Coronary MR angiography with steady-state free precession: individually adapted breath-hold technique versus free-breathing technique.

PURPOSE To compare image quality and coronary artery stenosis detection with breath-hold (BH) and free-breathing navigator-gated (NAV) coronary magnetic resonance (MR) angiography performed with the same imaging sequence (steady-state free precession) and identical spatial resolution in patients suspected of having coronary artery disease. MATERIALS AND METHODS Forty consecutive patients suspected of having coronary artery disease underwent steady-state free precession MR imaging of the left or the right coronary artery twice. Correction of breathing motion was performed once with NAV and again with BH. Maximal BH duration and coronary artery rest period were individually determined, and duration of data acquisition was adapted (parallel imaging with different sensitivity encoding factors was used). Quantitative analysis of coronary MR angiography data was performed with multiplanar reformatting software to determine visual score for image quality, vessel sharpness, visible vessel length, and number of visible side branches. Diagnostic accuracy for detection of coronary stenosis of 50% or greater was determined in comparison with results of conventional invasive angiography. The two techniques were compared regarding differences in angiographic parameters with paired Student t testing. chi(2) or Fisher exact testing was used when appropriate. RESULTS More coronary artery segments were assessable with NAV than with BH MR angiography (254 [79.4%] vs 143 [44.7%] of 320 segments). Overall sensitivity and specificity with NAV were 72% (26 of 36 segments) and 91.7% (200 of 218 segments), versus 63% (12 of 19 segments) and 82.3% (102 of 124 segments) with BH; NAV enabled correct diagnosis in 13% more segments. BH yielded nondiagnostic images in 14 patients, while NAV yielded diagnostic images in all patients. When these 14 patients were excluded, there was a significant increase in visual score for left (3.0 vs 2.4, P <.01) and right (3.3 vs 3.0, P <.05) coronary arteries and no significant difference in vessel sharpness but significant improvement in visible vessel length in left coronary artery (85.9 vs 71.4 mm, P =.003) and number of visible side branches in left (4.9 vs 3.9, P =.04) and right (2.8 vs 2.4, P =.04) coronary arteries on NAV images as compared with BH images. CONCLUSION Free-breathing NAV was superior to BH coronary MR angiography in terms of image quality and diagnostic accuracy of stenosis detection.