Diagnostic performance between RT-PCR and chest CT in outpatients with clinically suspected COVID-19

Objective: To investigate the diagnostic performance between chest computed tomography (CT) and reverse transcription-polymerase chain reaction (RT-PCR) in outpatients with suspected coronavirus disease 2019 (COVID-19). Patients and Methods: Between March and June 2020, a total of 812 patients with clinically suspected COVID-19 who underwent both chest CT and initial-single RT-PCR on admission to outpatient units were retrospectively enrolled. CT severity-score (CT-SS) was calculated and data were matched with PCR results. Results: Of 812 patients, 54% (439/812) had positive RT-PCR results, and 47% (425/812) had a positive chest CT scan. With RT-PCR results as reference, the sensitivity, specificity, accuracy of chest CT in defining COVID-19 infection were 60%, (95% CI 56-65%, 265/439 patients), 57% (95% CI 52-62%, 213/373), 59% (95% CI 55-62%, 478/812), respectively. Three hundred eighty-seven (47%) patients had no CT findings, 380/812 (46.8%) had mild, 45/812 (5.5%) had moderate, and no patients in the severe group Conclusion: Chest CT did not show high sensitivity for the diagnosis of COVID-19 for outpatients. We suggest RT-PCR should be the primary diagnostic tool. Chest CT might be considered if there is a strong clinical suspicion with repeatedly negative RT-PCR test results, ensuring infection prevention and control measures can be preserved.