Accuracy of Vessel-Encoded Pseudocontinuous Arterial Spin-Labeling in Identification of Feeding Arteries in Patients with Intracranial Arteriovenous Malformations

BACKGROUND AND PURPOSE: Identifying feeding arteries of intracranial AVMs is very important for preoperative evaluation. DSA remains the reference standard for diagnosis but is invasive. Our aim was to evaluate the diagnostic accuracy of vessel-encoded pseudocontinuous arterial spin-labeling in identifying feeding arteries of intracranial AVMs by using DSA as the criterion standard. MATERIALS AND METHODS: Eighteen patients with AVMs were examined with vessel-encoded pseudocontinuous arterial spin-labeling and DSA. Three postlabeling delays (postlabeling delay = 1, 1.3, and 1.6 seconds) were applied in 6 patients, and a single postlabeling delay (1 second) was applied in the remainder. Perfusion-weighted images were decoded into individual vascular territories with standard and relative tagging efficiencies, respectively. The supply fraction of each feeding artery to the AVM was calculated. The within-subject ANOVA was applied to compare supply fractions acquired across 3 postlabeling delays. Receiver operating characteristic analysis curves were calculated to evaluate the diagnostic accuracy of vessel-encoded pseudocontinuous arterial spin-labeling for identifying the feeding arteries of AVMs. RESULTS: There were no significant differences in supply fractions of the 3 major arteries to AVMs acquired with 3 postlabeling delays (P > .05). For vessel-encoded pseudocontinuous arterial spin-labeling with standard labeling efficiencies, the area under the receiver operating characteristic analysis curve was 0.942. The optimal cutoff of the supply fraction for identifying feeding arteries was 15.17%, and the resulting sensitivity and specificity were 84.62% and 93.33%, respectively. For vessel-encoded pseudocontinuous arterial spin-labeling with relative labeling efficiencies, the area under the receiver operating characteristic analysis curve was 0.957. The optimal cutoff of the supply fraction was 11.73%, which yielded an 89.74% sensitivity and 93.33% specificity. CONCLUSIONS: The contribution fraction of each feeding artery of the AVM can be reliably estimated by using vessel-encoded pseudocontinuous arterial spin-labeling. Vessel-encoded pseudocontinuous arterial spin-labeling with either standard or relative labeling efficiencies offers a high level of diagnostic accuracy compared with DSA for identifying feeding arteries.