Stanford and Rad Partners developed a structured framework for pre-deployment evaluation of radiology AI models to guide purchasing decisions.
Key Details
- 1Framework was developed by Stanford University and Rad Partners, detailed in the American Journal of Roentgenology.
- 2A workgroup of four radiologists evaluated 13 AI models from one vendor (Aidoc) between 2022 and 2024.
- 3Nearly 89,000 exams across multiple sites were used in the assessment.
- 4Attributes for evaluating value included task tediousness, likelihood of radiologist oversight, and clinical impact of misses.
- 5Five tasks were rated as high value, five as medium, and two as low based on the framework.
Why It Matters
This framework provides radiology groups with a practical, evidence-based method to evaluate AI models before investment, addressing the gap between AI marketing claims and real-world outcomes. Adoption of such structured assessment tools can improve clinical effectiveness and ROI for AI integration in imaging practices.
Source
Radiology Business
Related News
•AuntMinnie
Study: Computer Vision Models Best LLMs in Chest CT Breast Abnormality Detection
Computer vision models (CVMs) surpass large language models (LLMs) in accurately labeling incidental breast abnormalities on chest CT scans.
•AuntMinnie
Deep Learning Models Rival Radiologists for Pancreatic Cancer Detection on CT
Deep-learning models achieved comparable or superior accuracy to experienced radiologists in detecting pancreatic cancer on CT scans, especially for small tumors.
•Radiology Business
Radiology AI Devices at Elevated Risk for FDA Recalls, Study Finds
Radiology AI devices are more likely to face FDA recalls, largely due to deviations from intended use and incomplete clinical data.