Achieving robust labeling above the circle of Willis with vessel-encoded arterial spin labeling.

PURPOSE: To improve the robustness of noninvasive vessel-selective perfusion imaging and angiography using vessel-encoded arterial spin labeling (VEASL) when applied to complex vascular geometries, such as above the circle of Willis (CoW) in the brain. METHODS: Our proposed improved optimized encoding scheme (IOES) better accounts for vascular geometry and the VEASL encoding process, leading to more SNR-efficient encodings than previous approaches. Pseudo-continuous arterial spin labeling (PCASL) parameters were optimized for a thinner labeling region, allowing tortuous vessels to be more accurately treated as single points within the labeling plane. Our optimized approach was compared to the original OES method above the CoW in healthy volunteers, with preliminary application in two Moyamoya patients. RESULTS: In simulation, the IOES improved SNR efficiency by approximately 10% and used longer wavelength encodings that are less sensitive to subject motion. The effective labeling thickness was reduced using optimized PCASL parameters, which maintained high labeling efficiency. In healthy volunteers, these improvements allowed for the separation of at least nine arteries and their downstream tissues, with more accurate vessel decoding and closer alignment between the measured VEASL signal modulation and the encoding design. Vascular territories consistent with angiography were found in the Moyamoya patients. CONCLUSIONS: Combining IOES with optimized PCASL parameters, the vessel-decoding efficacy in a region with complex vascular geometry above the CoW was improved. The automated encoding design process and scan times under 6 min make it feasible to observe flow patterns above the CoW in clinical settings, particularly for studies of collateral circulation.