OBJECTIVES: A precise assessment of angioarchitectural characteristics using noninvasive imaging is helpful for serial follow-up and weighting risk of natural history in unruptured brain arteriovenous malformation (bAVM). This study aimed to test the hypothesis that susceptibility weighted imaging (SWI) would provide an accurate evaluation of angioarchitectural features of unruptured bAVM. METHODS: A total of 81 consecutive patients with unruptured bAVM were examined. Image quality of SWI for the assessment of bAVM angioarchitectural features was determined by a 5-point scale. The accuracy of SWI for detection of angioarchitectural features was evaluated using digital subtraction angiography as a standard reference and further compared among unruptured bAVMs with or without silent intralesional microhemorrhage on SWI to examine the potential confounding effect of microhemorrhage on image analysis. RESULTS: All lesions were identified on SWI. Image quality of SWI was judged to be at least adequate for diagnosis (range, 3-5) in all patients by both readers. Using digital subtraction angiography as a reference standard, the area under the receiver operating curve of detection of deep or posterior fossa location, exclusively deep venous drainage, venous ectasia, venous varices, and the presence of associated aneurysm on SWI was 1, 0.93, 0.94, 0.95, and 0.83, respectively. Silent intralesional microhemorrhage were detected in 39 patients (48.15%) on SWI and no significant difference (P > 0.05) was found in angioarchitectural features between patients with and without silent microhemorrhage. CONCLUSIONS: SWI might he a noninvasive alternative technique for angiography in the angioarchitectural assessment of unruptured bAVM.