Page 130 - Cardiac Electrophysiology | A Modeling and Imaging Approach

P. 130
Posterolateral pacing: Errors in determining pacing sites were 6, 6, 4 and 9 mm (average

6.25 mm) for pacing depths of 0, 3.2, 6.4 and 9.6 mm. The reconstructed potential captured
correctly the CCW rotation and expansion of the pattern. For all cases, epicardial electrograms
and breakthrough times were reconstructed with high accuracy.

Infarct and VT (Burnes, JACC 2001;38:2071) 265 . Based on pure Q-wave epicardial electro-
grams, ECGI identified the location and extent of an abnormal EP substrate associated with an
infarct scar (4 – day old infarction). EGM morphologies were very similar to measured ones, with
attenuation of amplitudes. 9 cycles of VT were reconstructed, accurately depicting the epicardial

activation sequence during VT.

Hybrid Dog – Modeling Studies (Burnes, Circulation 2000;101:533 268 and Circulation
2000;102:2152 269 ).

Infarct and VT. Epicardial potentials were recorded from a 490-electrode sock in an open-
chest dog. These potentials were used to generate potentials in a human torso model. Realistic
geometry errors and measurement noise were added to the torso potentials, which were then

used for ECGI input. Infarct was created by LAD occlusion and ethanol injection. Reconstructed
epicardial potential maps captured the pattern and greater than 50% reduction of potentials over
the infarct relative to the pre-infarct control; CC was 0.90 (control) and 0.91 (infarct) for LV pacing,
and 0.77 (control) and 0.79 (infarct) for RA pacing. Corresponding CC values for EGMs were 0.89,

0.90, 0.93, and 0.92, respectively. Reduced amplitude, Q-wave morphology and fractionation of
EGMs over the infarct were captured correctly. Inverse - estimated region of abnormal EP substrate
closely resembled the measured one. During VT, ECGI reconstructed the recorded reentry
pathway, including its key components: the central common pathway, VT exit site, lines of block,

and regions of slow conduction.

Effects of Torso Inhomogeneities (Ramanathan, J Cardiovasc Electrophysiol 2001; 12:241) 270 .
Absolute potential amplitudes were affected by assuming a homogeneous torso, but epicardial

potential patterns, EGM morphologies, isochrones and locations of pacing sites were reconstructed
with comparable accuracy.

Dependence of ECGI Accuracy on Computational and Regularization Methods

As stated in the introduction to this section, accuracy of ECGI depends on the choice of
methodology and details of its execution, including a choice between numerical methods (e.g.,
Boundary Element or Finite Element method) segmentation of surfaces, mesh construction, in-

terpolation, signal conditioning (baseline drift correction, filtering, etc.), choice of a regularization
method, choice of a regularization parameter, etc. In fact, one can tailor the scheme to the applica-
tion to optimize accuracy. Examples are provided on the next page.