On the efficiency and accuracy of the single equivalent moving dipole method to identify sites of cardiac electrical activation

Abstract

We have proposed an algorithm to guide radiofrequency catheter ablation procedures. This algorithm employs the single equivalent moving dipole (SEMD) to model cardiac electrical activity. The aim of this study is to investigate the optimal time instant during the cardiac cycle as well as the number of beats needed to accurately estimate the location of a pacing site. We have evaluated this algorithm by pacing the ventricular epicardial surface and inversely estimating the locations of pacing electrodes from the recorded body surface potentials. Two pacing electrode arrays were sutured on the right and left ventricular epicardial surfaces in swine. The hearts were paced by the electrodes sequentially at multiple rates (120–220 bpm), and body surface ECG signals from 64 leads were recorded for the SEMD estimation. We evaluated the combined error of the estimated interelectrode distance and SEMD direction at each time instant during the cardiac cycle, and found the error was minimum when the normalized root mean square (RMS[subscript n]) value of body surface ECG signals reached 15 % of its maximum value. The beat-to-beat variation of the SEMD locations was significantly reduced (p < 0.001) when estimated at 15 % RMS[subscript n] compared to the earliest activation time (EAT). In addition, the 5–95 % interval of the estimated interelectrode distance error decreased exponentially as the number of beats used to estimate a median beat increased. When the number of beats was 4 or larger, the 5–95 % interval was smaller than 3.5 mm (the diameter of a commonly used catheter). In conclusion, the optimal time for the SEMD estimation is at 15 % of RMS[subscript n], and at that time instant a median beat estimated from 4 beats is associated with a beat-to-beat variability of the SEMD location that is appropriate for catheter ablation procedures.