Alzheimer's disease heterogeneity revealed by neuroanatomical normative modeling

Abstract

Introduction: Overlooking the heterogeneity in Alzheimer's disease (AD) may lead to diagnostic delays and failures. Neuroanatomical normative modeling captures individual brain variation and may inform our understanding of individual differences in AD-related atrophy.

Methods: We applied neuroanatomical normative modeling to magnetic resonance imaging from a real-world clinical cohort with confirmed AD (n = 86). Regional cortical thickness was compared to a healthy reference cohort (n = 33,072) and the number of outlying regions was summed (total outlier count) and mapped at individual- and group-levels.

Results: The superior temporal sulcus contained the highest proportion of outliers (60%). Elsewhere, overlap between patient atrophy patterns was low. Mean total outlier count was higher in patients who were non-amnestic, at more advanced disease stages, and without depressive symptoms. Amyloid burden was negatively associated with outlier count.

Discussion: Brain atrophy in AD is highly heterogeneous and neuroanatomical normative modeling can be used to explore anatomo-clinical correlations in individual patients.

Keywords: Alzheimer's disease; MRI; amyloid PET; heterogeneity; neuroanatomical normative modeling; neurodegeneration.

FIGURE 1

Overall outlier distribution. (A) Distribution of outlier prevalence across the left (LH) and right (RH) hemispheres. (B) Outlier maps showing spatial distribution of outliers in the clinical cohort (n = 86). The superior temporal sulci (in green) featured the highest number of outliers (ie, regions with significantly reduced thickness compared to the norm) in both hemispheres. (C) Hamming distance plot illustrating dissimilarity between patients in the spatial distribution of outliers. Yellow indicates greater dissimilarity. (D) Outlier distance density illustrates the spread of outlier dissimilarity (calculated by Hamming distance).

FIGURE 2

Outlier profiles according to disease severity. (A) Outlier maps showing distribution of outliers according to disease severity. (B) Hamming distance plot illustrating dissimilarity between patients in the spatial distribution of outliers; the yellow color indicates greater dissimilarity. (C) Outlier distance density illustrates the spread of outlier dissimilarity (calculated by Hamming distance). AD, Alzheimer's disease; MCI, mild cognitive impairment.

FIGURE 3

Outlier profiles according to phenotype. (A) Outlier maps showing distribution of outliers according to phenotype. (B) Hamming distance plot illustrating dissimilarity between patients in the spatial distribution of outliers; the yellow color indicates greater dissimilarity. (C) Outlier distance density illustrates the spread of outlier dissimilarity (calculated by Hamming distance).

FIGURE 4

Case series. This short case series illustrates the possible use of outlier maps to gain insight into the association between atrophy profiles and clinical history. These four MCI patients had a similar clinical presentation, a positive amyloid PET imaging, but very heterogeneous patterns of outlier regions. Purple‐colored areas indicate outlier regions (z‐score < 1.96). This finding corroborates the large heterogeneity of AD atrophy profiles at presentation and indicates another possible application of normative modeling for a closer investigation of anatomo‐clinical associations. (A) A man in his 70s presenting to our clinic with a 3‐year history of memory problems, intact activities of daily living (ADLs) and preserved insight. Medical history review did not highlight significant comorbidities or depressive symptoms. On examination, he scored 94/100 on the ACE‐III and 26/30 on the MMSE. Clinical follow‐up revealed a slow progression of cognitive deficits. (B) A man in his late 60s presenting with a 4‐year history of memory problems and preserved ADLs. Insight into the cognitive difficulties was limited and collateral account reported behavioral features such as passivity and reduced empathy. No history of depression was recorded. On examination, ACE‐III score was 85/100. Follow‐up visits revealed slow progression of the cognitive deficits with relative sparing of ADLs. (C) A lady in her 70s presenting with a 2‐year history of memory problems with intact ADLs, preserved insight, and no history of depression. MMSE score was 26/30. Follow‐up visits revealed a steady decline with gradual involvement of ADLs. (D) A man in his mid‐60s presenting with a 2‐year history of memory problems and intact ADLs and no history of depression. The ACE‐III score was 78/100 and follow‐up visits highlighted clinical progression. The MMSE score at 2 years following the first examination was 22/30. ACE‐III, Addenbrooke's Cognitive Examination version III; AD, Alzheimer's disease; ADLs, activities of daily living; MCI, mild cognitive impairment; MMSE, mini‐mental state examination; MR, magnetic resonance; PET, positron emission tomography; totOC, total outlier count.