Noninvasive coronary angiography with multislice computed tomography (CT) scanners is feasible with high sensitivity and negative predictive value. The radiation exposure associated with this technique, however, is high and concerns in the widespread use of CT have arisen. We evaluated the diagnostic accuracy of coronary angiography using 320-row CT, which avoids exposure-intensive overscanning and overranging. We prospectively studied 118 unselected consecutive patients with suspected coronary artery disease (CAD) referred for invasive coronary angiography (ICA). All patients had 320-row CT within 1 week of ICA, which, together with quantitative analysis, served as the reference standard. Of the 65 out of 118 patients who were diagnosed as having CAD by ICA, 64 (98 %) were correctly identified at 320-row CT. Noteworthy, 320-row CT correctly detected CAD in 3 patients with atrial fibrillation and ruled out the disease in the other 8 patients. From 151 significant coronary stenoses detected on ICA, 137 (91 %) were correctly identified with 320-row CT. In the per-patient analysis, sensitivity and specificity of 320-row CT were 98 and 91 %, respectively. In the per-vessel analysis, sensitivity and specificity of 320-row CT were 93 and 95 %, respectively. In the per segment analysis, sensitivity and specificity of 320-row CT were 91 and 99 %, respectively. Diameter stenosis determined with the use of CT showed good correlation with ICA (P < 0.001, R = 0.81) without significant underestimation or overestimation (-3.1 ± 24.4 %; P = 0.08). Comparison of CT with ICA revealed a significantly smaller effective radiation dose (3.1 ± 2.3 vs. 6.5 ± 4.2 mSv; P < 0.05) and amount of contrast agent required (99 ± 51 vs. 65 ± 42 ml, P < 0.05) for 320 row CT. The present study in an unselected population including patients with atrial fibrillation demonstrates that 320-row CT may significantly reduce the radiation dose and amount of contrast agent required compared with ICA while maintaining a very high diagnostic accuracy.