A Combined Radiomics and Machine Learning Approach to Overcome the Clinicoradiologic Paradox in Multiple Sclerosis.
Journal
AJNR. American journal of neuroradiology
ISSN: 1936-959X
Titre abrégé: AJNR Am J Neuroradiol
Pays: United States
ID NLM: 8003708
Informations de publication
Date de publication:
11 2021
11 2021
Historique:
received:
07
11
2020
accepted:
12
07
2021
pubmed:
18
9
2021
medline:
25
11
2021
entrez:
17
9
2021
Statut:
ppublish
Résumé
Conventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images. In this retrospective cross-sectional study, 3T brain MR imaging studies of patients with multiple sclerosis, including 3D T1-weighted and T2-weighted FLAIR sequences, were selected from 2 institutions. T1-weighted images were processed to obtain volume, connectivity score (inferred from the T2 lesion location), and texture features for an atlas-based set of GM regions. The site 1 cohort was randomly split into training ( The selection procedure identified the 9 most informative variables, including age and secondary-progressive course and a subset of radiomic features extracted from the prefrontal cortex, subcortical GM, and cerebellum. The machine learning models predicted disability with high accuracy ( The multidimensional analysis of brain MR images, including radiomic features and clinicodemographic data, is highly informative of the clinical status of patients with multiple sclerosis, representing a promising approach to bridge the gap between conventional imaging and disability.
Sections du résumé
BACKGROUND AND PURPOSE
Conventional MR imaging explains only a fraction of the clinical outcome variance in multiple sclerosis. We aimed to evaluate machine learning models for disability prediction on the basis of radiomic, volumetric, and connectivity features derived from routine brain MR images.
MATERIALS AND METHODS
In this retrospective cross-sectional study, 3T brain MR imaging studies of patients with multiple sclerosis, including 3D T1-weighted and T2-weighted FLAIR sequences, were selected from 2 institutions. T1-weighted images were processed to obtain volume, connectivity score (inferred from the T2 lesion location), and texture features for an atlas-based set of GM regions. The site 1 cohort was randomly split into training (
RESULTS
The selection procedure identified the 9 most informative variables, including age and secondary-progressive course and a subset of radiomic features extracted from the prefrontal cortex, subcortical GM, and cerebellum. The machine learning models predicted disability with high accuracy (
CONCLUSIONS
The multidimensional analysis of brain MR images, including radiomic features and clinicodemographic data, is highly informative of the clinical status of patients with multiple sclerosis, representing a promising approach to bridge the gap between conventional imaging and disability.
Identifiants
pubmed: 34531195
pii: ajnr.A7274
doi: 10.3174/ajnr.A7274
pmc: PMC8583276
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
1927-1933Informations de copyright
© 2021 by American Journal of Neuroradiology.
Références
Nat Rev Neurol. 2015 Oct;11(10):597-606
pubmed: 26369511
Ann Neurol. 2018 Feb;83(2):223-234
pubmed: 29328531
Eur Neurol. 2014;72 Suppl 1:1-5
pubmed: 25278115
CNS Drugs. 2017 Mar;31(3):217-236
pubmed: 28185158
AJNR Am J Neuroradiol. 2016 Mar;37(3):394-401
pubmed: 26564433
Eur Radiol. 2019 Sep;29(9):4670-4677
pubmed: 30770971
Nat Rev Neurol. 2015 Aug;11(8):471-82
pubmed: 26149978
Brain Connect. 2013;3(5):451-63
pubmed: 23855491
Mult Scler Relat Disord. 2015 Jul;4(4):329-33
pubmed: 26195051
Curr Opin Neurol. 2002 Jun;15(3):239-45
pubmed: 12045719
Mult Scler. 1999 Aug;5(4):283-6
pubmed: 10467389
Radiology. 2016 Feb;278(2):563-77
pubmed: 26579733
Nat Rev Neurol. 2010 Aug;6(8):438-44
pubmed: 20625376
Hum Brain Mapp. 2014 Dec;35(12):5946-61
pubmed: 25053254
J Neurol. 2012 Oct;259(10):2151-60
pubmed: 22446893
Radiology. 2018 Nov;289(2):487-496
pubmed: 30015589
F1000Res. 2017 Oct 12;6:1828
pubmed: 29093810
Front Neuroinform. 2014 Feb 21;8:14
pubmed: 24600388
CNS Drugs. 2018 Jun;32(6):543-558
pubmed: 29926371
Neuroimage. 2002 Jan;15(1):273-89
pubmed: 11771995
Ann Neurol. 2011 Nov;70(5):764-73
pubmed: 22162059
Nat Rev Neurol. 2015 Jun;11(6):327-38
pubmed: 26009002
Ann Neurol. 2011 Feb;69(2):292-302
pubmed: 21387374
Bioinformatics. 2004 Oct 12;20(15):2479-81
pubmed: 15073010
PLoS One. 2015 Dec 22;10(12):e0145497
pubmed: 26693908
Radiology. 2018 Aug;288(2):554-564
pubmed: 29714673
BMC Neurol. 2014 Mar 25;14:58
pubmed: 24666846
Med Phys. 2019 Nov;46(11):5116-5123
pubmed: 31539450
Cancer Res. 2017 Nov 1;77(21):e104-e107
pubmed: 29092951
Diagn Interv Imaging. 2020 Dec;101(12):795-802
pubmed: 32651155
Ann Neurol. 2008 Feb;63(2):254-62
pubmed: 18257039
AJNR Am J Neuroradiol. 2019 Jan;40(1):99-106
pubmed: 30573464
Neurology. 2004 Feb 24;62(4):586-90
pubmed: 14981175
J Neurol Neurosurg Psychiatry. 2017 Dec;88(12):1065-1072
pubmed: 28844067
Ann Neurol. 2018 Feb;83(2):210-222
pubmed: 29331092