新型冠状病毒肺炎转归期肺感染吸收曲线的建立及临床应用

Establishment and clinical application of pulmonary infection absorption curve in prognostic period of COVID-19 patients

目的通过建立与分析新型冠状病毒肺炎(COVID-19)转归期患者肺感染吸收曲线以明确其CT复查的合适时间。方法回顾分析2020年2月9日至3月16日火神山医院放射科行CT检查患者,通过人工智能(AI)软件自动分析全肺感染体积,分别计算不同CT复查时间点与相应的感染吸收变化量(前后两次CT检查的差值),最终916例经过CT复查证实为转归期患者纳入本研究,并绘制相应感染吸收量随时间变化曲线。结果肺感染吸收曲线显示,在转归期患者肺炎治疗过程中,感染吸收量最高的点位于第15日,其次为第14日和第12日,前者与后二者间无统计学差异(P>0.05),但与之前第4~11日结果均有统计学差异(P<0.05)。结论AI大数据分析显示,对于转归期患者,间隔2周左右进行CT复查是一个比较合适的复查周期。

Objective To establish and analyze the absorption curve in prognostic period of coronavirus disease 2019(COVID-19) patients and determine the proper time of CT re-examination. Methods Retrospective analysis was carried out on the data of 2 548 COVID-19 patients examined by CT in the radiology department of Wuhan Huoshenshan Hospital from February 9, 2020 to March 16, 2020. The volume of the whole lung infection was calculated automatically by artificial intelligence(AI) software. The prognostic period was determined by the calculation of the absorption amount(the alternation between the first and second CT examinations) at the different time points of repeated CT. Finally, 916 patients in the prognostic period were involved in the study. Their curve of infection absorption was created and analyzed. Results The curve of infection absorption showed that the biggest change of infection absorption was found on the 15 th day, followed by the 14 th day and the 12 th day. There was no significant difference between the former and the latter two(P>0.05), but there was significant difference between the former and the 4 th-11 th days(P<0.05). Conclusion The analysis of AI big data suggests that CT re-examination at two-week intervals can show marked change of infection absorption for the patients in prognostic period, which may be a quite suitable CT re-examination cycle.