CT-Derived Chest Muscle Metrics for Outcome Prediction in COVID-19 Patients

Purpose or Learning Objective: Decreased muscle mass is a predictor of unfavorable outcome in several conditions, but its prognostic impact on COVID-19 patients is unknown. We assessed and compared the contribution of computed tomography (CT)-derived muscle status and lung parenchymal features in predicting clinical outcomes in COVID-19 patients. Methods or Background: Clinical/laboratory data and clinical outcomes (intensive care unit [ICU] admission and death) were retrieved retrospectively for patients with COVID-19 confirmed by reverse transcriptase polymerase chain reaction who underwent chest CT on admission in four hospitals in northern Italy from February 21 to April 30, 2020. Extent and type of pulmonary involvement, mediastinal lymphadenopathy, and pleural effusion were assessed. Cross-sectional areas and attenuation of paravertebral muscles were measured on axial CT images at the T5 and T12 vertebral levels. Multivariate linear and binary logistic regression were used to find associations between variables in predicting ICU admission or death and to obtain predictive models, including odds ratios (ORs), tested and compared using receiver operating characteristic curve (ROC) analysis. Results or Findings: A total of 552 patients (364 men;median age: 65 years;interquartile range: 54-75) were included. In a CT-based model (including lung and muscle status), reduced paravertebral muscle area at the T5 level showed the highest ORs for ICU admission (4.83;p < .001) and death (2.25;p = .027). When this model was extended to include clinical variables, reduced paravertebral muscle area at the T5 level still showed the highest ORs both for ICU admission (4.34;p < .001) and death (2.28;p = .001). At ROC analysis comparing these models, the chest CT-based model had the same area under the curve (AUC) for ICU admission prediction (0.83;p = .380) and a nonsignificantly lower AUC for death prediction (0.86 versus 0.87;p =. 282). Conclusion: The combination of CT-derived lung parenchymal features and muscle status allowed us to predict outcomes reliably of COVID-19 patients, without a relevant contribution from clinical variables.