Seven-gene biomarkers reveal prognostic and immune signatures in lung adenocarcinoma.

Reliable prognostic biomarkers for lung adenocarcinoma (LUAD) remain limited because of inter-cohort heterogeneity across transcriptomic studies. This study aimed to identify survival-related gene signatures associated with immune characteristics in LUAD. Three independent Gene Expression Omnibus (GEO) datasets were analyzed separately to identify consistently dysregulated genes in LUAD. Functional enrichment, protein-protein interaction network, survival, receiver operating characteristic (ROC) curve analysis, principal component analysis (PCA), expression validation, and immune infiltration analyses were performed. A total of 68 overlapping differentially expressed genes were identified. Functional enrichment analysis showed that these genes were mainly involved in vascular-related biological processes. Network and survival analyses identified seven significantly downregulated genes, including AGER, CAV1, EDNRB, ROBO4, EMCN, TEK, and PTPRB, which were associated with overall survival in LUAD. ROC analysis showed favorable diagnostic performance across the three GEO datasets, with area under the curve (AUC) values ranging from 0.873 to 1.000. In the larger datasets, AUCs ranged from 0.932 to 0.952 in GSE19188 and from 0.873 to 0.949 in GSE30219, with corresponding 95% CI ranges of 0.870-1.000 and 0.710-1.000, respectively. PCA further supported separation between LUAD and normal samples. Immune infiltration analysis showed associations between the seven-gene signature, tumor purity, and multiple immune cell populations. A seven-gene vascular-related signature was associated with prognosis and immune infiltration patterns in LUAD. These findings may support future biomarker development and studies of tumor microenvironment remodeling in LUAD.