Tc-99m ethylene cysteinate dimer SPECT in the differential diagnosis of parkinsonism.

Positron emission tomography (PET) and network analysis have been used to identify a reproducible pattern of regional metabolic covariation that is associated with idiopathic Parkinson's disease (PD). The activity of this PD-related pattern can be quantified in individual subjects and used to discriminate PD patients from atypical parkinsonians. Because PET is not commonly available, we sought to determine whether similar discrimination could be achieved using more routine single photon emission computed tomography (SPECT) perfusion methods. Twenty-three subjects with PD (age, 63 +/- 9 years), 22 subjects with multiple system atrophy (MSA; age, 64 +/- 7 years), and 20 age-matched healthy controls (age, 62 +/- 13 years) underwent SPECT imaging of regional cerebral perfusion with Tc-99m ethylene cysteinate dimer (ECD). Using network analysis, we determined whether a PD-related pattern existed in the SPECT data, and whether its expression discriminated PD from MSA patients. Additionally, we compared the accuracy of group discrimination achieved by this pattern with that of the PET-derived PD-related pattern applied to the SPECT data. Network analysis of the SPECT data identified a significant pattern characterized by relative increases in cerebellar, lentiform, and thalamic perfusion covarying with decrements in the frontal operculum and in the medial temporal cortex. Subject scores for this pattern discriminated PD patients from controls (P < 0.01) and from MSA patients (P < 0.03). Subject scores for the PET-derived PD-related pattern computed in the individual SPECT scans more accurately distinguished PD patients from controls (P < 0.005) and from MSA patients (P = 0.0002). A significant PD-related covariance pattern can be identified in SPECT perfusion data. Moreover, the disease related pattern identified previously with PET can be applied to individual SPECT perfusion scans to provide group discrimination between PD patients, healthy controls, and individuals with MSA. Because of significant individual subject overlap between groups, however, the clinical utility of this method in the differential diagnosis of Parkinsonism remains uncertain.