Prognostic significance of chest CT severity score in mortality prediction of COVID-19 patients, a machine learning study

Abstract Background The high mortality rate of COVID-19 makes it necessary to seek early identification high-risk patients with poor prognoses. Although the association between CT-SS and was reported, its prognosis significance in combination other prognostic parameters not evaluated yet. Methods This retrospective single-center study reviewed a total 6854 suspected referred Imam Khomeini hospital, Ilam city, west Iran, from February 9, 2020 December 20, 2020. performances k-Nearest Neighbors (kNN), Multilayer Perceptron (MLP), Support Vector Machine (SVM), J48 decision tree algorithms were based on most important relevant predictors. metrics derived confusion matrix used determine performance ML models. Results After applying exclusion criteria, 815 hospitalized cases entered into study. Of these, 447(54.85%) male mean (± SD) age participants 57.22(± 16.76) years. results showed that improved when they are fed by dataset data. kNN model an accuracy 94.1%, sensitivity 100. 0%, precision 89.5%, specificity 88.3%, AUC around 97.2% had best among three techniques. Conclusions integration data demographics, risk factors, clinical manifestations, laboratory algorithms. An comprehensive collection predictors could identify more efficiently lead optimal use hospital resources.