The purpose of this study was to determine the technical capacity and diagnostic accuracy of 3D time-of-flight magnetic resonance angiography (MRA) in suspected cerebral vasculitis in a retrospective analysis of MRA and digital subtraction angiography (DSA) in 14 young patients with clinical and/or radiological suspicion of cerebral vasculitis. A total of nine arteries were evaluated in each patient. Consensus review of DSA by three observers was the reference standard. The sensitivity for detecting a stenosis varied from 62 to 79% for MRA and from 76 to 94% for DSA, depending on the observer. The specificity for detecting a stenosis varied from 83 to 87% for MRA and from 83 to 97% for DSA. Using the criterion "more than two stenoses in at least two separate vascular distributions" to consider the examination as being true positive, the false-positive rates for MRA and DSA were comparable. MRA plays a role as the first angiographical examination in the diagnostic work-up of suspected cerebral vasculitis. When more than two stenoses in at least two separate vascular distributions are depicted on MRA, DSA is not expected to add a significant diagnostic contribution in a patient with suspected cerebral vasculitis. DSA remains necessary when MRA is normal or when less than three stenoses are seen.