Accuracy of multislice computed tomography coronary angiography: a pooled estimate

A PUBMED search was performed for peer-reviewed studies published in English from 2002 through August 2006. Accuracy parameters for detection of obstructive coronary artery disease by MSCT (multislice computed tomography) coronary angiography were analyzed on patient and segment basis. Pooled estimates of sensitivity (SN), specificity, positive predictive value, and negative predictive value (NPV) in patient-based analyses were 93%, 82%, 83%, and 92%, respectively. For 64/40-slice MSCT the respective accuracy estimates were 96%, 91%, 93%, and 96% and were better compared with 16-slice MSCT. Also, more segments were evaluated by 64-slice MSCT (96%) compared with 16-slice MSCT (86%). The SN and NPV of MSCT coronary angiography for patients with low coronary artery calcium (CAC) score were 92% and 99% compared with 77% and 89% for patients with high CAC score. The accuracy estimates for MSCT in studying coronary artery by-pass grafts and intracoronary stents showed a SN and NPV of 97% and 97% for graft occlusion or stenosis and only 71% and 93% respectively for in-stent restenosis. Diagnostic accuracy of MSCT coronary angiography has improved with the newer 64-slice versions. High CAC scores can affect the accuracy of MSCT coronary angiography. Although the accuracy of MSCT in evaluating native vessel and graft disease has improved significantly, its reliability in studying intracoronary stents remains uncertain.