AUTHOR=Li Dongdong , Ding Liting , Luo Jiao , Li Qiu-Gen 

TITLE=Prediction of mortality in pneumonia patients with connective tissue disease treated with glucocorticoids or/and immunosuppressants by machine learning

JOURNAL=Frontiers in Immunology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2023.1192369

DOI=10.3389/fimmu.2023.1192369

ISSN=1664-3224

ABSTRACT=Objectives
The assessment of accurate mortality risk was essential for managing pneumonia patients with connective tissue disease (CTD) treated with glucocorticoids or/and immunosuppressants. This study aims to construct a nomogram for predicting 90-day mortality in pneumonia patients by using machine learning. 
Methods
Data were obtained from the DRYAD database. Pneumonia patients with CTD were screened. The samples were randomly divided training cohort (70%) and validation cohort (30%). Univariate Cox regression analysis were used to screen for prognostic variables in training cohort. Prognostic variables were entered into the least absolute shrinkage and selection operator (Lasso) and random survival forest (RSF) analysis were used to screen important prognostic variables, respectively. The overlapping prognostic variables of the two algorithms were entered into the stepwise Cox regression analysis to screen the hub prognostic variables and construct model. Model predictive power was assessed using C-index, calibration curve and clinical subgroup analysis (age, gender, interstitial lung disease, diabetes mellitus). Clinical benefits of the model were assessed using decision curve analysis (DCA). Similarly, the C-index was calculated and the calibration curve was plotted to verify the model stability in the validation cohort.
Results
A total of 368 pneumonia patients with CTD (training cohort: 247; validation cohort: 121) treated with glucocorticoids or/and immunosuppressants were included. Univariate Cox regression analysis obtained 19 prognostic variables. Lasso and RSF algorithms obtained to eight overlapping variables. The overlapping variables were entered into stepwise Cox regression to obtain five variables (fever, cyanosis, blood urea nitrogen, need for ganciclovir treatment and anti-pseudomonas treatment), and prognostic model was constructed based on the five variables. The C-index of construction nomogram of the training cohort was 0.808. Calibration curve, DCA results and clinical subgroup analysis showed that the model also had good predictive power. Similarly, the C-index of the model in the validation cohort is 0.762 and the calibration curve has good predictive value.
Conclusion
In this study, the nomogram developed in performed well in predicting the 90-day risk of death in pneumonia patients with connective tissue disease treated with glucocorticoids or/and immunosuppressants.