人工智能辅助不同年资医师对隐匿性肋骨骨折诊断价值的应用研究  

The application of artificial intelligence in assisting physicians of different experience levels in the diagnosis of occult rib fractures

作　　者：李三保 王山林 杜利娟 李梓嫣 单爱国 Li Sanbao;Wang Shanlin;Du Lijuan;Li Ziyan;Shan Aiguo(Department of Radiology,Miyun District Traditional Chinese Medicine Hospital,Beijing 101500,China)

机构地区：[1]北京市密云区中医医院放射科,北京101500

出　　处：《实用医技杂志》2025年第4期260-264,共5页Journal of Practical Medical Techniques

基　　金：北京市密云区中医医院(北中医三附院密云院区)院级科研项目(KJY2023-005)。

摘　　要：目的探讨人工智能(AI)辅助不同年资医师对隐匿性肋骨骨折诊断价值。方法回顾性分析2023年1月至2024年6月420例在北京市密云区中医医院诊断为隐匿性肋骨骨折患者的CT初诊图像,以复诊CT骨痂形成为金标准,比较低年资和高年资医师分别在传统阅片和AI辅助阅片下诊断肋骨骨折的诊断性能、诊断信心和阅片时间。结果共入组420例,隐匿性骨折856处。低年资和高年资医师在AI辅助阅片后的诊断灵敏度(低年资42.87%和74.53%;高年资52.57%和77.92%)、诊断信心明显提升(5分,低年资17.29%和50.70%,高年资医师21.50%和54.32%)、阅片时间明显缩短[低年资(547±78)s和(353±76)s;高年资(476±68)s和(305±67)s];准确率(低年资91.90%和92.56%;高年资93.17%和95.04%)和假阳性率(低年资3.56%和5.77%;高年资3.07%和3.37%)均存在一定的提升(P<0.05),AI辅助阅片后低年资和高年资的诊断性能趋同(准确率92.56%和95.04%)。结论AI辅助阅片可以提高不同年资医师对隐匿性骨折的诊断性能,避免漏诊和误诊,并减轻工作负担,提高工作效率。Objective To explore the diagnostic value of artificial intelligence(AI)in assisting physicians of different seniority in the diagnosis of hidden rib fractures.Methods A retrospective analysis was conducted on CT images of 420 patients with occult rib fractures from January 2023 to June 2024.The 420 patients were randomly divided into 14 groups,with 30 patients in each group.Two junior physicians(resident physicians with less than 5 years of work experience)and two senior physicians(attending physicians with more than 10 years of work experience)were assigned to perform traditional and AI assisted image reading,respectively Evaluate the diagnostic performance,diagnostic confidence,and reading time of low seniority and senior physicians under traditional and AI assisted reading modes.Results A total of 420 cases were enrolled,with 856 hidden fractures.The diagnostic sensitivity(42.87%vs.74.53%for junior and senior physicians,52.57%vs.77.92%for senior physicians),diagnostic confidence(5 points,17.29%vs.50.70%for junior physicians,21.50%vs.54.32%for senior physicians),and reading time were significantly reduced after AI assisted film review[(547±78)s and(353±76)s for junior physicians,(476±68)s vs.(305±67)s for senior physicians];The accuracy rate(91.90%and 92.56%for junior physicians;93.17%vs.95.04%for senior physicians)and false positive rate(3.56%and 5.77%for junior physicians;3.07%vs.3.37%for senior physicians)have both shown some improvement(P<0.05).After AI assisted reading,the diagnostic performance of junior and senior physicians is similar(accuracy:92.56%vs.95.04%).Conclusion AI assisted image reading significantly improves the diagnostic performance of physicians with different seniority for occult rib fractures,enhances diagnostic confidence,shortens image reading time,and helps improve clinical work efficiency.

关 键 词：人工智能 体层摄影术 X线计算机 隐匿性骨折 

分 类 号：R683.1[医药卫生—骨科学]