Construction of a prognostic prediction model for concurrent radiotherapy in cervical cancer using GEO and TCGA databases with preliminary validation analysis.

INTRODUCTION: Radiotherapy is a primary treatment for intermediate and advanced cervical cancer (CC). Resistance to radiotherapy is a principal reason for treatment failure in synchronous applications, yet the molecular mechanisms remain poorly understood. Identifying reliable prognostic markers to predict and evaluate patient outcomes is essential for advancing therapeutic strategies. This study aims to address this need by developing a prognostic prediction model for concurrent radiotherapy in CC, utilizing both single-cell RNA sequencing (scRNA-seq) and bulk RNA sequencing data. METHODS: The research began by screening for co-expressed genes using samples from two GEO datasets (GSE236738 and GSE56363). To pinpoint target genes that exhibit significant co-expression, both univariate and multivariate Cox regression analyses were conducted, facilitating the development of prognostic prediction models. The clinical significance of these models was confirmed through the analysis of 144 CC samples sourced from the TCGA database, utilizing Kaplan-Meier survival curves, ROC curve analyses, and Spearman's correlation tests to investigate the relationships between gene expression and the levels of immune cell infiltration. IHC assays were conducted to further validate the prognostic potential of the identified target genes in CC patients. RESULTS: Prognostic models for four target genes-MPP5, SNX7, LSM12, and GALNT3-showed significant predictive value for the outcomes of CC patients undergoing radiotherapy, as demonstrated using the GSE236738 and GSE56363 datasets. The prognostic efficacy of the model was illustrated through scatter plots and calibration curves. Additionally, the model exhibited significant associations with tumor immune infiltration, immune checkpoints, and chemotherapeutic drug sensitivity. Immunohistochemistry (IHC) on clinical tumor samples confirmed that the protein expression levels of MPP5, SNX7, LSM12, and GALNT3 were distinctively predictive for CC patients. CONCLUSION: The results indicate that MPP5, SNX7, LSM12, and GALNT3 are significantly associated with radiotherapy sensitivity in CC cells. A prognostic risk model based on these genes demonstrated strong predictive capabilities for patient outcomes in radiotherapy, suggesting these genes as effective predictors and potential therapeutic targets for treating CC.