Abstract 297 Local conduction velocities determined by non-invasive electrocardiographic imaging predict response to atrial fibrillation ablation

Background
Long-term recurrence rates after atrial fibrillation (AF) ablation are still not satisfactory, particularly in patients with persistent AF. Therefore, predictive tools that allow for an a priori discrimination of therapy responders and non-responders are needed. Several predictors and predictive models have been proposed in the past, but data on their predictive value is conflictive and their capability to identify therapy responders limited. Here we validate a novel non-invasive electrocardiographic imaging (ECGi) technology and test its predictive value regarding AF recurrences.

21 healthy controls and 38 consecutive patients scheduled for AF ablation received a preprocedural ECGi with 128 electrodes placed on the torso (Corify Care, Inc.). Subsequently, a 3D model of the torso was acquired as an anatomical reference using a 3D reconstruction camera. A personalised 3D atrial geometry derived from a data base of human atria was computed by an AI-based algorithm. Local conduction velocities were then derived from reconstructed local unipolar epicardial electrograms and validated in patients undergoing AF ablation using invasive high-densitiy mapping, which showed a moderate correlation. ECGi-determined conduction velocities were significantly lower in AF patients than in healthy controls (1,43 ± 0.16 m/s vs. 1.64 ± 0.15 m/s; p < 0.001). Focusing only on the „slowest“, potentially most arrhythmogenic, of 19 predefined left and right atrial regions of each individual, the difference between the AF and the healthy group was even more pronounced (AF 0.79 ± 0.22 m/s vs. 1.09 ± 0.24 m/s; p<0.001). While the global mean conduction velocity was predictive of arrhythmia-free survival at 6 months, the mean conduction velocity of the „slowest“ region showed the highest predictive value of all parameters. Of note, the predictive value was higher than that of previously proposed predictors like LA diameter, AF type or total LA activation time, which did not improve the predictive model.
Applying a cutoff 1.5 standard deviations below the mean conduction velocity of the „slowest“ region in healthy controls (0.72 m/s), this parameter was not only predictive of arrhythmia-free survival (odds ratio 26.8, CI 2.7 – 262.2, p=0.005, specificity 79%, sensitivity 88%), but also capable of discriminating responders from non-responders: Patients with a mean conduction velocity of >0.72 m/s in all atrial regions showed a very high arrhythmia-free survival of 96%, whereas patients with one or more atrial regions with a mean conduction velocity <0.72 m/s had a poor outcome with an arrhythmia-free survival rate of only 46%. While a regional conduction velocity <0.72 m/s does certainly not exclude ablation therapy, the indication, particularly in case of repeat procedures, should be reconsidered more carefully and maybe more restrictively. It is tempting to speculate that those patients may benefit from more extensive ablation beyond PVI-only.

In this study we validated a novel ECGi technology and its ability to determine local conduction velocities. We found local atrial conduction velocities determined by ECGi to predict outcome after AF ablation. The regional conduction velocitiy threshold we defined based on a healthy cohort discriminated AF ablation responders from non-responders and may therefore guide selection of suitable candidates and those that are unlikely to benefit from ablation or re-ablation, respectively.

Clin Res Cardiol (2022). https://doi.org/10.1007/s00392-022-02087-y
Local conduction velocities determined by non-invasive electrocardiographic imaging predict response to atrial fibrillation ablation
E. Invers¹, J. Reventos¹, J. M. Tolosana¹, E. Guasch¹, A. Porta-Sanchez¹, I. Roca¹, E. Arbelo¹, J. Brugada¹, L. Mont¹, T. Althoff¹
¹Arrhythmia Section, Cardiovascular Institute (ICCV), Hospital Clínic, University of Barcelona, Barcelona, ES;
Background Long-term recurrence rates after atrial fibrillation (AF) ablation are still not satisfactory, particularly in patients with persistent AF. Therefore, predictive tools that allow for an a priori discrimination of therapy responders and non-responders are needed. Several predictors and predictive models have been proposed in the past, but data on their predictive value is conflictive and their capability to identify therapy responders limited. Here we validate a novel non-invasive electrocardiographic imaging (ECGi) technology and test its predictive value regarding AF recurrences. Methods and results 21 healthy controls and 38 consecutive patients scheduled for AF ablation received a preprocedural ECGi with 128 electrodes placed on the torso (Corify Care, Inc.). Subsequently, a 3D model of the torso was acquired as an anatomical reference using a 3D reconstruction camera. A personalised 3D atrial geometry derived from a data base of human atria was computed by an AI-based algorithm. Local conduction velocities were then derived from reconstructed local unipolar epicardial electrograms and validated in patients undergoing AF ablation using invasive high-densitiy mapping, which showed a moderate correlation. ECGi-determined conduction velocities were significantly lower in AF patients than in healthy controls (1,43 ± 0.16 m/s vs. 1.64 ± 0.15 m/s; p < 0.001). Focusing only on the „slowest“, potentially most arrhythmogenic, of 19 predefined left and right atrial regions of each individual, the difference between the AF and the healthy group was even more pronounced (AF 0.79 ± 0.22 m/s vs. 1.09 ± 0.24 m/s; p<0.001). While the global mean conduction velocity was predictive of arrhythmia-free survival at 6 months, the mean conduction velocity of the „slowest“ region showed the highest predictive value of all parameters. Of note, the predictive value was higher than that of previously proposed predictors like LA diameter, AF type or total LA activation time, which did not improve the predictive model. Applying a cutoff 1.5 standard deviations below the mean conduction velocity of the „slowest“ region in healthy controls (0.72 m/s), this parameter was not only predictive of arrhythmia-free survival (odds ratio 26.8, CI 2.7 – 262.2, p=0.005, specificity 79%, sensitivity 88%), but also capable of discriminating responders from non-responders: Patients with a mean conduction velocity of >0.72 m/s in all atrial regions showed a very high arrhythmia-free survival of 96%, whereas patients with one or more atrial regions with a mean conduction velocity <0.72 m/s had a poor outcome with an arrhythmia-free survival rate of only 46%. While a regional conduction velocity <0.72 m/s does certainly not exclude ablation therapy, the indication, particularly in case of repeat procedures, should be reconsidered more carefully and maybe more restrictively. It is tempting to speculate that those patients may benefit from more extensive ablation beyond PVI-only. Conclusion In this study we validated a novel ECGi technology and its ability to determine local conduction velocities. We found local atrial conduction velocities determined by ECGi to predict outcome after AF ablation. The regional conduction velocitiy threshold we defined based on a healthy cohort discriminated AF ablation responders from non-responders and may therefore guide selection of suitable candidates and those that are unlikely to benefit from ablation or re-ablation, respectively.
https://dgk.org/kongress_programme/ht2022/aP747.html