Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

MRI-Derived AD Signature of Cortical Thinning and Plasma P-Tau217 for Predicting Alzheimer Dementia Among Community-Dwelling Older Adults

Yu, Lei ; Wang, Tianhao ; Hansson, Oskar LU orcid ; Janelidze, Shorena LU ; Lamar, Melissa ; Arfanakis, Konstantinos ; Bennett, David A. ; Schneider, Julie A. and Boyle, Patricia A. (2024) In Neurology: Clinical Practice 14(3).
Abstract

Background and Objectives Structural brain MRI and blood-based phosphorylated tau (p-tau) measures are among the least invasive and least expensive Alzheimer's disease (AD) biomarkers to date. The extent to which these biomarkersmay outperformone another in predicting future Alzheimer dementia diagnosis is poorly understood, however. This study investigated 2 specific AD biomarkers, i.e., a cortical thickness signature of AD (AD-CT) and plasma p-tau217, for predicting Alzheimer dementia. Methods Data came from community-dwelling older participants of the Religious Orders Study or the Rush Memory and Aging Project. AD-CT was obtained from 3T MRI scans using a magnetization-prepared rapid acquisition gradient echo sequence and by... (More)

Background and Objectives Structural brain MRI and blood-based phosphorylated tau (p-tau) measures are among the least invasive and least expensive Alzheimer's disease (AD) biomarkers to date. The extent to which these biomarkersmay outperformone another in predicting future Alzheimer dementia diagnosis is poorly understood, however. This study investigated 2 specific AD biomarkers, i.e., a cortical thickness signature of AD (AD-CT) and plasma p-tau217, for predicting Alzheimer dementia. Methods Data came from community-dwelling older participants of the Religious Orders Study or the Rush Memory and Aging Project. AD-CT was obtained from 3T MRI scans using a magnetization-prepared rapid acquisition gradient echo sequence and by averaging thickness from previously identified cortical regions implicated in AD. Plasma p-tau217 was quantified using an immunoassay developed by Lilly Research Laboratories on the MSD platform. Both MRI scans and blood specimens were collected at the same visits, and subsequent diagnoses of Alzheimer dementia were determined through annual detailed clinical evaluations. Cox proportional hazards models examined the associations of the 2 biomarkers with incident Alzheimer dementia, and prediction accuracy was assessed using c-statistics. Results A total of 198 older adults, on average 84 years of age, were included. Over a mean follow-up of 4 years, 60 (30%) individuals developed Alzheimer dementia. AD-CT (hazard ratio: 1.71, 95% CI 1.26-2.31) and separately plasma p-tau217 (hazard ratio: 2.57, 95% CI 1.83-3.61) were associated with incident Alzheimer dementia. The c-statistic for prediction accuracy was consistently higher for plasma p-tau217 (between 0.74 and 0.81) than AD-CT (between 0.70 and 0.75) across a range of time horizons. Furthermore, with both biomarkers included in the same model, there was only modest improvement in the c-statistic due to AD-CT. Discussion Plasma p-tau217 outperforms an imaging-based cortical thickness signature of AD in predicting future Alzheimer dementia diagnosis. Furthermore, the AD cortical thickness signature adds little to the prediction accuracy above and beyond plasma p-tau217.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
in
Neurology: Clinical Practice
volume
14
issue
3
article number
e200291
publisher
Lippincott Williams & Wilkins
external identifiers
  • pmid:38720951
  • scopus:85206600684
ISSN
2163-0402
DOI
10.1212/CPJ.0000000000200291
language
English
LU publication?
yes
id
a4abcc25-3475-41bd-88a3-6d8880b73b75
date added to LUP
2024-12-11 11:57:36
date last changed
2025-07-10 05:02:51
@article{a4abcc25-3475-41bd-88a3-6d8880b73b75,
  abstract     = {{<p>Background and Objectives Structural brain MRI and blood-based phosphorylated tau (p-tau) measures are among the least invasive and least expensive Alzheimer's disease (AD) biomarkers to date. The extent to which these biomarkersmay outperformone another in predicting future Alzheimer dementia diagnosis is poorly understood, however. This study investigated 2 specific AD biomarkers, i.e., a cortical thickness signature of AD (AD-CT) and plasma p-tau217, for predicting Alzheimer dementia. Methods Data came from community-dwelling older participants of the Religious Orders Study or the Rush Memory and Aging Project. AD-CT was obtained from 3T MRI scans using a magnetization-prepared rapid acquisition gradient echo sequence and by averaging thickness from previously identified cortical regions implicated in AD. Plasma p-tau217 was quantified using an immunoassay developed by Lilly Research Laboratories on the MSD platform. Both MRI scans and blood specimens were collected at the same visits, and subsequent diagnoses of Alzheimer dementia were determined through annual detailed clinical evaluations. Cox proportional hazards models examined the associations of the 2 biomarkers with incident Alzheimer dementia, and prediction accuracy was assessed using c-statistics. Results A total of 198 older adults, on average 84 years of age, were included. Over a mean follow-up of 4 years, 60 (30%) individuals developed Alzheimer dementia. AD-CT (hazard ratio: 1.71, 95% CI 1.26-2.31) and separately plasma p-tau217 (hazard ratio: 2.57, 95% CI 1.83-3.61) were associated with incident Alzheimer dementia. The c-statistic for prediction accuracy was consistently higher for plasma p-tau217 (between 0.74 and 0.81) than AD-CT (between 0.70 and 0.75) across a range of time horizons. Furthermore, with both biomarkers included in the same model, there was only modest improvement in the c-statistic due to AD-CT. Discussion Plasma p-tau217 outperforms an imaging-based cortical thickness signature of AD in predicting future Alzheimer dementia diagnosis. Furthermore, the AD cortical thickness signature adds little to the prediction accuracy above and beyond plasma p-tau217.</p>}},
  author       = {{Yu, Lei and Wang, Tianhao and Hansson, Oskar and Janelidze, Shorena and Lamar, Melissa and Arfanakis, Konstantinos and Bennett, David A. and Schneider, Julie A. and Boyle, Patricia A.}},
  issn         = {{2163-0402}},
  language     = {{eng}},
  number       = {{3}},
  publisher    = {{Lippincott Williams & Wilkins}},
  series       = {{Neurology: Clinical Practice}},
  title        = {{MRI-Derived AD Signature of Cortical Thinning and Plasma P-Tau217 for Predicting Alzheimer Dementia Among Community-Dwelling Older Adults}},
  url          = {{http://dx.doi.org/10.1212/CPJ.0000000000200291}},
  doi          = {{10.1212/CPJ.0000000000200291}},
  volume       = {{14}},
  year         = {{2024}},
}