Objective: To establish in which clinical contexts cerebral MR angiography (MRA) is routinely carried out in a neurological university department and to describe its clinical impact. Material and Methods: Medical records, reports of findings and documentation of imaging examinations carried out in all 69 patients referred to the Department of Radiology from the Department of Neurology between 1995 and 1998 for cerebral MRA were evaluated. The clinical impact of all imaging findings was assessed on the basis of the medical records. Results: Circulatory disturbances in the vertebrobasilor arteries (n=4) were the most frequent indication for investigation. MRA followed CT or duplex sonography in 66 of the 69 patients with a mean delay of 8 days. MRA was considered diagnostically inferior to conventional MR in 11 cases, comparable in 30 and superior in 25. Comparing MRA and duplex sonography, the corresponding figures were 12, 29 and 23. In retrospect, 56 MRAs were judged unnecessary. Conclusion: Controlled clinical studies on optimal use of MRA are needed to avoid wasting resources and to exploit the method's full diagnostic potential in appropriate cases.