AI Should Read Mammograms Only When Confident: A Hybrid Breast Cancer Screening Reading Strategy

Key Points

• AI reduces workload in mammogram readings without compromising performance, enhancing efficiency.
• The AI model's interpretations rely on prediction uncertainty, ensuring confidence in results when utilized.
• A hybrid reading strategy streamlines mammography interpretations by involving AI when it demonstrates high confidence.
• Improved screening methods highlight the potential for AI to advance diagnostic capabilities in breast cancer detection.

Abstract