Radiomics of the Airway (RadAr): Multi-Scale Airway Phenotyping for Disease Characterization on Routine CT Imaging
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Purpose
Airway remodeling is a convergent feature across respiratory diseases, yet current CT tools provide limited characterization of the airway tree. We present Radiomics of the Airway (RadAr), an automated framework for multi-scale airway phenotyping from routine chest CT.
Methods
RadAr extracts multi-scale, interpretable airway measurements capturing luminal dimensions, tapering, architectural distortion, and global morphology and provides an interactive web portal for analysis and visualization. It was evaluated across four settings: 63-week mortality prediction in fibrotic interstitial lung disease (fILD; N=147), COVID-19 severity prediction (N=1164), structure-function association in progressive pulmonary fibrosis (PPF; N=9) and structure-inflammation markers in pediatric cystic fibrosis (CF; N=11). Unsupervised clustering identified airway phenotypes across the fILD and COVID-19 cohorts.
Results
In fILD, lower-lobe architectural distortion was associated with mortality (balanced accuracy 0.654). In COVID-19, severe disease was independently associated with luminal dilation (AUC 0.719, odds ratio 2.32, p=0.017). In PPF, airway phenotypes correlated with forced vital capacity (ρ=0.83), mid-expiratory flow (ρ=0.87), and ¹²⁹Xe MRI alveolar gas exchange impairment (ρ=0.70). In pediatric CF, reduced tapering and increased cylindricity were associated with prior exacerbations and bronchoalveolar lavage neutrophilia (ρ=-0.64 to −0.78). Five phenotypes were identified from extensive, tapered airway trees to sparse, dilated, thick-walled, tortuous trees, with increasing COVID-19 severity and fILD mortality across this spectrum.
Conclusions
RadAr identified interpretable, disease-specific airway signatures associated with function and outcomes across restrictive, obstructive, and mixed lung diseases in adult and pediatric settings. It provides a scalable framework that may support diagnosis, risk stratification, and longitudinal monitoring across pulmonary diseases.
