Post-acute sequelae of COVID-19 pneumonia: Six-month chest CT follow-up

Abstract

Background: The long-term post-acute pulmonary sequelae of COVID-19 remain unknown. Purpose: To evaluate lung injury in patients affected by COVID-19 pneumonia at the 6-month follow-up CT examination compared with the baseline chest CT examination. Materials and Methods: From March 19, 2020, to May 24, 2020, patients with moderate to severe COVID-19 pneumonia who had undergone baseline chest CT were prospectively enrolled at their 6-month follow-up. The CT qualitative findings, semiquantitative Lung Severity Score (LSS), and the well-aerated lung volume at quantitative chest CT (QCCT) analysis were analyzed. The performance of the baseline LSS and QCCT findings for predicting fibrosis-like changes (reticular pattern and/or honeycombing) at the 6-month follow-up chest CT examination was tested by using receiver operating characteristic curves. Univariable and multivariable logistic regression analyses were used to test clinical and radiologic features that were predictive of fibrosis-like changes. The multivariable analysis was performed with clinical parameters alone (clinical model), radiologic parameters alone (radiologic model), and the combination of clinical and radiologic parameters (combined model). Results: One hundred eighteen patients who had undergone baseline chest CT and agreed to undergo follow-up chest CT at 6 months were included in the study (62 women; mean age, 65 years 6 12 [standard deviation]). At follow-up chest CT, 85 of 118 (72%) patients showed fibrosis-like changes and 49 of 118 (42%) showed ground-glass opacities. The baseline LSS (.14) and QCCT findings (3.75 L and 80%) showed excellent performance for predicting fibrosis-like changes at follow-up chest CT. In the multivariable analysis, the areas under the curve were 0.89 (95% CI: 0.77, 0.96) for the clinical model, 0.81 (95% CI: 0.68, 0.9) for the radiologic model, and 0.92 (95% CI: 0.81, 0.98) for the combined model. Conclusion: At 6-month follow-up chest CT, 72% of patients showed late sequelae, in particular fibrosis-like changes. The baseline Lung Severity Score and the well-aerated lung volume at quantitative chest CT (QCCT) analysis showed excellent performance for predicting fibrosis-like changes at the 6-month chest CT (area under the curve, .0.88). Male sex, cough, lymphocytosis, and the well-aerated lung volume at QCCT analysis were significant predictors of fibrosis-like changes at 6 months, demonstrating an inverse correlation (area under the curve, 0.92).