Multiple system atrophy (MSA) and progressive supranuclear palsy (PSP) are the most common atypical parkinsonisms. They are neurodegenerative diseases, altering the metabolic brain activity. Because of similar early simptoms among parkinosnisms, clinical diagnosis is wrong in 25 % of cases.
Scaled subprofile model/principle component analysis (SSM/PCA) is a statistical method for identification of specific metabolic brain networks, which are suitable biomarker for diseases where metabolic brain activity is altered. SSM/PCA can be used with different imaging modalities, which are sensitive to brain metabolism, such as $^{18}$F-fluorodeoxy glucose positron emission tomography (FDG-PET), functional magnetic resonance imaging data (fMRI), resting state fMRI (rs-fMRI), voxel based morphometry structural MRI data (VBM MRI), arterial spin labeling MRI perfusion imaging (ASL), H$_2$$^{15}$O positron emission tomography (H$_2$$^{15}$O-PET) and 99mTc-ethylcysteinate dimer single photon emission computed tomography (99mTc-ECD SPECT). So far metabolic brain networks were found to be a sucessful biomarker for Parkinson's disease (PD), other parkinsonisms - MSA, PSP and corticobasal degeneration (CBD) and Alheimer's disease (AD).
The purpose of this master thesis is to use equal protocol for identification of metabolic brain networks for MSA and PSP with FDG-PET. The thesis gives a simplified description of FDG-PET imaging and statistical methods. All steps of protocol for biomarker identification are introduced, from selection of appropriate groups for pattern identification, image preprocessing and pattern identification and validation. Tests of pattern stability with respect to critical protocol parameters will be described. The second part of the thesis is the protocol application on slovenian identification groups with MSA and PSP, with resulting biomarker patterns: multiple system atrophy related pattern (MSARP) and progressive supranuclear palsy related pattern (PSPRP).
|
