For further insight into the physiology and pathogenesis of the developing brain, quantification of the cerebral glucose metabolism is needed. Arterial blood sampling or sampling of great volumes of blood is not justified for the purpose of PET studies in children. Therefore, we have developed simplified PET approaches to analyze brain FDG examinations during infancy.
Methods: The study consisted of 18 FDG-PET examinations chosen from our research protocols concerning hypoxicischemic encephalopathy and severe neonatal hypoglycemia. The input function for graphical analysis according to Patlak was derived in two ways: (1) a combined time-activity curve derived from the left ventricular activity concentration (first 7-17 min of the study) and radioactivity concentration in venous whole-blood samples and; (2) activity concentration measured in whole-blood venous blood samples (arterial plasma in one case). As an alternative for semiquantitation, the standardized uptake values (SUV) were calculated and correlated to local cerebral metabolic rates for glucose (LCMRGlc).
Results: The influx rate constants (Ki) and LCMRGlc values obtained using the combined curve versus venous curve did not differ statistically (p > 0.05). There was a good correlation between the SUV and LCMRGlc values (r = 0.83, p < 0.001).
Conclusion: Local cerebral metabolic rates for glucose can be accurately calculated by using the combined curve (left ventricular activity concentration during first 5 min of the study and 2-3 venous whole-blood samples at the end of the study) for even the smallest pediatric patients. When blood samples cannot be obtained, SUV values provide an alternative for estimation of the cerebral glucose uptake and interindividual comparison of the patients.