Artificial intelligence-supported double reading in European population breast cancer screening: A systematic review and meta-analysis of prospective programs.
Romuald Ferre, Thad Benefield, Cherie M Kuzmiak
✦ AI-curated · Sources linked
In 30 seconds
This meta-analysis evaluated the impact of artificial intelligence (AI) on breast cancer screening in European populations, synthesizing data from three large studies. The analysis included 597,419 examinations and found a pooled cancer detection rate (CDR) increase of +0.9 per 1000 examinations, indicating a modest benefit of AI integration in double reading screening programs without a consistent increase in recall rates.
Key findings
- Pooled cancer detection rate (CDR) increased by +0.9 per 1000 examinations.
- Pooled recall rate decreased by -0.6 per 1000 examinations.
- Positive predictive value (PPV) was higher with AI-supported screening.
- Efficiency signals included 44.3% fewer total readings in the MASAI study.
Why it matters
The integration of AI in breast cancer screening may enhance detection rates while maintaining or reducing recall rates, potentially alleviating radiologist workload and improving screening efficiency. Understanding these outcomes is crucial for optimizing screening protocols.