Subarachnoid space disease: diagnosis with fluid-attenuated inversion-recovery MR imaging and comparison with gadolinium-enhanced spin-echo MR imaging--blinded reader study.

Abstract

To evaluate fluid-attenuated inversion-recovery (FLAIR) magnetic resonance (MR) imaging in a blinded reader study for the detection of proved subarachnoid space (SAS) disease. FLAIR MR imaging was performed in 62 patients (21 with proved SAS or meningeal disease) and 41 control patients. A subset of 24 patients (eight patients with proved SAS disease and 16 control patients) also underwent gadolinium-enhanced T1-weighted MR imaging. FLAIR images were interpreted blindly and independently by two neuroradiologists. For SAS disease, the overall sensitivity, specificity, and accuracy of FLAIR for both readers were 85%, 93%, and 90%. In the 15 patients with inflammatory or neoplastic meningitis only (six patients with acute subarachnoid hemorrhage [SAH] excluded), the sensitivity, specificity, and accuracy of FLAIR for both readers were 82%, 93%, and 90%. All six acute SAH cases were interpreted as abnormal on FLAIR images by both readers. In the 24 patients who underwent both FLAIR and gadolinium-enhanced T1-weighted MR imaging, the sensitivity, specificity, and accuracy of FLAIR imaging were 86%, 91%, and 89%; the sensitivity, specificity, and accuracy of gadolinium-enhanced T1-weighted imaging were 43%, 88%, and 74%. FLAIR is highly sensitive and specific for the diagnosis of SAS disease. Unenhanced FLAIR is superior to gadolinium-enhanced T1-weighted MR imaging for the diagnosis of SAS disease. These data have important implications, because FLAIR is performed without the costs and inherent risks of intravenous contrast agents. FLAIR also appears to be highly sensitive but nonspecific for acute SAH.