Comparison of three-dimensional volume-targeted thin-slab FIESTA magnetic resonance angiography and 64-multidetector computed tomographic angiography for the identification of proximal coronary stenosis

Background: Based on recent clinical data, an imaging strategy of identifying proximal coronary disease allows further management decisions in patients with stable angina pectoris. We aimed to compare diagnostic accuracy of non-contrast fast steady-state (FIESTA) magnetic resonance angiography (MRA) with 64-multidetector computed tomographic angiography (CTA), using conventional coronary angiography (CA) as the reference standard.

Methods: Thirty patients with suspected coronary artery disease consented to participate in an institutional review board-approved protocol. Coronary MRA was performed at 1.5 T using a respiratory navigator and electrocardiogram-gated three-dimensional FIESTA pulse sequence. CTA images were acquired using a 64-multidetector computed tomographic scanner, using beta blockade to reduce the heart rate to less than 70 bpm. Coronary luminal stenosis >50% was identified. Plaques were classified as non-calcified, mixed, or calcified on CTA, and as high-, intermediate-, or low-signal on FIESTA MRA.

Results: Compared to CA, the sensitivity, specificity, and overall accuracy for detection of >50% proximal coronary stenoses were 83.0%, 86.9%, and 86.1% for MRA and 85.1%, 87.2%, and 86.8% for CTA, respectively. For the 24 calcified stenoses, MRA corrected 16 segments that overestimated on CTA and MRA had an accuracy of 75% in evaluating calcified plaques.

Conclusions: High-resolution three-dimensional FIESTA MRA and CTA have a similar accuracy in detecting proximal coronary stenosis. The clinical impact of identification of proximal disease in patients with stable CAD needs to be examined in future studies.