摘要
目的:利用食管癌的可变剪切数据,构建食管癌预后风险模型并建立列线图。方法:分别从癌症基因图谱(The Cancer Genome Atlas,TCGA)SpliceSeq数据库和TCGA数据库下载食管癌的可变剪切数据和食管癌患者的临床资料,将2个数据集合并,得到食管癌可变剪切的生存数据。采用单因素COX回归分析筛选与预后相关的可变剪切,为避免模型过度拟合,采用最小绝对值收敛和选择算子(least absolute shrinkage and selection operator,LASSO)回归分析筛选变量;随后将筛选得到的可变剪切纳入多因素COX回归分析构建食管癌的预后风险模型。基于该模型计算每位患者的风险评分并根据风险评分的中位数将患者分为高风险组和低风险组,采用Kaplan-Meie进行生存分析和受试者接受特征(receiver operating characteristic,ROC)曲线对模型效能进行评价。最终将风险模型与临床病理特征合并,采用COX回归分析探究食管癌的独立预后因素并建立列线图。采用一致性指数(concordance index,C-index)、校准图和决策曲线来评价列线图的预测准确度。结果:从TCGA SpliceSeq和TCGA数据库下载得到185例食管癌患者的可变剪切数据和临床资料。采用单因素COX回归分析筛选得到2389个与食管癌预后相关的可变剪切,随后采用Lasso回归分析筛选得到17个可变剪切,使用多因素COX回归分析建立了基于10个基因(CHRDL2、ERBB2、IAH1、C16orf13、C19orf82、RNF150、PNKP、ZNF467、TMPRSS4和HPS1)的可变剪切的食管癌预后风险模型。基于模型风险评分,将样本分为高风险组和低风险组;Kaplan-Meier生存分析显示,高风险组的总生存(overall survival,OS)率较低风险组差(P<0.05),ROC分析结果提示,该食管癌预后风险模型预测性良好[曲线下面积(area under curve,AUC)=0.865]。COX回归分析结果显示,病理分期和风险评分是食管癌的独立预后因子。结合这些预后因子构建的列线图显示出较好的区分度(C-index为0.79;95%可信区间为0.752~0.843),校准曲线及决策曲线提示该列线图具有较好的预测食管癌患者OS期的能力。结论:基于可变剪切的预后风险模型能够明显区分高风险和低风险组食管癌患者的生存率。基于模型风险评分及病理分期构建的列线图能够有效地预测食管癌患者的OS率。
Objective:To establish a prognostic risk model and nomogram with alternative splicing data of esophageal cancer.Methods:The data of alternative splicing and clinical information of esophageal cancer patients were downloaded from The Cancer Genome Atlas(TCGA)SpliceSeq and TCGA database respectively.Then the two data sets were combined to obtain the survival data of alternative splicing of esophageal cancer.Univariate COX regression analysis was used to screen the alternative splicing associated with the prognosis of esophageal cancer patients.The variables were screened with the least absolute shrinkage and selection operator(LASSO)regression analysis in order to avoid overfitting.The screened variables were analyzed by multivariate COX regression analysis as well as established the prognostic risk model.Based on the model,patients were divided into high-risk and low-risk groups according to the median risk score.The model effectiveness was evaluated by Kaplan Meier survival analysis and receiver operating characteristic(ROC)curve.Finally,the risk model was combined with clinicopathological features.COX regression analysis was used to explore the independent prognostic factors of esophageal cancer and to establish a nomogram.Concordance index(C-index),calibration plot and decision curve were used to evaluate the prediction accuracy of nomogram.Results:Data of 185 patients with the alternative splicing data and clinical information were download from TCGA SpliceSeq and TCGA,respectively.A total of 2389 survival associated alternative splicing were obtained by univariate COX regression.Using LASSO regression analysis,17 alternative splicing were obtained.Multivariate COX regression analysis was used to establish the prognostic risk model for esophageal cancer based on the alternative splicing of 10 genes,including CHRDL2,ERBB2,IAH1,C16orf13,C19orf82,RNF150,PNKP,ZNF467,TMPRSS4 and HPS1.Based on the risk score of the prognostic model,the patients were divided into high-risk group and low-risk group.Kaplan Meier survival analysis showed that the overall survival(OS)of high-risk group was worse than that of low-risk group(P<0.05).ROC analysis showed that the predictive effectiveness of the prognosis risk model of esophageal cancer was satisfactory[area under curve(AUC)=0.865].COX regression analysis suggested that pathological stage and risk score were the independent prognostic factors of esophageal cancer.The nomogram combined with these prognostic factors showed a good discrimination(C-index was 0.79;95%confidence interval=0.752-0.843).The calibration curve and decision curve indicted that the nomogram could predict OS of esophageal cancer patients effectively.Conclusion:The prognostic risk model based on alternative splicing can better distinguish the survival rate of esophageal cancer patients between high and low risk groups.The nomogram based on model risk score and pathological stage can effectively predict the OS of esophageal cancer patients.
作者
孔晨帆
孙建荣
杨佳潞
孙劲晖
KONG Chenfan;SUN Jianrong;YANG Jialu;SUN Jinhui(Graduate School,Beijing University of Chinese Medicine,Beijing 100029,China;Department of Gastroenterology,Dongzhimen Hospital,Beijing University of Chinese Medicine,Beijing 100700,China)
出处
《肿瘤》
CAS
CSCD
北大核心
2020年第12期834-845,共12页
Tumor
关键词
食管肿瘤
RNA加工
转录后
可变剪切
预后
决策支持技术
列线图
Esophageal neoplasms
RNA processing,post-transcriptional
Alternative splicing
Prognosis
Decision support techniques
Nomograms
作者简介
Correspondence to:SUN Jinhui(孙劲晖),E-mail:lidaosi@sina.com。
