Neurometabolic Disruption Following Sports Concussion Assessed with PET/MRI

Objectives: Animal studies have defined a neuropathological cascade following concussion consisting of large and transient disruptions of homeostasis in the acute phase, including glucose metabolism. Because this has never been observed in humans, our goal is to establish the existence, magnitude, and regional distribution of the fluctuations in glucose metabolism and related hemodynamic effects. A better understanding of the early pathology of concussion may lead to improved strategies for objective diagnosis and treatment.

Methods: Athletes from a single NCAA Division I university consented to undergo PET/MRI brain imaging in the event of concussion, using a within-subjects design in which subjects are imaged as soon as possible after injury and again when symptoms have resolved (>3 months later). We also recruited a non-injured control group from the same population. A 20 min PET/MRI scan commenced 40 min after i.v. injection of 3-5 mCi of FDG. MRI included standard structural scans and quantitative measures of cerebral blood flow (CBF) and cerebral metabolic rate of oxygen (CMRO2). PET images were registered to the Hammers brain atlas and regions of interest extracted. Clinical data regarding severity of injury and duration of recovery were collected as well.

Results: We have completed acute phase scans on 7 subjects within 31-94 hours of clinically diagnosed concussion injury (all male, ages 18-22). Three of these subjects have already had follow-up scans, with the remaining 4 subjects awaiting the recovery period. Additionally, 3 age- and sex-matched control athletes without injury were scanned. No clinically relevant functional or structural abnormalities were observed. Whole-brain SUV values were lower in acute phase compared to follow-up at 3 months (p=0.02, unpaired, n=7 acute, n=3 follow-up). Paired analysis (acute vs follow-up, n=3 for each) of individual regions normalized to whole brain indicated acute phase hypermetabolism in left hippocampus (a priori hypothesis, p=0.04). In post-hoc exploratory analysis, only hypometabolism was observed, also only on left side (left caudate nucleus, medial orbitofrontal gyrus and posterior cingulate gyrus) as well as both left and right posterior temporal cortex, although these did not survive familywise error correction (uncorrected p<0.05). Using MIM software, clinical reads of cortical surface maps, supported by semi-quantitative z score analysis, indicated mild medial temporal hypometabolism (primarily left side) in acute condition compared to follow-ups and controls, while quantitative analysis of these regions was not statistically significant. Comparing PET results to injury-relevant parameters, the ratio of acute SUV to follow-up SUV (in both gray matter GM and and white matter WM) decreased with delay from injury to scan time (n=3, R²> 0.99). The best correlation with clinical severity scores was the ratio of acute GM SUV normalized to WM to follow-up value of the same quantity, which also had a negative correlation (n=3, R²= 0.97-0.99).

Conclusions: Trends in PET suggest that acute phase (31-94 hours after injury) is hypometabolic across whole brain, with a similar trend in a small number of gray matter areas on the left side. The degree of hypometabolism (as assessed by acute/follow-up SUV) increases across this time window, suggesting that the lowest metabolic state occurs at even later times. The degree of hypometabolism (as assessed by acute/follow-up GM/WM ratio) may reflect clinical severity. MRI analysis of CBF and CMRO2 is ongoing. We acknowledge a FUSION Award from the Renaissance School of Medicine at Stony Brook University and the Office of the Vice President for Research and additional support from the Departments of Radiology and Orthopedics.

• Download figure
• Open in new tab
• Download powerpoint