Researchers and developers have introduced an AI-based diagnostic classifier that processes digital histopathology images (biopsy slides) from men diagnosed with prostate cancer to generate a “risk score” for future metastasis and cancer-specific mortality. 

Unlike standard pathology review (e.g. Gleason grading) or solely genomic tests, the AI assesses architectural tissue patterns, tumor microenvironment and other morphological cues across the entire slide  capturing complexity often missed by human eyes. 

Clinical-validation data (from multiple retrospective cohorts and post-surgery patients) show the AI classifier’s risk predictions align with long-term outcomes: men flagged “high-risk” had significantly increased rates of metastasis and mortality compared with those marked “low-risk.”  In some comparisons, the AI’s predictive performance rivalled or exceeded established tests based on genomics or conventional risk scores. 

Proponents hope this development will help tailor treatment more precisely  enabling aggressive therapy for high-risk patients, while avoiding overtreatment for those with indolent disease. However, experts note that integration into routine care requires further validation in larger and more diverse patient populations to ensure generalisability and avoid bias.