Segmentation and Separation of Cerebral Aneurysms : A Multi-Phase Approach

The purpose of this study is to develop a novel segmentation approach for saccular aneurysm extraction using computed tomography angiography (CTA). Manual segmentation is a tedious procedure, not precise and prone to inter-and intra-observer variability. Therefore, there is a need for an automated method, which is accurate, reproducible and reliable. The approach consists of four stages. At the first stage, the cerebral arteries are segmented and presented to the user to select a seed point within the aneurysm. The seed point is used at the second stage, to initially segment the aneurysm and its feeding vessels by using a 3D conditional morphology segmentation method. This result can be inspected by a neurosurgeon and radiologist to assess the status of the aneurysm with respect to the parent artery and to consider options for treatment. At the third stage, the coarse segmented aneurysm and feeding vessels are re-segmented via 3D edge based level set method to have an exact and fine segmentation. Finally, the aneurysm is separated from the parent artery based on the conditional morphology like first stage but with the optimum iteration value. The proposed method has been evaluated in 15 CTA data sets. A very good agreement with the visual judgment was obtained in both normal and aneurysmal CT scans. These CTA images (with pixel spacing [0.4, 0.4] mm and 0.6 mm slice thickness) were collected at Lausanne University (Switzerland) and St George's Hospital, London. In this study the novel hybrid region growing and level set techniques has been used for the segmentation of the brain aneurysm. The technique provides promising results for cerebral artery tree segmentation followed by aneurysm segmentation from CT angiography.