Relative Hypermetabolism of Vermis Cerebelli in Traumatic Brain Injured Patients Studied with 18FDG PET: A Descriptor of Brain Damage and a Possible Predictor of Outcome

Abstract

The almost constant presence of apparent metabolic hypermetabolism of cerebellar vermis seen on 18FDG PET in a population of injured brains has been reported in a previous paper. Aim of this paper is to determine a) whether there is a correlation between the entity of this sign, semi quantitatively determined, and the severity of the trauma at its onset, and b) whether the entity of the relative enhancement correlates with the medium and long term clinical outcome. A group of 45 consecutive patients admitted to the Acquired Brain Injury Unit of our Hospital for recent, major head trauma, underwent a basal 18FDG PET/CT scan of the brain; the presence of relative hypermetabolism of the vermis cerebelli was visually assessed and semi quantitatively determined (vermis/cerebellum ratio: V/C); the median V/C value was used as a divide between low V/C ratios (group A) and high V/C ratios (group B). During one year after trauma, every patient from both groups received an extensive testing to evaluate cognitive and behavioral performances and evolution: Disability Rating Scale (DRS) and Levels of Cognitive Function (LCF) were administered monthly from month 1 to month 6, and at 12 months from the trauma; Glasgow Outcome Scale (GOS) was administered at 3, 6 and 12 months from the head trauma. Numerical scores from each of these performance-testing protocols were cross-matched with values derived from the V/C 18FDG PET/CT determinations. A relative risk estimate via Chi-square testing was performed on the results of both groups for LCF and DRS scales at 1, 6 and 12 months from trauma. At one month after trauma, overall LCF (LCF1) values ranged from 2 to 8, avg. 3.77, SD ± 2.10; the average value in group A was 5.21, SD ± 2.09, in B group 2.47, SD ± 0.98 (F=17.5, P = 0). At this time, overall average DRS (DRS1) was 6.7, SD ± 2.05, ranging from 2 to 9; the average value was 5.52, SD ± 0.47 in group A, and 7.72, SD ± 0.30 in group B (F = 6.3, P = 0.01). Relative risk estimates for patients with higher V/C ratios for poor performance in DRS scale were: 2.46 at 1 month (confidence boundaries 1.66 - 3.64), 3.75 (c.b. 1.64 - 8.64) at 6 months, 5.17 (c.b. 1.76 - 15.16) at one year. Relative risk estimates for LCF scale were: 3.20 (c.b. 1.74 - 5.90) at 1 month, 6.909 (c.b. 1.03 - 46.15) at 6 months, 4.22 (c.b. 0.65 - 27.10) at 12 months. A) there is a strong correlation between the semi quantitatively determined values of vermian relative hypermetabolism and the severity of trauma as determined by standard cognitive and performance testings; the V/C ratio may therefore be considered a reliable, although non-specific, index of brain suffering. B) there is a good statistical correlation between the semi quantitative vermian/cerebellar ratio determined shortly after the trauma, and the clinical outcome of the patients, evaluated by standard clinical performance tests and relative risk estimates.