Objective: Examine utility of a factor analysis approach to neuroimaging data in Lewy body disease (LBD) and Alzheimer’s disease (AD).

Background: Significant heterogeneity exists among patients with LBD and AD. While neuroimaging and clinical evaluation can assist in differentiating the two syndromes, there remains significant overlap. Efforts to better understand the neurobiological underpinnings associated with biomarkers, cognition, and clinical symptoms within these diseases is needed.

Method: Participants with a diagnosis of AD or LBD were recruited from movement and memory disorders clinics. Participants underwent neuropsychological testing, neuroimaging (FDG PET, amyloid PET, DaTscan) and completed symptom questionnaires. A principal components analysis (PCA) was used to characterize neuroimaging data based on FDG PET and DatScan ROI ratings. The relationship between pathology (amyloid positivity, DaT positivity), neuroimaging factor scores, and clinical variables was examined using nonparametric statistics.

Results: 48 participants (Mean age: 69, range 47-81) with all imaging were included in present analyses. The PCA revealed 5 distinct factors: 1) subcortical FDG, 2) frontoparietal FDG, 3) basal ganglia(BG) DaT, 4) occipital FDG, and 5) right frontal FDG. The frontoparietal FDG and BG DaT factors were related to amyloid positivity (p < .001 and p = .019, respectively) and DaT Positivity (p=.03 and p < .001). The occipital FDG factor was related to DaT positivity only (p = .014).  Follow up analyses within the DaT positive subset, revealed presence of amyloid (i.e. mixed pathology) was related to frontoparietal FDG (p = .044) and BG DaT (p < .001). Correlational analyses revealed the subcortical FDG factor correlated positively with the Modified Somatic Perception Questionnaire (p = .01), while attention and executive function was positively associated with the frontoparietal FDG factor and negatively associated with BG DaT factor.

Conclusion: Overall, our approach revealed distinct FDG and DaT factors can be identified when considered within the same statistical model and were differentially related to pathology and clinical presentation measures in a complex clinical sample of older adults presenting with diagnoses of AD or LBD. In particular, this approach may hold utility in clinical settings given that the individual ROIs included within the model were based on radiology ratings.

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/neuroimaging-factors-differentially-associated-with-pathology-and-clinical-presentation/