Entropy-Based Algorithm to Detect Life Threatening Cardiac Arrhythmias Using Raw Electrocardiogram Signals

Ventricular tachycardia or fibrillation (VT-VF) as fatal cardiac arrhythmia is the major cause leading to sudden cardiac death. It is crucial for the patient to receive immediate medical intervention when either VT or VF occurs. The aim of this study is to investigate the possibility of predicting ventricular arrhythmia from electrocardiogram (ECG) signals. We use symbolic dynamics together with an entropy based complexity measure to discriminate between normal sinus rhythm (NSR) and life threatening arrhythmias like, ventricular tachycardia/fibrillation (VT/VF). The statistical analyses show that either, Renyi or Shannon Entropy measure is found to have potential in discriminating normal and VT/VF subjects and thus can significantly add to the prognostic value of traditional cardiac analysis. However, it is found that Renyi entropy outperforms Shannon entropy. The receiver operating characteristic curve (ROC) analysis confirms the robustness of this new approach and exhibits an average sensitivity of about 91.8% (94.9%), specificity of about 95.4% (97.5%), positive predictivity of around 96.1% (95.6%) and accuracy of about 93.4% (96.6%), with Renyi entropy to distinguish between normal and VT (VF) subjects. The presented method is simple, but with high detection quality, computationally fast and well suited for real time implementation in automated external or implantable cardioverter-defibrillators.