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{boldsymbol{ % }} Confidence Interval (CI): 0.84–0.90) and 0.84 (95{boldsymbol{ % }} CI: 0.76–0.91) at slice level and patient level, respectively.
Highlights
Prostate cancer is the most common form of cancer among males in the United States
With convolutional neural networks (CNNs)’ promising results in computer vision field[15,21], the medical imaging research community has shifted their interest toward deep learning-based methods for designing computer-aided detection and diagnosis (CAD) tools for cancer detection
Tsehay et al.[22] conducted a 3× 3 pixel level analysis by 5 convolution layers deep VGGNet[20] inspired CNN with 196 patients. They fine-tuned their classifier by cross-validation method within the training set with 144 patients and achieved area under receiver operating characteristic curve (ROC) curve (AUC) of 0.90 AUC on a separated test set of 52 patients
Summary
Prostate cancer is the most common form of cancer among males in the United States. In 2017, it was the third leading cause of death from cancer in men in the United States, with around 161,360 new cases which represented 19% of all new cancer cases and 26,730 deaths, which represented 8% of all cancer deaths[1]. Deep learning methods have shown promising results in a variety of computer vision tasks such as segmentation, classification, and object-detection[15,16,17] These methods consist of convolution layers that are able to www.nature.com/scientificreports extract different features from low-level local features to high-level global features from input images. To achieve convincing performance, an optimal combinations and structures of the layers as well as precise fine-tuning of hyper-parameters are required[15,17,20] This remains as one of the main challenges of deep learning-based methods when applied to different fields such as medical imaging. The result was based on 3× 3 windows of pixels extracted from MRI slices of DWI, T2-weighted images (T2w), and b-value images of 2000s mm−2
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
