Journal Article Ppublish
PubMed: 40684252 DOI: 10.1002/alz.70446
Staging of Alzheimer's disease progression in Down syndrome using mixed clinical and plasma biomarker measures with machine learning.
Alzheimers Dement 21(7):e70446
Authors: Mina Idris, Fedal Saini, Phoebe Ivain, R Asaad Baksh, Leda A Bianchi, Sarah E Pape, Henrik Zetterberg, Peter A Wijeratne, André Strydom
Affiliations
- Department of Forensic and Neurodevelopmental Sciences, Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, Greater London, UK.
- Natbrainlab, Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, Greater London, UK.
- Department of Forensic and Neurodevelopmental Sciences, Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, Greater London, UK. Natbrainlab, Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, Greater London, UK.
- Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, the Sahlgrenska Academy at the University of Gothenburg, Mölndal, Sweden. Clinical Neurochemistry Laboratory, Sahlgrenska University Hospital, Mölndal, Sweden. Department of Neurodegenerative Disease, UCL Institute of Neurology, London, UK. UK Dementia Research Institute at UCL, London, UK. Hong Kong Center for Neurodegenerative Diseases, InnoHK, Hong Kong, China. Wisconsin Alzheimer's Disease Research Center, University of Wisconsin School of Medicine and Public Health, University of Wisconsin-Madison, Madison, Wisconsin, USA.
- Sussex Artificial Intelligence Centre, School of Engineering and Informatics, University of Sussex, Falmer, Brighton, UK.
Abstract
INTRODUCTION
Adults with Down syndrome (DS) have a high risk for Alzheimer's disease (AD). Although the sequence of plasma biomarker and cognitive changes in AD in DS is well studied, their related trajectories are not fully characterized. Data-driven methods can estimate disease progression from cross-sectional data.
METHODS
In 57 adults with DS and no AD, we used the event-based model to sequence plasma biomarker and cognitive changes in preclinical AD. Generalized additive models assessed the relationship between age and plasma biomarkers.
RESULTS
The earliest changes occurred in the amyloid beta 42/40 ratio, followed by memory changes. Later alterations in neurofilament light and tau concentrations preceded executive and visuomotor function changes, with glial fibrillary acidic protein levels changing last. The highest rate of plasma biomarker changes occurred between ages 39 and 52.
CONCLUSION
Changes in DS follow a pattern similar to that of sporadic and familial AD. Event-based modeling offers individual-level staging, potentially optimizing diagnosis and clinical trial patient selection.
HIGHLIGHTS
The pre-clinical stages of Alzheimer's disease (AD) development in Down syndrome (DS) are not well defined, despite the extremely high prevalence of AD. Better understanding of early AD progression would aid in diagnostics and treatment. Data-driven methods, such as the event-based model, can aid in clarifying the sequence of cognitive and plasma biomarker changes in the early stages of AD while accounting for baseline variability. We find that plasma amyloid beta 42/40 ratio and memory changes precede changes in plasma biomarker levels of neurodegeneration, with changes in executive and visuomotor functions occurring later, before neuroinflammatory marker changes. Combining plasma biomarkers could be a useful measure of preclinical AD for trials, particularly in individuals between 39 and52 years of age.
MeSH
Dementia, Down Syndrome, Alzheimer Disease, Disease Progression, Cognitive Dysfunction, Biomarkers, Machine Learning, Amyloid, Learning, Memory, Cognition, Plasma