Brain glucose metabolism in patients with traumatic brain injury undergoing rehabilitation: a longitudinal 18F-FDG PET study.

Abstract

Background: Measuring 18F-FDG PET-detected brain glucose uptake provides reliable information on metabolic tissue abnormalities, cells dysfunction, and neurovascular changes after traumatic brain injury (TBI).

Objectives: We aimed to study the relationship between post-traumatic brain glucose metabolism and functional outcomes in the so far unexplored field of longitudinally 18F-FDG PET-monitored patients undergoing rehabilitation after moderate-to-severe TBI.

Methods: Fourteen patients consecutively admitted to our unit in the post-acute phase after TBI underwent 18F-FDG-PET scans performed before and 6 months after inpatient rehabilitation program. The Glasgow Coma Scale (GCS) for neurological status, and the Functional Independence Measure (FIM) plus the Glasgow Outcome Scale-Extended (GOSE) scales for the rehabilitation outcome, were applied on admission and discharge. Voxel-wise analyses were performed, with the Statistical Parametric Mapping (SPM12) software, to investigate pre- vs. post-rehabilitation changes of brain metabolism, and their relationships with clinical indices.

Results: In the whole sample, 18F-FDG uptake significantly increased in the following five regions that were hypometabolic before rehabilitation: inferior frontal gyrus bilaterally, alongside right precentral gyrus, inferior parietal lobule, and cerebellum. However, only for the right precentral gyrus the median voxel peak-value at baseline resulted a significant predictor of both cognitive (FIM cognitive subscale, p = 0.012), and functional (GOS-E, p = 0.02; post- vs. pre-treatment GOS-E difference, p = 0.009) improvements. ROC curve analysis showed that a peak voxel-value of 1.7998 was the optimal cut-off for favorable rehabilitation outcome. Unfavorable functional outcomes were predicted by increased 18F-FDG uptake in the inferior frontal gyrus (GOS-E, p = 0.032) and precentral gyrus (FIM cognitive subscale, p = 0.017; GOS-E, p = 0.015).

Conclusion: This proof-of-principle study enlightens the metabolic changes occurring in moderate-to-severe TBI course. Notably, such changes preferentially involve definite frontal brain areas regardless of TBI localization and entity. These findings pave the way for further studies with translational purposes.