空军军医大学学报2025，Vol.46Issue(6)：785-792,8.DOI:10.13276/j.issn.2097-1656.2025.06.013
胰腺癌吉西他滨治疗反应性预测模型建立及免疫综合分析
Establishment of a predictive model for gemcitabine response in the treatment of pancreatic cancer and a comprehensive analysis of immunity
摘要
Abstract
Objective To establish a predictive model for gemcitabine(GEM)response and a model for tumor immune proportion score(IPScore)in pancreatic cancer based on TCGA sequencing data and investigate the impact of tumor immune status on GEM response in the treatment of pancreatic cancer,thereby guiding the screening of GEM-sensitive patients and the application of GEM-based combined therapies.Methods Transcriptome and GEM-response data of patients with pancreatic cancer from TCGA database were used to identify differentially expressed genes(DEGs)related to GEM response.Prognosis-related DEGs were screened via univariate Cox regression analysis,and least absolute shrinkage and selection operator(LASSO)regression was used to further screen and establish a risk score model(GEMScore value).According to GEMScore value,patients were divided into high GEMScore group and low GEMScore group,and gene set enrichment analysis(GSEA)was used to analyze the functional distribution of DEGs in tumor samples of the two groups.The ESTIMATE,tumor stemness,and ssGSEA were used to evaluate the differences in tumor immune status between the two groups.Genes related to GEM response were screened from the immune gene set,an IPScore model was constructed,and the relationship between tumor immune status and GEM treatment response was analyzed using the above two models.Results A total of 1 083 DEGs were identified in pancreatic cancer tissue samples from the GEM-responsive and non-responsive groups(P＜0.05),and 29 prognosis-related DEGs were identified by univariate Cox regression analysis.An 8-gene(XDH,CLDN1,SIX2,SUSD2,FOXP3,ADAMTS8,CCL23,and PRUNE2)predictive model for GEM treatment response was established using LASSO regression,and its predictive capability was validated via ROC analysis.GSEA analysis revealed that GEM treatment response was associated with the immune status of the tumor microenvironment(P＜0.01).Compared to the low GEMScore group,the high GEMScore group exhibited higher tumor purity(P＜0.01)and tumor stemness scores(P＜0.01),but with lower infiltration of immune and stromal cells(P＜0.01).The combined analysis of GEMScore and IPScore showed a negative correlation between GEMScore value and IPScore value(R=-0.45,P＜0.001),namely,lower IPScore value was associated with higher GEMScore value,indicating poorer GEM treatment response.Conclusion This study establishes a multi-gene predictive model for GEM response and a model for tumor IPScore,which systematically evaluates the relationship between tumor immune status and GEM treatment response in patients with pancreatic cancer,providing a basis for identifying GEM-sensitive patients and selecting individualized therapeutic strategies.关键词
胰腺癌/吉西他滨/治疗反应性/预测模型/肿瘤免疫微环境/疾病预后Key words
pancreatic cancer/gemcitabine/treatment response/predictive model/tumor immune microenvironment/disease prognosis分类
医药卫生引用本文复制引用
田也,马庆智,汪浩,陈志南,王珂..胰腺癌吉西他滨治疗反应性预测模型建立及免疫综合分析[J].空军军医大学学报,2025,46(6):785-792,8.基金项目
国家自然科学基金(82273226) （82273226）
