Development and validation of COVID-19 Radiological Risk Score (COVID-RRS): a multivariable radiological score to estimate the in-hospital mortality risk in COVID-19 patients.

OBJECTIVE: To develop and validate in-hospital mortality risk score comprising radiological aberrances in chest computed tomography (CT) performed on admission. PATIENTS AND METHODS: Single-center, longitudinal cohort study in adult patients admitted with Coronavirus Disease 2019 (COVID-19) to our ward. Patients were followed-up during hospitalization until discharge or death. Eligibility criteria for the study comprised positive real-time reverse transcription-polymerase chain reaction test (RT-PCR) for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and ground-glass opacities in chest CT. In-hospital death was the outcome of interest. Radiological, laboratory, and clinical data were analyzed. Radiological determinants of mortality were used as variables in multivariate logistic regression analysis, and results were used to build a radiological risk score. RESULTS: 371 patients were enrolled in development and validation cohorts (181 and 190 respectively), with a total of 47 non-survivors. Univariate analysis data determined 12 predictive factors (nine risk and three protective). In multivariate analysis, we developed COVID-RRS (COVID-19 Radiological Risk Score) - a radiological score predicting in-hospital COVID-19 mortality risk comprising estimated lung involvement percentage, pleural effusion, and domination of consolidation-type changes in chest CT. Our score was superior in the prediction of COVID-19 mortality to the percentage of lung involvement alone, Chest Computed Tomography Severity Score (CTSS), and Total Severity Score (TSS) in both groups with AUC of 0.910 and 0.902, respectively (p <0.001). CONCLUSIONS: Additional imaging features independently contribute to COVID-19 mortality risk. Our model comprising lung involvement estimation, pleural effusion, and domination of consolidations performed significantly better than scores based on the extent of the changes alone. COVID-RRS is a simple, reliable, and ready-to-use tool for clinical practice.

CLINICAL RESPONSE PREDICTORS of TOCILIZUMAB THERAPY in PATIENTS with SEVERE COVID-19

Background: Aberrant immune response is hallmark of severe COVID-19, irrespectively from viral replication. Immunomodulatory therapies such as interleukin-6 (IL-6) receptor inhibitors were proven to be benefcial in reducing in-hospital mor-tality1. Yet, it remains unclear which patients can beneft most from such therapy. Objectives: To identify predictors of clinical response to tocilizumab (TCZ) added to dexamethasone in patients hospitalized with severe COVID-19. Methods: We prospectively assessed clinical and laboratory details of 120 patients hospitalized due to severe COVID-19 treated with TCZ (two doses of 8 mg/kg 24h apart) in our ward between 1st Feb 2021 and 31st Dec 2021. Severe COVID-19 was defned as SpO2 <94% on room air with ground glass opacities in chest computed tomography (CT). Clinical response was defned as respiratory improvement on day 5 after TCZ infusion compared to day of treatment initiation, no further deterioration and survival. Decision of adding TCZ to dexamethasone as emergency therapy was made collectively by rheumatologists experienced in COVID-19 treatment. Laboratory and clinical parameters from hospital admission day and from TCZ institution day were analyzed. Statistical analysis was conducted with PQStat v.1.8.2 and predictors were identifed in univariate logistic regression. Results: We identifed 86 (71.7%) clinical responders and 34 (28.3%) non-re-sponders. 20 (58.8%) of the second group needed ICU admission, 18 (52.9%) died on ICU and 2 patients (5.9%) died on the ward. Responders were signif-cantly younger (mean age 56.1 vs. 63.5 years, p=0.006), had lesser comorbidity burden (median Charleson Comorbidity Index 2 vs. 3, p=0.005), lower median lung involvement (50 vs. 70%, p<0.001), higher median baseline PaO2/FiO2 index (203 vs. 106, p<0.001) and less of them needed high-fow oxygen therapy on TCZ initiation day (12.7% vs 32.4%, p=0.025). Identifed predictors of clinical response are shown in Table 1. Conclusion: Administration of TCZ early in severe disease, with moderate IL-6 concentration and low organ damage indices is most benefcial in patients with severe COVID-19, especially in younger patients without respiratory and cardiac comorbidities.