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

打开文本图片集

Theapplicationof artificial inteligence inassistingphysicians of differentexperience levels inthe diagnosis ofoccult rib fractures

LiSanbao，Wanghanlin，Duijuan，LiZiyan，haniguo Departmentofadiology，iyunDistrictTraditionalChineseedicineHospital，eijingO5oo，ina

【Abstract】Objective To explore thediagnostic value of artificial intelligence （AI） in assisting physicians of diffrent seniority inthe diagnosis of hidden rib fractures.MethodsAretrospective analysis was conducted on CT imagesof 420patients with occult rib fractures from January 2023 to June2O24.The420 patients were randomly divided into14 groups，with30 patients in each group.Two junior physicians （resident physicians with less than 5 years of work experience）andtwo senior physicians （atendingphysicians with more than1O years of work experience）were assigned to perform traditional and AI assisted image reading，respectively Evaluate the diagnostic performance，diagnosticconfidence，and reading timeoflowseniorityand seniorphysicians under traditionaland AI assisted reading modes.ResultsA total of 42Ocases were enroled，with 856 hidden fractures.Thediagnostic 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% Us. 50.70% for junior physicians， 21.50% Us. 54.32% for senior physicians）， and reading time were significantlyreduced after AI assisted film review[（547±78）sand（353±76）sfor 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% forjunior 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% ） ConclusionAI assisted image reading significantly improves the diagnostic performance of physicians with differentseniority for occultrib fractures，enhancesdiagnosticconfidence，shortens image reading time，and helps improve clinical work efficiency.

【Keywords】 Artificial intelligence;Tomography，X-ray computed; Occult fracture Fund program：Beijing Miyun District Traditional Chinese Medicine Hospital （Miyun Branch of North Third Affiliated Hospitalof Traditional Chinese Medicine）Hospital levelScientific Research Project （KJY2023-005） DOI：10.19522/j.cnki.1671-5098.2025.04.004

肋骨骨折是最常见的胸部外伤，通常由车祸、跌倒和跌倒引起。（剩余8603字）