Background: This paper presents a three-dimensional (3D) method for segmenting corpus callosum in normal\nsubjects and brain cancer patients with glioblastoma.\nMethods: Nineteen patients with histologically confirmed treatment na�¯ve glioblastoma and eleven normal control\nsubjects underwent DTI on a 3T scanner. Based on the information inherent in diffusion tensors, a similarity\nmeasure was proposed and used in the proposed algorithm. In this algorithm, diffusion pattern of corpus callosum\nwas used as prior information. Subsequently, corpus callosum was automatically divided into Witelson subdivisions.\nWe simulated the potential rotation of corpus callosum under tumor pressure and studied the reproducibility of\nthe proposed segmentation method in such cases.\nResults: Dice coefficients, estimated to compare automatic and manual segmentation results for Witelson\nsubdivisions, ranged from 94% to 98% for control subjects and from 81% to 95% for tumor patients, illustrating\ncloseness of automatic and manual segmentations. Studying the effect of corpus callosum rotation by different\nEuler angles showed that although segmentation results were more sensitive to azimuth and elevation than skew,\nrotations caused by brain tumors do not have major effects on the segmentation results.\nConclusions: The proposed method and similarity measure segment corpus callosum by propagating a hypersurface\ninside the structure (resulting in high sensitivity), without penetrating into neighboring fiber bundles\n(resulting in high specificity).
Loading....