Investigating associations of an MRI‐based measure of Alzheimer's disease neuroanatomic risk with incident cognitive impairment and β‐amyloid burden across race and sex in the ARIC cohort: Neuroimaging / Optimal neuroimaging measures for early detection. (7th December 2020)

• Record Type: Journal Article
• Authors: Casanova, Ramon; Hayden, Kathleen M.; Barnard, Ryan; Anderson, Andrea; Hsu, Fang‐Chi; Talluri, Rajesh; Whitlow, Christopher T.; Griswold, Michael E; Hughes, Timothy M.; Gottesman, Rebecca F; Wagenknecht, Lynne E.
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 5
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.040984
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: MRI based biomarkers are of increasing interest due to availability, low cost, and low patient burden compared to other biomarkers. Alzheimer's disease pattern similarity (AD‐PS) scores, an MRI‐based AD risk measure, have been previously investigated in predominantly white cohorts. The Atherosclerosis Risk in Communities (ARIC) study provides an opportunity to further evaluate the AD‐PS score in a diverse setting. Method: The ARIC study has collected cognitive data and MRI scans at visit 5 (2011‐2013) and visit 6 (2015‐2017). Participants were aged 76.0 (5.3) yrs at visit 5 and included both blacks and whites. MRI scans from the ARIC and Alzheimer's Disease Neuroimaging Initiative (ADNI) studies were aligned and segmented using the Advanced Normalization Toolbox software. A high‐dimensional classifier was trained using gray matter images (N = 598) from the ADNI study. Then ARIC MRI images were provided as input to the classifier to obtain the scores. Incident cognitive impairment (ICI) analyses were based on individuals diagnosed as cognitively normal (CN) (N = 868) at visit 5. Logistic regression considered ICI (either a diagnosis of mild cognitive impairment or dementia at visit 6) as the outcome. In PET analyses logistic regression was used to investigate associations of the AD‐PS scores with cross‐sectional β‐amyloid lesion burden (global cortical standardized uptake value ratio dichotomized at 1.2) derived from F18‐Florbetapir PET images available in a subset (N = 330, MCI=89). All analyses were adjusted for age, race, sex, education, site, and intracranial volume. The relevant variables were standardized. Analyses were further stratified by race and sex. Result: Among participants who were CN at visit 5, 125 were diagnosed with MCI and 28 with dementia at visit 6. The AD‐PS scores were significantly associated with ICI in: the full sample, whites, blacks, males and females (Table 1). The scores were also significantly associated with β‐amyloid burden in: the full sample, blacks and males. Conclusion: AD‐PS scores, an MRI based biomarker of AD, were associated with ICI and β‐amyloid burden derived from PET in the ARIC study. Significant associations were found in the overall sample, as well as by race and sex.