Thèse d'exercice
This is not COVID : a descriptive multicentric observational study
Français
Travail non accessible
Objective: Diagnosis of COVID-19 infection remains clinically difficult. During this worldwide pandemic, CT has been largely used to evaluate clinical suspicion of COVID-19. The aim of our study was defined CT differential diagnosis.
Materials & methods: A multicenter study was performed between March 18 and April 23, 2020. Study was approved by the local institutional review board and recorded on the clinicaltrial.gov website (NCT04339686). All consecutive symptomatic patients referred for COVID-19 suspicion that underwent both RT-PCR and Chest CT within 24 hours were included. Atypical CT findings for COVID-19 were analyzed and compared with final diagnosis according to discharge report.
Results: 401 patients were finally included. Among them, 161 had chest CT typical for COVID-19 pneumonia (40%, 161/401). Of the remaining 240 patients, three main CT patterns were observed: respiratory tract infection for 73 (0.55, 73/137), heart congestion signs for 39 (0.28, 39/137), chest neoplasm for 8 (0.06, 8/137). Twelve patients had other diseases. The probability for COVID-19 pneumonia if none of these CT patterns was present was 93% (161/173).
Conclusions: 95% of the patients referred for COVID-19 pneumonia suspicion without typical finding(s) will have either no abnormal finding, respiratory tract infection, pulmonary congestion or chest neoplasm. If none of the item is present, then 99% of the patients will be COVID-19 positive.
Keywords : COVID-19 pneumonia, computed tomography, diagnosis, performance.
Menu :
Annexe :
Université de Poitiers - 15, rue de l'Hôtel Dieu - 86034 POITIERS Cedex - France - Tél : (33) (0)5 49 45 30 00 - Fax : (33) (0)5 49 45 30 50
petille@support.univ-poitiers.fr -
Crédits et mentions légales