Acute changes in regional cerebral (18)F-FDG kinetics in patients with traumatic brain injury

During the acute phase after traumatic brain injury (TBI), the metabolic state is regionally heterogeneous. The purpose of this study was to characterize contusional, pericontusional, and remote regions of TBI by estimating glucose transporter and hexokinase activities on the basis of (18)F-FDG kinetic modeling.

A standard 2-compartment model was used to measure (18)F-FDG kinetic parameters in 21 TBI patients with cerebral contusions studied during the acute phase (3.1 +/- 2.1 [mean +/- SD] d after injury). Nineteen patients also underwent (15)O-water PET to measure regional cerebral blood flow (CBF). A control study ((18)F-FDG and (15)O-water) was done with 18 healthy volunteers. The rate constants K(i), K(1), and k(3) were assumed to represent the uptake, transport, and hexokinase activity of (18)F-FDG, respectively; K(i) was calculated as K(1) x [k(3)/(k(2) + k(3))].

The areas of contusional and pericontusional tissues located 4.5, 13.5, and 22.5 mm away from the contusion (PC(4.5), PC(13.5), and PC(22.5), respectively) demonstrated significantly reduced K(1) values, whereas the K(1) values for remote areas remained normal. The k(3) values were significantly reduced regardless of the distance from the contusion. Pericontusional areas with CT- or MRI-evidenced tissue damage showed significantly lower K(i) (P < 0.001), CBF (P < 0.01), and K(1) (P < 0.0001) values than did areas without such damage, whereas the k(3) values did not differ significantly. Seven patients showed regionally increased (18)F-FDG uptake (hot spots) in pericontusional areas. The k(3) value for the hot spots (0.086 +/- 0.024/min) was significantly higher than that for the remote cortex (P < 0.01), whereas the K(i), CBF, and K(1) values did not show significant differences. Patients with hot spots showed significantly higher K(i) and k(3) values in PC(4.5) (P < 0.05) and higher k(3) values in PC(22.5) (P < 0.05) than did patients without hot spots, whereas the K(1) and CBF values did not differ significantly.

Brain tissue (18)F-FDG kinetics in TBI patients were consistent with reduced hexokinase activity in the whole brain (including apparently uninjured cortex), whereas glucose transport was impaired only in the area immediately around the contusion. Pericontusional high levels of (18)F-FDG uptake observed in a subgroup of patients could have been the result of regionally increased hexokinase activity.