ABSTRACT
Prostate cancer is the second most commonly diagnosed malignancy in men, and is the sixth most common cause of cancer-related male death worldwide. Early diagnosis of prostate cancer enhances the success rate of survival for male patients. As a result, it is critical to identify and create methods to facilitate early diagnosis of prostate cancer, especially noninvasive techniques, such as medical imaging and computer-aided diagnostic (CAD) systems. In this chapter, a novel noninvasive framework for the early diagnosis of prostate cancer from diffusion-weighted magnetic resonance imaging (DW-MRI) is proposed. The proposed framework consists of three main steps. In the first step, the prostate is localized and segmented based on a level-set model. In the second step, the apparent diffusion coefficient (ADC) of the segmented prostate volume is mathematically calculated for different b-values. To preserve continuity, the calculated ADC values are normalized and refined using a generalized Gauss-Markov random field (GGMRF) model. The cumulative distribution function (CDF) of refined ADC for the prostate tissues at different b-values are then constructed. These CDFs are considered as global features describing water diffusion, which can be used to distinguish between benign and malignant tumors. Finally, a deep learning network, namely a stacked nonnegativity constraint autoencoder (SNCAE), is used to classify the prostate as benign or malignant based on the extracted CDFs from the previous step. Preliminary experiments on 53 clinical DW-MRI datasets showed the high accuracy of the proposed CAD system.
