Improving prostate biopsy protocol with a computer aided detection tool based on semi-supervised learning

Prostate cancer is one of the most frequently diagnosed neoplasy and its presence can only be con¯rmed by biopsy. Due to the high number of false positives, Computer Aided Detection (CAD) systems can be used to reduce the number of cores requested for an accurate diagnosis. This work proposes a CAD procedure for cancer detection in Ultrasound images based on a learning scheme which exploits a novel semi-supervised learning (SSL) algorithm for reducing data collection export and avoiding collected data wasting. The ground truth database comprises the RF signals acquired during biopsies and the corresponding tissue samples histopathological outcome. A comparison to a state-of-art CAD scheme based on supervised learning demonstrates the effectiveness of the proposed SSL procedure at enhancing CAD performance. Experiments on ground truth images from biopsy findings show that the proposed CAD scheme is effective at improving the efficiency of the biopsy protocol.