Mass General Brigham researchers developed and validated an AI model that automatically identifies and measures aggressive prostate cancer lesions in MRI scans, with the goal of helping clinicians make more consistent and precise treatment decisions.

The AI model was trained and validated on MRI scans from more than 700 prostate cancer patients at Mass General Brigham. It was able to identify and demarcate the edges of 85% of the most radiologically aggressive prostate lesions (those with a PI-RADS score of 5). Tumor volume estimates from the model were independently associated with higher risks of treatment failure and metastasis, for both surgical and radiation therapy patients. Research findings were published in the journal Radiology in October 2024. Researchers planned to validate the model with a larger, multi-institutional dataset; routine clinical deployment was not confirmed in available sources.