Blinded prospective evaluation of sensitivity of MR angiography to known intracranial aneurysms: importance of aneurysm size.

PURPOSE: To determine the sensitivity of time-of-flight and phase-contrast MR angiography for the detection of intracranial aneurysms. METHODS: Sixteen patients with 27 intracranial aneurysms previously identified with conventional angiography and 19 control patients were examined with three-dimensional time-of-flight, three-dimensional phase-contrast MR angiography, and standard MR imaging. Subvolumes of the carotid and posterior circulations, source images, and standard MR images were blindly interpreted by three experienced neuroradiologists. RESULTS: Detection of an aneurysm by a given sequence was defined as at least two of the three blinded readers identifying the aneurysm. The sensitivities of the sequences based on all 27 aneurysms were: transaxial T1, 25.9%; T2, 48.1%; PC, 44.4%; and TF, 55.6%. Two of 3 aneurysms detected with T2 but not MR angiography had adjacent blood products. Five millimeters appeared to be a critical size; the sensitivities for aneurysms greater than or equal to 5 mm were: T1, 37.5%; T2, 62.5%; PC, 75%; and TF, 87.5%. CONCLUSIONS: Three-dimensional time-of-flight MR with 512 x 256 matrix is more sensitive than three-dimensional phase-contrast or standard MR imaging for detection of aneurysms. Retrospectively, aneurysms 3 mm or larger can be identified with MR angiography; however, prospectively, 5 mm is the critical size for detection.