Prostate Cancer Detection using Deep Convolutional Neural Networks

Abstract

Prostate cancer is one of the most common forms of cancer and the third leading cause of cancer death in North America. As an integrated part of computer-aided detection (CAD) tools, diffusion-weighted magnetic resonance imaging (DWI) has been intensively studied for accurate detection of prostate cancer. With deep convolutional neural networks (CNNs) significant success in computer vision tasks such as object detection and segmentation, different CNN architectures are increasingly investigated in medical imaging research community as promising solutions for designing more accurate CAD tools for cancer detection. In this work, we developed and implemented an automated CNN-based pipeline for detection of clinically significant prostate cancer (PCa) for a given axial DWI image and for each patient. DWI images of 427 patients were used as the dataset, which contained 175 patients with PCa and 252 patients without PCa. To measure the performance of the proposed pipeline, a test set of 108 (out of 427) patients were set aside and not used in the training phase. The proposed pipeline achieved area under the receiver operating characteristic curve (AUC) of 0.87 (95[Formula: see text] Confidence Interval (CI): 0.84-0.90) and 0.84 (95[Formula: see text] CI: 0.76-0.91) at slice level and patient level, respectively.

Figure 1

Block diagram of the proposed pipeline for prostate cancer detection. The inputs to each CNN are 66 $\times$ 66 $\times$ 6 (ADC, b0, b100, b400, b1000, b1600) MRI slices. The output is the slice level and patient level results.

Figure 2

The structural difference between original residual network and fully pre-activated residual network.

Figure 3

Block diagram of the proposed first-order statistical feature extractor. PCa Set: probabilistic output set from each CNN which is associated with PCa class. Non PCa Set: probabilistic output set from each CNN which is associated with non PCa class.

Figure 4

Slice-level ROC curve of the proposed ResNet inspired deep learning architecture (AUC: 0.87, CI: 0.84–0.90).

Figure 5

Patient-level ROC curve of the proposed pipeline: Random Forest classifier trained on the features extracted by the CNNs (AUC: 0.84, CI: 0.76–0.91).