PURPOSE: To compare the findings of fluid-attenuated inversion recovery(FLAIR) MR imaging in the detection of subarachnoid hemorrhage(SAH), with those of precontrast CT and T1-weighted MR imaging. MATERIALS AND METHODS: In 13 patients (14 cases) with SAH, FLAIR MR images were retrospectively analyzed and compared with CT(10 patients, 11 cases) and T1-weighted MR images(9 cases). SAH was confirmed on the basis of high density along the subarachnoid space,as seen on precontrast CT, or lumbar puncture. MR imaging was performed on a 1.0T unit. FLAIR MR and CT images were obtained during the acute stage(less than 3 days after ictus) in 10 and 9 cases, respectively, during the subacute stage(4 -14 days after ictus) in two cases and one, respectively, and during the chronic stage(more than 15 days after ictus) in two cases and one, respectively. CT was performed before FLAIR MR imaging, and the interval between CT and FLAIR ranged from 24hours(6 cases) to 2 -3 (2 cases) or 4 -7 days(3 cases). In each study, the conspicuity of visualization of SAH was graded as excellent, good, fair, or negative at five locations(sylvian fissure, cortical sulci, anterior basal cistern, posterior basal cistern, and perimesencephalic cistern). RESULTS: In all cases, subarachnoid hemorrhages were demonstrated as high signal intensity areas on FLAIR images. The detection rates for SAH on CT and T1-weighted MR images were 100 %(11/11) and 89%(8/9), re-spectively. FLAIR was superior to T1-weighted imaging in the detection of SAH at all sites except the anterior basal cistern (p<0.05) and superior to CT in the detection of SAH at the cortical sulci(p<0.05). CONCLUSION: On FLAIR MR images, subarachnoid hemorrhages at all stages are demonstrated as high signal intensity areas; the FLAIR MR sequence is thus considered useful in the detection of SAH. In particular, FLAIR is more sensitive than CT for the detection of SAH in the cortical sulci.