Towards a Quantitative Assessment of Dural Ectasia in Patients with Marfan Syndrome

摘要

In this paper an automatic segmentation algorithm and a surface rendering technique for volumetric analysis of dural ectasia are presented. The segmentation algorithm, based on an unsupervised Fuzzy C-Means method, was applied on T2-weighed Magnetic Resonance images of 6 patients affected by Marfan Syndrome. The segmented images were then rendered using a marching cubes algorithm, a powerful technique producing very high-quality images. The proposed methodology is a preliminary step towards a morphological characterization of the pathological dilatation of the dura related to this particular autosomal disorder. The fidelity of the generated 3D reconstructions comes close to anatomical reality and provides the clinician with a tool for a panoramic view of the structures and a means for a more accurate diagnosis of ectasia.