In the present study, we found that the most common CT imaging features in patients with COVID-19 were: bilateral, multifocal GGO, peripheral distribution; the predominant lower lobe; pleural effusion and lymphadenectasis were rare, which is consistent with previous reports[6-8]. In addition, “feather sign”was found in 22 patients (65.2%), “dandelion sign” in 18 (52.9%), “pomegranate sign” in nine (26.1%), and “rime sign” in seven (20.6%) patients, which could be considered as new features in patients with COVID-19. The “feather sign” and “dandelion sign” on the CT image included stripe or round GGO, thickened blood vessels, and small-thickened interlobular septa. GGO shows diffuse alveolar damage under the microscope, which is histologically caused by alveoli filled with blood, pus, water, or cells[9, 10]. The reason for the thickening of blood vessels in the lesion may be the following:under the effect of inflammatory factors, the increased permeability of the vascular wall may lead to the dilation of capillaries and the corresponding thickening of the pulmonary artery[11]. The incidence of “feather sign” and “dandelion sign” in this study was 65.2% and 52.9%, respectively.
In this study, nine patients (26.1%) presented with “pomegranate sign”, which is an atypical chest CT feature of COVID-19. A “pomegranate sign” can be characterized as a further increase of the range of ground-glass opacity that occupies part of the lung sub-segment, the more significant thickening of the interlobular septum, complicated with a small amount of punctate alveolar hemorrhage[9, 12], and lesions that are in imbricate arrangement, and are similar to a pomegranate. Moreover, among seven patients (20.6%) who developed a “rime sign”, two were critically ill. A “rime sign” is characterized by numerous alveolar edemas. Hemorrhagic necrosis can be observed in some alveoli. Moreover, mucus and hemorrhagic exudate diffusely cover the bronchiole wall. Microscopically, the wall thickening, small vascular proliferation, luminal stenosis, and occlusion, accompanied by interstitial infiltration of inflammatory cells, such as lymphocytes, plasma cells, and monocytes[12], as well as numerous pulmonary interstitial fibrosis and partial hyaline degeneration are observed. This type of lesion has a wide range and looks like a white rime attached to abranch.
All patients were followed up for two weeks. Among 13 hospitalized patients, 11 had stable conditions and gradually recovered, two reported rapid progression during the hospital stay and were in critical condition, and underwent extracorporeal membrane oxygenation (ECMO). Thirty patients were cured and discharged. Fourteen underwent re-examination after 5-14 days;four had a cough, stomachache and chest distress. Very few patients developed weakness and shortness of breath. Laboratory tests revealed a significant increase in T-lymphocyte counts in four patients (flow cytometry: 762/ul, 899 /ul, respectively; the normal value was 0), accompanied by a significant increase in absolute counts of helper and cytotoxic T-lymphocytes, while alanine aminotransferase, creatinine and total bilirubin levels were increased to varying degrees, which suggested that the patient's immune function was deteriorated, and liver and renal functions were impaired; these data were consistent with previous reports[13]. These findings suggest that liver injury may be caused by SARS-CoV-2 infection or induced by drug treatment during hospitalization.
Among patients who underwent CT re-examination, two patients showed no obvious changes; in four patients, lesions were completely absorbed; in eight cases, lesions were partially absorbed; and one had mediastinal lymphadenectasis. These data suggest that CT can be used to monitor changes during disease progression, which is consistent with the findings of Hosseiny et al[14].
Multiple NAT (including nasal, anal, and throat swabs) were performed after 5-14 days. In two patients, the results of NAT of nasal and throat swabs were negative, while the result of NAT of the anal swab was positive, which is why the patients were immediately readmitted to the hospital for treatment. The remaining 12patients reported negative results for multiple NAT. A recent study[15] revealed that four patients with COVID-19 showed “positive results” for nucleic acid test 5-13 days after discharge. This suggested that current discharge standards should be revised; nasal, anal, and throat swabs should be combined, as well as supplemented by a variety of other testing methods. COVID-19 patients should be monitored during treatment, rehabilitation, and quarantine, so as to fundamentally control the occurrence of the “positive results” after discharge[16].
This study had some limitations. Firstly, no children are enrolled in this study, and the clinical, epidemiological, and imaging features of children with COVID-19 are lacking. Secondly,The number of patients collected in this study is so small that study results have certain limitations. The reliability of the conclusion needs to be further expanded to verify the sample size. Thirdly, sufficient pathological specimens are currently unavailable for comparison with imaging features. We will collect more patients data and pathological specimens to observe the evolution and outcome of the disease and determine the correlation between the imaging and pathology.