Segmentation and 3D visualization of pheochromocytoma in contrast-enhanced CT images

As a kind of adrenal tumors, pheochromocytoma is commonly present with serious and potentially lethal cardiovascular complications. In this paper, a novel image segmentation and three-dimensional (3D) visualization framework is proposed to extract and visualize the pheochromocytoma and intratumoral necrosis in multiphase contrast-enhanced computed tomography (CECT) images in “Digital Imaging and Communications in Medicine” (DICOM) format. Firstly, a localized region-based level set method (LRLSM) is applied to segment regions of interest (ROIs). Then, a CT histogram analysis method is implemented and a novel tumor-necrosis volume ratio is proposed. Finally, an enhanced marching cubes (MC) algorithm is used for 3D isosurface rendering. Experimental results indicate that the segmented contours are very close to the manual delineations, and the extracted features are in accordance with clinical works. It is concluded that the important clinical statistics and intuitive 3D visualization obtained by the proposed framework can provide accurate and effective references in the diagnosis of pheochromocytoma.