Non-invasive assessment of tetralogy of fallot isthmuses
E Guillen Buisan, J Reventos-Presmanes, M Correas, J Vicente, I Hernandez, S Monteiro, S Laranjo, M Oliveira, S Ros, F Atienza, M S Guillem, A M Climent, A Porta-Sanchez, I Roca-Luque, T RosseelAbstract
Background
Patients with repaired Tetralogy of Fallot (rToF) can present four anatomical isthmuses (AI) resulting from surgical correction. Earlier studies have demonstrated that these isthmuses are critical in sustaining ventricular arrhythmias (VA), being the leading cause of mortality in rToF. Risk stratification remains challenging, as invasively measured conduction velocity is the only strong predictor of VA. Non-invasive volumetric source imaging (VSI) reconstructs electrical activation throughout the cardiac volume, allowing visualization of conduction isthmuses and quantification of activation dynamics.
Purpose
To evaluate whether VSI can qualitatively identify the characteristic isthmuses of Tetralogy of Fallot and to assess quantitative VSI-derived metrics that differentiate ToF patients from control subjects.
Methods
Eight patients with rToF (75% male, 39.9 ± 15.4 years) and five controls with right bundle branch block (RBBB; 80% male, 71.8 ± 25.9 years) were included in the study. Due to spatial resolution limits in VSI, three isthmuses were considered: AI1 (lateral), AI2 (anterior) and AI3+4 (septal, combining two patch-related isthmuses due to their anatomical proximity). VSI reconstructions from 128-lead body-surface signals during sinus rhythm were compared with 3D electroanatomic maps (EAM) to identify slow-conducting isthmuses. Regional activation dispersion (rAD) was also computed in the identified isthmus regions for rToF and control RBBB subjects.
Results
VSI accurately identified the anatomical isthmuses observed with EAM, achieving 94% sensitivity, 71% specificity, 89% positive predictive value, and 83% negative predictive value (Figure 1). The lower specificity was due to a few false-positive detections in non-critical regions, supporting VSI as a potential screening tool for non-invasive identification of conduction isthmuses. Ventricular tachycardia (VT) was induced in five patients; in three with complete EAM, VSI clearly identified the critical circuit isthmus. In the remaining two, single-beat VSI activation maps were obtained, but EAM mapping was incomplete due to hemodynamic instability and non-sustained VT. Quantitative rAD analysis revealed significantly greater activation dispersion in rToF patients compared to controls (110 ± 18.6 ms vs. 73.1 ± 26.8 ms, p < 0.001; Figure 2), indicating increased conduction heterogeneity in the reconstructed myocardium, although rAD did not differentiate rToF patients with and without inducible VT.
Conclusions
This non-invasive VSI approach proved capable of accurately delineating rToF conduction isthmuses, achieving high sensitivity and good specificity. These findings highlight the potential of VSI as a non-invasive tool for substrate screening and pre-procedural assessment in rToF. Future studies including a larger cohort will determine whether quantitative VSI metrics, such as rAD, can further stratify rToF patients according to their risk of VA.