Researchers developed machine learning models that outperform PSA testing in predicting abnormal prostate MRI findings for suspected prostate cancer.
Key Details
- 1Study involved 11,879 prostate MRI scans for suspected prostate cancer.
- 2Two ML models were trained, with AUCs of 0.711 (Model A) and 0.616 (Model B), compared to 0.593 for PSA testing alone.
- 3Model A included prostate volume in addition to clinical variables; Model B did not.
- 4Model A had higher specificity (28.3%) and comparable sensitivity (89%) versus PSA testing (>4 ng/mL).
- 5False negative rates: 8% for Model A, 16.8% for Model B; most were clinically insignificant or benign.
Why It Matters
By improving risk stratification with prostate MRI, machine learning models can help triage patients more effectively, optimizing resource use and potentially reducing unnecessary imaging. This supports more personalized and efficient approaches in prostate cancer diagnosis within radiology.
Source
AuntMinnie
Related News
•AuntMinnie
AI Model Uses Ultrasound to Assess Fetal Lung Maturity
Researchers demonstrated an AI model's strong accuracy in measuring fetal lung maturity from ultrasound images.
•AuntMinnie
AI Model Predicts Dosimetry for Lu-177 PSMA Therapy Using PET/CT
A machine learning PET/CT model shows promise for predicting radiation dose prior to Lu-177 PSMA therapy in prostate cancer patients.
•AuntMinnie
AI Advances in Ultrasound Highlighted at AIUM 2026 Keynote
AI is increasingly enhancing ultrasound imaging, clinical workflows, and education, though challenges in trust and implementation remain.