Hyperpolarized 129Xe MRI Features Associated with Interstitial Lung Disease Identified Using an Interpretable Diagnostic Algorithm
A new MRI technique using xenon gas distinguished lung disease types with 93.5% accuracy, offering promise for earlier detection.
Duke University researchers developed a 4-metric interpretable decision tree using hyperpolarized 129Xe MRI to distinguish ILD, COPD, and healthy controls with 93.5% accuracy in a 155-patient cohort. The fully interpretable, physiologically grounded algorithm provides a clinically translatable framework for early ILD detection and individualized pulmonary assessment.
What the study was
- Study design
- Diagnostic accuracy study with interpretable ML (L1-regularized logistic regression + SHAP + decision tree)
- Population
- ILD (n=84), COPD (n=21), and healthy controls (n=50); age 56.6±17.8 years, 80 females
- Sample size
- 155
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- Academic Radiology
Why it surfaced
Strong diagnostic AI study with interpretable algorithm (SHAP + decision tree), validated with bootstrapping, 93.5% accuracy with very high COPD specificity (100%). Interpretability is a key translational advantage. Sample size is moderate (n=155); bootstrap stability confirmed.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.