Volume 17, Issue 12 p. 1938-1949
FEATURED ARTICLE

Personalized prediction of progression in pre-dementia patients based on individual biomarker profile: A development and validation study

Line Kühnel

Line Kühnel

H. Lundbeck A/S, Copenhagen, Denmark

Department of Mathematical Sciences, University of Copenhagen, Copenhagen, Denmark

Search for more papers by this author
Vincent Bouteloup

Vincent Bouteloup

Inserm, Population Health Research Center, University of Bordeaux, Bordeaux, France

CHU de Bordeaux, Pole Santé Publique, Talence, France

Search for more papers by this author
Jérémie Lespinasse

Jérémie Lespinasse

Inserm, Population Health Research Center, University of Bordeaux, Bordeaux, France

CHU de Bordeaux, Pole Santé Publique, Talence, France

Search for more papers by this author
Geneviève Chêne

Geneviève Chêne

Inserm, Population Health Research Center, University of Bordeaux, Bordeaux, France

CHU de Bordeaux, Pole Santé Publique, Talence, France

Search for more papers by this author
Carole Dufouil

Carole Dufouil

Inserm, Population Health Research Center, University of Bordeaux, Bordeaux, France

CHU de Bordeaux, Pole Santé Publique, Talence, France

Search for more papers by this author
José Luis Molinuevo

José Luis Molinuevo

H. Lundbeck A/S, Copenhagen, Denmark

Search for more papers by this author
Lars Lau Raket

Corresponding Author

Lars Lau Raket

H. Lundbeck A/S, Copenhagen, Denmark

Clinical Memory Research Unit, Department of Clinical Sciences, Lund University, Lund, Sweden

Correspondence

Lars Lau Raket, Department of Data Science, H. Lundbeck A/S, Ottiliavej 9, 2500 Valby, Denmark.

E-mail: [email protected]

Search for more papers by this author
for the MEMENTO study group and the Alzheimer's Disease Neuroimaging Initiative

for the MEMENTO study group and the Alzheimer's Disease Neuroimaging Initiative

Data used in preparation of this article were obtained from the MEMENTO cohort study and the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within both studies contributed to the design and implementation of the respective studies and provided data but did not participate in analysis or writing of this report. A complete listing of MEMENTO investigators can be found at http://memento-cohort.org/ and a complete listing of ADNI investigators can be found at http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf.

Search for more papers by this author
First published: 28 September 2021
Citations: 8

Abstract

Introduction

The prognosis of patients at the pre-dementia stage is difficult to define. The aim of this study is to develop and validate a biomarker-based continuous model for predicting the individual cognitive level at any future moment. In addition to personalized prognosis, such a model could reduce trial sample size requirements by allowing inclusion of a homogenous patient population.

Methods

Disease-progression modeling of longitudinal cognitive scores of pre-dementia patients (baseline Clinical Dementia Rating ≤ 0.5) was used to derive a biomarker profile that was predictive of patient's cognitive progression along the dementia continuum. The biomarker profile model was developed and validated in the MEMENTO cohort and externally validated in the Alzheimer's Disease Neuroimaging Initiative.

Results

Of nine candidate biomarkers in the development analysis, three cerebrospinal fluid and two magnetic resonance imaging measures were selected to form the final biomarker profile. The model-based prognosis of individual future cognitive deficit was shown to significantly improve when incorporating biomarker information on top of cognition and demographic data. In trial power calculations, adjusting the primary analysis for the baseline biomarker profile reduced sample size requirements by ≈10%. Compared to conventional cognitive cut-offs, inclusion criteria based on biomarker-profile cut-offs resulted in up to 28% reduced sample size requirements due to increased homogeneity in progression patterns.

Discussion

The biomarker profile allows prediction of personalized trajectories of future cognitive progression. This enables accurate personalized prognosis in clinical care and better selection of patient populations for clinical trials. A web-based application for prediction of patients’ future cognitive progression is available online.