Free-breathing 3D coronary MRA: The impact of "isotropic" image resolution

During conventional x-ray coronary angiography, multiple projections of the coronary arteries are acquired to define coronary anatomy precisely. Due t o time constraints, coronary magnetic resonance angiography (MRA) usually p rovides only one or two views of the major coronary vessels. A coronary MRA approach that allowed for reconstruction of arbitrary isotropic orientatio ns might therefore be desirable. The purpose of the study was to develop a three-dimensional (3D) coronary MRA technique with isotropic image resoluti on in a relatively short scanning time that allows for reconstruction of ar bitrary views of the coronary arteries without constraints given by anisotr opic voxel size. Eight healthy adult subjects were examined using a real-ti me navigator-gated and corrected free-breathing interleaved echoplanar (TFE -EPI) 3D-MRA sequence. Two 3D datasets were acquired for the left and right coronary systems in each subject, one with anisotropic (1.0 x 1.5 x 3.0 mm , 10 slices) and one with "near" isotropic (1.0 x 1.5 x 1.0 mm, 30 slices) image resolution. All other imaging parameters were maintained. In all case s, the entire left main (LM) and extensive portions of the left anterior de scending (LAD) and the right coronary artery (RCA) were visualized. Objecti ve assessment of coronary vessel sharpness was similar (41% +/- 5% vs. 42% +/- 5%; P = NS) between in-plane and through-plane views with "isotropic" v oxel size but differed (32% +/- 7% vs. 23% +/- 4%; P < 0.001) with nonisotr opic voxel size. In reconstructed views oriented in the through-plane direc tion, the vessel border was 86% more defined (P < 0.01) for isotropic compa red with anisotropic images. A smaller (30%; P < 0.001) improvement was see n for in-plane reconstructions. Vessel diameter measurements were view inde pendent (2.81 +/- 0.45 mm vs. 2.66 +/- 0.52 mm; P = NS) for isotropic, but differed (2.71 +/- 0.51 mm vs. 3.30 +/- 0.38 mm: P < 0.001) between anisotr opic views. Average scanning time was 2:31 +/- 0:57 minutes for anisotropic and 7:11 +/- 3:02 minutes for isotropic image resolution (P < 0.001). We p resent a new approach for "near" isotropic 3D coronary artery imaging, whic h allows for reconstruction of arbitrary views of the coronary arteries. Th e good delineation of the coronary arteries in all views suggests that isot ropic 3D coronary MRA might be a preferred technique for the assessment of coronary disease, although at the expense of prolonged scan times. Comparat ive studies with conventional x-ray angiography are needed to investigate t he clinical utility of the isotropic strategy. (C) 2000 Wiley-Liss, Inc.