Modeling and mapping of prostate cancer

Abstract Prostate cancer is one of the most common cancers for men. Biopsy of the prostate is carried out following protocols that designate locations and number of needles for tissue sampling as prostate cancer mostly does not show up in the ultrasound images or other modalities. Current protocols for prostate biopsy, however, were created largely based on qualitative clinical experiences and pathological data. They are thus sub-optimal in terms of rate of cancer detection. In fact, the most commonly used protocol (the sextant biopsy) has only about a 30% rate of cancer detection. This leads to more frequent repeat biopsies and potentially runs the risk of not detecting a significant cancer in early stage. In this paper, we aim at improving the performance of prostate needle biopsy by modeling and mapping prostate cancer using real prostate specimens. Surface models of the prostate are reconstructed using deformable modeling techniques. A 3D visualization system is developed to simulate the process of prostate needle biopsy. A 3D distribution map of prostate cancer is built and used to develop optimal biopsy protocols. The 3D map can later be superimposed to online ultrasound images to guide biopsy process of real patients. Experimental results are also provided.