FDA/CE 认证乳腺X线筛查AI诊断效能的系统综述与荟萃分析
本研究系统检索 2018 年 1 月至 2024 年 7 月评估 FDA/CE 认证乳腺X线筛查 AI 的研究,共纳入 42 篇。荟萃分析显示数字乳腺X线中 AI 合并 AUC 为 0.89,敏感度 76.3%,特异度 89.6%。证据支持 AI 在分诊和替代第二阅片者等场景减少工作量,但支持完全替代放射科医生的前瞻性证据仍缺乏。
要点
- AI 合并 AUC 0.89、敏感度 76.3%、特异度 89.6%,可作为阅片辅助工具参考
- AI 分诊和第二阅片者替代模式在不劣于双阅片的同时可明显减少工作量
- 仅一项前瞻性研究提示 AI 作为额外审阅者可增加检出且召回增加很小
核心结论
- FDA/CE 认证 AI 系统在乳腺X线筛查中表现良好,工作流程减负证据最强,但完全替代放射科医生的前瞻性证据缺乏
研究设计与样本
- 系统检索 Scopus、Embase、PubMed,纳入 2018 年 1 月至 2024 年 7 月发表的 42 项研究
- 纳入研究评估 FDA/CE 认证商用 AI,使用数字乳腺X线(DM)和/或数字乳腺断层合成(DBT)进行乳腺癌检测
- 33 项研究评估 DM,6 项评估 DBT,3 项同时评估 DM 和 DBT;39 项为回顾性,3 项为前瞻性
- 采用改良 QUADAS-2 工具评估研究质量,随机效应荟萃分析合并独立 AI 的敏感度、特异度和 AUC
诊断标准与关键数值
- DM 中独立 AI 合并 AUC 为 0.89(95% CI: 0.85-0.92)
- DM 中独立 AI 合并敏感度为 76.3%(95% CI: 66.2-84.2%)
- DM 中独立 AI 合并特异度为 89.6%(95% CI: 84.5-93.2%)
- AI 整合工作流程的前瞻性评估显示癌症检出不劣于标准双阅片,同时显著减少工作量
- 一项前瞻性研究显示 AI 作为额外审阅者增加了癌症检出,召回增加幅度很小
临床应用与阅片要点
- AI 可用于分诊或替代第二阅片者,在维持检出水平的同时降低放射科医生工作量
- AI 作为额外审阅者可能增加额外检出,但需关注召回率变化
- 研究偏倚风险较高,尤其在选择和病例流程方面,临床应用需结合本地验证
局限性
- 纳入研究偏倚风险常较高,尤其在病例流程与时间偏倚方面
- 支持 AI 完全替代放射科医生的前瞻性证据仍然缺乏
- 多数研究设计为回顾性,影响对真实临床场景的推广
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
OBJECTIVE: To evaluate the diagnostic performance of FDA/CE-cleared AI systems for mammography screening and assess their roles across clinical workflows.
METHODS: A systematic search was conducted in Scopus, Embase, and PubMed for full-text articles published from January 2018 to July 2024 evaluating commercially available FDA/CE-cleared AI for breast cancer detection using digital mammography (DM) and/or digital breast tomosynthesis (DBT). Included studies specified the AI system and reported diagnostic performance metrics (e.g. sensitivity, specificity, AUC) and/or clinical impact (e.g. cancer detection rate [CDR], recall rate, workload). Study quality was assessed using a modified QUADAS-2 tool. A random-effects meta-analysis pooled standalone AI performance (sensitivity, specificity, and AUC); other workflow roles were synthesised narratively.
RESULTS: In total, 42 studies were included (39 retrospective; 3 prospective): 33 evaluated DM, 6 DBT, and 3 both DM and DBT. Risk of bias was frequently high, particularly for flow and timing. In DM, pooled standalone AI AUC was 0.89 (95% CI: 0.85-0.92), with sensitivity of 76.3% (95% CI: 66.2-84.2%) and specificity of 89.6% (95% CI: 84.5-93.2%). Prospective evaluations of AI-integrated workflows, including triage and second-reader replacement, showed non-inferior cancer detection with significant workload reduction compared with standard double reading. A single prospective study of AI as an additional reviewer reported additional cancer detection with minimal increase in recalls.
CONCLUSIONS: FDA/CE-cleared AI systems demonstrate promising diagnostic performance in breast screening. The strongest evidence supports workflow configurations that reduce radiologist workload, particularly triage and independent/supporting reader models, while prospective evidence supporting complete replacement of radiologists remains lacking.
引用
- Wulandari, Gandomkar, Rickard, et al. Diagnostic performance of FDA-cleared and CE-marked AI systems for mammography screening: a systematic review and meta-analysis. Eur J Radiol. 2026. PMID: 42561891. doi:10.1016/j.ejrad.2026.113136
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。