Quantitative Analysis of the Cerebral Vasculature on Magnetic Resonance Angiography

The arterial connections in the Circle of Willis are a central source of collateral blood flow and play an important role in pathologies such as stroke and mental illness. Analysis of the Circle of Willis and its variants can shed light on optimal methods of diagnosis, treatment planning, surgery, and quantification of outcomes. We developed an automated, standardized, objective, and high-throughput approach for categorizing and quantifying the Circle of Willis vascular anatomy using magnetic resonance angiography images. This automated algorithm for processing of MRA images isolates and automatically identifies key features of the cerebral vasculature such as branching of the internal intracranial internal carotid artery and the basilar artery. Subsequently, physical features of the segments of the anterior cerebral artery were acquired on a sample and intra-patient comparisons were made. We demonstrate the feasibility of using our approach to automatically classify important structures of the Circle of Willis and extract biomarkers from cerebrovasculature. Automated image analysis can provide clinically-relevant vascular features such as aplastic arteries, stenosis, aneurysms, and vessel caliper for endovascular procedures. The developed algorithm could facilitate clinical studies by supporting high-throughput automated analysis of the cerebral vasculature.