摘要
目的初步探索人工智能肺炎辅助诊断系统在新型冠状病毒肺炎(COVID-19)疑似病例CT筛查中的应用价值。方法回顾性分析2020年1-3月郑州大学第一附属医院收治的206例COVID-19疑似病例的胸部CT及临床资料。其中COVID-19确诊患者61例(轻型1例,普通型29例,重型19例,危重型12例),排除COVID-19患者145例(85例为其他类型肺炎,60例胸部CT未见明显异常)。利用人工智能肺炎辅助诊断系统对206例COVID-19疑似患者做出"是否为病毒性肺炎"的初步诊断,计算灵敏度、特异度、阳性预测值、Youden指数及AUC等评价该系统对COVID-19的辅助筛查作用。利用该系统定量计算61例COVID-19确诊患者胸部CT肺炎感染区域的体积和密度,并进行随访。结果 (1)人工智能肺炎辅助诊断系统对COVID-19疑似患者辅助筛查的灵敏度、特异度、阳性预测值、Youden指数和AUC分别为75.4%、53.1%、40.4%、0.285和0.643。(2)肺炎辅助诊断系统可为COVID-19确诊患者提供客观的量化评价指标:全肺病灶体积、全肺病灶体积占全肺容积的百分比等。(3)经过2~34 d的随访,有3例危重型患者死亡,末次胸部CT的全肺病灶体积占全肺容积的百分比分别为54.5%、74.0%及84.6%。结论人工智能肺炎辅助诊断系统对COVID-19疑似病例的胸部CT有一定的筛查价值,可以定量评估肺炎感染区域的体积和密度,为患者病情评估或随访提供客观的量化指标。
Objective To explore the application value of artificial intelligence pneumonia assistant diag⁃nosis system in CT screening of suspected COVID⁃19 patients.Methods Retrospective analysis was performed on the chest CT imaging and clinical data of 206 suspected COVID⁃19 patients in the First Affiliated Hospital of Zhengzhou University from January 2020 to March 2020.61 patients were diagnosed as COVID⁃19(1 patient with mild pneumonia,29 patients with ordinary pneumonia,19 patients with severe pneumonia,and 12 patients with critical⁃severe pneumonia),and 145 patients were excluded from COVID⁃19(85 patients with other types of pneu⁃monia and 60 patients with normal chest CT).A preliminary diagnosis of“whether it is viral pneumonia”was made in 206 suspected COVID⁃19 patients by using the artificial intelligence pneumonia assistant diagnostic system,and the screening effect of this system on COVID⁃19 was evaluated by sensitivity,specificity,positive predictive val⁃ue,Youden index and AUC.The system was used to calculate the volume and density of the infected area of chest CT in 61 confirmed COVID⁃19 patients.All patients were followed up for a period of time.Results(1)Sensitivi⁃ty,specificity,positive predictive value,Youden index and AUC of the artificial intelligence pneumonia assistant diagnostic system for the screening of suspected COVID⁃19 patients were 75.4%,53.1%,40.4%,0.285 and 0.643,respectively.(2)This pneumonia assistant diagnostic system could provide quantitative evaluation indica⁃tors for confirmed COVID⁃19 patients objectively:total lung lesion volume and the percentage of lesion volume in total lung volume.(3)COVID⁃19 patients were followed up for 2 to 34 days.3 patients died eventually,with the percentage of lesion volume in total lung volume(the last chest CT)54.5%,74.0%and 84.6%,respectively.Conclusion The artificial intelligence pneumonia assistant diagnosis system has certain screening value for sus⁃pected COVID⁃19 cases.It can assess the volume of the infect
作者
郭和合
詹鹤凤
张永高
高剑波
GUO Hehe;ZHAN Hefeng;ZHANG Yonggao;GAO Jianbo(Department of Radiology,the First Affiliated Hospital of Zhengzhou University,Zhengzhou 450052,China)
出处
《实用医学杂志》
CAS
北大核心
2020年第13期1729-1732,共4页
The Journal of Practical Medicine
