Background: Intravascular ultrasound (IVUS) can detect luminal dimension relatively accurately but IVUS is more invasive, expensive and laborious than multidetector computed tomography(MDCT). However accuracy of MDCT is not proved, so we compared the quantitative analysis of luminal dimension by IVUS and MDCT and detection of significant luminal stenosis on IVUS and MDCT. Methods: Twenty patients with intermediate stenosis by MDCT underwent IVUS assessment of the culprit artery. The results from IVUS and MDCT were compared using the two-tailed unpaired t-test. We defined significant luminal stenosis as a minimum luminal area (MLA) ��� 4 mm2. RESULTS: There was a poor correlation between IVUS and MDCT with respect to minimum luminal diameter(correlation 0.21, p=0.37) and minimal luminal area(correlation 0.18, p=0.449). In contrast, there was a significant difference in the assessment of distal reference segment luminal diameter(2.63짹0.65 mm by IVUS versus 2.88짹0.41 mm by MDCT, p=0.028) and distal reference segment luminal area(6.70짹1.93 mm2 by IVU versus 5.70짹2.64 mm2 by MDCT, p=0.028). A total of twenty stenotic coronary lesions were classified as intermediate by QCA; 85% of these were found to be significant luminal stenosis on IVUS vs 70% of these were found to be significant luminal stenosis on MDCT. Conclusions: The luminal dimension detected by MDCT is higher than that of IVUS and the frequency of severe stenosis detected by MDCT was low. In conclusion, MDCT underestimated the severity of stenosis.