前列腺癌預后影響因素的Cox模型分析.pdf

1、汕頭大學碩士學位論文前列腺癌預后影響因素的Cox模型分析姓名：黃耀林申請學位級別：碩士專業(yè)：臨床醫(yī)學（泌尿外科）指導教師：葉炯賢20100520汕頭大學醫(yī)學院2003級臨床醫(yī)學（本碩連讀）碩士研究生學位論文VEnglishAbstractObjectiveTheaimofthestudyistoexpletheprognosticfactsofcarcinomaoftheprostate(CaP)establishCoxproption

2、alhazardsmodelfpredictingprognosisofCaPprovidingprooffthetreatmentprognosispredictionofCaP.MethodsDataof145patientswhowerediagnosedwithCaPinPeikingUniversityShenzhenHospitalfromJanuary2003toAugust2009werecollected.Allcas

3、ewereoutpatienttelephonefollowupeduntilFebruary2010.12caseswereeliminatedbecauseofincompleteclinicdatafailureoffollowupearlydyingofotherdiseasestheother133caseswereincluedinthestudy.KaplanMeiermethodLogranktestwereusedfs

4、urvivalanalysis29factspossibllyimpactingonprognosisofCaPwereunivariatemultivarateanalysedbaseontheCoxproptionalhazardsregressionmodel.ThenindependentprognosisfactsofCaPwereedpredictivemodalwasestablished.PI(Prognosticind

5、ex)thresholdwasidentifiedverified.ResultsThediseasespecificsurvivalratesoftheCaPat135yearwere90.9%78.4%65.4%.CoxmutivarateanalysisindicatedthatthemostimprotantindependentprognosticfactsimpactingonCaPwereX4bonepain(RR=4.7

6、07P=0.03)X9HGB(RR=8.050P=0.005)X20Tstage(RR=10.245P=0.001)X29treatments(RR=10.964P=0.001).Therefthepredictivemodalwascreatedasthefollowing:PI=2.8691.687X41.756X92.326X201.779X29.ChisquaretestshowedX2=71.026(P0.001)whichi

7、llustratedtheequationreasonable.EachPIofthepatientwascalculated.ThenallthepatientsweredividedintohighriskmedianrisklowriskgroupLogranktestindicatedthatthesurvivalratesamongthesegroupweresignificantdifference(P0.001).Conc

8、lusionTstageHGBbonepaintreatmentswereallindependentprognosticfactsofCaP.TreatmentswasprotectivefacttheothersweredangerfactsofCaPprognosis.ThepredictivemodalincludingthesefourfactswasabletopredictthesurvivalofCaPeffective

9、ly.InclinicwkwiththeprostatecancerpatientswecaneasillyknowthepatientsprognosisbaseonthePImodaljustneedingasimplecalculation.ItwasindeedprovideausefultoolfclinictreatmentpredictingprognosisofCaPwithhighpracticalvalue.Keyw