Machine learning based quantification of ejection and filling parameters by fully automated dynamic measurement of left ventricular volumes from cardiac magnetic resonance images.
Adult
Aged
Algorithms
Automation
Endocardium
/ diagnostic imaging
Female
Heart
/ diagnostic imaging
Heart Ventricles
/ diagnostic imaging
Humans
Image Processing, Computer-Assisted
/ methods
Machine Learning
Magnetic Resonance Imaging
Male
Middle Aged
Prospective Studies
Reproducibility of Results
Software
Ventricular Function, Left
Artificial intelligence
Left ventricle
Time-volume curves
Journal
Magnetic resonance imaging
ISSN: 1873-5894
Titre abrégé: Magn Reson Imaging
Pays: Netherlands
ID NLM: 8214883
Informations de publication
Date de publication:
04 2020
04 2020
Historique:
received:
28
08
2019
revised:
13
11
2019
accepted:
07
12
2019
pubmed:
16
12
2019
medline:
3
11
2020
entrez:
16
12
2019
Statut:
ppublish
Résumé
Although analysis of cardiac magnetic resonance (CMR) images provides accurate and reproducible measurements of left ventricular (LV) volumes, these measurements are usually not performed throughout the cardiac cycle because of lack of tools that would allow such analysis within a reasonable timeframe. A fully-automated machine-learning (ML) algorithm was recently developed to automatically generate LV volume-time curves. Our aim was to validate ejection and filling parameters calculated from these curves using conventional analysis as a reference. We studied 21 patients undergoing clinical CMR examinations. LV volume-time curves were obtained using the ML-based algorithm (Neosoft), and independently using slice-by-slice, frame-by-frame manual tracing of the endocardial boundaries. Ejection and filling parameters derived from these curves were compared between the two techniques. For each parameter, Bland-Altman bias and limits of agreement (LOA) were expressed in percent of the mean measured value. Time-volume curves were generated using the automated ML analysis within 2.5 ± 0.5 min, considerably faster than the manual analysis (43 ± 14 min per patient, including ~10 slices with 25-32 frames per slice). Time-volume curves were similar between the two techniques in magnitude and shape. Size and function parameters extracted from these curves showed no significant inter-technique differences, reflected by high correlations, small biases (<10%) and mostly reasonably narrow LOA. ML software for dynamic LV volume measurement allows fast and accurate, fully automated analysis of ejection and filling parameters, compared to manual tracing based analysis. The ability to quickly evaluate time-volume curves is important for a more comprehensive evaluation of the patient's cardiac function.
Sections du résumé
BACKGROUND
Although analysis of cardiac magnetic resonance (CMR) images provides accurate and reproducible measurements of left ventricular (LV) volumes, these measurements are usually not performed throughout the cardiac cycle because of lack of tools that would allow such analysis within a reasonable timeframe. A fully-automated machine-learning (ML) algorithm was recently developed to automatically generate LV volume-time curves. Our aim was to validate ejection and filling parameters calculated from these curves using conventional analysis as a reference.
METHODS
We studied 21 patients undergoing clinical CMR examinations. LV volume-time curves were obtained using the ML-based algorithm (Neosoft), and independently using slice-by-slice, frame-by-frame manual tracing of the endocardial boundaries. Ejection and filling parameters derived from these curves were compared between the two techniques. For each parameter, Bland-Altman bias and limits of agreement (LOA) were expressed in percent of the mean measured value.
RESULTS
Time-volume curves were generated using the automated ML analysis within 2.5 ± 0.5 min, considerably faster than the manual analysis (43 ± 14 min per patient, including ~10 slices with 25-32 frames per slice). Time-volume curves were similar between the two techniques in magnitude and shape. Size and function parameters extracted from these curves showed no significant inter-technique differences, reflected by high correlations, small biases (<10%) and mostly reasonably narrow LOA.
CONCLUSION
ML software for dynamic LV volume measurement allows fast and accurate, fully automated analysis of ejection and filling parameters, compared to manual tracing based analysis. The ability to quickly evaluate time-volume curves is important for a more comprehensive evaluation of the patient's cardiac function.
Identifiants
pubmed: 31838116
pii: S0730-725X(19)30523-5
doi: 10.1016/j.mri.2019.12.004
pmc: PMC7135920
mid: NIHMS1565659
pii:
doi:
Types de publication
Journal Article
Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
28-32Subventions
Organisme : NHLBI NIH HHS
ID : T32 HL007381
Pays : United States
Informations de copyright
Copyright © 2019 Elsevier Inc. All rights reserved.
Déclaration de conflit d'intérêts
Declaration of competing interest ARP has received research support from Philips Healthcare and Neosoft. Akhil Narang was funded by a T32 Cardiovascular Sciences Training Grant (5T32HL7381) from the National Institutes of Health (USA).
Références
J Cardiol. 2015 Jul;66(1):50-6
pubmed: 25242299
Med Image Comput Comput Assist Interv. 2012;15(Pt 2):535-43
pubmed: 23286090
Eur Heart J Cardiovasc Imaging. 2018 Jul 1;19(7):730-738
pubmed: 29538684
Echocardiography. 2011 Jul;28(6):676-83
pubmed: 21718353
Echocardiography. 2008 Oct;25(9):968-73
pubmed: 18771555
IEEE Trans Med Imaging. 2017 Nov;36(11):2228-2238
pubmed: 28708550
Int J Cardiovasc Imaging. 2013 Feb;29(2):285-93
pubmed: 22752362
J Am Soc Echocardiogr. 2003 Jan;16(1):29-37
pubmed: 12514632
Med Image Anal. 2017 Jul;39:78-86
pubmed: 28437634
Conf Proc IEEE Eng Med Biol Soc. 2015 Aug;2015:683-6
pubmed: 26736354
J Cardiovasc Magn Reson. 2018 Sep 14;20(1):65
pubmed: 30217194
Eur Heart J Cardiovasc Imaging. 2019 May 1;20(5):541-549
pubmed: 30304500
Med Image Anal. 2016 May;30:108-119
pubmed: 26917105
Conf Proc IEEE Eng Med Biol Soc. 2015;2015:7292-5
pubmed: 26737975
Eur J Echocardiogr. 2001 Mar;2(1):31-9
pubmed: 11882423
J Cardiovasc Magn Reson. 2019 Apr 25;21(1):24
pubmed: 31023305
Med Phys. 2018 Apr;45(4):1471-1480
pubmed: 29389013
Eur J Radiol. 2017 Jul;92:78-83
pubmed: 28624024
Med Image Anal. 2017 Jan;35:159-171
pubmed: 27423113
Med Image Anal. 2012 Jul;16(5):933-51
pubmed: 22465077
Eur J Radiol. 2018 May;102:61-67
pubmed: 29685546