人工智能软件检测胸部CT肋骨骨折的效能探讨
Effectiveness study of artificial intelligence software in detecting rib fractures on chest CT images
李玲 1李海梅 1何悦明1
作者信息
- 1. 首都医科大学附属复兴医院放射科 北京 100038
- 折叠
摘要
目的 探讨评价人工智能(artificial intelligence,AI)软件辅助放射科医师在胸部CT图像检测肋骨骨折的应用效能.方法 选取因胸部外伤行急诊胸部CT检查的患者 301 例.将 6 位放射科医师分为低年资组(≤5 年)和高年资组(≥10 年)对CT图像阅片;1 个月后,医师+AI软件联合进行第 2 次阅片.采用卡方检验比较不同组间检测肋骨骨折敏感度的差异.配对样本t检验比较结合AI前后医师诊断时间是否有统计学意义.结果 低年资组+AI、高年资组+AI结合阅片在全部肋骨骨折、错位型、隐匿型肋骨骨折的检测敏感度均优于医师单独阅片,差异有统计学意义(P<0.001).应用AI阅片后,高低年资组医师的阅片时间均明显缩短,差异有统计学意义(P<0.001).结论 AI软件辅助诊断有助于放射科医师提高肋骨骨折检测效能,优化诊断流程并有效减少诊断时间.
Abstract
Objective To evaluate the effectiveness of artificial intelligence(AI)software in assisting radiologists in detect-ing rib fractures on chest CT images.Methods 301 cases of emergency chest CT due to chest trauma were collected.Six imag-ing physicians were divided into a low seniority group(≤5 years)and a high seniority group(≥10 years)to review CT images;After one month,the physician and AI software jointly reviewed the film for the second time.Chi-square test was used to compare the sensitivity differences between different groups in detecting rib fractures.Paired sample t-tests were used to compare whether there was statistical significance in physician diagnostic time before and after AI.Results The detection sensitivity of low se-niority group+AI and high seniority group+AI combined with film reading in all rib fractures,dislocation type,and concealed rib fractures was better than that of physicians alone,with a statistically significant difference(P<0.001).After using AI to read films,the reading time of senior and junior physicians was significantly shortened,with a statistically significant difference(P<0.001)Conclusion AI software-assisted diagnosis can help radiologists improve the detection efficiency of rib fractures,opti-mize the diagnostic process,and effectively reduce diagnostic time.
关键词
人工智能/肋骨骨折/体层摄影术,X线计算机Key words
Artificial intelligence/Rib fractures/Tomography,X-ray computed引用本文复制引用
出版年
2024
NETL
NSTL
万方数据