Discussion
Patient demographics in terms of age and gender distributions are similar to other published studies [5, 6]. Clinical symptoms and laboratory findings also match those of other initial reports with smaller cohorts [5, 6] with fever and cough in the majority of the patients. The majority patients had normal white blood cell count and lymphocyte counts at presentation though 26.4% of lymphocyte counts below normal range. Of patients in this study, 85.2% had traveled from Wuhan or known contacts with infected individuals. Notably, 14.8% of patients in this study based in Shaanxi did not have known or identifiable exposures. The source of infection may become increasingly difficult to identify if the disease becomes more widespread.
Depending on the severity of clinical symptoms, patients do not always present within the first few days of symptom onset. In our study cohort, the interval between symptom onset and first chest CT ranged from 1 to 15 days. Hence, familiarity with evolution of CT findings is useful to radiologists. Ground glass opacities are the most common CT findings within 0–3 days of symptom onsets as described in other published reports focused on initial presentations [11]. Over time, GGO remains a common finding and consolidation occurs with higher frequencies than in the early phase of disease, which means the disease is progressing rapidly. It is different from what happened with SARS [12]. Frequency of consolidation decreases 2 weeks after symptom onset. Reticulations and linear opacities, signs of interstitial involvement, and fibrosis become increasing prevalent later in the disease course. The total severity score shows a slight decrease in the third week. Cavitations were present in a small percentage of patients and likely present pre-existing conditions as none of the patients in the study cohort were observed to develop cavitation during the course of COVID-19. Lymphadenopathy and pleural effusions were absent on all the CTs analyzed, even on scans obtained 15–21 days after symptom onset. These findings are also similar to other reports [5]. Also similar to other published studies [13], we have observed a lower lobe predominance of pulmonary involvement. Bilateral involvement is found in the majority of patients though a significant number of patients do have unilateral involvement. Awareness of a significant minority of patients with only unilateral findings on CT at various time intervals is important so that radiologists do not exclude the possibility of COVID-19 simply because the findings were unilateral. Finally, in some patients, the CT findings can be minimal or even negative.
In areas outside of China, influenza remains significantly more prevalent at the current time. Common imaging findings of influenza virus–associated pneumonia include GGO, consolidation, and a combination of both. Pleural effusion, if present, is usually minimal, and lymphadenopathy is rare [14, 15]. Organizing pneumonia has also been reported [16]. There are recent publications that have suggested that CT findings of COVID-19 are similar to the organizing pneumonia (OP) pattern observed from post viral or atypical infections and other etiologies [17]. Despite a study discussing the nonspecific nature of COVID-19 pneumonia [18], a recent publication suggests that radiologists are able to differentiate COVID-19 from other viral pneumonia on CT [5, 6, 18, 19]. Laboratory testing for the novel coronavirus is time-consuming, and there are reports of shortage of test kits in some locations [20]. As we learned increasingly, initial laboratory tests for coronavirus can be falsely negative [20]. Therefore, the presence of suspicious CT findings in patients, correlated with a number of days after symptom onset, should prompt repeat laboratory testing and consideration of respiratory isolation in patients with appropriate travel and exposure history.
Limitations of our study include the retrospective nature of the study where all patients in the cohort presented to healthcare setting for evaluation. It is possible that there are infected individuals within the population with subclinical or mild clinical symptoms who did not present for care and the findings reported here are skewed toward those who were more symptomatic. Additionally, many of the more critically ill patients were transferred to other hospitals which were not included in this study. Our data were collected from six sites with variable CT scanning parameters though the data heterogeneity reflects the different practice parameters and settings. The patients did not have CT studies at regular intervals as often is the case in the clinical setting where the timing of imaging is dependent on the clinical course and different for each patient. Finally, even though the study describes evolution of CT findings, the chronic alterations in the pulmonary parenchyma remain to be determined. A longer follow-up would be needed.
In conclusion, CT findings of patients with mild COVID-19 outside of Wuhan include predominance of GGO in the early phase with an increase in the frequency of consolidation and linear opacities over the 3 weeks post onset of symptoms. A small percentage of patients can have no or minimal CT abnormalities. Given that laboratory tests for the novel coronavirus can be falsely negative, radiologists play a key role in identifying suspicious CT findings based on time interval from the onset of symptoms and guide further evaluation and management of patients.