Ventricular arrhythmia localisation by computer processing of fragmented cardiac myopotentials

Abstract

Accurate localisation of arrhythmogenic origin is necessary for successful surgery on patients with ventricular arrhythmias, and fragmentation mapping is one of the principal techniques used intraoperatively for this purpose. Fragmentation mapping is normally based on subjective characterisation of fragmented cardiac myopotentials. These fragmented electrograms are an important feature of arrhythmogenic scar and have been used with moderate success to direct surgery. However, very little is known or understood about fragmented electrograms. The aim of the authors' study is to produce a fragmentation mapping system based on the automated classification of electrograms. Therefore, the two main issues that are addressed are the extraction of electrograms from recorded and occasionally noisy signals and evaluating the characteristics applicable to discriminating between normal and abnormal electrograms. The output of the fragmentation mapping system is in the form of a graphical map representing 53 grid positions on the left ventricular epicardium. This will provide a useful tool that will potentially localise the arrhythmogenic origin more rapidly than manual schemes, thus allowing a more effective treatment of arrhythmias.<>