Segmentation and modeling of the blood vessels in the neck area from patient Computed Tomography Angiogram (CTA) images is of increasing clinical importance. The models are used for the diagnosis of stenosis, aneurisms, and for planning catheter endovascular surgeries. We have developed a nearly automatic method for patient-specific modeling of the aorta and the carotid, vertebral, and sub-clavian arteries for patient-specific simulations before the surgery. We use the graph cut method to automatically generate a vessels segmentation which is then refined with an smart interactive editing tool to produce a geometric mesh. Our results on over 60 clinical CTA datasets show that our method is accurate, robust, and easy to use. The method is being integrated into the ANGIO Mentor, a training simulator for endovascular procedures developed by Simbionix Ltd, Israel. Surgeons can train on the simulator before the catheterization to significantly reduce their learning curve, shorten the procedure time, and improve their performace.

+ Joint work with: 1. Jacob Sosna MD, Phillip Berman, Dept of Radiology, Hadassah University Hospital 2. Ofek Shilon and Einav Nammer, Simbionix Ltd, Lod. 3. Moti Freiman, Noah Broide, Miriam Natanzon, Judith Frank, Lior Weizman School of Engineering and Computer Science, The Hebrew University of Jerusalem

This project is funded by a MAGNETON grant from the Ministry of Trade and Industry, Israel.