Detection of coronary artery stenoses using multi-detector CT with 16 x 0.75 collimation and 375 ms rotation

Aims Insufficient spatial and temporal resolutions have limited image quality and accuracy of multi-detector CT (MDCT) for coronary artery visualization and detection of stenoses. We assessed the accuracy of a new 16-slice scanner with 370 ms rotation and 0.75 mm collimation for detection of coronary stenoses using an analysis approach based on coronary artery segments. Methods and results Fifty consecutive patients scheduled for diagnostic coronary angiography in stable clinical condition and sinus rhythm were enrolled. All patients with a heart rate > 60 b.p.m. received 100 mg atenolol p.o. and up to four doses of 5 mg metoprolol i.v. before the scan. MDCT was performed using 16x0.75 mm collimation, 120 kV, and ECG-gated tube current modulation. Ninety millilitres of contrast agent was injected intravenously. MDCT images were visually analysed using the 16-segment coronary artery model of the American Heart Association and compared with invasive, quantitative coronary angiography in a blinded fashion. A significant stenosis was assumed if the diameter reduction was >= 50%. Mean heart rate was 58 b.p.m. during MDCT. After exclusion of two patients with not fully evaluable data sets, MDCT correctly identified at least one coronary stenosis in all 25 patients with significant coronary lesions in angiography and correctly demonstrated the absence of stenoses in 19/23 patients (sensitivity 100%, specificity 83%). Sensitivity and specificity for all 50 patients were 93 and 83%, respectively. On a per-segment basis, nine coronary segments distal of total occlusions and 128 coronary segments with a reference diameter < 1.5 mm were excluded from the analysis. Twenty-eight of the included 663 segments (4%) were unevaluable due to calcification or motion artefact. In the remaining 635 segments, 50/53 stenoses were detected by MDCT (sensitivity 94%, specificity 96%, negative predictive value 99%, positive predictive value 69%). Conclusion Increasing temporal and spatial resolutions of MDCT lead to improved evaluation and diagnostic accuracy for detection of coronary stenoses.