The differences between electromagnetic-based mapping (EM) and impedance-based mapping (IM) in 3D anatomical reconstruction have not been fully clarified. We aimed to investigate the anatomical accuracy between EM (MediGuide™) and IM (EnSite Velocity™) systems.

We investigated 15 consecutive patients (10 males, mean age 58±9 years) who underwent pulmonary veins (PVs) isolation for paroxysmal atrial fibrillation (PAF). Contrast-enhanced computed tomography (CT) image of the left atrium (LA) was acquired before ablation and the 3D geometry of the LA was constructed using EM during ablation procedure. We measured the 4 PV angles between the main trunk of each PV and the posterior LA after field scaling. Additionally, the posterior LA surface area was measured. The variables were compared to those of CT-based geometry. A control group of 40 patients who underwent conventional PVs isolation using IM were also evaluated.

The actual and relative changes of EM and CT-based geometry in all PV angles and posterior LA were significantly smaller compared to those of IM and CT-based geometry. Intraclass correlation coefficient (ICC) between EM and CT-based geometry were 0.871 (right superior pulmonary vein [RSPV]), 0.887 (right inferior pulmonary vein [RIPV]), 0.853 (left superior pulmonary vein [LSPV]), 0.911 (left inferior pulmonary vein [LIPV]), and 0.833 (posterior LA). On the other hand, ICC between IM and CT-based geometry were 0.548 (RSPV), 0.639 (RIPV), 0.691 (LSPV), 0.706 (LIPV), and 0.568 (posterior LA).

Image integration with EM enables high accurate visualization of cardiac anatomy compared to IM in PAF ablation.