1. Modeling autosomal dominant Alzheimer's disease with machine learning
- Author
-
Hiroshi Mori, Laura Swisher, Sarah B. Berman, Yi Su, Aylin Dincer, Mathias Jucker, James M. Noble, Colin L. Masters, Jonathan Vöglein, Robert A. Koeppe, Bernardino Ghetti, Tammie L.S. Benzinger, DS Marcus, Jasmeer P. Chhatwal, Lisa Cash, Jason Hassenstab, Eric McDade, Randall J. Bateman, David M. Cash, Brian A. Gordon, Richard J. Perrin, Michael W. Weiner, Ari Stern, Nelly Joseph-Mathurin, Austin A. McCullough, Qing Wang, Johannes Levin, Karin L. Meeker, Deborah Koudelis, Michael J. Fulham, Jeremy F. Strain, Anne M. Fagan, Chengjie Xiong, Russ C. Hornbeck, Nick C. Fox, Adam M. Brickman, Stephen Salloway, Beau M. Ances, Clifford R. Jack, Celeste M. Karch, Carlos Cruchaga, William S. Brooks, Martin R. Farlow, Peter R. Schofield, Patrick Luckett, Hwamee Oh, John E. McCarthy, Todd A. Kuffner, William E. Klunk, John C. Morris, and Shaney Flores
- Subjects
Male ,magnetic resonance imaging (MRI) ,Epidemiology ,Precuneus ,genetics [Alzheimer Disease] ,Disease ,computer.software_genre ,030218 nuclear medicine & medical imaging ,Machine Learning ,pathology [Alzheimer Disease] ,chemistry.chemical_compound ,0302 clinical medicine ,autosomal dominant Alzheimer's disease (ADAD) ,pathology [Atrophy] ,Aniline Compounds ,fluorodeoxyglucose (FDG) ,medicine.diagnostic_test ,Health Policy ,Magnetic Resonance Imaging ,Pittsburgh compound B (PiB) ,Psychiatry and Mental health ,medicine.anatomical_structure ,Positron emission tomography ,Mutation (genetic algorithm) ,Female ,medicine.drug ,Adult ,Amyloid ,genetics [Mutation] ,Machine learning ,Article ,03 medical and health sciences ,Cellular and Molecular Neuroscience ,Atrophy ,Developmental Neuroscience ,Alzheimer Disease ,Fluorodeoxyglucose F18 ,metabolism [Fluorodeoxyglucose F18] ,medicine ,Humans ,ddc:610 ,metabolism [Amyloid] ,Fluorodeoxyglucose ,business.industry ,medicine.disease ,Thiazoles ,chemistry ,Positron-Emission Tomography ,Mutation ,Neurology (clinical) ,Artificial intelligence ,Geriatrics and Gerontology ,Pittsburgh compound B ,business ,computer ,030217 neurology & neurosurgery - Abstract
INTRODUCTION: Machine learning models were used to discover novel disease trajectories for autosomal dominant Alzheimer's disease. METHODS: Longitudinal structural magnetic resonance imaging, amyloid positron emission tomography (PET), and fluorodeoxyglucose PET were acquired in 131 mutation carriers and 74 non-carriers from the Dominantly Inherited Alzheimer Network; the groups were matched for age, education, sex, and apolipoprotein e4 (APOE e4). A deep neural network was trained to predict disease progression for each modality. Relief algorithms identified the strongest predictors of mutation status. RESULTS: The Relief algorithm identified the caudate, cingulate, and precuneus as the strongest predictors among all modalities. The model yielded accurate results for predicting future Pittsburgh compound B (R2 = 0.95), fluorodeoxyglucose (R2 = 0.93), and atrophy (R2 = 0.95) in mutation carriers compared to non-carriers. DISCUSSION: Results suggest a sigmoidal trajectory for amyloid, a biphasic response for metabolism, and a gradual decrease in volume, with disease progression primarily in subcortical, middle frontal, and posterior parietal regions.
- Published
- 2021
- Full Text
- View/download PDF