Metabolic rates in small brain nuclei determined by high-resolution PET

Identification of small nuclei in the brain by PET has been limited by the spatial resolution of conventional scanners. The new detector technology and advanced signal analysis of a high-resolution research tomograph (HRRT) has improved 3-dimensional spatial resolution to 2.2 mm at sufficient efficiency and permitted the quantification of tracer concentrations in small volumes. Methods: In 9 healthy volunteers, cerebral glucose metabolism was investigated after intravenous injection of 370 MBq of 18F-FDG, and regional cerebral metabolic rates for glucose (rCMRGlc) were determined in various structures of the brain identified on coregistered MR images using stereotactic and topographic anatomic information. Results: rCMRGlc values (in μmol/100 g/min) were higher in the cerebral cortex (33.5 ± 2.98), the basal ganglia (32.6 ± 3.04 in the nucleus caudatus and 40.2 ± 3.50 in the putamen), the thalamus (36.6 ± 4.72), and the cerebellum (29.8 ± 2.20) and were lower in the cerebral white matter (12.3 ± 1.45) than those reported previously with conventional scanners. This resulted in an increased ratio of cortical values to white-matter values. Various nuclei in the basal frontal lobe (21.4 ± 3.19 in the basal forebrain and 32.3 ± 2.39 in the nucleus accumbens), the temporal lobe (22.2 ± 1.74 in the corpus amygdalae), the hippocampus (25.7 ± 2.11), the diencephalon (23.1 ± 3.33 in the corpus geniculatum laterale, 20.2 ± 2.87 in the corpus geniculatum mediale, and 25.2 ± 3.29 in the nucleus subthalamicus), and the brain stem (24.4 ± 2.47 in the colliculus superior, 31.4 ± 3.63 in the colliculus inferior, 31.0 ± 3.10 in the nucleus ruber, and 22.8 ± 2.35 in the substantia nigra) could be identified, and the metabolic rate was assessed in these structures. The effect of improved spatial resolution on quantified metabolic rates could directly be demonstrated in a few cases investigated on scanners of different generations. Conclusion: The improved spatial resolution of the HRRT decreased partial-volume effects in the quantification of metabolic rates in the brain and increased the accuracy of rCMRGlc values in large structures. For the first time, this scanner has permitted the determination of metabolic rates in small nuclei that are involved in various neurodegenerative disorders.