Yan Wang1, Florent Seguro1, Farshid Faraji1, Chengcheng Zhu1, Henrik Haraldsson1, Michael Hope1, Jing Liu1, and David Saloner2
1Radiology, UCSF, San Francisco, CA, United States, 2Radiology, UCSF/VA, San Francisco, CA, United States
An abdominal aortic aneurysm (AAA) may rupture if left untreated. Automatic segmentation methods allow quick and reproducible AAA morphology quantification, which is favorable for the diagnosis of AAA disease. This study developed a novel registration based geometric active contour model to segment vessel wall of AAA from 3D black blood MRI. This method was initially tested on six patients. The proposed method agreed well with manual segmentation with an average Dice value of 91.22%, demonstrating good segmentation accuracy. These methods can be potentially used for AAA evaluation in clinical setting.
