Purpose: To investigate the accuracy of different computed tomographic (CT) reformation techniques in assessing the coronary arteries.
Materials and methods: Sixty-four patients undergoing both multi-detector row CT and invasive coronary angiography were consecutively included in a retrospective study. CT scans were obtained with collimation of 4 x 1 mm, pitch of 1.5, and rotation time of 500 msec. Retrospective electrocardiographic gating was used for image reconstruction, with 1.25-mm section thickness and 0.5-mm increment. The CT data set of each patient was evaluated by independent observers using transverse scanning, virtual endoscopic, and three-dimensional reformation and multiplanar reformation.
Results: Hemodynamically relevant stenoses (>50%) were detected with highest sensitivity at transverse scanning (58 of 79 [73.4%] stenoses), followed by virtual endoscopic (38 of 79 [48.1%] stenoses) and three-dimensional reformation (34 of 79 [43.0%] stenoses), and multiplanar reformation (37 of 79 [46.8%] stenoses). Atherosclerotic plaques were identified with comparable sensitivities at transverse scanning (143 of 218 plaques [65.6%]) and at three-dimensional (139 of 218 [63.8%] plaques) and virtual endoscopic reformation (136 of 218 [62.4%] plaques). Multiplanar reformation had distinctly poorer results (217 of 218 [58.3%] plaques). Combined interpretation with all four techniques increased sensitivity to 74.7% (59 of 79) for stenosis and 71.6% (156 of 218) for atherosclerosis. Calculated overall specificity was 91.4% or greater. Sufficient vascular evaluation was possible only in vessels larger than 1.6 mm in diameter. Thus, even in patients with heart rates below 60 bpm, only 80.0% of all coronary segments could be visualized, while at higher frequencies, visibility decreased to 66.2%.
Conclusion: Although multi-detector row CT is a favorable alternative procedure in evaluating coronary arteries, its clinical value still is restricted to low heart rates and proximal coronary arterial segments.