Machine learning-powered, device-embedded heart sound measurement can optimize AV delay in patients with CRT

Philip Westphal; Hongxing Luo; Mehrdad Shahmohammadi; Frits W Prinzen; Tammo Delhaas; Richard N Cornelussen

doi:10.1016/j.hrthm.2023.05.025

Machine learning-powered, device-embedded heart sound measurement can optimize AV delay in patients with CRT

Heart Rhythm. 2023 Sep;20(9):1316-1324. doi: 10.1016/j.hrthm.2023.05.025. Epub 2023 May 27.

Authors

Philip Westphal¹, Hongxing Luo², Mehrdad Shahmohammadi³, Frits W Prinzen², Tammo Delhaas³, Richard N Cornelussen⁴

Affiliations

¹ Department of Physiology, Cardiovascular Research Institute Maastricht (CARIM), Maastricht, The Netherlands; Bakken Research Center, Medtronic, plc, Maastricht, The Netherlands.
² Department of Physiology, Cardiovascular Research Institute Maastricht (CARIM), Maastricht, The Netherlands.
³ Department of Biomedical Engineering, Cardiovascular Research Institute Maastricht (CARIM), Maastricht, The Netherlands.
⁴ Department of Physiology, Cardiovascular Research Institute Maastricht (CARIM), Maastricht, The Netherlands; Bakken Research Center, Medtronic, plc, Maastricht, The Netherlands. Electronic address: Richard.Cornelussen@medtronic.com.

PMID: 37247684
DOI: 10.1016/j.hrthm.2023.05.025

Abstract

Background: Continuous optimization of atrioventricular (AV) delay for cardiac resynchronization therapy (CRT) is mainly performed by electrical means.

Objective: The purpose of this study was to develop an estimation model of cardiac function that uses a piezoelectric microphone embedded in a pulse generator to guide CRT optimization.

Methods: Electrocardiogram, left ventricular pressure (LVP), and heart sounds were simultaneously collected during CRT device implantation procedures. A piezoelectric alarm transducer embedded in a modified CRT device facilitated recording of heart sounds in patients undergoing a pacing protocol with different AV delays. Machine learning (ML) was used to produce a decision-tree ensemble model capable of estimating absolute maximal LVP (LVP_max) and maximal rise of LVP (LVdP/dt_max) using 3 heart sound-based features. To gauge the applicability of ML in AV delay optimization, polynomial curves were fitted to measured and estimated values.

Results: In the data set of ∼30,000 heartbeats, ML indicated S1 amplitude, S2 amplitude, and S1 integral (S1 energy for LVdP/dt_max) as most prominent features for AV delay optimization. ML resulted in single-beat estimation precision for absolute values of LVP_max and LVdP/dt_max of 67% and 64%, respectively. For 20-30 beat averages, cross-correlation between measured and estimated LVP_max and LVdP/dt_max was 0.999 for both. The estimated optimal AV delays were not significantly different from those measured using invasive LVP (difference -5.6 ± 17.1 ms for LVP_max and +5.1 ± 6.7 ms for LVdP/dt_max). The difference in function at estimated and measured optimal AV delays was not statiscally significant (1 ± 3 mm Hg for LVP_max and 9 ± 57 mm Hg/s for LVdP/dt_max).

Conclusion: Heart sound sensors embedded in a CRT device, powered by a ML algorithm, provide a reliable assessment of optimal AV delays and absolute LVP_max and LVdP/dt_max.

Keywords: Artificial intelligence; Cardiac resynchronization therapy; Clinical study; Heart sounds; Hemodynamics; Machine learning; Remote monitoring.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Cardiac Resynchronization Therapy Devices
Cardiac Resynchronization Therapy* / methods
Electrocardiography / methods
Heart Failure* / diagnosis
Heart Failure* / therapy
Heart Sounds*
Humans
Ultrasonography