Session S71.4

Cardiac Arrhythmia Spectral Analysis of Electrogram Signals Using Fourier Organization Analysis

O Barquero-Pérez*, JL Rojo-Álvarez, J Requena-Carrión,
F Alonso-Atienza, E Everss, R Goya-Esteban, JJ Sánchez-Muñoz, A García-Alberola

University Rey Juan Carlos
Fuenlabrada, Spain

Background: Dominant Frequency Analysis (DFA) and Organization Analysis (OA) of cardiac electrograms (EGM) aim to establish clinical targets for cardiac arrhythmia ablation. Previous spectral descriptions of the EGM have often discarded relevant information in the spectrum, such as the harmonic structure or the spectral envelope.
Aim: We propose a generalized and simple description of the spectral features in EGM recordings, called Fourier Organization Analysis (FOA), which accounts for the organization and periodicity in the EGM in terms of the fundamental frequency (f0). The method consists of a Least Squares projection of the EGM into a Hilbert signal space of harmonically related sinusoids accounting for near-periodic fluctuations. Parameter f0 can be readily estimated according to the minimum approximation error of the EGM by the model as a function of the f0 in the range 1-15 Hz.
Methods: EGM were first analyzed in a simulation computer model for a plane wavefront stimulated tissue and a reentry anchored to an infarcted tissue region. Also, can-coil and tip-ring EGM stored in Implantable Cardioverter Defibrillator (ICD) were analyzed (5 sinus rhythm –SR-, 8 supraventricular tachycardia –SVT-, 8 ventricular –VT-, and 7 ventricular fibrillation –VF-).
Results: Similar values of f0 were obtained with both methods for simulated plane wavefront (DFA 2.54 Hz, FOA 2.46 Hz), whereas it was different in anchored rotor (DFA 3.91 Hz, FOA 2.20 Hz). The EGM reconstructed with f0 from FOA was in close agreement with the original, whereas the EGM reconstructed with Fourier Series and f0 from DFA had a significantly different morphology, and apparently reduced organization parameters were obtained due to the error in the estimation of f0. In ICD recordings, there were differences in the estimation of f0 with DFA (can-coil SR 2.94+-0.76, SVT 2.58+-0.16, VT 3.25+-0.67, VF 5.10+-0.87 Hz; tip-ring SR 3.66+-0.33, SVT 3.51+-1.2, VT 3.25+-0.66, VF 5.25+-0.82) or with FOA (can-coil SR 1.15+-0.17, SVT 2.58+-0.16, VT 2.74+-0.46, VF 5.26+-0.89 Hz; tip-ring SR 1.22+-0.11, SVT 2.58+-0.16, VT 2.70+-0.52, VF 5.27+-0.87 Hz). In all cases, only EGM reconstructed with f0 from FOA were always in close agreement with real EGM.
Conclusion: We conclude that FOA yields an adequate framework for analyzing ICD stored EGMs.

(Abstract Control Number: 42)