An additional and easy algorithm based on morphological analysis of unipolar electrogram: misclassification in implantable cardioverter-defibrillators

摘要

A classification method based on electrogram (EGM) morphology is proposed in order to distinguish among Ventricular Tachycardias (VT), Supraventricular Tachycardias (SVT) and Ventricular Fibrillations (VF), which are the most difficult to separate using only rate. The authors' algorithm starts to work only when rate criterion is fulfilled. The authors have broached a classification of isolated beats and afterwards, applying very simple rules, the full episode is classified. The study has been carried out using a wide database composed of unipolar EGMs (fs=200 Hz) from ICDs, recorded during clinical episodes, where rate criterion had been fulfilled. For every beat of an episode a number of morphological parameters, normalized using sinus rhythm morphology and morphology during the arrhythmia, is calculated and powerful statistic tools such as logistic regression are used to achieve a good diagnosis. Logistic regression produces functions, which can discriminate effectively among different morphological complexes. They are simple and have a low computational complexity, which means that they can be used without delay and almost without an additional cost. The algorithm shows a really good level of performance at both levels of classification (isolated complex and episode) with diagnosis sensitivity and specificity above 90%.