Novel insights into relationships between metabolic covariance patterns of FDG-PET data and clinical status in Parkinson's disease using partial least squares correlation analysis

IntroductionMetabolic covariance patterns derived from imaging data help characterize disease-related physiological changes in several neurodegenerative disorders, including Parkinson's disease (PD), but their relevance to different disease stages and/or clinical variables related to disease or disease predisposition, such as age, is often unclear.MethodsWe incorporated clinical information in deriving metabolic covariance patterns relevant to different aspects of PD: disease initiation, disease progression, and physiological similarities in PD and healthy aging. This was achieved by combining Partial Least Squares Correlation analysis with Scaled Subprofile Modeling (SSM-PLSC).ResultsWhen combining PD and HC data, SSM-PLSC identified a spatial pattern similar to the well-known PD-related disease pattern as expected; when applied to PD-only data-thus emphasizing disease progression-the spatial pattern became characterized by expanding putaminal hypermetabolism and reduced emphasis on cerebellar hypermetabolism. Finally, when applied to PD and HC data but permitting a different dependence on clinical variables, SSM-PLSC identified a spatial pattern with relative hypermetabolism in the basal ganglia, brain stem, and white matter together with relative hypometabolism in frontal cortex; in HC this pattern solely related to age, while in PD the same pattern significantly correlated with both age and disease duration.ConclusionWe identified metabolic patterns that are more closely related with different aspects of PD, directly derived from relationships between metabolic alterations and clinical variables. We also revealed metabolic signatures common to PD and aging, which may highlight age-related metabolic changes that form a predisposition to PD, as age is the single highest risk factor for PD.Plain language summary titleBrain changes in Parkinson's disease and their relationship to clinical aspects of disease.