Non-invasive late activation mapping identifies atrial deceleration zones
S Ros, C Fambuena, M Correas, I Martin-Martinez, L Gonzalez, G Rios-Munoz, A Arenal, E Gonzalez-Torrecilla, P Avila, A Carta-Bergaz, P Martinez-Legazpi, J Bermejo, A M Climent, M S Guillem, F AtienzaAbstract
Background
Atrial deceleration zones—regions of slow conduction—are emerging as critical targets for atrial fibrillation (AF) ablation. However, identifying them with conventional sequential mapping is time-consuming and technically challenging. A non-invasive tool that provides a global map of these zones could guide and streamline the procedure.
Purpose
To determine if non-invasive late activation mapping with ECGI can accurately localize atrial deceleration zones, using high-density invasive electroanatomic mapping (EAM) as the gold-standard.
Methods
The relationship between slow conduction and ECGI-derived late activation was first established in a comprehensive in-silico study. We then prospectively validated the method in 19 persistent AF patients undergoing ablation. During coronary sinus pacing (600 ms), invasive high-density EAM was used to generate reference maps of deceleration zones. Simultaneously, non-invasive ECGI maps of late activation were generated and co-registered with EAM for direct spatial comparison.
Results
In patients, regions identified by ECGI as having late activation corresponded to areas with significantly greater invasively-measured deceleration on EAM . As shown in Figure 1 (middle panel), deceleration was nearly twice as high in ECGI-identified late activation areas compared to other regions (12.1 ± 2.9 m/s² vs 7.2 ± 2.8 m/s²; p<0.001). The figure also includes a representative patient example illustrating how areas of high deceleration on the EAM map spatially overlap with regions identified by ECGI as showing late activation (left column, upper panel) and absence of late activation (lower panel). These clinical findings were strongly supported by our in-silico validation (right panel), where ECGI correctly identified simulated slow conduction regions by mapping a significantly larger area of late activation compared to healthy tissue (19.0 ± 22.8% vs 0 ± 0%; p<0.001).
Conclusion
Non-invasive late activation mapping with ECGI reliably identifies and localizes atrial deceleration zones, validated against both high-density invasive EAM and in-silico models. This provides a powerful tool to create a global, pre-procedural map of the arrhythmogenic substrate, with the potential to guide clinicians to specific targets and streamline AF ablation procedures.ECGI late activation mapping validation