Researchers developed an explainable AI model to improve breast cancer detection in MRI with lower false positive rates.
Key Details
- 1AI model was trained on over 10,000 contrast-enhanced breast MRI exams from 2005 to 2022.
- 2The dataset included diverse cases and varying categories of breast density for robust training.
- 3Unlike prior models, this AI was evaluated on both low- and high-prevalence cancer cohorts, improving clinical relevance.
- 4The model features explainability to provide evidence-backed decisions, critical for clinical adoption.
- 5A key aim is to reduce false positive rates, a common issue in breast MRI diagnostics.
Why It Matters
False positives in breast MRI can lead to unnecessary anxiety and follow-up tests. Explainable, robust AI models like this can aid radiologists in accurate cancer detection and support clinical integration through transparency.
Source
Health Imaging
Related News
•AI in Healthcare
Literature Review Highlights Gaps in Economic Evaluation of Healthcare AI
A Finnish review finds significant gaps in economic evaluation reporting of AI technologies in Western healthcare.
•AI in Healthcare
Economic Evaluations of AI in Healthcare Face Major Gaps
A Finnish review finds significant deficiencies in how studies evaluate and report the economic impact of healthcare AI.
•Health Imaging
AI Platform Triples Functional Independence in UK Stroke Patients
AI brain imaging software deployed by NHS has significantly improved stroke outcomes and functional independence rates in England.