Using Machine Learning to Identify Organ System Specific Limitations to Exercise via Cardiopulmonary Exercise Testing

Area of Science:

• Physiology
• Medical Informatics
• Machine Learning

Background:

• Cardiopulmonary Exercise Testing (CPET) is crucial for assessing exercise response and identifying limitations.
• Interpreting complex CPET data requires specialized training and significant clinical experience.
• Current data visualization methods for CPET are challenging for widespread clinical adoption.

Purpose of the Study:

• To develop a simplified data interpretation and visualization tool for CPET using machine learning.
• To enhance the accessibility of CPET results for clinicians.
• To augment existing diagnostic procedures with an AI-driven platform.

Main Methods:

• Investigated machine learning algorithms for CPET data analysis.
• Developed a visualization tool displaying cardiac, pulmonary, and other limitations.
• Utilized three independent random forest classifiers to define limitation values.
• Created an interactive dashboard with scores and interpretability plots for clinician understanding.

Main Results:

• Successfully developed a machine learning platform for CPET data interpretation.
• The visualization tool categorizes limitations into cardiac, pulmonary, and other.
• An interactive dashboard provides interpretable scores and plots for clinicians.
• The platform demonstrates potential for simplifying complex CPET data.

Conclusions:

• The developed machine learning tool can make CPET more accessible to clinicians.
• This platform has the potential to augment current diagnostic procedures.
• Simplified interpretation and visualization can improve the clinical utility of CPET.