Abstract
BACKGROUND: Mounting evidence indicates that lung adenocarcinoma (LUAD) patients are at elevated risk for venous thromboembolism (VTE), presenting a major clinical challenge. This study utilized public databases to identify crosstalk genes (CGs) between LUAD and VTE, applied machine learning methods to discover shared diagnostic biomarkers, and explored their underlying mechanisms. METHODS: Disease-specific genes for VTE were extracted from comprehensive genomic databases (CTD, DisGeNET, GeneCards, OMIM), while transcriptomic profiles of LUAD and VTE cohorts were retrieved from GEO via GEOquery implementation. Molecular crosstalk analysis identified candidate genes through differential expression algorithms and disease-association metrics. Functional annotation employed GO and KEGG analyses to elucidate the biological significance of identified CGs. LASSO regression analysis of VTE and LUAD matrices yielded overlapping diagnostic biomarkers. Immune contexture was characterized via CIBERSORT deconvolution, followed by correlation analyses between hub genes and immune infiltration profiles. Hub genes expression was corroborated through independent cohort validation and serological quantification. Diagnostic utility was evaluated through receiver operating characteristic (ROC) curve and nomogram. Therapeutic potential was assessed via DSigDB-based drug sensitivity profiling. RESULT: Through transcriptomic analysis, we identified 381 CGs, which demonstrated significant enrichment in inflammatory cascades, immunological processes, and coagulation pathways. LASSO regression analysis of LUAD and VTE cohorts revealed TIMP1 and TMEM132A as putative shared diagnostic biomarkers. TMEM132A exhibited significant correlation with immune cell infiltration patterns across both diseases, modulating the immune microenvironment. Validation cohorts and serological assessment confirmed elevated TMEM132A expression in LUAD and LUAD combined with VTE phenotypes. The diagnostic accuracy of TMEM132A was substantiated by ROC curves and nomogram analyses. Pharmacological sensitivity analysis indicated that TMEM132A may serve as a potential target for the therapeutic agents birabresib and abemaciclib. CONCLUSION: TMEM132A demonstrates diagnostic utility as a predictive biomarker for VTE occurrence in LUAD, suggesting its potential role as a susceptibility gene in this patient cohort.