High-density sequential mapping of repetitive atrial conduction patterns during atrial fibrillation

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – EU funding. Main funding source(s): This work was supported by PersonalizeAF project. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 860974. Background Localized AF drivers are considered candidate ablation targets for patients with persistent atrial fibrillation (AF). These drivers are expected to be associated with repetitive atrial conduction patterns during AF. Thus, tools that localize atrial sites with repetitive electrical activity might be instrumental in guiding ablation. Purpose High-density mapping catheters cover only a small portion of the atria. Combining sequential recordings from those catheters could provide a more complete picture of repetitive conduction patterns, and enable AF driver localization. We hypothesize that the repetitive activity generated by local AF drivers can be detected by means of high-coverage composite activation maps generated from spatially overlapping sequential recordings. Methods Repetitive conduction patterns were detected in a goat model of AF (249-electrode epicardial mapping array, 2.4mm inter-electrode distance, n=16) by exploiting recurrence plots (Fig 1A-C). Cross-recurrences of repetitive patterns in sequential recordings were detected in spatially overlapping recording locations. Using this information, local activation maps were aligned and combined into larger composite average activation maps (Fig. 1D-F). The proposed algorithm was tested on a dataset formed by segmenting the epicardial mapping area into four spatially overlapping regions. The proposed algorithm was then used to merge these segmented regions back together to reconstruct the original mapping area. Reconstruction accuracy was measured as the correlation between original and reconstructed average activation patterns (Fig. 2.). Statistical analyses were performed to investigate a possible relation between reconstruction accuracy and pattern properties such as duration, size, complexity, and cycle length. Patterns were classified as single peripheral, multiple waves, focal source, or re-entry based on the preferential conduction velocity directions. Results Among 1021 detected repetitive patterns, 328 spatiotemporally stable- patterns were present in all four artificially segmented recordings. In 32% of these, repetitiveness was associated with a local driver-either focal or re-entrant. Composite maps could be generated in 75% of the cases, and mainly in case of larger patterns (p<0.01). The average correlation between the actual activation maps and the composite maps was 0.86 ±0.16. Only pattern duration showed a statistically significant low correlation with reconstruction accuracy of composite maps (r=0.126, p<0.05). There was no significant difference in the reconstruction accuracy for multiple waves, focal sources and re-entries. Conclusion(s) The proposed framework could align sequentially recorded repetitive epicardial patterns over different atrial regions, to produce high-fidelity composite maps. The performance was minimally affected by pattern properties, thus suggesting potential use with a diverse range of AF patterns.