Deformation-Based Morphometry Adapted for Lung CT

• Activity Number: 297 - SPEED: Food, Environment, Biomedical Imaging and Physical System Visualization/Learning, Part 1
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Imaging
• Author(s): Sarah Ryan* and Tasha Fingerlin and Nichole E Carlson and Lisa Maier
• Companies: and National Jewish Health and University of Colorado Anschutz and National Jewish Health
• Keywords: lung; computed tomography; deformation-based morphometry; volume; pulmonary sarcoidosis

Abstract

The human lung is a highly flexible organ whose structure is influenced by demographics and disease. In this work, we adapt Deformation-Based Morphometry (DBM), a unified statistical framework for analyzing structural differences in anatomy, to computed tomography scans of the lungs. Our method, implemented in the lungct R package, includes segmentation of the left and right lungs, and non-linear deformation of the segmentations to the standard lung mask. We capture differences in lung structure by computing the rate of change on the deformation field at every voxel. To show the functionality of our method, we apply it to a population of subjects with pulmonary sarcoidosis enrolled in the NHLBI-funded GRADS study (N=330), as this disease is known to alter lung structure. We found that sarcoidosis subjects have significantly smaller lungs than healthy controls (p< 0.05), with the largest lung shrinkage found near the hila region of the right lung. Furthermore, we found that certain visual patterns, such as traction bronchiectasis, result in more severe lung shrinkage patterns. We conclude that DBM is a useful tool to understand anatomical differences in the lung.