Evaluation of Artificial Intelligence-Based Gleason Grading Algorithms "in the Wild".

The biopsy Gleason score is an important prognostic marker for prostate cancer patients. It is, however, subject to substantial variability among pathologists. Artificial intelligence (AI)-based algorithms employing deep learning have shown their ability to match pathologists' performance in assigning Gleason scores, with the potential to enhance pathologists' grading accuracy. The performance of Gleason AI algorithms in research is mostly reported on common benchmark data sets or within public challenges. In contrast, many commercial algorithms are evaluated in clinical studies, for which data are not publicly released. As commercial AI vendors typically do not publish performance on public benchmarks, comparison between research and commercial AI is difficult. The aims of this study are to evaluate and compare the performance of top-ranked public and commercial algorithms using real-world data. We curated a diverse data set of whole-slide prostate biopsy images through crowdsourcing containing images with a range of Gleason scores and from diverse sources. Predictions were obtained from 5 top-ranked public algorithms from the Prostate cANcer graDe Assessment (PANDA) challenge and 2 commercial Gleason grading algorithms. Additionally, 10 pathologists (A.C., C.R., J.v.I., K.R.M.L., P.R., P.G.S., R.G., S.F.K.J., T.v.d.K., X.F.) evaluated the data set in a reader study. Overall, the pairwise quadratic weighted kappa among pathologists ranged from 0.777 to 0.916. Both public and commercial algorithms showed high agreement with pathologists, with quadratic kappa ranging from 0.617 to 0.900. Commercial algorithms performed on par or outperformed top public algorithms.

Predicting prostate cancer grade reclassification on active surveillance using a deep Learning-Based grading algorithm.

Deep learning (DL)-based algorithms to determine prostate cancer (PCa) Grade Group (GG) on biopsy slides have not been validated by comparison to clinical outcomes. We used a DL-based algorithm, AIRAProstate, to re-grade initial prostate biopsies in two independent PCa active surveillance (AS) cohorts. In a cohort initially diagnosed with GG1 PCa using only systematic biopsies (n = 138), upgrading of the initial biopsy to ≥GG2 by AIRAProstate was associated with rapid or extreme grade reclassification on AS (odds ratio 3.3, p = .04), whereas upgrading of the initial biopsy by contemporary uropathologist reviews was not associated with this outcome. In a contemporary validation cohort that underwent prostate magnetic resonance imaging before initial biopsy (n = 169), upgrading of the initial biopsy (all contemporary GG1 by uropathologist grading) by AIRAProstate was associated with grade reclassification on AS (hazard ratio 1.7, p = .03). These results demonstrate the utility of a DL-based grading algorithm in PCa risk stratification for AS.

Predicting Prostate Cancer Molecular Subtype With Deep Learning on Histopathologic Images.

Microscopic examination of prostate cancer has failed to reveal a reproducible association between molecular and morphologic features. However, deep-learning algorithms trained on hematoxylin and eosin (H&E)-stained whole slide images (WSI) may outperform the human eye and help to screen for clinically-relevant genomic alterations. We created deep-learning algorithms to identify prostate tumors with underlying ETS-related gene (ERG) fusions or PTEN deletions using the following 4 stages: (1) automated tumor identification, (2) feature representation learning, (3) classification, and (4) explainability map generation. A novel transformer-based hierarchical architecture was trained on a single representative WSI of the dominant tumor nodule from a radical prostatectomy (RP) cohort with known ERG/PTEN status (n = 224 and n = 205, respectively). Two distinct vision transformer-based networks were used for feature extraction, and a distinct transformer-based model was used for classification. The ERG algorithm performance was validated across 3 RP cohorts, including 64 WSI from the pretraining cohort (AUC, 0.91) and 248 and 375 WSI from 2 independent RP cohorts (AUC, 0.86 and 0.89, respectively). In addition, we tested the ERG algorithm performance in 2 needle biopsy cohorts comprised of 179 and 148 WSI (AUC, 0.78 and 0.80, respectively). Focusing on cases with homogeneous (clonal) PTEN status, PTEN algorithm performance was assessed using 50 WSI reserved from the pretraining cohort (AUC, 0.81), 201 and 337 WSI from 2 independent RP cohorts (AUC, 0.72 and 0.80, respectively), and 151 WSI from a needle biopsy cohort (AUC, 0.75). For explainability, the PTEN algorithm was also applied to 19 WSI with heterogeneous (subclonal) PTEN loss, where the percentage tumor area with predicted PTEN loss correlated with that based on immunohistochemistry (r = 0.58, P = .0097). These deep-learning algorithms to predict ERG/PTEN status prove that H&E images can be used to screen for underlying genomic alterations in prostate cancer.

A deep learning system for prostate cancer diagnosis and grading in whole slide images of core needle biopsies.

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.

Carcino-Net: A Deep Learning Framework for Automated Gleason Grading of Prostate Biopsies.

Gleason scoring for prostate cancer grading is a subjective examination and suffers from suboptimal interobserver and intraobserver variability. To overcome these limitations, we have developed an automated system to grade prostate biopsies. We present a novel deep learning architecture Carcino-Net, which improves semantic segmentation performance. The proposed network is a modified FCN8s with ResNet50 backbone. Using Carcino-Net, we not only report best performance in separating the different grades, we also offer greater accuracy over other state-of-the-art frameworks. The proposed system could expedite the pathology workflow in diagnostic laboratories by triaging high-grade biopsies.Clinical relevance- Carcinoma of the prostate is the second most common cancer diagnosed in men, with approximately one in nine men diagnosed in their lifetime. The tumor staging via Gleason score is the most powerful prognostic predictor for prostate cancer patients.