Chest CT Imaging Signature of Coronavirus Disease 2019 Infection In Pursuit of the Scientific Evidence
Chest CT Imaging Signature of Coronavirus Disease 2019 Infection In Pursuit of the Scientific Evidence
复制标题
DOI：
10.1016/j.chest.2020.06.025
复制
发表时间：
2020-11-01
期刊：
影响因子：
9.6
通讯作者：
Kwee, Robert M.
中科院分区：
文献类型：
--
作者：
Adams, Hugo J. A.;Kwee, Thomas C.;Kwee, Robert M.
背景：胸部CT可用于冠状病毒病2019 （COVID-19）的诊断，但缺乏明确的科学证据。因此，我们系统回顾并荟萃分析了COVID-19的胸部CT影像特征。研究问题：COVID-19感染的胸部CT成像特征是什么？研究设计和方法：系统检索COVID-19患者胸部CT影像学表现的原始研究。评估了研究的方法学质量。在研究间异质性的情况下（预定义为I-2 =50），使用随机效应模型计算胸部CT成像结果的合并患病率；否则，采用固定效应模型。结果：共纳入28项研究。每项研究的COVID-19患者中位数为124例（范围50-476），共3466例患者。有症状患者的中位患病率为99%（范围为76.3%-100%）。27项研究（96%）采用回顾性设计。方法学质量方面存在转诊偏倚风险或实际偏倚（13项研究）、患者谱偏倚（8项研究）、疾病进展偏倚（26项研究）、观察者变异性偏倚（27项研究）和试验评价偏倚（14项研究）。正常胸部CT表现的总患病率为10.6%。后叶病变的总患病率为90.0%，毛玻璃混浊为81.0%，双侧病变为75.8%，左下叶受累73.1%，血管增厚72.9%，右下叶受累72.2%。胸膜积液的总患病率为5.2%，淋巴结病为5.1%，气道分泌物/芽状树征为4.1%，中心病变分布为3.6%，心包积液为2.7%，空化/囊性改变为0.7%。其他CT表现的总患病率在10.5% ~ 63.2%之间。解释：对COVID-19胸部CT成像结果的研究存在方法学质量问题。需要更多高质量的研究来建立COVID-19的CT诊断标准。根据现有证据（需要谨慎解释），一些胸部CT成像结果似乎提示COVID-19，但正常的胸部CT成像结果并不排除COVID-19，即使在有症状的患者中也是如此。
BACKGROUND: Chest CT may be used for the diagnosis of coronavirus disease 2019 (COVID-19), but clear scientific evidence is lacking. Therefore, we systematically reviewed and meta-analyzed the chest CT imaging signature of COVID-19.RESEARCH QUESTION: What is the chest CT imaging signature of COVID-19 infection?STUDY DESIGN AND METHODS: A systematic literature search was performed for original studies on chest CT imaging findings in patients with COVID-19. Methodologic quality of studies was evaluated. Pooled prevalence of chest CT imaging findings were calculated with the use of a random effects model in case of between-study heterogeneity (predefined as I-2 =50); otherwise, a fixed effects model was used.RESULTS: Twenty-eight studies were included. The median number of patients with COVID-19 per study was 124 (range, 50-476), comprising a total of 3,466 patients. Median prevalence of symptomatic patients was 99% (range, >76.3%-100%). Twenty-seven of the studies (96%) had a retrospective design. Methodologic quality concerns were present with either risk of or actual referral bias (13 studies), patient spectrum bias (eight studies), disease progression bias (26 studies), observer variability bias (27 studies), and test review bias (14 studies). Pooled prevalence was 10.6% for normal chest CT imaging findings. Pooled prevalences were 90.0% for posterior predilection, 81.0% for ground-glass opacity, 75.8% for bilateral abnormalities, 73.1% for left lower lobe involvement, 72.9% for vascular thickening, and 72.2% for right lower lobe involvement. Pooled prevalences were 5.2% for pleural effusion, 5.1% for lymphadenopathy, 4.1% for airway secretions/tree-in-bud sign, 3.6% for central lesion distribution, 2.7% for pericardial effusion, and 0.7% for cavitation/cystic changes. Pooled prevalences of other CT imaging findings ranged between 10.5% and 63.2%.INTERPRETATION: Studies on chest CT imaging findings in COVID-19 suffer from methodologic quality concerns. More high-quality research is necessary to establish diagnostic CT criteria for COVID-19. Based on the available evidence that requires cautious interpretation, several chest CT imaging findings appear to be suggestive of COVID-19, but normal chest CT imaging findings do not exclude COVID-19, not even in symptomatic patients.