Objective To conduct a meta-analysis-based evaluation of dual-source CT (DSCT) in the diagnosis of coronary artery disease (CAD).
Methods English literatures were searched in Medline. A study was identified if it: (1) used DSCT angiography as a diagnostic test for detection of significant coronary stenosis (≥50% diameter stenosis) in patients with suspected CAD, and (2) used coronary angiography as the reference standard. Sensitivity, specificity with 95% CI on the patient basis was pooled by bivariate random-effects model. Hierarchical weighted symmetric summary receiver-operating curve was also estimated. Likelihood ratios were calculated on the basis of the pooled estimates of sensitivity and specificity. By means of Bayes’ theorem integrated with pooled likelihood ratios, the clinical utility of the results was estimated.
Results 10 studies included 1271 patients were identified. Mean rate of nonevaluable patient and segment was 5.6% (33/590) and 2.3% (271/11745) respectively. The pooled results were as follows: sensitivity of 99% (97%–99%), specificity of 86% (79%–90%), positive likelihood ratio of 6.84, and negative likelihood ratio of 0.01. Given a pre-test probability below 84%, the likelihood of disease was less than 5% based on a negative DSCT. If a pre-test probability was above 13%, the likelihood of disease was more than 50% on a positive DSCT.
Conclusions DSCT angiography has good diagnostic performance in the detection of CAD, however, it could not replace conventional coronary angiography totally.