Purpose: Progressive supranuclear palsy (PSP) is an akinetic-rigid syndrome that can be difficult to differentiate from Parkinson's disease (PD), particularly at an early stage. [99mTc]ethyl cysteinate dimer (ECD) SPECT could represent a widely available tool to assist in the differential diagnosis. In this study we used voxel-based analysis and Computerised Brain Atlas (CBA)-based principal component analysis (PCA) of [99mTc]ECD SPECT data to test whether: (1) specific patterns of rCBF abnormalities can differentiate PSP from controls and PD; (2) networks of dysfunctional brain regions can be found in PSP vs controls and PD.
Methods: Nine PD patients, 16 PSP patients and ten controls were studied with [99mTc]ECD SPECT using a brain-dedicated device (Ceraspect). Voxel-based analysis was performed with statistical parametric mapping. PCA was applied to volume of interest data after spatial normalisation to CBA.
Results: The voxel-based analysis showed hypoperfusion of the anterior cingulate and medial frontal cortex in PSP compared with controls and PD. In PSP patients the rCBF impairment extended to the pre-supplementary motor area and prefrontal cortex, areas involved in executive function and motor networks. Compared with PSP patients, PD patients showed a mild rCBF decrease in associative visual areas which could be related to the known impairment of visuospatial function. The PCA identified three principal components differentiating PSP patients from controls and/or PD patients that included groups of cortical and subcortical brain regions with relatively decreased (cingulate cortex, prefrontal cortex and caudate) or increased (parietal cortex) rCBF, representing distinct functional networks in PSP.
Conclusion: Anterior cingulate hypoperfusion seems to be an early, distinct brain abnormality in PSP as compared with PD.