Active Surveillance for Intermediate-risk Prostate Cancer: Implications of AI-enabled Pattern Quantification.

Abstract

Active surveillance (AS) mitigates overtreatment in prostate cancer. Although men with pure Gleason pattern (GP) 3 disease are well-established candidates, criteria for AS in men with GP4 disease remain heterogeneous and weakly evidence based. Recent studies have demonstrated the absolute amount of GP4 to be the dominant predictor of outcome in men with GP4, raising the possibility that GP4 volume could refine AS selection. Clinical translation is limited by inconsistent measurement and interobserver variability, particularly in small-volume disease. Artificial intelligence and digital pathology offer a potential route to reproducible, scalable quantification but are not yet sufficiently validated for clinical decision-making. PATIENT SUMMARY: Active surveillance allows some men with prostate cancer to delay treatment and avoid side effects. For intermediate-risk disease, it remains unclear who can be safely monitored. Measuring higher-grade tumour volume may help but is difficult and not yet standardised. Computer-based analysis may improve this but is not ready to guide treatment decisions.