Assessment of Left Ventricular Diastolic Function using Phonocardiogram Signals: A Comparison with Echocardiography

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Medical Device Technology

Background:

• Left ventricular (LV) diastolic dysfunction is a critical indicator of heart health.
• Echocardiography is the standard for assessing LV diastolic function but requires specialized equipment and expertise.
• There is a need for accessible, non-invasive methods for continuous monitoring of diastolic function.

Purpose of the Study:

• To develop and validate a fully-automated, end-to-end wearable phonocardiogram (PCG)-based system.
• To derive echocardiography-like proxy metrics for LV diastolic function assessment from PCG signals.
• To evaluate the clinical utility of these proxy metrics against established guidelines.

Main Methods:

• Physiologically-motivated features were extracted from PCG signals.
• Noise-subtraction, heartbeat-segmentation, and quality-assurance algorithms were employed.
• Proxy metrics for five echocardiographic parameters were calculated and validated using ASE/EACVI guidelines.

Main Results:

• The system successfully identified LV diastolic dysfunction in 29 out of 34 patients with 87.5% accuracy.
• Elevated LV filling pressures were identified in 17 out of 34 patients with 75% accuracy.
• The developed proxy metrics demonstrated significant clinical value in assessing diastolic function.

Conclusions:

• A novel wearable PCG-based system offers a promising, automated approach for LV diastolic function assessment.
• The derived proxy metrics show high accuracy in identifying diastolic dysfunction and elevated filling pressures.
• This technology has the potential to enhance non-invasive cardiac monitoring and early diagnosis.