Identification of endometrial cancer biomarkers using weighted gene coexpression network analysis and machine learning.

BACKGROUND: Endometrial carcinoma (UCEC) exhibits a rising incidence in China, imposing a substantial burden on both women and society. Identifying biomarkers for UCEC is critical for precise diagnosis and treatment. METHODS: RNA-seq data for UCEC and normal endometrial tissues were sourced from the GEO and TCGA databases. The Limma package was used to analyze differentially expressed genes (DEGs) from GEO datasets, while weighted gene co-expression network analysis (WGCNA) was employed to identify gene modules associated with UCEC. Machine learning algorithms (GBM, KNN, LASSO, SVM, NNET, RF, DT, GLM) were subsequently applied for further gene screening. Key biomarkers were identified by analyzing overall survival (OS) and progression-free survival (PFS) using TCGA-UCEC data on the Xiantao Academic online analysis platform.Functional enrichment analysis, clinical significance assessment, COX regression analysis, prognostic nomogram construction, single-gene logistic regression analysis, potential mechanism exploration, and immune characterization were all conducted via the Xiantao Academic online analysis platform. Drug sensitivity profiling was performed using the CPADS database, followed by molecular docking validation. In vitro functional assays included the EdU assay for evaluating cell proliferation and the Transwell assay for assessing cell invasion. RESULTS: Kaplan-Meier survival analysis showed that both overall survival (OS) and progression-free survival (PFS) were significantly shorter in the high-MAL-expression group than in the low-expression group, suggesting a close association between high MAL expression and poor prognosis. Further multivariate Cox regression analysis, which included clinical stage and tumor grade, revealed that the independent predictive value of MAL for OS was attenuated, indicating that its prognostic significance may be partially confounded by traditional clinicopathological factors. MAL may serve as a potential biomarker in UCEC patients, with its high expression promoting tumor progression, correlating negatively with prognosis, and modulating immune characteristics. In vitro experiments demonstrated that elevated MAL expression promoted the proliferation and invasion of ECC-1 cells, while MAL knockdown suppressed these phenotypes. CONCLUSION: This study found that high MAL expression is significantly correlated with poor prognosis in endometrial cancer patients, which is consistent with its functional role in promoting tumor cell proliferation and invasion. However, after adjusting for clinical factors such as stage and grade, the independent prognostic effect of MAL on overall survival was not prominent, suggesting that it may more likely reflect the extent of tumor progression and adverse clinical features, rather than serving as a sole determinant of prognosis. Therefore, MAL can be regarded as a molecular marker closely associated with disease progression and survival outcomes, and it holds potential for providing auxiliary value in risk stratification and personalized management when combined with clinical parameters.