Development and validation of the automated imaging differentiation in parkinsonism (AID-P): a multicentre machine learning study

Derek B. Archer, Justin T. Bricker, Winston T. Chu, Roxana G. Burciu, Johanna L. McCracken, Song Lai, Stephen A. Coombes, Ruogu Fang, Angelos Barmpoutis, Daniel M. Corcos, Ajay S. Kurani, Trina Mitchell, Mieniecia L. Black, Ellen Herschel, Tanya Simuni, Todd B. Parrish, Cynthia Comella, Tao Xie, Klaus Seppi, Nicolaas I. BohnenMartijn LTM Müller, Roger L. Albin, Florian Krismer, Guangwei Du, Mechelle M. Lewis, Xuemei Huang, Hong Li, Ofer Pasternak, Nikolaus R. McFarland, Michael S. Okun, David E. Vaillancourt

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Background: Development of valid, non-invasive biomarkers for parkinsonian syndromes is crucially needed. We aimed to assess whether non-invasive diffusion-weighted MRI can distinguish between parkinsonian syndromes using an automated imaging approach. Methods: We did an international study at 17 MRI centres in Austria, Germany, and the USA. We used diffusion-weighted MRI from 1002 patients and the Movement Disorders Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS III) to develop and validate disease-specific machine learning comparisons using 60 template regions and tracts of interest in Montreal Neurological Institute space between Parkinson's disease and atypical parkinsonism (multiple system atrophy and progressive supranuclear palsy) and between multiple system atrophy and progressive supranuclear palsy. For each comparison, models were developed on a training and validation cohort and evaluated in an independent test cohort by quantifying the area under the curve (AUC) of receiving operating characteristic curves. The primary outcomes were free water and free-water-corrected fractional anisotropy across 60 different template regions. Findings: In the test cohort for disease-specific comparisons, the diffusion-weighted MRI plus MDS-UPDRS III model (Parkinson's disease vs atypical parkinsonism had an AUC 0·962; multiple system atrophy vs progressive supranuclear palsy AUC 0·897) and diffusion-weighted MRI only model had high AUCs (Parkinson's disease vs atypical parkinsonism AUC 0·955; multiple system atrophy vs progressive supranuclear palsy AUC 0·926), whereas the MDS-UPDRS III only models had significantly lower AUCs (Parkinson's disease vs atypical parkinsonism 0·775; multiple system atrophy vs progressive supranuclear palsy 0·582). These results indicate that a non-invasive imaging approach is capable of differentiating forms of parkinsonism comparable to current gold standard methods. Interpretations: This study provides an objective, validated, and generalisable imaging approach to distinguish different forms of parkinsonian syndromes using multisite diffusion-weighted MRI cohorts. The diffusion-weighted MRI method does not involve radioactive tracers, is completely automated, and can be collected in less than 12 min across 3T scanners worldwide. The use of this test could positively affect the clinical care of patients with Parkinson's disease and parkinsonism and reduce the number of misdiagnosed cases in clinical trials. Funding: National Institutes of Health and Parkinson's Foundation.

Original languageEnglish (US)
Pages (from-to)e222-e231
JournalThe Lancet Digital Health
Volume1
Issue number5
DOIs
StatePublished - Sep 2019

Fingerprint

Parkinsonian Disorders
Diffusion Magnetic Resonance Imaging
Progressive Supranuclear Palsy
Multiple System Atrophy
Area Under Curve
Parkinson Disease
Movement Disorders
Radioactive Tracers
Machine Learning
Imaging
Parkinson's disease
Machine learning
Water
Austria
Anisotropy
National Institutes of Health (U.S.)
Diagnostic Errors
Germany
Patient Care
Biomarkers

All Science Journal Classification (ASJC) codes

  • Medicine (miscellaneous)
  • Health Informatics
  • Decision Sciences (miscellaneous)
  • Health Information Management

Cite this

Archer, D. B., Bricker, J. T., Chu, W. T., Burciu, R. G., McCracken, J. L., Lai, S., ... Vaillancourt, D. E. (2019). Development and validation of the automated imaging differentiation in parkinsonism (AID-P): a multicentre machine learning study. The Lancet Digital Health, 1(5), e222-e231. https://doi.org/10.1016/S2589-7500(19)30105-0
Archer, Derek B. ; Bricker, Justin T. ; Chu, Winston T. ; Burciu, Roxana G. ; McCracken, Johanna L. ; Lai, Song ; Coombes, Stephen A. ; Fang, Ruogu ; Barmpoutis, Angelos ; Corcos, Daniel M. ; Kurani, Ajay S. ; Mitchell, Trina ; Black, Mieniecia L. ; Herschel, Ellen ; Simuni, Tanya ; Parrish, Todd B. ; Comella, Cynthia ; Xie, Tao ; Seppi, Klaus ; Bohnen, Nicolaas I. ; Müller, Martijn LTM ; Albin, Roger L. ; Krismer, Florian ; Du, Guangwei ; Lewis, Mechelle M. ; Huang, Xuemei ; Li, Hong ; Pasternak, Ofer ; McFarland, Nikolaus R. ; Okun, Michael S. ; Vaillancourt, David E. / Development and validation of the automated imaging differentiation in parkinsonism (AID-P) : a multicentre machine learning study. In: The Lancet Digital Health. 2019 ; Vol. 1, No. 5. pp. e222-e231.
@article{23c0a6c0383e4890997863180b45faeb,
title = "Development and validation of the automated imaging differentiation in parkinsonism (AID-P): a multicentre machine learning study",
abstract = "Background: Development of valid, non-invasive biomarkers for parkinsonian syndromes is crucially needed. We aimed to assess whether non-invasive diffusion-weighted MRI can distinguish between parkinsonian syndromes using an automated imaging approach. Methods: We did an international study at 17 MRI centres in Austria, Germany, and the USA. We used diffusion-weighted MRI from 1002 patients and the Movement Disorders Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS III) to develop and validate disease-specific machine learning comparisons using 60 template regions and tracts of interest in Montreal Neurological Institute space between Parkinson's disease and atypical parkinsonism (multiple system atrophy and progressive supranuclear palsy) and between multiple system atrophy and progressive supranuclear palsy. For each comparison, models were developed on a training and validation cohort and evaluated in an independent test cohort by quantifying the area under the curve (AUC) of receiving operating characteristic curves. The primary outcomes were free water and free-water-corrected fractional anisotropy across 60 different template regions. Findings: In the test cohort for disease-specific comparisons, the diffusion-weighted MRI plus MDS-UPDRS III model (Parkinson's disease vs atypical parkinsonism had an AUC 0·962; multiple system atrophy vs progressive supranuclear palsy AUC 0·897) and diffusion-weighted MRI only model had high AUCs (Parkinson's disease vs atypical parkinsonism AUC 0·955; multiple system atrophy vs progressive supranuclear palsy AUC 0·926), whereas the MDS-UPDRS III only models had significantly lower AUCs (Parkinson's disease vs atypical parkinsonism 0·775; multiple system atrophy vs progressive supranuclear palsy 0·582). These results indicate that a non-invasive imaging approach is capable of differentiating forms of parkinsonism comparable to current gold standard methods. Interpretations: This study provides an objective, validated, and generalisable imaging approach to distinguish different forms of parkinsonian syndromes using multisite diffusion-weighted MRI cohorts. The diffusion-weighted MRI method does not involve radioactive tracers, is completely automated, and can be collected in less than 12 min across 3T scanners worldwide. The use of this test could positively affect the clinical care of patients with Parkinson's disease and parkinsonism and reduce the number of misdiagnosed cases in clinical trials. Funding: National Institutes of Health and Parkinson's Foundation.",
author = "Archer, {Derek B.} and Bricker, {Justin T.} and Chu, {Winston T.} and Burciu, {Roxana G.} and McCracken, {Johanna L.} and Song Lai and Coombes, {Stephen A.} and Ruogu Fang and Angelos Barmpoutis and Corcos, {Daniel M.} and Kurani, {Ajay S.} and Trina Mitchell and Black, {Mieniecia L.} and Ellen Herschel and Tanya Simuni and Parrish, {Todd B.} and Cynthia Comella and Tao Xie and Klaus Seppi and Bohnen, {Nicolaas I.} and M{\"u}ller, {Martijn LTM} and Albin, {Roger L.} and Florian Krismer and Guangwei Du and Lewis, {Mechelle M.} and Xuemei Huang and Hong Li and Ofer Pasternak and McFarland, {Nikolaus R.} and Okun, {Michael S.} and Vaillancourt, {David E.}",
year = "2019",
month = "9",
doi = "10.1016/S2589-7500(19)30105-0",
language = "English (US)",
volume = "1",
pages = "e222--e231",
journal = "The Lancet Digital Health",
issn = "2589-7500",
publisher = "Elsevier Ltd",
number = "5",

}

Archer, DB, Bricker, JT, Chu, WT, Burciu, RG, McCracken, JL, Lai, S, Coombes, SA, Fang, R, Barmpoutis, A, Corcos, DM, Kurani, AS, Mitchell, T, Black, ML, Herschel, E, Simuni, T, Parrish, TB, Comella, C, Xie, T, Seppi, K, Bohnen, NI, Müller, MLTM, Albin, RL, Krismer, F, Du, G, Lewis, MM, Huang, X, Li, H, Pasternak, O, McFarland, NR, Okun, MS & Vaillancourt, DE 2019, 'Development and validation of the automated imaging differentiation in parkinsonism (AID-P): a multicentre machine learning study', The Lancet Digital Health, vol. 1, no. 5, pp. e222-e231. https://doi.org/10.1016/S2589-7500(19)30105-0

Development and validation of the automated imaging differentiation in parkinsonism (AID-P) : a multicentre machine learning study. / Archer, Derek B.; Bricker, Justin T.; Chu, Winston T.; Burciu, Roxana G.; McCracken, Johanna L.; Lai, Song; Coombes, Stephen A.; Fang, Ruogu; Barmpoutis, Angelos; Corcos, Daniel M.; Kurani, Ajay S.; Mitchell, Trina; Black, Mieniecia L.; Herschel, Ellen; Simuni, Tanya; Parrish, Todd B.; Comella, Cynthia; Xie, Tao; Seppi, Klaus; Bohnen, Nicolaas I.; Müller, Martijn LTM; Albin, Roger L.; Krismer, Florian; Du, Guangwei; Lewis, Mechelle M.; Huang, Xuemei; Li, Hong; Pasternak, Ofer; McFarland, Nikolaus R.; Okun, Michael S.; Vaillancourt, David E.

In: The Lancet Digital Health, Vol. 1, No. 5, 09.2019, p. e222-e231.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Development and validation of the automated imaging differentiation in parkinsonism (AID-P)

T2 - a multicentre machine learning study

AU - Archer, Derek B.

AU - Bricker, Justin T.

AU - Chu, Winston T.

AU - Burciu, Roxana G.

AU - McCracken, Johanna L.

AU - Lai, Song

AU - Coombes, Stephen A.

AU - Fang, Ruogu

AU - Barmpoutis, Angelos

AU - Corcos, Daniel M.

AU - Kurani, Ajay S.

AU - Mitchell, Trina

AU - Black, Mieniecia L.

AU - Herschel, Ellen

AU - Simuni, Tanya

AU - Parrish, Todd B.

AU - Comella, Cynthia

AU - Xie, Tao

AU - Seppi, Klaus

AU - Bohnen, Nicolaas I.

AU - Müller, Martijn LTM

AU - Albin, Roger L.

AU - Krismer, Florian

AU - Du, Guangwei

AU - Lewis, Mechelle M.

AU - Huang, Xuemei

AU - Li, Hong

AU - Pasternak, Ofer

AU - McFarland, Nikolaus R.

AU - Okun, Michael S.

AU - Vaillancourt, David E.

PY - 2019/9

Y1 - 2019/9

N2 - Background: Development of valid, non-invasive biomarkers for parkinsonian syndromes is crucially needed. We aimed to assess whether non-invasive diffusion-weighted MRI can distinguish between parkinsonian syndromes using an automated imaging approach. Methods: We did an international study at 17 MRI centres in Austria, Germany, and the USA. We used diffusion-weighted MRI from 1002 patients and the Movement Disorders Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS III) to develop and validate disease-specific machine learning comparisons using 60 template regions and tracts of interest in Montreal Neurological Institute space between Parkinson's disease and atypical parkinsonism (multiple system atrophy and progressive supranuclear palsy) and between multiple system atrophy and progressive supranuclear palsy. For each comparison, models were developed on a training and validation cohort and evaluated in an independent test cohort by quantifying the area under the curve (AUC) of receiving operating characteristic curves. The primary outcomes were free water and free-water-corrected fractional anisotropy across 60 different template regions. Findings: In the test cohort for disease-specific comparisons, the diffusion-weighted MRI plus MDS-UPDRS III model (Parkinson's disease vs atypical parkinsonism had an AUC 0·962; multiple system atrophy vs progressive supranuclear palsy AUC 0·897) and diffusion-weighted MRI only model had high AUCs (Parkinson's disease vs atypical parkinsonism AUC 0·955; multiple system atrophy vs progressive supranuclear palsy AUC 0·926), whereas the MDS-UPDRS III only models had significantly lower AUCs (Parkinson's disease vs atypical parkinsonism 0·775; multiple system atrophy vs progressive supranuclear palsy 0·582). These results indicate that a non-invasive imaging approach is capable of differentiating forms of parkinsonism comparable to current gold standard methods. Interpretations: This study provides an objective, validated, and generalisable imaging approach to distinguish different forms of parkinsonian syndromes using multisite diffusion-weighted MRI cohorts. The diffusion-weighted MRI method does not involve radioactive tracers, is completely automated, and can be collected in less than 12 min across 3T scanners worldwide. The use of this test could positively affect the clinical care of patients with Parkinson's disease and parkinsonism and reduce the number of misdiagnosed cases in clinical trials. Funding: National Institutes of Health and Parkinson's Foundation.

AB - Background: Development of valid, non-invasive biomarkers for parkinsonian syndromes is crucially needed. We aimed to assess whether non-invasive diffusion-weighted MRI can distinguish between parkinsonian syndromes using an automated imaging approach. Methods: We did an international study at 17 MRI centres in Austria, Germany, and the USA. We used diffusion-weighted MRI from 1002 patients and the Movement Disorders Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS III) to develop and validate disease-specific machine learning comparisons using 60 template regions and tracts of interest in Montreal Neurological Institute space between Parkinson's disease and atypical parkinsonism (multiple system atrophy and progressive supranuclear palsy) and between multiple system atrophy and progressive supranuclear palsy. For each comparison, models were developed on a training and validation cohort and evaluated in an independent test cohort by quantifying the area under the curve (AUC) of receiving operating characteristic curves. The primary outcomes were free water and free-water-corrected fractional anisotropy across 60 different template regions. Findings: In the test cohort for disease-specific comparisons, the diffusion-weighted MRI plus MDS-UPDRS III model (Parkinson's disease vs atypical parkinsonism had an AUC 0·962; multiple system atrophy vs progressive supranuclear palsy AUC 0·897) and diffusion-weighted MRI only model had high AUCs (Parkinson's disease vs atypical parkinsonism AUC 0·955; multiple system atrophy vs progressive supranuclear palsy AUC 0·926), whereas the MDS-UPDRS III only models had significantly lower AUCs (Parkinson's disease vs atypical parkinsonism 0·775; multiple system atrophy vs progressive supranuclear palsy 0·582). These results indicate that a non-invasive imaging approach is capable of differentiating forms of parkinsonism comparable to current gold standard methods. Interpretations: This study provides an objective, validated, and generalisable imaging approach to distinguish different forms of parkinsonian syndromes using multisite diffusion-weighted MRI cohorts. The diffusion-weighted MRI method does not involve radioactive tracers, is completely automated, and can be collected in less than 12 min across 3T scanners worldwide. The use of this test could positively affect the clinical care of patients with Parkinson's disease and parkinsonism and reduce the number of misdiagnosed cases in clinical trials. Funding: National Institutes of Health and Parkinson's Foundation.

UR - http://www.scopus.com/inward/record.url?scp=85071720507&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071720507&partnerID=8YFLogxK

U2 - 10.1016/S2589-7500(19)30105-0

DO - 10.1016/S2589-7500(19)30105-0

M3 - Article

AN - SCOPUS:85071720507

VL - 1

SP - e222-e231

JO - The Lancet Digital Health

JF - The Lancet Digital Health

SN - 2589-7500

IS - 5

ER -