Application Of Quantitative CT In The Diagnosis Of Tracheobronchial Tuberculosis

Objective: To analyze the characteristics of tracheobronchial tuberculosis(TBTB)using artificial intelligence quantitative CT,and to explore the diagnostic value of the diagnostic model based on quantitative CT indicators,general information,clinical symptoms and laboratory tests for TBTB,so as to provide help for the diagnosis and evaluation of TBTB.Method: picked by trained researchers the second people’s hospital of yan ’an April2020 to April 2022 the second people’s hospital of yan ’an hospital treatment line of bronchoscopy tuberculosis patients,according to whether the merger tracheal bronchus TB,They were divided into simple pulmonary tuberculosis group and tracheal-bronchial tuberculosis group.General information,clinical symptoms,routine biochemistry(WBC,NEU%,NEUT,LYMPH,LYMPH%,LYM,LYM%,Crea,etc.),etiology of mycobacterium tuberculosis(sputum smear,sputum culture,sputum TB-DNA),chest imaging,bronchoscopy and pathology data of the above patients were collected.SPASS software was used for statistical processing.The ROC curve was used to evaluate the diagnostic efficacy of a single imaging index for bronchial tuberculosis,and the cut-off value was obtained,according to which binary assignment was performed.The diagnostic models constructed by combining general data,clinical symptoms,laboratory examination,multiple quantitative CT indicators,and all the above indicators were respectively constructed.The consistency of the diagnostic results of the three models with those under bronchoscopy was compared,and sensitivity,specificity,accuracy,positive predictive value and negative predictive value were calculated.To explore the application value of quantitative CT in the diagnosis of bronchial tuberculosis.Results:1.This study found that the most frequently invaded site of TBTB was the upper bronchus.Bronchial thickening was more common in right upper segmental bronchus and left upper segmental bronchus.Stenosis was more common in the right upper and lower segmental bronchi and the left upper and lower segmental bronchi.The reduction of bronchial wall area caused by tracheal remodeling was more common in grade 2 bronchi of the right upper lobe,grade 3 of the right lower lobe,grade 2-3 bronchi of the left upper lobe and the left lower lobe.2.Univariate analysis showed that compared with the PTB group,TBTB In group A,patients had age,sex,cough,sputum smear,sputum TB-DNA,average density of both lungs,average mass of both lungs,volume of both lungs,ratio of bronchial wall thick-outer diameter(Grade 3 in upper lobe of both lungs),ratio of bronchial lumen to body surface area(Grade 2-3 in upper lobe of right lung,grade 3 in middle lobe of right lung,grade 4 in lower lobe of right lung,grade 3-4 in upper lobe of left lung,and grade 4in lower lobe of left lung),and average bronchi There were significant differences in the wall area(Grade 3 in the upper and lower lobes of the right lung and grade 3 in the upper lobe of the left lung)(p < 0.05).3.Logistic regression analysis showed that the average density of bilateral lungs,average mass of bilateral lungs,bronchial wall thickness/external diameter(level 3 in the right upper lobe),and average bronchial lumen diameter/body surface area(level 4 in the left upper lobe)were independent risk factors for the diagnosis of PTB combined with TBTB.The AUC value of ROC curve of multiple combined imaging index models for predicting TBTB patients was 0.803(95%CI=0.736-0.870,P=0.000),sensitivity was65.5%,specificity was 83.1%,the maximum entry index was 0.486 and the corresponding critical value was 0.696.4.The AUC value of ROC of the diagnostic model combined with general data,clinical symptoms and laboratory examination was 0.703(95%CI= 0.6233-0.783,P=0.000),sensitivity was 80.1%,specificity was 46.2%,the maximum entry index was0.345 and the corresponding critical value was 0.57.5.The AUC value of ROC curve of the model combined with quantitative CT indicators,general data,clinical symptoms and laboratory examination was 0.868(95%CI=0.816-0.919,P=0.000),sensitivity was 74.8%,specificity was 83.1%,the maximum entry index was 0.579 and the corresponding critical value was 0.670.Conclusions: 1.This study found that the most common site of TBTB invasion was the upper lobe bronchus.Bronchial thickening was more common in the bronchus of the upper segmental segments of both lungs.The stenosis was more common in the right upper and lower bronchus,and the left upper and lower bronchus.The reduction of bronchial wall area caused by tracheal remodeling was more common in grade 2 bronchi in the right upper lobe,grade 3 in the right lower lobe,and grade 2-3 bronchi in the left upper and lower lobe.2.Whole lung indexes(bilateral average lung density,bilateral average lung mass)and local airway indexes(bronchial wall thickness/external diameter,bronchial average lumen diameter/body surface area,etc.)were predictive factors for the diagnosis of PTB combined with TBTB.3.The risk prediction model based on the combined quantitative CT indicators,general data,clinical symptoms and laboratory examination has high efficacy in the diagnosis of TBTB,and can provide a convenient and rapid evaluation index for clinical.