Whole-Heart Coronary Magnetic Resonance Angiography at 1.5 Tesla

Abstract

To evaluate the impact of the blood-pool contrast agent gadofosveset trisodium on diagnostic accuracy of whole-heart coronary magnetic resonance angiography (CMRA) at 1.5 Tesla. Thirty consecutive patients with suspected coronary artery disease underwent free-breathing whole-heart CMRA at 1.5 Tesla. CMRA was performed with a T2-prepared steady-state free precession sequence (unenhanced CMRA) and an inversion-recovery-prepared steady-state free precession sequence after administration of gadofosveset trisodium (contrast-enhanced CMRA). Two readers independently performed a per-segment evaluation of CMRA (8 proximal and mid coronary segments) for detection of significant stenosis (≥50%) using invasive coronary angiography as reference. Disagreement was settled by consensus reading and interobserver variability was assessed using an unweighted kappa statistic. Whole-heart CMRA was successfully performed in 27 patients. The percentage of assessable segments was significantly lower on unenhanced CMRA compared with contrast-enhanced CMRA (Reader 1: 79% [170/216] vs. 89% [192/216], respectively; Reader 2: 73% [157/216] vs. 87% [188/216], respectively; P < 0.001). Intention-to-diagnose analysis of the consensus reading yielded sensitivity, specificity, and diagnostic accuracy of unenhanced versus contrast-enhanced CMRA as follows: 73.1% versus 73.1% (P = 1.0), 68.3% versus 80.2% (P = 0.002), and 68.9% versus 79.3% (P = 0.004), respectively. The kappa value for interobserver agreement was 0.61 (95% confidence interval = 0.50-0.72) for unenhanced CMRA and 0.72 (95% confidence interval = 0.62-0.82) for contrast-enhanced CMRA. The blood-pool contrast agent gadofosveset trisodium increased the number of assessable coronary segments on whole-heart CMRA in comparison to unenhanced whole-heart CMRA. The impact of gadofosveset trisodium on diagnostic accuracy, however, was only minor.