Mean Pressure Gradient Prediction Based on Chest Angular Movements and Heart Rate Variability Parameters.

Area of Science:

• Biomedical Engineering
• Cardiovascular Research
• Machine Learning Applications

Background:

• Aortic stenosis (AS) diagnosis relies on expensive ultrasound echocardiography.
• There is a need for cost-effective and accessible methods for AS severity classification.

Purpose of the Study:

• To develop a low-cost framework for classifying aortic stenosis severity using wearable sensors.
• To evaluate the efficacy of machine learning models in analyzing gyroscopic data for AS classification.

Main Methods:

• Feature extraction from gyroscopic readings, including cardiac timing intervals and heart rate variability (HRV) parameters.
• Classification of AS severity (mild, moderate, severe) using state-of-the-art machine learning (ML) methods.
• Analysis of feature importance using game theory, identifying isovolumetric contraction time (IVCT) and isovolumetric relaxation time (IVRT) as key indicators.

Main Results:

• The Light Gradient-Boosted Machine (Light GBM) model achieved the highest performance with 94.44% accuracy and 94.29% F1-score.
• Isovolumetric contraction time (IVCT) and isovolumetric relaxation time (IVRT) were identified as the most representative features for AS severity.
• The framework demonstrated the potential for accurate classification of AS severity using readily available sensor data.

Conclusions:

• The proposed framework offers a promising low-cost alternative to traditional echocardiography for aortic stenosis severity assessment.
• Wearable sensor technology combined with machine learning can effectively classify AS severity.
• This approach has the potential to improve accessibility and reduce healthcare costs associated with AS management.