Biopsy-guided learning with deep convolutional neural networks for Prostate Cancer detection on multiparametric MRI

Yohannes Tsehay, Nathan Lay, Xiaosong Wang, Jin Tae Kwak, Baris Turkbey, Peter L. Choyke, Peter A. Pinto, Bradford J. Wood, Ronald M. Summers

Abstract: Prostate Cancer (PCa) is highly prevalent and is the second most common cause of cancer-related deaths in men. Multiparametric MRI (mpMRI) is robust in detecting PCa. We developed a weakly supervised computer-aided detection (CAD) system that uses biopsy points to learn to identify PCa on mpMRI. Our CAD system, which is based on a deep convolutional neural network architecture, yielded an area under the curve (AUC) of 0.903±0.009 on a receiver operation characteristic (ROC) curve computed on 10 different models in a 10 fold cross-validation. 9 of the 10 ROCs were statistically significantly different from a competing support vector machine based CAD, which yielded a 0.86 AUC when tested on the same dataset (α = 0.05). Furthermore, our CAD system proved to be more robust in detecting high-grade transition zone lesions.