Because of beam-hardening and blooming artifacts, it is difficult to determine the degree of stenosis caused by calcified coronary artery plaques at coronary CT angiography (CTA). Our goal was to determine how accurate coronary CTA is in evaluating these lesions. Thirty-one patients who had one or more calcified coronary artery plaques at coronary CTA underwent invasive coronary angiography. The size of the calcified coronary artery plaques was graded subjectively as small, moderate, or large. Using postprocessing techniques such as segmentation and tracking, we attempted to determine whether the calcified lesions were obstructive (> or = 50% diameter narrowing) or nonobstructive (< 50% diameter narrowing). Concordance with invasive coronary angiography was then determined. Calcified coronary artery plaques were graded by coronary CTA as small at 61 locations, moderate at 22 locations, and large at 43 locations. Of the 61 small calcified coronary artery plaques, 5% were obstructive at invasive coronary angiography; of the 22 moderate-sized calcified coronary artery plaques, 14% were obstructive; and of the 43 large calcified coronary artery plaques, 42% were obstructive. Concordance between coronary CTA and invasive coronary angiography occurred in 58 of 61 (95%) small calcified coronary artery plaques, 20 of 22 (91%) moderate-sized coronary artery plaques, and 29 of 43 (67%) large coronary artery plaques. Coronary CTA underestimated the degree of stenosis in one small calcified coronary artery plaque but overestimated the degree of stenosis in two small, two moderate, and 14 large coronary artery plaques. In detecting obstructive lesions caused by the 43 large calcified coronary artery plaques, coronary CTA had a sensitivity of 100%, specificity of 44%, positive predictive value of 56%, negative predictive value of 100%, and accuracy of 67%. Coronary CTA can be used to accurately predict the presence of obstructive disease in > 90% of small and moderate-sized calcified coronary artery plaques. With large calcified coronary artery plaques, CTA correctly predicts the presence of obstructive disease in approximately two thirds of the cases. When errors occur, they are usually due to overestimation of the degree of stenosis.