ISSN 1004-4140
CN 11-3017/P
张凯, 柴军, 刘瑞, 等. 人工智能定量分析新型冠状病毒感染不同病毒变异株CT特征[J]. CT理论与应用研究, 2023, 32(5): 595-602. DOI: 10.15953/j.ctta.2023.043.
引用本文: 张凯, 柴军, 刘瑞, 等. 人工智能定量分析新型冠状病毒感染不同病毒变异株CT特征[J]. CT理论与应用研究, 2023, 32(5): 595-602. DOI: 10.15953/j.ctta.2023.043.
ZHANG K, CHAI J, LIU R, et al. Quantitative Analysis of Computed Tomography Features of Different COVID-19 Infection Virus Variants Using Artificial Intelligence[J]. CT Theory and Applications, 2023, 32(5): 595-602. DOI: 10.15953/j.ctta.2023.043. (in Chinese).
Citation: ZHANG K, CHAI J, LIU R, et al. Quantitative Analysis of Computed Tomography Features of Different COVID-19 Infection Virus Variants Using Artificial Intelligence[J]. CT Theory and Applications, 2023, 32(5): 595-602. DOI: 10.15953/j.ctta.2023.043. (in Chinese).

人工智能定量分析新型冠状病毒感染不同病毒变异株CT特征

Quantitative Analysis of Computed Tomography Features of Different COVID-19 Infection Virus Variants Using Artificial Intelligence

  • 摘要: 目的:人工智能(AI)定量分析比较新型冠状病毒感染德尔塔(Delta)和奥密克戎(Omicron)变异株感染患者的胸部CT影像学特征。方法:回顾性分析2022年2月20日至2022年4月19日在内蒙古自治区第四医院确诊的294例新型冠状病毒Delta变异株感染患者及2022年12月1日至12月30日在内蒙古自治区人民医院确诊的222例Omicron变异株感染患者的临床资料及首次CT影像学资料进行分析,分为Delta组和Omicron组,应用推想预测肺部感染辅助诊断软件进行定量计算,比较分析组间CT影像学征象及CT定量数据。结果:磨玻璃斑片影、磨玻璃结节影、索条、实变、铺路石征、小叶间隔增厚及病灶内增粗血管影等影像学征象在两组之间比较无统计学意义。Omicron组病灶分布较Delta组更容易出现沿支气管血管束分布;Delta组的全肺病灶体积、体积占比、右肺中叶病灶体积、体积占比、右肺下叶病灶体积、体积占比均高于Omicron组;Delta组患者病灶分布于 -570~-470 HU体积、-470~-370 HU体积、-370~-270 HU体积、-270~-170 HU体积均高于Omicron组。结论:Delta变异株感染患者肺炎早期CT病灶体积及体积占比Omicron组高,Omicron组病灶分布较Delta组更容易出现不典型的沿支气管血管束分布,人工智能肺炎辅助诊断系统对COVID-19患者定量评估肺炎感染区域体积及体积占比,为患者病情评估提供客观的参考数据。

     

    Abstract: Objective: To quantitatively analyze and compare the chest computed tomography (CT) imaging features of patients infected with delta and omicron variants of COVID-19 using artificial intelligence (AI). Method: The clinical data of 294 patients infected with the novel coronavirus delta variant diagnosed at the Fourth Hospital of Inner Mongolia Autonomous Region from February 20, 2022 to April 19, 2022 and 222 patients infected with the omicron variant diagnosed at the People's Hospital of Inner Mongolia Autonomous Region from December 1, 2022 to December 30, 2022 were retrospectively analyzed. CT imaging data were analyzed and divided into delta and omicron groups. Quantitative calculation was performed using deductive predictive pulmonary infection auxiliary diagnostic software, and CT imaging signs and quantitative CT data between groups were compared and analyzed. Results: No statistical significance was noted between the delta and omicron groups in imaging signs, such as ground-glass opacity, ground-glass nodule, cord-like lesion, consolidation, paving stone sign, thickened interlobular septum, and thickened vessels in the lesion. The distribution of lesions along the bronchial vascular bundle was more likely in the omicron than in the delta group. The total lung lesion volume, volume proportion, right middle lobe lesion volume, volume proportion, right inferior lobe lesion volume, and volume proportion in the delta group were higher than those in the omicron group. The proportions of lesions in the delta group in −570 ~ −470 HU, −470 ~ −370 HU, −370 ~ −270 HU, and −270 ~ −170 HU volumes were higher than those in the omicron group. Conclusion: In the early stage of COVID-19, the volume of CT lesions in the patients infected with the delta variant was higher than that in the omicron group, and the distribution of lesions in the omicron group was more likely to have atypical distribution along the bronchial vascular bundle than that in the delta group. The volume and volume proportion of the pneumonia-infected area in patients with COVID-19 were quantitatively evaluated using the AI-assisted diagnosis system for COVID-19 to provide objective reference data for patients' condition assessment.

     

/

返回文章
返回