Identification of COPD From Multi-View Snapshots of 3D Lung Airway Tree via Deep CNN

Abstract

Chronic obstructive pulmonary disease (COPD) is associated with morphologic abnormalities of airways with various patterns and severities. However, the way of effectively representing these abnormalities is lacking and whether these abnormalities enable to distinguish COPD from healthy controls is unknown. We propose to use deep convolutional neural network (CNN) to assess 3D lung airway tree from the perspective of computer vision, thereby constructing models of identifying COPD. After extracting airway trees from CT images, snapshots of their 3D visualizations are obtained from ventral, dorsal and isometric views. Using snapshots of each view, one deep CNN model is constructed and further optimized by Bayesian optimization algorithm to indentify COPD. The majority voting of three views presents the final prediction. Finally, the class-discriminative localization maps have been drawn to visually explain the CNNs' decisions. The models trained with single view (ventral, dorsal and isometric) of colorful snapshots present the similar accuracy (ACC) (86.8%, 87.5% and 86.7%) and the model after voting achieves the ACC of 88.2%. The ACC of the final voting model using gray and binary snapshots achieves 88.6% and 86.4%, respectively. Our specially designed CNNs outperform the typical off-the-shelf CNNs and the pre-trained CNNs with fine tuning. The class-discriminative regions of COPD are mainly located at central airways; however, regions in HC are scattering and located at peripheral airways. It is feasible to identify COPD using snapshots of 3D lung airway tree extracted from CT images via deep CNN. The CNNs can represent the abnormalities of airway tree in COPD and make accurate CT-based diagnosis of COPD. INDEX TERMS Chronic obstructive pulmonary disease (COPD), deep learning, convolutional neural networks, computed tomography (CT), airway, image classification.

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

“…The advantage of the proposed 3D-cPRM approach, which achieved a detection accuracy of 89.3%, was confirmed above. Our result is comparable to that of a previous CNN study 17 and far exceeded that of multi-view snapshots of a 3D lung-airway tree (24/596 FPs in our method, vs. 26/280 FPs and an accuracy of 88.6% 18 ). Therefore, a 3D-CNN can effectively eliminate the false-positive results generated by 2D-CNN.…”

“…A deep CNN better represents these abnormalities from 3D PRM images than from 2D PRM images; in the former case, the classification accuracy of COPD versus non-COPD reached 89.3%. To our knowledge, our method identifies COPD patients at least as accurately as previous classification approaches [17][18][19]27 . A strong positive correlation was found between some combinations of PRM phenotypes and 3D CNNs.…”

“…Our 3D-cPRM method avoids these difficulties by efficiently distinguishing PRM abnormalities in COPD. The developed CNN fully exploits the available CT data, rather than pre-selecting 2D slices or snapshot images for 2D CNNs 18 , or directly inputting CT slices to a 2D-CNN 17 .…”

“…Our 3D-cPRM method avoids these difficulties by efficiently distinguishing PRM abnormalities in COPD. The developed CNN fully exploits the available CT data, rather than pre-selecting 2D slices or snapshot images for 2D CNNs 18 , or directly inputting CT slices to a 2D-CNN 17 . www.nature.com/scientificreports/ Table 3 overviews the datasets, methods, and classification results of some recent deep-learning studies that distinguish COPD from non-COPD subjects.…”

“…Using CNN models and four canonical CT slices at predefined anatomic landmarks in the COPDGene and ECLIPSE testing cohorts, González et al 17 achieved a COPD detection accuracy of 0.773. Du et al 18 applied a CNN on 2D multi-view snapshots of a 3D lung-airway tree, and classified COPD and non-COPD cases. They reported an accuracy of 88.6% on grayscale snapshots.…”