SS 304 - Interstitial lung disease (ILD) and COPD: quantification and function
SS 304 - Interstitial lung disease (ILD) and COPD: quantification and functionWednesday, March 1, 14:00 - 15:30 Room: O Session Type: Scientific Session Topics: Computer Applications, Chest Moderators: J. Broncano (Cordoba/ES), P. A. Grenier (Paris/FR) Add session to my schedule In your schedule (remove)
Is CT visual score able to demonstrate significant differences in distribution and progression of IPF abnormalities on the basis of smoking habit and treatment?
Purpose: To evaluate the role of smoking habit in patients with idiopathic pulmonary fibrosis (IPF) regard to the distribution of the main CT abnormalities (honeycombing and reticulation) and their progression during the follow-up. To assess the mortality of these patients in relation to smoking habit and treatment.
Methods and Materials: 56 patients with multidisciplinary diagnosis of IPF were distributed into three groups on the basis of smoking habit (18 non-smokers, 14 ex-smokers less than 20 pack-year (PY) and 24 ex-smokers more than 20 pack-year). We also classified the population in two groups; treated with Pirfenidone and untreated. The distribution of different abnormalities were visually scored by a radiologist at baseline and during the follow-up. The correlation with mortality rate were assessed.
Results: In the 3 groups no significant differences in the distribution of reticulation and honeycombing were found. A significant increase of honeycombing and reticulation in ex-smokers was demonstrated. The percentage of reticulation and honeycombing increases faster in untreated patients; the progression was quantified in 1,6 versus 2,3 points/year for honeycombing and 0,5 vs 3 points/year for reticulation. No significant difference in mortality rate was observed.
Conclusion: In our study CT visual score was not able to demonstrate any significant difference of IPF abnormalities distribution on the basis of smoking habit. Conversely a greater progression of abnormalities was showed in heavy smokers and untreated patients.
Volumetric texture analysis in the characterisation of interstitial lung diseases in lung transplant patients
Purpose: Diagnosing usual interstitial pneumonia (UIP) with chest CT is fundamental because of the new treatment options that could modify the prognosis of this disease; however, it still represents a challenge for radiologists. The aim of this work was to investigate the possibility of volume texture analysis (TA) in differentiating UIP from non-UIP fibrosis, comparing TA parameters with histological specimens.
Methods and Materials: We selected the pre-transplant chest CT examinations of 47 patients (35 males); eighteen of them (38.3%) suffered from UIP. The lung was segmented into three regions of interest (ROI), the apex (A-ROI), the middle (MD-ROI) and the lower lobe region (B-ROI), using the lung autosegmentation of the FocalPro software, Elekta ©. All the ROIs were analysed with LifeX Software © and TA parameters were extracted, including the grey level co-occurrence matrix feature (GLCM), neighbourhood grey-level dependence matrix (NGLDM), grey-level run length matrix (GLRLM), grey-level zone length matrix (GLZLM), sphericity and indices from the grey-level histogram. Univariate and multivariate analyses, with ROC curve calculation, between all these parameters and the pathological evidence of UIP/non-UIP fibrosis were performed.
Results: Only B-ROI TA parameters resulted significantly correlated with the diagnosis in univariate analysis (histogram kurtosis, p:0.016; GLCM-contrast, p:0.016; GLCM-dissimilarity, p:0.011; NGLDM-contrast, p:0.034; GLZLM-zp, p:0.042). The logistic regression analysis produced significant results for TA parameters B-ROI histogram skewness (p:0.011), GLCM-dissimilarity (p:0.005), LGRE (p:0.020) and A-ROI GLRLM-LRE (p:0.020), with an R2 of 0,645 and an AUC of 0.897.
Conclusion: Volumetric TA seems to improve the differential diagnosis of pulmonary fibrosis.
Utility of 3D computer-aided diagnosis system for pulmonary functional loss and treatment response assessments in connective tissue disease patients
Purpose: To evaluate the capability of a 3D computer-aided diagnosis (CAD) system of thin-section CT for quantitative pulmonary functional loss and treatment response assessment in connective tissue disease (CTD) patients.
Methods and Materials: 37 consecutive CTD patients underwent thin-section CT and pulmonary function test as follow-up examinations. In this study, total 135 follow-up examinations were divided as following three groups at each time point: stable (n=103), acute exacerbation (n=16), and after treatment (n=16) phases. In this study, all CT data were analysed by our proprietary software, and percentages of following six volume extents to total lung volume were automatically calculated in each CTD patient: normal lung, GGO and reticulation, nodular, honeycombing, consolidation, and emphysema. To determine the capability of pulmonary functional loss assessments, step-wise regression analyses were performed. To evaluate the capability for treatment response assessment, each volume and pulmonary functional changes were compared among three groups by Tukey’s HSD test.
Results: In the step-wise regression test, VC change was significantly affected by the following two factor changes: the first-step factor was GGO and reticulation, and the second-step factor was honey comb (r=0.39, p<0.05). On comparison of each volume change, stable and after treatment phase groups had significant differences with acute exacerbation phase group (normal lung, GGO and reticulation, honeycomb and consolidation: p<0.0001). In addition, consolidation volume change had significant difference between stable and after treatment phase groups (p=0.0002).
Conclusion: 3D CAD system for thin-section CT has a potential for pulmonary functional loss and treatment response assessments in CTD patients.
Interstitial lung disease in systemic sclerosis: comparison between quantitative CT and visual-scoring system before and after therapy
Purpose: To compare quantitative computed tomography (qCT) analysis with semiquantitative visual analysis for the evaluation of interstitial lung disease (ILD) in systemic sclerosis (SS). To evaluate the role of radiological scores during treatment follow-up.
Methods and Materials: 15 patients with ILD in SS were evaluated both clinically and radiologically with high-resolution computed tomography (HRCT) scans before and during treatment with immunosuppressive therapy. Semiquantitative evaluation was performed by 2 chest radiologists independently using Goh’s visual-scoring system. Qualitative analysis was made using texture-analysis software CALIPER (Imbio, Minnesota), which can assess extent of 5 different patterns: normal, ground-glass, reticular, honeycombing, and hyperlucent. Results were compared with respiratory function tests (RFT).
Results: 30 HRCT scans were evaluated. For the semiquantitative analysis overall interobserver agreement between radiologists was good, whereas agreement between Goh’s scoring system and CALIPER was fair for ground-glass and poor for reticular. The extent of ground-glass and reticular patterns were inversely correlated with RFT both with Goh’s score and CALIPER, while the latter showed greater correlation at the analysis of Pearson’s coefficient (P<0.001 for FVC and P=0.001 for DLCO). In particular, patients with FVC<79% and DLCO<50% showed significantly greater extent of each pattern of disease for both scores (P<0.05). ROC curve analysis did not identify any cutoff of radiological involvement extent to predict functional deterioration.
Conclusion: Both scores showed significant correlation between the extent of ILD and RFT. Despite the fair agreement with Goh’s score, CALIPER showed good performance in assessing disease extent and a greater correlation with RFT than Goh’s score.
Three dimensional structures of alveoli and alveolar ducts on magnified 3D print model based on micro CT of the lung specimen
Purpose: Virtual microscopic images of the peripheral lung specimen by micro CT disclosed structures of the alveolar duct and alveolus. However, fine 3D structures of the alveoli and alveolar ducts have not been reported. Author reported those structures in fine visible form at a glance by magnified 3D print model.
Methods and Materials: Inflated fixed lung specimen was prepared by Heitzman’s method. Size of 5 mm cubic sample was used. FOV of micro CT was 5 or 10 mm in diameter. Data were processed by the original application named as micro-NewVES and NewVES system for 3D vision and virtual endoscopy, created by co-author KM. Magnified 3D model was printed from micro CT data. Non-destructive analyses of internal structures were performed with virtual image of the magnified print model by high resolution CT for clinical use.
Results: Virtual micro endoscopic observation disclosed the orifice of the respiratory bronchiole, and then the alveolar duct was appeared. Magnified model and its virtual images disclosed structures of the alveolar ducts. The alveolar duct was constructed circular arrangement of 6 alveoli as one unit. Several units were successively connected one after another. At the dead end of the alveolar duct, there is one alveolus in the center that surrounded by 6 alveoli were found.
Conclusion: Magnified 3D model and virtual endoscopy of the model give us easy to understand complicated structures of peripheral lung at a glance. This method is useful to establish basics for image analysis of COPD and many diseases with alveolar involvement.
Quantitative CT analysis in COPD: do coexisting fibrotic changes influence pulmonary function in patients with advanced emphysema?
Purpose: Aim of this retrospective study was to find out if there is significant correlation of fibrotic changes with pulmonary function tests (FEV1) in patients with severe pulmonary emphysema (GOLD 3-4).
Methods and Materials: 86 patients with pulmonary emphysema underwent CT followed by pulmonary function tests. All CT-scans were performed with identical scanning parameters. Quantitative analysis of CT-data was performed with MeVisPULMO 3D v3.42 (Fraunhofer MEVIS, Bremen, Germany) to detect low (LAV) and high (HAV) attenuation volumes, defined by a threshold (LAV: <-950HU; HAV: >-700HU). HAV was considered the measurable correlate of fibrotic changes of lung tissue. HAV-extent was classified in three groups (<7%, 7-10%, >10% of lung volume). Peripheral lung volumes within a subpleural space of 2cm width were considered as the lungs “peel”. In multivariate regression-analysis results were correlated with FEV1 and LAV.
Results: There was no significant correlation of HAV and FEV1 compared with LAV. Even when more than 10% of all lung voxels accounted for HAV, LAV showed a significant higher negative correlation with FEV1 (r= -.309, R²=.096, p=.003) than HAV (no correlation, p=.786). The highest decrease of FEV1 was registered in expiration-CT and core volumes (r= -.377, p<.001 vs. HAV with p=.376). Neither in in-/expiration series nor in core-/peel-volumes HAV showed a significant correlation with FEV1.
Conclusion: Only emphysematous changes have a significant influence on FEV1 whilst coexisting fibrotic changes do not. In contrast to CPFE (chronic pulmonary fibrosis and emphysema) emphysema-score seems to show best correlation with lung function in COPD.
Xenon-enhanced ADCT: utility for functional and morphological assessments of smokers as compared with ventilation SPECT/CT
Purpose: To prospectively and directly compare the utility of xenon-contrast enhanced area-detector CT (Xe-ADCT) for pulmonary functional and disease severity assessments in smokers, when compared with ventilation SPECT/CT.
Methods and Materials: 46 consecutive smokers underwent prospective unenhanced and xenon-enhanced ADCTs, krypton ventilation SPECT/CT and pulmonary function tests. Xe-ADCT was generated from unenhanced and xenon-enhanced ADCT. Then, all smokers were divided into 4 groups as follows: ‘Non-COPD’, ‘Mild COPD’, ‘Moderate COPD’ and ‘Severe or Very Severe COPD’ groups. For each method, regional ventilation was assessed by 10-point scoring system on a per-lobe basis. Then, ventilated lung volume (VLV) on each method, functional lung volume (FLV) and wall area percent (WA%) in each subject were calculated according to past literatures. To evaluate the capability of each index for pulmonary functional loss assessment, all indexes were correlated with %FEV1 by step-wise regression analyses. To compare each index among all groups, Tukey’s HSD test were performed.
Results: In the step-wise regression test, %FEV1 (r=0.62, p<0.001) was significantly affected by the following three factors: the first-step factor, VLV on Xe-ADCT; the second-step factor, WA%; the third-step factor, FLV. All indexes of ‘Non-COPD’ and ‘Mild COPD’ groups had significant difference with those of ‘Severe or Very Severe COPD’ groups (p<0.05), and all indexes except VLV on SPECT/CT had significant difference between ‘Moderate COPD’ and ‘Severe or Very Severe COPD’ groups.
Conclusion: Xenon-enhanced ADCT had equal to or better capabilities for pulmonary functional and disease severity assessments in smokers as compared with ventilation SPECT/CT.
Quantitative CT analysis of pulmonary vessels using virtual gradationally peeled off lung in COPD patients: interrelation with emphysema, air trapping and pulmonary function
Purpose: To analyse quantitatively pulmonary vessel using virtual peeled off lung in COPD patients and interrelation with emphysema index (EI), air trapping index (ATI) and PFT.
Methods and Materials: 100 COPD patients (M:F=93:7; 65.2±0.8years) have non-contrast 3D inspiration/expiration CTs, PFT and DLco were enrolled. AView Lung software was used to analyse pulmonary vessel. Closed plane having an even distance (6mm~24mm, 3mm interval) from the fissure and surface regarding whole lung were automatically generated. Number of vessels per square centimeter, mean diameter of vessels, and percent of vessel area per surface were semi-automatically measured on all generated planes. EI (percentage of area HU≤-950), ATI (percentage of area subtracted HU≤60HU between inspiration CT and registered expiration CT), and mean of bronchiolar wall area percentage (WA%) were also measured. Correlations between parameters of vessels and EI, ATI, WA%, PFT, DLco were evaluated.
Results: All level of number of vessels per square centimeter and percent of vessel area per surface showed significant negative correlation with EI (r=-0.606~-0.811), ATI (r=-0.310~-0.423), RV (r=-0.331~-0.463) and positive correlation with WA% (r=0.264~0.340), and PFT parameters (r=0.243~0.552). They also showed significant positive correlation with DLco (r=0.241~0.443) except vessel area per surface on 6mm plane. Mean diameter of vessels showed significant negative correlation with EI, and ATI on 9~18mm level plane (r=-0.216~-0.370). All detailed results are showed in Table 1.
Conclusion: CT measured parameters of pulmonary vessels are related with CT measured emphysema, air trapping, airway wall thickness and clinical pulmonary functions.
Purpose: The classic cardiac complication of those with chronic obstructive pulmonary disease (COPD) is that of cor pulmonale. However this has typically been described in those with end stage disease or based on echo measures which is known to be poor for right ventricular (RV) assessment and even worse in a COPD cohort. Thus the aim of the current study was to use cardiac MRI to assess cardiac remodelling in COPD.
Methods and Materials: 21 healthy volunteers (HV) free from cardiovascular disease and 60 patients with COPD were recruited. Those with COPD were screened with echocardiography, and those with silent left ventricular (LV) systolic dysfunction excluded. All patients underwent spirometry and cardiac MRI for mass and volume assessment on the same visit.
Results: With increasing severity of COPD both left and right ventricular end diastolic volumes (EDV) fell (LVEDV: 60±10ml in HV vs. 44±18ml in severe COPD, p=0.011; RVEDV 60±13 in HV vs. 42±16 in severe COPD, p=0.017) and as did stroke volumes (SV) (LVSV: 36±5ml in HV vs. 25±10ml in severe COPD, p=0.001; RVSV 37±6 in HV vs. 24±11 in severe COPD, p=0.001). Left ventricular mass (LVM) also fell with increasing severity of COPD (LVM: 41±7g in HV vs. 33±14g in severe COPD, p=0.024), while RV mass remained unchanged (p=0.24).
Conclusion: Increasing severity of COPD results in reduced cardiac volumes and mass, suggesting the predominant cause of cardiac remodelling in COPD is due to reduced preload rather than the classical cor pulmonale model of increased afterload.
Purpose: To assess the prevalence and risk factors for CT signs of chronic lung disease in a large cohort of never-smoking HIV-infected patients.
Methods and Materials: 329 never-smoking HIV-infected patients (49.6±8.9 years, 73.9% men) underwent ECG-gated chest CT scan for the evaluation of coronary artery calcium score (CACs) and CT slices at the level of L4-L5 for the assessment of subcutaneous and visceral adipose tissue (SAT-VAT). CT signs of chronic lung disease (emphysema, respiratory bronchiolitis (RB), bronchiectasis, bronchial wall thickening, fibrosis and non-calcified nodules) were collected. Lung emphysema was graded with a semi-quantitative visual score (0-4 for each lobe). Epicardial adipose tissue (EAT) was calculated in a subset of 153 patients. Univariable and multivariable logistic regression analyses were performed to identify factors independently associated with the presence of emphysema.
Results: Emphysema was found in 52 patients (15.6%, 13 patients with score>4), RB in 39 (11.7%), bronchiectasis in 52 (15.8%) and nodules in 33 (10%). Univariable analysis showed significant associations between the presence of emphysema and: male sex (p=0.09), age (p<0.01), BMI (p=0.05), Framingham risk score (p<0.01), CACs (p<0.01), VAT (p<0.01), and EAT (p<0.01). No significant association was found with HIV-related factors and inflammatory markers. In a multivariable model, significant predictors of emphysema were: age (OR=1.098; p=0.01), and EAT (OR=1.015; p=0.03).
Conclusion: CT signs of chronic lung diseases are common in never-smoking HIV patients. Emphysema was associated with epicardial and visceral adiposity, but not with HIV-related variables, underlying a common pathogenetic mechanism linking lung CT abnormalities and ectopic fat accumulation.