Objectively Measured Chronic Lung Injury on Chest CT

Area of Science:

• Radiology and Pulmonary Medicine
• Artificial Intelligence in Medical Imaging

Background:

• Tobacco smoke exposure is a known cause of irreversible lung diseases like emphysema and pulmonary fibrosis.
• Early detection of lung tissue changes preceding these diseases is crucial for intervention.

Purpose of the Study:

• To develop and validate an objective CT analysis tool using machine learning to identify early lung changes (NormHA) in visually normal lung tissue.
• To assess the association of NormHA with clinical severity and mortality in individuals exposed to tobacco smoke.

Main Methods:

• A machine learning classification tool was trained on 34,528 training points from 297 COPDGene participants' chest CT scans.
• The tool classified lung tissue in 9,038 participants into normal, emphysema, fibrotic/interstitial, or NormHA categories.
• Associations between NormHA quartiles and biomarkers, clinical severity (FEV1, 6-min walk distance), and mortality were analyzed.

Main Results:

• Increased NormHA percentage in the lung was significantly associated with higher C-reactive protein and intercellular adhesion molecule 1.
• Higher NormHA quartiles correlated with significantly lower percent predicted FEV1, reduced 6-min walk distance, and increased mortality risk.
• These associations persisted even in individuals without visually apparent interstitial lung abnormalities.

Conclusions:

• A novel CT-based metric, NormHA, identifies a distinct tissue class associated with adverse pulmonary outcomes.
• NormHA may represent an early indicator of lung disease progression, independent of established emphysema and fibrosis.
• This tool offers potential for earlier detection and risk stratification in individuals exposed to tobacco smoke.