The purpose of this study was to evaluate the feasibility of absolute quantification of regional cerebral glucose utilization (rCMR(glc)) in mice by use of (18)F-FDG and a small animal PET scanner. rCMR(glc) determined with (18)F-FDG PET was compared with values determined simultaneously by the autoradiographic 2-(14)C-DG method. In addition, we compared the rCMR(glc) values under isoflurane, ketamine and xylazine anesthesia, and awake states.
Methods: Immediately after injection of (18)F-FDG and 2-(14)C-DG into mice, timed arterial samples were drawn over 45 min to determine the time courses of (18)F-FDG and 2-(14)C-DG. Animals were euthanized at 45 min and their brain was imaged with the PET scanner. The brains were then processed for 2-(14)C-DG autoradiography. Regions of interest were manually placed over cortical regions on corresponding coronal (18)F-FDG PET and 2-(14)C-DG autoradiographic images. rCMR(glc) values were calculated for both tracers by the autoradiographic 2-(14)C-DG method with modifications for the different rate and lumped constants for the 2 tracers.
Results: Average rCMR(glc) values in cerebral cortex with (18)F-FDG PET under normoglycemic conditions (isoflurane and awake) were generally lower (by 8.3%) but strongly correlated with those of 2-(14)C-DG (r(2) = 0.95). On the other hand, under hyperglycemic conditions (ketamine/xylazine) average cortical rCMR(glc) values with (18)F-FDG PET were higher (by 17.3%) than those with 2-(14)C-DG. Values for rCMR(glc) and uptake (percentage injected dose per gram [%ID/g]) with (18)F-FDG PET were significantly lower under both isoflurane and ketamine/xylazine anesthesia than in the awake mice. However, the reductions of rCMR(glc) were markedly greater under isoflurane (by 57%) than under ketamine and xylazine (by 19%), whereas more marked reductions of %ID/g were observed with ketamine/xylazine (by 54%) than with isoflurane (by 37%). These reverse differences between isoflurane and ketamine/xylazine may be due to competitive effect of (18)F-FDG and glucose uptake to the brain under hyperglycemia.
Conclusion: We were able to obtain accurate absolute quantification of rCMR(glc) with mouse (18)F-FDG PET imaging as confirmed by concurrent use of the autoradiographic 2-(14)C-DG method. Underestimation of rCMR(glc) by (18)F-FDG in normoglycemic conditions may be due to partial-volume effects. Computation of rCMR(glc) from (18)F-FDG data in hyperglycemic animals may require, however, alternative rate and lumped constants for (18)F-FDG.