Segmentation of macroscopic slice images for colorectal cancer evaluation

Colorectal cancer is the third most common cancer. Depending on the extent of the disease, radiotherapy and surgery are currently the only curative therapies. When surgery is performed, the shortest distance from affected regions to the resection boundary is an important predictive factor for prognosis of the disease. This article will present segmentation of the rectal wall in macroscopic slices from resected specimens. The algorithm uses texture information for characterizing different regions in the macroscopic slices. Segmentation is achieved using Level Sets. The segmentation results provide landmarks for volume reconstruction and information for comparison between pathology and MRI.