临床-CT特征模型对耐多药肺结核的预测价值

打开文本图片集

[Abstract]Objective：Toexplore thevalueoftheclinic-CTfeatures model inpredictingmultidrug-resistant pulmonary tuberculosis（MDR-PTB）.Methods：Atotalof 43Opulmonarytuberculosispatientswithdrugsensitivitytestresultswere retrospectivelycolected，including223casesofMDR-PTB（MDR-PTBgroup），and207casesofdrugsensitivepulmonary tuberculosis（DS-PTBgroup）．Theclinicaldataandchest CTfeaturesof thetwogroupswerecomparedandanalyzed.A one-wayanalysswasperformed，clinicalandCTfeatureswithstatisticallsignificantdiferenceswereselectedasindependent variables，andlogisticregressionanalysiswasfurtherperformedwithmultidrugpulmonaryresistanceasthedependentvariable. Thentheclinic-CTfeaturesmodelwasestablished，andthediagnosticeficacyofthemodelwasalsoverifiedwithROC curves.Results：There-treatmentrate，bronchiectasis，lesioninvolvingmorethan3lunglobes，multiplecavities，thick-waled cavities，ncapsulatedefusion，pleuralthickeninganddiseaseprogresionoftheMDR-PTBgroupwereallhigherthanthoseof the DS-PTB group （all P<0.05 ）.The logistic regression analysis indicated that re-treatment rate，multiple cavities，thick-walled cavitiesanddiseaseprogressionweretheindependentrisk factorsforMDR-PTB.Theregressonequationmodelbasedon theabove four independent risk factors hadan AUC of O.799，a specificity of 89.9%，a sensitivity of 61.0% ，with the optimal criticalvalueofO.570.Conclusions：TheclinicalandchestCTfeaturesofMDR-PTBandDS-PTBhavecertaindiferences. Whenmultiplemanifestationssuchasrtreatment，utiplecavities，tick-walledcaviis，anddiseaseprogressionceistt highlysuggestsMDR-PTB.Basedontheclinic-CTfeaturesmodel，itcanprovideimagingvaluefortheearlydiagnosisof MDR-PTB patients.

KeyWords] Pulmonary tuberculosis;Multidrug-resistant;Tomography，X-ray computed;Diagnos

耐多药肺结核（multidrug-resistantpulmonarytuberculosis，MDR-PTB）是指患者感染的结核分枝杆菌（mycobacteriumtuberculosis，MTB）经体外药物敏感性试验（drug sensitivetest，DST）证实至少对异烟肼和利福平耐药的结核病[1-2]。（剩余6240字）