In magnetocardiographic (MCG) localization of arrhythmia substrates, a model of the thorax as volume conductor is a crucial component of the calculations. In this study, we investigated different models of the thorax, to determine the most suitable to use in the computations. Our methods and results are as follows. We studied 11 patients with overt Wolff-Parkinson-White syndrome, scheduled for catheter ablation. The MCG registrations were made with a 37-channel "superconducting quantum interference device" system. The underlying equivalent current dipole was computed for the delta-wave. Three models of the thorax were used: the infinite halfspace, a sphere and a box. For anatomical correlation and to define the suitable sphere and box, magnetic resonance images were obtained. As reference we used the position of the tip of the catheter, at successful radio-frequency-ablation, documented by cine-fluoroscopy. Nine patients could be evaluated. The mean errors (range) when using the infinite halfspace, the sphere and the box were 96 (49-125), 21 (5-39), and 36 mm (20-58 mm), respectively (p < 0.0001). In conclusion, the sphere was significantly better suited than the other models tested in this study, but even with this model the accuracy of MCG localization must further improve to be clinically useful. More realistic models of the thorax are probably required to achieve this goal.

Ventricular arrhythmia may in myocardial failure arise as a consequence of remodeling related to hypertrophy and dilatation. In surgically repaired tetralogy of Fallot (TOF), which carries a substantial risk for ventricular arrhythmias and sudden death, the situation is even more complex and several potential arrhythmia mechanisms exist. The authors wanted to test a completely noninvasive localization technique, magnetocardiography (MCG) and magnetic resonance imaging (MRI), to define the origin of ventricular ectopic beats (VEBs) in this model of nonischemic cardiomyopathy. The study included 84 patients with surgically repaired TOF, all 11 subjects with VEBs Lown grade > or = 2 on a 24-hour Holter were included. From 37-channel MCG registrations, the underlying current dipole was computed for the VEBs, and for anatomic correlation MRIs were produced. Eight patients had VEBs of altogether 11 morphologies during the MCG recording. The RVOT was the origin of 6 VEBs, while 4 originated from the RV free wall, and 1 VEB could not be localized. Applying this completely noninvasive technique, it was possible to define different origins of RV ectopies in a complex heart model. Thus, VEBs originated from the RVOT, suggesting a relation to surgical scars and from the nonoperated parts of the RV, supporting that ventricular remodeling might be of equal importance for arrhythmogenesis in this model of cardiomyopathy. This technique can assist in answering the important question of whether there is any difference in prognostic information depending on the origin and related mechanism of VEBs in this and other high risk conditions.