key: cord-0936994-h9sfpbij
authors: Wen, Zeying; Chi, Yonge; Zhang, Liang; Liu, Huan; Du, Kun; Li, Zhengxing; Chen, Jie; Cheng, Liuhui; Wang, Daoqing
title: Coronavirus Disease 2019: Initial Detection on Chest CT in a Retrospective Multicenter Study of 103 Chinese Subjects
date: 2020-04-06
journal: Radiol Cardiothorac Imaging
DOI: 10.1148/ryct.2020200092
sha: 5bf8fae52c0600c18aa288990cadeed9dd9eaf3d
doc_id: 936994
cord_uid: h9sfpbij

BACKGROUND: Coronavirus disease 2019 (COVID-19) is a new viral respiratory disease that has recently emerged from China, becoming a pandemic. However, few studies have analyzed data regarding the clinical performance of chest computed tomography (CT) obtained in subjects with suspected COVID-19 at the initial presentation to medical facilities. OBJECTIVE: The purpose of the present study was to evaluate the performance of chest CT the initial presentation of patients with suspected COVID-19. METHODS: Data from 103 patients who were under investigation for COVID-19 based on inclusion criteria according to WHO Interim Guidance were retrospectively collected from January 21, 2020 to February 14, 2020. All patients underwent chest CT scanning and reverse transcription polymerase chain reaction testing (RT-PCR) for COVID-19 at hospital presentation. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) (with 95% confidence intervals) were calculated to evaluate the performance of CT. Subgroup analyses were also performed based on the geographical distribution of these cases in the province of Henan, China. RESULTS: There were 88 /103 (85%) patients with COVID-19 confirmed by RT-PCR. The overall sensitivity, specificity, PPV, and NPV were 93% (85-97%), 53% (27-77%), 92% (83-96%), and 42% (18-70%), respectively. Similar results were shown in both geographic regions. The respective sensitivity, specificity, PPV, and NPV for chest CT in the districts of Xinyang and Zhumadian (n = 56) were 92% (80-97%), 63% (26-90%), 93% (81-98%), and 56% (23-85%), while these indicators in the district of Anyang (n = 47) were 95% (81-99%), 43% (12-80%), 90% (76-97%), and 60% (17-93%). There were no significant differences in the prevalence of positive exams in the two geographic subgroups for CT (P=0.423) or RT-PCR (P=0.931). CONCLUSION: Although initial chest CT obtained at hospital presentation showed high sensitivity in patients under investigation for COVID-19 in the two geographic regions in Henan province, the NPV was only modest, suggesting low value of CT as a screening tool.

In December 2019, a major outbreak of viral pneumonia caused by a new coronavirus started in Wuhan, China (1) (2) (3) . The disease caused by this novel virus has been named Coronavirus Disease 2019 (COVID-19) by the World Health Organization (WHO) (4) . The disease has spread widely throughout China, followed by other countries worldwide, and it is now considered a pandemic (5) (6) (7) (8) (9) . As the disease spread in the communities has increased, many of the infected persons may not have an obvious history of exposure (10) . The rapid rate of spread confirms the highly infectious nature of the novel coronavirus, which is transmitted from person to person by means of droplets, direct contact, and inhaled air (11) .

Given that early detection and isolation of COVID-19 cases could potentially decrease the spread rate, initial efforts focused on the use of chest CT in the forefront of diagnosis, in addition to reverse transcriptase polymerase chain reaction (RT-PCR) testing (12) . Most common CT features of COVID-19 include multilobar ground-glass opacities with or without consolidation in the periphery of the lungs (13) (14) (15) (16) . Although early studies evaluating the clinical performance of chest CT obtained at initial presentation in the city of Wuhan, China showed high sensitivity (17) , it is still questionable if chest CT should have any role in routine screening for COVID-19 in patients with suspected infection. Therefore, the purpose of the present study was to evaluate the performance of chest CT in patients under investigation for COVID-19 from two centers in the province of Henan, China, which is located near the center of this pandemic.

This analysis was retrospective and approved by the relevant institutional review boards. The need for written patient consent was waived. The authors reviewed the medical records from 103 patients who were considered under investigation for COVID-19 between January 21, 2020, and February 14, 2020, from two areas in Henan Province, China. The inclusion criteria of patient under investigation are summarized in Table 1 (4). The following exclusion criteria was applied: (1) A person with a history of fever for more than 14 days, but without symptoms and signs for acute respiratory infection (ARI) or exposure history within 14 days.

(2) A person with history of ARI for more than 14 days, but without exposure history. These patients were also divided into two subgroups based on the geographical distribution. Subgroup 1 included 56 residents from Xinyang and Zhumadian districts 7 (population estimated in 11.6 million inhabitants), which are located in the southernmost area of Henan Province, adjacent to the city of Wuhan. Subgroup 2 included 47 residents from Anyang district (population estimated in 5.2 million inhabitants), which is located in the northernmost area of Henan Province, relatively distant from Wuhan.

Chest Computed tomography (CT) was performed in all patients at initial clinical presentation by using multi-slice CT systems (Philips Brilliance 16-detector CT and Philips

Brilliance 128-detector iCT, Philips, Best, Netherlands; GE Lightspeed 16-detector CT, GE Healthcare, Waukesha, WI), without intravenous contrast, using standard clinical protocols.

Parameters used included a peak x-ray tube voltage of 120 kV, automatic tube current modulation (145-300mAs) and a slice thickness of 2 to 3 mm without interslice gap. Studies were obtained in supine position during inspiratory pause (except in one elderly patient, who was unable to hold his breath). The images were reviewed by 3 radiologists with 8-15 years of experience in chest CT imaging, who were blinded to the laboratorial status (including the RT-PCR testing) of the patients, using a clinical Picture Archiving and Communication System (PACS) workstation. All diverging interpretations were resolved by discussion until a consensus was reached.

All CT images were assessed according to previous reports of radiologic findings in COVID-19 (13) (14) (15) (16) , using the Fleischner Society lexicon as reference (22) . The distribution of the lung opacities was defined as peripheral if the involvement was limited to the outer one-third of the lungs and it was defined as central if the involvement was limited to the inner 8 two thirds of the lungs. Lymphadenopathy was defined as the presence of hilar or mediastinal lymph nodes ≥ 10 mm in short-axis dimension.

The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for diagnosing COVID-19 are reported with 95% confidence interval (95% CI), using RT-PCR as reference. The continuous demographic features and CT radiation dose were described as mean (SD). Pearson Chi-square analysis was used to test differences in CT performance and RT-PCR positivity prevalence between geographic regions. Statistical analysis was performed using SPSS 26.0 software (International Business Machine, IBM, Armonk, NY). The statistical significance level was set to 0.05. The sensitivity, specificity, PPV, and NPV (with 95% CI) were 93% (85-97%), 53%

(27-77%), 92% (83-96%), and 42% (18-70%) in the entire cohort, respectively ( Table 3) .

Similar results were seen in both geographic subgroups. The sensitivity, specificity, PPV, and NPV in Subgroup 1 were 92% (80-97%), 63% (26-90%), 93% (81-98%), and 56% (23-85%), while in in Subgroup 2 they were 95% (81-99%), 43% (12-80%), 90% (76-97%), and 60% (17-93%), respectively. In addition, there were no significant differences in the distribution of 1 normal and abnormal CT or RT-PCR positivity in the two geographic subgroups (respectively, P=0.423 and P=0.931) ( Table 4 and Table 5 , respectively).

In the present study, the sensitivity and specificity of chest CT in patients under investigation for COVID-19 was 93% and 53%, respectively, using the results of RT-PCR as the standard of reference, with no significant differences between two distinct regions in the Chinese province of Henan, one of which is geographically closer to the center of the disease outbreak in Wuhan. The sensitivity in this study is similar to a previous study (97%) reported by Ai et al, who focused on the subjects in Wuhan (17) Noticeably, the sensitivity of the first RT-PCR testing was only about 42% in our study, increasing to 75% on the second round of testing. In fact, this observation has been also seen in reports from other medical centers (18, 19) . As a result, many experts in China had In this study, we also noted that none of the patients had mediastinal or hilar lymphadenopathy. Those CT findings are aligned with the report by Chung et al. (14) . This special CT features accompanied with the decrease number of peripheral blood lymphocytes may contribute to disclose the pathogenesis of COVID-19 infection. Huang, et al. reported that the virus of COVID-19 can consume lymphocytes, which may also be an important substrate for the rapid proliferation of the virus in early stages of the disease (3). We contemplate that this cytopathic effect on lymphocytes could be one of the causes for the lack of lymphadenopathy. 1 This study is not without limitations. First, CT readers were aware of the high suspicion level for COVID-19 in the population cohorts, which could have led to observer bias.

Nevertheless, they were still blinded to the RT-PCR status of the patients. Second, the PPV and NPV could be affected by local prevalence of COVID-19 in our sample of patients under investigation. To balance the magnitude of such effects, we included patients from distinct geographic regions from our province, one closer and another distant from the center of the outbreak. Last, we used RT-PCR in different specimens as the standard of reference in our study. Although the sensitivity of RT-PCR has been reported vary significantly, such effects could have been balanced in our study by obtaining multiple sequential tests during hospital admission.

We conclude that: (1) although initial chest CT showed a high sensitivity for detecting COVID-19, the NPV observed in both geographic regions (56% and 60%) were too low for proposing this exam as a screening tool in patients under investigation; (2) The specificity of CT findings was also relatively low, reinforcing the need to consider all the clinical information, exposure history, and alternative diagnoses. 

Outbreak of pneumonia of unknown etiology in Wuhan China: the mystery and the miracle

The continuing 2019-nCoV epidemic threat of novel coronaviruses to global health -the latest 2019 novel coronavirus outbreak in Wuhan

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Clinical management of severe acute respiratory infection when novel coronavirus (nCoV) infection is suspected -Interim guidance

Novel coronavirus -Republic of Korea (ex-China)

First travel-related case of 2019 novel coronavirus detected in United States

Clinical characteristics and imaging manifestations of the 2019 novel coronavirus disease (COVID-19): A multi-center study in Wenzhou city

2019 novel coronavirus of pneumonia in Wuhan, China: emerging attack and management strategies

Evolution of the novel coronavirus from the ongoing Wuhan outbreak and modeling of its spike protein for risk of human transmission

Office of state administration of traditional Chinese medicine. Notice on the issuance of a program for the diagnosis and treatment of novel coronavirus (2019-nCoV) infected pneumonia

Novel Coronavirus Pneumonia Outbreak in 2019: Computed Tomographic Findings in Two Cases

CT Imaging Features of 2019 Novel Coronavirus (2019-nCoV)

CT Imaging of the 2019 Novel Coronavirus

Pneumonia Associated with 2019 Novel Coronavirus: Can Computed Tomographic Findings Help Predict the Prognosis of the Disease?

Correlation of Chest CT and RT-PCR Testing in Coronavirus Disease 2019 (COVID-19) in China: A Report of 1014 Cases

Sensitivity of Chest CT for COVID-19: Comparison to RT-PCR

Chest CT for Typical 2019-nCoV Pneumonia: Relationship to Negative RT-PCR Testing

Patients with RT-PCR Confirmed COVID-19 and Normal Chest CT

Thin-Section CT of Severe Acute Respiratory Syndrome: Evaluation of 73 Patients Exposed to or with the Disease

Fleischner Society: glossary of terms for thoracic imaging

 Fig 1: 44-year-old woman with fever for three days, lymphocytopenia and a traveling history to Wuhan in January 2020.Axial chest CT image shows bilateral peripheral multi-focal ground-glass opacities (arrows). 

An ARI with history of fever or measured temperature ≥38 C° and cough; onset within the last ~10 days; and requiring hospitalization. However, the absence of fever does NOT exclude viral infection.

Severe acute respiratory infection (SARI) in a person, with history of fever and cough requiring admission to hospital, with no other etiology that fully explains the clinical presentation (clinicians should also be alert to the possibility of atypical presentations in patients who are immunocompromised);AND any of the following: A history of travel to Wuhan, Hubei Province, China in the 14 days prior to symptom onset; or the disease occurs in a health care worker who has been working in an environment where patients with severe acute respiratory infections are being cared for, without regard to place of residence or history of travel; or the person develops an unusual or unexpected clinical course, especially sudden deterioration despite appropriate treatment, without regard to place of residence or history of travel, even if another etiology has been identified that fully explains the clinical presentation. A person with acute respiratory illness of any degree of severity who, within 14 days before onset of illness, had any of the following exposures: close physical contact with a confirmed case of nCoV infection, while that patient was symptomatic; or a healthcare facility in a country where hospital-associated nCoV infections have been reported;2