Cortical atrophy pattern–based subtyping predicts prognosis of amnestic MCI: an individual-level analysis

Hee Jin Kim, Jong Yun Park, Sang Won Seo, Young Hee Jung, Yeshin Kim, Hyemin Jang, Sung Tae Kim, Jun Kyung Seong, Duk L. Na

Research output: Contribution to journalArticle

Abstract

We categorized patients with amnestic mild cognitive impairment (aMCI) based on cortical atrophy patterns and evaluated whether the prognosis differed across the subtypes. Furthermore, we developed a classifier that learns the cortical atrophy pattern and predicts subtypes at an individual level. A total of 662 patients with aMCI were clustered into 3 subtypes based on cortical atrophy patterns. Of these, 467 patients were followed up for more than 12 months, and the median follow-up duration was 43 months. To predict individual-level subtype, we used a machine learning–based classifier with a 10-fold cross-validation scheme. Patients with aMCI were clustered into 3 subtypes: medial temporal atrophy, minimal atrophy (Min), and parietotemporal atrophy (PT) subtypes. The PT subtype had higher prevalence of APOE ε4 carriers, amyloid PET positivity, and greater risk of dementia conversion than the Min subtype. The accuracy for binary classification was 89.3% (MT vs. Rest), 92.6% (PT vs. Rest), and 86.6% (Min vs. Rest). When we used ensemble model of 3 binary classifiers, the accuracy for predicting the aMCI subtype at an individual level was 89.6%. Patients with aMCI with the PT subtype were more likely to have underlying Alzheimer's disease pathology and showed the worst prognosis. Our classifier may be useful for predicting the prognosis of individual aMCI patients.

LanguageEnglish
Pages38-45
Number of pages8
JournalNeurobiology of Aging
Volume74
DOIs
Publication statusPublished - 2019 Feb 1

Fingerprint

Atrophy
Amyloid
Dementia
Cognitive Dysfunction
Alzheimer Disease
Pathology

Keywords

  • Alzheimer's disease
  • Classifier
  • Cortical atrophy pattern
  • Mild cognitive impairment

ASJC Scopus subject areas

  • Neuroscience(all)
  • Ageing
  • Clinical Neurology
  • Developmental Biology
  • Geriatrics and Gerontology

Cite this

Cortical atrophy pattern–based subtyping predicts prognosis of amnestic MCI : an individual-level analysis. / Kim, Hee Jin; Park, Jong Yun; Seo, Sang Won; Jung, Young Hee; Kim, Yeshin; Jang, Hyemin; Kim, Sung Tae; Seong, Jun Kyung; Na, Duk L.

In: Neurobiology of Aging, Vol. 74, 01.02.2019, p. 38-45.

Research output: Contribution to journalArticle

Kim, Hee Jin ; Park, Jong Yun ; Seo, Sang Won ; Jung, Young Hee ; Kim, Yeshin ; Jang, Hyemin ; Kim, Sung Tae ; Seong, Jun Kyung ; Na, Duk L. / Cortical atrophy pattern–based subtyping predicts prognosis of amnestic MCI : an individual-level analysis. In: Neurobiology of Aging. 2019 ; Vol. 74. pp. 38-45.
@article{780d57d7f35f40719df268526597db53,
title = "Cortical atrophy pattern–based subtyping predicts prognosis of amnestic MCI: an individual-level analysis",
abstract = "We categorized patients with amnestic mild cognitive impairment (aMCI) based on cortical atrophy patterns and evaluated whether the prognosis differed across the subtypes. Furthermore, we developed a classifier that learns the cortical atrophy pattern and predicts subtypes at an individual level. A total of 662 patients with aMCI were clustered into 3 subtypes based on cortical atrophy patterns. Of these, 467 patients were followed up for more than 12 months, and the median follow-up duration was 43 months. To predict individual-level subtype, we used a machine learning–based classifier with a 10-fold cross-validation scheme. Patients with aMCI were clustered into 3 subtypes: medial temporal atrophy, minimal atrophy (Min), and parietotemporal atrophy (PT) subtypes. The PT subtype had higher prevalence of APOE ε4 carriers, amyloid PET positivity, and greater risk of dementia conversion than the Min subtype. The accuracy for binary classification was 89.3{\%} (MT vs. Rest), 92.6{\%} (PT vs. Rest), and 86.6{\%} (Min vs. Rest). When we used ensemble model of 3 binary classifiers, the accuracy for predicting the aMCI subtype at an individual level was 89.6{\%}. Patients with aMCI with the PT subtype were more likely to have underlying Alzheimer's disease pathology and showed the worst prognosis. Our classifier may be useful for predicting the prognosis of individual aMCI patients.",
keywords = "Alzheimer's disease, Classifier, Cortical atrophy pattern, Mild cognitive impairment",
author = "Kim, {Hee Jin} and Park, {Jong Yun} and Seo, {Sang Won} and Jung, {Young Hee} and Yeshin Kim and Hyemin Jang and Kim, {Sung Tae} and Seong, {Jun Kyung} and Na, {Duk L.}",
year = "2019",
month = "2",
day = "1",
doi = "10.1016/j.neurobiolaging.2018.10.010",
language = "English",
volume = "74",
pages = "38--45",
journal = "Neurobiology of Aging",
issn = "0197-4580",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Cortical atrophy pattern–based subtyping predicts prognosis of amnestic MCI

T2 - Neurobiology of Aging

AU - Kim, Hee Jin

AU - Park, Jong Yun

AU - Seo, Sang Won

AU - Jung, Young Hee

AU - Kim, Yeshin

AU - Jang, Hyemin

AU - Kim, Sung Tae

AU - Seong, Jun Kyung

AU - Na, Duk L.

PY - 2019/2/1

Y1 - 2019/2/1

N2 - We categorized patients with amnestic mild cognitive impairment (aMCI) based on cortical atrophy patterns and evaluated whether the prognosis differed across the subtypes. Furthermore, we developed a classifier that learns the cortical atrophy pattern and predicts subtypes at an individual level. A total of 662 patients with aMCI were clustered into 3 subtypes based on cortical atrophy patterns. Of these, 467 patients were followed up for more than 12 months, and the median follow-up duration was 43 months. To predict individual-level subtype, we used a machine learning–based classifier with a 10-fold cross-validation scheme. Patients with aMCI were clustered into 3 subtypes: medial temporal atrophy, minimal atrophy (Min), and parietotemporal atrophy (PT) subtypes. The PT subtype had higher prevalence of APOE ε4 carriers, amyloid PET positivity, and greater risk of dementia conversion than the Min subtype. The accuracy for binary classification was 89.3% (MT vs. Rest), 92.6% (PT vs. Rest), and 86.6% (Min vs. Rest). When we used ensemble model of 3 binary classifiers, the accuracy for predicting the aMCI subtype at an individual level was 89.6%. Patients with aMCI with the PT subtype were more likely to have underlying Alzheimer's disease pathology and showed the worst prognosis. Our classifier may be useful for predicting the prognosis of individual aMCI patients.

AB - We categorized patients with amnestic mild cognitive impairment (aMCI) based on cortical atrophy patterns and evaluated whether the prognosis differed across the subtypes. Furthermore, we developed a classifier that learns the cortical atrophy pattern and predicts subtypes at an individual level. A total of 662 patients with aMCI were clustered into 3 subtypes based on cortical atrophy patterns. Of these, 467 patients were followed up for more than 12 months, and the median follow-up duration was 43 months. To predict individual-level subtype, we used a machine learning–based classifier with a 10-fold cross-validation scheme. Patients with aMCI were clustered into 3 subtypes: medial temporal atrophy, minimal atrophy (Min), and parietotemporal atrophy (PT) subtypes. The PT subtype had higher prevalence of APOE ε4 carriers, amyloid PET positivity, and greater risk of dementia conversion than the Min subtype. The accuracy for binary classification was 89.3% (MT vs. Rest), 92.6% (PT vs. Rest), and 86.6% (Min vs. Rest). When we used ensemble model of 3 binary classifiers, the accuracy for predicting the aMCI subtype at an individual level was 89.6%. Patients with aMCI with the PT subtype were more likely to have underlying Alzheimer's disease pathology and showed the worst prognosis. Our classifier may be useful for predicting the prognosis of individual aMCI patients.

KW - Alzheimer's disease

KW - Classifier

KW - Cortical atrophy pattern

KW - Mild cognitive impairment

UR - http://www.scopus.com/inward/record.url?scp=85056167379&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056167379&partnerID=8YFLogxK

U2 - 10.1016/j.neurobiolaging.2018.10.010

DO - 10.1016/j.neurobiolaging.2018.10.010

M3 - Article

VL - 74

SP - 38

EP - 45

JO - Neurobiology of Aging

JF - Neurobiology of Aging

SN - 0197-4580

ER -