Objective: To prospectively compare the accuracy of coronary CT angiography (CCTA) and conventional coronary angiography (CCA) for stenosis detection using composite findings from both tests as an enhanced reference standard.
Methods: One hundred thirteen patients underwent CCTA and CCA. Per-segment and per-patient accuracy of CCTA compared with initial CCA interpretation were determined. Angiographers were then unblinded to the CCTA results and re-evaluation of the CCA studies was performed with knowledge of CCTA findings, which was used as an enhanced reference standard to compare the diagnostic accuracy of CCTA versus CCA.
Results: When using the enhanced reference standard instead of initial CCA interpretation, CCTA accuracy for identifying segments (patients) with ≥50% stenosis increased from 97.7% (96.5%) to 98.1% (98.2%), sensitivity from 90.5% (100%) to 90.8% (100%), and specificity from 98.4% (94.3%) to 98.9% (97.1%). CCTA identified six segments and two patients with stenoses ≥50% missed on initial CCA interpretation. Compared with the enhanced reference standard the accuracies of CCTA and of initial CCA interpretation were not different (p = 0.87).
Conclusion: CCTA compares favourably with CCA for stenosis detection. Use of a composite reference standard combining findings from both tests can control for the effect of false-negative CCA results when evaluating the accuracy of CCTA.