Identifying multiple sclerosis subtypes using unsupervised machine learning and MRI data

Arman Eshaghi; Alexandra L. Young; Peter A. Wijeratne; Ferran Prados; Douglas L. Arnold; Sridar Narayanan; Charles R. G. Guttmann; Frederik Barkhof; Daniel C. Alexander; Alan J. Thompson; Declan Chard; Olga Ciccarelli

doi:https://doi.org/10.1038/s41467-021-22265-2

Identifying multiple sclerosis subtypes using unsupervised machine learning and MRI data

Arman Eshaghi, Alexandra L. Young, Peter A. Wijeratne, Ferran Prados, Douglas L. Arnold, Sridar Narayanan, Charles R. G. Guttmann, Frederik Barkhof, Daniel C. Alexander, Alan J. Thompson, Declan Chard, Olga Ciccarelli

Research output: Contribution to journal › Article › Academic › peer-review

85 Citations (Scopus)

Abstract

Multiple sclerosis (MS) can be divided into four phenotypes based on clinical evolution. The pathophysiological boundaries of these phenotypes are unclear, limiting treatment stratification. Machine learning can identify groups with similar features using multidimensional data. Here, to classify MS subtypes based on pathological features, we apply unsupervised machine learning to brain MRI scans acquired in previously published studies. We use a training dataset from 6322 MS patients to define MRI-based subtypes and an independent cohort of 3068 patients for validation. Based on the earliest abnormalities, we define MS subtypes as cortex-led, normal-appearing white matter-led, and lesion-led. People with the lesion-led subtype have the highest risk of confirmed disability progression (CDP) and the highest relapse rate. People with the lesion-led MS subtype show positive treatment response in selected clinical trials. Our findings suggest that MRI-based subtypes predict MS disability progression and response to treatment and may be used to define groups of patients in interventional trials.

Original language	English
Article number	2078
Journal	Nature communications
Volume	12
Issue number	1
DOIs	https://doi.org/10.1038/s41467-021-22265-2
Publication status	Published - 1 Dec 2021

Access to Document

https://doi.org/10.1038/s41467-021-22265-2

Cite this

@article{6dc943914719477f8f92151cc16fdda0,

title = "Identifying multiple sclerosis subtypes using unsupervised machine learning and MRI data",

abstract = "Multiple sclerosis (MS) can be divided into four phenotypes based on clinical evolution. The pathophysiological boundaries of these phenotypes are unclear, limiting treatment stratification. Machine learning can identify groups with similar features using multidimensional data. Here, to classify MS subtypes based on pathological features, we apply unsupervised machine learning to brain MRI scans acquired in previously published studies. We use a training dataset from 6322 MS patients to define MRI-based subtypes and an independent cohort of 3068 patients for validation. Based on the earliest abnormalities, we define MS subtypes as cortex-led, normal-appearing white matter-led, and lesion-led. People with the lesion-led subtype have the highest risk of confirmed disability progression (CDP) and the highest relapse rate. People with the lesion-led MS subtype show positive treatment response in selected clinical trials. Our findings suggest that MRI-based subtypes predict MS disability progression and response to treatment and may be used to define groups of patients in interventional trials.",

author = "Arman Eshaghi and Young, {Alexandra L.} and Wijeratne, {Peter A.} and Ferran Prados and Arnold, {Douglas L.} and Sridar Narayanan and Guttmann, {Charles R. G.} and Frederik Barkhof and Alexander, {Daniel C.} and Thompson, {Alan J.} and Declan Chard and Olga Ciccarelli",

note = "Funding Information: A.E. has received speaker{\textquoteright}s honoraria from Biogen and At The Limits educational programme. He has received travel support from the National Multiple Sclerosis Society and honorarium from the Journal of Neurology, Neurosurgy and Psychiatry for Editorial Commentaries. In the last 3 years D.C. has received honoraria from Excemed (2017) for faculty-led education work; had meeting expenses funded by the IMSCOGS (2019), EAN (2018), ECTRIMS (2018) and Soci{\'e}t{\'e} des Neurosciences (2017). He is a consultant for Biogen and Hoffmann-La Roche. He has received research funding from the International Progressive MS Alliance, the MS Society, and the National Institute for Health Research (NIHR) University College London Hospitals (UCLH) Biomedical Research Centre. He is a member of the MS Society{\textquoteright}s Biomedical Grant Review Panel and a trustee of the MS Trust. O.C. has received research grants from the MS Society of Great Britain & Northern Ireland, National Institute for Health Research (NIHR) University College London Hospitals Biomedical Research Centre, EUH2020, Spinal Cord Research Foundation, and Rosetrees Trust. She serves as a consultant for Novartis, Teva, and Roche and has received an honorarium from the American Academy of Neurology as Associate Editor of Neurology and serves on the Editorial Board of Multiple Sclerosis Journal. CRGG has received research grants form Sanofi and the National Multiple Sclerosis Society. F.B. has received compensation for consulting services and/or speaking activities from Bayer Schering Pharma, Biogen Idec, Merck Serono, Novartis, Genzyme, Synthon BV, Roche, Teva, Jansen research and IXICO and is supported by the NIHR Biomedical Research Centre at UCLH. A.J.T. has received honoraria/support for travel for consultancy from Eisai, Hoffman La Roche, Almirall, and Excemed, and support for travel for consultancy as chair of the International Progressive MS Alliance Scientific Steering Committee, and from the National MS Society (USA) as a member of the Research Programs Advisory Committee. He receives an honorarium from SAGE Publishers as Editor-in-Chief of Multiple Sclerosis. Journal and a free subscription from Elsevier as a board member for the Lancet Neurology. D.L.A. has received research grant funding and/ or personal compensation for consulting from Acorda, Adelphi, Alkermes, Biogen, Celgene, Frequency Therapeutics, Genentech, Genzyme, Hoffman-La Roche, Immuene Tolerance Network, Immunotec, MedDay, EMD Serono, Novartis, Pfizer, Receptos, Roche, Sanofi-Aventis, Canadian Institutes of Health Research, MS Society of Canada, and International Progressive MS Alliance; and holds an equity interest in NeuroRx Research. F.B., D.C.A. and A.E. hold equity stake in Queen Square Analytics. S.N. has received research funding from the Canadian Institutes of Health Research, the International Progressive MS Alliance, the Myelin Repair Foundation and Immunotec. He has received honoraria/travel support from Genentech and MedDay, and personal compensation from NeuroRx Research. The remaining authors declare no competing interests. Funding Information: This study was supported by the International Progressive MS Alliance (IPMSA, award reference number PA-1603-08175). We are grateful to all the IPMSA investigators who have contributed trial data to this study as part of EPITOME: Enhancing Power of Intervention Trials Through Optimized MRI Endpoints network (see the list of investigators in the appendix). This study was also supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre. O.C. is a National Institute of Health Research (NIHR) Professor (grant code: RP-2017-08-ST2-004). We are grateful to Professor Geraint Rees for his comments. We thank Rozie Arnaoutellis, Istvan Morocz, and Caramanos Zografos for coordinating and organising this study. We thank Jonathan Steel for IT support during this work. This research in part has been conducted using the UK Biobank Resource under Application Number 47233. Data have also been provided in part by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University. D.C.A. has received funding for this work from Engineering and Physical Sciences Research Council Grants M020533, M006093, and J020990. This project has received funding from the European Union{\textquoteright}s Horizon 2020 Research and Innovation Programme under Grant Agreements 666992. MS-SMART is an investigator-led project sponsored by University College London (UCL). This project (reference 11/30/11) is funded by the Efficacy and Mechanism Evaluation (EME) Programme, a Medical Research Council (MRC) and National Institute for Health Research (NIHR) partnership. The views expressed in this publication are those of the author(s) and not necessarily those of the MRC, NIHR, or the Department of Health and Social Care. A.J.T. is a National Institute of Health Research (NIHR) Emeritus Senior Investigator. A.L.Y. is supported by an MRC skills development fellowship (MR/T027800/1). Publisher Copyright: {\textcopyright} 2021, The Author(s). Copyright: Copyright 2021 Elsevier B.V., All rights reserved.",

year = "2021",

month = dec,

day = "1",

doi = "https://doi.org/10.1038/s41467-021-22265-2",

language = "English",

volume = "12",

journal = "Nature communications",

issn = "2041-1723",

publisher = "Nature Publishing Group",

number = "1",

}

TY - JOUR

T1 - Identifying multiple sclerosis subtypes using unsupervised machine learning and MRI data

AU - Eshaghi, Arman

AU - Young, Alexandra L.

AU - Wijeratne, Peter A.

AU - Prados, Ferran

AU - Arnold, Douglas L.

AU - Narayanan, Sridar

AU - Guttmann, Charles R. G.

AU - Barkhof, Frederik

AU - Alexander, Daniel C.

AU - Thompson, Alan J.

AU - Chard, Declan

AU - Ciccarelli, Olga

N1 - Funding Information: A.E. has received speaker’s honoraria from Biogen and At The Limits educational programme. He has received travel support from the National Multiple Sclerosis Society and honorarium from the Journal of Neurology, Neurosurgy and Psychiatry for Editorial Commentaries. In the last 3 years D.C. has received honoraria from Excemed (2017) for faculty-led education work; had meeting expenses funded by the IMSCOGS (2019), EAN (2018), ECTRIMS (2018) and Société des Neurosciences (2017). He is a consultant for Biogen and Hoffmann-La Roche. He has received research funding from the International Progressive MS Alliance, the MS Society, and the National Institute for Health Research (NIHR) University College London Hospitals (UCLH) Biomedical Research Centre. He is a member of the MS Society’s Biomedical Grant Review Panel and a trustee of the MS Trust. O.C. has received research grants from the MS Society of Great Britain & Northern Ireland, National Institute for Health Research (NIHR) University College London Hospitals Biomedical Research Centre, EUH2020, Spinal Cord Research Foundation, and Rosetrees Trust. She serves as a consultant for Novartis, Teva, and Roche and has received an honorarium from the American Academy of Neurology as Associate Editor of Neurology and serves on the Editorial Board of Multiple Sclerosis Journal. CRGG has received research grants form Sanofi and the National Multiple Sclerosis Society. F.B. has received compensation for consulting services and/or speaking activities from Bayer Schering Pharma, Biogen Idec, Merck Serono, Novartis, Genzyme, Synthon BV, Roche, Teva, Jansen research and IXICO and is supported by the NIHR Biomedical Research Centre at UCLH. A.J.T. has received honoraria/support for travel for consultancy from Eisai, Hoffman La Roche, Almirall, and Excemed, and support for travel for consultancy as chair of the International Progressive MS Alliance Scientific Steering Committee, and from the National MS Society (USA) as a member of the Research Programs Advisory Committee. He receives an honorarium from SAGE Publishers as Editor-in-Chief of Multiple Sclerosis. Journal and a free subscription from Elsevier as a board member for the Lancet Neurology. D.L.A. has received research grant funding and/ or personal compensation for consulting from Acorda, Adelphi, Alkermes, Biogen, Celgene, Frequency Therapeutics, Genentech, Genzyme, Hoffman-La Roche, Immuene Tolerance Network, Immunotec, MedDay, EMD Serono, Novartis, Pfizer, Receptos, Roche, Sanofi-Aventis, Canadian Institutes of Health Research, MS Society of Canada, and International Progressive MS Alliance; and holds an equity interest in NeuroRx Research. F.B., D.C.A. and A.E. hold equity stake in Queen Square Analytics. S.N. has received research funding from the Canadian Institutes of Health Research, the International Progressive MS Alliance, the Myelin Repair Foundation and Immunotec. He has received honoraria/travel support from Genentech and MedDay, and personal compensation from NeuroRx Research. The remaining authors declare no competing interests. Funding Information: This study was supported by the International Progressive MS Alliance (IPMSA, award reference number PA-1603-08175). We are grateful to all the IPMSA investigators who have contributed trial data to this study as part of EPITOME: Enhancing Power of Intervention Trials Through Optimized MRI Endpoints network (see the list of investigators in the appendix). This study was also supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre. O.C. is a National Institute of Health Research (NIHR) Professor (grant code: RP-2017-08-ST2-004). We are grateful to Professor Geraint Rees for his comments. We thank Rozie Arnaoutellis, Istvan Morocz, and Caramanos Zografos for coordinating and organising this study. We thank Jonathan Steel for IT support during this work. This research in part has been conducted using the UK Biobank Resource under Application Number 47233. Data have also been provided in part by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University. D.C.A. has received funding for this work from Engineering and Physical Sciences Research Council Grants M020533, M006093, and J020990. This project has received funding from the European Union’s Horizon 2020 Research and Innovation Programme under Grant Agreements 666992. MS-SMART is an investigator-led project sponsored by University College London (UCL). This project (reference 11/30/11) is funded by the Efficacy and Mechanism Evaluation (EME) Programme, a Medical Research Council (MRC) and National Institute for Health Research (NIHR) partnership. The views expressed in this publication are those of the author(s) and not necessarily those of the MRC, NIHR, or the Department of Health and Social Care. A.J.T. is a National Institute of Health Research (NIHR) Emeritus Senior Investigator. A.L.Y. is supported by an MRC skills development fellowship (MR/T027800/1). Publisher Copyright: © 2021, The Author(s). Copyright: Copyright 2021 Elsevier B.V., All rights reserved.

PY - 2021/12/1

Y1 - 2021/12/1

N2 - Multiple sclerosis (MS) can be divided into four phenotypes based on clinical evolution. The pathophysiological boundaries of these phenotypes are unclear, limiting treatment stratification. Machine learning can identify groups with similar features using multidimensional data. Here, to classify MS subtypes based on pathological features, we apply unsupervised machine learning to brain MRI scans acquired in previously published studies. We use a training dataset from 6322 MS patients to define MRI-based subtypes and an independent cohort of 3068 patients for validation. Based on the earliest abnormalities, we define MS subtypes as cortex-led, normal-appearing white matter-led, and lesion-led. People with the lesion-led subtype have the highest risk of confirmed disability progression (CDP) and the highest relapse rate. People with the lesion-led MS subtype show positive treatment response in selected clinical trials. Our findings suggest that MRI-based subtypes predict MS disability progression and response to treatment and may be used to define groups of patients in interventional trials.

AB - Multiple sclerosis (MS) can be divided into four phenotypes based on clinical evolution. The pathophysiological boundaries of these phenotypes are unclear, limiting treatment stratification. Machine learning can identify groups with similar features using multidimensional data. Here, to classify MS subtypes based on pathological features, we apply unsupervised machine learning to brain MRI scans acquired in previously published studies. We use a training dataset from 6322 MS patients to define MRI-based subtypes and an independent cohort of 3068 patients for validation. Based on the earliest abnormalities, we define MS subtypes as cortex-led, normal-appearing white matter-led, and lesion-led. People with the lesion-led subtype have the highest risk of confirmed disability progression (CDP) and the highest relapse rate. People with the lesion-led MS subtype show positive treatment response in selected clinical trials. Our findings suggest that MRI-based subtypes predict MS disability progression and response to treatment and may be used to define groups of patients in interventional trials.

UR - http://www.scopus.com/inward/record.url?scp=85103936047&partnerID=8YFLogxK

U2 - https://doi.org/10.1038/s41467-021-22265-2

DO - https://doi.org/10.1038/s41467-021-22265-2

M3 - Article

C2 - 33824310

SN - 2041-1723

VL - 12

JO - Nature communications

JF - Nature communications

IS - 1

M1 - 2078

ER -

Identifying multiple sclerosis subtypes using unsupervised machine learning and MRI data

Abstract

Access to Document

Other files and links

Cite this