Title
Qrs Fragmentation Index As A New Discriminator For Early Diagnosis Of Heart Diseases
Abstract
In the past few years, the presence of fragmentation in the QRS complex has been demonstrated to be related to diseases such as myocardial fibrosis, cardiac sarcoidosis, arrythmogenic cardiopathies, acute coronary syndrome, and Brugada syndrome, among others. The detection of fragmentation in the QRS is usually carried out manually, which represents a subjective pattern recognition task that demands an effort by the clinician, increasing with the number of patients. These problems have made the process of fragmentation detection a good candidate to its automatization. In this work, we used a database with over six-thousand 12-lead ECG from Hospital Virgen de la Arrixaca de Murcia (Spain), which where digitally recorded with GE MAC5000. Affected and non-affected patients records were extracted for computerized analysis. Clinical supervision was performed for gold-standard development and for signal classification. Fragmentation detection algorithms were developed using first and second derivatives calculation in the pre-qualified segments of the signal, after fiducial point detection. The obtained results were 96.88% sensitivity, 72.92% specificity, and 82.50% accuracy. These results confirm that it is possible to automatically detect fragmentation, constituting a relevant tool to pre-qualify patients for further diagnostic-tests, and it also opens new opportunities for computerized diagnosis.
Year
DOI
Venue
2017
10.22489/CinC.2017.267-258
2017 COMPUTING IN CARDIOLOGY (CINC)
Field
DocType
Volume
Cardiac sarcoidosis,Fiducial marker,Discriminator,Internal medicine,Acute coronary syndrome,Brugada syndrome,Cardiology,Fragmentation (computing),Signal classification,QRS complex,Medicine
Conference
44
ISSN
Citations 
PageRank 
2325-8861
0
0.34
References 
Authors
0
8