Abstract

Background The ability to predict Alzheimer’s disease (AD) before diagnosis is a topic of intense research. Early diagnosis would aid in improving treatment and intervention options, however, there are no current methods that can accurately predict AD years in advance. This study examines a novel machine learning approach that integrates the combined effects of vascular (white matter hyperintensities, WMHs), and structural brain changes (gray matter, GM) with clinical factors (cognitive status) to predict post-mortem neuropathological outcomes.

Methods Healthy older adults, participants with mild cognitive impairment, and AD from the Alzheimer’s Disease Neuroimaging Initiative dataset with both post-mortem neuropathology data and antemortem MRI and clinical data were included. Longitudinal data were analyzed across three intervals before death (post-mortem data): 0-4 years, 4-8 years, and 8-14 years. Additionally, cross-sectional data at the last visit or interval (within four years, 0-4 years) before death were also examined. Machine learning models including gradient boosting, bagging, support vector regression, and linear regression were implemented. These models were applied towards feature selection of the top seven MRI, clinical, and demographic data to identify the best performing set of variables that could predict postmortem neuropathology outcomes (i.e., neurofibrillary tangles, neuritic plaques, diffuse plaques, senile/amyloid plaques, and amyloid angiopathy).

Results A total of 94 participants (55-90 years of age) were included in the study. At last visit, the best-performing model included total and temporal lobe WMHs and achieved r=0.87(RMSE=0.62) during cross-validation for neuritic plaques. For longitudinal assessments across different intervals, the best-performing model included regional GM (i.e., hippocampus, amygdala, caudate) and frontal lobe WMH and achieved r=0.93(RMSE=0.59) during cross-validation for neurofibrillary tangles. For MRI and clinical predictors and clinical-only predictors, t-tests demonstrated significant differences at all intervals before death (t[-13.60-7.90], p-values<0.001). Overall, post-mortem neuropathology outcome were predicted up to 14 years before death with high accuracies (∼90%).

Conclusions Prediction accuracy was higher for post-mortem neuropathology outcomes that included MRI (WMHs, GM) and clinical features compared to clinical-only features. These findings highlight that MRI features are critical to successfully predict AD-related pathology years in advance which will improve participant selection for clinical trials, treatments, and intervention options.

Competing Interest Statement

The authors have declared no competing interest.

Funding Statement

This research was supported by a grant from the Canadian Institutes of Health Research (CIHR). Dr. Kamal is supported by a scholarship from Fonds de Recherche du Quebec - Sante (FRQS). Dr. Dadar reports receiving research funding from the Quebec Bio-Imaging Network and Fonds de Recherche du Quebec - Sante (FRQS), Natural Sciences and Engineering Research Council of Canada (NSERC), Healthy Brains for Healthy Lives (HBHL), Alzheimer Society Research Program (ASRP), CIHR, and Douglas Research Centre (DRC). Dr. Morrison is supported by CIHR.

Author Declarations

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Data used in preparation of this article were also obtained from the Alzheimers Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu).

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

Data Availability

Data used in preparation of this article were also obtained from the Alzheimers Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu).

AbbreviationsADAlzheimer’s diseaseGMgray matterWMwhite matterWMHwhite matter hyperintensityMRIMagnetic resonance imagingNFTneurofibrillary tangle burdenMLmachine learningSVMsupport vector machineADNIAlzheimer’s Disease Neuroimaging InitiativeCDR-SBClinical Dementia Rating-Sum of BoxesSCDsubjective cognitive declineRMSEroot mean square errorFDRfalse discovery rate