1.1. Background: Recently, typical CT features of coronavirus disease (COVID-19) pneumonia have been reported. However, the prognostic value of CT evaluations has not been fully elucidated. Methods:In this retrospective, multi-center cohort study, we reviewed 645 patients who had died or recovered from laboratory-confirmed COVID-19 between January 23, 2020 and March 6, 2020 in two hospitals in Wuhan and Shenyang, China. Demographic, clinical, laboratory, and CT data were extracted and compared between the non-survivor and survivor group. Multivariable logistic regression analysis identified risk factors for in-hospital death. Results:This study enrolled 253 patients (63 died in the hospital, 190 were discharged). Compared to survivors, non-survivors were older, mostly male, had a higher prevalence of preexisting comorbidity, higher incidences of hypoxemia, lymphopenia and bacterial coinfection (p<0.001 for each). Regarding CT evaluations, non-survivors had higher CT scores (14.3±3.4 vs. 8.1±2.9), higher incidences of bronchial dilation with mosaic (34.9% vs. 10.5%), emphysema (28.6% vs. 10.5%), and diffuse opacity distribution (76.1% vs. 36.8%; all p<0.001). Multivariable regression analysis showed that increasing odds of in-hospital death were associated with preexisting comorbidity (OR=12.48, 95% CI: 1.48-105.07; p=0.020), lower peripheral capillary oxygen saturation (per 1% increase OR=0.681, 95%CI: 0.50-0.93; p=0.016), bacterial coin-Volume 2 | Issue 7 ajsccr.org 2 fection (OR=10.73, 95%CI: 1.01-114.01; p=0.049), and higher CT involvement scores (per 0.5-point increase OR=1.41, 95%CI: 1.04-1.91; p=0.028). Conclusion:The CT involvement score is a potential predictor of short-term outcomes in patients with COVID-19. A thorough assessment combining CT evaluation with demographic and clinical information may help establish risk stratifications and optimize treatment decisions at an early stage.