A prospective blinded comparison of two-dimensional (2D) time-of-flight (TF) magnetic resonance (MR) angiography and color duplex flow ultrasound (US) with conventional angiography as a standard of reference was performed in 50 patients with hemispheric ischemic symptoms. The guidelines of the North American Symptomatic Carotid Endarterectomy Trial for measuring stenosis of the internal carotid artery were utilized for maximum-intensity-projection (MIP) images and conventional angiograms. While the 2D TF MIP technique overrepresented the degree of stenosis in the internal carotid arteries, it matched the performance of color duplex flow US as measured by means of receiver operating characteristic curves. A signal void on the MIP images corresponded to a 70% or greater internal carotid artery stenosis in 17 of 20 arteries. At its current state of development, the accuracy of 2D TF MR angiography equals that of US in characterizing the degree of carotid stenosis but cannot be considered a replacement for conventional angiography.