Background: Metabolic reprogramming, particularly toward oxidative phosphorylation (OXPHOS), is a hallmark of lung squamous cell carcinoma (LUSC) and contributes to its aggressive phenotype and immunosuppressive microenvironment. While OXPHOS activation is increasingly recognized as a key metabolic feature in LUSC, its prognostic implications and associated gene signatures remain underexplored. This study aimed to identify OXPHOS-related differentially expressed genes (DEGs) and construct a robust prognostic signature for LUSC.
Methods: Using GEO datasets, we developed an OXPHOS-related prognostic signature via ssGSEA, differential analysis, and LASSO-Cox regression.
Results: An 8-gene OXPHOS-related signature (LTBP1, MFGE8, ACTN1, CD59, CDC25C, SAAL1, SFXN4, PTTG1) was identified. High-risk patients exhibited significantly shorter overall survival than low-risk patients across all cohorts. The model demonstrated strong predictive accuracy for 1-, 3-, and 5-year survival. Notably, the high-risk group showed enriched pathways related to tumor stemness and immunosuppression.
Conclusion: We developed and validated a novel OXPHOS-based gene signature that effectively stratifies LUSC patients by risk. This signature highlights the clinical relevance of OXPHOS in LUSC prognosis and may guide personalized therapeutic strategies targeting metabolic vulnerabilities. Study limitations include its retrospective design and lack of experimental validation.
Keywords: LASSO regression; Lung squamous cell carcinoma; bioinformatics; gene expression; nomogram; oxidative phosphorylation; prognostic signature; survival analysis.
A novel prognostic signature based on oxidative phosphorylation-related genes was constructed specifically for lung squamous cell carcinoma (LUSC).Gene expression data from four GEO datasets (GSE157011, GSE30219, GSE37745, GSE42127) were integrated and analyzed using ssGSEA, univariate Cox regression, and LASSO regression methods.Eight signature genes (LTBP1, MFGE8, ACTN1, CD59, CDC25C, SAAL1, SFXN4, and PTTG1) were identified and used to develop a risk score model that effectively stratified patients by survival outcomes.The model demonstrated robust predictive performance across training, internal testing, and external validation cohorts, with AUC values above 0.60 for 1-, 3-, and 5-year survival.Subgroup analyses confirmed the prognostic value of the signature across various clinical characteristics including age, gender, and tumor stage.A nomogram incorporating the risk score and clinical features showed good calibration and may support individualized prognosis estimation in clinical practice.