Source record
AI-supported reading in the MASAI mammography trial
この言語の本文は未訳のため、英語で表示しています。
The source is the MASAI trial’s prespecified interim safety analysis by Lång and colleagues in The Lancet Oncology, 2023. It compares AI-supported reading with Transpara against standard double reading in Swedish population mammography. Among about 80,000 women, cancer detection was 6.1 per 1,000 with AI support and 5.1 per 1,000 with standard reading, false positives were 1.5% in both arms, and screen-reading workload fell by 44.3%. Interval cancer, the primary endpoint, was left to later follow-up.
- Source date
- 2023年8月1日 · 年 2023
- Evidence label
- 論文
- Source type
- 論文
Citation
Lång et al., The Lancet Oncology
Lång et al. Artificial intelligence-supported screen reading versus standard double reading in the Mammography Screening with Artificial Intelligence trial (MASAI): prespecified interim safety analysis. The Lancet Oncology, 2023. https://www.thelancet.com/journals/lanonc/article/PIIS1470-2045(23)00298-X/fulltext
Key public facts
- The trial randomly assigned screening mammograms at four sites in southwest Sweden. Participants were women aged 40–80 inside the existing invitation.
- Transpara 1.7.0 scored examinations. Lower scores were read by one reader and the highest score by two readers, with detection marks on higher-risk examinations.
- In the interim analysis of about 80,000 women, cancer detection was 6.1 per 1,000 with AI support and 5.1 per 1,000 with standard reading (ratio 1.2, p=0.052).
- False-positive rates were 1.5% in both arms. Screen-reading workload fell by 44.3%.
- Interval cancer was the primary endpoint and was not reported in this safety analysis.
教育と市場把握のための条目。医療助言でも推奨でもありません。