Pulmonary MRI and Cluster Analysis Help Identify Novel Asthma Phenotypes

Area of Science:

• Pulmonary Medicine
• Radiology
• Biomarkers

Background:

• Current asthma phenotypes lack strong links to disease mechanisms or treatment outcomes.
• While CT scans offer insights, functional MRI provides complementary phenotypic data.
• Eosinophilia is a known asthma phenotype, but others require better definition.

Purpose of the Study:

• To develop data-driven asthma clusters using functional MRI (magnetic resonance imaging) airway biomarkers.
• To better characterize airway pathologies in asthma patients through novel phenotyping.

Main Methods:

• Utilized retrospective data from 45 asthma patients undergoing post-bronchodilator 129Xe MRI, CT, spirometry, and plethysmography.
• Assessed MRI ventilation defect percent (VDP) and CT-derived airway measurements (WA%, WT, LA, D, TAC).
• Applied k-means cluster analysis to MRI and CT biomarkers to generate phenotypic subgroups.

Main Results:

• Four distinct asthma clusters were identified using MRI VDP, TAC, WT*, and D*.
• Significant differences in imaging measurements and lung function (FEV1, RV/TLC, airway resistance) were observed across clusters.
• No significant differences in asthma control, quality of life, or severe asthma proportions were found between clusters.

Conclusions:

• Generated four proof-of-concept asthma clusters using 129Xe MRI, revealing distinct structure-function pathologies.
• Demonstrated the feasibility of cluster analysis using 129Xe MRI and CT biomarkers for asthma phenotyping.
• This imaging-based approach may challenge current asthma phenotyping and treatment paradigms.