Acute ischemic coronary syndromes often result from the

Transcription

1 Noninvasive In Vivo Human Coronary Artery Lumen and Wall Imaging Using Black-Blood Magnetic Resonance Imaging Zahi A. Fayad, PhD; Valentin Fuster, MD, PhD; John T. Fallon, MD, PhD; Timothy Jayasundera, MD; Stephen G. Worthley, MD; Gerard Helft, MD; J. Gilberto Aguinaldo, MD; Juan J. Badimon, PhD; Samin K. Sharma, MD Background High-resolution MRI has the potential to noninvasively image the human coronary artery wall and define the degree and nature of coronary artery disease. Coronary artery imaging by MR has been limited by artifacts related to blood flow and motion and by low spatial resolution. Methods and Results We used a noninvasive black-blood (BB) MRI (BB-MR) method, free of motion and blood-flow artifacts, for high-resolution (down to 0.46 mm in-plane resolution and 3-mm slice thickness) imaging of the coronary artery lumen and wall. In vivo BB-MR of both normal and atherosclerotic human coronary arteries was performed in 13 subjects: 8 normal subjects and 5 patients with coronary artery disease. The average coronary wall thickness for each cross-sectional image was mm (range, 0.55 to 1.0 mm) in the normal subjects. MR images of coronary arteries in patients with 40% stenosis as assessed by x-ray angiography showed localized wall thickness of mm (range, 3.30 to 5.73 mm). The difference in maximum wall thickness between the normal subjects and patients was statistically significant (P ). Conclusions In vivo high-spatial-resolution BB-MR provides a unique new method to noninvasively image and assess the morphological features of human coronary arteries. This may allow the identification of atherosclerotic disease before it is symptomatic. Further studies are necessary to identify the different plaque components and to assess lesions in asymptomatic patients and their outcomes. (Circulation. 2000;102: ) Key Words: atherosclerosis magnetic resonance imaging coronary disease Acute ischemic coronary syndromes often result from the rupture of a mildly to moderately stenotic coronary artery plaque, leading to thrombus formation. 1 3 Currently available imaging techniques for the diagnosis of coronary artery disease are limited. For example, coronary angiography demonstrates only the degree of luminal narrowing and fails to visualize the arterial wall. Moreover, arteries accommodate plaque growth through outward displacement of the vessel wall, thereby preserving lumen cross-sectional area. 4 Other imaging techniques, such as intravascular ultrasound, 5 fast CT, 6 and angioscopy, 7 have all advanced our understanding of atherosclerosis, but these techniques are invasive and yield limited information about plaque composition. Recent in vivo studies of atherosclerotic plaques in animal models, 8,9 carotid arteries, 10,11 and aorta 12 demonstrate that high-resolution MRI can noninvasively image the artery wall and assess plaque composition. Preliminary studies in a porcine model of atherosclerosis showed that the major difficulties of MR coronary wall imaging are due to the combination of cardiac and respiratory motion artifacts, the nonlinear course of the coronary arteries, and their relatively small size and location. 13 Thus, an effective in vivo MRI technique for coronary artery imaging must overcome artifacts related to blood flow and cardiac, respiratory, and vessel wall motion to achieve high-resolution and high-contrast imaging. Current white-blood non contrast-enhanced MR coronary angiography (gradient-echo, 14 echo planar, 15 spiral, 16 etc) provide no information about the coronary wall structure or atherosclerotic plaque characteristics. In this context, the concept of black-blood MRI (BB-MR) is promising, because the signal from static tissue is maximized and the transverse magnetization of flowing blood is made intentionally incoherent, leading to blood signal void. 17,18 Therefore, by combining BB-MR with high-spatialresolution and fast-data-acquisition imaging, both lumen and wall imaging of the coronary arteries should be possible. A number of different methods are available for BB-MR However, none of these methods have been used for coronary lumen and wall imaging. In this in vivo study of normal and atherosclerotic human coronary arteries, we use long-echo-train-length (ETL) fastspin-echo (FSE) imaging with velocity-selective inversion preparatory pulses 22,23 to nullify the signal from flowing Received November 15, 1999; revision received February 22, 2000; accepted February 29, From the Zena and Michael A. Wiener Cardiovascular Institute (Z.A.F., V.F., J.T.F., T.J., S.G.W., G.H., J.G.A., J.J.B., S.S.) and the Departments of Radiology (Z.A.F.), Medicine (V.F., J.T.F., J.J.B., S.S.), and Pathology (J.T.F.), Mount Sinai School of Medicine, New York, NY. Correspondence to Zahi A. Fayad, PhD, Mount Sinai School of Medicine, Box 1234, New York, NY zahi.fayad@mssm.edu 2000 American Heart Association, Inc. Circulation is available at 506

2 Fayad et al In Vivo MRI of Human Coronary Atherosclerosis 507 blood. A cardiac phased-array surface coil for high-resolution coronary imaging (460- to 750- m in-plane spatial resolution) is also used. 24 The results of this study clearly demonstrate that normal and atherosclerotic human coronary wall imaging can be performed with high-resolution BB-MR methods. Methods Subjects We studied 13 subjects: 8 healthy subjects (mean age, 30 years; range, 25 to 37 years; 4 men) without a history of cardiovascular disease and 5 consecutive coronary artery disease patients (mean age, 68 years; range, 50 to 78 years; 3 men) with 40% stenosis as documented by x-ray angiography. The MRI studies were performed within 24 hours of the coronary contrast x-ray angiogram. Written informed consent was obtained from all subjects, and the institutional review board approved the protocol. MR Imaging MRI was performed on a 1.5-T whole-body MRI system (General Electric Medical Systems, Signa) equipped with high-performance gradient (40-mT/m amplitude, 150-mT m 1 ms 1 slew rate) and a multichannel receiver with a maximal bandwidth of 250 khz. A 4-element (2 anterior and 2 posterior) specially designed cardiac phased-array receiver surface coil was used for signal reception. 24 ECG electrodes were attached to trigger data acquisition and minimize cardiac and vessel motion. All subjects were instructed to withhold breathing at end expiration to minimize respiratory motion. Breath-holding was confirmed by a bellows respiratory monitor. Imaging Sequence Fast scout gradient-echo images were acquired initially in the coronal, sagittal, and transverse planes for localization. All imaging was subsequently performed with a BB-MR sequence. The major features of the BB-MR sequence are (1) velocity-selective inversion preparatory pulses; (2) short RF excitation pulses; (3) optimized fat saturation; and (4) long-etl FSE imaging. Velocity-Selective Inversion Preparatory Pulses Velocity-selective inversion preparatory pulses were used to suppress the signal from flowing blood, thereby avoiding possible flow artifacts and providing a strong contrast between the dark flowing blood and bright wall signal of the coronary arteries. The flow suppression consisted of 2 inversion pulses: 1 nonselective 180 inversion recovery (IR) pulse followed immediately by a sliceselective 180 inversion pulse. 22 The first IR pulse inverts the magnetization of the entire body, including all of the blood. Next, the second IR pulse reinverts the imaging slice but leaves the blood outside the slice inverted. The section thickness of the selective inversion pulse was set to 3 times the slice thickness to accommodate possible misregistration of tissue between the preparatory pulses and data acquisition. The nonselective pulse consisted of a rectangular hard pulse 1024 s long. The selective pulse was a hyperbolicsecant pulse 8640 s long. 25 This provided good B1 insensitivity and inversion profile. The velocity-selective inversion pulses were placed at end diastole (after the detection of the ECG trigger), and the data acquisition occurred during diastole. This process maximized the blood flow suppression due to outflow and also minimized artifacts due to vessel motion. Image acquisition started after a predetermined inversion time (TI). The delay time or TI for the velocity-selective inversion preparatory pulses was determined close to the null point of the blood signal (see equation). TI is based on the T1 relaxation value of the blood and the TR interval: TI T1 1 e TR T1. 2 With TR 2 RR 1000 ms (heart rate 60 bpm) and T ms, from the equation, TI is 625 ms. Short Optimized RF Pulses The data acquisition is performed with an FSE sequence. As usual for the FSE sequence, the time between the 90 excitation pulse and the first refocusing pulse is half the time between the neighboring refocusing pulses (the echo spacing; ESP). The strong and fast gradients made possible very compact echo trains. To further shorten the ESP, short radiofrequency (RF) pulses optimized by use of the Shinnar-LeRoux algorithm were used. 26 The RF excitation pulses were 1.2 ms long. The refocusing pulse had a flip angle of 155. With a data acquisition sampling of 125 khz and 256 frequency points, an ESP as short as 3.9 ms was achievable. These pulses provided reduced power deposition and reduced echo amplitude unstabilities Optimized Fat Suppression In acquiring images of the coronary artery wall, the velocityselective inversion pulses were immediately followed by a chemical shift selective (CHESS) pulse. This pulse eliminated the epicardial fat signal and thus enhanced the definition of the outer boundary of the arteries. To take account of the multicomponent nature of the fat signal, the CHESS pulse was optimized according to Kuroda et al 30 and resulted in improved fat suppression and coronary wall visualization. Long-ETL FSE Imaging The preparatory pulses (velocity-selective inversion pulses and CHESS pulse) were followed by an ECG-gated, long-etl FSE imaging sequence. The short ESP allowed the use of long-etl data acquisition without the disadvantage of T2 relaxation blurring. 31 From the initial scout images of the coronary arteries, 5 contiguous transverse (cross-sectional) images of the lumen and wall of the proximal segments of the right (RCA) and left anterior descending (LAD) coronary arteries were acquired in 13 subjects. Imaging was performed during short periods of suspended respiration of 12 to 18 heartbeats per slice. One way to restrict the number of phaseencoding steps is to reduce the field of view (FOV) in the phaseencoding direction. However, for a small FOV, this may result in back-folding artifacts. Therefore, when this was the case, we disabled the 2 posterior coil elements of our 4-coil-elements anterior and posterior phased-array coil by a user-controlled variable just before imaging. The imaging parameters were TR 2 RR intervals, TE 40 ms, asymmetric (3/4) FOV in the phase encoding direction (in some of the images), 18- to 29-cm FOV, 3- to 5-mm slice thickness, no interslice gap, or acquisition matrix, number of signals averaged (NSA) 1, 32 ETL, 125-kHz data sampling. The in-plane resolution was 0.46 to 0.75 mm. Image Analysis The MR images were transferred to a Macintosh computer for analysis. The inner (ie, lumen) and outer (eg, adventitial-medial) boundaries of the vessels were traced semiautomatically with ImagePro Plus (Media Cybernetics). The semiautomatic tracing tool works by following an edge (ie, boundary) of significant contrast. The maximal wall thickness was determined from each crosssectional image. The data were then analyzed with a 2-tailed unpaired Student s t test. A value of P 0.05 was considered to be statistically significant. Values are mean SEM. Results All the images demonstrated excellent flow suppression and high contrast and signal-to-noise in the coronary arteries and clearly delineated the coronary wall. Cross-sectional images of normal coronary artery wall showed a circular lumen surrounded by a uniform thin wall (Figure 1). A transverse lumen image obtained without fat saturation (Figure 1A) and wall image obtained with fat saturation (Figure 1B) of the

3 508 Circulation August 1, 2000 Figure 1. In vivo cross-sectional BB-MR images of lumen (A) and wall (B) of proximal LAD from normal subject (see arrow). Lumen image is obtained without fat saturation; wall image is obtained with fat saturation to better delineate coronary artery wall. Blood flow in coronary artery lumen is suppressed with velocity-selective inversion preparatory pulses. Images show normal circular lumen surrounded by uniform thin coronary wall. Average wall thickness of LAD measures 0.8 mm. B, Inset, Magnified view of LAD. Some imaging parameters were TR 2 RR intervals, TE 40 ms, cm FOV, 3-mm slice thickness, acquisition matrix, NSA 1, 32 echo train length, 125- khz data sampling. LV indicates left ventricle; RV, right ventricle; RVOT, right ventricular outflow tract; and CW, chest wall. proximal LAD from a normal subject are shown. In this subject, the maximum wall thickness of the LAD measures 0.8 mm. Figure 2 shows the ectatic atherosclerotic coronary arteries and thickened coronary wall of a 45-year-old male patient. The BB-MR cross-sectional lumen image reveals a circular lumen and an anterior plaque (arrow, Figure 2A). The cross-sectional image of the wall clearly reveals a variably thick proximal RCA, with the wall thinner around the 6 o clock position and thicker in the other sectors (Figure 2B). In that patient, the maximum wall thickness is 3.3 mm. Figure Figure 2. In vivo cross-sectional BB-MR images of lumen (A) and wall (B) of RCA from 45-year-old male patient with ectatic atherosclerotic coronary arteries and thickened coronary wall. Lumen image is obtained without fat saturation; wall image is obtained with fat saturation to better delineate coronary artery wall. Blood flow in coronary artery lumen is suppressed with velocity-selective inversion preparatory pulses. Maximum wall thickness is 3.3 mm. BB-MR cross-sectional lumen image reveals circular lumen and anterior plaque (arrow, A). Crosssectional image of wall clearly reveals variably thick proximal RCA, with wall thinner around 6 o clock position and thicker in other sectors (B). B, Inset, Magnified view of RCA. Some imaging parameters were TR 2 RR intervals, TE 40 ms, cm FOV, 5-mm slice thickness, acquisition matrix, NSA 1, 32 echo train length, 125-kHz data sampling. LA indicates left atrium; RA, right atrium; other abbreviations as in Figure 1. Figure 3. X-ray angiogram from 78-year-old female patient with mild disease on x-ray angiography in proximal LAD (arrow, A). BB-MR cross-sectional lumen image reveals circular lumen (B); wall shows uniformly thickened LAD wall (B) with concentric plaque (maximum thickness 4.13 mm) (C). C, Inset, Magnified view of LAD. Blood flow in coronary artery lumen is suppressed with velocity-selective inversion preparatory pulses. Some imaging parameters were TR 2 RR intervals, TE 40 ms, cm FOV, 4-mm slice thickness, acquisition matrix, NSA 1, 32 echo train length, 125-kHz data sampling. Abbreviations as in Figure 1. 3A shows mild disease in the proximal LAD as seen on x-ray angiography in a 78-year-old female patient. The BB-MR cross-sectional lumen image reveals a circular lumen (Figure 3B), and the wall image shows a concentric plaque (maximum thickness of 4.13 mm) (Figure 3C). Figure 4A shows high-grade stenosis in the proximal LAD on the x-ray angiogram in a 76-year-old male patient. The cross-sectional coronary image at that location shows an obstructed lumen (elliptical shape) on the BB-MR lumen image (Figure 4B). The BB-MR wall image obtained with fat saturation reveals a large eccentric plaque measuring 5.73 mm with heterogeneous signal intensity, possibly due to the different tissue composition (Figure 4C). In the normal subjects, the average maximum coronary wall thickness was mm (range, 0.55 to 1.0 mm; n 40). MR images of coronary arteries in coronary artery disease patients showed atherosclerotic plaques 3.30 to 5.73 mm in maximum wall thickness ( mm; n 25). The difference in maximum coronary wall thickness between the normal subjects and patients was statistically significant (P ). Discussion This study demonstrates, for the first time, that in vivo MRI can provide high-spatial-resolution images of the coronary artery wall in normal and diseased human arteries. Vessel lumen and wall morphology in normal and atherosclerotic human coronary arteries was assessed with a high-resolution

4 Fayad et al In Vivo MRI of Human Coronary Atherosclerosis 509 Figure 4. X-ray angiogram from 76-year-old male patient shows high-grade stenosis in proximal LAD (arrows, A). In vivo crosssectional BB-MR images of LAD lumen (B) shows obstructed lumen (elliptical lumen shape); wall image (C) shows large eccentric plaque with heterogeneous signal intensity (maximum thickness 5.73 mm). Blood flow in coronary artery lumen is suppressed with velocity-selective inversion preparatory pulses. Some imaging parameters were TR 2 RR intervals, TE 40 ms, cm FOV, 4-mm slice thickness, acquisition matrix, NSA 1, 32 echo train length, 125-kHz data sampling. Abbreviations as in Figure 1. BB sequence consisting of the long-etl FSE imaging method with velocity-selective inversion preparatory pulses to eliminate the signal from the flowing blood. The coronary artery wall was clearly seen in all patients and had a distinct MR signal characteristic of surrounding tissue. Data obtained from comparative studies with standard histology and epivascular and intravascular ultrasound are in close agreement with the results of our study Moreover, a recent study in pigs using a similar MRI technique showed an excellent correlation between matched in vivo coronary wall images and histopathology sections for average wall thickness. 36 As seen in Figures 1 through 4, the in-plane spatial resolution used in this study (460 to 750 m) was adequate for the distinction between normal and diseased portions of the coronary arteries and to quantify coronary artery wall and plaque thickness. Possible Further Improvements The slice thickness of the MR image (3 to 5 mm) causes volume averaging (ie, partial-volume effect) and can contribute to an overestimation of the coronary wall. Thinner slice thickness, as used with 3D acquisition techniques, could further improve our coronary artery wall imaging. 18,37 Moreover, zero-filled interpolation of the images can be used to create images to reduce the partial-volume effects in imaging pixels. 38 This can be achieved by appending zeros on each side of the data before Fourier transformation. A relatively high spatial resolution of 0.46 to 0.75 mm was achieved by use of a specially tailored phased-array coil. Other coil designs, such as a smaller anterior 4-element phased-array coil, may improve the spatial resolution and allow the identification of the substructures within atherosclerotic coronary lesions. The BB FSE sequence used in this study has flexible multicontrast capabilities (ie, proton-density or T2 weighting through direct manipulation of TE), which, with improvements in spatial resolution and image contrast, may allow the characterization of the different coronary plaque components. 9,11,12,44,45 A misalignment of the imaging plane and the long axis of the vessel can lead to inaccurate cross-sectional images and lead to errors in wall thickness measurements and plaque imaging. Careful planning in this study minimized these errors. A 3D imaging sequence 18,37 will allow image reformatting in any desired plane direction and thus ensure the proper alignment between the imaging plane and the course of the coronary arteries. The effect of slowly flowing blood near the vessel walls is another phenomenon that could potentially degrade the accuracy of vessel wall imaging with BB techniques. However, preliminary results in our study and with a similar BB-MR sequence in the coronary arteries 36 and in the brain 46 suggest that this effect is minimal. Breath-holding was used to suppress respiratory motion. This limits the maximal duration of the scan and may not be possible in certain patients. We have limited the breathholding duration to 12 to 18 heartbeats (12 to 18 seconds for heartbeats of 60 bpm), which was well tolerated by all subjects. Adequate breath-holding was confirmed by respiratory bellows. Short breath-holding limits the achievable spatial resolution and data sampling for each image, which in turn leads to vessel wall imaging. These problems can be overcome only by a prolongation of the duration of the breath-holds, which is well tolerated in some patients, or by the use of navigator techniques to avoid breath-holding altogether. 47,48 The reduction of the ESP will also lead to shorter breath-holds and reduction of vessel motion blurring. We have visualized the major epicardial coronary arteries but not the side branches. However, coronary atherosclerosis most often involves the proximal portion of the coronary arteries, usually at or near branch sites. 49 Evaluation of the whole extent of the epicardial coronary arteries will be developed in future studies. Moreover, validation and repeatability studies of the MRI findings need to be performed, possibly in patients undergoing intravascular ultrasound. Clinical Implications Atherosclerotic coronary artery plaque rupture is a key event leading to acute coronary syndromes. In vivo MRI provides a means to noninvasively image and assess the morphological features of atherosclerotic and normal human coronary arteries. Future work will certainly aim at the identification of the different plaque components. This may allow the identification of the vulnerable plaques before they rupture and may provide a way to target pharmacological intervention to reduce or prevent cardiovascular disease.

5 510 Circulation August 1, 2000 Acknowledgments This study was supported in part by a Radiological Society of North America seed grant (Dr Fayad) and NIH grants P50-HL and R01-HL We acknowledge the help of Karen Metroka, Stella Palencia, and Mary Ann Whelan-Gales in patient selection and recruitment. We thank Paul Wisdom and John Abela for help in MRI. References 1. Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation. 1995;92: Fuster V, Fayad ZA, Badimon JJ. Acute coronary syndromes: biology. Lancet. 1999;353(suppl 2):SII5 SII9. 3. Davies MJ. Stability and instability: two faces of coronary atherosclerosis: the Paul Dudley White Lecture, Circulation. 1996;94: Glagov S, Weisenberg E, Zarins CK, et al. Compensatory enlargement of human atherosclerotic coronary arteries. N Engl J Med. 1987;316: Ge J, Chirillo F, Schwedtmann J, et al. Screening of ruptured plaques in patients with coronary artery disease by intravascular ultrasound. Heart. 1999;81: Callister TQ, Raggi P, Cooil B, et al. Effect of HMG-CoA reductase inhibitors on coronary artery disease as assessed by electron-beam computed tomography. N Engl J Med. 1998;339: Uchida Y, Nakamura F, Tomaru T, et al. Prediction of acute coronary syndromes by percutaneous coronary angioscopy in patients with stable angina. Am Heart J. 1995;130: Skinner MP, Yuan C, Mitsumori L, et al. Serial magnetic resonance imaging of experimental atherosclerosis detects lesion fine structure, progression and complications in vivo. Nat Med. 1995;1: Fayad ZA, Fallon JT, Shinnar M, et al. Noninvasive in vivo highresolution magnetic resonance imaging of atherosclerotic lesions in genetically engineered mice. Circulation. 1998;98: Yuan C, Beach KW, Smith LH Jr, et al. Measurement of atherosclerotic carotid plaque size in vivo using high resolution magnetic resonance imaging. Circulation. 1998;98: Toussaint JF, LaMuraglia GM, Southern JF, et al. Magnetic resonance images lipid, fibrous, calcified, hemorrhagic, and thrombotic components of human atherosclerosis in vivo. Circulation. 1996;94: Fayad ZA, Tamana N, Badimon JJ, et al. In-vivo MR characterization of plaques in the thoracic aorta. Circulation. 1998;98(suppl I):I-515. Abstract. 13. Worthley SG, Helft G, Fuster V, et al. High resolution ex vivo magnetic resonance imaging of in situ coronary and aortic atherosclerosis plaque in a porcine model. Atherosclerosis. 2000;150: Manning WJ, Li W, Edelman RR. A preliminary report comparing magnetic resonance coronary angiography with conventional angiography. N Engl J Med. 1993;328: Slavin GS, Riederer SJ, Ehman RL. Two-dimensional multishot echoplanar coronary MR angiography. Magn Reson Med. 1998;40: Meyer CH, Hu BS, Nishimura DG, et al. Fast spiral coronary artery imaging. Magn Reson Med. 1992;28: Jara H, Yu BC, Caruthers SD, et al. Voxel sensitivity function description of flow-induced signal loss in MR imaging: implications for black-blood MR angiography with turbo spin-echo sequences. Magn Reson Med. 1999;41: Alexander AL, Buswell HR, Sun Y, et al. Intracranial black-blood MR angiography with high-resolution 3D fast spin echo. Magn Reson Med. 1998;40: Edelman RR, Mattle HP, Wallner B, et al. Extracranial carotid arteries: evaluation with black blood MR angiography. Radiology. 1990;177: Schmalbrock P, Hacker VA, Rao A. Three-dimensional steady-state MR angiography of the lower extremities. J Magn Reson Imaging. 1994;4: Le Roux P, Gilles RJ, McKinnon GC, et al. Optimized outer volume suppression for single-shot fast spin-echo cardiac imaging. J Magn Reson Imaging. 1998;8: Edelman RR, Chien D, Kim D. Fast selective black blood MR imaging. Radiology. 1991;181: Simonetti OP, Finn JP, White RD, et al. Black blood T2-weighted inversion-recovery MR imaging of the heart. Radiology. 1996;199: Fayad ZA, Connick TJ, Axel L. An improved quadrature or phased-array coil for MR cardiac imaging. Magn Reson Med. 1995;34: Silver MS, Joseph RI, Hoult DI. Highly selective /2 and pulse generation. J Magn Res. 1984;59: Shinnar M, Eleff S, Subramanian H, et al. The synthesis of pulse sequences yielding arbitrary magnetization vectors. Magn Reson Med. 1989;12: Alsop DC. The sensitivity of low flip angle RARE imaging. Magn Reson Med. 1997;37: Le Roux P, Hinks RS. Stabilization of echo amplitudes in FSE sequences. Magn Reson Med. 1993;30: Hennig J. Multiecho imaging sequences with low refocusing flip angles. J Magn Reson. 1988;78: Kuroda K, Oshio K, Mulkern RV, et al. Optimization of chemical shift selective suppression of fat. Magn Reson Med. 1998;40: Listerud J, Einstein S, Outwater E, et al. First principles of fast spin echo. Magn Reson Q. 1992;8: McPherson DD, Hiratzka LF, Lamberth WC, et al. Delineation of the extent of coronary atherosclerosis by high-frequency epicardial echocardiography. N Engl J Med. 1987;316: Potkin BN, Bartorelli AL, Gessert JM, et al. Coronary artery imaging with intravascular high-frequency ultrasound. Circulation. 1990;81: McPherson DD, Johnson MR, Alvarez NM, et al. Variable morphology of coronary atherosclerosis: characterization of atherosclerotic plaque and residual arterial lumen size and shape by epicardial echocardiography. J Am Coll Cardiol. 1992;19: Podesser BK, Neumann F, Neumann M, et al. Outer radius-wall thickness ratio, a postmortem quantitative histology in human coronary arteries. Acta Anat. 1998;163: Worthley SG, Helft G, Fuster V, et al. In vivo high-resolution MRI non-invasively defines coronary lesion size and composition in a porcine model. Circulation. 1999;100(suppl I):I-521. Abstract. 37. Luk-Pat GT, Gold GE, Olcott EW, et al. High-resolution threedimensional in vivo imaging of atherosclerotic plaque. Magn Reson Med. 1999;42: Du YP, Parker DL, Davis WL, et al. Reduction of partial-volume artifacts with zero-filled interpolation in three-dimensional MR angiography. J Magn Reson Imaging. 1994;4: Gold GE, Pauly JM, Glover GH, et al. Characterization of atherosclerosis with a 1.5-T imaging system. J Magn Reson Imaging. 1993;3: Shinnar M, Fallon JT, Wehrli S, et al. The diagnostic accuracy of ex vivo magnetic resonance imaging for human atherosclerotic plaque characterization. Arterioscler Thromb Vasc Biol. 1999;19: Toussaint JF, Southern JF, Fuster V, et al. Water diffusion properties of human atherosclerosis and thrombosis measured by pulse field gradient nuclear magnetic resonance. Arterioscler Thromb Vasc Biol. 1997;17: Pachot-Clouard M, Vaufrey F, Darrasse L, et al. Magnetization transfer characteristics in atherosclerotic plaque components assessed by adapted binomial preparation pulses. Magma. 1998;7: Lin W, Abendschein DR, Haacke EM. Contrast-enhanced magnetic resonance angiography of carotid arterial wall in pigs. J Magn Reson Imaging. 1997;7: Martin AJ, Gotlieb AI, Henkelman RM. High-resolution MR imaging of human arteries. J Magn Reson Imaging. 1995;5: Yuan C, Murakami JW, Hayes CE, et al. Phased-array magnetic resonance imaging of the carotid artery bifurcation: preliminary results in healthy volunteers and a patient with atherosclerotic disease. J Magn Reson Imaging. 1995;5: Melhem ER, Jara H, Yucel EK. Black blood MR angiography using multislab three-dimensional TI-weighted turbo spin-echo technique: imaging of intracranial circulation. Am J Roentgenol. 1997;169: Ehman RL, Felmlee JP. Adaptive technique for high-definition MR imaging of moving structures. Radiology. 1989;173: Botnar RM, Stuber M, Danias PG, et al. Improved coronary artery definition with T2-weighted, free-breathing, three-dimensional coronary MRA. Circulation. 1999;99: Hochman JS, Phillips WJ, Ruggieri D, et al. The distribution of atherosclerotic lesions in the coronary arterial tree: relation to cardiac risk factors. Am Heart J. 1988;116: