Identification of supraventricular tachycardia mechanisms with surface electrocardiograms using a convolutional neural network.
Artificial intelligence
Convolutional neural network
Electrocardiogram
Long RP tachycardia
Machine learning
Supraventricular tachycardia
Journal
Heart rhythm O2
ISSN: 2666-5018
Titre abrégé: Heart Rhythm O2
Pays: United States
ID NLM: 101768511
Informations de publication
Date de publication:
Aug 2023
Aug 2023
Historique:
medline:
30
8
2023
pubmed:
30
8
2023
entrez:
30
8
2023
Statut:
epublish
Résumé
It remains difficult to definitively distinguish supraventricular tachycardia (SVT) mechanisms using a 12-lead electrocardiogram (ECG) alone. Machine learning may identify visually imperceptible changes on 12-lead ECGs and may improve ability to determine SVT mechanisms. We sought to develop a convolutional neural network (CNN) that identifies the SVT mechanism according to the gold standard of SVT ablation and to compare CNN performance against experienced electrophysiologists among patients with atrioventricular nodal re-entrant tachycardia (AVNRT), atrioventricular reciprocating tachycardia (AVRT), and atrial tachycardia (AT). All patients with 12-lead surface ECG during sinus rhythm and SVT and had successful SVT ablation from 2013 to 2020 were included. A CNN was trained using data from 1505 surface ECGs that were split into 1287 training and 218 test ECG datasets. We compared the CNN performance against independent adjudication by 2 experienced cardiac electrophysiologists on the test dataset. Our dataset comprised 1505 ECGs (368 AVNRT, 304 AVRT, 95 AT, and 738 sinus rhythm) from 725 patients. The CNN areas under the receiver-operating characteristic curve for AVNRT, AVRT, and AT were 0.909, 0.867, and 0.817, respectively. When fixing the specificity of the CNN to the electrophysiologist adjudicators' specificity, the CNN identified all SVT classes with higher sensitivity: (1) AVNRT (91.7% vs 65.9%), (2) AVRT (78.4% vs 63.6%), and (3) AT (61.5% vs 50.0%). A CNN can be trained to differentiate SVT mechanisms from surface 12-lead ECGs with high overall performance, achieving similar performance to experienced electrophysiologists at fixed specificities.
Sections du résumé
Background
UNASSIGNED
It remains difficult to definitively distinguish supraventricular tachycardia (SVT) mechanisms using a 12-lead electrocardiogram (ECG) alone. Machine learning may identify visually imperceptible changes on 12-lead ECGs and may improve ability to determine SVT mechanisms.
Objective
UNASSIGNED
We sought to develop a convolutional neural network (CNN) that identifies the SVT mechanism according to the gold standard of SVT ablation and to compare CNN performance against experienced electrophysiologists among patients with atrioventricular nodal re-entrant tachycardia (AVNRT), atrioventricular reciprocating tachycardia (AVRT), and atrial tachycardia (AT).
Methods
UNASSIGNED
All patients with 12-lead surface ECG during sinus rhythm and SVT and had successful SVT ablation from 2013 to 2020 were included. A CNN was trained using data from 1505 surface ECGs that were split into 1287 training and 218 test ECG datasets. We compared the CNN performance against independent adjudication by 2 experienced cardiac electrophysiologists on the test dataset.
Results
UNASSIGNED
Our dataset comprised 1505 ECGs (368 AVNRT, 304 AVRT, 95 AT, and 738 sinus rhythm) from 725 patients. The CNN areas under the receiver-operating characteristic curve for AVNRT, AVRT, and AT were 0.909, 0.867, and 0.817, respectively. When fixing the specificity of the CNN to the electrophysiologist adjudicators' specificity, the CNN identified all SVT classes with higher sensitivity: (1) AVNRT (91.7% vs 65.9%), (2) AVRT (78.4% vs 63.6%), and (3) AT (61.5% vs 50.0%).
Conclusion
UNASSIGNED
A CNN can be trained to differentiate SVT mechanisms from surface 12-lead ECGs with high overall performance, achieving similar performance to experienced electrophysiologists at fixed specificities.
Identifiants
pubmed: 37645266
doi: 10.1016/j.hroo.2023.07.004
pii: S2666-5018(23)00163-0
pmc: PMC10461210
doi:
Types de publication
Journal Article
Langues
eng
Pagination
491-499Informations de copyright
© 2023 Heart Rhythm Society. Published by Elsevier Inc.
Références
JAMA. 2016 Dec 13;316(22):2402-2410
pubmed: 27898976
Heart Rhythm. 2004 Oct;1(4):393-6
pubmed: 15851189
Heart Rhythm. 2015 Jul;12(7):1667-76
pubmed: 25828600
Nat Med. 2019 Jan;25(1):70-74
pubmed: 30617318
J Am Coll Cardiol. 2008 Nov 18;52(21):1711-7
pubmed: 19007691
JAMA Cardiol. 2021 Nov 1;6(11):1285-1295
pubmed: 34347007
Science. 2006 Jul 28;313(5786):504-7
pubmed: 16873662
Am J Cardiol. 2003 May 1;91(9):1084-9
pubmed: 12714151
Circ Cardiovasc Qual Outcomes. 2019 Sep;12(9):e005289
pubmed: 31525078
Heart. 2022 Jun 24;108(14):1107-1113
pubmed: 34635482
Acta Cardiol. 2010 Apr;65(2):171-6
pubmed: 20458824
J Am Coll Cardiol. 1997 Feb;29(2):394-402
pubmed: 9014995
Europace. 2009 Jul;11(7):944-8
pubmed: 19525496
J Am Coll Cardiol. 1993 Jan;21(1):85-9
pubmed: 8417081
Circulation. 1983 Aug;68(2):368-73
pubmed: 6861312
J Am Coll Cardiol. 2001 Oct;38(4):1163-7
pubmed: 11583898
Circ Cardiovasc Qual Outcomes. 2020 Oct;13(10):e006556
pubmed: 33079589
Europace. 2022 Jul 21;24(7):1186-1194
pubmed: 35045172
Nat Med. 2019 Jan;25(1):65-69
pubmed: 30617320
Europace. 2008 May;10(5):624-8
pubmed: 18400770
JACC Clin Electrophysiol. 2021 Nov;7(11):1358-1365
pubmed: 34217658
Am J Cardiol. 2003 Dec 15;92(12):1424-8
pubmed: 14675578
Lancet. 2019 Sep 7;394(10201):861-867
pubmed: 31378392
Circ Arrhythm Electrophysiol. 2013 Jun;6(3):597-605
pubmed: 23629734
Ann Noninvasive Electrocardiol. 2002 Apr;7(2):120-6
pubmed: 12049683