基于免疫相关基因的子宫内膜癌预后模型构建

Construction of prognostic model for endometrial cancer based on immune-related genes

目的 建立子宫内膜癌免疫相关基因预后模型,为子宫内膜癌免疫治疗提供参考。方法 从TCGA数据库下载子宫内膜癌转录组数据和临床信息,免疫相关基因列表从IMMPORT数据库获得。利用差异分析、功能富集分析评估差异免疫相关基因的功能,在训练队列中利用单因素、多因素Cox回归和最小绝对收缩和选择算法(least absolute shrinkage and selection operator,LASSO)构建预后评估模型,利用时间依赖的受试者工作特征(receiver operating characteristic,ROC)曲线、校准图、决策曲线(decision curve analysis,DCA)评估模型的效能。结果 本研究分析了504例子宫内膜癌组织和22例癌旁组织的免疫相关基因表达谱,获得394个差异表达基因,其中上调基因111个,下调基因283个;功能富集分析发现,差异基因主要参与免疫应答、信号转导和免疫反应等生物过程。在训练队列中通过单因素Cox回归分析确定了31个预后相关基因,通过LASSO-Cox回归建立了6个免疫基因(ANGPT4、HGF、PLA2G2A、SPAG11B、SST、VGF)的预测模型。该模型在训练队列中的1年、3年和5年ROC曲线下面积分别为0.986、0.846和0.849;在验证队列中分别为0.667、0.629和0.693;在整体队列中分别为0.798、0.735和0.770。结论 本研究基于6个免疫相关基因建立的预后模型可有效预测子宫内膜癌患者预后,为免疫治疗提供参考。

ObjectiveTo establish a prognostic model of immune-related genes in endometrial cancer to provide the reference for the immunotherapy of endometrial cancer.MethodsEndometrial cancer transcriptome data and clinical information were downloaded from TCGA database,and the immune-related genes were acquired from IMMPORT database.Subsequently,the differential analysis and functional enrichment analysis were used to evaluate the function of differential expression immune-related genes.The prognostic assessment models was constructed in training sets using univariable,multivariable Cox regression and least absolute shrinkage and selection operator(LASSO).The receiver operating characteristic(ROC) curve,calibration plot,decision curve analysis(DCA) curve were used to evaluate the performance of models.ResultsThe immune-related gene expression profiles for 504 cases of endometrial cancer and 22 cases of adjacent tumor tissues were analyzed,and 394 differentially expressed genes were obtained,including 111 up-regulated genes and 283 down-regulated genes.Functional enrichment analysis showed that differential genes were mainly involved in biological processes such as immune response,signal transduction and immune response.A total of 31 prognosis-related genes were identified in the training cohort by the univariable Cox regression analysis.The predictive model of 6 immune-related genes(ANGPT4,HGF,PLA2G2A,SPAG11B,SST,VGF) were constructed by LASSO-Cox regression.The area under the ROC curve of the model in the training cohort were 0.986,0.846 and 0.849 at 1,3 and 5 years,respectively;the area under the curve of the model at 1,3 and 5 years were0.667,0.629 and 0.693,respectively,in the validation cohort;and the area under the curve of the model at 1,3 and 5 years were 0.798,0.735 and 0.770,respectively,in total cohort.ConclusionsThe prognostic model is developed based on 6 immune-related genes,and can be effectively used to predict the prognosis of patients with endometrial cancer,providing a reference for immunotherapy.