·149CHINESEJOURNALOFCTANDMRI,FEB.2023,Vol.21,No.02TotalNo.160【第一作者】李邦凤，女，本科生，主要研究方向：影像组学。E-mail：1967016975@qq.com【通讯作者】彭云，男，医师，主要研究方向：双能CT应用、影像组学。E-mail：penghh12580@163.com论著TheValueoftheCTRadiomicsCombinedwithSupportVectorMachinesinDifferentBetweenIncidentalAcuteandOldvertebralCompressionFractures*LIBang-feng1,2,FUYu-ping1,2,GONGLiang-geng1,PENGYun1,*,LINHua-shan3.1.DepartmentofRadiology,TheSecondAffiliatedHospitalofNanchangUniversity,Nanchang330006,JiangxiProvince,China2.TheSecondClinicalCollegeofMedicine,NanchangUniversity,Nanchang330006,JiangxiProvince,China3.GEPharmaceuticalGEHealthcare,Changsha410000,HunanProvince,ChinaABSTRACTObjectiveToinvestigatethevalueofvertebraltextureanalysisbasedonthoracicCTimagescombinedwithsupportvectormachine(SVM)machinelearningmethodinidentifyingacuteandoldvertebralcompressionfractures.MethodsThedataof132patientswithincidentalvertebralcompressionfracturesdetectedonroutinechestCTandconfirmedbyMRIfromMay2018toMay2021wereretrospectivelyanalyzed.163vertebraeincluding98acutefracturesand65oldwereincluded.TheMazdasoftwarewasusedtoextracttexturefeaturesofeachvertebrainaxialandsagittalorientation.ThentheIPMSsoftwarewasusedtofurtherdimensionalityreductionandmodelbuilding.Theoldandacutefracturedvertebraewererandomlydividedintothetrainingandvalidationsamplesaccordingtotheratioof7:3.Fortrainingsample,theTtest,WilcoxonranksumtestandPearsoncorrelationanalysiswereperformedtoscreenthetexturefeaturesinbothorientations.SVMmodelswerebuiltbasedontheselectedaxialandsagittalparametersrespectively.Thenthediagnosticvaluewastestedbythevalidationsampleandthereceiveroperatingcharacteristiccurves(ROC)wereobtained.ResultsForeachvertebra,294featureswereextractedfromsagittalandaxialimagingrespectively.8parameterswerefinallyobtainedforthesagittalorientation,theAUCoftheSVMmodelwas0.78and0.68forthetrainingandvalidationsample,respectively;7parameterswerefinallyobtainedfortheaxialorientation,theAUCofth...
