The Relationship Between Quantitative CT Parameters And Small Airway Disease In Patients With Mild To Moderate COPD

ObjectiveTo explore the correlation between the quantitative parameters of CT and the pulmonary function indicators(including spirometer indicators and pulse oscillation indicators)reflecting small airway disease by COPD semi-automatic analysis software.In additon,to construct the nomogram of the mild to moderate COPD diagnosis prediction model based on quantitative CT parameters by R software,and the model was evaluated and verified.Method34 patients with mild to moderate COPD diagnosed by pulmonary function examination and 42 healthy controls were enrolled for pulmonary function examination.The end-inspiratory and end-expiratory multi-slice high resolution CT(HRCT)were performed within 7 days.Quantitative analysis was made by COPD Analysis software:①emphysema parameters(LAA%);②bronchial parameters:right upper lobe apical segment 5th grade bronchial wall thickness(WT),wall area(WA),lumen diameter(LD),lumen area(LA),WA/LA;③air trapping parameters (RVC-860 to-950,MLDE/I).The of the two groups were compared between groups,and the correlation between quantitative CT parameters and pulmonary function was discussed.The variables with statistical differences in the comparison between the two groups were included in the bivariate multifactor logical regression analysis to establish a regression equation.The prediction nomogran model was constructed by R software 4.0.2,and the discrimination,calibration and clinical efficiency of the prediction model were evaluated and validated.Result1.The LAA%of whole lung,right lung,left lung,middle lobe of right lung,lower lobe of right lung,upper lobe of left lung and lower lobe of left lung had significant negative correlation with FEV1%pred,FEV1/FVC,MEF75,MEF50 and MMEF.Except for the weak positive correlation between LD5 and MEF25,there was no significant correlation between airway measurement parameters and spirometer parameters.2.There was no significant correlation between LAA%and all IOS parameters in whole lung,left lung,right lung,right upper lobe,right middle lobe,left upper lobe and left lower lobe.For bronchial parameters,LD5 and LA5 were weakly negatively correlated with Z5,Z5%pred,R5,R5%pred,R20 and R20%pred.WA5/LA5 was weakly positively correlated with Fres,R5-R20 and AX.There was no significant correlation between WT5,RVC-860 to-950 and MLDE/I and all 10 S parameters.3.In the binary multi-factor logical regression analysis based on CT quantitative parameters with groups as dependent variables,the results show that LAA%LL,LAA%RUL,WT5 and MLDE/I are independent predictors for the diagnosis of mild to moderate COPD.4.According to the analysis of ROC curve,the nomogram prediction model AUC of this study is 0.961.The sensitivity,specificity and accuracy of the line chart prediction model are 88.2%,92.9%and 90.8%,respectively.The model shows good differentiation,calibration,clinical efficiency and high C index(C-index=0.946)in internal verification.Conclusion1.The quantitative CT parameters of patients with mild to moderate COPD have a certain correlation with lung function indicators reflecting small airway disease.Among them,emphysema parameters can well evaluate of the degree of airflow limitation,and bronchial parameters can more reflect the change of small airway resistance than emphysema parameters.2.In the multivariate logistic regression analysis based on quantitative CT parameters,LAA%LL,LAA%RUL,WT5 and MLDE/I were independent predictors of COPD diagnosis,indicating that emphysema in left and right upper lobe,increase of bronchial wall thickness and air trapping were the most important CT features in the diagnosis of mild to moderate COPD.3.The nomogram prediction model of COPD based on CT quantitative parameters is developed and verified in this study.The model has good differentiation,calibration and clinical efficiency,so it can be used as a relatively accurate clinical diagnosis tool for mild to moderate COPD based on CT examination.