DIFFERENTIAL DIAGNOSIS OF COVID-19 AND BACTERIAL COMMUNITY-ACQUIRED PNEUMONIA USING STANDARD CHEST COMPUTED TOMOGRAPHY

Abstract

Background: Coronavirus disease caused by SARS-CoV-2 presents imaging features that overlap with bacterial community-acquired pneumonia (CAP), making differential diagnosis clinically challenging. Objective: To evaluate the diagnostic performance of standard-dose CT (SDCT) and develop a multivariate predictive model for differentiating COVID-19 from bacterial CAP. Methods: Retrospective analysis of 200 patients (110 COVID-19, 90 CAP). CT features were assessed independently by two radiologists. Multivariate logistic regression and ROC analysis were performed. Results: Independent predictors of COVID-19 included ground-glass opacity (OR=7.9), bilateral involvement (OR=6.4), and peripheral distribution (OR=4.8) (p<0.001). The logistic regression model demonstrated AUC=0.972 (95% CI 0.948–0.992). Sensitivity 95.5%, specificity 91.2%. Hosmer–Lemeshow test p=0.64. Conclusion: Standard CT provides excellent diagnostic accuracy. Multivariate modeling significantly improves discrimination between COVID-19 and CAP.