Improved Estimation of Cerebral Artery Branch Territories Using Cluster-Based Segmentation of Vessel-Encoded Pseudocontinuous ASL Data

Akash P. Kansagra¹, Eric C. Wong²

Vessel encoded pseudo-continuous ASL allows non-invasive estimation of multiple cerebrovascular territories above the circle of Willis. Segmentation of vascular territories has traditionally been achieved by clustering of highly correlated tagging data. Here, we assess improvements in vascular territory estimation that occur when tagging efficiency data are supplemented with spatial data in the clustering scheme. These results offer a simple and straightforward means to enhance the accuracy of mapping of small cerebral artery branch territories.