e16605 Background: Prostate Cancer is the most commonly diagnosed male cancer in the U.S. Multiparametric magnetic resonance imaging (mpMRI) is increasingly used for both prostate cancer evaluation and biopsy guidance. The PI-RADS v2 scoring paradigm was developed to stratify prostate lesions on MRI and to predict lesion grade. Prostate organ and lesion segmentation is an essential step in pre-biopsy surgical planning. Deep learning convolutional neural networks (CNN) for image recognition are becoming a more common method of machine learning. In this study, we develop a comprehensive deep learning pipeline of 3D/2D CNN based on U-Net architecture for automatic localization and segmentation of prostates, detection of prostate lesions and PI-RADS v2 lesion scoring of mpMRIs. Methods: This IRB approved retrospective review included a total of 303 prostate nodules from 217 patients who had a prostate mpMRI between September 2014 and December 2016 and an MR-guided transrectal biopsy. For each T2 weighted image, a board-certified abdominal radiologist manually segmented the prostate and each prostate lesion. The T2 weighted and ADC series were co-registered and each lesion was assigned an overall PI-RADS score, T2 weighted PI-RADS score, and ADC PI-RADS score. After a U-Net neural network segmented the prostate organ, a mask regional convolutional neural network (R-CNN) was applied. The mask R-CNN is composed of three neural networks: feature pyramid network, region proposal network, and head network. The mask R-CNN detected the prostate lesion, segmented it, and estimated its PI-RADS score. Instead, the mask R-CNN was implemented to regress along dimensions of the PI-RADS criteria. The mask R-CNN performance was assessed with AUC, Sørensen–Dice coefficient, and Cohen’s Kappa for PI-RADS scoring agreement. Results: The AUC for prostate nodule detection was 0.79. By varying detection thresholds, sensitivity/PPV were 0.94/.54 and 0.60/0.87 at either ends of the spectrum. For detected nodules, the segmentation Sørensen–Dice coefficient was 0.76 (0.72 – 0.80). Weighted Cohen’s Kappa for PI-RADS scoring agreement was 0.63, 0.71, and 0.51 for composite, T2 weighted, and ADC respectively. Conclusions: These results demonstrate the feasibility of implementing a comprehensive 3D/2D CNN-based deep learning pipeline for evaluation of prostate mpMRI. This method is highly accurate for organ segmentation. The results for lesion detection and categorization are modest; however, the PI-RADS v2 score accuracy is comparable to previously published human interobserver agreement.

e16600 Background: Prostate cancer is the most common cancer of men in the United States, with over 200,000 new cases diagnosed in 2018. Multiparametric MRI of the prostate (mpMRI) has emerged as valuable adjunct for the detection and characterization of prostate cancer as well as for guidance of prostate biopsy. As mpMRI progresses towards widespread clinical use, major challenges have been identified, arising from the need to increase accuracy of mpMRI localization of prostate lesions, improve in lesion categorization, and decrease the time and technical complexity of mpMRI evaluation by radiologists or urologists. Deep learning convolutional neural networks (CNN) for image recognition are becoming a more common method of machine learning and show promise in evaluation of complex medical imaging. In this study we describe a deep learning approach for automatic localization and segmentation of prostates organ on clinically acquired mpMRIs. Methods: This IRB approved retrospective review included patients who had a prostate MRI between September 2014 and August 2018 and an MR-guided transrectal biopsy. For each mpMRI the prostate was manually segmented by a board-certified abdominal radiologist on T2 weighted sequence. A hybrid 3D/2D CNN based on U-Net architecture was developed and trained using these manually segmented images to perform automated organ segmentation. After training, the CNN was used to produce prostate segmentations autonomously on clinical mpMRI. Accuracy of the CNN was assessed by Sørensen–Dice coefficient and Pearson coefficient. Five-fold validation was performed. Results: The CNN was successfully trained and five-fold validation performed on 411 prostate mpMRIs. The Sørensen–Dice coefficient from the five-fold cross validation was 0.87 and the Pearson correlation coefficient for segmented volume was 0.99. Conclusions: These results demonstrate that a CNN can be developed and trained to automatically localize and volumetrically segment the prostate on clinical mpMRI with high accuracy. This study supports the potential for developing an automated deep learning CNN for organ segmentation to replace clinical manual segmentation. Future studies will look towards prostate lesion localization and categorization on mpMRI.