Comprehensive analysis and experimental validation of BST1 as a novel diagnostic biomarker for pediatric sepsis using multiple machine learning algorithms.

Bone marrow stromal cell antigen-1 (BST1) expression is elevated in a variety of human diseases, but its relationship with pediatric sepsis is unclear. This study aimed to investigate the expression of BST1 in pediatric sepsis patients and its value for diagnosing pediatric sepsis. Key neutrophil extracellular traps (NETs)-related genes that accurately diagnose pediatric sepsis patients were identified using various machine learning methods, including LASSO, RF, XGBoost, LightGBM, and Boruta algorithms. GSE13904, GSE26378, and GSE26440 datasets were used to examine BST1 expression in pediatric sepsis and to evaluate the suitability of BST1 for the diagnosis of pediatric sepsis using receiver operating characteristic curves. Additionally, the usefulness of diagnostic BST1 for identifying pediatric sepsis in actual clinical samples was evaluated by measuring serum BST1 levels using the enzyme-linked immunosorbent assay (ELISA) on 30 normal samples and 35 pediatric sepsis samples. Furthermore, a more thorough analysis of the connection between immune cells and BST1 expression was conducted utilizing the CIBERSORT database. Finally, we used ELISA to quantify myeloperoxidase (MPO)-DNA complex levels to assess the establishment of NETs. Integrating machine learning approaches, we identified BST1 as the most robust NET-associated diagnostic biomarker for pediatric sepsis. Comparative analysis revealed a significant elevation of BST1 expression in septic patients versus healthy controls, with pronounced discriminative capacity in a 65 clinical samples cohort (AUC value = 0.873), outperforming the effects of conventional inflammatory indicators such as procalcitonin (PCT), white blood cell count (WBC), and C-reactive protein (CRP). Lastly, we discovered a substantial correlation between the BST1 level and the MPO-DNA complex level in the serum of pediatric sepsis children. Conclusion: Serum BST1 has the potential to be an effective diagnostic biomarker for pediatric sepsis. What is Known: • Pediatric sepsis is a major cause of morbidity, death, financial burden, and negative effects on health systems throughout the world. What is New: • Serum BST1 has the potential to be an effective diagnostic biomarker for pediatric sepsis.