We determined the accuracy of angiography in the diagnosis of internal carotid ulcers by comparing the angiographic reports with the pathologic findings in 36 endarterectomy specimens. Eighteen of these specimens had microscopic ulcerations, and the observer with the highest accuracy rate read 12, of which ten were ulcerated. These results revealed a sensitivity of 56%, a specificity of 89%, and an overall accuracy of 61% for angiography. The ulcers were classified into types A, B, and C to assess the interobserver agreement rate among three readers. This resulted in a 4% interobserver agreement among a total of 75 ulcers. Because of the high interobserver disagreement and the poor correlation between angiographic and pathologic findings in the surgical specimens, we conclude that the diagnosis of carotid artery ulceration by angiography is not reliable.