Assessing PEP and LVET from heart sounds: algorithms and evaluation

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Signal Processing

Background:

• Accurate estimation of systolic time intervals is crucial for assessing cardiac function.
• Traditional methods for measuring PEP and LVET can be invasive or complex.
• Heart sounds offer a non-invasive acoustic window into cardiac mechanics.

Purpose of the Study:

• To develop and validate algorithms for estimating pre-ejection period (PEP) and left ventricular ejection time (LVET) using phonocardiograms (heart sounds).
• To compare heart sound-based LVET estimation with photoplethysmography (PPG) measurements.
• To assess the feasibility of using heart sounds to detect aortic valve movement onset.

Main Methods:

• A Bayesian approach was employed to estimate PEP, utilizing instantaneous amplitude and valve timing.
• LVET estimation involved analyzing the high-frequency signature of the second heart sound (S2) to determine aortic valve closure.
• LVET was also calculated from peripheral PPG signals for comparative analysis.

Main Results:

• The algorithms were evaluated on 658 heartbeats.
• Average absolute error for PEP estimation was 10.32 msec (SD 7.3 msec).
• Average absolute error for LVET estimation was 15.8 msec (SD 13.6 msec).

Conclusions:

• Heart sounds can be reliably used to estimate key systolic time intervals (PEP and LVET).
• The developed methods demonstrate good accuracy, supporting their clinical potential.
• Acoustic analysis of heart sounds shows promise for detecting the initiation of aortic valve dynamics.