Computed tomography-based identification of ganglionated plexi to guide cardioneuroablation for vasovagal syncope

Cardioneuroablation (CNA) is an effective treatment for cardio-inhibitory vaso-vagal syncope (CI-VVS) that is under evaluation as a potential alternative to cardiac pacing in selected patients.1 The cornerstone of CNA is targeting groups of autonomic ganglia known as ganglionated plexi (GPs) embedded in epicardial fat pads (EFP) that are interconnected via intrinsic nerves with the sino-atrial and atrio-ventricular nodes. Ganglionated plexi are commonly identified by high frequency stimulation, fractionated electrogram (fEGM), or based on anatomical landmarks.1–5 Although varying in efficacy, these approaches may be hampered by limited sensitivity and specificity, unintended AF induction and interpatient variability. The CT-assisted CNA6 and CT-based identification of EFP to target GPs during AF ablation7 have been reported. We sought to assess feasibility of CT-based EFP-guided CNA in patients with CI-VVS. We enrolled 12 patients (six males, 53 ± 13 years) with multiple episodes of CI-VVS. Ten had head-up tilt-test (HUT) with cardio-inhibitory response (type 2A in three patients, type 2B in five and mixed response in two). One patient had a negative HUT, and one did not undergo the test. These two patients experienced spontaneous transient symptomatic high-degree AV block. Patients underwent electrophysiology study, including atropine test that ruled out organic sinus and AV node dysfunction. After discussion with patients on management options, CNA was proposed prior to committing to pacing. The study was approved from the Ethic Committee, and informed consent was obtained. Contrast-enhanced CTs displaying right, left atrial (LA) anatomy, aorta and EFP near the area of anticipated GPs8 with attenuation −190 to −30 HU were segmented and exported using ADAS3D software (Galgo Inc). The following GPs were identified as EFP: left superior GP between left superior pulmonary vein (PV) and LA appendage, Marshall tract GP in the carina between left PVs, left inferior GP posteriorly to the left inferior PV, inferior paraseptal ganglionated plexus (IPSGP) between the posterior wall of LA and coronary sinus, superior paraseptal ganglionated plexus (SPSGP) between the right superior PV and superior vena cava, right inferior GP (RIGP) between the two right PVs, and aorta-superior vena cava GP (Ao-SVC GP). On the day of CNA, anatomical map of the LA was acquired using CARTO 3 (Biosense Webster, Diamond Bar, CA) and fused with CT. The EFP areas were searched for fEGMs (≥4 deflections) but ablated independently from fragmentation. The ablation end-point for each GP was defined as abolition of radiofrequency (RF)-induced vagal response for left PV GPs, an increase in basal HR ≥25% for SPSGP, RIGP and Ao-SVC GPs, and shortening of AH interval for IPSGP. In case RF over EFP did not elicit autonomic response, we targeted the closest GP area based on anatomical landmarks. Left PV GPs were ablated first, followed by right PV GPs, IPSGP and Ao-SVC GP. A total of 84 EFP were ablated in 12 patients (median RF applications: 2, IQR: 1–3; time: 57 s, IQR: 33–77). Response to RF for each EFP is shown in Figure 1. Overall, 25 (30%) EFP responded to RF. Superior paraseptal ganglionated plexus displayed the highest rate of RF response (100%) and the most consistent correlation between EFP position and post-RF HR increase (Figure 1). Five additional GPs (6%) responded close but outside the EFP, at the level of the predicted area (EFP-RF distance: 5.5 ± 1 mm). Therefore, out of 30 GPs that responded to RF, 25 (83%) were effectively ablated over the EFP. The RF delivery did not induce any response in 54 (64%) GPs either at the level of the EFP or at the predicted anatomical location. After CNA, atropine was administered showing flattened HR response (from 75 ± 24 to 78 ± 24 bpm) without substantial changes of sinus or AV node function in all but one patient, who still displayed significant HR increase (from 72 to 126 bpm). There were no procedure-related complications. Nine patients had follow-up >4 months (median: 5, IQR: 4–9), without syncope recurrences. Anatomic localization of epicardial fat pads (EFP) and corresponding ganglionated plexi (GPs). For each GP, radiofrequency (RF) response is reported according to the application site (over EFP, outside EFP, no response). (A) Disagreement between EFP and effective RF site at the level of the Ao-SVC GP. (B) In the same patient, effective RF over the EFP at the level of SPSGP. Ao-SVC GP, aorta-superior vena cava GP; IPSGP, inferior paraseptal GP; LIGP, left inferior GP; LSGP, left superior GP; MTGP, Marshall tract GP; RIGP, right inferior GP; SPSGP, superior paraseptal GP. This study shows that CT-based EFP-guided CNA for CI-VVS is feasible, can assist RF delivery with high precision and has the potential to overcome the interpatient variability that affects CNA when performed solely by anatomic landmarks. Some EFP may not show autonomic response to RF due to previous successful ablation of a GP concealing subsequent GPs’ response. Alternatively, some CT-identified EFP might not contain GPs, as we did not use direct methods (i.e. extracardiac vagal stimulation) to prove EFP innervation. Of note, we did not target areas with vagal innervation (i.e. left interatrial septum)9 not associated with EFP. Further larger studies with longer follow-up are required to improve CT-based identification of GPs and our understanding of GP pathophysiology. Data are available upon reasonable request.