Abstract
The unfolded protein response (UPR) plays a pivotal role in the pathogenesis and progression of acute myeloid leukemia (AML). This study aims to investigate the prognostic value of UPR-related genes (URGs) and establish a UPR-related gene signature (URGsig) to enhance prognosis prediction and guide therapeutic decision-making in AML. Gene expression profiles of AML patients were obtained from the GDC and GEO databases. Cox regression and LASSO regression analyses were applied to identify key genes for the construction of URGsig. Comprehensive bioinformatics analyses were conducted to elucidate the biological and clinical implications of URGsig. A nomogram integrating URGs and clinical prognostic features was developed to predict survival probability for AML patients. Additionally, the differential expression of core genes within the URGsig was validated in clinical samples. Notably, two distinct UPR-related subtypes were identified, and they displayed significant heterogeneity in clinical outcomes and tumor microenvironment (TME). The URGsig, comprising six URGs, showed a strong correlation with survival outcomes and exhibited robust predictive capabilities. Importantly, patients categorized into the high-risk subgroup based on URGsig were predicted to show lower chemosensitivity but a better response to immunotherapy. The nomogram performed well in prognosis prediction, with an area under the curve (AUC) of 0.912 for 5-year overall survival. In summary, our findings highlight the URGsig as a promising prognostic biomarker and provide novel insights into the mechanism by which UPR influences the immune landscape of AML. This paradigm may lay a foundation for the development of personalized treatment strategies for AML patients.