Effect of genetic factors on [18F]FDG PET metabolic phenotypes in dementia with Lewy bodies

Abstract

BACKGROUND: F]FDG PET can support the diagnosis of Dementia with Lewy Bodies (DLB), but it remains unclear how genetic factors influence metabolic phenotypes. OBJECTIVES: F]FDG PET metabolic patterns in DLB. METHODS: F]FDG PET scans from 43 patients with DLB stratified by GBA1 and APOE ε4 status, and 35 from healthy subjects. Analyses included the cingulate island sign (CIS), regions of interest, SSM/PCA disease patterns, and a machine learning multi-class model. We evaluated the similarity of the DLB patient scans in our cohort with respect to typical DLB, Alzheimer's disease (AD) and Parkinson's disease (PD)-like patterns. RESULTS: APOE ε4 status mainly influenced the CIS, with APOE ε4-negative patients showing greater preservation of the DLB-typical CIS pattern (p = 0.03) compared to APOE ε4 carriers. In contrast, GBA1 status influenced global metabolic phenotype. GBA1 carriers showed a more homogeneous PD/DLB-like metabolic pattern in the machine learning model (p = 0.003) compared to GBA1 non-carriers, whereas the latter group demonstrated greater heterogeneity and higher expression of the AD-related metabolic pattern (p = 0.004). These effects were observed along a metabolic spectrum rather than as distinct clusters. CONCLUSIONS: APOE ε4 and GBA1 modulate distinct aspects of the metabolic phenotype in DLB. GBA1 non-carriers and APOE ε4 carriers showed a higher rate of atypical metabolic signatures, which may contribute to biological heterogeneity and increase the risk of diagnostic misclassification.