Bullous Parametric Response Map for Functional Localization of COPD

Advanced bronchoscopic lung volume reduction treatment (BLVR) is now a routine care option for treating patients with severe emphysema. Patterns of low attenuation clusters indicating emphysema and functional small airway disease (fSAD) on paired CT, which may provide additional insights to the target selection of the segmental or subsegmental lobe of the treatments, require further investigation. The low attenuation clusters (LACS) were segmented to identify the scalar and spatial distribution of the lung destructions, in terms of 10 fractions scales of low attenuation density (LAD) located in upper lobes and lower lobes. The LACs of functional small airway disease (fSAD) were delineated by applying the technique of parametric response map (PRM) on the co-registered CT image data. Both emphysematous LACs of inspiratory CT and fSAD LACs on expiratory CT were used to derive the coefficients of the predictive model for estimating the airflow limitation. The voxel-wise severity is then predicted using the regional LACs on the co-registered CT to indicate the functional localization, namely, the bullous parametric response map (BPRM). A total of 100 subjects, 88 patients with mild to very severe COPD and 12 control participants with normal lung functions (FEV1/FVC % > 70%), were evaluated. Pearson’s correlations between FEV1/FVC% and LAV%HU-950 of severe emphysema are − 0.55 comparing to − 0.67 and − 0.62 of LAV%HU-856 of air-trapping and LAV%fSAD respectively. Pearson’s correlation between FEV1/FVC% and FEV1/FVC% predicted by the proposed model using LAD% of HU-950 and fSAD on BPRM is 0.82 (p < 0.01). The result of the Bullous Parametric Response Map (BPRM) is capable of identifying the less functional area of the lung, where the BLVR treatment is aimed at removing from a hyperinflated area of emphysematous regions.

Interactively visualizing distributional regression models with distreg.vis

A newly emerging field in statistics is distributional regression, where not only the mean but each parameter of a parametric response distribution can be modelled using a set of predictors. As an extension of generalized additive models, distributional regression utilizes the known link functions (log, logit, etc.), model terms (fixed, random, spatial, smooth, etc.) and available types of distributions but allows us to go well beyond the exponential family and to model potentially all distributional parameters. Due to this increase in model flexibility, the interpretation of covariate effects on the shape of the conditional response distribution, its moments and other features derived from this distribution is more challenging than with traditional mean-based methods. In particular, such quantities of interest often do not directly equate the modelled parameters but are rather a (potentially complex) combination of them. To ease the post-estimation model analysis, we propose a framework and subsequently feature an implementation in R for the visualization of Bayesian and frequentist distributional regression models fitted using the bamlss, gamlss and betareg R packages.

A 3D-CNN model with CT-based parametric response mapping for classifying COPD subjects

Chronic obstructive pulmonary disease (COPD) is a respiratory disorder involving abnormalities of lung parenchymal morphology with different severities. COPD is assessed by pulmonary-function tests and computed tomography-based approaches. We introduce a new classification method for COPD grouping based on deep learning and a parametric-response mapping (PRM) method. We extracted parenchymal functional variables of functional small airway disease percentage (fSAD%) and emphysema percentage (Emph%) with an image registration technique, being provided as input parameters of 3D convolutional neural network (CNN). The integrated 3D-CNN and PRM (3D-cPRM) achieved a classification accuracy of 89.3% and a sensitivity of 88.3% in five-fold cross-validation. The prediction accuracy of the proposed 3D-cPRM exceeded those of the 2D model and traditional 3D CNNs with the same neural network, and was comparable to that of 2D pretrained PRM models. We then applied a gradient-weighted class activation mapping (Grad-CAM) that highlights the key features in the CNN learning process. Most of the class-discriminative regions appeared in the upper and middle lobes of the lung, consistent with the regions of elevated fSAD% and Emph% in COPD subjects. The 3D-cPRM successfully represented the parenchymal abnormalities in COPD and matched the CT-based diagnosis of COPD.