Date of Award
Doctor of Philosophy (PhD)
In this thesis, a method that extracts the time course of radioactivity concentration in the blood from brain PET images is developed as a non-invasive alternative to arterial blood sampling. Even though the methodology is developed for 6-[18 F]fluoro-L-m -tyrosine (Fm T), a presynaptic dopaminergic radiotracer, it is generalisable to other radiotracers. The method is based on predefined regions of interest (ROI) drawn on the largest cerebral blood vessels, the venous sinuses, which are visible in the PET images. The time course of radioactivity in the cerebral blood was initially corrected for partial volume and spillover using a population value for the calibre of the blood vessel. The method was refined by parametrising the ROIs in order to extract the calibre of the blood vessel directly from PET images and to simultaneously correct the radioactivity concentration for partial volume and spillover. In a validation study, the radioactivity concentration was recovered to 100 ± 4% in syringes filled with an 11 C solution and inserted into a water-filled cylindrical phantom. Even though the diameter of syringes was estimated with an accuracy of half a pixel (1mm) in the phantom studies, the method systematically overestimated the blood vessel calibre by 2-3mm compared with measurements made in magnetic resonance venograms in human studies. The application of the parametric ROIs method to clinical studies awaits the development of more accurate scatter correction methods and the implementation of correction for head motion. The between-subject variations in the blood vessel calibre were measured to be comparable to the bias given by the parametric ROIs, thereby justifying the use of a population value over current subject-specific values. A compartmental model relating the radiotracer Fm T, its main radiolabelled metabolite and their exchange between plasma and erythrocytes was also developed in order to transform the time course of total radioactivity measured in whole blood into the time course of the unmetabolised radiotracer in plasma. Presynaptic dopamine function is severely disturbed in Parkinson's disease (PD) with a pattern of striatal involvement (posterior putamen dopamine levels decreased compared to that in caudate nucleus on the side opposite to the clinically impaired limbs) that distinguishes it from other movement disorders. Compared to arterial blood sampling, the image-derived plasma input function enabled the discrimination of normal and PD subjects as well as the identification of the affected and unaffected sides of the PD subject. The image-derived plasma input function also classified 20 of the 21 consecutive patients suspected of suffering from a variety of movement disorders into the same categories as determined using an indirect cerebellar input function. Using image-derived input functions, quantitative analyses of PET/Fm T studies are feasible in a clinical setting and, in combination with striatal patterns of Fm T uptake, the method provides useful diagnostic information in individual patients. (Abstract shortened by UMI.)
Asselin, Marie-Claude, "Image-derived plasma input function for the quantification of positron tomography brain studies with 6-[fluorine-18]-fluoro-L-meta-tyrosine" (2004). Open Access Dissertations and Theses. Paper 2755.