摘要
目的探讨人工智能(AI)辅助低剂量CT在肺结节筛查中的价值。方法回顾性分析在嘉兴市中医医院放射科行胸部低剂量CT体检人员100例,分人工组、AI组及人工联合AI组,分别对入组病例进行阅片并记录肺结节,以2名主任医师联合AI读片结果为金标准,比较三组的肺结节检出率、漏诊率、假阳性率、阅片时间。结果人工组、AI组及人工联合AI组三组在检出率、假阳性率及阅片用时上,差异均有统计学意义(χ^(2)分别=145.00、52.00,H=263.00,P均<0.05)。结论人工联合AI筛查肺结节,能提高结节检出率,降低漏诊及假阳性率,缩短阅片时间。
Objective To explore the value of artificial intelligence(AI)assisted low dose CT in pulmonary nodule screening.Methods A retrospective analysis was made on 100 cases of low-dose CT in the radiology department of Jiaxing TCM hospital.They were divided into artificial group,AI group and artificial combined AI group.The pulmonary nodules were read and recorded respectively.The detection rate,missed diagnosis rate,false positive rate and reading time of pulmonary nodules were compared among the three groups according to the results of two chief physicians combined with AI.Results There were significant differences in detection rate,false positive rate and reading time among artifi⁃cial group,AI group and artificial combined AI group(χ^(2)=145.00,52.00,H=263.00,P<0.05).Conclusion Manual co⁃mbined with AI screening for pulmonary nodules can improve the detection rate of nodules,reduce missed diagnosis and false positive rate,and shorten the reading time.
作者
刘传现
周冰
LIU Chuanxian;ZHOU Bing(Department of Radiology,Jiaxing TCM Hospital Affiliated to Zhejiang Chinese Medical University,Jiaxing 314000,China.)
出处
《全科医学临床与教育》
2023年第4期317-319,共3页
Clinical Education of General Practice
基金
浙江省卫生健康委员会面上项目(2021KY1123)
浙江省嘉兴市科技计划公益类项目(2021AD30032)。
关键词
人工智能
低剂量CT
肺结节
筛查
artificial intelligence
low-dose CT
pulmonary nodules
screening
