Exploring radiation resistance-related genes in pancreatic cancer and their impact on patient prognosis and treatment

Abstract

Background: Pancreatic cancer is a highly lethal disease with increasing incidence worldwide. Despite surgical resection being the main curative option, only a small percentage of patients are eligible for surgery. Radiotherapy, often combined with chemotherapy, remains a critical treatment, especially for locally advanced cases. However, pancreatic cancer's aggressiveness and partial radio resistance lead to frequent local recurrence. Understanding the mechanisms of radiotherapy resistance is crucial to improving patient outcomes.

Methods: Pancreatic cancer related gene microarray data were downloaded from GEO database to analyze differentially expressed genes before and after radiotherapy using GEO2R online tool. The obtained differentially expressed genes were enriched by GO and KEGG to reveal their biological functions. Key genes were screened by univariate and multivariate Cox regression analysis, and a risk scoring model was constructed, and patients were divided into high-risk group and low-risk group. Subsequently, Kaplan-Meier survival analysis was used to compare the survival differences between the two groups of patients, further analyze the differential genes of the two groups of patients, and evaluate their sensitivity to different drugs.

Results: Our model identified 10 genes associated with overall survival (OS) in pancreatic cancer. Based on risk scores, patients were categorized into high- and low-risk groups, with significantly different survival outcomes and immune profile characteristics. High-risk patients showed increased expression of pro-inflammatory immune markers and increased sensitivity to specific chemotherapy agents, while low-risk patients had higher expression of immune checkpoints (CD274 and CTLA4), indicating potential sensitivity to targeted immunotherapies. Cross-dataset validation yielded consistent AUC values above 0.77, confirming model stability and predictive accuracy.

Conclusion: This study provides a scoring model to predict radiotherapy resistance and prognosis in pancreatic cancer, with potential clinical application for patient stratification. The identified immune profiles and drug sensitivity variations between risk groups highlight opportunities for personalized treatment strategies, contributing to improved management and survival outcomes in pancreatic cancer.

Keywords: immune microenvironment; pancreatic cancer; personalized immunotherapy; prognostic scoring model; radiotherapy resistance.

Figure 1

Identification and functional analysis of DEGs related to RT resistance in PAAD. (A) Volcano plots showing gene expression changes in PAAD tissues after RT compared to before RT in the GSE179351 datasets, with |log₂FC| ≥1 and p < 0.05 set as cutoff values. In the DEGs analysis, red dots represent upregulated genes, and blue dots represent downregulated genes. (B) Volcano plots showing gene expression changes in PAAD tissues after RT compared to before RT in the GSE225767 datasets, with |log₂FC| ≥1 and p < 0.05 set as cutoff values. In the DEGs analysis, red dots represent upregulated genes, and blue dots represent downregulated genes. (C) Venn diagram of upregulated genes in GSE179351 and GSE225767. (D) Venn diagram of downregulated genes in GSE179351 and GSE225767. (E) Enrichment analysis of 121 co-upregulated genes and 27 co-downregulated genes performed using the Metascape website.

Figure 2

Prognostic model for PAAD Patients based on radiation-resistant genes. (A) Multivariate Cox regression analysis identified 10 genes associated with OS in patients with PAAD. (B) Optimal cutoff value of risk score determined by the “surv_cutpoint” function. 178 PAAD patients were divided into high-risk and low-risk groups. (C) Heat map displaying the expression levels of the 10 key genes in individual patients. (D) The Kaplan-Meier OS curve shows the survival differences among patients in different risk groups. (E) The AUC values corresponding to these gene combinations were calculated by multiple logistic regression model. The AUC value is 0.7664. (F) Bar plot showing the variable significance of 10 filtered genes in random forest model.

Figure 3

Validation of the scoring model’s predictive performance and clinical relevance. (A) ROC curves for three GEO datasets. (B) Risk scores compared by survival status and age group. (C) Sankey diagram depicting relationships among gender, risk group, and cancer stage. *p < 0.05; **p < 0.01.

Figure 4

DEGs and functional enrichment analysis in high- and low-risk groups. (A) Volcano plot of 933 DEGs with 348 upregulated (red) and 585 downregulated (blue) genes. (B) GO enrichment of upregulated genes. (C) KEGG enrichment of upregulated genes in pathways such as pancreatic secretion, neuroactive ligand-receptor interaction, and protein digestion, supporting tumor growth and metabolic demands in high-risk patients. (D) GO enrichment of downregulated genes. (E) KEGG enrichment of downregulated genes in pathways like neuroactive signaling, cytoskeletal organization, and metabolism, indicating reduced proliferation and migration potential in low-risk patients.

Figure 5

Survival analysis of immune-related DEGs, immune function, and immune infiltration differences between high-risk and low-risk PAAD patients. (A) Kaplan-Meier survival curves for 11 immune-related DEGs significantly associated with OS in PAAD patients. (B) Comparison of immune function scores between high-risk (purple) and low-risk (orange) groups. (C) Immune cell infiltration scores comparing high-risk and low-risk groups. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 6

Immune checkpoint and drug sensitivity analysis. (A) Comparison of CD274 expression between high-risk (red) and low-risk (blue) groups. (B) Comparison of CTLA4 expression between high-risk (red) and low-risk (blue) groups. (C) TIDE scores between low-risk (blue) and high-risk (red) groups. (D) Drug sensitivity analysis between high-risk (red) and low-risk (blue) groups across multiple anti-cancer drugs. ***p < 0.001.