Background: In several studies with 16-slice or 64-slice MDCT, the sensitivity and specificity for diagnosis of significant left main disease were 100%, 100%, respectively. However frequency of left main disease in those studies was ranged from 1 to 6. We sought to determine the accuracy of 64-slice CT for diagnosis of left main disease in real world. Methods: We enrolled 579 patients with symptoms of coronary artery disease, who underwent coronary CT angiography with 64-MDCT and conventional coronary angiography (CCA). Significant stenosis was defined as diameter narrowing more than 50%. Left main disease was defined as disease involving proximal vessel to the line perpendicular to the longitudinal axis of proximal LAD or proximal LCx, which meets bifurcation point (the proximal point where proximal LAD and proximal LCx meet together). Accuracy for diagnosis of left main disease was calculated. Results: The mean age was 62 years (range 21-89). Men were 342, women 237. Fourty-one had significant left main disease in CCA. Twenty-six of them were correctly diagnosed with MDCT (Sensitivity 63.4%). 529 patients without significant left main disease were correctly diagnosed with MDCT (specificity 91.3%). Positive and negative predictive values were 74.3%, 97.2%, respectively. In patients with left main disease in CCA, 12 patients underwent coronary artery bypass surgery (CABG) and 12 patients got stenting in left main artery. MDCT diagnosed 10 patients with CABG and 7 patients with stenting in left main artery as having left main disease. Conclusion: The accuracy of 64-MDCT in diagnosis of left main disease was lower than previously reported. It could be explained by the different view of bifurcation lesion between CCA and MDCT. Further study is warranted to help to plan treatment for left main disease with MDCT.