AF driver detection in pulmonary vein area by electropcardiographic imaging: Relation with a favorable outcome of pulmonary vein isolation

Figures

• 

  FIGURE 1. Information about simulations 1 and 2 is presented in the first and second panel rows. Panels (A, D) show the simulated action potentials in an instant previous to the ablation, while panels (B, E) show the action potentials after ablation. In (C, F), ECGI-derived rotor histograms of the pre-ablation periods are shown.

• 

  FIGURE 2. (A) Recall, precision and F-score values obtained in the validation dataset. (B,C), Manually labelled and automatically detected rotor histograms obtained from recordings 9 (top) and 3 (bottom) respectively. The recall, precision and F-score values calculated in recording 9 were 0.87, 0.92, and 0.87 while the recall, precision and F-score values of recording 3 are 0.66, 1, and 0.7 respectively.

• 

  FIGURE 3. (A) and (B) Illustration of the regional anatomical division of the atria; (C) and (D) relative errors obtained in each of the atrial regions; (E) Mean number of annotated and detected PSs in different atrial regions across the validation dataset. The mean absolute error between these two magnitudes is also presented in black.

• 

  FIGURE 4. Number of rotors obtained in all the patients: In blue are represented patients staying in sinus rhythm (No recurrence) and in orange patients returning to arrhythmia 6 months after PVI (Recurrence). The number of rotors was calculated in the whole atria in two different scenarios regarding the number of turns used to filter the PS: (A) Number of rotors obtained without any turn filtering; (B) Number of rotors spinning for at least one turn.

• 

  FIGURE 5. (A,B) Proportion of PSs per second detected in different regions in each studied group: in blue is presented the data from patients that remained in sinus rhythm 6 months after ablation (Non-recurrence), and in orange patients that returned to arrhythmia (Recurrence). In panel (A), are presented the results obtained when all the reentries are considered in the analysis. In panel (B), the same results are shown, when only the reentries lasting for more than 1 turn. The distribution of the statistical p/values obtained for each regions are presented din panel (C).

• 

  FIGURE 6. (A,B) PS histograms with and without filtering. (C) Percentage of PSs found in all the different atrial regions when no filtering strategies are applied. No statistically significant differences are observed.