A NEW artificial intelligence model could transform prostate cancer diagnosis by offering a highly accurate, non-invasive alternative to biopsies, researchers announced this week.
Prostate cancer is one of the most common health concerns for men, yet assessing tumor aggressiveness typically depends on tissue biopsies, which are invasive and prone to variation among clinicians. Addressing this challenge, scientists have developed the MRI-based Predicted Transformer for Prostate Cancer (MRI-PTPCa), a foundation model trained on multiparametric MRI scans linked with pathology data.
The model was trained using nearly 1.3 million image–pathology pairs from more than 5,500 patients across multiple clinical cohorts. By applying contrastive learning, MRI-PTPCa was able to differentiate cancer grades directly from imaging data with remarkable accuracy.
In real-world clinical testing, the AI system achieved an area under the curve above 0.978, with a grading accuracy of 89.1%, outperforming both standard clinical assessments and other predictive models. Importantly, the predictions produced by MRI-PTPCa were consistently aligned with pathology results, suggesting a significant step toward reducing the need for invasive diagnostic procedures.
Researchers say the model’s scalability and reliability position it as a powerful decision-support tool for clinicians. By integrating advanced imaging and AI, MRI-PTPCa could streamline early cancer detection, improve patient outcomes, and reduce reliance on biopsies, which carry risks of infection and discomfort.
Experts believe this innovation marks a milestone in precision oncology, opening the door for AI-driven, non-invasive cancer diagnostics across other tumor types in the future.
Reference
Shao L et al. An MRI–pathology foundation model for noninvasive diagnosis and grading of prostate cancer. Nat Cancer. 2025; doi: 10.1038/s43018-025-01041-x.
