Artificial Intelligence-Enabled Prostate Cancer Diagnosis and Prognosis: Current State and Future Implications.

Abstract

In this modern era of digital pathology, artificial intelligence (AI)-based diagnostics for prostate cancer has become a hot topic. Multiple retrospective studies have demonstrated the benefits of AI-based diagnostic solutions for prostate cancer that includes improved prostate cancer detection, quantification, grading, interobserver concordance, cost and time savings, and a potential to reduce pathologists' workload and enhance pathology laboratory workflow. One of the major milestones is the Food and Drug Administration approval of Paige prostate AI for a second review of prostate cancer diagnosed using core needle biopsies. However, implementation of these AI tools for routine prostate cancer diagnostics is still lacking. Some of the limiting factors include costly digital pathology workflow, lack of regulatory guidelines for deployment of AI, and lack of prospective studies demonstrating the actual benefits of AI algorithms. Apart from diagnosis, AI algorithms have the potential to uncover novel insights into understanding the biology of prostate cancer and enable better risk stratification, and prognostication. This article includes an in-depth review of the current state of AI for prostate cancer diagnosis and highlights the future prospects of AI in prostate pathology for improved patient care.