Five-CpG-based prognostic signature for predicting survivalin hepatocellular carcinoma patients

(整期优先)网络出版时间:2018-04-14
/ 1
Objective:Hepatocellularcarcinoma(HCC)isacommonmalignancyassociatedwithhighmorbidityandmortalityratesworldwide.EarlydiagnosisplaysanimportantroleintheimprovementofHCCprognosis.Methods:Inthisstudy,weconductedacomprehensiveanalysisofHCCDNAmethylationandgeneexpressiondatasetsinTheCancerGenomeAtlas(TCGA),toidentifyaprognosticsignatureforHCCdiagnosisandsurvivalprediction.First,weidentifieddifferentialmethylationCpG(dmCpG)sitesinHCCsamplesandcomparedthemwiththoseinadjacentnormallivertissues;thiswasfollowedbyunivariateanalysisandSureIndependenceScreening(SIS)inthetrainingset.Therobustnessoftheidentifiedprognosticsignaturewasevaluatedusingthetestingset.ToexplorethebiologicalprocessesinvolvedinHCCprogression,wealsoperformedfunctionalenrichmentanalysisforoverlappinggenesbetweengenescontainingdmCpGsites(DMGs)anddifferentialexpressiongenes(DEGs)inHCCpatients,usingdatafromtheDatabaseforAnnotation,Visualization,andIntegratedDiscovery(DAVID).Results:Asaresult,weidentifiedfiveCpGsitesthatweresignificantlyassociatedwithHCCsurvivalthroughunivariateanalysisandSIS.Univariateanalysisofclinicalcharacteristicsidentifiedageandriskfactors(includingalcoholconsumptionandsmoking)asindependentfactorsthatindicatedHCCsurvival.Multivariateanalysisindicatedthattheintegratedprognosticsignature(weightedcombinationofthefiveCpGsites)thattookageandriskfactorsintoconsiderationresultedinmoreaccuratesurvivalprediction.Conclusions:ThisstudyprovidesanovelsignatureforpredictingHCCsurvival,andshouldbehelpfulforearlyHCCdiagnosisandpersonalizedtreatment.