ReviewDiagnostic accuracy of dual-source CT coronary angiography in patients with atrial fibrillation: Meta analysis

Rationale and objectiveTo synthesize the available data to underscore the diagnostic accuracy of dual-source CT (DSCT) coronary angiography in patients with atrial fibrillation (AF).Materials and methodsWe searched in the electronic databases of PubMed for all published studies that examined patients with AF using DSCT. We used an exact binomial rendition of the bivariate mixed-effects regression model to synthesize the diagnostic data.ResultsThe positive and negative likelihood ratios (LRs) at the patient level were 6.0 (CI, 3.6-10.1) and 0.03(CI, 0.004-0.2), respectively. The negative predictive values higher than 90% were available for a CAD prevalence <78%. The pooled vessel- and segment-level estimates showed higher positive and negative LRs than the patient-level estimates (15.3 [CI, 9.8-23.9] and 0.1 [CI, 0.07-0.3]; 25.1 [CI, 10.8-58.5] and 0.2 [CI, 0.2-0.3], respectively). No statistically significant heterogeneity between studies and publication bias were found at the patient level estimate. A sensitivity analysis showed that no study influenced the pooled results larger than 0.02.ConclusionsCardiac angiography with DSCT can be applied as an imaging test for ruling out CAD in patient with AF. However, DSCT angiography may be not an effective tool for risk stratification for the high negative LR at the artery and segment levels.