基于生物信息学构建胰腺癌患者预后相关miRNA预测模型及其应用价值  

Construction and application value of prognosis associated miRNA prediction model based on bioinforma-tics analysis in pancreatic cancer patients

作　　者：顾江宁 罗海峰[1] 王晨琦 宁振[1] 杜渐[1] 马驰[1] 陈云龙 崔士猛 林志坤 刘一平 谭广[1] Gu Jiangning;Luo Haifeng;Wang Chenqi;Ning Zhen;Du Jian;Ma Chi;Chen Yunlong;Cui Shimeng;Lin Zhikun;Liu Yiping;Tan Guang(Department of Hepatobiliary Surgery,the First Affiliated Hospital of Dalian Medical University,Dalian 116000,China)

机构地区：[1]大连医科大学附属第一医院肝胆外科,116000

出　　处：《中华消化外科杂志》2020年第4期421-430,共10页Chinese Journal of Digestive Surgery

基　　金：国家自然科学基金青年基金项目(81902382);辽宁省自然科学基金博士启动基金(2019BS 077);大连市医学科学研究计划项目(1912029)。

摘　　要：目的探讨基于生物信息学构建胰腺癌患者预后相关微小RNA(miRNA)预测模型及其应用价值。方法采用回顾性队列研究方法。收集肿瘤基因图谱计划(TCGA)数据库(https://cancergenome.nih.gov/)自建库起至2017年9月171例胰腺癌患者的临床病理资料;男93例,女78例;中位年龄为65岁,年龄范围为35~88岁。171例患者中,64例临床病理资料完整。171例患者采用随机抽样法按7∶3比例分为训练集123例和测试集48例。训练集用于构建预测模型,测试集用于验证预测模型效能。从GEO基因公共表达数据库中下载包含9对胰腺癌及对应癌旁组织的miRNA测序数据的数据集GSE41372,从癌组织及癌旁组织差异表达的miRNA中筛选出候选差异表达miRNA,基于训练集患者信息,进行LASSO-COX回归分析,从候选差异表达miRNA中筛选出与生存相关miRNA,将其拟合成一个相对精简的预后相关miRNA模型。分别在训练集和测试集中对构建的预后相关miRNA模型的预测效能进行验证,以受试者工作曲线下面积(AUC)评价模型准确性,以一致性指数(C-index值)评价模型效能。观察指标:(1)患者生存情况。(2)差异表达miRNA筛选结果。(3)预后相关miRNA模型的构建。(4)预后相关miRNA模型的验证。(5)胰腺癌患者临床病理因素比较。(6)影响胰腺癌患者预后的相关因素分析。(7)预后相关miRNA模型与第8版TNM分期预测效能的比较。正态分布的计量资料以±s表示,两组间比较采用Student-t检验,多组间比较采用方差分析。偏态分布的计量资料以M(范围)表示,组间比较采用Mann-Whitney U检验。计数资料以绝对数或百分比表示,组间比较采用χ2检验。等级资料分析采用秩和检验。采用计数资料相关性分析,挖掘预后相关miRNA模型与患者临床病理参数之间的相关性。采用COX进行单因素及多因素分析,并判断相关性,结果以风险比及95%可信区间表示,风险比<1时,证明该因素为�Objective To construct a prognosis associated micro RNA(miRNA)prediction model based on bioinformatics analysis and evaluate its application value in pancreatic cancer patients.Methods The retrospective cohort study was conducted.The clinicopathological data of 171 pancreatic cancer patients from the Cancer Genome Atlas(TCGA)(https://cancergenome.nih.gov/)between establishment of database and September 2017 were collected.There were 93 males and 78 females,aged from 35 to 88 years,with a median age of 65 years.Of the 171 patients,64 had complete clinicopathological data.Patients were allocated into training dataset consisting of 123 patients and validation dataset consisting of 48 patients using the random sampling method,with a ratio of 7∶3.The training dataset was used to construct a prediction model,and the validation dataset was used to evaluate performance of the prediction model.Nine pairs of miRNA sequencing data(GSE41372)of pancreatic cancer and adjacent tissues were downloaded from Gene Expression Omnibus database.The candidate miRNAs were selected from differentially expressed miRNAs in pancreatic cancer and adjacent tissues for LASSO-COX regression analysis based on the patients of training dataset.A prognosis associated miRNA prediction model was constructed upon survival associated miRNAs which were selected from candidate differentially expressed miRNAs.The performance of prognosis associated miRNA prediction model was validated in training dataset and validation dataset,the accuracy of model was evaluated using the area under curve(AUC)of the receiver operating characteristic curves and the efficiency was evaluated using the consistency index(C-index).Observation indicarors:(1)survival of patients;(2)screening results of differentially expressed miRNAs;(3)construction of prognosis associated miRNA model;(4)validation of prognosis associated miRNA model;(5)comparison of clinicopathological factors in pancreatic cancer patients;(6)analysis of factors for prognosis of pancreatic cancer patients;(7)c

关 键 词：胰腺肿瘤 总体生存率 微小RNA LASSO回归模型 受试者工作曲线 TNM分期 

分 类 号：R735.9[医药卫生—肿瘤] Q811.4[医药卫生—临床医学]