Application of the software tools developed by PVEOut to the data sets available within the consortium have provided preliminary results in several pathologies. The implementation of a multiparametric segmentation technique capable of segmenting multiple sclerosis lesions, beside the apparently normal brain tissues, was applied to 50 patients with Relapsing-remitting multiple sclerosis (RR-MS) and to 54 NV, allowed to demonstrate that brain atrophy in RR-MS is mainly related to GM loss, and correlating to faWM [Quarantelli M, et al. Neuroimage 2003;18:360]. A similar analysis carried out in 6 patients with CADASIL (Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy) [Quarantelli M, et al. ISMRM Proceedings 2001, abs. 1452] demonstrated the applicability of the same techniques to the analysis of brain MRI from patients with cerebrovascular disorders and the lack of correlation between brain atrophy and ischemic lesion load in CADASIL.
The application of the methods for brain MRI segmentation and automated volume of interest (VOI) definition to MRI studies of 25 normal volunteers (NV), 14 patients with deficit schizophrenia (DS) and 14 with non-deficit schizophrenia (NDS), allowed to confirm the atrophy in schizophrenic patients, and to highlight structural greater structural brain abnormalities in NDS, adding to the evidence that deficit schizophrenia does not represent just the more severe end of a schizophrenia continuum [Quarantelli M , et al. NeuroImage 2002;17:373].
While the conventional analysis of SPET CBF studies from 8 patients with fronto-temporal dementia (FTD) and of 21 AD patients confirmed the preferential involvement of the frontotemporal regions in FTD patients and of the temporoparietal regions in AD patients [Varrone A, et al. Eur J Nucl Med Mol Imaging 2002;29:1447], an analysis with PVE-correction of 10 AD patients compared to 12 subjects with Mild Cognitive Impairment (MCI) [Quarantelli M, et al. Proc. of the 2004 meeting of the Organisation for Human Brain Mapping], beside confirming the involvement of temporal lobes and fronto-parietal association cortex in AD, showed that only the decreases in posterior cingulate remained significant after PVE-correction. A similar finding was obtained when analysing FDG-PET studies from 23 mild AD patients and 13 aged NV [K Berkouk, et al. Proc. of the 2003 meeting of the Organisation for HBM; Abtract No. 1017].
Furthermore, an analysis with PVE-correction of SPECT studies of GABA-receptors (a synaptic marker) in 5 AD patients 8 MCI patients and 3 NV demonstrated, before PVE-correction, a significant fronto-temporo-parietal cortical reduction bilaterally, in AD vs. both MCI and NV, while after PVE-correction this decrease remained significant in posterior cingulate, suggesting that the reduction in cGABA-receptors in AD parallels but also exceeds structural changes as measured by MRI segmentation [Pappatà S, et al. Proc. of the 2004 meeting of the Organisation for Human Brain Mapping].
A novel method for analysing FDG uptake and underlying brain tissue volumes was also developed and applied to PET studies from 18 NV, showing a relative hyper-activity, as compared to GM volume, of most GM structures of the dorsal part of the brain, except the thalamus, and a relative hypo-activity, as compared to GM volume, of the hippocampus and the cerebellum, consistent with autoradiographic data from non-human primates [K. Berkouk, et al. Proc. of the 2002 meeting of the Organisation for Human Brain Mapping; Abstr. No. 10127].
Regarding Neuroreceptor PET studies, PVE correction software was also used to evaluate the impact of partial volume effect on the quantification of dopamine-D2 receptors with PET using the high-affinity radioligand [11C]FLB 457, and to identify those receptor-binding parameters most and least susceptible to PVE. Also the performances of different PVE correction algorithms were compared, to identify the most suitable for imaging of dopaminergic tracers.
This study confirmed that the regions most susceptible to PVE are the small or thin regions (amygdala and temporal cortex). Vd was much more underestimated (increased up to +55% when PVE correction was applied) without PVE correction than BP (changes with PVE-correction ranging between -10 and +10%). Comparison of different PVE-correction techniques implemented in PVELab proved the ROI method to be the most suitable technique for PVE correction in dopamine-receptor imaging.