20120205SPIEPoster.pdf

Cerebral arteriovenous malformations (AVMs) are a type of vascular anomaly consisting of large intertwined vascular growth (the nidus) that are prone to serious hemorrhaging and can result in patient death if left untreated. Intervention through surgical clipping of feeding and draining vessels to the nidus is a common treatment. However, identification of which vessels to clip is challenging even to experienced surgeons aided by conventional image guidance systems. In this work, we describe our methods for processing static preoperative angiographic images in order to effectively visualize the feeding and draining vessels of an AVM nidus. This is one by processing the vessel images with level-set front propagation and using the resulting maps to label the vessels by colour and decluttering the image by vessel topological distances. In order to aid the surgeon in the vessel clipping decision-making process during surgery, the results are displayed to the surgeon using augmented virtuality.