Artificial intelligence and optical coherence tomography for the automatic characterisation of human atherosclerotic plaques.
Journal
EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology
ISSN: 1969-6213
Titre abrégé: EuroIntervention
Pays: France
ID NLM: 101251040
Informations de publication
Date de publication:
17 May 2021
17 May 2021
Historique:
pubmed:
3
2
2021
medline:
19
5
2021
entrez:
2
2
2021
Statut:
ppublish
Résumé
Intravascular optical coherence tomography (IVOCT) enables detailed plaque characterisation in vivo, but visual assessment is time-consuming and subjective. This study aimed to develop and validate an automatic framework for IVOCT plaque characterisation using artificial intelligence (AI). IVOCT pullbacks from five international centres were analysed in a core lab, annotating basic plaque components, inflammatory markers and other structures. A deep convolutional network with encoding-decoding architecture and pseudo-3D input was developed and trained using hybrid loss. The proposed network was integrated into commercial software to be externally validated on additional IVOCT pullbacks from three international core labs, taking the consensus among core labs as reference. Annotated images from 509 pullbacks (391 patients) were divided into 10,517 and 1,156 cross-sections for the training and testing data sets, respectively. The Dice coefficient of the model was 0.906 for fibrous plaque, 0.848 for calcium and 0.772 for lipid in the testing data set. Excellent agreement in plaque burden quantification was observed between the model and manual measurements (R2=0.98). In the external validation, the software correctly identified 518 out of 598 plaque regions from 300 IVOCT cross-sections, with a diagnostic accuracy of 97.6% (95% CI: 93.4-99.3%) in fibrous plaque, 90.5% (95% CI: 85.2-94.1%) in lipid and 88.5% (95% CI: 82.4-92.7%) in calcium. The median time required for analysis was 21.4 (18.6-25.0) seconds per pullback. A novel AI framework for automatic plaque characterisation in IVOCT was developed, providing excellent diagnostic accuracy in both internal and external validation. This model might reduce subjectivity in image interpretation and facilitate IVOCT quantification of plaque composition, with potential applications in research and IVOCT-guided PCI.
Sections du résumé
BACKGROUND
BACKGROUND
Intravascular optical coherence tomography (IVOCT) enables detailed plaque characterisation in vivo, but visual assessment is time-consuming and subjective.
AIMS
OBJECTIVE
This study aimed to develop and validate an automatic framework for IVOCT plaque characterisation using artificial intelligence (AI).
METHODS
METHODS
IVOCT pullbacks from five international centres were analysed in a core lab, annotating basic plaque components, inflammatory markers and other structures. A deep convolutional network with encoding-decoding architecture and pseudo-3D input was developed and trained using hybrid loss. The proposed network was integrated into commercial software to be externally validated on additional IVOCT pullbacks from three international core labs, taking the consensus among core labs as reference.
RESULTS
RESULTS
Annotated images from 509 pullbacks (391 patients) were divided into 10,517 and 1,156 cross-sections for the training and testing data sets, respectively. The Dice coefficient of the model was 0.906 for fibrous plaque, 0.848 for calcium and 0.772 for lipid in the testing data set. Excellent agreement in plaque burden quantification was observed between the model and manual measurements (R2=0.98). In the external validation, the software correctly identified 518 out of 598 plaque regions from 300 IVOCT cross-sections, with a diagnostic accuracy of 97.6% (95% CI: 93.4-99.3%) in fibrous plaque, 90.5% (95% CI: 85.2-94.1%) in lipid and 88.5% (95% CI: 82.4-92.7%) in calcium. The median time required for analysis was 21.4 (18.6-25.0) seconds per pullback.
CONCLUSIONS
CONCLUSIONS
A novel AI framework for automatic plaque characterisation in IVOCT was developed, providing excellent diagnostic accuracy in both internal and external validation. This model might reduce subjectivity in image interpretation and facilitate IVOCT quantification of plaque composition, with potential applications in research and IVOCT-guided PCI.
Identifiants
pubmed: 33528359
pii: EIJ-D-20-01355
doi: 10.4244/EIJ-D-20-01355
pmc: PMC9724931
pii:
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
41-50Références
JACC Cardiovasc Interv. 2015 Jun;8(7):937-45
pubmed: 26088511
J Cardiovasc Comput Tomogr. 2019 Sep - Oct;13(5):274-280
pubmed: 31029649
IEEE Trans Med Imaging. 2015 Sep;34(9):1867-78
pubmed: 25794387
Eur Heart J. 2015 Dec 14;36(47):3346-55
pubmed: 26242713
Cardiol J. 2020;27(4):350-361
pubmed: 32436590
Eur Heart J. 2020 Jan 14;41(3):383-391
pubmed: 31504405
Rev Cardiovasc Med. 2019 Sep 30;20(3):171-177
pubmed: 31601091
N Engl J Med. 2011 Jan 20;364(3):226-35
pubmed: 21247313
J Am Coll Cardiol. 1998 Jan;31(1):126-33
pubmed: 9426030
Circulation. 2002 Sep 24;106(13):1640-5
pubmed: 12270856
Lancet. 2016 Nov 26;388(10060):2618-2628
pubmed: 27806900
IEEE Trans Med Imaging. 2019 Feb;38(2):426-434
pubmed: 30130180
EuroIntervention. 2020 Sep 18;16(7):568-576
pubmed: 31951207
Eur Heart J Cardiovasc Imaging. 2012 May;13(5):370-84
pubmed: 22330231
Cardiol J. 2018;25(4):443-458
pubmed: 29774520
EuroIntervention. 2019 Jun 20;15(2):189-197
pubmed: 31147309
Biomed Opt Express. 2017 Jan 30;8(2):1203-1220
pubmed: 28271012
EuroIntervention. 2007 Nov;3(3):365-70
pubmed: 19737719