评估深度学习人工智能对儿科急诊肘关节骨折检测的支持作用
这项回顾性研究纳入2019年1月至2020年4月法国一所大学医院急诊科收治的755名0-15岁创伤患儿,以两名独立专家盲法判读为参考标准,评估急诊临床医生在无和有深度学习算法辅助时检测肘关节骨折的诊断性能,并外部测试算法独立性能。理论AI辅助使医生敏感度从77.3%升至98.9%(提高21.6%,p<0.001),特异度从88.3%降至63.5%(降低24.8%,p<0.001);算法独立敏感度98.0%(95%CI 96.0-99.0),特异度70.0%(95%CI 65.3-74.2)。
要点
- 理论AI辅助使急诊医生敏感度从77.3%升至98.9%,特异度从88.3%降至63.5%。
- 算法独立敏感度98.0%(95%CI 96.0-99.0),特异度70.0%(95%CI 65.3-74.2)。
- 755名患儿中352名(47%)存在肘关节骨折、关节积液和/或脱位。
核心结论
- AI算法在检测儿科肘关节骨折方面表现出高性能,显著提高急诊临床医生的敏感度并有助于减少漏诊率。
研究设计与样本
- 回顾性研究,纳入2019年1月至2020年4月法国一所大学医院急诊科收治的所有0-15岁创伤后开具肘关节正侧位X线片的儿童,共755名(中位年龄8岁),参考标准由两名对AI算法结果不知情的独立专家建立,并外部测试算法独立性能。
关键影像表现/征象
- 原文未报告具体影像征象描述。
诊断标准与关键数值
- 参考标准为两名独立专家盲法判读结果。
- 急诊医生无AI辅助时敏感度77.3%、特异度88.3%,理论AI辅助时敏感度98.9%、特异度63.5%。
- 算法独立敏感度98.0%(95%CI 96.0-99.0),特异度70.0%(95%CI 65.3-74.2)。
临床应用与阅片要点
- 原文未据此给出推荐。
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
OBJECTIVES: Interpreting elbow radiographs in pediatric trauma cases is challenging for emergency clinicians due to anatomical peculiarities of this joint in children and the risk of complications from missed fractures. Artificial intelligence (AI), particularly deep learning algorithms, has the potential to assist in fracture detection. We aimed to assess the performance of pediatric emergency clinicians in detecting elbow fractures in children without and with the assistance of a deep learning algorithm.
PATIENTS AND METHODS: This retrospective study included all children aged 0-15 years admitted to the emergency department of a French university hospital between January 2019 and April 2020, for whom frontal and lateral elbow radiographs were ordered following trauma. The reference standard was established by two independent experts blind to the AI algorithm results. The diagnostic performance of emergency clinicians was evaluated and compared without and with the theoretical AI assistance. Additionally, the performance of the stand-alone AI algorithm was externally tested.
RESULTS: Out of 755 children included (median age: 8 years), 352 (47 %) had an elbow fracture, joint effusion, and/or dislocation. The theoretical AI assistance improved clinician's sensitivity by 21.6 % (from 77.3 % to 98.9 %; p < 0.001), though this was accompanied by a 24.8 % decrease in specificity (from 88.3 % to 63.5 %; p < 0.001). The stand-alone AI algorithm achieved a sensitivity of 98.0 % (95 % CI: 96.0-99.0) and a specificity of 70.0 % (95 % CI: 65.3-74.2).
CONCLUSION: The AI algorithm demonstrated high performance in detecting pediatric elbow fractures, significantly improving sensitivity for emergency clinicians and helping to reduce the rate of missed diagnoses.
引用
- Costa JD, Vrignaud B, Frampas E, et al. Assessing deep learning artificial intelligence support for detecting elbow fractures in the pediatric emergency department. Eur J Radiol. 2025. PMID: 41197552. doi:10.1016/j.ejrad.2025.112498
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。