The usefulness of 64-multidetector row computed tomography (MDCT) for the evaluation of aortic atherosclerotic disease (AAD), a potential source of cerebral emboli, has recently been suggested. We aimed to assess the significance of AAD by using MDCT in patients who are suspected to have had embolic stroke clinically or radiologically but without evident cardioembolic source (possibly embolic stroke, or PES). From 5/2007 to 10/2007, patients who were presented with acute ischemic stroke, met predefined criteria for PES, and underwent MDCT, were found in a prospective stroke registry (PES group). High-risk AAD was defined as thrombus, ulceration, and a >or=4 mm thickness of atherosclerotic plaque in the ascending aorta and aortic arch on MDCT. For comparison, patients who were hospitalized due to non-embolic stroke (NES) during the study period and who underwent MDCT for the purpose of screening of coronary artery disease were selected (NES group). Among a total of 336 patients with acute ischemic stroke, 57 (20.0%) satisfied the criteria for PES, and MDCT was performed in 50 of these 57 patients. One-hundred six patients were selected as the NES group. The PES group had significantly higher prevalence of high-risk AAD than the NES group did (38.0 vs. 13.2%, P < 0.01). The odds ratio of high-risk AAD was 4.03 (95% confidence interval, 1.81-8.98) and this result remained significant after adjustment for risk factors of atherosclerosis. This study suggests the etiologic role of aortic atherosclerosis detected by MDCT in patients who are assumed to have had embolic stroke but without evident embolic source.