Precision measurement of cardiac structure and function in cardiovascular magnetic resonance using machine learning.
Cardiac magnetic resonance
Cardiovascular imaging
Image processing
Machine learning
Ventricular function
Journal
Journal of cardiovascular magnetic resonance : official journal of the Society for Cardiovascular Magnetic Resonance
ISSN: 1532-429X
Titre abrégé: J Cardiovasc Magn Reson
Pays: England
ID NLM: 9815616
Informations de publication
Date de publication:
10 03 2022
10 03 2022
Historique:
received:
11
10
2021
accepted:
03
02
2022
entrez:
11
3
2022
pubmed:
12
3
2022
medline:
7
5
2022
Statut:
epublish
Résumé
Measurement of cardiac structure and function from images (e.g. volumes, mass and derived parameters such as left ventricular (LV) ejection fraction [LVEF]) guides care for millions. This is best assessed using cardiovascular magnetic resonance (CMR), but image analysis is currently performed by individual clinicians, which introduces error. We sought to develop a machine learning algorithm for volumetric analysis of CMR images with demonstrably better precision than human analysis. A fully automated machine learning algorithm was trained on 1923 scans (10 scanner models, 13 institutions, 9 clinical conditions, 60,000 contours) and used to segment the LV blood volume and myocardium. Performance was quantified by measuring precision on an independent multi-site validation dataset with multiple pathologies with n = 109 patients, scanned twice. This dataset was augmented with a further 1277 patients scanned as part of routine clinical care to allow qualitative assessment of generalization ability by identifying mis-segmentations. Machine learning algorithm ('machine') performance was compared to three clinicians ('human') and a commercial tool (cvi42, Circle Cardiovascular Imaging). Machine analysis was quicker (20 s per patient) than human (13 min). Overall machine mis-segmentation rate was 1 in 479 images for the combined dataset, occurring mostly in rare pathologies not encountered in training. Without correcting these mis-segmentations, machine analysis had superior precision to three clinicians (e.g. scan-rescan coefficients of variation of human vs machine: LVEF 6.0% vs 4.2%, LV mass 4.8% vs. 3.6%; both P < 0.05), translating to a 46% reduction in required trial sample size using an LVEF endpoint. We present a fully automated algorithm for measuring LV structure and global systolic function that betters human performance for speed and precision.
Sections du résumé
BACKGROUND
Measurement of cardiac structure and function from images (e.g. volumes, mass and derived parameters such as left ventricular (LV) ejection fraction [LVEF]) guides care for millions. This is best assessed using cardiovascular magnetic resonance (CMR), but image analysis is currently performed by individual clinicians, which introduces error. We sought to develop a machine learning algorithm for volumetric analysis of CMR images with demonstrably better precision than human analysis.
METHODS
A fully automated machine learning algorithm was trained on 1923 scans (10 scanner models, 13 institutions, 9 clinical conditions, 60,000 contours) and used to segment the LV blood volume and myocardium. Performance was quantified by measuring precision on an independent multi-site validation dataset with multiple pathologies with n = 109 patients, scanned twice. This dataset was augmented with a further 1277 patients scanned as part of routine clinical care to allow qualitative assessment of generalization ability by identifying mis-segmentations. Machine learning algorithm ('machine') performance was compared to three clinicians ('human') and a commercial tool (cvi42, Circle Cardiovascular Imaging).
FINDINGS
Machine analysis was quicker (20 s per patient) than human (13 min). Overall machine mis-segmentation rate was 1 in 479 images for the combined dataset, occurring mostly in rare pathologies not encountered in training. Without correcting these mis-segmentations, machine analysis had superior precision to three clinicians (e.g. scan-rescan coefficients of variation of human vs machine: LVEF 6.0% vs 4.2%, LV mass 4.8% vs. 3.6%; both P < 0.05), translating to a 46% reduction in required trial sample size using an LVEF endpoint.
CONCLUSION
We present a fully automated algorithm for measuring LV structure and global systolic function that betters human performance for speed and precision.
Identifiants
pubmed: 35272664
doi: 10.1186/s12968-022-00846-4
pii: 10.1186/s12968-022-00846-4
pmc: PMC8908603
doi:
Types de publication
Journal Article
Research Support, N.I.H., Intramural
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
16Subventions
Organisme : British Heart Foundation
ID : FS/19/35/34374
Pays : United Kingdom
Organisme : British Heart Foundation
ID : SP/20/2/34841
Pays : United Kingdom
Organisme : British Heart Foundation
ID : AA/18/6/34223
Pays : United Kingdom
Organisme : Department of Health
Pays : United Kingdom
Organisme : British Heart Foundation
ID : FS/18/21/33447
Pays : United Kingdom
Organisme : Intramural NIH HHS
ID : ZIA HL006242
Pays : United States
Organisme : British Heart Foundation
ID : FS/17/82/33222
Pays : United Kingdom
Informations de copyright
© 2022. The Author(s).
Références
J Cardiovasc Magn Reson. 2020 Dec 14;22(1):87
pubmed: 33308262
J Cardiovasc Magn Reson. 2015 Jul 28;17:63
pubmed: 26215273
J Cardiovasc Magn Reson. 2013 May 01;15:35
pubmed: 23634753
Lancet Digit Health. 2019 Oct;1(6):e271-e297
pubmed: 33323251
JACC Cardiovasc Imaging. 2020 Jan;13(1 Pt 1):44-54
pubmed: 31103587
J Cardiovasc Magn Reson. 2017 Feb 3;19(1):18
pubmed: 28178995
Eur Heart J. 2016 Sep 21;37(36):2768-2801
pubmed: 27567406
Stat Methods Med Res. 2015 Feb;24(1):68-106
pubmed: 24919829
J Am Coll Cardiol. 2013 Dec 17;62(24):2284-93
pubmed: 23994420
Nature. 2020 Jan;577(7788):89-94
pubmed: 31894144
Eur Heart J. 2010 Nov;31(21):2677-87
pubmed: 20801924
Circ Cardiovasc Imaging. 2019 Oct;12(10):e009214
pubmed: 31547689
Radiology. 2019 Jan;290(1):81-88
pubmed: 30299231
J Am Coll Cardiol. 1999 Jun;33(7):1948-55
pubmed: 10362198
Front Cardiovasc Med. 2020 Mar 05;7:25
pubmed: 32195270
Lancet. 1986 Feb 8;1(8476):307-10
pubmed: 2868172
Nature. 2015 May 28;521(7553):436-44
pubmed: 26017442
Eur Heart J. 2016 Jul 14;37(27):2129-2200
pubmed: 27206819
Radiology. 2000 Aug;216(2):597-602
pubmed: 10924592
J Cardiovasc Magn Reson. 2013 Oct 08;15:91
pubmed: 24103764
IEEE Trans Pattern Anal Mach Intell. 2017 Apr;39(4):640-651
pubmed: 27244717
Circulation. 2017 Aug 8;136(6):e137-e161
pubmed: 28455343
Am J Cardiol. 2002 Jul 1;90(1):29-34
pubmed: 12088775
Circulation. 1962 Apr;25:674-85
pubmed: 13893859
PLoS Med. 2018 Nov 6;15(11):e1002689
pubmed: 30399149
Biometrics. 1989 Mar;45(1):255-68
pubmed: 2720055
J Cardiovasc Magn Reson. 2018 Sep 14;20(1):65
pubmed: 30217194