Alternative normalization methods demonstrate widespread cortical hypometabolism in untreated de novo Parkinson's disease

Abstract

Aim: Previous positron emission tomography (PET) [18F]fluorodeoxyglucose ([18F]FDG) studies in Parkinson's disease (PD) demonstrated that moderate to late stage patients display widespread cortical hypometabolism, whereas early stage PD patients exhibit little or no cortical changes. However, recent studies suggested that conventional data normalization procedures may not always be valid, and demonstrated that alternative normalization strategies better allow detection of low magnitude changes. We hypothesized that these alternative normalization procedures would disclose more widespread metabolic alterations in de novo PD.

Methods: [18F]FDG PET scans of 26 untreated de novo PD patients (Hoehn & Yahr stage I-II) and 21 age-matched controls were compared using voxel-based analysis. Normalization was performed using gray matter (GM), white matter (WM) reference regions and Yakushev normalization.

Results: Compared to GM normalization, WM and Yakushev normalization procedures disclosed much larger cortical regions of relative hypometabolism in the PD group with extensive involvement of frontal and parieto-temporal-occipital cortices, and several subcortical structures. Furthermore, in the WM and Yakushev normalized analyses, stage II patients displayed more prominent cortical hypometabolism than did stage I patients.

Conclusion: The use of alternative normalization procedures, other than GM, suggests that much more extensive cortical hypometabolism is present in untreated de novo PD patients than hitherto reported. The finding may have implications for our understanding of the basic pathophysiology of early-stage PD.

Figure 1

SPM clusters showing significant relative CMRglc reduction in the comparison between de novo PD patients and NL by different types of normalization (P<0.05, FWE corrected). Note that a separate t-value scale was used in the bottom row, due to the very extreme t-values found in the Yakushev analysis. Abbreviations. GM: global grey matter mean; WM: white matter.

Figure 2

SPM clusters of significant relative hypermetabolism in de novo PD patients compared to NL after GM normalization (P<0.05, FWE corrected).

Figure 3

Normalized CMRglc data for frontal, parietal, temporal and occipital lobes, using the three different types of normalization. Abbreviations. GM: global grey matter mean normalization; NL: normal controls; PD_I: Parkinson’s disease patients with H&Y score I; PD_II: Parkinson’s disease patients with H&Y score II; SPM CMRglc-_norm: normalized values of cerebral metabolic rate of glucose; WM: white matter normalization; Yakushev: Yakushev normalization; *: significant difference, P<0.05.