ss- missbildungen und Gefaessgeschwulste des Ge- hirns. Leipzig (Germany): Thieme; 1936.

26 Pediatric MR Angiography: Principles and Applications 56. Bonnet P, Dechaume JP, Blanc E (1937) Lanevrisme cir-soide de la retine. Ses relations avec lanevrysme cirsoide du cerveau [in French]. Journal de Me decine de Lyon. 18:165–178. 57. Wyburn-Mason R. Arteriovenous aneurysm of midbrain and retina, facial naevi and mental changes. Brain. 1943;66:163–203. 58. Bhattacharya JJ, Luo CB, Alvarez H, et al. PHACES syndrome: a review of eight previously unreported cases with late arterial occlusions. Neuroradiology. 2004;46:227–33. 59. Lasjaunias P, Ter Brugge K, Berenstein A. Surgical Neuroangiography. Berlin: Springer; 2006. 60. Griffiths PD. Sturge-Weber syndrome revisited: the role of neuroradiology. Neuropediatrics. 1996;27:284–294. 61. Truhan AP, Filipek PA. Magnetic resonance imag- ing. Its role in the neuroradiologic evaluation of neurofibromatosis, tuberous sclerosis, and Sturge-Weber syndrome. Arch Dermatol. 1993;129: 219–226. 62. Hu J, Yu Y, Juhasz C, et al. MR susceptibility weighted imaging (SWI) complements conventional contrast enhanced T1 weighted MRI in characterizing brain abnormalities of Sturge-Weber Syndrome. J Magn Reson Imaging. 2008;28:300–307. 63. Ziyeh S, chumacher M, Streker R, Rossler J, Hochmuth A, Klisch J. Head and neck vascular malformations: time-resolved MR projection angiography. Neuroradiology. 2003;45:681–686. 64. Rollins N, Ison C, Reyes T, Chia J. Cerebral MR venography in children: comparison of 2D time-o-flight and gadolinium enhanced 3D gradient-echo techniques. Radiology. 2005;235:1011–1017. 65. Widjaja E, Shroff M, Blaser S, Laughlin S, Raybaud C. 2D time-offlight MR venography in neonates: anatomy and pitfalls. AJNR Am J Neuroradiol. 2006;27:1913–1918. 66. Meckel S, Reisinger C, Bremerich J, et al. Cerebral venous thrombosis: diagnostic accuracy of combined, dynamic and static, contrast-enhanced 4D MR venography. AJNR Am J Neuroradiol. 2010;31:527–535. 67. Schoenberg BS, Mellinger JF, Schoenberg DG. Cerebrovascular disease in infants and children: a study of incidence, clinical features, and survival. Neurology. 1978;28:763–768. 68. Broderick J, Talbot GT, Prenger E, et al. Stroke in children within a major metropolitan area: the surprising importance of intracerebral hemorrhage. J Child Neurol. 1993;8:250–255. 69. Giroud M, Lemesle M, Gouyon JB, et al. Cerebrovascular disease in children under 16 years of age in the city of Dijon, Franc: a study of incidence and clinical features from 1983 to 1993. J Clin Epidemiol. 1995;48:1343–1348. 70. deVeber G. Stroke and the child’s brain: an overview of epidemiology, syndromes and risk factors. Curr Opin Neurol. 2002;15:133–138. 71. Fullerton HJ, Wu YW, Sidney S et al. (2007) Risk of recurrent childhood arterial ischemic stroke in a population-based cohort: the importance of cerebrovascular imaging. Pediatrics. 119(3); 495–501. 72. Kirkham F, Sebire G, Steinlin M, et al. Arterial ischemic stroke in children. Thromb Haemost. 2004;92:697–706. 73. Braun KP, Bulder MM, Chabrier S, Kirkham FJ, Uiterwaal CS, Tardieu M, Sébire G. The course and outcome of unilateral

379

74. 75.

76. 77.

78.

79.

80.

81.

82.

83.

84.

85. 86.

87.

88.

intracranial arteriopathy in 79 children with ischaemic stroke. Brain. 2009;132(Pt 2):544–557. Scott RM, Smith ER. Moyamoya disease and Moyamoya syndrome. N Engl J Med. 2009; 360:1226–1237. Singla M, John E, Hidalgo G, et al. Moyamoya vasculopathy in a child after hemolytic uremic syndrome: a possible etiopathogenesis. Neuropediatrics. 2008;39:128–130. Rea D, Brandsema J, Armstrong D, et al. Cerebral arteriopathy in children with neurofibromatosis type 1. Pediatrics. 2009;124:e476-e483. Morioka M, Hamada JI, Kawano T, Todaka T, Yano S, Kai Y, Ushio Y. Angiographic dilatation and branch extension of the anterior choroidal and posterior communicating arteries are predictors of hemorrhage in adult Moyamoya patients. Stroke. 2003;34;90–95. Fushimi Y, Miki Y, Kikuta K, et al. Comparison of 3.0- and 1.5-T Three-dimensional time-of-flight MR angiography in Moyamoya disease: preliminary experience. Radiology. 2006;239:232–239. Yoon HK, Shin HJ, Lee M, Byun HS, Na DG, Han BK. MR angiography of Moyamoya disease before and after Encephaloduroarteriosynangiosis. AJR Am J Roentgenol. 2000;174:195–2000. Neff W, Horn P, Schmiedek P, Duber C, Dinter DJ. 2D Cine phasecontrast MRI for volume flow evaluation of the brain-supplying circulation in Moyamoya disease. AJR Am J Roentgenol. 2006; 187:W107-W116. Aviv RI, Benseler SM, DeVeber G, Silverman ED, Tyrrell PN, Tsang LM, Armstrong D. Angiography of primary central nervous system angiitis of childhood: conventional angiography versus magnetic resonance angiography at presentation. Am J Neuroradiol. 2007;28:9–15. Eleftheriou D, Cox T, Saunders D, Klein NJ, Brogan PA, Ganesan V. Investigation of childhood central nervous system vasculitis: magnetic resonance angiography versus catheter cerebral angiography. Dev Med Child Neurol. 2010;52:863–867. Mascalchi M, Quilici N, Ferrito G, et al. Identification of the feeding arteries of spinal vascular lesions via phase-contrast MR angiography with three-dimensional acquisition and phase display. AJNR Am J Neuroradiol. 1997;18:351–358. Saraf-Lavi E, Bowen BC, Quencer RM, et al. Detection of spinal dural arterio-venous fistulae with MR imaging and contrastenhanced MR angiography: sensitivity, specificity, and prediction of vertebral level. AJNR Am J Neuroradiol. 2002;23:858–867. Backes WH, Nijenhuis RJ. Advances in spinal cord MR angiography. Am J Neuroradiol. 2008;29:619–631. Mull M, Nijenhuis RJ, Backes WH, Krings T, Wilmink JT, Thron A. Value and limitations of contrast-enhanced MR angiography in spinal arteriovenous malformations and dural arteriovenous fistulas. AJNR Am J Neuroradiol. 2007;28:1249–1258. Jaspers K, Nijenhuis RJ, Backes WH. Differentiation of spinal cord arteries and veins by time-resolved MR angiography. J Magn Reson Imaging. 2007;26:31–40. Hyodoh H, Shirase R, Akiba H, et al. Double-subtraction maximum intensity projection MR angiography for detecting the artery of Adamkiewicz and differentiating it from the drainage vein. J Magn Reson Imaging. 2007;26:359–365.

wwwwwwwwwwww

Contrast Agents for MR Angiography Christoph U. Herborn

Recent years have seen the rapid development of techniques and applications for contrast-enhanced magnetic resonance angiography (CE-MRA) and the growing acceptance of the method in clinical routine. In addition to advances in hardware and software design, especially the development in the field of contrast media for CE-MRA has triggered the modality to increasingly become the diagnostic standard of reference for vascular imaging. Despite the majority of MR contrast agents is still lacking the direct approval for CE-MRA, this chapter reviews the properties and characteristics of both currently available agents and those which have been developed recently and are now on their way to the clinical market. Two major groups of MR contrast agents for CE-MRA can be distributed: On the one hand, paramagnetic agents that mostly rely on gadolinium and on the other hand superparamagnetic agents that are based on iron oxide particles. Paramagnetic agents can be further divided into a group with weak or strong interaction with serum proteins and a group without such interaction [1–6]. In addition, there are macromolecular agents that tend to stay longer in the blood pool. Superparamagnetic agents, on the contrary, can be divided upon their particle size and coating into groups of ultrasmall size (ultrasmall particles of iron oxide, USPIO) or groups coated with dextran or starch [7–12]. Figure 27.1 and Table 27.1 summarize the various agents and their brand names and physical properties, respectively.

Paramagnetic Contrast Agents At present, up to nine gadolinium-based contrast agents are approved worldwide and might be used for CE-MRA. Six of these agents do not show any interaction with blood serum C.U. Herborn, MD, MBA () University Medical Center Hamburg-Eppendorf, Martinistrasse 52, D-20251 Hamburg, Germany e-mail: [email protected]

27

and are therefore mainly used for first pass CE-MRA, i.e., acquisition of data during the first pass of the agents through the arterial bed after peripheral venous injection. A rather new agent, Gadobutrol (Gadovist), possesses a higher molarity than the other agents (1.0 M as compared to 0.5 M) and might have advantages over the other agents due to a smaller bolus size [7]. Two other agents, Gadobenate Dimeglumine (MultiHance) and Gadofosveset trisodium (Ablavar), have either a weak and transient or a strong interaction to serum in the blood, respectively. Thereby, these two agents remain longer in the blood pool than any of the aforementioned agents, which tend to diffuse into the interstitium after the first transit through the arterial and especially the venous bed [13–21]. Chemical structures of the agents are shown in Fig. 27.2.

Non-protein-Binding MR Contrast Agents The agents without any interactions with serum proteins are considered nonspecific and extracellular; they are excreted through the kidneys by glomerular filtration. With regard to T1 relaxation rates, a strong indicator for signal intensities in CE-MRA, these agents are comparable with values ranging between 4.3 and 5.6 L/mmol/s. Therefore, all these agents show more or less equivocal results in their vascular imaging performance with good quality and diagnostic performance. The decision for or against a certain dye has been made on reports about adverse events or comparatively subjective results at various institutions. The vast majority of scientific evaluations of CE-MRA has been made with these nonprotein-binding chelates and – as a matter of fact – these agents still are mostly used for CE-MRA examinations in routine practice. As mentioned before, Gadobutrol (Gadovist) has several advantages over the other non-protein-binding agents due to its higher molarity, the lower viscosity of the dye and finally the compact and concentrated bolus that achieves higher intravascular signal.

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_27, © Springer Science+Business Media, LLC 2012

381

382

C.U. Herborn

Fig. 27.1 Paramagnetic and superparamagnetic agents

Table 27.1 Brand names and physical properties of paramagnetic and superparamagnetic agents Molarity (mol/L) Molecular structure Thermodynamic stability constant (log Keq) Osmolarity (Osm/kg) Viscosity (mPas @ 37°C) T1 relaxivity (L/mmol/s), plasma

Magnevist Dotarem ProHance Omniscan Gadovist OptiMARK 0.5 0.5 0.5 0.5 1.0 0.5 Linear, ionic Cyclic, ionic Cyclic, non-ionic Linear, non-ionic Cyclic, non-ionic Linear, non-ionic

MultiHance 0.5 Linear, ionic

22.1

25.8

23.8

16.9

21.8

16.6

22.6

1.96

1.35

0.63

0.65

1.6

1.11

1.97

2.9

2.0

1.3

1.4

4.96

2.0

5.3

4.9

4.3

4.6

4.8

5.6

N/A

9.7

While early CE-MRA examinations were focused on single stations in the pelvis or the lower leg, the introduction of multistation CE-MRA with moving tables and bolus-chase techniques, have pushed the application toward the coverage of large vascular territories, including whole-body MRA from head to toe. The typical dose used for these examinations ranges between 0.1 and 0.3 mmol/kg bodyweight. Recent guidelines reflecting the risk associated with higher doses of gadolinium-based agents (see below) recommend

doses between 0.1 and 0.2 mmol/kg bodyweight, which still permits good image quality. All non-protein-binding agents have been used for various CE-MRA indications and have produced comparable image quality to that achieved with selective DSA. However, rapid loss of intravascular signal due to leakage of the agents into interstitial tissue has lead to both refinement of patient preparation with cuffs for femoral or even pelvic compression and to the development of contrast agents with protein interaction.

27

Contrast Agents for MR Angiography

383

Ablavar was previously known as vasovist. Both agents present with a significantly higher T1 relaxivity after their binding to albumin than compared to values achieved by the non-protein-binding chelates. This finding translates into more signal from the vascular bed which is potentially helpful for the assessment of small peripheral vessels. In addition, smaller doses of gadolinium seem to be applicable for virtually all vascular territories from the carotids to the feet [22–29]. Gadofosveset causes an extended signal increase from the vascular bed due to a strong and enduring noncovalent tie with albumin in the blood thereby permitting CE-MRA during a more prolonged timeframe after the initial first arteriovenous transit. Due to this binding, a gadolinium dose of merely 0.03 mmol/kg bodyweight is sufficient for diagnostic quality comparable to DSA examinations of the lower limb arteries. The phase of equilibrium distribution of contrast material in both arteries and veins is called steady-state and might open new windows for artery display with high spatial resolution. Furthermore, the agent might be used for the ever promising approach of CE-MRA toward the coronary arteries [30, 31].

Macomolecular Blood Pool Agents All aforementioned paramagnetic agents have a similar small size which allows for a rapid diffusion out of the vascular bed in case of lacking interaction with serum protein. Some gadolinium-based agents, however, have a macromolecular structure that prevents a fast leakage into the interstitium. Representatives of this group are P792 and gadomer-17 with molecular weights between 6.5 and 35 kDa, respectively, which is vastly bigger than the size of the other agents that range between 0.56 and 1.0 kDa. Nevertheless, the macromolecular agents are filtered through the glomerulus and excreted unmetabolized. These agents are still in a preclinical phase of evaluation, very first results hint at a potential use for CE-MRA of the coronary arteries.

Fig. 27.2 Chemical structures of 6 paramagnetic contrast agents. (a) Gadopentate. (b) Gadobutrol. (c) Gadodiamide. (d) Gadoteridol. (e) Gd-DOTA. (f) Gd-DTPA

Protein-Binding MR Contrast Agents Generally, Gadobenate Dimeglumine (MultiHance) with a transient and weak interaction with serum protein must be discriminated from Gadofosveset trisodium (ablavar), which shows a relatively strong binding interaction with protein.

Risk Profile of Paramagnetic Contrast Agents for MRA Gadolinium-based contrast agents have proved to be among the safest available for clinical imaging. In particular, with the doses used for CE-MRA, there are low rates of occurrence of nephrotoxicity and allergic reactions compared with those rates for iodinated contrast agents used for DSA or computed tomography angiography (CTA). Until recently, gadolinium-based paramagnetic contrast agents were used especially in patients with renal failure and patients undergoing

384

dialysis. In these patients, doses above 0.3 mmol/kg bodyweight have been injected. Interestingly, it was demonstrated that some paramagnetic contrast agents, namely gadodiamide and gadoversetamide interfere with the colorimetric test for serum calcium, resulting in spurious hypocalcemia in routine laboratory testing. This false measure was caused by free gadolinium that binds with calcium. As calcium metabolism becomes more important in renal failure, this observation is critical in this subgroup of frequently severely ill patients. In addition, this finding was shown to be a greater problem with higher doses of gadolinium, e.g., as given for CE-MRA of the renal arteries. Stability of gadolinium chelates also matters when dealing with potential tetention of the ion in the body following possible transmetallation and the release of a free gadolinium3+ ion. The recent discovery of an association between the administration of gadolinium-based contrast agents and nephrogenic systemic fibrosis (NSF) has also changed the way that gadolinium is used. High doses above 0.2 mmol/kg bodyweight are now rare because most applications have been adjusted to a standard dose of 0.1 mmol/kg, especially in patients undergoing dialysis or patients with a glomerular filtration rate (GFR) of less than 30 mL/min. When patients are undergoing dialysis, CE-MRA with gadolinium is ideally scheduled for just before the next dialysis appointment to facilitate prompt clearance. NSF is a rare fibrosing condition occurring in patients with profound renal failure or patients undergoing dialysis. In NSF, patches of skin become thickened and tethered to the underlying tissue, reducing range of motion and leading to contractures. The fibrosing process can also involve internal organs, including the lungs, heart, and muscles. Although many cases are mild and limited to dermatologic manifestations, an estimated 5% of cases have a more fulminant course resulting in death. Because treatment options are limited, an emphasis on prevention has been under way, limiting gadolinium exposure in patients with severe renal failure (estimated GFR 30 mL/min/1.73 m2. Therefore, it is critical to identify patients with severe CRF and those in ARF prior to administering GBCAs. In otherwise healthy and stable patients, as applies to most outpatients and some inpatients who are medically stable, routine serum creatinine screening is probably not warranted and is not recommended by either the Food and Drug Administration (FDA) [21] or the Canadian Association of Radiologists (CAR) [23]. However, a detailed screening questionnaire to identify potential risk factors for CRF should be obtained in all patients. If a risk factor (i.e., history of renal disease, dialysis, solitary kidney, transplant; age > 60; hypertension; diabetes; ischemic heart disease; stroke; vascular disease; organ transplant; myeloma; chemotherapy) is identified, then screening with serum creatinine and calculation of eGFR using standard formulae (such as a Cockroft-Gault) should be performed. If the eGFR is >30 mL/min/1.73 m2 in this group, the risk of NSF is either absent or extremely low and it is generally felt that it is appropriate to proceed with GBCA, and detailed patient counseling regarding risk is probably not needed. If a highdose GBCA exam is going to be performed, consideration should be given to routine screening. In the case of inpatients, particularly unstable patients, consideration should be given to screening even if the questionnaire is negative. Finally, in any patient where there is suspicion of acute renal injury, serum Cr should be drawn and a nephrology consultation should be strongly considered keeping in mind that the eGFR may be unreliable. In patients at high risk for NSF (eGFR < 30 and those with ARF), it must be first determined whether GBCA-enhanced exam is really necessary and whether other imaging tests may provide the needed information. Specifically pertaining to vascular MR imaging, there has been significant improvement in noncontrast MRA techniques, including time-offlight MRA, phase-contrast imaging, arterial spin labeling, and balanced steady-state free precession techniques. In some cases, Doppler ultrasound or functional imaging with scintigraphy may provide adequate information. Consideration should also be given to CT angiography which is performed using iodinated contrast agents. In patients with chronic endstage (anuric) renal failure on dialysis, CT angiography is usually a viable alternative; however, in patients with CRF with residual renal function, the risk of iodinated contrast

28

CE-MRA in the Age of Nephrogenic Systemic Fibrosis

media-induced nephropathy must be weighed against the risks posed by GBCAs. Finally, in some instances (i.e., where intervention is highly likely), proceeding directly to DSA with iodinated contrast agent may also be appropriate. The technique of CO2 angiography is also occasionally used, but has many limitations. However, if the information from CEMRA is deemed necessary, i.e., the risks associated with CEMRA are outweighed by the benefit of the information that is obtained, at-risk patients should not be simply denied necessary imaging. At our institution, a nephrology consultation is obtained and the decision to proceed is based on collaborative decision making among referring team, nephrology, and radiology, taking account of the risks and benefits of the procedure. A detailed consent discussion with the patient is mandatory. Once the decision to proceed has been made, it is suggested that Gadodiamide, Gadopentetate dimeglumine, and Gadoversetamide be avoided as these are the agents that have been convincingly associated with NSF. The lowest possible dose of GBCA providing a diagnostic study should be used and repeated CEMR studies in short time intervals ought to be avoided. GBCA should not be used for CT or DSA studies as an alternative to iodinated contrast in patients with renal dysfunction, as was occasionally done in the past when the risk of NSF was not known. Patients should be instructed of the signs and symptoms of NSF and to seek medical attention should they develop. Published data suggests that prompt hemodialysis enhances GBCA elimination (over 98% free Gd is removed after three sessions) while it appears that peritoneal dialysis is not very effective in removing GBCA. It is, therefore, strongly recommended that prompt hemodialysis following administration of GBCAs be considered, although it remains unproven whether this reduces the risk of development of NSF. It is also unclear whether longer dialysis sessions would be of benefit. At our institution, GBCA exams in at-risk patients are coordinated with the nephrology team to facilitate hemodialysis immediately following the exam whenever possible. There is currently no data to support the use of renal protective protocols, such as bicarbonate administration or aggressive hydration in at-risk patients [13].

Conclusion In this chapter, we have reviewed the clinical features of NSF, risk factors for its development, and its relationship to GBCAs and severe renal dysfunction. Current recommendations and prevention strategies have been discussed with particular attention to how this relates to MRA applications. Although the emergence of NSF has increased the complexity of decision making in CEMRA imaging, it is reassuring that measures taken to reduce the use of GBCA in at-risk

393

patients seem to have stemmed the tide of the disease, evidenced by the dwindling number of cases reported in the literature. The NSF story serves as an important reminder, however, that even the most seemingly safe and widely used medical/imaging procedures are accompanied by risk, some that may be unknown – and therefore should be used judiciously.

References 1. LeBoit PE. What nephrogenic fibrosing dermopathy might be. Arch Dermatol. 2003;139:928–930. 2. Cowper SE, Robin HS, Steinberg SM, Su LD, Gupta S, LeBoit PE. Scleromyxoedema-like cutaneous diseases in renal-dialysis patients. Lancet. 2000;356:1000–1001. 3. Mayr M, Burkhalter F, Bongartz G. Nephrogenic systemic fibrosis: clinical spectrum of disease. J Magn Reson Imaging. 2009;30: 1289–1297. 4. Jimenez SA, Artlett CM, Sandorfi N, et al. Dialysis associated systemic fibrosis (nephrogenic fibrosing dermopathy): study of inflammatory cells and transforming growth factor beta1 expression in affected skin. Arthritis Rheum. 2004;50:2660–2666. 5. Juluru K, Vogel-Claussen J, Macura KJ, Kamel IR, Steever A, Bluemke DA. MR imaging in patients at risk for developing nephrogenic systemic fibrosis: protocols, practices, and imaging techniques to maximize patient safety. Radiographics. 2009;29:9–22. 6. Prince MR, Zhang HL, Prowda JC, Grossman ME, Silvers DN. Nephrogenic systemic fibrosis and its impact on abdominal imaging. Radiographics. 2009;29:1565–1574. 7. Grobner T. Gadolinium – a specific trigger for the development of nephrogenic fibrosing dermopathy and nephrogenic systemic fibrosis? Nephrol Dial Transplant. 2006;21:1104–1108. 8. High WA, Ayers RA, Chandler J, Zito G, Cowper SE. Gadolinium is detectable within the tissue of patients with nephrogenic systemic fibrosis. J Am Acad Dermatol. 2007;56:21–26. 9. Wiginton CD, Kelly B, Oto A, et al. Gadoliniumbased contrast exposure, nephrogenic systemic fibrosis, and gadolinium detection in tissue. AJR Am J Roentgenol. 2008;190:1060–1068. 10. Weinreb JC, Abu-Alfa AK. Gadolinium-based contrast agents and nephrogenic systemic fibrosis: why did it happen and what have we learned? J Magn Reson Imaging. 2009;30:1236–1239. 11. Idée JM, Port M, Robic C, Medina C, Sabatou M, Corot C.Role of thermodynamic and kinetic parameters in gadolinium chelate stability. J Magn Reson Imaging. 2009;30:1249–1258. 12. Aime S, Caravan P. Biodistribution of Gadolinium-based contrast agents, including gadolinium deposition. J Magn Reson Imaging. 2009;30:1259–1267. 13. Leiner T, Kucharczyk W. NSF prevention in clinical practice: summary of recommendations and guidelines in the United States, Canada, and Europe. J Magn Reson Imaging. 2009;30:1357–1363. 14. Sieber MA, Lengsfeld P, Frenzel T, Golfier S, Schmitt-Willich H, Siegmund F, Walter J, Weinmann HJ, Pietsch H. Preclinical investigation to compare different gadolinium-based contrast agents regarding their propensity to release gadolinium in vivo and to trigger nephrogenic systemic fibrosis-like lesions. Eur Radiol. 2008;18:2164–2173. 15. Sieber MA, Steger-Hartmann T, Lengsfeld P, Pietsch H. Gadoliniumbased contrast agents and NSF: evidence from animal experience. J Magn Reson Imaging. 2009;30:1268–1276. 16. Varani J, DaSilva M, Warner RL, Deming MO, Barron AG, Johnson KJ, Swartz RD. Effects of gadolinium-based magnetic resonance imaging contrast agents on human skin in organ culture and human skin fibroblasts. Invest Radiol. 2009;44:74–81.

394 17. Newon BB, Jimenez SA. Mechanism of NSF: New evidence challenging the prevailing theory. J Magn Reson Imaging. 2009; 30:1277–1283. 18. Prince MR, Zhang HL, Roditi GH et al. Risk factors of NSF: a literature review. J Magn Reson Imaging. 2009;30:1298–1308. 19. Bryant BJ 2nd, Im K, Broome DR. Evaluation of the incidence of nephrogenic systemic fibrosis in patients with moderate renal insufficiency administered gadobenate dimeglumine for MRI. Clin Radiol. 2009;64:706–713. 20. United States Food and Drug Administration. Information for Healthcare Professionals Gadolinium-Based Contrast Agents for Magnetic Resonance Imaging (marketed as Magnevist, MultiHance, Omniscan, OptiMARK, ProHance). Washington, D.C. http://www. fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationfor PatientsandProviders/ucm142884.htm. Accessed March 22, 2010. 21. American College of Radiology (ACR). Manual on Contrast Media, 6.0 edition. American College of Radiology; 2008:54–56. www.acr.

A. Bharatha et al.

22.

23.

24.

25.

org/SecondaryMainMenuCategories/quality_safety/contrast_manual. aspx. Accessed March 22, 2010. Padilla-Thornton A, Rafat Zand K, Barrett B, Stein L, Andrew G, Forster BB. Canadian Association of Radiologists national advisory on gadolinium administration and nephrogenic systemic fibrosis. Can Assoc Radiol J. 2008;59:237–240. European Society of Urogenital Radiology. ESUR guideline: gadolinium-based contrast media and nephrogenic systemic fibrosis. Vienna, Austria. http://www.esur.org/Nephrogenic_Fibrosis.39.0. html. Accessed March 22, 2010. Miki Y, Isoda H, Togashi K. Guideline to use gadolinium-based contrast agents at Kyoto University Hospital. J Magn Reson Imaging. 2009;30:1364–1365. Laurent S, Elst LV, Muller RN. Comparative study of the physicochemical properties of six clinical low molecular weight gadolinium contrast agents. Contrast Media Mol Imaging. 2006;1: 128–137.

Emerging Interventional MR Applications

29

Clifford R. Weiss, Aravindan Kolandaivelu, Jeff Bulte, and Aravind Arepally

Introduction The desire to help patients without doing harm has driven medicine to develop minimally invasive methods for diagnosing and treating disease. It is not surprising, then, that over the past three decades medicine has evolved an increasing emphasis on image-guided intervention. Traditionally, these interventions have been performed using fluoroscopy, ultrasound, and computed tomography. Most recently, however, radiologists’ interventional skills and trends toward minimally invasive surgery have converged to create a burgeoning interest in the use of magnetic resonance imaging for guidance in interventional procedures, including the delivery of cellular therapeutics [1]. Initial clinical applications were seen for neurosurgery, where intraoperative magnetic resonance imaging (MRI) systems with movable magnets provided unique glimpses into the operative field [2, 3]. With the advent of 3D gradient echo techniques, further applications for vascular and nonvascular frontiers have appeared [4]. The feasibility of angioplasty of the aorta and iliac arteries with MR tracking has been demonstrated in animal models [5, 6]. C.R. Weiss, MD () Division of Cardiovascular and Interventional Radiology, The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins Hospital, Blalock 544, 600 N. Wolfe St., Baltimore, MD 21287, USA A. Kolandaivelu, MD Division of Cardiology, Cardiac Arrhythmia Service, Johns Hopkins Hospital, Carnegie 568, 600 N. Wolfe St., Baltimore, MD 21287, USA J. Bulte, PhD Division of MR Research, The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins Medical Institutes, Blalock 544, 600 N. Wolfe Street, Baltimore, MD 21287, USA A. Arepally, MD Division of Interventional Radiology, Piedmont Healthcare, 1984 Peachtree Road, Suite 505, Atlanta, GA 30309, USA

However, despite the significant technological leaps with this modality, one main criticism has been the lack of any major clinical implementation. Initial efforts to replicate prior fluoroscopic procedures have failed and have not resulted in any clinical impact. More recently, there have been advancements in vascular interventional MRI for emerging clinical applications that previously were not feasible or were limited when performing conventional X-ray imaging. This chapter discusses two emerging vascular applications for interventional vascular MRI: MR-guided atrial fibrillation (AF) ablation and MR-guided cellular therapy.

Emerging Clinical Applications MR-Guided Atrial Fibrillation Ablations Atrial Fibrillation Background Atrial fibrillation is the most common clinically relevant arrhythmia, affecting 1 in 200 people in the general population and 1 in 10 people over the age of 75. The principal morbidities associated with AF are stroke due to embolization of atrial thrombus and symptoms related to poor heart rate regulation with resting heart rates commonly over 110 beats per minute. In 1998, Haissaguerre and colleagues [7] reported that triggering foci for AF typically arise from one or more pulmonary veins. Circumferential ablation around the pulmonary vein connections to the atrium can block the exit of these triggers and has emerged as a primary goal of atrial fibrillation catheter ablation [8]. AF ablation can achieve success rates of 80% in patients with intermittent AF and an otherwise normal heart. However, multiple procedures are commonly required to achieve this and the success rate drops to 50% or less for the more chronic forms of AF associated with age and ischemic, hypertensive, and valvular heart disease [9]. There also remains a 4% risk of significant complications, including cardiac perforation, pulmonary vein stenosis, stroke, and the rare but potentially lethal risk of atrial-esophageal fistula formation [10].

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0_29, © Springer Science+Business Media, LLC 2012

395

396

Intraprocedural MRI Applications Performing electrophysiology procedures within the MRI scanner has a number of potential advantages toward the goal of improving ablation efficacy and safety. First is the ability to visualize ablation lesions with high spatial resolution. The ability to identify regions of incomplete ablation could permit targeting of additional treatment during the procedure. Second, MRI permits catheter position and contact to be visualized relative to soft tissue structures. Catheter– tissue contact is an important factor for efficient lesion formation that is poorly assessed by X-ray fluoroscopy [11]. The ability to provide continuous visualization of soft tissue anatomy without ionizing radiation exposure may also permit more precise and safe catheter manipulation. Third, the 3D MRI anatomy and myocardial scar images used to guide ablation procedures can be acquired at the time of the procedure in the same coordinate system as the “real-time” images used for catheter guidance. This avoids the errors and time delay currently introduced by registering preacquired 3D image data to a separate electrospatial mapping (ESM) catheter tracking system. Over the past 15 years, the basic techniques to enable fully MRI-guided EP procedures have been developed. Real-Time Imaging Lardo and colleagues [12] introduced the potential of using “real-time” MRI (rtMRI) for guiding EP procedures in 2000. Continuous MRI at 1 fps was used to guide a nonferromagnetic EP catheter from an internal jugular vein to selected locations in the right atrium and right ventricle. They also demonstrated the ability to perform and monitor creation of an ablation lesion in the MRI scanner. Delivery of RF ablation energy during imaging caused significant MR image degradation. However, this noise could be adequately suppressed by low-pass filtering of the ablation energy source. After ablation, imaging showed the lesion position and extent using both DEMRI and T2-weighted imaging techniques. Nazarian et al. [13] subsequently demonstrated the ability to use rtMRI to direct a catheter to standard electrophysiology study locations and record satisfactory intracardiac electrograms during rtMRI scanning. Importantly, this was the initial report of performing rtMRI-guided electrophysiology procedure in patients. Recently, Hoffmann and colleagues [14] reported the feasibility of performing a full ablation protocol under five image per second rtMRI guidance. The cavotricuspid isthmus (CTI) is a region between the tricuspid valve and inferior vena cava that can be ablated to cure the typical atrial flutter arrhythmia. Using a nonferromagnetic ablation catheter, guidance to and ablation of the entire CTI region were performed. Completion of CTI ablation was guided by T2-weighted imaging. After MRI-guided ablation, the current clinical end point of conduction block across the CTI was demonstrated in 15 of 18 procedures.

C.R. Weiss et al.

Lesion Imaging These promising but imperfect results allude to the need for better intraprocedure MR lesion assessment techniques. Acute interstitial edema likely leads to the hyperintense region on T2-weighted MRI that corresponds to acute RF ablation damage; however, this region may overestimate the region of ablation-induced tissue necrosis [15]. RF ablation lesions may also be visualized by noncontrast T1-weighted imaging, but the lesion contrast-to-noise ratio is poor compared with T2-weighted imaging or DEMRI [15]. Gadolinium contrast DEMRI appears capable of providing good lesion visualization; however, the limit on total gadolinium administration limits its use for serial lesion monitoring during the procedure [16]. MR thermography is a promising technique that utilizes the decrease in proton resonance frequency with increasing temperature to estimate the region of heatinginduced tissue necrosis [17]. This technique has been used to follow tumor ablation in the uterus, liver, prostate, and brain using diverse energy sources, including RF, but is sensitive to motion. Its use for following RF ablation in the beating heart is being investigated. Device Tracking Another area of active investigation is improved techniques for rtMRI guidance of complex arrhythmia ablation. While fluoroscopy provides projection images where the entire catheter body and tip are easily visualized, standard 2D rtMRI only depicts a slice through the body that is 5–10-mm thick. Curved devices, such as catheters, may pass in and out of the MR image plane and lead to misinterpretation of the device position. One approach to this problem is to track the catheter tip position to provide similar information to that of current ESM systems. The currently favored method was first described by Dumoulin et al. [18] and uses rapid 1D projection rtMRI to identify the 3D position of a small receiver coil located in the catheter tip. Using this method, multicoil catheter designs have been used for MRI catheter guidance in the atrium and ventricles to target ablation and obtain surface maps of myocardial electrical properties [19, 20]. This position tracking technique has also been interleaved with rtMRI sequences to automatically move the image plane position to the catheter tip location during device manipulation. This technique likely represents the next step in rtMRI catheter guidance. Safety/Devices MR-guided intravascular procedures raise a number of safety concerns beyond the standard SAR tissue heating concerns associated with exposure to scanning-associated electromagnetic radiation [21]. In addition to avoiding ferromagnetic materials that could experience significant forces when brought close to the scanner, catheters must be designed to avoid RF transmission-induced current that can lead to significant

29

Emerging Interventional MR Applications

heating [22]. A number of techniques have been developed to avoid MRI-induced heating of interventional EP devices, including the use of nonmetallic polymer materials for structural elements, high-resistance alloys, RF chokes, and transformer transmission lines [23–26]. Development of devices for safe MRI-guided EP procedures requires cooperation among academic centers, imaging companies, and regulatory agencies.

MR-Guided Cellular Therapy Overview of Cellular Therapy Analogous to MR-guided ablation of atrial fibrillation, image-guided cellular therapies necessitate visualization of the delivery system and also the agents being delivered. Physicians have used image guidance for delivery of various therapeutic agents for over three decades. The main goal of all local delivery is to increase the concentration of a specific therapeutic agent in a target tissue with minimal nontarget distribution. Compared to systemic therapy, local delivery provides a high level of therapeutic efficacy with minimal systemic effects. In a similar manner, cellular therapies can take advantage of these techniques to enhance the delivery process. Although some cellular therapies are adequately managed with systemic delivery, certain organs and conditions require a targeted approach for maximum affect. In fact due to the anatomical constraints of the vascularity of specific organs, systemic therapy is precluded and image-guided techniques are necessary to overcome some of these barriers. For example, the vasculature of the brain, liver, and pancreas has been shown to have unique anatomical boundaries that create a barrier to conventional therapies and therefore do not respond to conventional systemic therapies. With the brain, the blood–brain barrier acts as a physiologic barrier that prevents the migration and the transport of agents from the systemic vasculature. With both the liver and pancreas, there is a separate anatomic venous vascular supply, termed portomesenteric system, that is completely isolated from systemic circulation. Therefore, with disorders involving these organs (i.e., stroke, cirrhosis, and diabetes), accessing these secluded organs is critical to the success of cellular therapy. Finally, another relevant opportunity with image-guided therapy is the ability to not only deliver to target organs, but also to administer only into injured tissues. As demonstrated by multiple studies, substantial mobilization of stem cells has been demonstrated after myocardial infarction and also with liver injury. Clearly, the homing of cells to injured tissues is a major mechanism of regeneration and there has been extensive work in utilizing image-guided therapy to help foster this process by providing targeted delivery into injured tissue.

397

Cell Tracking Using MRI The clinical use of novel experimental immune and stem cell therapies calls for suitable methods that can monitor the cellular biodistribution noninvasively following administration. Among the different clinically used imaging techniques, MRI has superior spatial resolution with excellent soft tissue contrast. In order for exogenous therapeutic cells to be detected, they need to have a different contrast from endogenous cells. There are several different approaches to endow cells with MR-visible properties [27]. The most sensitive and widely used MR labels today are the superparamagnetic iron oxide nanoparticles or SPIOs. SPIOs are clinically approved and create strong, local magnetic field disturbances that spoil the MR signal leading to hypointense contrast. As of 2009, at least four clinical MRI cell tracking studies have been performed, reviewed in detail elsewhere [1]. Cells can be labeled with SPIOs, notably Feridex®, using simple incubation [28, 29], following coating with transfection agents [30] or by magnetoelectroporation [31]. The optimal labeling technique depends on the cell type (i.e., is the cell phagocytic or very hard to transfect) and its application. Paramagnetic gadolinium chelates can also be used [32], although the induced contrast can be ambigious – the positive contrast decreases at higher fields while negative, susceptibilitybased contrast takes over. It has not been used clinically and there are serious concerns about its safety for cell tracking – gadolinium is toxic once dechelated, which may occur following prolonged retention by cells in acidic compartments (i.e., lysosomes). In addition, the NSF issue has dampened enthusiasm to initiate preliminary clinical studies. Conventional approaches require the cells to be prelabeled with contrast agent before injection. An example of homing of bone marrow (BM) stem cells is shown in Fig. 29.1. Hematopoietic bone marrow cells participate in the formation of atherosclerosis. Following transplantation of Feridexlabeled BM stem cells, large MR signal voids of the aorta walls were seen that were attributed to the “blooming” effect of migrated Feridex-BM stem cells in the plaques. Homing of Feridex-labeled BM stem cells has also been demonstrated for injured arteries [33]. In the recipient mice, the left femoral arteries were injured using a cuff-constriction or endothelium-damage approach while the right femoral arteries were uninjured to serve as controls. MRI showed larger regions of hypointensity with Feridex-labeled cells at the sites of the injured arteries as compared to control arteries (p < 0.01) (see Fig. 29.2). Both studies clearly demonstrate evidence that supports the potential use of MRI to detect homing of intravenously injected BM cells to vascular abnormalities. For MRI cell tracking, an emerging application is to perform 1H MRI only for anatomical information in conjunction with the use of 19F as a tracer molecule. As there is no endogenous background signal, “hot-spot” images of the tracer can

Fig. 29.1 (a–c) Cross-sectional view of representative in vivo 4.7-T MR images of aortas from atherosclerotic ApoE mice fed a high-cholestrol diet. Insets outline the ascending aorta, showing large MR signal void of the aortic wall in animal c following Feridex-labeled bone marrow (BM) stem cell transplantation while the aortae in animals a and b appear as bright rings. (d–f) Magnification of insets of a–c. In d and e (controls, treated with no BM stem cells and unlabeled BM stem cells), the thickened aortic walls due to atherosclerotic plaques (arrows) are visualized, with the aortic walls appearing as bright rings. In f (treated

with Feridex-labeled BM stem cells), larger signal voids (open arrows) of the aortic wall are seen. (g–i) Histochemical staining for Feridex with Prussian blue (scale bar = 50 mm) and (j–l) histochemical staining for LacZ with X-gal (scale bar = 20 mm). Numerous Feridex- and LacZpositive cells (blue color, arrows) are detected in aortic tissues of mice receiving BM stem cell transplants (in k, i, and l), which are not visualized in the control aortic tissues (in g, j and h) (reproduced, with permission, from Qiu B et al. [43])

Fig. 29.2 (a–d) (upper panel) Cross-sectional view of representative in vivo 4.7-T MR images of different arteries with various treatments. Images a and c are taken from the right legs of mice, which are vertically flipped for convenient comparison. Insets outline the femoral artery areas, showing larger hypointense areas at the site of the injured femoral artery in animal d with both Feridex-labeled BM stem cell transplantation and the cuff placement while no such findings appear at the sites of the

uninjured femoral arteries in animals a and c. In animal b, the cuff itself with circulated blood creates a small hypointense circle. (e–f) (middle panel) Histochemical staining for iron with Prussian blue. (i–l) (lower panel) Histochemical staining for LacZ with X-gal. Numerous ironpositive cells (blue color in h) and LacZ-positive cells (blue color in l) are detected, which are not visualized in the control tissues (e–g and i–k). 40× (reproduced, with permission, from Gao F et al. [33])

29

Emerging Interventional MR Applications

be obtained and superimposed on the anatomical 1H images [34–36]. The strategies for intracellular labeling is similar to that used for labeling with SPIO particles: in this case, 19F particles can be simply phagocytized [37], mixed with transfection agents [35], or electroporated into cells. Finally, novel semipermeable microcapsules have been developed that immunoprotect cells and are visible with multiple modalities. Initially, X-ray-visible alginate capsules were developed that contained barium or bismuth (“X caps”) [38], but capsules containing SPIO (“magnetocapsules”) have now also been described [39]. In principle, any type of contrast agent can be coencapsulated allowing multimodality tracking. Thus, depending on the specific cellular imaging application, a quite complete set of tools exists for MRI cell tracking.

Device Tracking for Cell Therapy A promising and exciting new development for cell and drug delivery is the fusion of transvascular and percutaneous approaches for a new procedure that is best described as a hybrid technique. In this method, the vascular system (arterial or venous) is used as a conduit to perform punctures of the target tissues for local drug/cellular delivery. Compared to the previously discussed methods, this technique is technically more invasive, requires sophisticated imaging support/ devices, and experience of the operator in performing endovascular procedures. However, the main advantages are that it is less invasive than current surgical options, provides access to certain tissues and organs that are difficult to reach, and minimizes potential systemic side effects by delivering only to pathological tissues. The advantage of this technique is best seen for procedures, such as direct intramyocardial injections, which currently require surgical exposure. Due to the results of phase I clinical trials that have shown that in vivo tissue engineering of the myocardium is feasible with local surgical intramyocardial delivery, there has been tremendous research dedicated to finding a percutaneous option. The challenge with performing this procedure is to deliver the therapeutic agent only to the damaged myocardium. MRI with its inherent ability to provide real-time visualization of the myocardium without radiation or iodinated contrast has become the modality of choice for this procedure. The initial feasibility of MR-guided hybrid procedures was demonstrated by Lederman et al. [40], where a commercially available injection catheter (Stiletto™; Boston Scientific, Natick, MA) was modified for real-time, MR-guided intramyocardial injections. Using commercial real-time imaging software and a 1.5-T MR scanner, the modified Stileto™ system was readily visible during advancement, successfully oriented in the left ventricle followed by delivery of dilute gadolinium–DTPA into the myocardium of swine. Additionally, Dick et al. [41] further

399

modified the Stiletto™ injection catheter system so that the guide catheters were arranged as one RF antenna. The second RF antenna, a microcoil, was built into the distal tip of the injection needle system which created a high-intensity signal at the distal tip in order to enhance positioning before myocardial injections. Based upon this work, several other groups have also demonstrated the accuracy of the hybrid technique. Saeed et al. [42] used an XMR system along with a modified clinical catheter for myocardial delivery; in their study, X-ray was used for 2D guidance into the left ventricle from the femoral artery and 3D MR fluoroscopy was utilized for injections. Additionally, they were able to demonstrate the fundamental benefit of MRI by delivering gadolinium chelate agents only to an artificially created target of 1.5–2 cm in the myocardium. To further demonstrate this accuracy, other authors have successfully delivered iron particles (which have T1/T2 effects) and dysprosium-based contrast agents (T2/T2* effects) only into the infracted myocardium of animal models. Despite the ability of these investigators to modify commercially available catheter/injection systems, a key limitation is that current drug delivery devices are designed for fluoroscopy and not fully optimized for MR delivery. Due to those concerns, novel catheter designs for hybrid delivery under MR have now been developed. At our institution, we have developed a steerable intramyocardial injection catheter with a defl ectable distal section, which can be actively tracked and used to deliver therapeutics to target tissue under MR guidance. The components of the catheter are arranged to form a “loopless-antenna” RF receiver coil that provides a region of high signal along the length of the coil to enable active tracking. The distal tip of the catheter was modified to create a “coiled tip” that provides high-intensity signal at the distal tip. Therefore, the position of the distal tip as it apposes the target tissue can be visualized before the needle is advanced. Using this steerable myocardial injection catheter, successful targeted delivery of gadolinium contrast and ironlabeled mesenchymal stromal cells to myocardial infarct border targets was performed.

Summary As demonstrated, recently, there has been significant improvement in both imaging and hardware design that now allows for highly targeted delivery of agents. Cardiac applications, such as ablation of the atrium or delivery of stem cells to the myocardium, require a complex imaging platform to ensure proper placement of the therapeutic modality. In addition to precise delivery, MRI provides multiparametric imaging that yields anatomic, physiologic, and functional data that is currently not feasible with other imaging modalities.

400

Furthermore, with real-time tracking and feedback mechanisms from the target tissue, MRI technology significantly impacts and enables such disruptive technologies. As research continues to further integrate MRI with novel clinical procedures, including the advent of cellular therapeutics, the use of interventional vascular MRI will continue to grow.

References 1. Bulte JW. In vivo MRI cell tracking: clinical studies. AJR Am J Roentgenol. 2009;193:314–325. 2. Tronnier VM, Wirtz CR, Knauth M, et al. Intraoperative diagnostic and interventional magnetic resonance imaging in neurosurgery. Neurosurgery. 1997;40:891–900; discussion 900–892. 3. Maurer CR, Jr., Hill DL, Martin AJ, et al. Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: preliminary results. IEEE Trans Med Imaging. 1998;17:817–825. 4. Ladd ME, Quick HH, Debatin JF. Interventional MRA and intravascular imaging. J Magn Reson Imaging. 2000;12:534–546. 5. Godart F, Beregi JP, Nicol L, et al. MR-guided balloon angioplasty of stenosed aorta: in vivo evaluation using near-standard instruments and a passive tracking technique. J Magn Reson Imaging. 2000;12:639–644. 6. Becker GJ. 2000 RSNA annual oration in diagnostic radiology: the future of interventional radiology. Radiology. 2001;220:281–292. 7. Haissaguerre M, Jais P, Shah DC, et al. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Engl J Med. 1998;339:659–666. 8. Pappone C, Rosanio S, Oreto G, et al. Circumferential radiofrequency ablation of pulmonary vein ostia: A new anatomic approach for curing atrial fibrillation. Circulation. 2000;102:2619–2628. 9. Calkins H. Further insight into the technique and outcomes of “curative” catheter ablation of atrial fibrillation. Circ Arrhythm Electrophysiol. 2008;1:238–239. 10. Dong J, Calkins H. Technology insight: catheter ablation of the pulmonary veins in the treatment of atrial fibrillation. Nat Clin Pract Cardiovasc Med. 2005;2:159–166. 11. Wittkampf FH, Nakagawa H. RF catheter ablation: Lessons on lesions. Pacing Clin Electrophysiol. 2006;29:1285–1297. 12. Lardo AC, McVeigh ER, Jumrussirikul P, et al. Visualization and temporal/spatial characterization of cardiac radiofrequency ablation lesions using magnetic resonance imaging. Circulation. 2000; 102:698–705. 13. Nazarian S, Kolandaivelu A, Zviman MM, et al. Feasibility of realtime magnetic resonance imaging for catheter guidance in electrophysiology studies. Circulation. 2008;118:223–229. 14. Hoffmann BA, Koops A, Rostock T, et al. Interactive real-time mapping and catheter ablation of the cavotricuspid isthmus guided by magnetic resonance imaging in a porcine model. Eur Heart J. 2010;31:450–456. 15. Dickfeld T, Kato R, Zviman M, et al. Characterization of acute and subacute radiofrequency ablation lesions with nonenhanced magnetic resonance imaging. Heart Rhythm. 2007;4:208–214. 16. Dickfeld T, Kato R, Zviman M, et al. Characterization of radiofrequency ablation lesions with gadolinium-enhanced cardiovascular magnetic resonance imaging. J Am Coll Cardiol. 2006;47: 370–378. 17. Rieke V, Butts Pauly K. MR thermometry. J Magn Reson Imaging. 2008;27:376–390.

C.R. Weiss et al. 18. Dumoulin CL, Souza SP, Darrow RD. Real-time position monitoring of invasive devices using magnetic resonance. Magn Reson Med. 1993;29:411–415. 19. Dukkipati SR, Mallozzi R, Schmidt EJ, et al. Electroanatomic mapping of the left ventricle in a porcine model of chronic myocardial infarction with magnetic resonance-based catheter tracking. Circulation. 2008;118:853–862. 20. Schmidt EJ, Mallozzi RP, Thiagalingam A, et al. Electroanatomic mapping and radiofrequency ablation of porcine left atria and atrioventricular nodes using magnetic resonance catheter tracking. Circ Arrhythm Electrophysiol. 2009;2:695–704. 21. Shellock FG, Crues JV. MR procedures: biologic effects, safety, and patient care. Radiology. 2004;232:635–652. 22. Nitz WR, Oppelt A, Renz W, Manke C, Lenhart M, Link J. On the heating of linear conductive structures as guide wires and catheters in interventional MRI. J Magn Reson Imaging. 2001;13:105–114. 23. Weiss S, Vernickel P, Schaeffter T, Schulz V, Gleich B. Transmission line for improved RF safety of interventional devices. Magn Reson Med. 2005;54:182–189. 24. Krueger S, Schmitz S, Weiss S, et al. An MR guidewire based on micropultruded fiber-reinforced material. Magn Reson Med. 2008; 60:1190–1196. 25. Wirtz D, Lips O, David B, Krueger S, Wiess S. Diagnostic MR-electrophysiology catheter with highly resistive wires for reduction of RF-heating. In: Joint Annual Meeting ISMRM-ESMRMB. Berlin, 2007. 26. Ladd ME, Quick HH. Reduction of resonant RF heating in intravascular catheters using coaxial chokes. Magn Reson Med. 2000;43: 615–619. 27. Long CM, Bulte JW. In vivo tracking of cellular therapeutics using magnetic resonance imaging. Expert Opin Biol Ther. 2009;9: 293–306. 28. de Vries IJ, Lesterhuis WJ, Barentsz JO, et al. Magnetic resonance tracking of dendritic cells in melanoma patients for monitoring of cellular therapy. Nat Biotechnol. 2005;23:1407–1413. 29. Long CM, van Laarhoven HW, Bulte JW, Levitsky HI. Magnetovaccination as a novel method to assess and quantify dendritic cell tumor antigen capture and delivery to lymph nodes. Cancer Res. 2009;69:3180–3187. 30. Frank JA, Miller BR, Arbab AS, et al. Clinically applicable labeling of mammalian and stem cells by combining superparamagnetic iron oxides and transfection agents. Radiology. 2003;228:480–487. 31. Walczak P, Kedziorek DA, Gilad AA, Lin S, Bulte JW. Instant MR labeling of stem cells using magnetoelectroporation. Magn Reson Med. 2005;54:769–774. 32. Anderson SA, Lee KK, Frank JA. Gadolinium-fullerenol as a paramagnetic contrast agent for cellular imaging. Invest Radiol. 2006;41:332–338. 33. Gao F, Kar S, Zhang J, et al. MRI of intravenously injected bone marrow cells homing to the site of injured arteries. NMR Biomed. 2007;20:673–681. 34. Bulte JW. Hot spot MRI emerges from the background. Nat Biotechnol. 2005;23:945–946. 35. Ahrens ET, Flores R, Xu H, Morel PA. In vivo imaging platform for tracking immunotherapeutic cells. Nat Biotechnol. 2005;23: 983–987. 36. Partlow KC, Chen J, Brant JA, et al. 19 F magnetic resonance imaging for stem/progenitor cell tracking with multiple unique perfluorocarbon nanobeacons. Faseb J. 2007;21:1647–1654. 37. Ruiz-Cabello J, Walczak P, Kedziorek DA, et al. In vivo “hot spot” MR imaging of neural stem cells using fluorinated nanoparticles. Magn Reson Med. 2008;60:1506–1511.

29

Emerging Interventional MR Applications

38. Barnett BP, Kraitchman DL, Lauzon C, et al. Radiopaque alginate microcapsules for X-ray visualization and immunoprotection of cellular therapeutics. Mol Pharm. 2006;3:531–538. 39. Barnett BP, Arepally A, Karmarkar PV, et al. Magnetic resonanceguided, real-time targeted delivery and imaging of magnetocapsules immunoprotecting pancreatic islet cells. Nat Med. 2007;13: 986–991. 40. Lederman RJ, Guttman MA, Peters DC, et al. Catheter-based endomyocardial injection with real-time magnetic resonance imaging. Circulation. 2002;105:1282–1284.

401 41. Dick AJ, Guttman MA, Raman VK, et al. Magnetic resonance fluoroscopy allows targeted delivery of mesenchymal stem cells to infarct borders in swine. Circulation. 2003;108:2899–2904. 42. Saeed M, Lee R, Martin A, et al. Transendocardial delivery of extracellular myocardial markers by using combination X-ray/MR fluoroscopic guidance: feasibility study in dogs. Radiology. 2004;231:689–696. 43. Qiu B, Gao F, Walczak P, et al. In vivo MR imaging of bone marrow cells trafficking to atherosclerotic plaques. J Magn Reson Imaging. 2007;26:339–343.

wwwwwwwwwwww

Index

A Abdominal aorta and mesenteric vessels anatomy abdominal aorta, 273 celiac artery, 273 inferior mesenteric artery, 273–274 mesenteric arterial system, 273 superior mesenteric artery, 273 black blood techniques, 272 contrast-enhanced MRA contrast administration and bolus timing, 271–272 hardware considerations, 270 pulse sequence, 270–271 mesenteric arterial system, MRA acute mesenteric arterial thrombosis, 277 acute mesenteric ischemia, 275–276 acute SMA embolism, 276 aortic dissection, 277 atherosclerotic CMI, 277–279 chronic mesenteric ischemia, 277 mesenteric venous thrombosis, 277 nonatherosclerotic vascular pathology, 279 non-occlusive mesenteric ischemia, 277 MRA, abdominal aorta abdominal aortic aneurysm, 274 aortic dissections, 275 aortic occlusive disease, 274–275 aortitis, 275 congenital anomalies, 275 MRA at 3 T, 272–273 noncontrast MRA, 270 phase-contrast (PC) MRA, 269 steady-state free precession, 269–270 time-of-flight MRA, 269 time-resolved MRA, 272 transplant surgery, MRA, 279–280 T1-weighted gradient-echo fat-saturated imaging, 272 Acceleration errors, 63 Acute ischemic stroke, pediatric MRA moyamoya disease, 374–375 transient cerebral vasculopathy, 373–374 vascular dissection and stroke, 375 vasculitis, 375 Acute mesenteric arterial thrombosis, 277 Allison, J.W., 369 Alsaid, H., 205 Aneurysm, MRA techniques, 217–219 Anomalous venous drainage, 265

Anson, B.J., 298 Aortic stenoses, 248 Arepally, A., 395 Arterial disease, MRA techniques aneurysm, 217–219 atherosclerotic disease, 214–215 dissection, 216–217 moyamoya disease, 216 vasculitis, 215–216 Arterial spin labeling (ASL) noncontrast coronary artery imaging, 135–136 noncontrast LE-MRA techniques, 326 renal vascular diseases, 288 Arteriovenous fistulas, 309 Atherosclerotic CMI, 277–279 Atherosclerotic disease, 214–215 Atherosclerotic plaques, 199 Atrial fibrillation clinical applications, MRI device tracking, 396 electrophysiology procedures, 396 lesion imaging, 396 real-time imaging, 396 safety/devices, 396–397 pulmonary vascular imaging, 264–265 Aviv, R.I., 375 Axel, L., 40

B Backes, W.H., 376 Balanced steady-state free precession (bSSFP) lower extremity MRA, 326 renal vascular diseases, 287 thoracic aorta, 245 Balu, N., 113 Barnes, S., 157 Beddy, P., 253 Beilvert, A., 202 Bernstein, M.A., 61 Bharatha, A., 387 Black-blood techniques, 113 abdominal aorta and mesenteric vessels, 272 coronary MRA, 134 thoracic aorta, 241 Blatter, D.D., 177 Block, W., 169 Block, W.F., 172

J.C. Carr and T.J. Carroll (eds.), Magnetic Resonance Angiography: Principles and Applications, DOI 10.1007/978-1-4419-1686-0, © Springer Science+Business Media, LLC 2012

403

404 Blood pool contrast agents, 315, 355 Boada, F.E., 175 Bornert, P., 172 Botnar, R., 203 Briley-Saebo, K.C., 125, 202 Bronchial arteries imaging, 262 Buerger’s disease, 332–333 Buis, D.R., 369 Bulte, J., 395

C CAD visualization anomalous coronary arteries and coronary aneurysms, 133 coronary artery bypass graft assessment, 133 coronary flow imaging, 134 coronary stenosis identification, 132–133 coronary vessel wall imaging, 134 Cardiac motion suppression, 129–130 Carotid and vertebral circulation, clinical applications contrast-enhanced 3D Fourier transform MRA efficacy, 229 limitations, 229 technique, 228 3 T scanners, 229 craniocervical dissections imaging findings, 231 MRA technique, 231 posttraumatic and spontaneous cases, 230–231 extracranial carotid atherosclerosis carotid endarterectomy, 225 carotid stenosis measurement, 226 CASANOVA, 226 contrast-enhanced MRA, 229 stroke, 225 TOF MRA, 228 facial/trigeminal nerve, vascular compression, 235–236 fibromuscular dysplasia, 232–233 head and neck neoplasms, 236–237 intracranial carotid atherosclerosis, 233–234 moyamoya disease and moyamoya syndrome, 234–235 MRA techniques, 226 sickle cell disease, 235 subclavian steal syndrome, 233 Time-of-flight MRA application, 226–227 efficacy, 228 limitations, 227–228 vertebral artery atherosclerosis, 229–230 Carr, J., 239, 351 Cartesian methods, CE-MRA k-space sampling, 76–78 MRI data acquisition, 83 parallel acquisition, 78–79 parallel imaging, angiography, 187–188 receiver coils, 80–81 sampling, 76 signal enhancement effect, 81–82 temporal fidelity, 82 temporal footprint analysis, 79–80 in vivo results, 83–84 Catheter angiography, 257 Celiac artery, anatomy, 273 Cellular therapy, MRI

Index cell tracking, 397–399 device tracking, 399 CE-MRA. See Contrast-enhanced MR angiography (CE-MRA) Cerebral arteriovenous malformations (cAVM), 219–220 Chooi, W.K., 367, 371 Cho, Z.H., 151 Cine SSFP, 240 Coleman, S.S., 298 Collins, J., 351 Collins, J.D., 319 Common Carotid (CC) method, 226 Contrast agents MRA. See also Targeted contrast agents, molecular imaging gadolinium-based agents, 383–384 lower extremity MRA, 323–324 macromolecular agents, 383 non-protein-binding agents, 381–382 paramagnetic agents, 381 protein-binding agents, 383 superparamagnetic agents, 384 upper extremity and hand vessels, MRA, 315 Contrast-enhanced 3D Fourier transform MRA carotid and vertebral circulation efficacy, 229 limitations, 229 technique, 228 3 T scanners, 229 Contrast-enhanced MR angiography (CE-MRA). See also Low-dose contrast-enhanced MR angiography abdominal aorta and mesenteric vessels contrast administration and bolus timing, 271–272 hardware considerations, 270 pulse sequence, 270–271 best guess technique, 70 bolus timing considerations, 69–70 Cartesian methods (See Cartesian methods, CE-MRA) contrast agents, 67 contrast dose and injection rate, 68 coronary arteries (See Coronary artery, CE-MRA) Fourier ( k-space) considerations, 66–67 hand vs. power injection, 69 intracranial arterial and venous disease, 213–214 MR fluoroscopy, 70–71 nephrogenic systemic fibrosis (See Nephrogenic systemic fibrosis (NSF), CE-MRA) patient preparation, 68–69 postproccessing and display, 71–72 pulmonary embolism, 257 pulse sequence, 66 test bolus technique, 70 theory, 65 thoracic aorta, 241–245 triggering, automatic, 70 whole-body MRA, 72 Contrast media, gadolinium-based, 285, 287 Contrast-to-noise (CNR), 341–343 Conventional CE-MRA, 107–108, 242–243 Coronary artery, CE-MRA blood vs. surrounding tissue, contrast, 142 contrast agent administration, 143 contrast agents benefits, 142 contrast-enhanced EPI whole-heart technique, 143 contrast-enhanced whole-heart CMRA, 144 high-field coronary imaging, 144 k-space trajectories, 143 motion compensation, 141

Index parallel imaging benefits, 143–144 spatial resolution, 141–142 SSFP whole-heart CMRA, 143 whole-heart coronary MRA, 3T clinical applications, 145–147 navigator efficiency, 145 patient training, 144–145 survey scanning, 145 vector electrocardiogram, 145 Coronary artery disease, noninvasive imaging clinical applications anomalous coronary arteries, 338 bypass graft assessment, 339–340 vasculitides, 339 coronary vein MRI, 343–345 MRI vs. MDCT, 337–338 technical advances and impediments coronary motion, 340–341 SNR and CNR, 341–343 Coronary vessel wall imaging, 124 Craniocervical dissections imaging findings, 231 MRA technique, 231 posttraumatic and spontaneous cases, 230–231 Cross-linked iron oxides (CLIO), 201 Cystic adventitial disease, 331–332

D Dale, B., 176 Davarpanah, A.H., 351 DeBakey, M.E., 245, 246 Debrey, S.M., 229 Deistung, A., 160 3D fast spin echo (FSE) imaging thoracic aorta, 245 upper extremity and hand vessels, MRA, 300–301 Dick, A.J., 399 Diffusion-weighted imaging (DWI), 214 Double inversion recovery (DIR), 114–116 2D TOF MRA, 227 3D TOF MRA, 227 Dumoulin, C.L., 396 Dural AV fistula, 220–221 Duyn, J.H., 177, 178, 180 Du, Y.P., 160 Dynamic contrast-enhanced MRI, 288–290

E ECST method, 226 Eddy currents, 61–62 Eleftheriou, D., 375 Embolic disease, 306–307 Erdman, W.A., 357 Evangelista, A., 243 Evans, A.J., 356 Extracranial carotid atherosclerosis MRA technique, clinical uses carotid endarterectomy, 225 carotid stenosis measurement, 226 CASANOVA, 226 contrast-enhanced MRA, 229 stroke, 225 TOF MRA, 228

405 F Facial/trigeminal nerve, vascular compression, 235–236 Fasulakis, S., 368 Fat suppression, 119 Fenchel, M., 287 Ferumoxytol, 315 Fibromuscular dysplasia, 232–233, 279 Filtered back projection (FBP), 175 Finn, J.P., 107, 361 Flow compensation, 42 Flow-dependent noncontrast MR angiography in abdomen, 93 2D/3D partial-Fourier FSE flow-in technique, 98–99 flow-out technique, 99 tag-on and tag-off alternate subtraction, 99–100 fresh blood imaging, 91–92 noncontrast MR venography, 96–97 peripheral MRA FBI-Navi, 94 MIP procession, 95 readout (RO)/frequency direction, 93 STIR pulse, 96 systolic and diastolic triggering, 93 phase contrast, 91 principles, 92–93 SSFP Imaging b noncontrast MRA, 101 subtractive methods, 101–102 quiescent interval single-shot MRA, 102–104 time-of-flight (TOF) MOTSA, 89 SORS pulse, 90 TONE, 89 venography, 96–98 Flow enhancement effect (FRE), 41, 42 Flow quantification, 56. See also Phase-contrast MRI Flow-sensitive 4D MRI, 58–60 Fourier ( k-space) considerations, 66 Frahm, J., 172 Francois, C.J., 362 Fraser, D.G., 356 Frias, J., 202 First-order gradient moment nulling, 42 Fushimi, Y., 235

G Gadobenate dimeglumine (Gd-BOPTA), 315 Gadobenate dimeglumine (multihance), 381, 383 Gadobutrol (gadovist), 381, 383 Gadofosveset trisodium (ablavar), 381, 383 Gadolinium-based contrast agents (GBCA), 383–384 nephrogenic systemic fibrosis, CE-MRA chelate, solution stability of, 389 chemical properties and kinetic stability, 390 dosing limit, 391 history, 389 inherent toxicity, 389 nomenclature, 390 putative mechanisms, 389 thermodynamic/chemical stability, 390 Gadolinium chelates, 201 Gadolinium contrast agents, 355 Gadolinium-enhanced MRA lower extremity MRA

406 Gadolinium-enhanced MRA (continued) adaptations, 322–323 dosing, 323 high-resolution, 320 k-space, 321 parallel imaging technologies, 321 spatial resolution, 320 three-dimensional acquisitions, 320 time-resolved, 321 Gadolinium-enhanced MRV, 303, 304 Galen vein aneurysmal dilatation, 371 aneurysmal malformation, 370–371 malformation, 369 MRA role, malformation, 371–373 Garovic, V.D., 286 Generalized autocalibrating partially parallel acquisitions (GRAPPA), 189–191 Glover, G.H., 172 Griswold, M., 185 Gullberg, G.T., 39

H Haacke, E.M., 157 Haage, P., 259 Hadizadeh, D.R., 213 Hagspiel, K.D., 269 Haider, C.R., 75, 80 Hand vessels, MRA. See Upper extremity and hand vessels, MRA Hand vs. power injection, 69 Hays, A., 129 Head and neck neoplasms, 236–237 Herborn, C.U., 381 Heverhagen, J.T., 271 Hinshaw, W.S., 39 Hodnett, P., 351 Hoffmann, B.A., 396 Hong, S.N., 337 Hu, J., 372 Hunt and Hess classification, intracranial aneurysm, 218 Hu, P., 337 Hurst, D.R., 259 Hypothenar hammer syndrome, 308–309

I Ichinose, N., 89 Inferior mesenteric artery, anatomy, 273–274 Intracerebral hemorrhage, 221–222 Intracerebral vasculitis, 215 Intracranial arterial and venous disease arterial disease, stenoocclusive disease/stroke aneurysm, 217–219 atherosclerotic disease, 214–215 dissection, 216–217 moyamoya disease, 216 vasculitis, 215–216 cerebral arteriovenous malformations, 219–220 dural AV fistula, 220–221 intracerebral hemorrhage, differential diagnosis, 221–222 MRA techniques clinical recommendations, 214 contrast-enhanced, 213–214 phase-contrast, 213 time-of-flight, 213

Index venous disease thrombosis, 219 Intracranial carotid atherosclerosis, 233–234 Intramural hematoma (IMH), 246 Intravascular ultrasound (IVUS), 199–200 Invasive angiography, cardiac catheterization, 199 Iron-based blood pool agents, 355 Ito, K. 100

J Jaarsveld, B.C., 291 Jagadeesan, B.D., 365 Jaspers, K., 376 Johnson, C.P., 75, 80, 87 Joseph, P.M., 175 Jung, B., 51

K Kanazawa, H., 89, 98, 99 Kassai, Y., 89 Kelly, K., 205 Kim, C.Y., 361 Kim, S.-E., 39 King, K.F., 172 Klippel-Trénaunay syndrome, 358 Kluge, A., 260 Kolandaivelu, A., 395 Koopmans, P.J., 154 Kramer, U., 68, 272 Kreitner, K.F., 261 Krings, T., 371 Krum, H., 292 Kucharczyk, W., 387

L Ladd, M.E., 149 Laissy, J.P., 359 Lancelot, E., 205 Lardo, A.C., 396 Larmor equation, 51 Laub, G., 107 Lederman, R.J., 399 Lee, V.S., 297 Leiner, T., 283 Leriche’s syndrome, 330 Li, D., 141, 143 Li, F., 121 Lim, R.P., 297 Linear Eddy current correction, 176–177 Li, R., 113 Liu, J., 179 Li, W., 315, 356 Lohan, D.G., 108 Low-dose contrast-enhanced MR angiography abdomen–pelvis, 110 chest, 110 contrast agent type, 108–109 contrast timing and injection protocol, 109 conventional CE-MRA, 107–108 fast imaging tools, 108 gadolinium association, 107 head and neck, 109–110 lower extremity runoff, 110–111

Index Lower extremity MRA (LE-MRA) arterial spin labeling (ASL), 326 balanced steady-state free precession (bSSFP), 326 clinical applications, 329–330 contrast agents, 323–324 cystic adventitial disease, 331–332 ECG-gated 3D partial Fourier FSE, 326–327 ECG-gated flow-sensitive dephasing (FSD) bSSFP, 327 gadolinium-enhanced MRA adaptations, 322–323 dosing, 323 high-resolution, 320 k-space, 321 parallel imaging technologies, 321 spatial resolution, 320 three-dimensional acquisitions, 320 time-resolved, 321 imaging processing, 328–329 noncontrast techniques, 324 phase-contrast angiography, 325–326 plaque imaging, 328 popliteal entrapment, 330–331, 335 protocols, 334 quiescent interval single-shot (QISS) MRA, 327–328 renal insufficiency, 335 thromboangiitis obliterans/Buerger’s disease, 332–333 time-of-flight angiography, 324–325 3 T scanners, 328 Loy, D.N., 365 Lumenal stenosis, 199 Lustig, J., 181

M Macromolecular contrast agents, 383 Magnetic resonance angiography (MRA) bSSFP, 34–36 CE-MRA, 21 contrast materials, 21–22 data acquisition order, 24–25 imaging sequence, 22–24, 26–27 mask subtraction, 25 multistation exams, 26 phase-contrast, 29–32 processing methods, 32–33 scan synchronization, 24 three-dimensional TOF MRA, 28–29 time-of-flight MRA, 26 two-dimensional TOF MRA, 27–28 velocity encoding and velocity aliasing, 34 Magnetic resonance imaging (MRI) atrial fibrillation ablations device tracking, 396 electrophysiology procedures, 396 lesion imaging, 396 real-time imaging, 396 safety/devices, 396–397 Bloch equation, 10 bulk magnetization, 5–6 cellular therapy cell tracking, 397–399 device tracking, 399 contrast mechanisms T1 relaxation time, 8–9 T2 relaxation time, 6–8 T1 vs. T2 relaxation rates, 9–10

407 description, 3 frequency-encoding, 14–16 imaging parameters, 10–11 mechanism, 5 MR data ( k-space), 17–20 MR image formation, 11–13 parallel imaging, 21 phase-encoding, 16–17 signal-to-noise ratio (SNR), 20–21 slice selection, 13–14 types, 3, 4 Manning, W.J., 337 Markl, M., 51 Maxwell terms, 60–61 May-Thurner Syndrome, 359 McAteer, M., 206 Meade, T. J., 199 Meaney, J.F.M., 253 Meckel, S., 373 Median arcuate ligament syndrome, 279 Mesenteric aneurysms, 279 Mesenteric arterial system, 273 Mesenteric arterial system, MRA acute mesenteric arterial thrombosis, 277 acute mesenteric ischemia, 275–276 acute SMA embolism, 276 aortic dissection, 277 atherosclerotic CMI, 277–279 chronic mesenteric ischemia, 277 mesenteric venous thrombosis, 277 nonatherosclerotic vascular pathogies, 279 nonocclusive mesenteric ischemia, 277 Mesenteric ischemia, 275–277 Mesenteric venous thrombosis, 277 Michaely, H., 283 Micron-sized particles of iron oxide (MPIO), 200, 201 Mittal, S., 164 Miyazaki, M., 89, 92, 316 Moghari, M.H., 337 Mohajer, 321 Moody, A.R., 261 Morawski, A., 203 Mostardi, P.M., 75, 78, 82 Motion artifact reduction, 119–120 Moyamoya disease MRA technique, clinical uses, 234–235 MRA techniques, 216 MR angiography and high field strength benefits, 151–153 contrast-enhanced MRA, 7 T, 153–154 high-field MR, 154–155 7 T non-neuro MRA techniques, 155 MR fluoroscopy, 70–71 MRV techniques contrast agents, 355 contrast-enhanced MRV, 353–354 3D fast spin echo, 352–353 MR-directed thrombus imaging, 353 noncontrast-enhanced MRV, 351–352 STARFIRE, 352 steady-state free precession, 352 susceptibility-weighted imaging, 353 time of flight, 352 time-resolved MRV, 354 Mukherjee, S., 225 Mulder, W.J., 202

408 Mull, M., 376 Multicontrast techniques, 116–117 Multiple overlapping thin 3D slab acquisition (MOTSA) technique, 39, 89, 227

N Nael, K., 107 NASCET method carotid stenosis measurement, 226 Nayler, G.L., 51 Nazarian, S., 396 Neff, W., 375 Nephrogenic systemic fibrosis (NSF), CE-MRA clinical findings, 387 diagnosis, 388 gadolinium-based contrast agents chelate, solution stability of, 389 chemical properties and kinetic stability, 390 dosing limit, 391 history, 389 inherent toxicity, 389 nomenclature, 390 putative mechanisms, 389 thermodynamic/chemical stability, 390 prevention measurement, 392–393 renal dysfunction, 391 risk factors, 391–392 treatment, 388 Nezafat, R., 337 Nishimura, D.G., 39 Nonatherosclerotic vascular pathogies, 279 Non-cartesian MR angiography asymmetric FOVs, 171 field of view, 170 flow sensitivity, 171 gradient spoiling, 170 off resonance, 170–171 parallel imaging, angiography, 191–192 projection imaging 2D, 172–173 3D radial, 173–175 undersampled, 173 reconstruction filtered back projection (FBP), 175 gridding, 175–176 image degradation, k-space sampling errors, 176 linear Eddy current correction, 176–177 off-axis imaging, 177–178 steady-state free precession (SSFP), 178–179 sampling density, 171 sampling region, 170 spiral trajectory design, 171–172 SSFP Imaging, 101 b temporal processing, 180–181 time-resolved MRA quantitative velocity imaging, 180 temporal processing, 179–180 trajectory design, 170 Noncontrast coronary artery imaging arterial spin labeling, 135–136 CAD visualization anomalous coronary arteries and coronary aneurysms, 133 coronary artery bypass graft assessment, 133 coronary flow imaging, 134 coronary stenosis identification, 132–133

Index coronary vessel wall imaging, 134 cardiac motion suppression, 129–130 high field imaging, 136–137 parallel imaging, 136 radial and spiral imaging techniques, 135 respiratory motion compensation breath-hold technique, 130 contrast enhancement, 131 free-breathing technique, 130–131 whole-heart technique, 131–132 steady-state free-precession, 136 Noncontrast-enhanced MRA technique, 286–287 Non-contrast MRA (NC-MRA). See also Flow-dependent noncontrast MR angiography abdominal aorta and mesenteric vessels, 270 pediatric MRA, 366 thoracic aorta, 244–245 Nonocclusive mesenteric ischemia, 277 Non-protein-binding contrast agents, 381–382 Norton, P.T., 269 Nutcracker syndrome, 358–359

O Ohno, Y., 257, 259 O’ Keefe, 258 Ono, A., 97, 358 Oudkerk, M., 257

P Paget-Schroetter syndrome, 314 Pancreatic transplantation, 280 Pandey, T., 358 Papadias, A., 368 Parallel imaging, angiography Cartesian method, 187–188 conjugate gradient SENSE, 192–193 3D, 191 dynamic methods, 195 GRAPPA, 189–191 non-Cartesian, 191–192 PILS, 188 radial GRAPPA, 193–195 remarks, 187 sensitivity encoding, 188–189 SNR losses, 187 Parallel imaging with localized sensitivities (PILS), 188 Paramagnetic contrast agents, 381 Parker, D.L., 39 Parkes Weber syndrome, 358 PC angiography, 57–58 Pediatric MRA acute ischemic stroke moyamoya disease, 374–375 transient cerebral vasculopathy, 373–374 vascular dissection and stroke, 375 vasculitis, 375 arteriovenous malformations, 367–369 Galen vein aneurysmal dilatation, 371 aneurysmal malformation, 370–371 malformation, 369 MRA role, malformation, 371–373 noncontrast MRA, 366 phase contrast MRA, 366–367

Index principles, 365 spinal vascular malformations, 375–377 time-of-flight MRA, 366 time-resolved contrast-enhanced MRA, 367 Pelvic congestion syndrome (PCS), 358 Penetrating atherosclerotic ulcer (PAU), 247 Perfusion-weighted imaging (PWI) atherosclerotic disease, 214 Peters, D.C., 173 Phase contrast MRA (PC-MRA) abdominal aorta and mesenteric vessels, 269 intracranial arterial and venous disease, 213 lower extremity MRA, 325–326 pediatric MRA, 366–367 renal vascular diseases, 288 thoracic aorta, 244 upper extremity and hand vessels, MRA, 301 Phase-contrast MRI (PC-MRI) 2D CINE, applications, 56–57 flow-sensitive 4D MRI, 58–60 implementation and clinical protocols, 55–56 PC angiography, 57–58 principle bipolar gradient, 53 image reconstruction, 52 Larmor equation, 51 signal-to-noise ratio (SNR), 53 temporal footprint (TE), 52 velocity nois, 54 pulse sequences, 54–55 sources of errors acceleration errors, 63 Eddy currents, 61–62 gradient field nonlinearities, 61 Maxwell terms, 60–61 PIOPED III study, 258–259 Pipe, J.G., 175, 176 Plaque imaging, 121–122 Prince, M.R., 65, 351 Protein-binding contrast agents, 383 Pruessmann, K.P., 193 Pseudostenosis, 305 Pulmonary arteriovenous malformation (AVM), 263 Pulmonary artery sling, 263 Pulmonary embolism (PE) animal studies, MRA, 259 applied pulmonary artery anatomy, 253 blood-pool imaging, 260 catheter angiography, 254–255 CE-MRA and catheter angiography, 257 clinical considerations, 253–254 cross-sectional imaging, 255–256 direct thrombus imaging, 260–261 MRA, 256–257 vs. CTA, 257–258, 261 non-contrast MRA, 260 perfusion imaging, 259 PIOPED III study, 258–259 time-resolved MRA, 259 Pulmonary hypertension, 261 Pulmonary sequestration, 263 Pulmonary vascular imaging anomalous venous drainage, 265 atrial fibrillation, 264–265 bronchial arteries, 262 malignancies imaging, 262–263

409 pulmonary arteriovenous malformation (AVM), 263 pulmonary artery sling, 263 pulmonary embolism animal studies, MRA, 259 applied pulmonary artery anatomy, 253 blood-pool imaging, 260 catheter angiography, 254–255 CE-MRA and catheter angiography, 257 clinical considerations, 253–254 cross-sectional imaging, 255–256 direct thrombus imaging, 260–261 MRA, 256–257 MRA vs. CTA, 257–258, 261 non-contrast MRA, 260 perfusion imaging, 259 PIOPED III study, 258–259 time-resolved MRA, 259 pulmonary hypertension, 261 pulmonary sequestration, 263 pulmonary vein, 263–264 Pulmonary vein, 263–264

Q Qanadli, S.D., 254 Quick, H.H., 149 Quiescent interval single-shot MRA, 102–104

R Radial and spiral imaging techniques, 135 Rauscher, A, 154 Raynaud’s phenomenon, 309–310 Renal vascular diseases anatomical considerations, 283–284 ASL techniques, 288 ASTRAL trial, 291 dynamic contrast-enhanced MRI, 288–290 functional renal imaging, 290–291 gadolinium-based contrast media, 287 MR imaging, 283 noncontrast-enhanced MRA technique, 286–287 phase-contrast MRA, 288 renal artery stenosis functional significance assessment, 287–288 MR angiography, 284–286 Renal vein thrombosis, 359–360 Respiratory motion compensation breath-hold technique, 130 contrast enhancement, 131 free-breathing technique, 130–131 whole-heart technique, 131–132 Riederer, S.J., 75, 81 Ritt, M., 290 Rollins, N., 373 Ruehm, S., 201 Ruehm, S.G., 359

S Saam, T., 124 Saeed, M., 399 Saleh, R., 107 Scanlon, T., 319, 351 Seiberlich, N., 185 Seo, J.B., 259

410 Sheehan, J.J., 316 Shimizu, K., 173 Sickle cell disease, 235 Signal targeting alternative radiofrequency and flow-independent relaxation enhancement (STARFIRE), 352 Signal-to-noise (SNR), 341–343 Singh, N., 122 Single-shot two-dimensional SSFP, 240 Slice-selective off-resonance sinc (SORS) pulse, 90 SMA embolism, 276 SNR optimization, 118 Sodickson, D.K., 185 Soulez, G., 285 SPGRE pulse sequence, 41 Spuentrup, E., 287, 356 SSFP Imaging b noncontrast MRA, 101 subtractive methods, 101–102 Stafford, R.B., 328 Steady-state free precession abdominal aorta and mesenteric vessels, 269–270 thoracic aorta, 239–240 Steffens, J., 325 Steinman, D.A., 125 Steno-occlusive disease, 305–306 Stenoocclusive disease/stroke. See Arterial disease, MRA techniques Stuber, M., 129 Sturge-Weber (SWS) syndrome, 371–372 Subclavian steal syndrome, 233 Subclavian steal-syndrome, 215 Sugiura, S., 89 Superior mesenteric artery, anatomy, 273 Superparamagnetic contrast agents, 384 Superparamagnetic iron oxide nanoparticles (SPIO), 201 Suryan, G., 39 Susceptibility weighted imaging (SWI) blood properties, different field strengths, 162 caffeine and acetazolmide role, 163–164 cerebral microbleeds, 164–165 concepts, 157–159 flow compensated gradient echo sequence, 159–160 mapping and oxygen saturation measurement, 159 single echo approach, 160–162 T1 and T2* role, 162–163 Swoboda, N.A., 365 Symons, S.P., 387 Systemic-bladder drainage (SBD), 280

T Takayasu’s arteritis, 249–250 Targeted contrast agents, molecular imaging adhesion molecules, 205 cytokines and integrins, 205 integrins and angiogenesis, 207–208 lipid-rich plaque regions and lipoproteins Gd(III) chelates, 202 lipid-targeting method, 202 micelles, 202, 203 oxidized low-density lipoprotein (OxLDL), 202 macrophages, 201–202 matrix metalloproteinases, 205 MRI contrast agent classes iron oxide nanoparticles, 200 lipid-based nanostructure, 201 myeloperoxidase, 208–209

Index P-selectin and vascular cell adhesion molecule, 205–207 thrombus, plaque-associated activated platelets, 204 acute and subacute, 203 EP-1873, 203 fibrin, 203–204 human carotid plaques, 204 LIBS-MPIO, 204 Test bolus technique, 70 Thoracic aorta aneurysms, 247–248 aortic dissection, 245–246 computed tomography, 239 congenital abnormalities, 248–249 contrast-enhanced MRA balanced steady-state free precession, 245 conventional CE-MRA, 242–243 3D FSE, 245 non-contrast MRA, 244–245 phase contrast MRA, 244 time-resolved MRA, 241–242 giant cell arteritis, 250 inflammatory conditions, 250 intramural hematoma, 246 MRI techniques black-blood techniques, 241 cine SSFP, 240 contrast-enhanced MRA, 241–245 single-shot two-dimensional SSFP, 240 steady-state free precession, 239–240 three-dimensional SSFP, 240–241 T1-weighted gradient echo fat-saturated imaging, 241 penetrating atherosclerotic ulcer, 247 pseudoaneurysms, 247 stenoses, 248 Takayasu’s arteritis, 249–250 transesophageal echocardiography, 239 vasculitis, 249 Thoracic aortic aneurysm (TAA), 247–248 Thoracic outlet syndrome, 311–313 Three-dimensional SSFP, 240–241 Thromboangiitis obliterans/Buerger’s disease, 332–333 Tilted optimized non-saturating excitation (TONE) pulses, 46, 89 Time-of-flight (TOF) angiography 2D TOF, 44–45 3D TOF, 45–47 high magnetic fields, 49 k-space sampling strategies, 47 lower extremity MRA, 324–325 magnetization transfer, 47–48 MR signal flow sensitivity, 39 goal, 40 maximum intensity projection (MIP), 39 MOTSA acquisitions, 39 spoiled gradient echo (SPGRE) sequence, 39 multiple overlapping thin 3D slab acquisition, 48–49 phase dispersion and flow compensation 3D SPGR pulse sequence, 42, 43 first-order gradient moment nulling, 42 phase dispersion, 42 pulse sequence, 43 second-order gradient motion rephasing, 44 signal intensity, 42 velocity-dependent phase dispersion, 43

Index quantification arterial velocity, 42 Bloch equations, 40 flow enhancement effect (FRE), 41, 42 maximum transverse magnetization, 41 RF pulses, 40 speed definition, 40 SPGRE pulse sequence, 41 spoiled gradient sequence, 41 velocity, 41 Time-of-flight MR-angiography (TOF-MRA) abdominal aorta and mesenteric vessels, 269 carotid and vertebral circulation, clinical applications application, 226–227 efficacy, 228 limitations, 227–228 intracranial arterial and venous disease, 213 pediatric MRA, 366 upper extremity and hand vessels, MRA, 301 Time-resolved contrast-enhanced MRA, 367 Time-resolved 3D CE-MRA. See Cartesian methods, CE-MRA Time-resolved MRA abdominal aorta and mesenteric vessels, 272 pulmonary embolism, 259 thoracic aorta, 241–242 Transesophageal echocardiography, 239 Trattnig, S., 154 T1-weighted gradient-echo fat-saturated imaging abdominal aorta and mesenteric vessels, 272 thoracic aorta, 241

U Ultrasmall iron oxide nanoparticle (USPIO), 200 Underhill, H, 123 Upper extremity and hand vessels, MRA bright blood localizer, 299 clinical indications arteriovenous fistulas, 309 embolic disease, 306–307 hypothenar hammer syndrome, 308–309 Paget-Schroetter syndrome, 314 Raynaud’s phenomenon, 309–310 steno-occlusive disease, 305–306 thoracic outlet syndrome, 311–313 trauma, 307–308 vascular malformations, 311 vasculitis, 310–311 venous thrombosis, 313–314 contrast agents, 315 2D fast spin echo imaging, 300–301 extraluminal pathology, 305 flow-related artifacts, 305 gadolinium-enhanced MRA, 301 acquisition timing, 301–302 contrast dosage, 301 sequence parameters, 302 time-resolved, 302 gadolinium-enhanced MRV, 303, 304 high-field imaging, 315 image acceleration, 315 inaccurate timing, 303, 305 motion artifact, 305 noncontrast-enhanced techniques, 315–316 phase-contrast and time-of-flight MRA, 301 pseudostenosis, 305

411 scanning coil selection, 299 patient positioning, 299 vascular anatomy arterial anatomy, 297–298 venous anatomy, 298–299 vascular mimics, 305 Urata, J., 95 Urbach, H., 213

V Vasbinder, G.B., 285 Vasculitis, 215–216, 249, 310–311 Vector electrocardiogram, 145 Venkataraman, S., 275 Venous disease MRA techniques, 219 Venous imaging clinical applications DVT assessment, 356–358 Klippel-Trénaunay syndrome, 358 May-Thurner Syndrome, 359 nutcracker syndrome, 358–359 Parkes Weber syndrome, 358 pelvic congestion syndrome (PCS), 358 portal vein occlusion, 360 pulmonary vein evaluation, 361–362 renal vein thrombosis, 359–360 upper extremity and central vein evaluation, 360–361 MRV techniques contrast agents, 355 contrast-enhanced MRV, 353–354 3D fast spin echo, 353 MR-directed thrombus imaging, 353 noncontrast-enhanced MRV, 351–352 STARFIRE, 352 steady-state free precession, 352 susceptibility-weighted imaging, 353 time of flight, 352 time-resolved MRV, 354 Venous thrombosis, 313–314 Vertebral artery atherosclerosis, 229–230 Vessel wall imaging techniques applications natural history studies, 122–123 vessel morphology, 120–121 vulnerable plaque imaging, 121–122 black-blood techniques, 113 blood suppression, spin echo, 114 clinical trials, 124 contrast weightings contrast enhancement techniques, 117–118 multicontrast techniques, 116–117 coronary vessel wall imaging, 124 diffusion preparation, 116 double inversion recovery (DIR), 114–116 flow and relaxation properties, blood, 113–114 hemodynamic study, 125 higher resolution, 125 molecular imaging, 124–125 practical considerations fat suppression, 119 field strength, 118–119 image processing, 120 localization, 120

412 Vessel wall imaging techniques (continued) motion artifact reduction, 119–120 SNR optimization, 118 saturation band, 114 Von zur Muhlen, C., 204 Vulnerable plaque, 199

W Wang, J., 315 Ward, E., 239 Waters, E.A., 199 Weiss, C.R., 395 Weiss, R.G., 129 Whole-body MRA, 72 Whole-heart coronary MRA, 3T clinical applications, 145–147 navigator efficiency, 145 patient training, 144–145 SSFP, 143 survey scanning, 145 vector electrocardiogram, 145

Index Widjaja, W., 373 Wieben, O., 169 Willinek, W. A., 213 Willoteaux, S., 285 Wilson, G.J., 287 Windkessel effect, 56 Wintermark, M., 225 Winter, P., 207 Wittram, C., 254 Wong, H.I., 359

Y Yamada, I., 235 Yang, Q., 141 Yoon, H.K., 374 Yuan, C., 113, 121

Z Zhang, H., 65 Zhang, W., 65 Ziyeh, S., 373