Comprehensive analysis of potential biomarkers for the diagnosis and prognosis of Cervical squamous cell carcinoma -based on GEO and TCGA databases

Yufen Chen¹Qinghua Deng²Tengyue Fu³Yuxiang Huang⁴Houlin Li⁴Jingmu Xie⁴Feng Liao⁴Feimiao Zeng⁵Xingyi Fang^6*Ruiman Li^7*Zhuming Chen^4*

• ¹The First Clinical Medical School of Jinan University, Guangzhou, China
• ²Department of Gynaecology, The Second Affiliated Hospital of Guangdong Medical University, Zhanjiang, China
• ³Department of Neurosurgery, The First Affiliated Hospital of Shantou University Medical College, Shantou, China
• ⁴Guangdong-Hongkong-Macau Institute for CNS Regeneration, Jinan University, Guangzhou, Guangdong Province, China
• ⁵Diagnostic Pathology Center，The Second Affiliated Hospital of Guangdong Medical University, Zhanjiang, China
• ⁶Reproductive Medical Center, Affiliated Hospital of Guangdong Medical University, Zhanjiang, China
• ⁷Department of Obstetrics and Gynecology, First Affiliated Hospital of Jinan University, Guangzhou, China

Background: Cervical squamous cell carcinoma (CESC) constitutes a substantial global health burden, especially in resource-limited regions. The identification of reliable biomarkers is critical for developing a clinically applicable nomogram to predict survival outcomes and evaluate immune infiltration in CESC patients.Methods: This study integrated RNA-seq data from GEO and TCGA databases to identify key genes associated with CESC through differential expression analysis and machine learning techniques. Prognostic models were constructed and validated, with additional analyses exploring immune cell infiltration and gene function via GSEA and clinical correlation. Finally, key genes were validated via qRT-PCR in CESC tissues.Results: A total of 112 differentially expressed genes (DEGs) were identified through differential analysis of the GEO and TCGA datasets. EFNA1, CXCL8, and PPP1R14A emerged as prognostic biomarkers for CESC, showing significant associations with survival, tumor stage, and immune infiltration. EFNA1 may drive tumor progression via the MAPK signaling pathway, CXCL8 could influence immune evasion through NOD-like receptor signaling, and PPP1R14A may contribute to tumor invasion by modulating extracellular matrix remodeling. A nomogram integrating these genes demonstrated high predictive accuracy for overall survival (AUC>0.75) and calibration plots. Decision curve analysis (DCA) was performed to assess the nomogram's clinical utility and net benefit for application in clinical practice. Additionally, it was validated by qRT-PCR, showing elevated expression in tumors versus normal tissues (P<0.05). Conclusion: EFNA1, CXCL8, and PPP1R14A are promising biomarkers for CESC prognosis and immune regulation. The nomogram model provides a practical tool for personalized survival prediction, enhancing clinical decision-making for immunotherapy and risk stratification.