Biomedical Signal Analysis Laboratory  
 
     
       
   
Separation of Ventricular Tachycardia and Ventricular Fibrillation Using Two Unipolar Electrograms
 
Stephanie A. Caswell, Julie Thompson, Lorenzo DiCarlo, Janice M. Jenkins
 
In third-generation implantable antitachycardia devices (ATDs), immediate defibrillation is required for ventricular fibrillation (VF) while lower-energy therapies can convert many ventricular tachycardias (VTs). However, separation of these rhythms by rate measures is not always feasible. This study employed a measurement to quantify the coherence between two unipolar electrograms from the same catheter, where incoherence between signals indicates VF. A normalized cross correlation (CC) measured similarity between electrogram morphologies and standard deviation (STD) and interquartile range (IQR) (Q3 minus Q1) measured consistency over a passage. Two unipolar electograms were derived from the distal and proximal electrodes of a catheter located in the RVA (1 cm spacing). A wire located remotely in the femoral vein was used as the reference electrode. Sinus rhythm (SR), VT, and VF passages from 10 patients were recorded on FM tape at 3 3/4 ips (1-500 Hz) and digitized at 1000 Hz. The CC was computed for a 150 point window spanning the depolarization with +/-10 point shift and the peak CC value (PCC) was chosen. VF was separated from VT/SR by the STD and the IQR of the PCC values for a passage where a large STD or IQR indicates VF. \nRESULTS: Sensitivity (sens) and specificity (spec) were derived for all patients where sensitivity is with respect to VF only. The mean +/- std for PCC was 0.85+/-0.016 for SR, 0.79+/-0.046 for VT and 0.56+/-0.332 for VF. Patient-indepedent STD thresholds of 0.15-0.25 provided 100% sens and 100% spec. The IQR mean was 0.017 for SR, 0.10 for VT, and 1.03 for VF, providing 100% sens and 100% spec for IQR thresholds of 0.3-0.6.\nCONCLUSION: This method was able to successfully separate incoherent (VF) from coherent rhythms (SR, VT), improving upon ATD rate algorithms. Classification was achieved with patient-independent thresholds. In addition, 'polymorphic', but coherent rhythms (such premature ventricular depolarizations) would not be misclassified as VF. Proper distinction of VT from VF would allow defibrillation to be deferred for consideration of lower energy therapies, providing significant energy savings.
 
Winner of the Rosanna Degani Young Investigator Competition. Published in Proceedings of Computers in Cardiology 1996
 
Research Topics
 
Separation of Ventricular Tachycardia and Ventricular Fibrillation Using Two Unipolar Electrograms
 
Detection Algorithms in Implantable Cardioverter Defibrillators
 
Ventricular Tachycardia Versus Ventricular Fibrillation: Discrimination by Current Antitachycardia Devices
 
Chronic Bipolar Intraventricular Electrograms Are Stable During Changes in Body Position and Activity: Implications for Antitachycardia Devices
 
Ventricular Fibrillation Detection
 
Arrhythmia Classifier
 
Web Based Medical Signal Database
         
    Director: Dr. Stephanie Schuckers    Clarkson University    West Virginia University

Site designed by Aliki Cooper Web Design