Background: 64-slice computed tomography (CT) is highly accurate to diagnose coronary artery occlusive disease. The purpose of this study is to determine the accuracy of 64-slice CT for detecting significant stenoses as a diagnostic step in deciding to perform conventional coronary angiography (CCA). Methods:Coronary CT angiography had been being performed in patients with atypical chest pain or stable or unstable angina by using 64-slice CT (Siemens Sensation 64, Germany) since May, 2005. CCA was performed in 580 patients who had been diagnosed to have significant coronary artery disease (n=457) by CT and who had no significant coronary artery disease (n=123) until Jan, 2007. We excluded the patients who had had previous myocardial infarction or had undergone percutaneous coronary intervention or coronary artery bypass surgery, Beta-blocker was used to lower heart rate if heart rate > 65bpm. Analysis was retrospectively done with CT and CCA reports on 15 coronary artery segments according to AHA definition. Diagnostic accuracy of CT for detecting significant stenoses compared to CCA was determined on a per-segment and per-patient basis. Results:Heart rate was 56.7짹5.7 bpm (mean짹STD) during CT scanning. We included all CT scans and coronary artery segments in analysis. Sensitivity, specificity, and positive and negative predictive values were 49.9%, 95.4%, 62.6%, 92.5% on a per-segment basis and 91.3%, 59.0%, 87.0%, 69.1% on a per-patient basis, respectively. The low sensitivity and positive predictive value on a per-segment based analysis is due to high false-negative rates in small branches and significant distal or branch lesions neglected by readers when significant lesions existed in the proximal segments. Low negative predictive value on a per-patient based analysis was because most patients with negative CT did not undergo CCA. Conclusion:64-slice CT has relatively high sensitivity and positive predictive value in detecting significant stenoses on a per-patient basis in everyday practice and it is useful in selecting patients who need invasive coronary angiography.