Radiological diagnosis of carotid ulceration.

The standard radiological reports of 50 carotid angiograms were reviewed with regard to the presence of ulceration at the carotid bifurcation. The angiograms were then specifically re-examined retrospectively by a radiologist and the presence or absence of ulceration commented on. The macroscopic features of the carotid endarterectomy specimens of the same patients undergoing carotid endarterectomy were then examined by an independent observer. The results of these three examinations were then compared. There were 28 ulcers in 23 of the 50 specimens with 12 large ulcers (greater than three mm diameter) and 16 small ulcers (less than three mm). In the diagnosis of ulcers, the initial radiological report showed a sensitivity of 32% and a specificity of 97.2%, with an accuracy of 79%, whilst on the retrospective search for ulcers the sensitivity rose to 67.8% whilst the specificity fell to 55.5%, with an accuracy of 59%. The factors affecting the accuracy of the detection of carotid ulcers by angiography are discussed. Angiography in its present form has limitations in diagnosing carotid ulceration.