Implantable Cardioverter Defibrillator Stored Ecgs

Author : Siu-Ki Kwan
Publisher :
ISBN 13 : 9781361470954
Total Pages : pages
Book Rating : 4.4/5 (79 download)

DOWNLOAD NOW!

Download or read book Development of High Performance Implantable Cardioverter Defibrillator Based Statistical Analysis of Electrocardiography written by Siu-Ki Kwan and published by . This book was released on 2017-01-27 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: This dissertation, "Development of High Performance Implantable Cardioverter Defibrillator Based Statistical Analysis of Electrocardiography" by Siu-ki, Kwan, 關兆奇, was obtained from The University of Hong Kong (Pokfulam, Hong Kong) and is being sold pursuant to Creative Commons: Attribution 3.0 Hong Kong License. The content of this dissertation has not been altered in any way. We have altered the formatting in order to facilitate the ease of printing and reading of the dissertation. All rights not granted by the above license are retained by the author. Abstract: Abstract of thesis entitled Development of High Performance Implantable Cardioverter Defibrillator Based on Statistical Analysis of Electrocardiography Submitted by KWAN Siu-ki for the Degree of Doctor of Philosophy at The University of Hong Kong in May 2007 Sudden Cardiac Death is mainly caused by cardiac arrhythmia. An implantable Cardioverter Defibrillator ICD implanted in a patient at risk gives a corrective signal when arrhythmia is detected. Nevertheless, predicting the incidence of arrhythmia and preventing its occurrence remain a major concern, and the ability to distinguish atrial fibrillation and flutter from normal sinus rhythm in good time can save many lives worldwide. Fast and accurate detection is very critical, and hence research has been done to detect such heart conditions. Recently, a multi-feature Bayesian classifier was successfully developed to select five relevant features from raw electrocardiogram ECG and incorporated into a multi-feature Bayesian classifier for arrhythmia detection. A patent (number: 10/141,104) was granted to this multi-feature classifier development. Despite its success, the design was not optimized to fully utilize the information contained in the data set of size 364. In this project, a larger data set of 1568 raw intra-atrial ECG episodes were collected from 20 patients during electrophysiological procedures, so that an in-depth multi-variate statistical analysis could be performed. Three features out of the original five were identified and a new two-layer architecture was proposed to classify raw ECG into sinus rhythm SR, atrial fibrillation AF and atrial flutter AFL. The two-layered three-feature design was tested both on artificial neural network and multi-feature Bayesian classification. Finally, a novel two-layered multi-feature Bayesian was selected and realized on a single-chip microcomputer for ICD manufacturing. In comparing the new two-layered three-feature Bayesian ICD classifier to the five-feature one, the accuracies have been improved: from 98.72% to 99.62% in SR, from 97.32% to 98.21% in AF, and from 97.96% to 98.34% in AFL; the sensitivity from 96.57% to 99.14% in SR, from 97.95% to 98.71% in AF and from 95.67% to 96.13% in AFL; the specificities from 99.34% to 99.75% in SR, from 96.70% to 97.72% in AF and from 98.85% to 99.20% in AFL. In this connection, the increase in SR accuracy will reduce the number of painful inappropriate ICD shocks to patient and will make precious ICD battery power last longer. The increase in specificities of both AF and AFL will reduce the number of under-sensing of arrhythmia and avoid missed timely treatment to patient at risk. Both are vital measures in biomedical device. Hence, the proposed two-layered classifier performs more accurately than existing methods by employing fewer features, and the experimental results together with the microcomputer prototyping provide a solid foundation for designing low- power ICD. (406 words) DOI: 10.5353/th_b3832028 Subjects: Defibrillators Cardiovascular instruments, Implanted Arrhythmia - Diagnosis Bayesian statistical decision theory