Distinctive age‐related longitudinal dementia progression patterns using a machine‐learning‐based MRI biomarker. (20th December 2022)

• Record Type: Journal Article
• Authors: Ma, Da; Popuri, Karteek; Wang, Lei; Lockhart, Samuel N.; Craft, Suzanne; Gurcan, Metin Nafi; Beg, Mirza Faisal
• Is Part Of: Alzheimer's & dementia. Volume 18(2022)Supplement 1
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-12-20
• 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.067454
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Dementia of Alzheimer's type (DAT) is an age‐related neurodegenerative syndrome caused by Alzheimer's disease. The risk of developing dementia increases with age, but also depends on disease trajectory. In this study, we aim to evaluate the effect of normal aging on the MRI‐derived dementia risk score and investigate the distinctive age‐related longitudinal dementia progression patterns among individuals with different dementia trajectories. Methods: A total of 1609 participants with multi‐timepoint neuroimaging tests and diagnoses were recruited from the ADNI. Based on their dementia trajectories, the subjects were stratified into four groups: stable normal cognition (sNC=423), stable/progressive mild cognitive impairment (sMCI=535/pMCI=321), and stable dementia of Alzheimer's type (sDAT=330). Brain structures were automatically segmented and harmonized to control the confounding effect of sex and scanner, and feed into an extensively‐validated kernel‐based ensemble classifier to derive the score, which we termed longitudinal MR‐DAT‐Score (MRDATS). The CSF‐based biomarker was derived as the ratio between the total tau and the amyloid beta (t‐tau/Aβ1‐42). We analyzed the effect of aging on the biomarkers' distribution of each stratified group in terms of their empirical cumulative distribution function (ECDF). We also evaluated the difference in longitudinal MRDATS progression patterns among different groups. Results: Compared to the CSF biomarker, the MRDATS revealed a more distinctive age‐related risk increase among all stratified groups (Figure 1). In addition, the effect of aging persisted across both the non‐progression groups (sNC and sMCI) and the progression groups (pMCI) except for the DAT group, where the age‐related effect is only observable for the 80‐90 year old group. Distinctive longitudinal dementia risk progression patterns (in terms of MRDATS) were also observed for each stratified group (Figure 2). Specifically, for patients with mild cognitive impairment (MCI), distinctive patterns were observed between sMCI and pMCI across subjects from different aged populations. Conclusion: Our study demonstrated differential aging effect patterns as well as distinctive longitudinal patterns between AD‐progressive and non‐progressive subjects across different age groups. The results of this study emphasize the importance of disentangling the effect of aging with disease‐driven brain atrophy patterns.