Determining anatomical and electrophysiological detail requirements for computational ventricular models of porcine myocardial infarction.
Cardiac electrophysiology
Myocardial infarction
Patient-specific models
Ventricular tachycardia
Journal
Computers in biology and medicine
ISSN: 1879-0534
Titre abrégé: Comput Biol Med
Pays: United States
ID NLM: 1250250
Informations de publication
Date de publication:
02 2022
02 2022
Historique:
received:
16
08
2021
revised:
04
11
2021
accepted:
20
11
2021
pubmed:
17
12
2021
medline:
18
3
2022
entrez:
16
12
2021
Statut:
ppublish
Résumé
Computational models of the heart built from cardiac MRI and electrophysiology (EP) data have shown promise for predicting the risk of and ablation targets for myocardial infarction (MI) related ventricular tachycardia (VT), as well as to predict paced activation sequences in heart failure patients. However, most recent studies have relied on low resolution imaging data and little or no EP personalisation, which may affect the accuracy of model-based predictions. To investigate the impact of model anatomy, MI scar morphology, and EP personalisation strategies on paced activation sequences and VT inducibility to determine the level of detail required to make accurate model-based predictions. Imaging and EP data were acquired from a cohort of six pigs with experimentally induced MI. Computational models of ventricular anatomy, incorporating MI scar, were constructed including bi-ventricular or left ventricular (LV) only anatomy, and MI scar morphology with varying detail. Tissue conductivities and action potential duration (APD) were fitted to 12-lead ECG data using the QRS duration and the QT interval, respectively, in addition to corresponding literature parameters. Paced activation sequences and VT induction were simulated. Simulated paced activation and VT inducibility were compared between models and against experimental data. Simulations predict that the level of model anatomical detail has little effect on simulated paced activation, with all model predictions comparing closely with invasive EP measurements. However, detailed scar morphology from high-resolution images, bi-ventricular anatomy, and personalized tissue conductivities are required to predict experimental VT outcome. This study provides clear guidance for model generation based on clinical data. While a representing high level of anatomical and scar detail will require high-resolution image acquisition, EP personalisation based on 12-lead ECG can be readily incorporated into modelling pipelines, as such data is widely available.
Sections du résumé
BACKGROUND
Computational models of the heart built from cardiac MRI and electrophysiology (EP) data have shown promise for predicting the risk of and ablation targets for myocardial infarction (MI) related ventricular tachycardia (VT), as well as to predict paced activation sequences in heart failure patients. However, most recent studies have relied on low resolution imaging data and little or no EP personalisation, which may affect the accuracy of model-based predictions.
OBJECTIVE
To investigate the impact of model anatomy, MI scar morphology, and EP personalisation strategies on paced activation sequences and VT inducibility to determine the level of detail required to make accurate model-based predictions.
METHODS
Imaging and EP data were acquired from a cohort of six pigs with experimentally induced MI. Computational models of ventricular anatomy, incorporating MI scar, were constructed including bi-ventricular or left ventricular (LV) only anatomy, and MI scar morphology with varying detail. Tissue conductivities and action potential duration (APD) were fitted to 12-lead ECG data using the QRS duration and the QT interval, respectively, in addition to corresponding literature parameters. Paced activation sequences and VT induction were simulated. Simulated paced activation and VT inducibility were compared between models and against experimental data.
RESULTS
Simulations predict that the level of model anatomical detail has little effect on simulated paced activation, with all model predictions comparing closely with invasive EP measurements. However, detailed scar morphology from high-resolution images, bi-ventricular anatomy, and personalized tissue conductivities are required to predict experimental VT outcome.
CONCLUSION
This study provides clear guidance for model generation based on clinical data. While a representing high level of anatomical and scar detail will require high-resolution image acquisition, EP personalisation based on 12-lead ECG can be readily incorporated into modelling pipelines, as such data is widely available.
Identifiants
pubmed: 34915331
pii: S0010-4825(21)00855-6
doi: 10.1016/j.compbiomed.2021.105061
pmc: PMC8819160
pii:
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
105061Subventions
Organisme : Medical Research Council
ID : MR/N001877/1
Pays : United Kingdom
Organisme : Medical Research Council
ID : MR/N011007/1
Pays : United Kingdom
Informations de copyright
Copyright © 2021 The Authors. Published by Elsevier Ltd.. All rights reserved.
Références
Europace. 2021 Mar 4;23(23 Suppl 1):i12-i20
pubmed: 33437987
Heart Rhythm. 2020 Nov;17(11):1922-1929
pubmed: 32603781
Comput Cardiol (2010). 2013 Oct 1;40:373-376
pubmed: 24729986
JACC Clin Electrophysiol. 2021 Feb;7(2):238-249
pubmed: 33602406
Prog Biophys Mol Biol. 2016 Jan;120(1-3):222-35
pubmed: 26790342
Front Physiol. 2015 Oct 13;6:282
pubmed: 26528188
Heart Rhythm. 2007 Nov;4(11):1430-6
pubmed: 17954403
Heart Rhythm. 2019 Oct;16(10):1475-1483
pubmed: 30930329
Nat Commun. 2016 May 10;7:11437
pubmed: 27164184
Am J Physiol Heart Circ Physiol. 2006 Sep;291(3):H1088-100
pubmed: 16565318
SoftwareX. 2020 Jan-Jun;11:100454
pubmed: 32607406
IEEE Trans Biomed Eng. 2015 Sep;62(9):2251-2259
pubmed: 25872206
Physiol Rev. 2004 Apr;84(2):431-88
pubmed: 15044680
J Electrocardiol. 2003;36 Suppl:69-74
pubmed: 14716595
J Cardiovasc Magn Reson. 2019 Oct 10;21(1):62
pubmed: 31597563
Ann Biomed Eng. 2016 Jan;44(1):58-70
pubmed: 26424476
Circulation. 1970 Jun;41(6):899-912
pubmed: 5482907
J Magn Reson Imaging. 2018 Mar 9;:
pubmed: 29522262
Heart Rhythm. 2020 Apr;17(4):576-583
pubmed: 31751771
Circ Arrhythm Electrophysiol. 2009 Aug;2(4):433-40
pubmed: 19808500
Heart Rhythm. 2020 Aug;17(8):1262-1270
pubmed: 32272230
Heart. 2006 Apr;92 Suppl 1:i2-13
pubmed: 16543598
Ann Biomed Eng. 2012 Oct;40(10):2243-54
pubmed: 22648575
Biophys J. 2019 Dec 17;117(12):2361-2374
pubmed: 31521328
Circulation. 1963 Sep;28:333-8
pubmed: 14059452
J Comput Phys. 2017 Oct 1;346:191-211
pubmed: 28819329
Proc SPIE Int Soc Opt Eng. 2015 Feb 21;9413:
pubmed: 26633913
Med Image Anal. 2021 Jul;71:102080
pubmed: 33975097
Front Physiol. 2021 Feb 26;12:646023
pubmed: 33716795
PLoS One. 2017 Sep 1;12(9):e0184352
pubmed: 28863167
Front Physiol. 2018 Apr 09;9:356
pubmed: 29686626
Ann Biomed Eng. 2021 Dec;49(12):3143-3153
pubmed: 34431016
Europace. 2018 Nov 01;20(suppl_3):iii94-iii101
pubmed: 30476056
Med Image Anal. 2018 Apr;45:83-93
pubmed: 29414438
Med Image Anal. 2012 Jan;16(1):201-15
pubmed: 21920797
Med Image Anal. 2019 Oct;57:197-213
pubmed: 31326854
Nat Biomed Eng. 2018 Oct;2(10):732-740
pubmed: 30847259
Nat Rev Cardiol. 2019 Feb;16(2):100-111
pubmed: 30361497
Comput Biol Med. 2020 Aug;123:103895
pubmed: 32741753
Circ Res. 1972 Feb;30(2):217-43
pubmed: 5061320