LUMEN SEGMENTATION OF ATHEROSCLEROTIC CAROTID ARTERIES IN CTA

Lumen segmentation is relevant for quantification of the de-gree of stenoses and as preprocessing step for the analysis of the vessel wall, and for the simulation and planing of inter-ventions. Although lumen segmentation in CTA has received considerable interest, it remains a challenge for atheroscle-rotic vessels. In this paper we propose a semi-automatic lu-men segmentation method using a levelset approach, which utilizes both boundary information and regional intensity in-formation to accurately segment the lumen of atherosclerotic vessels. The method was trained and evaluated on a publicly available database of 56 carotid arteries. The average Dice similarity coefficient was 88.9 %, the mean absolute surface distance was 0.31 mm. Compared to using only boundary in-formation, this method yields better segmentation in severely atheroslerotic vessels. Index Terms — lumen segmentation, geodesic active con-tour, gradient, plaque, calcification, regional information 1