From March 1992 to November 1993 we used angioscopy and arteriography for intraoperative assessment of 103 carotid endarterectomies in 96 patients. The indication for surgery was asymptomatic stenosis in 55 cases and neurologic and/or ocular symptoms in 48. Intraoperative angioscopy and arteriography were performed to allow comparison of findings. Intraoperative angioscopic images were normal in 67 cases and abnormal in 36. The defect was an intimal flap in 26 cases, detachment of the distal plaque in seven cases, and an intimal wedge in five cases. In two cases both detachment and a wedge were observed. The defect was not considered severe enough to warrant revision in 31 cases and was corrected in five cases by either vein bypass (n = 1) or revision of the endarterectomy (n = 4). In the latter four cases repeat angioscopy showed normal findings. Arteriographic and angioscopic findings were compared in 102 cases. In the 71 cases in which angioscopic findings were normal, arteriography revealed a major abnormality in three cases: kinking in one and stenosis > 40% in two. Kinking was treated by attachment of the common carotid artery and stenosis by venous bypass. In the 31 cases in which angioscopy revealed defects not considered to warrant revision, arteriography revealed stenosis > 40% in three cases treated by either prosthetic bypass (n = 2) or revision of the endarterectomy (n = 1). The false negative rate for angioscopy was 5.9% and concordance between the two methods was 94.1%. The combined mortality-morbidity rate was 1.9% (one stroke and one death). Postoperative evaluation of anatomic findings by arteriography or Doppler ultrasonography revealed asymptomatic internal carotid occlusion in one and internal carotid stenosis < 30% in four cases. Angioscopy is a simple, low-cost method in intraoperative control that can be used either as an adjunct to arteriography or as an alternative if arteriography cannot be performed.