Microscopic Small Airway Abnormalities Identified in Early Idiopathic Pulmonary Fibrosis In Vivo Using Endobronchial Optical Coherence Tomography

Abstract

Rationale: Idiopathic pulmonary fibrosis (IPF) affects the subpleural lung but is considered to spare small airways. Micro-computed tomography (micro-CT) studies demonstrated small airway reduction in end-stage IPF explanted lungs, raising questions about small airway involvement in early-stage disease. Endobronchial optical coherence tomography (EB-OCT) is a volumetric imaging modality that detects microscopic features from subpleural to proximal airways. Objectives: In this study, EB-OCT was used to evaluate small airways in early IPF and control subjects in vivo. Methods: EB-OCT was performed in 12 subjects with IPF and 5 control subjects (matched by age, sex, smoking history, height, and body mass index). Subjects with IPF had early disease with mild restriction (FVC: 83.5% predicted), which was diagnosed per current guidelines and confirmed by surgical biopsy. EB-OCT volumetric imaging was acquired bronchoscopically in multiple, distinct, bilateral lung locations (total: 97 sites). IPF imaging sites were classified by severity into affected (all criteria for usual interstitial pneumonia present) and less affected (some but not all criteria for usual interstitial pneumonia present). Bronchiole count and small airway stereology metrics were measured for each EB-OCT imaging site. Measurements and Main Results: Compared with the number of bronchioles in control subjects (mean = 11.2/cm³; SD = 6.2), there was significant bronchiole reduction in subjects with IPF (42% loss; mean = 6.5/cm³; SD = 3.4; P = 0.0039), including in IPF affected (48% loss; mean: 5.8/cm³; SD: 2.8; P < 0.00001) and IPF less affected (33% loss; mean: 7.5/cm³; SD: 4.1; P = 0.024) sites. Stereology metrics showed that IPF-affected small airways were significantly larger, more distorted, and more irregular than in IPF-less affected sites and control subjects. IPF less affected and control airways were statistically indistinguishable for all stereology parameters (P = 0.36-1.0). Conclusions: EB-OCT demonstrated marked bronchiolar loss in early IPF (between 30% and 50%), even in areas minimally affected by disease, compared with matched control subjects. These findings support small airway disease as a feature of early IPF, providing novel insight into pathogenesis and potential therapeutic targets.

Keywords: idiopathic pulmonary fibrosis; interstitial lung disease; in vivo microscopy; small airways disease; usual interstitial pneumonia.

Figure 1.

Endobronchial optical coherence tomography (EB-OCT) volumetric imaging of idiopathic pulmonary fibrosis–affected (IPF-affected) and IPF-less affected imaging sites. (A) EB-OCT in an IPF-affected area shows dense, destructive fibrosis (F), with clusters of microscopic honeycomb cysts (HC). (B) Another example of EB-OCT in an IPF-affected airway, demonstrating destructive fibrosis (F), clusters of microscopic honeycomb cysts (HC), and traction bronchiolectasis. (C) EB-OCT in an IPF-less affected airway shows patchy fibrosis, with predominately nondestructive interstitial fibrosis (IF) and preserved alveoli (a) with no honeycombing.

Figure 2.

Comparison of bronchiole quantifications on endobronchial optical coherence tomography (EB-OCT) among idiopathic pulmonary fibrosis–affected (IPF-affected), IPF-less affected, and control imaging sites. Boxplot graphs display bronchiolar quantification per cubic centimeter (1 cm³) of each imaged airway (indicated by dots) across the IPF-affected, IPF-less affected, and control groups. Boxes delineate the middle 50% of data (ranging from the 25th to 75th percentile); the middle line indicates the median; and whiskers indicate the minimum and maximum numbers of bronchioles in each group, excluding outliers. Bronchiole count was significantly reduced in IPF-affected and IPF-less affected sites compared with matched control subject sites. There was no statistically significant difference in bronchiole count between IPF-affected and IPF-less affected sites.

Figure 3.

Volumetric endobronchial optical coherence tomography (EB-OCT) renderings displaying bronchioles and small airway lumen stereology for idiopathic pulmonary fibrosis–affected (IPF-affected), IPF-less affected, and control imaging sites. Small airway lumen segmentation is indicated in red, and bronchioles are indicated in blue. (A–C) Representative example volumetric EB-OCT images of (A) an IPF-affected site demonstrates a larger and more distorted lumen and bronchioles, compared with (B) IPF-less affected and (C) matched control airways. Airway lumen stereology of the IPF-less affected airway is comparable with that of the control airway, as demonstrated by the EB-OCT volumes in (B) and (C). Bronchiole number is visibly reduced in both (A) IPF-affected and (B) IPF-less affected sites, compared with (C) control subject sites.

Figure 4.

Comparison of small airway stereology metrics on endobronchial optical coherence tomography (EB-OCT) among idiopathic pulmonary fibrosis–affected (IPF-affected), IPF-less affected, and control imaging sites. (A–F) Boxplot graphs show cross-sectional stereology measurements averaged across each imaged small airway site (indicated by dots), demonstrating that IPF-affected airways had the largest and most irregular lumens compared with IPF-less affected and control small airways. Boxes delineate the middle 50% of data (ranging from the 25th to the 75th percentile); the middle line indicates the median; and whiskers indicate the minimum and maximum measurements in each group, excluding outliers. Kruskal-Wallis tests and post hoc Bonferroni adjustment revealed significant increases in (A) luminal area, (B) perimeter, (C) Feret diameter, and (F) surface irregularity in IPF-affected airways, compared with IPF-less affected and control small airways. Statistical analyses also revealed a significant reduction in (D) circularity and (E) roundness in IPF-affected airways, compared with IPF-less affected and control airways. IPF-less affected airways and control airways were statistically indistinguishable across all stereology parameters.