Statistical parametric mapping and cluster counting analysis of [18F] FDG-PET imaging in traumatic brain injury

In this study we investigated regional cerebral glucose metabolism abnormalities of [(18)F] fluorodeoxyglucose (FDG) positron emission tomography (PET) imaging in traumatic brain injury (TBI). PET images of 81 TBI patients and 68 normal controls were acquired and a word list learning task was administered during the uptake period. The TBI group included 35 patients with positive structural imaging (CT or MRI) findings soon after injury, 40 patients with negative findings, and 6 cases without structural imaging. Statistical parametric mapping (SPM) analysis was applied with several levels of spatial smoothing. Cluster counting analysis was performed for each subject to identify abnormal clusters with contiguous voxel values that deviated by two standard deviations or more from the mean of the normal controls, and to count the number of clusters in 10 size categories. SPM maps demonstrated that the 81 patients had significantly lower FDG uptake than normal controls, widely across the cortex (including bilateral frontal and temporal regions), and in the thalamus. Cluster counting results indicated that TBI patients had a higher proportion of larger clusters than controls. These large low-FDG-uptake clusters of the TBI patients were closer to the brain edge than those of controls. These results suggest that deficits of cerebral metabolism in TBI are spread over multiple brain areas, that they are closer to the cortical surface than clusters in controls, and that group spatial patterns of abnormal cerebral metabolism may be similar in TBI patients with cognitive deficits with and without obvious acute abnormalities identified on structural imaging.