Inverse Computations of Epicardial Potentials

Yoram Rudy

Cardiac Bioelectricity Research and Training Center, Case Western Reserve University,
Cleveland, Ohio, USA

Correspondence: Y Rudy, Cardiac Bioelectricity Center, 509 Wickenden Building, Cleveland, OH 44106-7207, USA.
E-mail: yxr@po.cwru.edu, phone 216-368-4051, fax 216-368-8672

1.  Introduction

Cardiac arrhythmias continue to be a leading cause of death and disability. Despite this fact, a true noninvasive imaging modality for cardiac electrical function is not yet available for clinical application. Efforts to develop such methodology have focused on the reconstruction of equivalent myocardial sources, activation times on the surface of the heart, or alternatively, potentials on the surface of the heart from which electrograms and activation times (isochrones) can also be derived. In our laboratory, the focus has been the reconstruction of potentials, electrograms and isochrones on the heart surface [we termed this approach Electrocardiographic Imaging (ECGI)]. This conference proceedings paper is a summary of my presentation at the ICE 2003 in Helsinki. It provides examples of reconstructions of epicardial potentials and electrograms in cases where epicardial activation times (isochrones) alone cannot provide the required information. These examples were published previously and include noninvasive reconstructions of: (1) Intramural activation before it reaches the epicardium [Oster et al., 1998], (2) Intramural ventricular tachycardia in infarcted hearts [Burnes et al., 2001], (3) Abnormal electrophysiological substrate associated with myocardial infarction [Burnes et al., 2000], and (4) Dispersion of myocardial repolarization [Ghanem et al., 2001].

2.  Materials and Methods

A detailed description of methods and protocols for each study is provided in each respective reference. Experiments were conducted in a torso-tank setup, where an isolated dog heart was placed in a human shaped torso-tank [studies (1) and (2)], or in an open-chest dog [studies (3) and (4)]. In all cases, epicardial potentials were computed using Tikhonov regularization of various orders. From the epicardial potential maps over the entire cardiac cycle, electrograms were constructed over the entire epicardial surface. The time of maximum negative derivative (“intrinsic deflection”) of each electrogram was taken as local activation time for the construction of epicardial isochrones. The reconstructed epicardial potentials, electrograms and isochrones were validated and evaluated by direct comparison with their measured counterparts.

3.  Results

3.1. Intramural Activation [Oster et al., 1998]

Cardiac excitation and arrhythmogenesis involve the three-dimensional ventricular wall, not only the epicardial and endocardial surfaces. Such intramural activity generates potentials on the heart surface even prior to arrival of the activation front (“breakthrough”) at the surface and before surface activation occurs. During this time the surface potentials reflect activity in the depth of the myocardial wall. However, since surface activation has not yet occurred, surface activation times (isochrones) cannot be defined, nor recovered, using inverse computations. In this study, we examined the ability of ECGI to characterize intramural myocardial activation, prior to surface activation, through noninvasive reconstruction of epicardial potentials. Intramural activation was initiated by pacing from five intramural depths. Reconstructed (and measured) epicardial potentials showed characteristic patterns: (1) early in activation, a central negative region with two flanking maxima aligned with the orientation of myocardial fibers at the depth of pacing (fibers rotate counterclockwise with increasing depth from epicardium to endocardium). (2) counterclockwise rotation of positive potentials with time for subepicardial pacing, as activation propagates towards the endocardium; (3) clockwise rotation of positive potentials for subendocardial pacing, as activation propagates towards the epicardium; (4) dual (clockwise and counterclockwise) rotation of positive potentials for midmyocardial pacing, as activation propagates in both transmural directions toward the endocardium and epicardium. These patterns reflect the fiber structure of the three-dimensional ventricular wall and allow us to estimate the depth of intramyocardial electrophysiological events (e.g., initiation sites). Thus, ECGI of epicardial potentials can provide information on the spread of excitation in the three-dimensional myocardium before activation of the surfaces has occurred.

3.2. Intramural Ventricular Tachycardia [Burnes et al., 2001]

In general, arrhythmogenic activity involves intramural excitation, including three-dimensional reentry and ectopic focal activation. In this study, we evaluated the ability of ECGI of epicardial potentials to determine the activation pattern during intramural reentry and locate the intramural components of the reentry pathway. The experiments were conducted in a dog heart with a four-day-old infarction. The heart was suspended in a human shaped torso-tank. Measured torso-surface potentials were used to noninvasively compute epicardial potentials, electrograms and isochrones. Reconstructions were performed during nine cycles of ventricular tachycardia (VT).

The infarct encompassed much of the anterior left ventricle (LV), as indicated by pure Q-wave epicardial electrograms (and verified with TTC staining). These electrograms showed no evidence of local activation, indicating complete conduction block. This ruled out the possibility of epicardial wavefront propagation across the infarcted area and indicated that the reentry involved intramural and septal activation. The reconstructed epicardial potential maps and epicardial electrograms provided sufficient information from which the intramural sequence of reentrant activation could be deduced. The results identified septal conduction in the posterior-anterior direction and partial activation of the conduction system as important components of the reentry.

3.3. Infarct Substrate [Burnes et al., 2000]

Myocardial infarction and subsequent remodeling creates altered electrophysiological (EP) substrate that is highly arrhythmogenic. Experiments were conducted in a normal (control) heart and in the same heart 2 hours after left anterior descending (LAD) coronary artery occlusion and ethanol injection to create an infarct. ECGI-reconstructed epicardial potential maps showed a large region of negative potentials covering the LV infarct region (indicating presence of underlying electrically-inactive necrotic tissue, which cannot support normal conduction). Reconstructed electrograms from the infarct region assumed a large Q-wave morphology, some with superimposed sharp biphasic RS waves of small magnitude. The Q-wave is indicative of the electrically-inactive tissue. The superimposed small deflections reflect local activation of small islands of surviving tissue within the infarct. Thus, ECGI in terms of epicardial potentials can noninvasively locate the infarct region and characterize its electrophysiological properties (e.g., “patchiness” of the substrate due to existence of isolated islands of activity).

3.4. Dispersion of Repolarization [Ghanem et al., 2001]

Repolarization abnormalities, which create large dispersions of myocardial repolarization, underlie many potentially fatal arrhythmias. In this study we demonstrated the ability of ECGI of epicardial potentials to noninvasively detect, localize and quantify regions of increased dispersion of repolarization in the heart. Through regional epicardial warming and cooling, we induced localized alterations in myocardial repolarization. From reconstructed epicardial electrograms, we determined local activation-recovery intervals (ARI). During control, a uniform ARI distribution existed over the entire heart surface. Regional warming induced short ARIs over the warmed region whereas, during cooling, prolonged ARIs were induced over the cooled region. Simultaneous adjacent warming and cooling created adjacent regions of short and long ARIs and steep repolarization gradients. In all cases, the ECGI reconstructed epicardial maps captured the regions of altered repolarization and the regional increase of ARI differences, reflecting the underlying dispersion of repolarization induced by the warming and cooling protocols. Similarly, epicardial QRST integral (integration of epicardial potentials over the QRST interval) maps were accurately reconstructed by ECGI, reflecting the localized alteration of repolarization and increased dispersion.

4.  Conclusion

I have tried, through examples from our previous work, to illustrate situations where reconstruction of the epicardial potentials over the entire cardiac cycle is necessary for deriving important physiological and clinical information. Direct reconstruction of epicardial activation times preserves only a single point in time on each electrogram, discarding all other potential information during the cardiac cycle. As illustrated here, the entire epicardial potential maps and electrogram waveforms are necessary for reconstruction of repolarization properties and of electrophysiological characteristics of abnormal arrhythmogenic substrate in diseased hearts (e.g. Q-wave electrograms and fractionated electrograms in infarct substrates). In addition, epicardial potentials and electrograms provide information on intramural activity in the ventricular wall, even before it arrives at the surface and surface activation times can be defined. In fact, direct reconstruction of surface activation times can erroneously assign activation times to inactive tissue, such as the surface aspect of an infarct or the intact epicardium (endocardium) during intramural activity that has not reached the surfaces.

Acknowledgements

Preparation of this manuscript was supported by NIH grants R37 HL-33343 and R01 HL- 49054. Experiments for studies (1) – (3) were conducted in Dr. Bruno Taccardi’s laboratory at the University of Utah, Salt Lake City. Experiments for study (4) were conducted in our laboratory at Case Western Reserve University in collaboration with Dr. Albert Waldo.

References

Oster HS, Taccardi B, Lux RL, Ershler PR, and Rudy Y, Electrocardiographic imaging: noninvasive characterization of intramural myocardial activation from inverse reconstructed epicardial potentials and electrograms, Circulation, 97:1496-1507, 1998.

Burnes JE, Taccardi B, Ershler P, and Rudy Y, Noninvasive ECG imaging of substrate and intramural ventricular tachycardia in infarcted hearts, J Amer College Cardiol (JACC), 38:2071-2078, 2001.

Burnes JE, Taccardi B, MacLeod RS, and Rudy Y, Noninvasive electrocardiographic imaging of electrophysiologically abnormal substrate in infarcted hearts: a model study, Circulation, 101:533-540, 2000.

Ghanem RN, Burnes JE, Waldo AL, and Rudy Y, Imaging dispersion of myocardial repolarization II. Noninvasive reconstruction of epicardial measures, Circulation, 104:1306-1312, 2001.