基于免疫相关基因的肾透明细胞癌预后分子标签的建立与验证

Establishment and validation of prognosis signature of renal clear cell carcinoma based on immune related genes

目的通过挖掘TCGA数据库中肾透明细胞癌(ccRCC)的免疫相关基因数据,构建ccRCC预后分子标签。方法从TCGA数据库中提取611例的RNA-Seq测序数据(正常样本72例,ccRCC样品539例)和530例ccRCC患者的临床数据,筛选出免疫相关基因(IRGS)表达数据,并计算正常肾组织与ccRCC组织表达差异的IRGS。在530例ccRCC肿瘤组织样品的IRGS表达数据中,随机抽取70%的样品应用LASSO-Cox回归构建预后分子标签(训练组),30%的样品验证构建的IRGS预后分子标签(测试组)。结果在ccRCC组织中筛选出差异表达的IRGS共775种。单因素Cox分析得到298种IRGS与ccRCC预后有相关性(P<0.05)。采用LASSO-Cox回归分析中筛选出12个与ccRCC预后相关的IRGS,分子标签值在训练组、测试组、所有样本数据与ccRCC患者的生存期呈显著相关,分子标签值越高患者的预后越差,其中所有样本数据中的1、3、5年受试者工作特征曲线(ROC)的曲线下面积分别为0.797、0.763、0.780。结论ccRCC的IRGS预后模型可用于预测ccRCC患者的预后,有利于临床治疗。

Objective To construct the prognostic molecular signature of renal clear cell carcinoma(ccRCC)by mining the data of immune related genes in TCGA database,so as to provide a new prognostic model for precise treatment of ccRCC.Methods The RNA-Seq sequencing data of 611 cases(72 normal samples,539 ccRCC samples)and the clinical data of 530 ccRCC patients were extracted from the TCGA database,and the expression data of Immune related genes(IRGS)were screened.IRGS of expression difference between normal kidney tissue and ccRCC tissue were calculated.Among 530 ccRCC tumor tissue samples with IRGS expression data,70%of the samples were randomly selected to construct the prognostic molecular label by LASSO-Cox regression(training group),and 30%of the samples were validated to construct the prognostic molecular label(test group).Results A total of 775 differentially expressed IRGS were identified in ccRCC tissues.Univariate Cox analysis showed that 298 IRGS were correlated with the prognosis of ccRCC(P<0.05).Twelve IRGS associated with the prognosis of ccRCC were screened out by LASSO-Cox regression analysis,which were ANGPTL3,APLNR,CCL22,CLDN4,FGF17,HCST,IL4,MIA,OSM,R3HDML,SEMA3A and SLPI.The molecular label value in the training group,the test group,and all sample data were significantly correlated with the survival of ccRCC patients,and the higher the molecular label value was,the worse the prognosis was.The area under the curve of the 1-year,3-year,and 5-year subjects′operating characteristic curves in all sample data were 0.797,0.763 and 0.780,respectively.Conclusions The IRGS prognostic model of ccRCC can be used to predict the prognosis of patients with ccRCC,which is beneficial to further guide clinical treatment.