Purpose. This study was undertaken to evaluate the diagnostic accuracy of computed tomography coronary angiography (CT-CA) for the detection of significant coronary artery stenosis (50% lumen reduction) compared with conventional coronary angiography (CCA) in a registry and to review major multicentre trials. Materials and methods. A total of 1,372 patients (882 men, 490 women; mean age 59.3 +/- 11.9 years) in sinus rhythm were studied with CT-CA (64-slice technology) and CCA. The diagnostic accuracy of CT-CA was evaluated against quantitative CCA as a reference standard for coronary artery stenosis. Positive and negative likelihood ratios and inter- and intraobserver agreement were calculated. Results. The prevalence of disease was 53%. CCA demonstrated the absence of significant coronary artery disease in 46.6% (639/1372), single-vessel disease in 24.7% (337/1372) and multivessel disease in 28.9% (396/1372) of patients. In per-patient analysis sensitivity, specificity and positive and negative predictive value of CT-CA were 99% [confidence interval (CI) 97-99], 92% (CI 89-94), 94% (CI 91-95) and 99% (CI 97-99), respectively. Per-patient and per-segment likelihood ratios (LR+=12.4 and LR =0.011; LR+=18.3 and LR-=0.064, respectively), were good. Inter- and intraobserver variability was 0.78 and 0.85, respectively. Conclusions. CT-CA is a reliable diagnostic modality both in terms of sensitivity and negative predictive value. Differences in trial results are also due to the different parameters used for patient inclusion.