RESUMEN
Objective: Determine whether the tomographic characteristics of patients with COVID-19 pneumonia at the hospital admission and the initial tomographic severity score (TSS) as well as some laboratory tests or clinical characteristics predict mortality. Methods: Retrospective analytical study that included patients with a clinical diagnosis of SARSCoV2 virus infection, performed by reverse transcriptase polymerase chain reaction (RT-PCR), serologic reactive test (IgM/IgG) and/or thoracic computed tomography (CT). Patients were divided into two groups: recovered and deceased. Two radiologists (blind evaluators) described the tomographic findings. TSS, clinical and laboratory parameters in relation to mortality were analyzed. Mortality predictions were made by binary logistic regression. Results: Hypertension was the most frequent associated disease, the most common clinical presentation included cough, discomfort, fever, and dyspnea. The ground glass opacity pattern was the most frequent, followed by consolidation and distortion of the architecture; however, they were not associated with higher mortality. The pattern of pleural effusion and bronchial dilation showed a significant difference from mortality (p <0.05). The binary logistic regression model showed that a moderate and high TSS (≥ 8), as well as a higher degree of lymphopenia, history of asthma and age were associated with an increased risk of death (p< 0.05). Conclusions: TSS is useful in the initial and comprehensive diagnostic evaluation of COVID-19 pneumonia, in conjunction with markers such as lymphopenia that can predict a poor short-term outcome. A high TSS score is a predictor of mortality.