Hemodynamic MRI parameters to predict asymptomatic unilateral carotid artery stenosis with random forest machine learning.
asymptomatic carotid artery stenosis
hemodynamics
individual watershed areas
machine learning
magnetic resonance imaging (MRI)
random forest–ensemble classifier
Journal
Frontiers in neuroimaging
ISSN: 2813-1193
Titre abrégé: Front Neuroimaging
Pays: Switzerland
ID NLM: 9918402387106676
Informations de publication
Date de publication:
2022
2022
Historique:
received:
28
09
2022
accepted:
20
12
2022
medline:
9
8
2023
pubmed:
9
8
2023
entrez:
9
8
2023
Statut:
epublish
Résumé
Internal carotid artery stenosis (ICAS) can cause stroke and cognitive decline. Associated hemodynamic impairments, which are most pronounced within individual watershed areas (iWSA) between vascular territories, can be assessed with hemodynamic-oxygenation-sensitive MRI and may help to detect severely affected patients. We aimed to identify the most sensitive parameters and volumes of interest (VOI) to predict high-grade ICAS with random forest machine learning. We hypothesized an increased predictive ability considering iWSAs and a decreased cognitive performance in correctly classified patients. Twenty-four patients with asymptomatic, unilateral, high-grade carotid artery stenosis and 24 age-matched healthy controls underwent MRI comprising pseudo-continuous arterial spin labeling (pCASL), breath-holding functional MRI (BH-fMRI), dynamic susceptibility contrast (DSC), T2 and T2 The most sensitive features in decreasing order were time-to-peak (TTP), cerebral blood flow (CBF) and cerebral vascular reactivity (CVR), all of these inside of iWSAs. Applying iWSAs combined with feature selection yielded significantly higher receiver operating characteristics areas under the curve (AUC) than whole GM/WM VOIs (AUC: 0.84 vs. 0.90, Random forest classifiers trained on multiparametric MRI data allow identification of the most relevant parameters and VOIs to predict ICAS, which may improve personalized treatments.
Sections du résumé
Background
UNASSIGNED
Internal carotid artery stenosis (ICAS) can cause stroke and cognitive decline. Associated hemodynamic impairments, which are most pronounced within individual watershed areas (iWSA) between vascular territories, can be assessed with hemodynamic-oxygenation-sensitive MRI and may help to detect severely affected patients. We aimed to identify the most sensitive parameters and volumes of interest (VOI) to predict high-grade ICAS with random forest machine learning. We hypothesized an increased predictive ability considering iWSAs and a decreased cognitive performance in correctly classified patients.
Materials and methods
UNASSIGNED
Twenty-four patients with asymptomatic, unilateral, high-grade carotid artery stenosis and 24 age-matched healthy controls underwent MRI comprising pseudo-continuous arterial spin labeling (pCASL), breath-holding functional MRI (BH-fMRI), dynamic susceptibility contrast (DSC), T2 and T2
Results
UNASSIGNED
The most sensitive features in decreasing order were time-to-peak (TTP), cerebral blood flow (CBF) and cerebral vascular reactivity (CVR), all of these inside of iWSAs. Applying iWSAs combined with feature selection yielded significantly higher receiver operating characteristics areas under the curve (AUC) than whole GM/WM VOIs (AUC: 0.84 vs. 0.90,
Conclusion
UNASSIGNED
Random forest classifiers trained on multiparametric MRI data allow identification of the most relevant parameters and VOIs to predict ICAS, which may improve personalized treatments.
Identifiants
pubmed: 37555162
doi: 10.3389/fnimg.2022.1056503
pmc: PMC10406220
doi:
Types de publication
Journal Article
Langues
eng
Pagination
1056503Informations de copyright
Copyright © 2023 Gleißner, Kaczmarz, Kufer, Schmitzer, Kallmayer, Zimmer, Wiestler, Preibisch and Göttler.
Déclaration de conflit d'intérêts
SK was employed by Philips GmbH Market DACH, Hamburg, Germany. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Références
Neuroimage Clin. 2020;28:102508
pubmed: 33395997
J Cereb Blood Flow Metab. 2016 May;36(5):842-61
pubmed: 26945019
Neurology. 2013 Jun 4;80(23):2145-50
pubmed: 23624562
Neuroimage Clin. 2014 Aug 28;6:115-25
pubmed: 25379423
Circ Res. 2017 Feb 3;120(3):573-591
pubmed: 28154105
J Cereb Blood Flow Metab. 2020 Feb;40(2):314-327
pubmed: 30480463
J Cereb Blood Flow Metab. 2014 Sep;34(9):1511-21
pubmed: 24938401
Stroke. 2000 Jul;31(7):1561-5
pubmed: 10884454
Neuroimage. 2019 Feb 15;187:209-225
pubmed: 29793062
Stroke. 2019 May;50(5):1081-1088
pubmed: 30943887
Neuroradiology. 2018 Mar;60(3):311-323
pubmed: 29299616
J Cereb Blood Flow Metab. 2019 Nov;39(11):2132-2143
pubmed: 29968499
JAMA. 1993 May 12;269(18):2386-91
pubmed: 8479064
Cancer. 1950 Jan;3(1):32-5
pubmed: 15405679
J Neuroradiol. 2019 Feb;46(1):44-51
pubmed: 29753641
AJNR Am J Neuroradiol. 2001 Jun-Jul;22(6):1056-61
pubmed: 11415897
Neuroepidemiology. 2013;40(1):36-41
pubmed: 23075828
Neuroimage. 2019 Oct 1;199:351-365
pubmed: 31173905
Stroke. 2014 Aug;45(8):2366-71
pubmed: 24994719
J Cereb Blood Flow Metab. 2021 Feb;41(2):380-396
pubmed: 32237952
AJNR Am J Neuroradiol. 2015 Jan;36(1):7-13
pubmed: 24788129
Arch Clin Neuropsychol. 2004 Mar;19(2):203-14
pubmed: 15010086
Stroke. 1991 Jun;22(6):711-20
pubmed: 2057968
Stroke. 1981 Jul-Aug;12(4):454-9
pubmed: 6976022
J Cereb Blood Flow Metab. 2018 Sep;38(9):1461-1480
pubmed: 28598243
Stroke. 2011 Jun;42(6):1550-5
pubmed: 21527764
J Cereb Blood Flow Metab. 2020 Apr;40(4):760-774
pubmed: 30952200
Phys Med Biol. 2017 Mar 21;62(6):2361-2375
pubmed: 28234631
Stroke. 2013 Apr;44(4):1002-7
pubmed: 23404720
Ann Transl Med. 2020 Oct;8(19):1277
pubmed: 33178809
J Magn Reson Imaging. 2021 Dec;54(6):1878-1889
pubmed: 34145686
J Vasc Surg. 2017 Oct;66(4):1083-1092
pubmed: 28712815
Biometrics. 1988 Sep;44(3):837-45
pubmed: 3203132
J Cereb Blood Flow Metab. 2016 Oct;36(10):1804-1812
pubmed: 26661219
NMR Biomed. 2014 Jul;27(7):853-62
pubmed: 24809665
Stroke. 2006 Mar;37(3):841-6
pubmed: 16424374
Chin Med Sci J. 2020 Dec 31;35(4):297-305
pubmed: 33413745