Longitudinal Radiological Findings in Patients With COVID-19 With Different Severities: From Onset to Long-Term Follow-Up After Discharge

Objective: This study aimed to investigate the evolution of radiological findings in the patients with coronavirus disease 2019 (COVID-19) pneumonia with different severities from onset to 1-year follow-up and identify the predictive factors for different pulmonary lesion absorption status in the patients infected with COVID-19. Methods: A retrospective study was performed on the clinical and radiological features of 175 patients with COVID-19 pneumonia hospitalized at three institutions from January 21 to March 20, 2020. All the chest CT scans during hospitalization and follow-ups after discharge were collected. The clinical and radiological features from the chest CT scans both at the peak stage and before discharge from the hospital were used to predict whether the pulmonary lesions would be fully absorbed after discharge by Cox regression. Then, these patients were stratified into two groups with different risks of pulmonary lesion absorption, and an optimal timepoint for the first CT follow-up was selected accordingly. Results: A total of 132 (75.4%) patients were classified into the non-severe group, and 43 (24.6%) patients were classified into the severe group, according to the WHO guidelines. The opacification in both the groups changed from ground-glass opacity (GGO) to consolidation and then from consolidation to GGO. Among the 175 participants, 135 (112 non-severe and 23 severe patients with COVID-19) underwent follow-up CT scans after discharge. Pulmonary residuals could be observed in nearly half of the patients (67/135) with the presentation of opacities and parenchymal bands. The parenchymal bands in nine discharged patients got fully absorbed during the follow-up periods. The age of patient [hazard ratio (HR) = 0.95, 95% CI, 0.95–0.99], level of lactate dehydrogenase (LDH) (HR = 0.99; 95% CI, 0.99–1.00), level of procalcitonin (HR = 8.72; 95% CI, 1.04–73.03), existence of diffuse lesions (HR = 0.28; 95% CI, 0.09–0.92), subpleural distribution of lesions (HR = 2.15; 95% CI, 1.17–3.92), morphology of residuals (linear lesion: HR = 4.58, 95% CI, 1.22–17.11; nodular lesion: HR = 33.07, 95% CI, 3.58–305.74), and pleural traction (HR = 0.41; 95% CI, 0.22–0.78) from the last scan before discharge were independent factors to predict the absorption status of COVID-19-related pulmonary abnormalities after discharge. According to a Kaplan–Meier analysis, the probability of patients of the low-risk group to have pulmonary lesions fully absorbed within 90 days reached 91.7%. Conclusion: The development of COVID-19 lesions followed the trend from GGO to consolidation and then from consolidation to GGO. The CT manifestations and clinical and laboratory variables before discharge could help predict the absorption status of pulmonary lesions after discharge. The parenchymal bands could be fully absorbed in some COVID-19 cases. In this study, a Cox regression analysis indicated that a timepoint of 3 months since onset was optimal for the radiological follow-up of discharged patients.