A comparison of simulated QRS isointegral maps resulting from pacing at adjacent sites - Implications for the spatial resolution of pace mapping using body surface potentials
R. Hren et Bb. Punske, A comparison of simulated QRS isointegral maps resulting from pacing at adjacent sites - Implications for the spatial resolution of pace mapping using body surface potentials, J ELCARDIOL, 31, 1998, pp. 135-144
The precise localization of ventricular tachycardia (VT) foci is a prerequi
site for the successful radiofrequency catheter ablation in patients. The p
urpose of this study was to systematically quantify over what distance adja
cent sites in the right ventricular (RV) and left ventricular (LV) epicardi
um and LV endocardium could be distinguished by inspecting morphological fe
atures of QRS isointegral maps using statistical methods. We investigated t
he spatial resolution of QRS isointegral maps by means of an anatomically a
ccurate computer model of the human ventricular myocardium that incorporate
s a bidomain model for simulating the realistic activation sequences and th
e oblique dipole model in combination with the boundary element method for
calculating extracardiac potentials. In this model, we initiated activation
sequences at a total of 183 epicardial and 75 LV endocardial pacing sites,
positioned in three levels (basal, middle, and apical). For each of the 25
8 pacing sites, we calculated a set of 10 QRS isointegral maps with added G
aussian noise at 117 leads (covering the anterior and posterior torso) and
at 32 leads (covering only the anterior torso), respectively. Sets of maps
were then cross correlated and root-mean-square (RMS) values of difference
maps were calculated for all possible pairs of pacing sites on the same lev
el. We applied the nonparametric unpaired Kolmogorov-Smirnov test and defin
ed the spatial resolution as the pacing site separation at which the differ
ences in correlation coefficients and RMS differences were significant (lev
el P < .05). We observed significant differences in maps when the distances
between pacing sites were on average (I SD) greater than 4.3 +/- 1.0 mm. Z
n more than 90% of pacing sites, the significant differences in maps were o
bserved within 4 mm even when using a 32-lead mapping system. The findings
of our study provide theoretical evidence that QRS isointegral maps may off
er noninvasive means for preinterventional planning of the ablative treatme
nt in localizing both endocardial and epicardial sites of origin of VT.