CT imaging changes of corona virus disease 2019(COVID-19): a multi-center study in Southwest China

An epidemic of the 2019 novel coronavirus (2019-nCoV) erupted in Wuhan, Hubei Province of China at the end of December 2019, and has quickly spread across the country, including overseas. As of February 23, 2020, 77,041 confirmed cases have been reported in China, with a cumulative death toll of 2445. As a newly identified member of the coronavirus-beta subfamily, the virus shares more than 85% homology with the bat severe acute respiratory syndrome (SARS)-like coronavirus (bat-SL-CoVZC45), belonging to the same class as the SARS-CoV found in 2003 and the Middle East respiratory syndrome virus (MERS-CoV) found in 2012. Whole-genome sequencing and genetic analysis of the strain has demonstrated that the virus is a different branch from SARS-CoV and MERS-CoV [1]. Currently, the epidemic that originated in Wuhan is ongoing. On February 11, 2020, the International Committee on Taxonomy of Virus (ICTV) officially named 2019-nCoV as “severe acute respiratory syndrome coronavirus 2” (SARS-CoV-2) [2]. Furthermore, the World Health Organization (WHO) has named the disease caused by infection with this virus as Coronavirus Disease 2019 (COVID-19). Generally, the incubation period of COVID-19 ranges from 1 to 14 days, with most people developing symptoms between 3–7 days; however, the longest incubation period can reach 24 days [3]. The clinical severity of COVID-19 varies greatly, from asymptomatic to death.

Currently, a positive result in nucleic acid testing (NAT) using reverse-transcriptase polymerase-chain-rection (RT-PCR) technology is the gold standard for diagnosing COVID-19. The assay has high specificity, but low sensitivity. Hence, chest CT has become a critical diagnostic tool for COVID-19, used in close combination with clinical manifestations and epidemiological evidence for disease confirmation. However, discrepancies between the results of NAT and imaging features have been reported in the literature. For example, the CT images of some patients showed apparent lesions in the lungs, yet NAT repeatedly showed negative results until eventually turning positive [4]. Recently, Fang et al. observed that in diagnosing COVID-19, the sensitivity of CT (98%) was significantly higher than that of RT-PCR (71%) [5]. China’s Diagnosis and Treatment Plan for Novel Coronavirus-infected Pneumonia (5th Trial Edition) has established CT imaging as the clinical diagnosis in Hubei Province. Therefore, there is an urgent need for further studies to provide data and evidence on the manifestations of COVID-19 in chest CT imaging.

In this study, the clinical manifestations, epidemiology, laboratory test results, and chest CT findings of 131 confirmed COVID-19 patients were summarized and investigated, to improve the understanding of the disease, thereby achieving effective control of the epidemic through early diagnosis, prompt treatment, and quarantine, beneficial to the timely implementation of measures for monitoring public health.

SARS-CoV-2 is a single-stranded RNA virus that can be transmitted from human to human; comparable to the transmission of SARS and MERS, SARS-CoV-2 is also transmitted primarily through respiratory droplets and contact [6, 7]. In this study, we had marginally more female subjects than male subjects, which could be related to the fact that the average age of men was lower than that of women. The major clinical symptoms observed included fever (65%), cough (65%), consistent with previous reports [3]. Notably, 76% of the patients had a history of close contact with people living in Wuhan. Although the contact history of the other 24% was unclear, these patients worked in crowded environments where numerous people gathered (such as restaurants and hotels).

In this group of patients, the WBC and neutrophil counts and the procalcitonin levels were normal in most cases, but the CRP was elevated, consistent with the report by Chen et al. [8]. Lymphocyte count was decreased in 74 patients (57%), and the average age were 51 ± 16 years; 57 patients (43%) presented normal lymphocyte counts, and the average age were 42 ± 11 years(P = 0.002). The results indicated that with increased patients’ age the immune cells were more susceptible to damage by the virus, and hence their immunity was weakened [9].

Initial chest CT scan failed to reveal any lesions in 6 of the 131 patients, while the CT findings of the remaining 125 cases showed that the majority of lesions (79%) involved bilateral lungs, 88% of the lesions involved two or more lobes simultaneously, and nearly 76% of the lesions were distributed in the periphery of the lung. The lesions were not confined to certain segments of lung lobes, possibly due to the small size of the virions that tend to deposit on the lobules in the periphery of the lung, thereby causing damage to the alveolar epithelium and affecting multiple adjacent lobules [10].

Common chest CT manifestations of COVID-19 included: (1) patchy ground-glass opacities with clear margins and visible interlobular septal thickening inside the lesions were observed in 106 cases (81%), some cases appeared a typical “ crazy paving pattern”, and visible vascular thickening. This can be explained by the virus-induced diffuse alveolar wall injury, vascular congestion, and alveolar septal inflammation [11]. (2) Increased lesion density, along with disease progression manifested as patchy consolidations, was detected in 91 cases (69%). These pathological changes could be attributed to alveolar wall collapse, causing the replacement of alveoli by exudates or products of other diseases (such as cells and epithelium) [12]. Recently, the autopsy of the first COVID-19 patient who died in China revealed significant shedding of the alveolar epithelium and the formation of the pulmonary hyaline membrane [13], consistent with the clinical symptoms of COVID-19, i.e. cough without substantial sputum. (3) Mixed lesions of ground-glass opacities and consolidations was observed in 79 patients (61%), with or without nodule. Diffused, bilateral pulmonary lesions similar to “white lungs” was observed in 10 patients. In these cases, consolidations constituted most of the lesions, with a small percentage demonstrating complicated pulmonary fibrotic foci, which might be related to the fact that the lesions were in the repairing phase [14].

Uncommon chest CT manifestations of COVID-19 included: (1) nodular lesions in 40 patients (31%). Only 7 of them had simple nodular lesions, whereas the rest had mixed nodular lesions with either ground-glass opacities or consolidations, and the other two signs are the main manifestations. (2) CT showing low central but high peripheral lesion density, or the so-called “reversed halo sign”, in the first CT scan of 1 patient. In another patient, the first CT scan showed nodular lesions, but the “reversed halo sign” emerged in the follow-up CT scan 2 days after. The pathological mechanism remains unknown, but this is not a specific imaging characteristic of COVID-19. Other diseases such as organizing pneumonia, cryptococcosis, and tuberculosis can also present the “reversed halo sign” with different pathological mechanisms [15]. (3) A small number of cases showed thickening of adjacent pleura, hydrothorax, pericardial effusion, and enlarged mediastinal lymph nodes, consistent with previous reports [16]. (4) In 1 case, a consolidation with cavity, it might be related to the bronchial discharge of necrotic material of the lesion. This was an extremely rare phenomenon.

Although the time interval between the two CT examinations for the 91 evaluated cases was short (average 3.5 days), but 66 (73%) patient’s chest CT results demonstrated rapid changes, appear significant progress or absorption. The other 25 (27%) cases was stable. From Table 3, it should note that mixed ground glass and consolidation with or without nodules in the first CT scan. These patients can be progressed, absorbed or stable. However, initial pure patchy ground-glass density imaging indicates a disease aggravated. The risk factors of disease progression are ambiguous. Recently, Liu [17] and Guo [18] have both reported that low lymphocyte counts and hypertension are predictors for worsening conditions, which could be related to the damaged immune system.

The imaging manifestations of COVID-19 are markedly similar to those of SARS and MERS [19, 20]. (1) SARS often presents as single lesion involving unilateral lung, and septal thickening is evident after the 2nd week [21]. (2) MERS is associated with high mortality. It’s difficult to distinguish the imaging from COVID-19, but pleural effusion and pneumothorax are more common in died patients cause by MERS [22]. (3) Early manifestation of influenza virus is primarily tracheobronchitis which presented as nodules or patchy shadows around the bronchial. Centrilobular nodules, pneumatocele formation and lymphadenopathy are often seen in influenza A [23]. To some extent, combined with epidemiology and chest CT can distinguish COVID-19 from other viruses. (4) For mycoplasma pneumonia, the major manifestations include reduced light transmittance of the lung lobes, thickened vascular, thickened bronchial walls, and visible peripheral nodules, presenting the “tree-in-bud” pattern. The disease is commonly complicated by pleural effusion and enlarged mediastinal and hilar lymph nodes [24]. (5) Bacterial pneumonia is often community-acquired or hospital-acquired, with increased WBCs and neutrophils and imaging often showing patchy, nodular, or consolidation shadows distributed along the bronchi or lung segments, which can be distinctly recognized [25]. (6) In cases of cryptococcus infections, lesions are mostly located under the pleura with single or multiple consolidations and nodular shadows, and the disease progresses slowly [26].

The mechanisms during the lesion mainly with ground glass no specific antiviral drugs against progression remain unclear in early stage. We speculate that antiviral drugs may only work at certain times. This phenomenon requires further antivirus study on early intervention. The radiomics will help identify microscopic features of ground glass in early stage. This work may be conducive to differentiate COVID-19 from other viral pneumonia and guide antiviral treatment evaluation.

Our study was limited in the following aspects: (1) Although we summarize the imaging feature of COVID-19, quantitative assessments of different pneumonia lesions need to be done in the future. (2) An in-depth study on whether there is a correlation between the disease course and its imaging manifestations was not conducted due to lack of prognostic data. (3) Not all relevant clinical information was collected. For some patients, the lab tests were incomplete, e.g. liver functions and blood coagulation were not examined.

In summary, the chest CT manifestations of COVID-19 often presented patchy ground-glass opacities or mixed ground-glass opacities and consolidation, involving the periphery of bilateral lungs, that can quickly change over a short period of time. COVID-19 is highly suspected if the patient also has reduced lymphocytes along with epidemiological evidence. Furthermore, it should be noted that some patients with normal chest CT imaging could demonstrate a positive NAT result.