头颈CT血管成像人工智能辅助阅片在住院医师规范化培训中的应用  

The Value of Artificial Intelligence-assisted Reading of Head and Neck CTA in Improving Teaching Effectiveness in Standardised Residency Training

作　　者：陈素月 郭荣 吉玉刚 陈望 CHEN Suyue;GUO Rong;JI Yugang;CHEN Wang(Department of Medical Imaging,Yancheng No.1 People’s Hospital,Affiliated Hospital of Medical School,Nanjing University,Yancheng,Jiangsu,224000)

机构地区：[1]南京大学医学院附属盐城第一医院、盐城市第一人民医院影像科,江苏盐城224000

出　　处：《江苏卫生事业管理》2025年第1期125-129,共5页Jiangsu Health System Management

基　　金：中国红十字基金会医学赋能公益专项基金-2022年领航菁英临床科研项目(XM_LHJY2022_05_33)。

摘　　要：目的:探讨头颈CT血管成像(CTA)人工智能(AI)辅助阅片在提升住院医师规范化培训教学效果中的应用价值。方法:将2022年1月至2023年3月在我科轮转的25名住院医师纳入试验组,利用AI自学;将2023年4月至2024年6月在我科轮转的28名住院医师纳入对照组,利用传统教学法进行培训。经过规范化培训后,为每人随机选择8份头颈CTA检查进行诊断报告书写,由上级医生记录其诊断报告书写时间,并对诊断报告书写质量进行评分。结果:试验组总成绩为优秀,对照组总成绩为良好,两组均未达到满分。试验组在关于解剖结构的描述、头颈血管病变的检出及定位、对斑块性质的描述及管腔狭窄程度的评估、对动脉瘤的识别及测量方面得分高于对照组(P<0.05);而对照组在头颈动脉以外病变的识别方面则优于试验组(P<0.05);试验组的诊断报告书写时间短于对照组诊断报告书写时间(P<0.05)。结论:AI与传统教学模式各有利弊,且不能相互取代,两者须有效结合,才能稳步提升住院医师规范化培训的教学效果。Objective:Exploring the value of artificial intelligence(AI)-assisted reading of head and neck computed tomography angiography(CTA)in enhancing teaching effectiveness in standardized residency training.Methods:25 residents rotating in our department from January 2022 to March 2023 were included in the experimental group and were self-taught using AI;28 residents rotating in our department from April 2023 to June 2024 were included in the control group and were trained using traditional teaching methods.After standardized training,8 head and neck CTA examinations were randomly selected for each person for diagnostic report writing,and their diagnostic report writing time was recorded by the supervising physician,who scored the quality of diagnostic report writing.Results:The overall score for the test group was excellent and the overall score for the control group was good,with neither group achieving a perfect score.The test group scored higher than the control group in the description of anatomical structures,detection and localization of head and neck vascular lesions,description of the nature of plaques and assessment of the degree of luminal stenosis,and identification and measurement of aneurysms(P<0.05),while the control group scored better than the test group in the identification of lesions other than the head and neck arteries(P<0.05),and the diagnostic report writing time was shorter than that of the control group(P<0.05).The diagnostic report writing time of the test group was shorter than that of the control group(P<0.05).Conclusion:AI and traditional teaching modes have their own advantages and disadvantages and cannot replace each other;they must be effectively combined to steadily improve the teaching effect of standardized residency training.

关 键 词：人工智能 医学影像 住院医师 规范化培训 教学质量 

分 类 号：R192.3[医药卫生—卫生事业管理]