Machine Learning for Automatic Paraspinous Muscle Area and Attenuation Measures on Low-Dose Chest CT Scans.
Chest CT
Machine learning
Muscle
Myosteatosis
Sarcopenia
Journal
Academic radiology
ISSN: 1878-4046
Titre abrégé: Acad Radiol
Pays: United States
ID NLM: 9440159
Informations de publication
Date de publication:
12 2019
12 2019
Historique:
received:
23
04
2019
revised:
21
06
2019
accepted:
26
06
2019
pubmed:
22
7
2019
medline:
23
6
2020
entrez:
22
7
2019
Statut:
ppublish
Résumé
To develop and evaluate an automated machine learning (ML) algorithm for segmenting the paraspinous muscles on chest computed tomography (CT) scans to evaluate for presence of sarcopenia. A convolutional neural network based on the U-Net architecture was trained to perform muscle segmentation on a dataset of 1875 single slice CT images and was tested on 209 CT images of participants in the National Lung Screening Trial. Low-dose, noncontrast CT examinations were obtained at 33 clinical sites, using scanners from four manufacturers. The study participants had a mean age of 71.6 years (range, 70-74 years). Ground truth was obtained by manually segmenting the left paraspinous muscle at the level of the T12 vertebra. Muscle cross-sectional area (CSA) and muscle attenuation (MA) were recorded. Comparison between the ML algorithm and ground truth measures of muscle CSA and MA were obtained using Dice similarity coefficients and Pearson correlations. Compared to ground truth segmentation, the ML algorithm achieved median (standard deviation) Dice scores of 0.94 (0.04) in the test set. Mean (SD) muscle CSA was 14.3 (3.6) cm The ML algorithm for measurement of paraspinous muscles compared favorably to manual ground truth measurements in the NLST. The algorithm generalized well to a heterogeneous set of low-dose CT images and may be capable of automated quantification of muscle metrics to screen for sarcopenia on routine chest CT examinations.
Identifiants
pubmed: 31326311
pii: S1076-6332(19)30323-X
doi: 10.1016/j.acra.2019.06.017
pmc: PMC6878160
mid: NIHMS1535302
pii:
doi:
Types de publication
Journal Article
Research Support, N.I.H., Extramural
Langues
eng
Sous-ensembles de citation
IM
Pagination
1686-1694Subventions
Organisme : NIA NIH HHS
ID : K25 AG058804
Pays : United States
Organisme : NIA NIH HHS
ID : P30 AG021332
Pays : United States
Organisme : NCATS NIH HHS
ID : UL1 TR001420
Pays : United States
Informations de copyright
Copyright © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
Références
J Nutr Health Aging. 2018;22(10):1148-1161
pubmed: 30498820
Med Image Anal. 2016 May;30:108-119
pubmed: 26917105
N Engl J Med. 2011 Aug 4;365(5):395-409
pubmed: 21714641
Comput Biol Med. 2017 Aug 1;87:95-103
pubmed: 28558319
Gynecol Oncol. 2016 Aug;142(2):311-6
pubmed: 27235857
AJR Am J Roentgenol. 2017 Jun;208(6):W208-W215
pubmed: 28267356
Orthopedics. 2016 Nov 1;39(6):e1159-e1164
pubmed: 27536954
AJR Am J Roentgenol. 2016 Jul;207(1):67-79
pubmed: 27101207
J Surg Res. 2018 Jan;221:161-166
pubmed: 29229123
Adv Exp Med Biol. 2017;987:213-218
pubmed: 28971460
J Orthop Translat. 2018 Oct 28;15:91-103
pubmed: 30533385
Support Care Cancer. 2016 Nov;24(11):4721-6
pubmed: 27364150
Med Phys. 2019 Jan;46(1):e1-e36
pubmed: 30367497
PeerJ. 2014 Jun 19;2:e453
pubmed: 25024921
Bioinformatics. 2012 Oct 1;28(19):2520-2
pubmed: 22908215
Calcif Tissue Int. 2017 Sep;101(3):229-247
pubmed: 28421264
IEEE Trans Med Imaging. 2016 May;35(5):1252-1261
pubmed: 27046893
JAMA Surg. 2017 Feb 15;152(2):e164604
pubmed: 28030710
Age Ageing. 2019 Jul 1;48(4):601
pubmed: 31081853
Radiology. 2011 Jan;258(1):243-53
pubmed: 21045183
Semin Musculoskelet Radiol. 2018 Jul;22(3):307-322
pubmed: 29791959
Phys Med Biol. 2018 Jun 27;63(13):135009
pubmed: 29851653
J Orthop Trauma. 2018 Jun;32(6):278-282
pubmed: 29533306
Comput Methods Programs Biomed. 2018 Mar;156:85-95
pubmed: 29428079
J Gerontol A Biol Sci Med Sci. 2019 Jun 18;74(7):1063-1069
pubmed: 30124775
Clin Nutr. 2019 Aug;38(4):1489-1495
pubmed: 30316536
Eur J Surg Oncol. 2019 May;45(5):728-735
pubmed: 30348603
J Am Med Dir Assoc. 2016 Dec 1;17(12):1164.e7-1164.e15
pubmed: 27816484
J Clin Med. 2018 Apr 08;7(4):null
pubmed: 29642478
Interact Cardiovasc Thorac Surg. 2019 Apr 1;28(4):542-549
pubmed: 30307497
Support Care Cancer. 2018 Jul;26(7):2353-2359
pubmed: 29417293
Int J Comput Assist Radiol Surg. 2018 Jul;13(7):977-986
pubmed: 29626280
Exp Gerontol. 2017 Oct 1;96:100-103
pubmed: 28647519
Neuroimage. 2012 Aug 15;62(2):782-90
pubmed: 21979382
Eur J Radiol. 2018 Dec;109:142-146
pubmed: 30527297
J Digit Imaging. 2018 Apr;31(2):262-274
pubmed: 29067570
World J Surg. 2015 Feb;39(2):373-9
pubmed: 25249011
J Digit Imaging. 2017 Aug;30(4):487-498
pubmed: 28653123
J Gerontol A Biol Sci Med Sci. 2016 Jun;71(6):811-6
pubmed: 26503375
IEEE Trans Med Imaging. 2016 Feb;35(2):512-20
pubmed: 26415164
Osteoporos Int. 2018 Jun;29(6):1253-1262
pubmed: 29500527
Aging Dis. 2018 Feb 1;9(1):151-164
pubmed: 29392090
Acad Radiol. 2017 Sep;24(9):1154-1161
pubmed: 28365235
Sci Rep. 2017 Sep 5;7(1):10425
pubmed: 28874743
PLoS One. 2017 Jan 17;12(1):e0169548
pubmed: 28095426
Curr Osteoporos Rep. 2018 Dec;16(6):730-737
pubmed: 30284195
Int J Surg. 2018 Nov;59:19-26
pubmed: 30266663
Radiology. 2019 Mar;290(3):669-679
pubmed: 30526356
Eur J Clin Nutr. 2014 Sep;68(9):1001-7
pubmed: 24961545
Crit Care. 2013 Sep 19;17(5):R206
pubmed: 24050662
Neuroimage. 2011 Feb 1;54(3):2033-44
pubmed: 20851191
Comput Methods Clin Appl Musculoskelet Imaging (2017). 2018;10734:25-35
pubmed: 30335867
J Diabetes Complications. 2018 Jun;32(6):558-564
pubmed: 29627372
Phys Med Biol. 2016 Sep 7;61(17):6553-69
pubmed: 27530679
AJR Am J Roentgenol. 2018 Feb;210(2):364-368
pubmed: 29220208
J Crit Care. 2018 Apr;44:117-123
pubmed: 29096229
J Gerontol A Biol Sci Med Sci. 2019 Sep 15;74(10):1671-1678
pubmed: 30726878
J Surg Oncol. 2015 Oct;112(5):503-9
pubmed: 26310812
Vasc Med. 2016 Jun;21(3):217-22
pubmed: 26850115
Eur J Cancer. 2016 Apr;57:58-67
pubmed: 26882087
Proc Nutr Soc. 2018 May;77(2):135-151
pubmed: 29745361
Maturitas. 2019 Jan;119:61-69
pubmed: 30502752