Session S61.2

Analysis of Cardiac Micro-Acceleration Signals for the Estimation of Systolic and Diastolic Time Intervals in Cardiac Resynchronization Therapy

L Giorgis*, AI Hernández, A Amblard, L Senhadji,
S Cazeau, E Donal, G Jauvert

Sorin Group
Le Plessis Robinson, France

Before the arrival of Doppler echocardiography, great efforts were made towards the use cardiac acoustic signals for the evaluation of the mechanical function of the heart. While a recent controversy questions the reproducibility and precision of echo-Doppler measurements, we notice a regain of interest in quantitative analysis of these acoustic signals, especially for the optimal delivery of Cardiac Resynchronization Therapy (CRT). In this context, previous studies have shown that endocardial acceleration (EA) may be useful for an online follow-up of the heart mechanical function. In particular, it has been demonstrated that changes in the peak-to-peak amplitude of the first component of the EA signal (PEA) are closely related to changes in the maximum rate of rise of left ventricular pressure (max LV dP/dt). The EA signal is already available in implantable device via a specific endocardial sensor embedded in a pacing lead (Sorin Group CRM).
The present study focuses on a non-invasive cardiac micro-acceleration (CMA) device, by cutaneous precordial application of an external version of the Sorin Group EA sensor, placed on the chest of the patient. The aim is to estimate the following systolic and diastolic time intervals from CMA signal features: i) QRS to mitral valve closure (tMC), ii) QRS to aortic valve opening (tAO), iii) QRS to aortic valve closure (tAC) and iv) QRS to mitral valve opening (tMO). For this matter, CMA, ECG and Doppler audio signals were acquired simultaneously on 50 CRT patients, paced in different ventricular configurations (right, left, and bi-ventricular pacing).
After appropriate pre-processing of the CMA signal, features extraction techniques, based mainly on classical envelograms (Shannon, Homomorphic envelogram, etc.), were used to segment its main components. An evaluation methodology is suggested to compare the different methods applied with different sets of parameters, considering the echo timings as a gold standard. In order to get reliable reference timings, we developed a Doppler signal processing platform to assist an experienced echocardiographist.
Features extracted from the CMA signal showed significant linear correlation with the markers under study. In particular, Shannon-based transforms provided the best results for tMC (r=0.74, p<0.001), tAO (r=0.73, p<0.001) and tMO (r=0.65, p<0.001), while a homomorphic envelogram showed the best performance for tAC (r=0.9, p<0.001). Results show the potential utility of CMA signals for monitoring systolic and diastolic time intervals, assessing cardiac performance and defining optimal, adaptive pacing configuration. The application of these algorithms in an implantable device, via the EA signal, could be a significant advance in CRT.

(Abstract Control Number: 227)