Abstract Severe asthma exhibits heterogeneity in airflow obstruction, driven by airway remodeling and air trapping, which can be noninvasively assessed via quantitative computed tomography (qCT). This study aimed to identify asthma phenotypes by clustering qCT measurements of airway dimensions, lung volumes, and densitometry, and to elucidate the underlying molecular pathways through sputum proteomics. We applied consensus clustering to qCT data from 239 asthma patients (severe and mild/moderate) and 68 healthy controls from the Chinese C-BIOPRED cohort. Four distinct qCT clusters emerged: Cluster 1, characterized by luminal dilation, severe air trapping, and reduced lung density; Cluster 2, with thickened airway walls and luminal narrowing without air trapping; Cluster 3, showing mild luminal dilation, preserved lung volumes, and optimal spirometry; and Cluster 4, featuring airway wall thickening, luminal narrowing, severe air trapping, and profound airflow obstruction. Sputum eosinophilia was elevated in Clusters 1 and 4. Proteomics revealed upregulated pathways in apoptosis execution and cornified envelope formation in Cluster 1, while Cluster 2 and 4 exhibited enhanced complement activation, fibrin formation, plasma lipoprotein assembly, and insulin-like growth factor (IGF) transport regulation. These findings delineate qCT-derived phenotypes and their associated underlying mechanisms of airway remodeling and airflow obstruction in severe asthma. This abstract is funded by: None
Deng et al. (2026) studied this question.