Anatomic predictors of recurrence after cryoablation for atrial fibrillation: a computed tomography based composite score.

Abstract

Effective pulmonary vein isolation (PVI) with cryoablation depends on adequate occlusion of pulmonary veins (PV) by the cryoballoon and is therefore likely to be affected by PV and left atrial (LA) anatomical characteristics and variants. Thus, the objective of this study was to investigate the effect of LA and PV anatomy, evaluated by computed tomography (CT), on acute and long-term outcomes of cryoablation for atrial fibrillation (AF). Fifty-eight patients (64.72+9.44years, 60.3% male) undergoing cryoablation for paroxysmal or early persistent AF were included. Pre-procedural CT images were analyzed to evaluate LA dimensions and PV anatomical characteristics. Predictors of recurrence were identified using regression analysis. 60.3% of patients had two PVs on each side with separate ostia, whereas 29.3% and 10.3% had right middle and left common PVs, respectively. The following anatomic characteristics were found to be independent predictors of recurrence: right superior PV ostial max:min diameter ratio > 1.32, left superior PV ostial max:min diameter ratio > 1.2, right superior PV antral circumference > 69.1mm, right inferior PV antral circumference > 61.38mm, right superior PV angle > 22.7°. Using these factors, LA diameter and right middle PV, a scoring model was created for prediction of "unfavorable" LA-PV anatomy (AUC = 0.867, p = 0.000009, score range = 0-7). Score of ≥ 4 predicted need for longer cryoenergy ablation (p = 0.039) and more frequent switch to radiofrequency energy (p = 0.066) to achieve PVI, and had a sensitivity of 83.3% and specificity of 82.5% to predict clinical recurrence. CT-based scoring system is useful to identify "unfavorable" anatomy prior to cryo-PVI, which can result in procedural difficulty and poor outcomes.