Prognostic Value of Clinical and Computed Tomography Score in Predicting Outcome of Patients with COVID-19: A prospective study in Iran

Abstract Background: Different clinical and radiologic factors predict poor outcomes in patients with Coronavirus Disease 2019 (COVID-19). Thus, we aimed to investigate the performance of two separate clinical and radiologic (CT) scoring systems in detecting the prognosis of patients with COVID-19 using a low-dose protocol for CT imaging. Methods: Eighty-six patients with confirmed COVID-19 were included in this prospective study. All patients underwent low-dose chest CT at the initial workup. By evaluating the extent of lung involvement on patients’ initial CT scan, scores from 0 to 4 were assigned to the five lobes and the lingula. Clinical score was based on the following factors: age, sex, presence of comorbidities, respiratory rate, and oxygen saturation. After at least 15 days of follow-up, the disease outcome was classified as either severe (intensive care unit admission, intubation, or death) or favorable. ROC analysis was used to evaluate the ability of each scoring system to predict patients' outcomes. Results: After follow-up, 80.2% and 15.1% of cases had developed favorable and severe outcomes (respectively), and 4.7% were lost to follow-up. Those with severe outcomes had a significantly higher clinical score and CT score than patients with favorable outcomes (p < 0.001 and p= 0.012, respectively). The intra-class correlation coefficient value for the CT score was 0.95. The optimal threshold of the CT score for identifying patients with severe outcomes was 7.5 (area under curve= 0.721) with 77% sensitivity and 65% specificity; the clinical score cut-off was 9.25 (area under curve= 0.832) with 92.3% sensitivity and 72.1% specificity. Conclusions: Both CT and clinical scoring systems displayed a quick, safe, and objective method for predicting outcomes in patients with COVID-19. However, compared with imaging, stratification of patients based on clinical factors seems to be a stronger predictor of outcome.