Independent validation of lung adenocarcinoma prognostic risk scores incorporating cholesterol and estrogen metabolism related transcriptional biomarkers

Abstract

Abstract We investigated the prognostic significance of cholesterol metabolism-related genes (CMRGs) and estrogen metabolism-related genes (EMRGs) in lung adenocarcinoma (LUAD). Transcriptomic and clinical data from the cancer genome atlas (TCGA) and gene expression omnibus (GEO) databases were analyzed. Iterative sure independence screening and least absolute shrinkage and selection operator (ISIS-LASSO) Cox regression identified prognostic CMRGs and EMRGs. Two risk scores—Cholescore and Estrogenscore—were constructed and validated. High Cholescore was associated with poor overall survival (OS), reduced immune infiltration, low immune checkpoint expression, and high tumor purity, suggesting an immunosuppressive tumor microenvironment. Estrogenscore also independently predicted LUAD prognosis. Mediation analysis revealed that estrogen-related pathways partially mediated the impact of cholesterol metabolism on prognosis. A significant interaction between Cholescore and Estrogenscore was identified, and patients with both high scores had the worst OS and lowest predicted immunotherapy benefit. Combining the two scores significantly improved the area under the curve (AUC) for 1–5 years OS prediction. These findings suggest that integrating cholesterol and estrogen metabolism signatures can improve LUAD prognostic stratification and provide molecular insights into tumor–immune interactions and immunotherapy response prediction.