Using a Deep Learning Algorithm to Predict Prostate Cancer Grade Reclassification on Active Surveillance

Machine learning is a field of artificial intelligence that enables computers to learn from data without being explicitly programmed. It involves developing algorithms that can identify patterns, make predictions, and improve performance over time based on experience.

Researchers tested a deep learning (DL) algorithm, called AIRAProstate, to evaluate its ability to grade prostate cancer compared with traditional methods. They studied two groups of patients undergoing active surveillance. In the first group, which included 138 men diagnosed with low-risk cancer (Grade Group 1), the DL algorithm predicted which cancers would later be found to be more serious. It was more effective than the current methods used by pathologists. In the second group of 169 men who had advanced imaging before their biopsy, the DL algorithm again correctly predicted which cancers would change to a higher grade over time. These findings suggest that DL tools like AIRAProstate could help doctors better understand a patient’s cancer risk and guide treatment decisions more accurately.

J Nat’l Cancer Inst 2024 Vol. 116. Pages 1683-1686