This study was conducted to investigate the feasibility of using 3D virtual intravascular endoscopy (VIE) as a novel approach for characterization of calcified coronary plaques with the aim of differentiating superficial from deep calcified plaques, thus improving assessment of coronary stenosis. A total of 61 patients with suspected coronary artery disease were included in the study. Minimal lumen diameter (MLD) was measured and compared between coronary CT angiography (CCTA) (>= 64-slice) and invasive coronary angiography (ICA) with regard to the measurement bias, whereas VIE findings were correlated with CCTA with respect to the diagnostic performance of coronary stenosis and the area under the curve (AUC) by receiver-operating characteristic curve analysis (ROC). In all 3 coronary arteries, the CCTA consistently underestimated the MLD relative to the ICA (P<0.001). On a per-vessel assessment, the sensitivity, specificity, positive predictive value, and negative predictive value and 95% confidence interval (CI) were 94%(95% CI: 61%, 100%), 27%(95% CI: 18%, 38%), 33%(95% CI: 23%, 43%), and 92%(95% CI: 74%, 99%) for CCTA, and 100% (95% CI: 89%, 100%), 85% (95% CI: 75%, 92%), 71% (95% CI: 56%, 84%), and 100%(95% CI: 95%, 100%) for VIE, respectively. The AUC by ROC analysis for VIE demonstrated significant improvement in analysis of left anterior descending calcified plaques compared with CCTA (0.99 vs 0.60, P<0.001), with better performance in the left circumflex and right coronary arteries (0.98 vs 0.84 and 0.77 vs 0.77, respectively; P = 0.07 and P = 0.96, respectively). There are no significant differences between 64-, 128-, and 640-slice CCTA and VIE in terms of sensitivity, specificity, positive and negative predictive value in the diagnosis of coronary stenosis. This study shows the feasibility of using VIE for characterizing morphological features of calcified plaques, therefore, significantly improving assessment of coronary stenosis.