Abstract
Chronic obstructive pulmonary disease (COPD) is a heterogeneous disease and the traditional variables extracted from computed tomography (CT) images may not be sufficient to describe all the topological features of lung tissues in COPD patients. We employed an unsupervised three-dimensional (3D) convolutional autoencoder (CAE)-feature constructor (FC) deep learning network to learn from CT data and derive tissue pattern-clusters jointly. We then applied exploratory factor analysis (EFA) to discover the unobserved latent traits (factors) among pattern-clusters. CT images at total lung capacity (TLC) and residual volume (RV) of 541 former smokers and 59 healthy non-smokers from the cohort of the SubPopulations and Intermediate Outcome Measures in the COPD Study (SPIROMICS) were analyzed. TLC and RV images were registered to calculate the Jacobian (determinant) values for all the voxels in TLC images. 3D Regions of interest (ROIs) with two data channels of CT intensity and Jacobian value were randomly extracted from training images and were fed to the 3D CAE-FC model. 80 pattern-clusters and 7 factors were identified. Factor scores computed for individual subjects were able to predict spirometry-measured pulmonary functions. Two factors which correlated with various emphysema subtypes, parametric response mapping (PRM) metrics, airway variants, and airway tree to lung volume ratio were discriminants of patients across all severity stages. Our findings suggest the potential of developing factor-based surrogate markers for new COPD phenotypes.
Highlights
Chronic obstructive pulmonary disease (COPD) is a heterogeneous disease and the traditional variables extracted from computed tomography (CT) images may not be sufficient to describe all the topological features of lung tissues in COPD patients
Since the introduction of the parametric response map (PRM)[4] that matches CT images at inspiration and expiration to identify voxels of functional small airway disease and emphysema (Emph; i.e. emphysematous region where CT intensity is less than − 950 HU on total lung capacity (TLC) image), a number of studies have focused on small airways using this voxel-wise imaging analysis technique[5]
In this study, we proposed an unsupervised three-dimensional (3D) convolutional autoencoder (CAE) model integrated with a feature constructor (FC) as a classifier to identify lung tissue pattern-clusters from one-dimensional (1D) representations of 3D regions of interests (ROIs) extracted from CT images of former smokers
Summary
Chronic obstructive pulmonary disease (COPD) is a heterogeneous disease and the traditional variables extracted from computed tomography (CT) images may not be sufficient to describe all the topological features of lung tissues in COPD patients. In this study, we proposed an unsupervised three-dimensional (3D) convolutional autoencoder (CAE) model integrated with a feature constructor (FC) as a classifier (a CAE-FC model) to identify lung tissue pattern-clusters from one-dimensional (1D) representations (embeddings) of 3D regions of interests (ROIs) extracted from CT images of former smokers. We applied an exploratory factor analysis (EFA) to explore the latent traits (factors) among these pattern-clusters and further assessed the capability of factors in predicting or diagnosing major clinical characteristics of COPD patients. These factors could be employed for unbiased sensitive subtyping of COPD progression without using human pre-defined features in the future
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
