基于综合生物信息学的宫颈癌预后模型建立

Prognostic Model of Cervical Cancer Based on Integrated Bioinformatics

目的:寻找与宫颈癌预后相关的免疫相关性长链非编码RNA(long non-conding RNA,IncRNA),构建宫颈癌预后评估模型,提高宫颈癌的生存率。方法:利用R4.2.1,采用Cox回归分析和最小绝对收缩和选择算子(lasso)Cox回归,寻找癌症基因组图谱(The Cancer Genome Atlas,TCGA)数据库中宫颈癌患者基因表达差异的长链非编码RNA(lncRNA)及生存信息差异,构建预后模型。利用R软件中“survminer”包(4.2.1版)生成Kaplan-Meier生存曲线,利用R软件中“rms”包(4.2.1版)生成列线图验证模型。结果:采用Cox回归分析和最小绝对收缩和选择算子(lasso)Cox回归模型构建免疫相关的22个lncRNA模型(包括SCP2D1.AS1、LINC00469、LINC00308、LINC00158、MIRLET7BHG、ARRDC1.AS1、SNHG32、DDC.AS1、NUP50.DT、LINC00658、LINC00629、LINC02802、LINC01281、SBF2.AS1、WWC2.AS2、USP30.AS1、IL21R.AS1、AA06、ASB16.AS1、ILF3.DT、RBAKDN、FAM27E3),预后模型预测宫颈癌患者的总生存期(HR=9.54;95%CI:6.27-14.50,P<0.01),将患者分为高、低风险组绘制K-M生存曲线,结果差异有统计学意义(P<0.01),该模型的ROC曲线的曲线下面积(AUC)为0.901,表明该模型具有良好的诊断效果。结论:本研究揭示了免疫相关的22个lncRNA模型是子宫颈癌预后预测的独立因素,可为临床上宫颈癌个体化治疗提供强有力的证据。

Objective:To search for immune-associated long non-conding RNA(IncRNA)which is associated with the prognosis of cervical cancer,construct the prognosis evaluation model of cervical cancer,and improve the survival rate of cervical cancer.Methods:Using R4.2.1,Cox regression analysis and minimum absolute contraction and selection operator(lasso)Cox regression,Long non-coding Rnas(lncrnas)with gene expression differences and survival information differences in patients with cervical Cancer were investigated in The Cancer Genome Atlas(TCGA)database,and a prognostic model was constructed.Results:The immune-associated 22-lncrnas(including SCP2D1.AS1,LINC00469,LINC00308,LINC00158,MIRLET7BHG,ARRDC1.AS1,SNH)were constructed by Cox regression analysis and minimum absolute contraction and selection operator(lasso)Cox regression model G32,(DDC.AS1,NUP50.DT,LINC00658,LINC00629,LINC02802,LINC01281,SBF2.AS1,WWC2.AS2,USP30.AS1,IL21R.AS1,AA06,ASB16.AS1,ILF3.D T,RBAKDN,FAM27E3)signature prognostic model predicted overall survival of patients with cervical cancer(HR=9.54;95%CI:6.27-14.50,P<0.01),patients were divided into high and low risk groups to draw K-M survival curves,and the results were statistically significant(P<0.01).The area under the curve(AUC)of ROC curve of this model was 0.901,indicating that this model had a good diagnostic effect.Conclusions:This study revealed that immune-related 22-lncRNA signatures are independent prognostic factors of cervical cancer,which can provide strong evidence for individualized clinical treatment of cervical cancer.