Drug-resistant Pulmonary Tuberculosis Versus Non-drug-resistant Pulmonary Tuberculosis:A Comparative Study Of CT Findings

Objective: To analyze the clinical data and chest CT findings of patients with drug-resistant pulmonary tuberculosis and non-drug-resistant pulmonary tuberculosis and investigate the chest CT features of drug-resistant pulmonary tuberculosis,so as to provide imaging evidence for timely laboratory examination of suspected drug-resistant pulmonary tuberculosis,shorten the diagnosis time of drug-resistant pulmonary tuberculosis,adjust the treatment plan in time,and improve the efficacy of therapy.Methods: A total of 1546 cases of pulmonary tuberculosis with complete clinical data,chest CT images and defined results of drug sensitivity test were collected from October 2008 to September 2017 in the Affiliated Hospital of Zunyi Medical University,including 516 cases of drug-resistant pulmonary tuberculosis(drug-resistant group)and 1030 cases of non-drug-resistant pulmonary tuberculosis(non-drug-resistant group or sensitive group).The CT images and clinical data of the drug-resistant and non-drug-resistant patients were recorded in detail and analyzed,and the information database was established,including clinical information(such as age,sex,diagnosis and treatment types,clinical manifestations,etc.)and imaging findings which included tracheobronchial lesions(such as bronchiectasis,bronchial dissemination,atelectasis,etc.),pulmonary lesions(such as exudation,proliferation,calcification,nodules,cavities,etc.),pleural lesions(such as pleural effusion,pneumothorax,etc.),complications(such as pneumoconiosis,collapsed lung,etc.)and the prognosis of disease.In statistics,the measurement data were analyzed by independent sample t-test and the enumeration data by independent sample chi-square test in the single-factor analysis,and the difference was statistically significant(p<0.05).Further,the logistics regression analysis was carried out in which the factors with a statistical difference(p<0.01)were selected as independent variables and drug resistance situation as dependent variable.In the meantime,the regression analysis equation model was constructed,on the foundation,ROC curve and Hosmer-Lemeshow test was used to verify the diagnostic efficiency and the level of calibration,respectively.Results:(1)Clinical data: the distribution of patients with retreatment and intermittent medication history in the drug-resistant group was higher than that in the non-drug-resistant group,and the difference was statistically significant(p<0.05).There was no significant difference in age,sex,diagnostic types and clinical manifestations between the two groups(p>0.05).(2)Imaging data: the following features were more frequently observed in the drug-sensitive group than in the drug-resistant group,including whole lung involvement,cavity,cavity-related manifestations(cavities with gas-liquid interface,cavities with thick wall,multiple cavities,multi-lobe involved cavities),collapsed lung,bronchial dissemination,calcification,bronchiectasis,emphysema,atelectasis,proliferation lesions,and encapsulated effusion,while the exudative lesions and pneumoconiosis were more common in the non-drug-resistant group than in the drug-resistant group,which indicated a statistically significant difference(p<0.05).(3)Logistics regression analysis showed that multiple cavities,thick-walled cavities,whole lung involvement,bronchial dissemination,bronchiectasis,emphysema,proliferative lesions,calcification and progression of the disease were the independent risk factors for drug resistance of pulmonary tuberculosis patients,and exudative lesions became protective factors of drug resistance.The regression equation model was produced through the above analysis,the total correct rate of prediction is 80.6%.(4)On this basis,the ROC curve was drawn whose area under the curve(AUC)is 82.6%,and the optimal cut-off value predicted by the regression model is 0.354,that the diagnostic efficiency was the best when the predicted value above the cut-off value would be predicted as drug-resistance,and vice versa.The Hosmer-Lemeshow test showed that there was no statistical significance between the observed value and the predicted value in the prediction model(p>0.05).Conclusion:(1)Chest CT manifestations of drug-resistant pulmonary tuberculosis may showed some certain characteristics,it significantly indicated the possibility of drug resistance in patients with pulmonary tuberculosis when there were multifarious image performance coexisting with each other,including multiple cavities,thick-walled cavities,bronchial dissemination,bronchiectasis,whole lung involvement,calcification,proliferative lesions,disease progression,etc.(2)CT examination and drug-resistance prediction model may offer imaging evidence for timely detection of drug resistance in patients with suspected drug-resistant pulmonary tuberculosis and contribute to the early diagnosis of drug-resistant pulmonary tuberculosis.