Incorrect AI suggestions, especially false-negatives, significantly reduce reader sensitivity and alter visual search in mammography interpretation.
Key Details
- 1Published in Radiology (July 2023), study assesses AI influence on mammography readers.
- 210 readers interpreted 60 cases with and without commercial AI assistance; eye tracking measured search behavior.
- 3False-negative AI prompts dropped median reader sensitivity from 71% (unassisted) to 39% (AI-assisted).
- 4Reader specificity improved with false-positive AI prompts (39% vs 21%).
- 5More visible AI prompts increased read times (25s to 34s).
- 6Eye-tracking showed fewer and shorter fixations when AI missed cancers.
Why It Matters
The study highlights how automation bias from AI errors can reduce diagnostic accuracy, stressing the need for careful AI design and radiologist awareness. Minimizing false-negative AI prompts is crucial in clinical workflow and training to prevent missed diagnoses.
Source
AuntMinnie
Related News
•Radiology Business
Meta Funds Ohio State’s Radiology AI Innovation Projects
Meta will fund Ohio State researchers to explore AI-driven advancements in radiology, including chest X-ray reporting and disease prediction.
•AuntMinnie
ChatGPT-5 Performs Poorly at Detecting Distal Radius Fractures on X-rays
ChatGPT-5 missed about half of distal radius fractures on x-rays, showing limited diagnostic capability in a real-world scenario.
•AuntMinnie
AI Triage Tool Speeds Up ICH Head CT Turnaround in ED
AI-based triage for intracranial hemorrhage on head CT reduced emergency department report turnaround by 13 minutes.