Polyp Segmentation on CT Colonography

Colorectal cancer is the third most common cancer and the second leading cause of cancer death in Americans. It is estimated that 50% of people over the age of 60 will have at least one polyp. Computed tomographic colonography (CTC) is a less invasive technique to examine the colon compared to the traditional optical colonoscopy (OC). Large-scale multicenter clinical trials conducted in several countries have demonstrated that CTC is an effective modality for colorectal cancer screening. During the past decade, computer-aided detection and diagnosis (CADe and CADx) for CTC has been an active area of investigation and development in academia and industry. Most CADe and CADx systems for CTC comprise the following modules: colon segmentation, polyp segmentation, and feature calculation. Among these modules, polyp segmentation and characterization are essential for CADe and CADx. This chapter reviews several state-of-the-art polyp segmentation techniques and details an advanced approach based on a combination of knowledge-guided intensity adjustment, fuzzy clustering, and adaptive deformable models. The chapter also presents advanced techniques for polyp feature characterization. The results were validated against manual measurement and segmentation on OC and CTC.