Dataset Information

Modeling autosomal dominant Alzheimer's disease with machine learning.

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 ε4 (APOE ε4). 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.

SUBMITTER: Luckett PH

PROVIDER: S-EPMC8195816 | biostudies-literature | 2021 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Modeling autosomal dominant Alzheimer's disease with machine learning.

Luckett Patrick H PH McCullough Austin A Gordon Brian A BA Strain Jeremy J Flores Shaney S Dincer Aylin A McCarthy John J Kuffner Todd T Stern Ari A Meeker Karin L KL Berman Sarah B SB Chhatwal Jasmeer P JP Cruchaga Carlos C Fagan Anne M AM Farlow Martin R MR Fox Nick C NC Jucker Mathias M Levin Johannes J Masters Colin L CL Mori Hiroshi H Noble James M JM Salloway Stephen S Schofield Peter R PR Brickman Adam M AM Brooks William S WS Cash David M DM Fulham Michael J MJ Ghetti Bernardino B Jack Clifford R CR Vöglein Jonathan J Klunk William W Koeppe Robert R Oh Hwamee H Su Yi Y Weiner Michael M Wang Qing Q Swisher Laura L Marcus Dan D Koudelis Deborah D Joseph-Mathurin Nelly N Cash Lisa L Hornbeck Russ R Xiong Chengjie C Perrin Richard J RJ Karch Celeste M CM Hassenstab Jason J McDade Eric E Morris John C JC Benzinger Tammie L S TLS Bateman Randall J RJ Ances Beau M BM

Alzheimer's & dementia : the journal of the Alzheimer's Association 20210121 6

<h4>Introduction</h4>Machine learning models were used to discover novel disease trajectories for autosomal dominant Alzheimer's disease.<h4>Methods</h4>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 ε4 (APOE ε4). A deep neural network was trained ...[more]

PMID: 33480178

Similar Datasets

Project description:Emerging plasma biomarkers of Alzheimer's disease might be non-invasive tools to trace early Alzheimer's disease-related abnormalities such as the accumulation of amyloid-beta peptides, neurofibrillary tau tangles, glial activation and neurodegeneration. It is, however, unclear which pathological processes in the CNS can be adequately detected by peripheral measurements and whether plasma biomarkers are equally applicable in both clinical and preclinical phases. Here we aimed to explore the timing and performance of plasma biomarkers in mutation carriers compared to non-carriers in autosomal dominant Alzheimer's disease. Samples (n = 164) from mutation carriers (n = 33) and non-carriers (n = 42) in a Swedish cohort of autosomal dominant Alzheimer's disease (APP p.KM670/671NL, APP p.E693G and PSEN1 p.H163Y) were included in explorative longitudinal analyses. Plasma phosphorylated tau (P-tau181), total tau (T-tau), neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) concentrations were measured with a single-molecule array method as previously described. Plasma biomarkers were additionally correlated to Alzheimer's disease core biomarkers in the CSF. Results from the longitudinal analyses confirmed that plasma P-tau181, NfL and GFAP concentrations were higher in mutation carriers compared to non-carriers. This change was observed in the presymptomatic phase and detectable first as an increase in GFAP approximately 10 years before estimated symptom onset, followed by increased levels of P-tau181 and NfL closer to expected onset. Plasma P-tau181 levels were correlated to levels of P-tau181 and T-tau in the CSF. Altogether, plasma P-tau181, GFAP and NfL seem to be feasible biomarkers to detect different Alzheimer's disease-related pathologies already in presymptomatic individuals. Interestingly, changes in plasma GFAP concentrations were detected prior to P-tau181 and NfL. Our results suggest that plasma GFAP might reflect Alzheimer's disease pathology upstream to accumulation of tangles and neurodegeneration. The implications of these findings need additional validation, in particular because of the limited sample size.

Dataset Information

Modeling autosomal dominant Alzheimer's disease with machine learning.

Introduction

Methods

Results

Discussion

Publications

Modeling autosomal dominant Alzheimer's disease with machine learning.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets