Machine Learning Prediction of Progression in Forced Expiratory Volume in 1 Second in the COPDGene® Study

Area of Science:

• Pulmonary Medicine
• Biostatistics
• Machine Learning in Healthcare

Background:

• Chronic obstructive pulmonary disease (COPD) is characterized by heterogeneity, making disease progression prediction challenging.
• Identifying predictors of COPD progression is crucial for effective disease management and intervention.

Purpose of the Study:

• To enhance the prediction of COPD progression using machine learning techniques.
• To incorporate a comprehensive set of phenotypic features for improved predictive accuracy.

Main Methods:

• Utilized data from 4496 smokers in the COPDGene study with 5-year follow-up.
• Developed and compared linear regression and random forest models to predict 5-year Forced Expiratory Volume in 1 second (FEV1) change.
• Employed cross-validation for training and testing, with validation on a 10-year follow-up dataset.

Main Results:

• Random forest models achieved an R-squared of 0.15 and an Area Under the ROC Curve of 0.71 for predicting top quartile progression.
• Random forest slightly outperformed linear regression.
• Prediction accuracy was highest for Global Initiative for Chronic Obstructive Lung Disease (GOLD) grades 1-2, with reduced accuracy in advanced stages.

Conclusions:

• Random forest models demonstrate reasonable accuracy in predicting FEV1 progression when combined with deep phenotyping.
• The developed model can identify smokers at higher risk of rapid COPD progression.
• Findings support patient stratification for clinical trials and targeted interventions.