Unenhanced Whole-Heart Coronary MRA: Prospective Intraindividual Comparison of 1.5-T SSFP and 3-T Dixon Water-Fat Separation GRE Methods Using Coronary Angiography as Reference

BACKGROUND. Coronary MRA is commonly performed at 1.5 T using SSFP acquisitions. Coronary MRA performed at 3 T using SSFP is limited due to impaired fat suppression and has been typically investigated using contrast-enhanced techniques. A Dixon fat-water separation gradient-recalled echo (GRE) method may enable high-quality unenhanced 3-T coronary MRA. OBJECTIVE. The purpose of this study was to compare 1.5-T SSFP and 3-T Dixon water-fat separation GRE methods for unenhanced whole-heart coronary MRA in patients with suspected coronary artery disease (CAD). METHODS. This prospective study included 44 patients (27 men and 17 women; mean age, 59 ± 8 [SD] years) with an intermediate to high risk of CAD who underwent both 1.5-T SSFP and 3-T Dixon GRE coronary MRA examinations before undergoing coronary angiography (CAG). Two radiologists independently assessed coronary arteries in terms of subjective image quality (on a scale of 1-5, with 5 denoting the highest image quality), number of visible segments, apparent contrast-to-noise ratio (CNR; vs myocardium), and presence of significant stenoses. Methods were compared using the mean of the readers' values for apparent CNR and using consensus interpretations for other measures. CAG served as the reference standard for detecting the presence of stenoses. RESULTS. Expressed as a kappa coefficient, interobserver agreement was 0.85 for image quality, 0.85 for segment visibility, and 0.83 for stenosis, and expressed as an intraclass correlation coefficient, interobserver agreement was 0.92 for apparent CNR. The mean overall image quality score was 4.0 ± 1.1 for 3-T Dixon GRE versus 3.0 ± 1.2 for 1.5-T SSFP. The percentage of visible segments for 3-T Dixon GRE versus 1.5-T SSFP was 96.7% versus 88.9% for all segments, 96.9% versus 90.1% for distal segments, and 93.1% versus 77.2% for branch segments. The mean overall apparent CNR was 93.2 ± 29.2 for 3-T Dixon GRE versus 80.8 ± 27.9 for 1.5-T SSFP. The 3-T Dixon GRE method, compared with the 1.5-T SSFP method, showed higher sensitivity and specificity in per-vessel analysis (87.9% vs 77.3% and 83.3% vs 60.6%, respectively), per-segment analysis (84.6% vs 74.8% and 90.9% vs 79.6%, respectively), and per-segment analysis of distal and branch segments (89.7% vs 75.9% and 89.7% vs 73.7%, respectively). CONCLUSION. For unenhanced coronary MRA, 3-T unenhanced Dixon GRE had better image quality and diagnostic performance than 1.5-T SSFP, particularly for distal and branch segments. CLINICAL IMPACT. The 3-T Dixon GRE technique may be preferred to the current clinical standard of the 1.5-T SSFP technique for unenhanced coronary MRA.