Objective: To evaluate the feasibility of using relative low-dose scan protocols in coronary imaging with 64-row MDCT. Materials and methods: A pulsating cardiac phantom was used to simulate coronary arteries of two sizes Q and 5 mm in diameter) with three stenosis degrees (25, 50 and 75%) at 55 bpm heart rate. Cardiac scans were performed on a 64-row MDCT scanner (GE LightSpeed VCT) with rotation time of 350 ms and pitch of 0.2 under six different scan protocols: 120 kV/650 mA, 1137.5 mAs (effective) (CTDIvol 121.69 mGy), 120 kV/550 mA, 962.5 mAs (CTDIvol 102.96 mGy), 120 kV/450 mA, 787.5 mAs (CTDIvol 84.24 mGy), 120 kV/350 mA, 612.5 mAs (CTDIvol 65.52 mGy), 100 kV/590 mA, 1032.5 mAs (CTDIvol 65.17 mGy) and 140 kV/390 mA, 682.5 mAs (CTDIvol 102.22 mGy). The simulative coronary arteries were filled with contrast media to reach 300 HU in the lumen. Background noise was measured to describe the basic image quality accordingly. CNR, SNR and contour sharpness represented in slope of CT density curve was calculated as well. Measured stenosis area and rates, described by the percentage area of stenosis on the cross-section images were also calculated. Results: The corresponding image noise levels described in standard deviation of background signals varied with radiation dose, CNR and SNR mainly varied with tube current. The contour sharpness, which can reflect actual spatial resolution, is affected mainly by tube voltage. The first five protocols depicted obviously steeper curves than the sixth one (P < 0.05). As for 25% stenosis, there was no significant difference among the stenosis rates of the six protocols (P > 0.05). As for evaluation on 50 and 75% stenosis, there was no significant difference between the first two protocols, and between the second two protocols as well. However, significant difference presented between these two groups (P > 0.05). When comparing the groups with similar radiation dose, protocols with lower tube voltage gain more accuracy in representing stenosis area and rate. Conclusion: Dose level and corresponding image quality is relevant to the accuracy of stenosis evaluation on simulated coronary arteries with 64-row MDCT. In this study, we find relative low-dose protocols with acceptable image quality showed a tendency of overestimating stenosis. Furthermore, using a lower tube voltage and higher tube current to gain accurate imaging result is more applicable than other protocols with the same radiation dose level. (c) 2007 Elsevier Ireland Ltd. All rights reserved.