Mass General radiologists developed a machine learning tool to identify patients at risk of intimate partner violence using imaging and clinical records.
Key Details
- 1Mass General Brigham experts created an AI program to detect domestic abuse risk.
- 2Inputs included radiology reports, imaging data, and diagnosis information.
- 3Models trained on records from 673 women experiencing domestic violence and 4,169 controls.
- 4Three AI models tested: using structured EHR data, unstructured reports, and a fusion model.
- 5Patterns in imaging—such as frequent visits and certain injury locations—inform risk prediction.
Why It Matters
Early identification of domestic violence victims remains challenging. Integrating AI decision support into radiology leverages imaging data for proactive and potentially life-saving intervention, highlighting radiology's evolving role in broader patient well-being.
Source
Radiology Business
Related News
•AuntMinnie
Radiology and AI Poised to Transform Endometriosis Diagnosis
Advances in imaging and AI are improving early detection and management of endometriosis.
•Radiology Business
AI Model Outperforms Radiologists in Early Pancreatic Cancer Detection
REMOD, a new AI model, detects pancreatic cancer on CT scans much earlier and more accurately than radiologists.
•AuntMinnie
AI in Radiology: Augmentation, Not Replacement – Risks and Realities Reviewed
AI is reshaping radiology workflows but is not ready to replace radiologists, as judgment, accountability, and clinical complexity remain unmet by current technology.