人工智能支持的乳腺X线摄影单读与双读方法:一项对照多阅片者研究
本回顾性研究纳入150例乳腺X线检查(30例病理 confirmed 恶性,120例无恶性经2年随访确认),26名不同经验放射科医生按五种方式阅片。结果显示同步AI支持将平均敏感度从69.2%提升至84.5%,模拟人-AI双读敏感度达91.8%,高于人-人双读的87.4%,特异度无显著差异。
要点
- 同步AI支持使平均敏感度从69.2%升至84.5%,特异度无显著差异(91.8% vs 90.8%,p=0.06)。
- AI作为独立第二阅片者(癌症概率≥10%标记)时,模拟人-AI双读敏感度91.8%,高于人-人双读87.4%(p=0.016)。
- AI支持对未支持单读中敏感度低于中位数的放射科医生提升最大(56.7%升至79.7%)。
核心结论
- AI支持在所有阅片方式中显著提升敏感度,尤其使表现较差的放射科医生获益;在模拟双读中,AI作为独立第二阅片者显著提高敏感度且不损害特异度。
研究设计与样本
- 回顾性研究,纳入150例乳腺X线检查(30例病理 confirmed 恶性,120例无恶性经2年随访确认),26名不同经验放射科医生参与,比较五种阅片方式:A)人工单读;B)AI单读(Lunit INSIGHT MMG);C)人工单读同步AI支持;D)模拟人-人双读;E)模拟人-AI双读(AI作为独立第二阅片者,标记癌症概率≥10%的病例)。
诊断标准与关键数值
- 同步AI支持单读平均敏感度84.5%(SD 8.1),无AI支持单读69.2%(SD 15.6),p<0.001。
- 同步AI支持单读平均特异度91.8%,无AI支持单读90.8%,p=0.06。
- 模拟人-AI双读敏感度91.8%(SD 3.4),模拟人-人双读87.4%(SD 8.8),p=0.016。
- 模拟人-AI双读平均特异度83.0%,模拟人-人双读84.0%,p=0.17。
- AI作为第二阅片者标记癌症概率≥10%的病例。
临床应用与阅片要点
- AI同步支持可提高乳腺X线单读敏感度,尤其对敏感度低于中位数的放射科医生提升更大(56.7%升至79.7%)。
- AI作为独立第二阅片者(癌症概率≥10%标记)的模拟双读敏感度高于人-人双读(91.8% vs 87.4%),特异度无显著差异。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
PURPOSE: To assess the impact of artificial intelligence (AI) on the diagnostic performance of radiologists with varying experience levels in mammography reading, considering single and simulated double reading approaches.
METHODS: In this retrospective study, 150 mammography examinations (30 with pathology-confirmed malignancies, 120 without malignancies [confirmed by 2-year follow-up]) were reviewed according to five approaches: A) human single reading by 26 radiologists of varying experience; B) AI single reading (Lunit INSIGHT MMG; C) human single reading with simultaneous AI support; D) simulated human-human double reading; E) simulated human-AI double reading, with AI as second independent reader flagging cases with a cancer probability ≥10 %. Sensitivity and specificity were calculated and compared using McNemar's test, univariate and multivariable logistic regression.
RESULTS: Compared to single reading without AI support, single reading with simultaneous AI support improved mean sensitivity from 69.2 % (standard deviation [SD] 15.6) to 84.5 % (SD 8.1, p < 0.001), providing comparable mean specificity (91.8 % versus 90.8 %, p = 0.06). The sensitivity increase provided by the AI-supported single reading was largest in the group of radiologists with a sensitivity below the median in the non-supported single reading, from 56.7 % (SD 12.1) to 79.7 % (SD 10.2, p < 0.001). In the simulated human-AI double reading approach, sensitivity further increased to 91.8 % (SD 3.4), surpassing that of the human-human simulated double reading (87.4 %, SD 8.8, p = 0.016), with comparable mean specificity (from 84.0 % to 83.0 %, p = 0.17).
CONCLUSION: AI support significantly enhanced sensitivity across all reading approaches, particularly benefiting worse performing radiologists. In the simulated double reading approaches, AI incorporation as independent second reader significantly increased sensitivity without compromising specificity.
引用
- Brancato B, Magni V, Saieva C, et al. AI-supported approaches for mammography single and double reading: A controlled multireader study. Eur J Radiol. 2025. PMID: 40262459. doi:10.1016/j.ejrad.2025.112101
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。