Accuracy of susceptibility-weighted imaging for the detection of arteriovenous shunting in vascular malformations of the brain.

BACKGROUND AND PURPOSE to determine the accuracy of susceptibility-weighted MRI (SWI) for the detection of arteriovenous shunting (AVS) in vascular malformations of the brain (BVM). METHODS we retrospectively identified 60 patients who had been evaluated for known or suspected BVM by both SWI and digital subtraction angiography, without intervening treatment, during a 3-year period. SWI images were retrospectively assessed by 2 independent reviewers for the presence of AVS as determined by the presence of signal hyperintensity within a venous structure in the vicinity of the BVM. Discrepancies were resolved by consensus among a panel of 3 neuroradiologists. Accuracy parameters of SWI for the detection of AVS were calculated using digital subtraction angiography as the reference standard. RESULTS a total of 80 BVM were identified in the 60 patients included in our study. Of the 29 BVM with AVS on digital subtraction angiography, 14 were untreated arteriovenous malformations, 10 were previously treated arteriovenous malformations, and 5 were untreated dural arteriovenous fistulas. Overall, SWI was 93% sensitive and 98% specific for the detection of AVS in BVM, with excellent interobserver agreement (κ=0.94). In the 14 previously treated arteriovenous malformations, SWI was 100% sensitive and specific for the detection of AVS. In the 28 BVM associated with intracerebral hemorrhage, SWI was 100% sensitive and 96% specific for the detection of AVS. CONCLUSIONS SWI is accurate for the detection of arteriovenous shunting in vascular malformations of the brain and, for some patients, SWI may offer a noninvasive alternative to angiography in screening for or follow-up of treated BVM.