自噬相关基因对老年女性乳腺癌预后预测的价值  

Predictive value of autophagy-related genes for prognosis of elderly female patients with breast cancer

作　　者：孟轲 耿一萌 杨珍珍 冯涵 李冰洁[1] 马望[1] MENG Ke;GENG Yimeng;YANG Zhenzhen;FENG Han;LI Bingjie;MA Wang(Department of Oncology,the First Affiliated Hospital,Zhengzhou University,Zhengzhou 450052)

机构地区：[1]郑州大学第一附属医院肿瘤科,郑州450052

出　　处：《郑州大学学报（医学版）》2022年第2期181-185,共5页Journal of Zhengzhou University(Medical Sciences)

基　　金：国家自然科学基金项目(820731680)。

摘　　要：目的:探索自噬相关基因(ARG)对老年女性乳腺癌预后预测的价值。方法:分别从TCGA和HADb数据库中下载得到老年女性乳腺癌504例的病例信息和ARG信息。将病例按7∶3的比例分为训练集(n=356)和验证集(n=148)。乳腺癌患者总生存期定义为从随机化开始至因任何原因死亡的时间。在训练集中,通过Cox回归分析筛选与老年女性乳腺癌患者预后相关的ARG,ARG相关风险评分=∑βn×基因n表达水平平均值。以风险评分的中位数作为界值,将乳腺癌患者分为高风险和低风险组,比较2组预后相关ARG的表达水平,并进行GO功能富集分析和KEGG通路分析。在训练集中,以年龄、临床分期、TNM分期作为协变量,以风险评分为自变量,进行Cox回归分析,构建老年女性乳腺癌预后预测模型。在训练集和验证集中,绘制预测模型的时依ROC曲线,用于评估模型的预测能力。结果:通过Cox回归分析获得了一个包括7个ARG的风险评分,IFNG、ATG4D、ATG4A、NKX2-3、MAPK9、ATG9B、VAMP3的表达水平平均值(β)分别为0.375(-0.619)、7.667(0.137)、6.210(0.143)、0.041(3.208)、7.846(0.104)、0.590(0.192)、34.929(0.054);GO和KEGG分析结果表明,这7个ARG参与的生物学过程主要包括巨噬细胞分化、蛋白质定位至前自噬体结构、自噬等,涉及的细胞成分主要有内体膜回收,涉及的分子功能主要包括半胱氨酸肽酶活性,与自噬(动物)等通路显著相关。基于ARG相关风险评分的预测模型中,年龄、临床分期、T分期、N分期、M分期及ARG相关风险评分的HR(95%CI)分别为1.072(1.030~1.116)、1.774(0.774~3.928)、0.830(0.523~1.316)、0.502(0.137~1.839)、1.430(0.870~2.350)、1.276(1.168~1.394);该模型在验证集中对患者3、5、10 a生存概率预测的ROC曲线下面积(95%CI)分别为0.769(0.546~0.992)、0.714(0.516~0.912)、0.803(0.615~0.992)。结论:筛选出的ARG对老年女性乳腺癌预后具有一定的预测价值。Aim:To explore the predictive value of autophagy-related gene(ARG)for the prognosis of the geriatric breast cancer(BRAC)patients.Methods:The data of BRAC patients were downloaded from the TCGA database and the ARG list was obtained from the HADb database.The geriatric BRAC patients were randomly allocated into training set(n=356)and a validation set(n=148)according to the ratio of 7∶3.Overall survival of BRAC patients was defined as the time from randomization to death from any cause.In the training set,Cox regression analysis was used to screen the ARGs related to the overall survival(prognosis)of geriatric BRAC patients.Risk score was equal to∑βn×average of ARGn expression level.Using the median risk score as the critical value,BRAC patients were divided into high-risk and low-risk groups.The expression levels of prognosis-related ARGs were compared between the two groups.GO and KEGG analysis were performed.A prognosis prediction model for BRAC was constructed by using Cox regression analysis with ARG-related risk score as independent variable,and age,clinical stage,and TNM stage as covariates.In the training set and validation set,the time-dependent ROC was conducted to determine the predictive performance of the model.Results:A risk score including seven ARGs was obtained and the average expression level(β)of IFNG,ATG4D,ATG4A,NKX2-3,MAPK9,ATG9B,and VAMP3 were 0.375(-0.619),7.667(0.137),6.210(0.143),0.041(3.208),7.846(0.104),0.590(0.192),and 34.929(0.054),respectively.GO and KEGG analysis showed that the seven ARGs were primarily enriched in the biological process of"macrophage differentiation","protein localization to pre-autophagosomal structure"and"autophagy";mainly enriched in cellular component of"recycling endosome membrane",mainly enriched in molecular function of"cysteine-type peptidase activity",and involved of autophagy-animal and autophagy-other pathways.In the prediction model based on ARG-related risk score,the HR(95%CI)of age,clinical stage,T stage,N stage,M stage,and ARG-related risk scor

关 键 词：乳腺癌 自噬相关基因 预后 

分 类 号：R737.9[医药卫生—肿瘤]