Machine Learning-Based Identification of a Neutrophil Extracellular Traps-Derived Signature for Improving Outcomes and Therapy Responses in Patients with Glioma

Abstract

Neutrophils, a cellular component within glioma, contribute to its heterogeneous microenvironment and affect clinical outcomes. In this study, our objective was to develop a signature associated with neutrophil extracellular traps (NETs) to stratify prognosis and predict the efficacy of immunotherapy in glioma. We employed expression data and clinical information from public databases to establish a prognostic risk model using LASSO and machine learning algorithms. The NETs-based signature not only exhibited superior accuracy compared to published signatures but also functioned as an independent prognostic factor. Additionally, we observed increased immune and stromal infiltration in high-risk patients and identified a correlation between NETs-related gene expression and drug sensitivity. We further pinpointed MMP9 as a hub gene linked to the prognostic model. Immunohistochemistry validation confirmed the unfavorable prognosis association of MMP9. Taken together, these findings highlight the potential of the NETs-based signature to facilitate precise patient stratification and offer guidance in making decisions regarding immunotherapy and chemotherapy for glioma.