Explainable artificial intelligence-based classification of asthma and COPD from xenon-129 ventilation magnetic resonance imaging
4 affiliations
- 1POLARIS, Section of Medical Imaging and Technologies, Division of Clinical Medicine, School of Medicine and Population Health, The University of Sheffield, Sheffield, UK.
- 2Insigneo Institute, The University of Sheffield, Sheffield, UK.
- 3Early Development Respiratory Medicine, AstraZeneca, Cambridge, UK.
- 4Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK.
Abstract
Background
Diagnosis of asthma and/or COPD is challenging due to overlapping clinical and physiological characteristics; however, due to varying treatment recommendations, determining the correct diagnosis is vital to clinical care. Hyperpolarised xenon-129 ventilation (¹²⁹Xe)-magnetic resonance imaging (MRI) provides high spatial resolution three-dimensional images of gas distribution within the lungs. Recently, artificial intelligence (AI)-based models have successfully classified pulmonary pathologies via medical images, but due to their black-box nature, clinical application is challenging. Here, we aim to develop an explainable AI framework to classify patients with asthma and/or COPD using ¹²⁹Xe-MRI alone.Methods
¹²⁹Xe-MRI from 207 patients with asthma and/or COPD were classified using a vision transformer (ViT) and a convolutional neural network (CNN). Regional visualisations of areas of importance (AOIs) associated with each diagnosis were generated. Regional patterns were investigated to develop a shape-based metric which quantified asthma-like and COPD-like features.Results
The ViT accurately classified asthma and/or COPD patients in 83% (n=25/30) of cases (area under the curve (AUC) 0.86, F₁ score 0.82), outperforming the CNN (AUC 0.71, F₁ score 0.62). Patients with asthma had more diffuse, elongated AOIs associated with disease grouping while patients with COPD had more focal, spherical AOIs. The proposed sphericity score weakly correlated with conventional lung function metrics in clinically understandable ways. When applied to patients with overlapping features of asthma and COPD (asthma+COPD), the classification framework informed the prevalence and regional distribution of asthma-like and COPD-like features.Conclusion
Explainable AI classification of ¹²⁹Xe-MRI could facilitate diagnosis of patients with asthma and/or COPD, including determining the prevalence of asthma-like or COPD-like features in patients with asthma+COPD, potentially informing management and treatment.
Related
- Surrogate Modeling and Explainable Artificial Intelligence for Complex Systems: A Workflow for Automated Simulation Exploration
- Deep learning-based self-supervised detection and localization of magnetic resonance imaging anomalies in artificial intelligence-supported radiotherapy workflows
- Using artificial intelligence for clinical trial quality assurance in a multi-institutional neurovascular-sparing prostate radiotherapy trial
- Future of Artificial Intelligence for Science in Japan 2024 Community Report
- Advancing Nuclear Physics with Machine Learning and Artificial Intelligence
BibTeX
@article{astley2026explainable,
title = {{Explainable artificial intelligence-based classification of asthma and COPD from xenon-129 ventilation magnetic resonance imaging}},
author = {Joshua R Astley and Laurie J Smith and Alberto M Biancardi and Rod Hughes and Rod Lawson and Bilal A Tahir and Helen Marshall and Jim M Wild},
journal = {ERJ open research},
year = {2026},
month = sep,
doi = {10.1183/23120541.00234-2026},
url = {https://europepmc.org/article/PMC/PMC13718910},
note = {CC BY},
}
RIS
TY - JOUR
TI - Explainable artificial intelligence-based classification of asthma and COPD from xenon-129 ventilation magnetic resonance imaging
AU - Joshua R Astley
AU - Laurie J Smith
AU - Alberto M Biancardi
AU - Rod Hughes
AU - Rod Lawson
AU - Bilal A Tahir
AU - Helen Marshall
AU - Jim M Wild
PY - 2026
DA - 2026/09/01
JO - ERJ open research
DO - 10.1183/23120541.00234-2026
UR - https://europepmc.org/article/PMC/PMC13718910
AB - Background Diagnosis of asthma and/or COPD is challenging due to overlapping clinical and physiological characteristics; however, due to varying treatment recommendations, determining the correct diagnosis is vital to clinical care. Hyperpolarised xenon-129 ventilation (¹²⁹Xe)-magnetic resonance imaging (MRI) provides high spatial resolution three-dimensional images of gas distribution within the lungs. Recently, artificial intelligence (AI)-based models have successfully classified pulmonary pathologies via medical images, but due to their black-box nature, clinical application is challenging. Here, we aim to develop an explainable AI framework to classify patients with asthma and/or COPD using ¹²⁹Xe-MRI alone.Methods ¹²⁹Xe-MRI from 207 patients with asthma and/or COPD were classified using a vision transformer (ViT) and a convolutional neural network (CNN). Regional visualisations of areas of importance (AOIs) associated with each diagnosis were generated. Regional patterns were investigated to develop a shape-based metric which quantified asthma-like and COPD-like features.Results The ViT accurately classified asthma and/or COPD patients in 83% (n=25/30) of cases (area under the curve (AUC) 0.86, F₁ score 0.82), outperforming the CNN (AUC 0.71, F₁ score 0.62). Patients with asthma had more diffuse, elongated AOIs associated with disease grouping while patients with COPD had more focal, spherical AOIs. The proposed sphericity score weakly correlated with conventional lung function metrics in clinically understandable ways. When applied to patients with overlapping features of asthma and COPD (asthma+COPD), the classification framework informed the prevalence and regional distribution of asthma-like and COPD-like features.Conclusion Explainable AI classification of ¹²⁹Xe-MRI could facilitate diagnosis of patients with asthma and/or COPD, including determining the prevalence of asthma-like or COPD-like features in patients with asthma+COPD, potentially informing management and treatment.
SN - 2312-0541
ER -