The aim is to compare virtual histology which uses spectral analysis of backscattered intravascular ultrasound (VH-IVUS) and multidetector-row computed tomography (MDCT) for the characterization of coronary atherosclerotic plaques obtained by directional coronary atherectomy (DCA). We performed DCA in 15 de novo native coronary stenotic lesions (15 patients) and selected one or two segments within the plaque from each patient (total 29 segments). Then, we evaluated the accuracy of the VH-IVUS findings in 50 sites among the 29 segments compared with the histopathology findings. MDCT was performed in all patients before percutaneous coronary intervention (PCI), and CT density values were measured. VH-IVUS data analysis correlated well with histopathological examination (predictive accuracy: 66.7% for fibrous, 100% for fibro-fatty, 100% for necrotic core, and 100% for dense calcium regions, respectively). In addition, CT density values between fibrous and fibro-fatty plaques classified by histopathology were 100.0 +/- 26.0 HU versus 110.4 +/- 67.9 HU, there were no difference among them (P = 0.594). These findings indicated that the validation of plaque characteristics using VH-IVUS correlates well with histopathology. While tissue characterization using CT density could be difficult to distinguish between fibro-fatty and fibrous tissue.