Evaluation of the ECGI Patchwork Method Using Experimental Data in Sinus Rhythm

Oumayma Bouhamama1, Lisl Weynans2, Laura Bear3
1Université de Bordeaux, IMB, UMR5251, 2Bordeaux University, 3IHU-LIRYC


Background: Noninvasive electrocardiographic imaging (ECGI) provides real-time panoramic images of epicardial electrical activity from potential measurements on the torso surface. We have recently developed a novel method that combines the solutions obtained with classic formulations in order to select the most accurate method based on their residuals. We hypothesize this Patchwork method (PM) improves the accuracy of reconstructions in sinus rhythm.

Method: Epicardial and body surface and potentials were recorded simultaneously in closed-chest pigs (n=5) during sinus rhythm. Torso and cardiac geometries were determined using MRI. To evaluate the PM, we compared the electrograms and activation times that we reconstructed against standard ECGI methods (FEM, MFS and BEM). Epicardial breakthrough sites and lines of conduction block were identified and compared, where lines of block were computed as a jump in local activation time of >20 ms between adjacent electrodes.

Results: Cardiac potentials and activation maps reconstructed with the PM were more correlated to those recorded than those obtained with the MFS, FEM and BEM (mean CC of 0.9 vs. 0.79-0.88 for ATs and 0.59 vs. 0.44-0.56 for potentials). The localization of breakthrough sites was more accurate using PM than standard ECGI methods (median LE of 17.16 vs. 26.84-30.52). No lines of block were identified in recorded maps. However, artificial lines of block were identified across all 5 derived activation maps using the FEM, BEM or MFS. These were present in only 2 activation maps obtained with the PM, and were substantially reduced in size.

Conclusion: The patchwork method demonstrates a higher level of accuracy in reconstructing activation maps and in locating breakthrough sites in sinus rhythm than standard ECGI methods. Importantly, these improvements include a reduction in the frequency of artificial block lines. This has important clinical implications as it may help reduce false diagnosis of conduction disorders.