[Comparative analysis of quantifying methods in FDG-PET scans of the brain]

The comparative analysis of three tracer kinetic methods most frequently applied for the quantization of the results of FDG-PET-brain scans was performed. The data of measurements on five healthy patients were evaluated by the most general method developed by Phelps, the Patlak-procedure and the SUV- (standard uptake value) method. It was demonstrated by the aid of correlation analysis that the applicability of the results of the more simple methods to estimate glucose metabolic rate (GMR) as calculated by the Phelps-method depended on the kind of the selected region of the brain. It was shown that the most considerable distortion occurred in the case of the same anatomical regions of the brain with both simplified methods. These regions were located either in the white matter or in the vicinity of larger size blood vessels or they were elements of the base of the skull [gyrus rectus (l. u.), pons, capsula interna (l. u.), cerebellum (l. u.), corpus callosum]. The distorted estimation is explained by the fact that the simpler models neglect dephosphorylation of the FDG-6P, and they also disregard the contribution of the intravascular activity to the tissue radioactivity as determined by the relatively low resolution PET measurement. The correlation coefficient between the GMR as calculated by the Phelps-method and glucose consumption data by the investigated simpler methods had very low values for regions located in the white matter, eventual close to blood vessels or being elements of the base of the skull.