Advances in artificial intelligence in prostate cancer pathology

Abstract

Prostate cancer is one of the most common cancers in men, and accurate pathology evaluation is crucial for risk stratification and treatment planning. Traditional approaches, such as Gleason grading, are limited by interobserver variability and restricted access to expert consultation and molecular testing worldwide. Advances in digital pathology and artificial intelligence (AI) present opportunities to overcome these challenges. Whole slide imaging (WSI) forms the foundation for AI research, enabling remote review, large-scale data sharing, and the development of automated tools for tumor detection, Gleason grading, and molecular prediction. AI applications now extend beyond diagnostic support to prognostic and predictive modeling, including risk of recurrence and treatment response. Despite rapid progress, limitations persist, including data size, ground truth variability, training sets predominantly sourced from high-resource settings, and the "black-box" nature of AI outputs. Pathologists generally view AI as supportive tools, emphasizing the need for transparency, explainability, standardization, and trust. Overall, AI in digital pathology shows strong potential to augment pathology expertise, improve diagnostic efficiency and accuracy, and expand access to advanced cancer care globally. Continued collaboration between pathologists, clinicians, and researchers will be essential to ensure safe, equitable adoption.