A clinical strategy to improve the diagnostic accuracy of 1.5-T non-contrast MR coronary angiography for detection of coronary artery disease: combination of whole-heart and volume-targeted imaging

Abstract

Objectives To evaluate the diagnostic performance of 1.5-T non-contrast MR coronary angiography (MRCA) for detection of coronary artery disease (CAD) using whole-heart imaging combined with volume-targeted imaging. Methods Forty-five patients scheduled for conventional coronary angiography (CAG) underwent 1.5-T free-breathing non-contrast steady-state free-precession MRCA, including whole-heart and subsequent three-vessel volume-targeted imaging. Coronary stenosis was evaluated as follows: (1) by whole-heart MRCA alone; (2) by combined MRCA (whole-heart plus volume-targeted images). The diagnostic performance for significant stenosis (>= 50% diameter reduction) was evaluated and compared using CAG as a reference standard. Results Combined MRCA was completed in all 45 patients with a total acquisition time of 16.6 +/- 3.3 min. The sensitivity, specificity, and accuracy of combined MRCA per patient were 97% (95% confidence interval 84-100%), 83% (52-98%), and 93% (82-98%), respectively. The areas under the receiver operating characteristic curve of combined MRCA were significantly higher than those of whole-heart MRCA on a per patient (0.97 versus 0.85,p= 0.0078) and per vessel (0.96 versus 0.86,p< 0.0001) basis. Compared with whole-heart MRCA, combined MRCA showed equally high sensitivity but significantly improved specificity on a per patient (83% versus 25%,p= 0.016) and per vessel (85% versus 50%,p< 0.0001) basis. Conclusions 1.5-T non-contrast MRCA combining whole-heart and volume-targeted imaging can detect significant CAD with high sensitivity and moderate specificity. Combined MRCA significantly improves specificity compared with whole-heart imaging alone.