Diagnostic accuracy of first generation dual-source computed tomography in the assessment of coronary artery disease: a meta-analysis from 24 studies

Abstract

The objective of this study is to evaluate the diagnostic accuracy of the first generation dual-source computed tomography (DSCT) in the diagnosis of coronary artery disease (CAD). We selected articles from four databases (Pubmed, Embase, the Cochrane central register of controlled trials (CENTRAL) and Chinese biomedical literature database. The strict study selection was made, and two reviewers independently extracted data back-to-back from included studies. Meta-Disc version 1.4 was used to obtain the pooled results. 24 studies were included in meta-analysis. A cut off point of≥50% stenosis was used in all the studies to define significant coronary artery stenosis. In patient-based analysis (n=801), pooled sensitivity was 0.980 [95% confidence interval (CI):0.970-0.990], specificity 0.870 (95% CI: 0.830-0.900), median positive predictive value (PPV) across studies 0.876 (range from 0.741 to 0.943) and negative predictive value (NPV) 0.964 (range from 0.900 to 1.000). In vessel-based analysis (n=3,620) DSCT pooled sensitivity was 0.957 (95% CI: 0.943-0.969), specificity 0.930 (95% CI: 0.910-0.940), median PPV across studies 0.838 (range from 0.534 to 0.964) and NPV 0.973 (range from 0.885 to 0.996). In segment-based analysis (n=6,177) DSCT pooled sensitivity was 0.915 (95% CI: 0.901-0.928), specificity 0.959 (95% CI: 0.956-0.963), median PPV 0.782 (range from 0.320 to 0.927) and NPV 0.985 (range from 0.929 to 0.999). In subgroups analysis, pooled sensitivity and specificity in segment based analysis were 93.1 and 92.3% when heart rate (HR) is beyond 70bpm; when HR was below 70bpm, the sensitivity was similar (93%), but specificity increased a little from 92.3 to 94%. When analysed based on segment with a cut-off calcium score of 400, the sensitivity was slightly higher in the subgroup with a score over 400 than in the subgroup with a score below 400 (94 vs. 91%), while the specificity was much lower in the subgroup with the high calcium score than the subgroup with the low calcium score (85 vs. 96%). For subgroups with heart rate beyond and below 65bpm in patient-based analysis, sensitivities were 0.95 (95% CI: 0.86-0.99) and 0.98 (95% CI 0.91-1.00), respectively, while the specificities were 0.88 (95% CI 0.81-0.94) and 0.85 (95% CI 0.77-0.91), respectively. The area under the receiver operating characteristic curve (AUC) in the two subgroups were 0.9608 and 0.9786, respectively. DSCT is highly sensitive for patient-based analysis of CAD and has high specificity and NPV for segment-based analysis of CAD. First generation DSCT may have a role in the evaluation of patients with chest pain as a simple non-invasive examination because of its ability to diagnose or exclude significant CAD.