As Positron Emission Tomography (PET) plays an increasing role in tumour diagnosis and research, a good image quality is crucial, which implies, among other properties, a good spatial resolution.
Thus, for this thesis, the spatial resolution was determined with a 3D–printed hot rod phantom with rod diameters between 3mmand 5 mm. As influencing parameters, gaussian filtering, Time of Flight (TOF) reconstruction, reconstructed measurement time and the reconstruction algorithms Filtered Backprojection (FBP) and two variations of the Ordered–Subset Expectation Maximisation (OSEM) algorithm were investigated. Furthermore, the influence of the radionuclides Fluorine–18, Gallium–68 and Copper–64 was examined, as well as the choice of scanner with the options Siemens Biograph mCT (PET/CT) and Siemens Biograph mMR (PET/MR). For comparison, several approaches were applied for the determination of spatial resolution: a visual and a mathematical approach, line profile evaluation and the determination of Full Width at Half Maximum (FWHM).
The results for the variation of most parameters lay within a 10 % deviation of the spatial resolution of 4.5mm given by Siemens, or were even better. The values for spatial resolution gained from the visual evaluation lay between 3.8mm and 5mm for F–18 measurements at the PET/CT, and were 3.8mm for the PET/MR. The spatial resolution of Ga–68 measurements at the PET/CT was lower, namely between 4.6mm and more than 5 mm.
FWHM determination was discarded as not practical for the evaluation of the acquired data, whereas the mathematical and the visual approaches confirmed the expected outcome. The measurement setup can be improved by ensuring a homogeneous activity distribution. PET/MR was found to be an interesting alternative to PET/CT; however, there is still room for improvement, particularly in the fields of attenuation correction and time resolution of detectors.