Interlaboratory Gleason grading variation affects treatment: a Dutch historic cohort study in 30 509 patients with prostate cancer

Artificial intelligence for detection of prostate cancer in biopsies during active surveillance

OBJECTIVES

To evaluate a cancer detecting artificial intelligence (AI) algorithm on serial biopsies in patients with prostate cancer on active surveillance (AS).

PATIENTS AND METHODS

A total of 180 patients in the Prostate Cancer Research International Active Surveillance (PRIAS) cohort were prospectively monitored using pre-defined criteria. Diagnostic and re-biopsy slides from 2011 to 2020 (n = 4744) were scanned and analysed by an in-house AI-based cancer detection algorithm. The algorithm was analysed for sensitivity, specificity, and for accuracy to predict need for active treatment. Prognostic properties of cancer size, prostate-specific antigen (PSA) level and PSA density at diagnosis were evaluated.

RESULTS

The sensitivity and specificity of the AI algorithm was 0.96 and 0.73, respectively, for correct detection of cancer areas. Original pathology report diagnosis was used as the reference method. The area of cancer estimated by the pathologists correlated highly with the AI detected cancer size (r = 0.83). By using the AI algorithm, 63% of the slides would not need to be read by a pathologist as they were classed as benign, at the risk of missing 0.55% slides containing cancer. Biopsy cancer content and PSA density at diagnosis were found to be prognostic of whether the patient stayed on AS or was discontinued for active treatment.

CONCLUSION

The AI-based biopsy cancer detection algorithm could be used to reduce the pathologists' workload in an AS cohort. The detected cancer amount correlated well with the cancer length measured by the pathologist and the algorithm performed well in finding even small areas of cancer. To our knowledge, this is the first report on an AI-based algorithm in digital pathology used to detect cancer in a cohort of patients on AS.

Pros and cons of artificial intelligence implementation in diagnostic pathology

The rapid introduction of digital pathology has greatly facilitated development of artificial intelligence (AI) models in pathology that have shown great promise in assisting morphological diagnostics and quantitation of therapeutic targets. We are now at a tipping point where companies have started to bring algorithms to the market, and questions arise whether the pathology community is ready to implement AI in routine workflow. However, concerns also arise about the use of AI in pathology. This article reviews the pros and cons of introducing AI in diagnostic pathology.

Use of the ISUP e-learning module improves interrater reliability in prostate cancer grading

AIMS

Prostate cancer (PCa) grading is an important prognostic parameter, but is subject to considerable observer variation. Previous studies have shown that interobserver variability decreases after participants were trained using an e-learning module. However, since the publication of these studies, grading of PCa has been enhanced by adopting the International Society of Urological Pathology (ISUP) 2014 grading classification. This study investigates the effect of training on interobserver variability of PCa grading, using the ISUP Education web e-learning on Gleason grading.

METHODS

The ISUP Education Prostate Test B Module was distributed among Dutch pathologists. The module uses images graded by the ISUP consensus panel consisting of 24 expert uropathologists. Participants graded the same 10 images before and after e-learning. We included those who completed the tests before and after training. We evaluated variation in PCa grading in a fully crossed study design, using linearly weighted kappa values for each pathologist, comparing them to other pathologists and to the ISUP consensus panel. We analysed the improvement in median weighted kappas before and after training, using Wilcoxon's signed rank-test.

RESULTS

We included 42 pathologists. Inter-rater reliability between pathologists improved from 0.70 before training to 0.74 after training (p=0.01). When compared with the ISUP consensus panel, five pathologists improved significantly, whereas the kappa of one pathologist was significantly lower after training. All pathologists who improved significantly, graded with less than substantial agreement before training.

CONCLUSIONS

ISUP Prostate Test B e-learning reduces variability in PCa grading. E-learning is a cost-effective method for standardisation of pathology.