Quantitative Study On The Degree Of Pulmonary Lobe Injury In COPD Patients Based On CT Parameter Response Mapping

Objective:The CT parameter response mapping（PRM）was used to quantitatively evaluate the damage of each lung lobe in COPD,combined with clinical parameters（age,smoking index,BMI）and pulmonary functions（FEV1%、FEV1/FVC）to explore the application value of parametric response mapping in the early diagnosis and severity assessment of COPD,so as to provide an objective evidence for COPD classification and the degree of damage to each lung lobe.The research contents include:Part one:Based on the parameters of biphasic quantitative CT and PRM parameters,multiple quantitative indexes were used to comprehensively analyze the lung injury in smokers without COPD（non-COPD group）and different subgroups of COPD,and correlated with lung function to find the reliable imaging index for the assessment of lobar small airway disease and emphysema.Part two:Based on the CT parameter response mapping PRM^fSAD%and PRM^Emph%,the lung lobe injury in smoking non-COPD group and different subgroups of COPD was quantitatively analyzed to explore the distribution characteristics of lobar small airway disease and emphysema with the increase of disease degree.Part three:The PRM^fSAD%and PRM^Emph%parameters of each lung lobe were combined with clinical parameters and pulmonary functions to analyze the clinical application value of PRM parameters.To explore the predictive value of PRM^fSAD%and PRM^Emph%of each lung lobe on lung function.Part one1.Materials and methods56 cases of smokers without COPD who underwent biphasic chest CT examination in our hospital from June 2018 to June 2021 were included;133 cases of COPD were diagnosed and divided into 18 cases of grade I,56 cases of GOLD II,39 cases of GOLD III and 20 cases of GOLD IV according to gold guidelines.In addition,25 non-smoking normal controls who participated in physical examination were included.Import the original DICOM data of biphasic CT scanning into the digital COPD test tool for image registration and calculate the CT threshold to obtain the PRM parameters:（1）Percentage of functional small airway disease(PRM^fSAD%)（2）Percentage of emphysema area(PRM^Emph%)（3）Percentage of normal lung tissue area(PRM^Normal%).Other CT observations include,emphysema parameters:（1）CT value corresponds to the 15th percentile of the inspiratory density histogram（IN-PD15）（2）Percentage of low-attenuation lung areas less than-950HU in inspiration phase（IN-950%）（3）Percentage of low-attenuation lung areas less than-910 HU in expiratory phase（EX-910%）.Air trapping parameters:（1）Percent of CT scan low attenuation area below-856 HU at end-expiratory（EX-856%）（2）the expiratory to inspiratory ratio of lung volume（LVEX/IN）（3）the expiratory to inspiratory ratio of mean lung density（MLDEX/IN）.Pulmonary function parameters included FVC,FEV1%and FEV1/FVC.Clinical parameters included age,BMI,and smoking index.The One-way analysis of variance or the Kruskal-Wallis H test was used to analyze differences in quantitative CT parameters and lung function between groups.Pairwise comparisons were performed by LSD method or pairwise comparison method.The correlation between the pulmonary function and CT quantitative parameters was analyzed by using Spearman.2.Results2.1 General information comparisonThere were significant differences in age,BMI,smoking index,and lung function（FVC,FEV1%,FEV1/FVC）between groups（P<0.001）.The COPD group was older than the smoking non-COPD group and the normal control group.The smoking index was higher in GOLD III/IV grades.BMI was the lowest in the GOLD IV grade.FEV1/FVC decreased with increasing disease severity.FVC was lower with GOLD III/IV.2.2 CT quantitative parameters comparisonThere were statistical differences in the CT quantitative parameters between the groups（P<0.001）.The results of CT emphysema parameters showed that all parameters in smoking non-COPD group were higher than those in normal control group.Except for the IN-950%parameter COPD I grade was lower than the non-COPD group,the remaining parameters IN-PD15 and EX-910%were higher in the COPD group.With the increase of gold grade,the EX-910%increases.The results of CT air trapping parameters showed that all parameters in the smoking non-COPD group were higher than those in the normal control group;EX-856%,LVEX/IN,MLDEX/IN were higher in COPD group than in non-COPD group.With the increase of GOLD grade,the LVEX/IN and MLDEX/IN of values increase.The PRM parameter results showed that the PRM^Emph%and PRM^fSAD%parameters gradually increased from the normal control group,smoking non-COPD,and GOLD I-IV grades.2.3 Correlation analysis of CT quantitative parameters and pulmonary functionEmphysema parameters were moderately to strongly correlated with FEV1%and FEV1/FVC（rs=-0.408～-0.648,P<0.001）.Air trapping parameters were moderately to strongly correlated with FEV₁%and FEV₁/FVC（rs=-0.441～-0.686,P<0.001）.PRM^Emph%and PRM^fSAD%were moderately to strongly correlated with FEV₁%and FEV₁/FVC（rs=-0.574～-0.665,P<0.001）.Except that there was no correlation between CT quantitative indexes and lung function（FEV1%,FEV1/FVC）in the normal control group,CT quantitative parameters were correlated with lung function（FEV1%,FEV1/FVC）in the smoking non-COPD group and different subgroups of COPD（P<0.05）.3.Brief summaryBased on biphasic CT,multiple quantitative indexes can reflect the characteristics of lung injury with different degrees of airflow restriction,and have a good correlation with lung function.PRM parameters can clearly distinguish small airway disease and emphysema areas,which can be used as reliable indicators to study the distribution characteristics of lung lobe injury.Part two1.Materials and methodsIncluded in the same as part one.Two-phase registration PRM parameters PRM^fSAD%,PRM^Emph%measured in each lung lobe,including right upper,middle and lower lobes（RUL,RML,RLL）and left upper and lower lobes（LUL,LLL）.Differences in PRM^fSAD%and PRM^Emph%（heterogeneity of variance）of each lung lobe between different groups were tested by Kruskal-Wallis H test.The paired comparison method was used for pairwise comparison.P<0.05 is considered statistically significant difference.2.ResultsThere were significant differences in the PRM^fSAD%and PRM^Emph%of each lung lobe in the normal control group,the smoking non-COPD group and the COPD group of GOLD I-IV grades（all P<0.001）.Among the different GOLD grades of PRM^fSAD%and PRM^Emph%,the damage to the right middle lobe and upper lobe of both lungs was severely injured.Pairwise comparison showed that there were significant differences in PRM^fSAD%of each lung lobe between the normal control group and the smoking non-COPD group（all P<0.05）.There was no statistically significant difference between right middle lobe PRM^fSAD%,but there were significant differences in other lung lobes between smoking non-COPD group and COPD group GOLD grade I（all P<0.05）.There was significant difference in PRM^fSAD%of the lower lobe of both lungs between GOLD I and GOLD II（P<0.05）.There was significant difference in PRM^fSAD%of the lower lobe of right lung between GOLD III and GOLD IV（P<0.05）.There were statistically significant differences in each lobe of PRM^Emph%between the normal control group and the smoking non-COPD group（all P<0.05）.There were statistically significant differences in each lobe of PRM^Emph%between the non-COPD group and GOLDⅠ（all P<0.05）.There was significant difference in PRM^Emph%of the lower lobe of both lungs between GOLD I and GOLD II（P<0.05）.There was significant difference in PRM^Emph%of the lower lobe of left lung between GOLD III and GOLD IV（P<0.05）.The comparison among the PRM^fSAD%and PRM^Emph%ofeach lobe showed that the differences in other groups were statistically significant（P<0.01）,except for the PRM^fSAD%and PRM^Emph%of GOLD IV（P=0.395and 0.840）.3.Brief summaryThe distribution of small airway disease and emphysema in each lung lobe was different in different degrees of airflow limitation groups.Smoking non COPD were mainly small airway disease in the middle lobe of the right lung.As COPD severity increases,in GOLDI/II grades,the small airway disease and emphysema mainly occurred in the right middle lobe and both lungs of upper lobes.In GOLDIII/IV grades,small airway disease and emphysema were mainly in all lobes of the whole lung.The degree of injury lesions in each lobe of GOLD grade IV was similar.Part three1.Materials and methodsThe included object is the same as the first part.Smoking non-COPD patients（56cases）and non-smoking normal controls（25 cases）were defined as non COPD patients（81 cases）.The patients diagnosed with COPD（133 cases）,GOLD I/II grades were defined as mild to moderate COPD group（74 cases）,GOLD III/IV grades were defined as severe COPD group（59 cases）.The clinical parameters and pulmonary function parameters are the same as the first part,and the PRM parameters of each lobe are the same as the second part.1.1 Spearman was used to analyze the correlation between PRM^fSAD%,PRM^Emph%and pulmonary function,BMI,smoking index and age.Stepwise regression analysis was performed with PRM^fSAD%and PRM^Emph%of each lung lobe as dependent variables and clinical parameters（BMI,smoking index,age）as independent variables.1.2 Taking FEV₁/FVC and FEV_l%as state variables and PRM^fSAD%and PRM^Emph%of each lung lobe of CT as test variables,draw the receiver performance curve（ROC）,calculate the area under the curve,and obtain the best cut-off value,sensitivity and specificity of lung lobe PRM^fSAD%and PRM^Emph%to predict FEV₁/FVC<0.7（distinguish COPD patients among all subjects）and FEVl%<50%（distinguish severe COPD among all COPD patients）.2.Results2.1 Correlation analysis results of PRM^fSAD%and PRM^Emph%of lung lobe with pulmonary function,Clinical parametersThere was a weak to strong negative correlation between PRM^fSAD%,PRM^Emph%and pulmonary function in each lobe（rs=-0.229～-0.725,P<0.001）;BMI and smoking index were highly correlated with PRM^Emph%of each lobe（rs=-0.415～-0.509,P<0.001）.There was a weak to moderate positive correlation between age and PRM^fSAD%,PRM^Emph%in each lobe（rs=0.291～0.439,P<0.001）.2.2 Multiple linear stepwise regression analysisMultiple linear stepwise regression analysis was used to explore the relationship between PRM parameters and clinical indexes.The results showed a linear relationship between age and lung lobes（RUL,RML,LUL）PRM^fSAD%（R²=0.163～237）.Age,smoking index,BMI were linearly associated with RLL PRM^fSAD%（R²=0.279,standardized beta values were 0.363,0.188,-0.182 respectively）.Age,smoking index were linearly associated with LLL PRM^fSAD%（R²=0.265,standardized beta values were 0.411and 0.210 respectively）.Smoking index and BMI were linearly correlated with lung lobes（RML,RLL,LLL）PRM^Emph%（R²=0.222～312）.Smoking index,BMI,age and RUL PRM^Emph%（R²=0.394,standardized beta values were 0.348,-0.330,0.216）,LUL PRM^Emph%（R²=0.404,standardized beta values were 0.348,-0.325 and 0.234 respectively）.All the above regression models were statistically significant in the ANOVA test（P<0.001）.2.3 The value of PRM^fSAD%and PRM^Emph%of each lobe in predicting lung function2.3.1 For all subjects,PRM parameters of all lung lobes have predictive value（AUC0.774～0.874）in the prediction of FEV1/FVC<0.7（distinguish patients with COPD）.The area under the PRM^Emph%（AUC 0.874）and PRM^fSAD%（AUC 0.868）curves of LLL is the largest and the accuracy is the highest.When the PRM^Emph%of LLL was 3.54%,the predictive sensitivity was 86.3%and the specificity was 74.40%.when the PRM^fSAD%of LLL was 14.37%,the sensitivity was 88.7%and the specificity was 78.90%.2.3.2 For all COPD patients,When the PRM parameters of each lobe predict FEV1%<50%（distinguishing severe COPD patients）,except for the PRM^fSAD%in the right upper lobe（AUC 0.596）,the remaining lobes have predictive value（AUC0.660～0.720）.Among them,the PRM^fSAD%of LLL（AUC 0.720）and PRM^Emph%of RLL（AUC 0.720）have the largest area under the curve,the highest accuracy.When the PRM^fSAD%of LLL was 27.13%,the predictive sensitivity was 74.60%and the specificity was 70.00%;when the PRM^Emph%of RLL was 20.56%,the predictive sensitivity was46.00%and the specificity was 91.40%.3.Brief summaryAge was mainly associated with PRM^fSAD%in each lobe,followed by PRM^Emph%in both upper lobes.BMI and smoking index were mainly associated with PRM^Emph%of each lung lobe,followed by PRM^fSAD%of lower lung.For all subjects,When the PRM^Emph%and PRM^fSAD%of the left lower lobe were respectively 3.54%and 14.37%,the AUC was 0.874 and 0.868,the sensitivity was 86.3%and 88.7%,and the specificity was 74.40%and 78.90%in the prediction of FEV1/FVC<0.7（distinguishing COPD patients）.For all COPD patients,When the PRM^fSAD%of the left lower lobe and the PRM^Emph%of the right lower lobe were respectively 27.13%and20.56%,the AUC was 0.720 and 0.720,and the sensitivity was 74.60%and 46.00%,specificity was 70.00%and 91.40%in the prediction of FEV1%<50%（distinguishing severe COPD patients）.Conclusions1)Based on dual gas phase CT,multiple quantitative indexes can reflect the characteristics of lung injury in smoking non-COPD and different GOLD grades of COPD.PRM parameters are superior to other parameters in clearly distinguishing small airway disease and emphysema which can be used as reliable indicators to study the distribution characteristics of lung lobe injury.2)PRM parameters PRM^fSAD%and PRM^Emph%can effectively evaluate the extent and progression of small airway disease and emphysema in different lobes in smoking non-COPD and different GOLD grades of COPD.Smoking non-COPD were mainly small airway disease in the middle lobe of the right lung.With the increase of COPD severity,in GOLDI/II grades,Small airway disease and emphysema mainly occurred in the right middle lobe and the upper lobes of both lungs.In GOLDIII/IV grades,small airway disease and emphysema are mainly in all lobes of the whole lung.In GOLDIV grade,the degree of injury lesions in each lobe was similar.3)The PRM parameters of each lobe had a good correlation with clinical parameters and pulmonary function indexes.The left lower lobe PRM^fSAD%and PRM^Emph%were more valuable in predicting FEV₁/FVC<0.7,and the right lower lobe PRM^Emph%and left lower lobe PRM^fSAD%were more valuable in predicting FEV₁%<50%.Age mainly affects lobe small airway disease,followed by bilateral upper lobe emphysema.BMI and smoking index mainly affected lobar emphysema,followed by lower lung small airway disease.