Prediction of Disease Progression of COVID-19 Based upon Machine Learning.

BACKGROUND: Since December 2019, COVID-19 has spread throughout the world. Clinical outcomes of COVID-19 patients vary among infected individuals. Therefore, it is vital to identify patients at high risk of disease progression. METHODS: In this retrospective, multicenter cohort study, COVID-19 patients from Huoshenshan Hospital and Taikang Tongji Hospital (Wuhan, China) were included. Clinical features showing significant differences between the severe and nonsevere groups were screened out by univariate analysis. Then, these features were used to generate classifier models to predict whether a COVID-19 case would be severe or nonsevere based on machine learning. Two test sets of data from the two hospitals were gathered to evaluate the predictive performance of the models. RESULTS: A total of 455 patients were included, and 21 features showing significant differences between the severe and nonsevere groups were selected for the training and validation set. The optimal subset, with eleven features in the k-nearest neighbor model, obtained the highest area under the curve (AUC) value among the four models in the validation set. D-dimer, CRP, and age were the three most important features in the optimal-feature subsets. The highest AUC value was obtained using a support vector-machine model for a test set from Huoshenshan Hospital. Software for predicting disease progression based on machine learning was developed. CONCLUSION: The predictive models were successfully established based on machine learning, and achieved satisfactory predictive performance of disease progression with optimal-feature subsets.

Lactic dehydrogenase-lymphocyte ratio for predicting prognosis of severe COVID-19.

ABSTRACT: To develop a useful score for predicting the prognosis of severe corona virus disease 2019 (COVID-19) patients.We retrospectively analyzed patients with severe COVID-19 who were admitted from February 10, 2020 to April 5, 2020. First, all patients were randomly assigned to a training cohort or a validation cohort. By univariate analysis of the training cohort, we developed combination scores and screened the superior score for predicting the prognosis. Subsequently, we identified the independent factors influencing prognosis. Finally, we demonstrated the predictive efficiency of the score in validation cohort.A total of 145 patients were enrolled. In the training cohort, nonsurvivors had higher levels of lactic dehydrogenase than survivors. Among the 7 combination scores that were developed, lactic dehydrogenase-lymphocyte ratio (LLR) had the highest area under the curve (AUC) value for predicting prognosis, and it was associated with the incidence of liver injury, renal injury, and higher disseminated intravascular coagulation (DIC) score on admission. Univariate logistic regression analysis revealed that C-reactive protein, DIC score ≥2 and LLR >345 were the factors associated with prognosis. Multivariate analysis showed that only LLR >345 was an independent risk factor for prognosis (odds ratio [OR] = 9.176, 95% confidence interval [CI]: 2.674-31.487, P < .001). Lastly, we confirmed that LLR was also an independent risk factor for prognosis in severe COVID-19 patients in the validation cohort where the AUC was 0.857 (95% CI: 0.718-0.997).LLR is an accurate predictive score for poor prognosis of severe COVID-19 patients.