Abstract
INTRODUCTION: Acute myeloid leukemia (AML) prognosis remains challenging due to limited biomarkers integrating tumor microenvironment (TME) dynamics. Neutrophils, key mediators of immune regulation, exhibit dual roles in cancer progression, yet their prognostic significance in AML is poorly defined. This study aimed to construct a neutrophil-related gene signature for AML risk stratification and explore its clinical and immunological implications. METHODS: Utilizing transcriptomic and clinical data from TCGA (The Cancer Genome Atlas), GEO (Gene Expression Omnibus), and OHSU cohorts (n=1537), we identified 148 neutrophil-related genes through literature mining. Prognostic genes were selected via univariate Cox regression and LASSO regression (R packages: survival, glmnet). A 5-gene model (CSF3R, BRAF, FFAR2, CD300A, CD37) was validated across internal (TCGA) and external cohorts (GSE10358, GSE14468, OHSU). Immune profiling, drug sensitivity analysis (GDSC database), and TIDE scoring were performed to assess immunotherapy relevance. RESULTS: The neutrophil-based model stratified AML patients into high- and low-risk groups with distinct overall survival (OS, p<0.0001 in TCGA). Multivariate Cox analysis confirmed its independence from age, FLT3, and TP53 mutations (HR=2.14, p=0.015). CD37 emerged as the strongest prognostic marker (AUC 5-year=0.680, p=0.0026), correlating with immunosuppressive TME features: elevated myeloid-derived suppressor cells (MDSCs, p<0.01), Treg infiltration (p <0.05), and upregulated immune checkpoints (PD1, CTLA4, LAG3; p<0.001). High CD37 expression predicted immunotherapy responsiveness (TIDE score, p=0.004) and interacted with 146 potential therapeutic agents (eg, BCL2 inhibitors). DISCUSSION: This study advances a novel 5-gene prognostic model integrating neutrophil biology into AML risk stratification. CD37, a key regulator of immune evasion, serves as a dual biomarker for prognosis and immunotherapy prediction. While validated across multiple cohorts, experimental studies are warranted to unravel CD37's mechanistic role. Our findings highlight the potential of neutrophil-centric biomarkers in guiding personalized AML therapy.