Identification of a m6A-immune-related risk model for predicting prognosis, immune microenvironment, and drug responses in acute myeloid leukemia

Abstract

Abstract This study utilized TCGA database to explore the role of m6A modification and immune infiltration in AML. Through unsupervised clustering and WGCNA analysis, 8 hub genes were identified, and a risk model with EHBP1L1 and ZNF385A was established using LASSO regression. A nomogram incorporating hub gene risk score and age showed satisfactory prognostic prediction. External validation of GEO confirmed the model’s effectiveness. TME analysis revealed correlations with monocytes and Treg cells, while immune checkpoints and HLA genes were associated with risk scores. Drug sensitivity analysis suggested potential responses to specific chemotherapy drugs. TIDE analysis indicated reduced ICI treatment benefit in high-risk patients. RT-qPCR validations revealed the significance of prognosis and risk stratification of ZNF385A . The noticeable trend of EHBP1L1 was observed. In addition, the accurate predictive capability of the risk model has been validated by clinical samples. Therefore, the risk model enables a quantitative evaluation of disease severity and progression risk in AML patients, based on their clinical and biological characteristics. This precise prediction not only informs treatment decisions but also guides the selection of chemotherapy regimens, overall improving patient outcomes.