An algorithm for imaging isochrones of ventricular activation on patient-specific epicardial surface

Shijie Zhou, John Sapp, Amir AbdelWahab, Milan Horacek
Dalhousie University


Abstract

Electrocardiographic imaging (ECGI) is a non-invasive imaging modality that has been shown to be able to provide information useful for pre-procedural planning of catheter ablation interventions. The methodology involves reconstruction of unipolar electrograms (EGMs) and isochronal maps on the epicardial surface from body-surface potentials. Isochronal activation maps visualize propagation of the activation wavefront in the myocardium and provide a summary of an entire activation sequence. We have developed an improved algorithm for evaluating the global activation times, constructed from measurements of propagation delay for pairs of computed EGMs at neighboring locations. First, the cross-correlation method is used to determine the time delay in the activation among pairs of neighboring locations. Next, a sparse linear system is constructed by using known activation relationships of neighboring locations with activation delays as inputs. To solve the sparse system, we use a sparse Bayesian learning method to calculate the global activation times. The principle aim of this study was to assess the proposed method in both structurally normal and scarred ventricular myocardium. Isochronal maps of calculated activation times were compared with local activation times (LAT) derived from directly-measured epicardial EGMs obtained by electroanatomic contact mapping, for pacing delivered by implantable cardioverter defibrillator at the endocardial right-ventricular apex. We found that even in the presence of infarct scar, isochronal maps calculated by the proposed method correlated closely with known LAT exported from an electroanatomic mapping system.