TL;DR
This study validates a method for identifying Alzheimer's disease (AD) subtypes and their progression patterns using neuroimaging data. Crucially, it demonstrates that these patterns are consistent across a diverse cohort (HABS-HD) and a predominantly white cohort (ADNI), suggesting these subtypes can generalize to a wider population. This is a significant step towards developing more inclusive and personalized treatments for AD.
What Problem Does This Paper Address?
Alzheimer's disease and related dementias (ADRDs) are complex, with significant biological and clinical variability. This heterogeneity makes it challenging to understand disease progression and develop effective treatments. Historically, much ADRD research has relied on datasets like the Alzheimer's Disease Neuroimaging Initiative (ADNI), which are primarily composed of non-Hispanic white participants. This raises a critical question: do the identified AD subtypes and their progression patterns hold true for more diverse populations, including Black/African American and Hispanic individuals? This paper addresses this gap by testing the generalizability of AD subtypes identified in ADNI to the Health and Aging Brain Study – Health Disparities (HABS-HD) cohort, which has a broader sociodemographic representation.
Methods
The researchers utilized structural Magnetic Resonance Imaging (MRI) data from both the HABS-HD and ADNI cohorts. Standard neuroimaging processing was performed using FreeSurfer to extract key brain metrics: cortical thickness and hippocampal volume.
The core of their analysis involved applying the Subtype and Stage Inference (SuStaln) algorithm. SuStaln is designed to identify distinct spatial atrophy patterns (subtypes) and map disease progression stages within these subtypes. To compare the identified subtypes and their progression trajectories between the two cohorts, the authors employed Positional Variance Diagrams (PVDs). These diagrams visually represent the variability and uncertainty in the disease progression patterns derived by SuStaln, allowing for a direct comparison of the spatial and temporal characteristics of the subtypes across HABS-HD and ADNI.
Datasets
- Health and Aging Brain Study – Health Disparities (HABS-HD): This dataset includes structural MRI data from a diverse population, with participants categorized by race/ethnicity:
- Black/African American (AA): n = 588
- Hispanic (HIS): n = 1005
- Non-Hispanic White (NHW): n = 1009
- Alzheimer's Disease Neuroimaging Initiative (ADNI): This dataset, primarily composed of non-Hispanic white participants, was used for comparison:
- n = 864
The HABS-HD dataset is presented as a valuable resource for understanding ADRDs in diverse populations, offering a broader sociodemographic perspective compared to ADNI.
Key Insights and Ideas
- Generalizability of AD Subtypes: The most significant insight is that the spatial atrophy subtypes and their progression patterns identified by SuStaln in the predominantly white ADNI cohort are remarkably similar to those found in the more diverse HABS-HD cohort. This suggests that these subtypes are not specific to a single demographic group but represent fundamental patterns of AD progression.
- Distinct Progression Pathways: The study clearly delineates three distinct spatial atrophy subtypes, each with a characteristic pattern of degeneration:
- Subtype 1: Starts with hippocampal degeneration, then spreads to posterior and anterior brain regions.
- Subtype 2: Begins with atrophy in the occipital lobe, progressing to the parietal regions.
- Subtype 3: Characterized by initial frontal lobe degeneration, followed by spread to posterior regions.
- Clinical Correlates of Subtypes: The research highlights that these subtypes are not just anatomical patterns but also have clinical implications. In the HABS-HD cohort, Subtype 3 is associated with the most severe cognitive impairment and the oldest age group, underscoring the clinical relevance of these identified patterns.
- Importance of Diverse Datasets: The study emphasizes the critical need for research in diverse populations. By using HABS-HD, the authors were able to validate findings from ADNI and demonstrate that AD progression patterns can be generalized, which is crucial for developing equitable diagnostic and therapeutic strategies.
Key Results
- The SuStaln algorithm successfully identified three distinct spatial atrophy subtypes and progression stages in both HABS-HD and ADNI cohorts.
- The Positional Variance Diagrams (PVDs) showed that the HABS-HD cohort's subtype progression patterns closely aligned with those observed in the ADNI cohort.
- Subtype 1: Characterized by initial hippocampal degeneration, followed by posterior and then anterior spread.
- Subtype 2: Characterized by initial occipital atrophy, progressing to parietal regions.
- Subtype 3: Characterized by initial frontal degeneration, with subsequent spread to posterior regions.
- In the HABS-HD cohort, significant demographic and clinical differences were observed across subtypes (p < 0.05). Specifically, Subtype 3 exhibited the most severe cognitive impairment and comprised the oldest age group.
- Table 1 shows significant differences in age, CDR, and MMSE scores across the HABS-HD subtypes, with Subtype 3 showing the lowest MMSE scores and highest CDR values.
Limitations and Future Work
- Sample Size for Subtype 1 in HABS-HD: While the paper states SuStaln identified three subtypes in HABS-HD, the table for HABS-HD subtypes shows N=104 for Subtype 1, N=114 for Subtype 2, and N=114 for Subtype 3. This suggests that Subtype 1 might have a smaller representation in the HABS-HD cohort compared to the other two, which could influence the robustness of findings for this specific subtype within HABS-HD.
- Further Validation: While the study demonstrates generalizability, further validation across even more diverse cohorts and with longitudinal data would strengthen these findings.
- Mechanism Exploration: The paper identifies distinct patterns and their clinical correlates. Future work could delve deeper into the underlying biological mechanisms driving these specific subtype progressions and how they might differ across racial and ethnic groups.
- Clinical Translation: The authors highlight the potential of these subtypes for advancing clinical trials and precision medicine. Future work will focus on translating these findings into practical applications, such as developing targeted interventions or biomarkers for specific subtypes.
Summary updated