Clinical correlation of multivariate spatial covariance analysis of 18F-FEOBV VAChT PET in Parkinson disease

Objectives We applied a multivariate spatial covariance analysis to ¹⁸F-FEOBV (ligand for the vesicular acetylcholine transporter [VAChT]) PET brain images of two groups of subjects (Parkinson Disease Patients, PD and Normal Controls, NC). The goal of our study was to investigate the clinical correlations of the resulting Subject Scaling Factor (SSF) scores based on the Spatial Covariance Pattern (SCP) in the PD patients.

Methods Multivariate spatial covariance analysis was performed on ¹⁸F-FEOBV PET brain images of 27 (PD, n=18 and NC, n=9) subjects. The analysis was based on Scaled Subprofile Model (SSM) using the generalized covariance analysis software suit (http://www.nitrc.org/projects/gcva_pca/). This resulted in an SCP, which successfully distinguished PD from NC subjects. The subject scaling factor (SSF) scores represent the extent to which each subject expresses the SCP. The PD subjects completed the Mini-BESTest (Mini Balance Evaluation Systems Test) and PDCRS (Parkinson’s disease cognitive rating scale) clinical tests. The correlation between the SSF scores and these clinical scores were calculated for these patients.

Results The most significant correlation coefficient between the SSF scores and the Mini-BESTest were for reactive postural control (r= -0.51, p = 0.045) and with the PDCRS for verbal learning (r=-0.49, p = 0.05). Also, the SSF scores showed a correlation of r = -0.48 with the average ¹⁸F-FEOBV values in the caudate nucleus.

Conclusions Multivariate spatial covariance analysis of ¹⁸F-FEOBV PET correlates with clinical deficits in postural control and cognition in PD. These data provide novel clinical information about impaired brain network functions of cholinergic terminals in PD. Research Support: NIH P50 NS091856