Gleason grading, a risk stratification method for prostate cancer, is subjective and dependent on experience and expertise of the reporting pathologist. Deep Learning (DL) systems have shown promise in enhancing the objectivity and efficiency of Gleason grading. However, DL networks exhibit domain shift and reduced performance on Whole Slide Images (WSI) from a source other than training data. We propose a DL approach for segmenting and grading epithelial tissue using a novel training methodology that learns domain agnostic features. In this retrospective study, we analyzed WSI from three cohorts of prostate cancer patients. 3741 core needle biopsies (CNBs) received from two centers were used for training. The κquad (quadratic-weighted kappa) and AUC were measured for grade group comparison and core-level detection accuracy, respectively. Accuracy of 89.4% and κquad of 0.92 on the internal test set of 425 CNB WSI and accuracy of 85.3% and κquad of 0.96 on an external set of 1201 images, was observed. The system showed an accuracy of 83.1% and κquad of 0.93 on 1303 WSI from the third institution (blind evaluation). Our DL system, used as an assistive tool for CNB review, can potentially improve the consistency and accuracy of grading, resulting in better patient outcomes.

No abstract

No abstract

No abstract

fusion software. We assessed diagnostic performance of elastic (EF) versus rigid fusion (RF) PB in a propensity score matched (PSM) analysis.METHODS: A total of 314 fusion PB were prospectively collected from 2 different centres. All patients were biopsy naïve and all mpMRI reported a single suspicious area. Overall, 211 PB were performed using a RF system and 103 using a EF software. The two groups were compared for the main clinical features. A 1:1 PSM analysis was employed to reduce covariate imbalance to <10% . Detection rate (DR) for any prostate cancer (PCa) and clinically significant (cs) PCa were compared and stratified for PI-RADS Score. Chi-square and Mann-Whitney test were used to compare categorical and continuous variables, respectively. A per target Univariable and Multivariable regression analysis were applied to identity predictors of any PCa and cs PCa.RESULTS: The two cohorts were compared for the main clinical variables. After applying the PSM two cohorts of 83 cases were selected (Table 1). DR of any PCa cancer and csPCa was comparable between the two cohorts (all p>0.077). DR of csPCa was comparable between the two cohorts for every PIRADS score. (Table 2). At univariable regression analysis lesion size, PI-RADS Score, PSA-Density and EF system were predictors of any PCa (all p<0.001), however at multivariable analysis only PI-RADS Score was independent predictor of any PCa (p[0.027). At univariable regression analysis PI-RADS Score (p<0.001) and PSA-density (p[0.02) were predictors of csPCa while at multivariable analysis only PI-RADS score was independent predictors of csPCa (Table 3).CONCLUSIONS: Fusion PB guarantee high diagnostic accuracy for csPCa, regardless the fusion technology. Prospective randomized study are needed to confirm these data. PI-RADS score remains the only independent predictor of csPCa.