Computational analysis of thoracic multidetector row HRCT for segmentation and quantification of small airway air trapping and emphysema in obstructive pulmonary disease

Abstract

Rationale and objectives: Obstructive pulmonary disease phenotypes are related to variable combinations of emphysema and small-airway disease, the latter manifested as air trapping (AT) on imaging. The investigators propose a method to extract AT information quantitatively from thoracic multi-detector row high-resolution computed tomography (HRCT), validated by pulmonary function testing (PFT) correlation.

Materials and methods: Seventeen patients with obstructive pulmonary disease who underwent HRCT and PFT within a 3-day interval were retrospectively identified. Thin-section volumetric HRCT in inspiration and expiration was registered and analyzed using custom-made software. Nonaerated regions of lung were segmented through exclusion of voxels > -50 Hounsfield units (HU); emphysematous areas were segmented as voxels < -950 HU on inspiratory images. Small-airway AT volume (ATV) was segmented as regions of lung voxels whose attenuation values increased by less than a specified change threshold (set from 5 to 300 HU in 25-HU increments) between inspiration and expiration. Inspiratory and expiratory total segmented lung volumes, emphysema volume (EV), and ATV for each threshold were subsequently calculated and correlated with PFT parameters.

Results: A strong positive correlation was obtained between total segmented lung volume in inspiration and total lung capacity (r = 0.83). A strong negative correlation (r = -0.80) was obtained between EV and the ratio between forced expiratory volume in 1 second and forced vital capacity. Stronger negative correlation with forced expiratory volume in 1 second/forced vital capacity (r = -0.85) was demonstrated when ATV (threshold, 50 HU) was added to EV, indicating improved quantification of total AT to predict obstructive disease severity. A moderately strong positive correlation between ATV and residual volume was observed, with a maximum r value of 0.72 (threshold, 25 HU), greater than that between EV and residual volume (r = 0.58). The benefit of ATV quantification was greater in a subgroup of patients with negligible emphysema compared to patients with moderate to severe emphysema.

Conclusions: Small-airway AT segmentation in conjunction with emphysema segmentation through computer-assisted methodologies may provide better correlations with key PFT parameters, suggesting that the quantification of emphysema-related and small airway-related components of AT from thoracic HRCT has great potential to elucidate phenotypic differences in patients with chronic obstructive pulmonary disease.