Chest computed tomography versus RT-PCR in early diagnostics of COVID-19 - a systematic review with meta-analysis.

Abstract

The purpose of this study was to compare the sensitivity and specificity of computed tomography (CT) scans of the chests of patients with the reference reverse-transcription real-time polymerase chain reaction (RT-PCR) in early diagnosis of COVID-19. A systematic review with meta-analysis for numerical outcomes was performed, including 10 studies (6528 patients). High risk of systematic bias (spectrum bias) was demonstrated in all studies, while in several studies research information bias was found to be possible. The sensitivity of CT examination ranged from 72% to 98%, and the specificity from 22% to 96%. The overall sensitivity of the CT scan was 91% and the specificity 87% (95% CI). Overall sensitivity of the RT-PCR reference test was lower (87%) than its specificity (99%) (95% CI). No clear conclusion could be drawn on the rationale of using CT scanning in the early diagnosis of COVID-19 in situations when specific clinical symptoms and epidemiological history would indicate coronavirus infection. The sensitivity of the CT test seems to be higher than that of the RT-PCR reference test, but this may be related to the mode of analysis and type of material analysed in genetic tests. CT scanning could be performed in symptomatic patients, with a defined time interval from symptom onset to performing CT or RT-PCR, and it should be explicitly included as an additional procedure when initial coronavirus genetic test results are negative, while clinical symptoms and epidemiological history indicate possible infection. However, a reference test showing the presence of coronavirus genetic material is essential throughout the diagnostic and treatment process.