C. Redies

Effect of Vascular Activity in the Determination of Rate Constants for the Uptake of 18F-Labeled 2-Fluoro-2-Deoxy-D-Glucose: Error Analysis and Normal Values in Older Subjects

Regional cerebral blood volume (CBV) can be calculated using data obtained during the kinetic analysis of 18F-labeled 2-fluoro-2-deoxy-d-glucose (FDG) uptake measured by positron emission tomography (PET). As a result the influence of vascular activity upon the determination of FDG rate constants can be minimized. The method is investigated by simulation experiments and by analysis of PET studies on seven older, healthy human volunteers aged 52–70 years. The accuracy of measured FDG rate constants k1, k2, and k3, obtained either by omitting the early portion of the uptake curve or by explicit inclusion of CBV as a fit parameter, is compared. The root mean square error in measured rate constant for the latter method is equivalent to that obtained by omitting the first 2.5–3 min of tissue data and neglecting the CBV term. Hence, added information about the physiological state of the tissue is obtained without compromising the accuracy of the (FDG) rate constant measurement. In hyperemic tissue the explicit determination of the vascular fraction results in more accurate estimates of the FDG rate constants. The ratio of CBV determined by this method to CBV obtained using C15O in six subjects with CBV in the normal range was 0.92 ± 0.32. A comparison of the CBV image obtained by this method with that obtained using C15O in an arteriovenous malformation case demonstrates the accuracy of the approach over a wide range of CBV values. The mean value for CBV fraction in gray matter obtained by this method in the older control group was 0.040 ± 0.014. Average gray matter rate constants obtained were k1 = 0.084 ± 0.012, k2 = 0.150 ± 0.071, and k3 = 0.099 ± 0.045 min−1.

In vivo measurement of [18F]fluorodeoxyglucose rate constants in rat brain by external coincidence counting.

The operational equation for the double-label deoxyglucose method described in the following paper requires the knowledge of the rate constants for transfer of fluorodeoxyglucose across the blood-brain barrier (K1* and K2*), and those for phosphorylation of fluorodeoxyglucose (K3*) and dephosphorylation of fluorodeoxyglucose-6-phosphate (k4*). These rate constants were determined in anesthetized rats by external coincidence counting. Radioactivity in parietal brain was measured for a 110 min experimental period after a bolus injection of 18F-labeled fluorodeoxyglucose. Apparent rate constants were obtained by fitting the resulting tissue radioactivity curves to the tissue radioactivity function of the deoxyglucose model modified to take into account the dephosphorylation of fluorodeoxyglucose-6-phosphate. The apparent fluorodeoxyglucose rate constants in rat brain are K1* = 0.195 ml g-1 min-1, k2* = 0.379 min-1, k3* = 0.088 min-1, and k4* = 0.009 min-1.

Estimation of cerebral oxygen utilization rate by single-bolus 15O2 inhalation and dynamic positron emission tomography.

This study shows that regional CMRO2 can be estimated by means of nonlinear regression using dynamic positron emission tomographic data acquired during 1 min following single-bolus inhalation of 15O2. The feasibility of simultaneous estimation of CBF, cerebral blood volume (CBV), oxygen extraction ratio (OER), and CMRO2 was assessed by simulations using the model of Mintun et al. Four oxygen metabolic measurements, each consisting of a CBF, CBV, and 15O2 bolus study, were carried out on three volunteers. Regional values for CBF, CBV, OER, and CMRO2 were derived in two ways: from the fits of the time-activity curves of the dynamic 15O2 bolus study alone [CMRO2(fit)] and from the three separate studies [CMRO2 (control)]. For the 56 regions of interest analyzed, using a fit interval of 60 s, CMRO2(fit) was 93.4 ± 7.8% of CMRO2(control) (mean ± SD) with a correlation coefficient of r = 0.95. CMRO2(control) ranged from 87 to 290 μmol/min/100 g. Individual simultaneous estimates of CBF, CBV, and OER were not reliable. Finally, we found that the validity of the model was limited in practice to the first minute after tracer inhalation.

Time-Dependent Changes of Lumped and Rate Constants in the Deoxyglucose Method in Experimental Cerebral Ischemia

Time-dependent changes in the lumped and rate constants in a bilateral middle cerebral artery (MCA) occlusion in cats were evaluated. These variables were measured in 11 cats after a sham operation, in five after a 1-h occlusion, in two after a 2-h occlusion, in five after a 4-h occlusion, and in four after a 16-h occlusion. The time course of the cerebral tissue radioactivity [Ci* (t)] was monitored by external coincidence counting during a programmed infusion of [18F]2-fluorodeoxyglucose (FDG). Arterial plasma concentration [Cp* (t)] of tracer was kept constant during the first 45 min. Comparison of k2* and k3* in the sham-operated group, estimated by external coincidence counting, and by the ratio of extraction fractions of glucose and [18F]2-FDG, demonstrated no significant difference between these rate constants in these two groups of animals. The rate and lumped constants were also estimated from Ci* (t) and Cp* (t), as well as from the ratio of extraction fractions of glucose and [18F]2-FDG, respectively, in the MCA occlusion group. Significant decrease in k3* was observed after 1 h of occlusion (20% lower than in the sham operation, p < 0.05); in ki* decrease occurred within 4 h of occlusion (21% lower than in the sham operation, p < 0.05). However, decrease in k2* was observed only after 16 h of occlusion (26% lower than in the sham operation, p < 0.05). Namely, decrease of rate constants occurred first in k3* then in k1* and k2*. In contrast to the rate constants where decrease was observed, a significant increase in the lumped constant occurred between 1 and 4 hours after occlusion (55% higher than in the 1-h occlusion, p < 0.05). However, the values of the lumped constant were not significantly different between the 4-h occlusion group and the 16-h occlusion group.

The Deoxyglucose Method in the Ferret Brain. I. Methodological Considerations

In the brain of the anesthetized ferret, the 2-deoxyglucose (2-DG) transfer rate constants required to determine cerebral glucose utilization by the deoxyglucose method were calculated from regional gray matter time-radioactivity curves measured for 180 min after tracer injection. Results suggest that loss of metabolized tracer from brain occurs at a rate of about 1%/min for the first 180 min after injection if the rate constant of the rate-limiting step for loss of metabolized tracer (k*4) represents a first-order kinetic process. A simulation experiment shows that, whether k*4 is assumed to be 0 or 0.01 min−1, has a negligible influence on glucose utilization rates obtained in conventional 45 min autoradiographic experiments provided that the entire analysis, including lumped constant determination, is carried out in a consistent way. The 2-DG lumped constant for k*4 = 0 is 0.54, and 0.68 for k*4 = 0.01 min−1.

The Deoxyglucose Method in the Ferret Brain. II. Glucose Utilization Images and Normal Values

To measure cerebral glucose utilization with the autoradiographic deoxyglucose method, the tracer transfer rate constants and lumped constants must be known. 2-Deoxyglucose (2-DG) and fluorodeoxyglucose (FDG) constants were determined in 18 gray and white matter brain structures of the anesthetized ferret. The ferret is a domestic carnivore particularly suitable for deoxyglucose studies because of its small brain size and low body weight. The average gray matter rate constants for tracer transfer across the blood-brain barrier are similar for 2-DG and FDG in the ferret brain (K*1 = 0.21 ml/g/min and k*2 = 0.39 min−1). The rate constant for the rate-limiting step of tracer phosphorylation, k*3, is 1.6 times higher for FDG than for 2-DG (0.21 vs. 0.13 min−1). Loss of metabolized tracer is about 1–1.5%/min throughout the ferret brain for both tracers as estimated for a 180 min experimental period. Taking into account this loss, the lumped constant is 0.92 for FDG and 0.68 for 2-DG. Glucose utilization values in the brain of the anesthesized ferret range from 33 μmol/100 g/min in the corpus callosum to 104 μmol/100 g/min in the caudate nucleus. Representative glucose utilization images of coronal sections of the ferret brain are shown. Brain structures are identified on the same slices counterstained with thionin.