Use of diffusion tensor imaging to assess the impact of normobaric hyperoxia within at-risk pericontusional tissue after traumatic brain injury

Abstract

Ischemia and metabolic dysfunction remain important causes of neuronal loss after head injury, and we have shown that normobaric hyperoxia may rescue such metabolic compromise. This study examines the impact of hyperoxia within injured brain using diffusion tensor imaging (DTI). Fourteen patients underwent DTI at baseline and after 1 hour of 80% oxygen. Using the apparent diffusion coefficient (ADC) we assessed the impact of hyperoxia within contusions and a 1 cm border zone of normal appearing pericontusion, and within a rim of perilesional reduced ADC consistent with cytotoxic edema and metabolic compromise. Seven healthy volunteers underwent imaging at 21%, 60%, and 100% oxygen. In volunteers there was no ADC change with hyperoxia, and contusion and pericontusion ADC values were higher than volunteers (P<0.01). There was no ADC change after hyperoxia within contusion, but an increase within pericontusion (P<0.05). We identified a rim of perilesional cytotoxic edema in 13 patients, and hyperoxia resulted in an ADC increase towards normal (P=0.02). We demonstrate that hyperoxia may result in benefit within the perilesional rim of cytotoxic edema. Future studies should address whether a longer period of hyperoxia has a favorable impact on the evolution of tissue injury.

Figure 1

Lesion regions of interest. Fluid-attenuated inversion recovery (FLAIR) and apparent diffusion coefficient (ADC) images with lesion core (red), contusion (green), and perilesion (yellow) identified on a single axial slice.

Figure 2

Traumatic penumbra. Fluid-attenuated inversion recovery (FLAIR), gradient echo, and apparent diffusion coefficient (ADC) images at normoxia and hyperoxia demonstrating contusions within the left frontal and temporal parietal regions. These lesions have a hemorrhagic core shown by low signal on the gradient echo corresponding to the presence of blood degradation products, surrounded by a region of ‘vasogenic edema' with high signal on FLAIR and ADC. Around these lesions is a hypointense rim consistent with ‘cytotoxic edema', an example of which is shown at higher magnification and identified by the arrows. The final image has a color map showing the ADC increase calculated from the difference between the ADC images after hyperoxia. This highlights that the increase in ADC occurs predominantly within this border zone immediately surrounding the contusions.

Figure 3

Lesion-based analysis. Apparent diffusion coefficient (ADC) within brain tissue identified as contusion (A) and pericontusion (B) at baseline and after normobaric hyperoxia. The shaded gray box represents the 95% confidence interval for healthy controls from a region of mixed gray and white matter.

Figure 4

Impact of hyperoxia within traumatic penumbra. Changes in apparent diffusion coefficient (ADC) for the rim of cytotoxic edema surrounding visible brain lesions in 13 subjects. The shaded gray box represents the 99% confidence interval for healthy controls from a region of mixed gray and white matter.