In patients with heart failure, mortality is increased in the presence of an ECG QRS pattern of left bundle branch block (LBBB) occurring during intrinsic conduction or provoked by right ventricular (RV) pacing.^1,2 In contrast, overall survival in these patients may improve with pre-excitation of the inferolateral left ventricle (LV) by cardiac resynchronization therapy (CRT).³ Thus, the electrical activation sequence in LV dysfunction imparts important prognostic effects. However, descriptions of cardiac electrical activation normally or in the presence of LV pathology, or during pacing, are remarkably scarce.^4,5,6,7 This information may be necessary to better understand effects of perturbed electrical activation on prognosis and also the diverse clinical responses observed in response to pacing. For example, RV pacing may increase mortality in heart-failure patients with LBBB despite similar QRS patterns on the surface ECG, and up to one third of patients may not respond to CRT.^8,9 Surface ECG recordings do not provide sufficient resolution to address these issues. Investigation requires detailed electrical mapping of ventricular activation. A method to perform this noninvasively, called electrocardiographic imaging (ECGI), has been recently described.¹⁰

ECGI provides noninvasive high-resolution electrical mapping of cardiac excitation on the epicardial surface.^10,11 ECGI images epicardial potentials, electrograms, isochrones (activation sequences), and repolarization patterns from body surface electrocardiographic measurements. The system has been validated extensively in animal models,^{12,13,14,15,16,17} and in humans by comparison to direct epicardial mapping during open-heart surgery¹⁸ and to catheter mapping.^19,20 Reconstruction accuracy superior to 10 mm was consistently obtained in human subjects. ECGI may therefore be used as a noninvasive imaging modality for evaluation of human ventricle cardiac activation under differing conditions. The methodology has been detailed previously.¹⁰ Briefly, ECGI acquires more than 200 channels of body surface electrocardiograms using a multi-electrode vest. Epicardial geometry and body-surface electrode positions are registered simultaneously by a thoracic CT scan. The body surface potential data and the geometry data are processed with algorithms developed to compute epicardial potentials over the entire epicardium, from which epicardial electrograms (typically 600 over the heart surface), isochrones, and repolarization patterns are constructed. All images are obtained during a single beat. This chapter is based on the authors’ previously published study²¹ where we tested the hypothesis that, during LV dysfunction, disease process may alter electrical substrate and modulate effects of LBBB and pacing.

This study evaluated eight patients (72 ± 11 years, 6 male) treated with CRT (Table 16.1). Using ECGI, intrinsic conduction (present in six patients) was compared to RV pacing and to biventricular pacing with optimal atrio-ventricular intervals (defined echocardiographically by the Ritter method²²). Ventricular activation times and sequences during intrinsic and paced rhythms were examined. Time zero was set at the beginning of QRS for all native rhythm episodes and at ventricular pacing stimuli for all paced episodes. Patients were categorized as responders if there was evidence of reverse modeling defined by more than 10% improvement in echocardiographic measures of LVEF (increase) and LV internal dimension (decrease). Echocardiographic assessment was performed at least 6 months after implantation to permit time for remodeling to occur in response to CRT.