基于TCGA数据库构建和评估口腔癌的免疫相关预后模型

Construction and Evaluation of Immune Related Prognosis Model of Oral Cancer Based on TCGA Database

目的构建和评估口腔癌免疫相关的预后模型,探索口腔癌肿瘤免疫微环境的分子特性。方法基于TCGA数据库中HNSC队列的mRNA表达数据、临床信息和ImmPort数据库中的免疫相关基因列表,利用R包limma分析得到免疫相关的差异表达基因列表,通过单因子回归和多因子回归分析构建口腔癌的免疫相关预后模型,并进一步结合临床特征构建列线图模型,综合评估预后模型的性能。结果共得到1533个差异表达基因,其中73个基因与免疫相关,通过单因子回归和多因子回归分析得到6个免疫相关的差异基因与预后相关;构建的预后模型ROC曲线下的面积在3年时为0.678,4年时为0.671,5年时为0.683;构建的列线图模型的C-index从0.63增加至0.67,基于建立的预后风险模型,高风险评分组的预后差于低风险评分组,差异有统计学意义(P<0.05);预后风险评分与N分期相关(P<0.05),与病理分期和T分期无关(P>0.05);高风险评分组的基因MUC16突变率偏高,而NOTCH1则相反(P<0.05),且高风险评分组的免疫治疗相关靶基因的表达水平偏低(P<0.05);较高的TMB与较差的预后相关(P<0.05)。结论本次构建的预后模型和对肿瘤免疫微环境的分子特性的探索可能有助于口腔癌患者的预后风险预测和免疫治疗。

Objective To construct and evaluate an immune-related prognostic model for oral cancer and explore the molecular characteristics of tumor immune microenvironment.Methods Based on the mRNA expression data and clinical information of HNSC cohort in TCGA database and the list of immune-related genes in ImmPort database,the list of immune-related differentially expressed genes was obtained by R-packet limma analysis.The immune-related prognosis model of oral cancer was constructed by single factor regression and multi-factor regression analysis,and the line graph model was further constructed by combining the clinical characteristics to comprehensively evaluate the performance of the prognosis model.Results A total of 1533 DEGs were obtained,of which 73 genes were immune related.Six immune-related DEGs were significantly correlated with prognosis by univariable Cox regression and multivariate Cox regression analysis(P<0.05).The area under the ROC curve of the constructed prognostic model was 0.678 at 3 years,0.671 at 4 years and 0.683 at 5 years.The C-index of the constructed nomogram model increased from 0.63 to 0.67.Based on the established prognostic risk model,the prognosis of the high-risk score group was worse than that of the low-risk score group(P<0.05).The prognostic risk score was significantly correlated with N stage(P<0.05),but not correlated with pathological stage and T stage(P>0.05).In addition,the mutation rate of gene MUC16 in the high-risk score group was significantly higher,while NOTCH1 was the opposite(P<0.05),and the expression level of immunotherapy related target genes in the high-risk score group was significantly lower(P<0.05),and higher TMB was significantly associated with poor prognosis(P<0.05).Conclusion The constructed prognostic model and the exploration of the molecular characteristics of tumor immune microenvironment will help to predict the prognostic risk and immunotherapy of oral cancer patients.