基于胸部CT的人工智能技术在COPD中的临床应用价值及进展  

Clinical application and advance of artificial intelligence technology based on chest CT in chronic obstructive pulmonary disease

作　　者：廖玺铭 王琨[1] 李强[1] Liao Ximing;Wang Kun;Li Qiang(Department of Respiratory and Critical Care Medicine,Shanghai Dongfang Hospital,Tongji University,Shanghai 200120,China)

机构地区：[1]同济大学附属上海市东方医院呼吸与危重症医学科,上海200120

出　　处：《国际呼吸杂志》2022年第2期97-103,共7页International Journal of Respiration

基　　金：国家重点研发计划(2018YFC1313700)。

摘　　要：慢性阻塞性肺疾病(COPD)是一种严重危害人类健康的常见病和多发病。目前对COPD的诊断和严重程度判定主要依赖于肺功能与计算机断层扫描影像结合。传统的人工阅片方式存在着人力负担重、主观性强等缺陷。基于人工智能的深度学习技术通过运用大量已知结果的个体数据进行训练、建立模型,从而对COPD的影像进行精准的识别和评估,大大提高了COPD的筛查和诊断效率。目前已经有多个模型应用胸部影像在COPD的识别、慢性阻塞性肺疾病全球倡议分级评估和急性加重风险预测等方面展现了良好的临床应用前景。本文希望通过对胸部影像学及人工智能技术在COPD诊断、严重程度分级、分型等方面的进展及临床应用价值进行综述,以帮助呼吸学界提高对COPD的认知。Chronic obstructive pulmonary disease(COPD)is a common and frequently-occurring disease that seriously endangers human health.At present,the diagnosis and severity grading of COPD mainly rely on the combination of lung function and computed tomography images.The traditional manual reading method has some disadvantages such as heavy burden and strong subjectivity.Deep learning technology based on artificial intelligence,through the use of a large number of the known results of individual data to train and build model,has accurately identified and evaluated COPD images,greatly improving the efficiency of COPD screening and diagnosis.In recent years,several models using chest imaging have shown good clinical application prospects in COPD identification,Global Initiative for Chronic Obstructive Lung Disease grading assessment,and acute exacerbation risk prediction.This paper aims to review the progress and clinical application value of chest imaging and artificial intelligence technology in COPD diagnosis,severity grading,classification,and other aspects,so as to help improve the understanding of COPD in the respiratory field.

关 键 词：肺疾病 慢性阻塞性 人工智能 计算机断层扫描 深度学习 

分 类 号：R563.9[医药卫生—呼吸系统]