Forty-four patients in sinus rhythm with suspected coronary artery disease underwent 16-row multislice computed tomography coronary angiography and conventional coronary angiography. Two protocols for image analysis were applied to the multislice computed tomographic images: standard projections versus interactive postprocessing. The diagnostic accuracy of both methods for the detection of significant lesions (>50% lumen reduction) was compared with quantitative coronary angiography. Sensitivity and specificity were 58% and 96% and 96% and 97%, for standard projections and interactive postprocessing protocol, respectively.