基于TSP家族构建胃癌预后评估模型  

Construction of prognostic evaluation model of gastric cancer based on TSP family

作　　者：刘航[1] 卢文卿 包博文 郑雪莹 李冬阳 王多 车晓芳[1] LIU Hang;LU Wenqing;BAO Bowen;ZHENG Xueying;LI Dongyang;WANG Duo;CHE Xiaofang(The First Hospital of China Medical University,Shenyang 110001,China)

机构地区：[1]中国医科大学附属第一医院,沈阳110001

出　　处：《山东医药》2021年第10期6-10,共5页Shandong Medical Journal

基　　金：辽宁省重点研发计划指导计划项目(2018225060);沈阳市科技计划项目(19-112-4-099)。

摘　　要：目的基于凝血酶原(TSP)家族构建胃癌预后评估模型。方法从GEO数据库下载胃癌样本量最大的GSE62254数据集,获取胃癌样本转录组测序表达数据和相关临床资料,在排除生存资料缺失样本后,最终纳入胃癌患者283例。利用单因素和多因素Cox回归模型筛选TSP家族中胃癌预后的独立危险因素,并建立基于TSP家族的风险评分模型RS1;分析RS1与胃癌患者临床病理特征的关系,结合RS1与胃癌其他独立预后因素,建立风险评分模型RS2,然后与传统基于临床病理特征的预后模型RS3进行比较。结果单因素Cox回归分析显示,TSP1、TSP3和TSP5可能与胃癌患者预后有关(P均<0.05);多因素Cox回归分析显示,TSP3、TSP5是胃癌患者预后的独立危险因素(P均<0.05)。基于TSP3、TSP5构建RS1,以RS1的中位数为临界值将患者分为高RS1者与低RS1者,结果发现,高RS1者生存期明显低于低RS1者(P<0.05),RS1预测胃癌患者预后的ROC曲线下面积(AUC)为0.663。RS1与肿瘤T分期和腹膜转移有关(P均<0.05),与性别、年龄、N分期、M分期及Lauren分型无关(P均>0.05)。单因素Cox回归分析显示,T分期、N分期、M分期、Lauren分型和RS1可能影响胃癌患者预后(P均<0.05)。多因素Cox回归分析显示,T分期、N分期、M分期和RS1是胃癌患者预后的独立危险因素(P均<0.05)。基于T分期、N分期、M分期和RS1构建了RS2,以RS2的中位数为临界值将患者分为高RS2者与低RS2者,高RS2者生存期明显低于低RS2者(P<0.01),RS2预测胃癌患者预后的AUC为0.795。仅基于T分期、N分期、M分期构建了RS3,RS3预测胃癌患者预后的AUC为0.767。DeLong检验发现,RS2预测胃癌患者预后的AUC明显高于RS1、RS3(P均<0.05)。结论本研究成功基于TSP家族构建了胃癌预后的评估模型,该模型能够较准确地判断胃癌患者预后。Objective To establish the prognostic evaluation model of gastric cancer based on prothrombin(TSP)family.Methods The GSE62254 dataset with the largest sample size of gastric cancer was downloaded from GEO data⁃base,and the sequencing expression data and clinical data of gastric cancer sample transcriptional group were obtained.After excluding the missing samples of survival data,283 patients with gastric cancer were included.Univariate and multi⁃variate Cox regression analyses were used to screen out the independent prognostic factors of gastric cancer in TSP family,and a risk score model RS1 based on TSP family was established,and we analyzed the relationship between RS1 and clini⁃copathological characteristics of patients with gastric cancer.We combined RS1 with other independent prognostic factors of gastric cancer to establish another risk score model RS2.Then,RS2 was compared with the traditional prognostic model RS3 based on clinicopathological characteristics.Results Univariate Cox regression analysis showed that TSP1,TSP3 and TSP5 might be associated with the prognosis of patients with gastric cancer(all P<0.05),while multivariate Cox regression analysis showed that TSP3 and TSP5 were independent risk factors for prognosis of patients with gastric cancer(both P<0.05).RS1 was constructed based on TSP3 and TSP5.According to the median of RS1,the patients were divid⁃ed into the high RS1 group and low RS1 group.The results showed that the overall survival(OS)of patients with high RS1 was significantly lower than that of patients with low RS1(P<0.05).The AUC of RS1 in predicting the prognosis of patients with gastric cancer was 0.663.RS1 was related to T stage and peritoneal metastasis(both P<0.05),but not to sex,age,N stage,M stage or Lauren classification(all P>0.05).Cox univariate analysis showed that T stage,N stage,M stage,Lauren classification and RS1 might affect OS of patients with gastric cancer(all P<0.05).Cox multivariate anal⁃ysis showed that T stage,N stage,M stage,and RS1 were independe

关 键 词：胃癌 凝血酶原 预后 风险评分 

分 类 号：R735.2[医药卫生—肿瘤]