EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Cross‐Dataset Evaluation of Dementia Longitudinal Progression Prediction Models

Datensatzübergreifende Bewertung von Vorhersagemodellen für die longitudinale Demenzprogression

Chen Zhang iD, Lijun An iD, Naren Wulan, Kim-Ngan Nguyen, Csaba Orban, Pansheng Chen, Christopher Chen, Juan Helen Zhou iD, Keli Liu, B T Thomas Yeo iD, Alzheimer's Disease Neuroimaging Initiative and Australian Imaging Biomarkers and Lifestyle Study of Aging

Human Brain Mapping · 2025-08 · 46 (11)

DOI: 10.1002/hbm.70280 ↗

Auf einen Blick

Fünf auf ADNI trainierte Modelle wurden in drei externen Datensätzen mit 2.312 Teilnehmenden und 13.200 Zeitpunkten getestet. L2C-FNN schnitt insgesamt am besten ab.

Fragestellung

Wie gut generalisieren fünf auf ADNI trainierte Modelle zur Vorhersage der Alzheimer-Progression auf externe Kohorten?

Methode

Vergleich von fünf Algorithmen in drei externen Datensätzen; Training auf ADNI, Test auf externen Kohorten.

Ergebnisse

Getestet wurden 5 Modelle in 3 externen Datensätzen mit 2.312 Teilnehmenden und 13.200 Zeitpunkten. L2C-FNN war insgesamt am besten; bei Kognition und Ventrikelvolumen waren L2C-FNN und AD-Map vorn, bei der klinischen Diagnose L2C-FNN.

Einschränkungen

Nur auf ADNI trainiert; die Übertragbarkeit auf andere Kohorten war unklar und wurde hier erst extern geprüft.

Schlussfolgerung der Autor:innen

L2C-FNN zeigt starkes Potenzial für die kurz- und langfristige Vorhersage der Demenzprogression.

Kernaussagen

  • L2C-FNN war insgesamt das beste Modell
  • Externe Tests mit 2.312 Teilnehmenden und 13.200 Zeitpunkten
  • Leistung blieb auch bei 0 bis 6 Jahren Vorhersagehorizont stabil

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

Kurz gefasst

Fachartikel in der Zeitschrift Human Brain Mapping, 2025-08. Verfasst von Chen Zhang und 10 weiteren Autor:innen. Beteiligte Einrichtung: Centre for Sleep and Cognition (CSC) & Centre for Translational Magnetic Resonance Research (TMR), Yong Loo Lin School of Medicine, National Universi…. Hauptthemen laut MEDLINE: Disease Progression, Alzheimer Disease, Neuroimaging, Algorithms. Die Arbeit wurde bisher 10-mal zitiert (Crossref) und nennt 88 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NIA NIH HHS, Singapore National Medical Research Council, NIMH NIH HHS.

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurately predicting Alzheimer's Disease (AD) progression is useful for clinical care. The 2019 TADPOLE (The Alzheimer's Disease Prediction Of Longitudinal Evolution) challenge evaluated 92 algorithms from 33 teams worldwide. Unlike typical clinical prediction studies, TADPOLE accommodates (1) a variable number of observed time points across patients, (2) missing data across modalities and visits, and (3) prediction over an open‐ended time horizon, which better reflects real‐world data. However, TADPOLE only used the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, so how well top algorithms generalize to other cohorts remains unclear. We tested five algorithms in three external datasets covering 2312 participants and 13,200 time points. The algorithms included FROG, the overall TADPOLE winner, which utilized a unique Longitudinal‐to‐Cross‐sectional (L2C) transformation to convert variable‐length longitudinal histories into feature vectors of the same length across participants (i.e., same‐length feature vectors). We also considered two FROG variants. One variant unified all XGBoost models from the original FROG with a single feedforward neural network (FNN), which we referred to as L2C‐FNN. We also included minimal recurrent neural networks (MinimalRNN), which were ranked second at publication time, as well as AD Course Map (AD‐Map), which outperformed MinimalRNN at publication time. All five models—three FROG variants, MinimalRNN, and AD‐Map—were trained on ADNI and tested on the external datasets. L2C‐FNN performed the best overall. In the case of predicting cognition and ventricle volume, L2C‐FNN and AD‐Map were the best. For clinical diagnosis prediction, L2C‐FNN was the best, while AD‐Map was the worst. L2C‐FNN also maintained its edge over other models, regardless of the number of observed time points and regardless of the prediction horizon from 0 to 6 years into the future. L2C‐FNN shows strong potential for both short‐term and long‐term dementia progression prediction. Pretrained ADNI models are available: https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/predict_phenotypes/Zhang2025_L2CFNN .

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Disease ProgressionAlzheimer DiseaseNeuroimagingAlgorithmsNeural Networks, ComputerHumansLongitudinal StudiesAgedMaleFemaleAged, 80 and over

Stichwörter der Autor:innen

Alzheimer's diseaseXGBoostdomain generalizationfeature engineeringlongitudinal progression modelingrecurrent neural networks

Autor:innen und Institutionen

  • Centre for Sleep and Cognition (CSC) & Centre for Translational Magnetic Resonance Research (TMR), Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
  • Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore.
  • N.1 Institute for Health, National University of Singapore, Singapore, Singapore.
  • Memory Aging and Cognition Centre, Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
  • Department of Medicine, Healthy Longevity Translational Research Programme, Human Potential Translational Research Programme & Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, National University…
  • Integrative Sciences and Engineering Programme (ISEP), National University of Singapore, Singapore, Singapore.

ORCID-Profile: Chen Zhang, Lijun An, Juan Helen Zhou, B T Thomas Yeo

Förderung

  • NIA NIH HHS
  • Singapore National Medical Research Council
  • NIMH NIH HHS
  • Temasek Foundation
  • NIBIB NIH HHS
  • Ministry of Health-Singapore
  • United States National Institutes of Health
  • NCATS NIH HHS

Zitieren

APA

Zhang, C., An, L., Wulan, N., Nguyen, K. N., Orban, C., Chen, P., Chen, C., Zhou, J. H., Liu, K., Yeo, B. T. T., & Alzheimer's Disease Neuroimaging Initiative and Australian Imaging Biomarkers and Lifestyle Study of Aging (2025). Cross‐Dataset Evaluation of Dementia Longitudinal Progression Prediction Models. Human Brain Mapping, 46(11). https://doi.org/10.1002/hbm.70280

Vancouver

Zhang C, An L, Wulan N, Nguyen KN, Orban C, Chen P, et al. Cross‐Dataset Evaluation of Dementia Longitudinal Progression Prediction Models. Human Brain Mapping. 2025;46(11). doi:10.1002/hbm.70280

BibTeX
@article{Zhang2025,
  author = {Zhang, Chen and An, Lijun and Wulan, Naren and Nguyen, Kim-Ngan and Orban, Csaba and Chen, Pansheng and Chen, Christopher and Zhou, Juan Helen and Liu, Keli and Yeo, B T Thomas and Alzheimer's Disease Neuroimaging Initiative and Australian Imaging Biomarkers and Lifestyle Study of Aging, Alzheimer's Disease Neuroimaging Initiative and Australian Imaging Biomarkers and Lifestyle Study of Aging},
  title = {Cross‐Dataset Evaluation of Dementia Longitudinal Progression Prediction Models},
  journal = {Human Brain Mapping},
  year = {2025},
  volume = {46},
  number = {11},
  doi = {10.1002/hbm.70280}
}
Kostenlos starten →