Reliable Deep Learning-Based Detection of Misplaced Chest Electrodes During Electrocardiogram Recording: Algorithm Development and Validation.
ECG
ECG interpretation
cardiovascular disease
deep learning
electrode misplacement
engineering
feature engineering
machine learning
myocardial
myocardial infarction
physicians
Journal
JMIR medical informatics
ISSN: 2291-9694
Titre abrégé: JMIR Med Inform
Pays: Canada
ID NLM: 101645109
Informations de publication
Date de publication:
16 Apr 2021
16 Apr 2021
Historique:
received:
29
10
2020
accepted:
27
02
2021
revised:
12
02
2021
entrez:
16
4
2021
pubmed:
17
4
2021
medline:
17
4
2021
Statut:
epublish
Résumé
A 12-lead electrocardiogram (ECG) is the most commonly used method to diagnose patients with cardiovascular diseases. However, there are a number of possible misinterpretations of the ECG that can be caused by several different factors, such as the misplacement of chest electrodes. The aim of this study is to build advanced algorithms to detect precordial (chest) electrode misplacement. In this study, we used traditional machine learning (ML) and deep learning (DL) to autodetect the misplacement of electrodes V1 and V2 using features from the resultant ECG. The algorithms were trained using data extracted from high-resolution body surface potential maps of patients who were diagnosed with myocardial infarction, diagnosed with left ventricular hypertrophy, or a normal ECG. DL achieved the highest accuracy in this study for detecting V1 and V2 electrode misplacement, with an accuracy of 93.0% (95% CI 91.46-94.53) for misplacement in the second intercostal space. The performance of DL in the second intercostal space was benchmarked with physicians (n=11 and age 47.3 years, SD 15.5) who were experienced in reading ECGs (mean number of ECGs read in the past year 436.54, SD 397.9). Physicians were poor at recognizing chest electrode misplacement on the ECG and achieved a mean accuracy of 60% (95% CI 56.09-63.90), which was significantly poorer than that of DL (P<.001). DL provides the best performance for detecting chest electrode misplacement when compared with the ability of experienced physicians. DL and ML could be used to help flag ECGs that have been incorrectly recorded and flag that the data may be flawed, which could reduce the number of erroneous diagnoses.
Sections du résumé
BACKGROUND
BACKGROUND
A 12-lead electrocardiogram (ECG) is the most commonly used method to diagnose patients with cardiovascular diseases. However, there are a number of possible misinterpretations of the ECG that can be caused by several different factors, such as the misplacement of chest electrodes.
OBJECTIVE
OBJECTIVE
The aim of this study is to build advanced algorithms to detect precordial (chest) electrode misplacement.
METHODS
METHODS
In this study, we used traditional machine learning (ML) and deep learning (DL) to autodetect the misplacement of electrodes V1 and V2 using features from the resultant ECG. The algorithms were trained using data extracted from high-resolution body surface potential maps of patients who were diagnosed with myocardial infarction, diagnosed with left ventricular hypertrophy, or a normal ECG.
RESULTS
RESULTS
DL achieved the highest accuracy in this study for detecting V1 and V2 electrode misplacement, with an accuracy of 93.0% (95% CI 91.46-94.53) for misplacement in the second intercostal space. The performance of DL in the second intercostal space was benchmarked with physicians (n=11 and age 47.3 years, SD 15.5) who were experienced in reading ECGs (mean number of ECGs read in the past year 436.54, SD 397.9). Physicians were poor at recognizing chest electrode misplacement on the ECG and achieved a mean accuracy of 60% (95% CI 56.09-63.90), which was significantly poorer than that of DL (P<.001).
CONCLUSIONS
CONCLUSIONS
DL provides the best performance for detecting chest electrode misplacement when compared with the ability of experienced physicians. DL and ML could be used to help flag ECGs that have been incorrectly recorded and flag that the data may be flawed, which could reduce the number of erroneous diagnoses.
Identifiants
pubmed: 33861205
pii: v9i4e25347
doi: 10.2196/25347
pmc: PMC8087970
doi:
Types de publication
Journal Article
Langues
eng
Pagination
e25347Informations de copyright
©Khaled Rjoob, Raymond Bond, Dewar Finlay, Victoria McGilligan, Stephen J Leslie, Ali Rababah, Aleeha Iftikhar, Daniel Guldenring, Charles Knoery, Anne McShane, Aaron Peace. Originally published in JMIR Medical Informatics (http://medinform.jmir.org), 16.04.2021.
Références
Circulation. 2000 Jun 13;101(23):E215-20
pubmed: 10851218
Am J Crit Care. 1997 Nov;6(6):457-62
pubmed: 9354224
J Electrocardiol. 1996;29 Suppl:5-9
pubmed: 9238370
Clin Cardiol. 1990 Sep;13(9):668-9
pubmed: 2208827
IEEE J Biomed Health Inform. 2018 May;22(3):642-652
pubmed: 28715343
Clin Cardiol. 1991 Jun;14(6):469-76
pubmed: 1810683
IEEE J Biomed Health Inform. 2017 Jul;21(4):1133-1145
pubmed: 27254875
Physiol Meas. 2012 Sep;33(9):1549-61
pubmed: 22903067
Med Biol Eng Comput. 2014 Feb;52(2):109-19
pubmed: 24142562
J Electrocardiol. 2016 Nov - Dec;49(6):911-918
pubmed: 27662775
IEEE J Biomed Health Inform. 2017 Nov;21(6):1719-1729
pubmed: 28287993
Int J Clin Pract. 2008 Jan;62(1):65-70
pubmed: 17764456
Eur J Intern Med. 2012 Oct;23(7):610-5
pubmed: 22939805
J Am Coll Cardiol. 2017 May 30;69(21):2657-2664
pubmed: 28545640
Circulation. 1991 Dec;84(6):2442-53
pubmed: 1835677
J Emerg Med. 2009 Jul;37(1):79-81; author reply 81-2
pubmed: 19321285
J Electrocardiol. 1996 Jul;29(3):179-84
pubmed: 8854328
J Electrocardiol. 2008 May-Jun;41(3):251-6
pubmed: 18433616
Proc IEEE Inst Electr Electron Eng. 2013 Dec 1;101(12):2470-2494
pubmed: 24431472
Postgrad Med J. 2005 Feb;81(952):122-5
pubmed: 15701746
Circulation. 1981 May;63(5):1166-72
pubmed: 7471378
Am J Emerg Med. 2003 Nov;21(7):574-7
pubmed: 14655240
BMJ. 2015 Oct 20;351:h5153
pubmed: 26487159