Abstract
1342
Objectives Partial-volume effects (PVEs) are a degrading phenomenon in emission tomography caused by limited spatial resolution. The objective of this work is to investigate the effects of data-driven partial-volume correction (PVC) on 123I-Ioflupane (DaTSCAN) SPECT images.
Methods Five normal controls underwent DaTSCAN SPECT (GE Infinia) and T1 MR. The DaTSCAN images were reconstructed with ordered-subset expectation maximisation (OSEM) and PV-corrected using reblurred Van-Cittert deconvolution. An isotropic 3-D point-spread function (PSF) of 8.6mm full-width at half maximum (FWHM) was assumed for the purpose of PVC. The DaTSCAN images were rigidly registered to the patient MR images using a block-matching approach. The striatal regions of interest (ROIs) were manually segmented using the patient MR data. Binding potential (BPND) values for the caudate and putamen were calculated using a template region in the posterior parietal lobe and also the segmented grey matter (GM) of the same region.
Results Partial-volume correction resulted in an average increase of BPND in the striatal regions of 15.66% (p=0.0035) when using the segmented grey matter (see Table rows 1 and 2) as a background region and 14.35% (p=0.013) using the template (see Table rows 3 and 4).
Conclusions These preliminary results suggest that PVC significantly increases BPND when using a segmented grey matter reference region. We now plan to apply these techniques to patients with movement disorders.
Research Support Benjamin Thomas is supported by a CASE studentship between GE Healthcare and the Engineering and Physical Sciences Research Council (EPSRC)
The effects of partial-volume correction and reference regions on binding potential.
CV = Coefficient of Variation, SD = Standard Deviation