SARS-CoV-2 pneumonia: association between findings of CT and SF ratios in inpatients and critically ill patients

Abstract Introduction We aim to evaluate the association between CT findings and SpO2/FiO2 (SF) in hospitalized patients with SARS-CoV-2 pneumonia. Our secondary purpose was to assess the predictive value of the SF ratio to identify critically ill patients and mortality. Methods A prospectively recorded electronic database for all hospitalized patients with SARS-CoV-2 pneumonia was used. CT findings and SF ratio were evaluated on admission. In the clinical course, they were divided into 2 groups as inpatients and critically ill patients to compare. Results During the study period, 256 patients with suspected SARS-CoV-2 were admitted to the hospital. The final cohort consisted of 87 patients. There is a significant moderate to strong negative correlation between the SF ratio and not-, and poorly-inflated lung ratios in critically ill patients (r = -0.591, p = < .001 and r = -0.549, p = < .001). There is a significant moderate negative correlation between the SF ratio and FWHM (r = -0.417, p = .024) in critically ill patients. The optimal cut-off point of the SF ratio for admission to the ICU was determined as 445 (87.1% sensitivity, 91.1% specificity, 84.4% PPV, and 92.7% NPV). Conclusions Lung involvement on chest CT in critically ill patients shows heterogeneity and FWHM could be used to estimate the lung heterogeneity. The SF ratio is significantly associated with CT findings and SF ratio on admission can be used alone to identify critically ill patients and mortality.