Towards automated electrocardiac map interpretation: An intelligent contouring tool based on spatial aggregation

Abstract

The interpretation of cardiac potential maps is essential for the localization of the anomalous excitation sites associated with electrical conduction pathologies, such as arrythmia, and its automation would make clinical practice both realistic and of great impact on health care. Conventional contouring tools allow us to efficiently visualize patterns of electrical potential distribution but they do not facilitate the automated extraction of general rules necessary to infer the correlation between pathopysiological patterns and wavefront structure and propagation. The Spatial Aggregation (SA) approach, which aims at interpreting a numeric input field, is potentially capable of capturing structural information about the underlying physical phenomenon, and of identifying its global patterns and the causal relations between events. These characteristics are due to its hierarchical strategy in aggregating objects at higher and higher levels. However, when dealing with a complex domain geometry and/or with non uniform data meshes, as in our context, the original version of SA may unsoundly perform contouring. Isocurve entanglement and/or segmentation phenomena may occur due to metric-based adjacency relations and a scarse density of isopoints. This paper discusses the problems above, and presents definitions and algorithms for a sound representation of the spatial contiguity between isopoints and for the construction of isocurves. © Springer-Verlag Berlin Heidelberg 2003.