Purpose

With the prognosis for childhood brain cancer patients constantly improving, renewed attention has been placed on the long-term effects of radiotherapy, such as cognitive impairment and white matter necrosis¹. To develop treatments for this radiation-induced brain injury (RIBI), animal models are critical to understand its pathological mechanism. Although prevalently used, mouse models do not replicate the MRI changes observed clinically because their brain structures are not similar to human brain structures. We chose to develop a pig model because its brain structures are more comparable to human brain structures and can therefore be used to accurately model the development of RIBI.

Methods

Two male 3-month old Yucatan mini-pigs were used to create our swine model. The animals were treated in a manner consistent with pediatric brain tumor patients. Intensity modulated radiation therapy (IMRT) was utilized for the radiation treatment planning based on CT images (GE Light Speed VCT, GE Medical Systems, Milwaukee, WI). A single-fraction dose of 25 Gy was delivered with a clinical 6 MV linear accelerator (Varian Medical Systems, Palo Alto, CA) to the left hemisphere of each mini-pig brain (not including the brain stem), while the right hemisphere was kept unirradiated as a control. MR images of the brain were acquired 1 week before irradiation, 3 months post-irradiation, and 4 months post-irradiation on a 3T Siemens MAGNETOM Prisma MRI scanner. T1-weighted and T2-weighted images were first acquired using a 3D MP-RAGE sequence and 3D FSE sequence, respectively. MR spectroscopy data was acquired using semi-LASER localization with the acquired T1-weighted and T2-weighted images as a reference for voxel placement. The mini-pig was then given an intravenous injection of 0.2 mL/kg of MultiHance, after which diffusion-weighted images were acquired using a SE EPI sequence to allow the contrast enough time to diffuse. Finally, we ran MP-RAGE again to acquire T1-weighted post-contrast images.

Results

Behavioral and neurological deficits (left head tilt and left circling) were observed as early as 2 months post-irradiation. There was observable pathology on MRI at our 3-month scan which matched the severity of neurological impairment. These neurological deficits progressed over time, leading to balance loss and inability to stand which ultimately led to early euthanasia of both pigs. After euthanizing one mini-pig at 3 months post-irradiation and the other at 4 months post-irradiation, the brains were processed for histological analysis. MRI and DWI scans of the last surviving mini-pig taken at 3 and 4 months post-irradiation are shown in Figure 1 and 2, respectively. An MRS spectrum of the same animal at 3 months post-irradiation is displayed in Figure 3.

Discussion

The progression of the morphological changes seen in our model is comparable to the disease development seen in humans after receiving radiotherapy for head-related cancers². White matter lesions were detected at 3 months post-irradiation (Figures 1a-c) and characterized by increased signal intensity in the T2-weighted image with no contrast enhancement in the T1-weighted post-contrast image³. An area of low signal intensity could be seen in the left hemisphere of the T1-weighted image that was not present in the contralateral hemisphere. This decrease in signal intensity could be interpreted as increased fluid due to edema. To support this finding, ADC and FA maps (Figures 1d-e) showed increased diffusion and decreased anisotropy, respectively. An overall midline shift to the right could be seen, assuming due to the edema. By 4 months post-irradiation, areas of radiation necrosis have also begun to form alongside the white matter lesions (Figures 2a-c). Necrotic areas were characterized by increased signal intensity in the T2-weighted image along with contrast enhancement in the T1-weighted post-contrast image. The edema previously observed at 3 months post-irradiation has worsened over the time span of a month, as indicated by the ADC and FA maps (Figures 2d-e), which further intensified the midline shift. An interesting discovery from our 3-month MRS scan of the irradiated hemisphere (Figure 3) was the detection of a lactate peak along with a decreased N-acetyl aspartate peak in relation to creatine. While this evidence typically points to infarct tissue⁴, the lactate concentration measured in our spectrum is not significant enough to conclude the presence of infarction, but rather suggests this region has sustained vascular injury. Additional optimization of the semi-LASER sequence is needed to improve SNR in our spectra to confirm this finding. Further work also needs to be done in determining the appropriate dose to specifically induce the long-term effects of radiation injury. Nevertheless, our preliminary data suggests that a swine model can replicate the development RIBI similarly to the clinical disease seen in humans.

Acknowledgements

We want to acknowledge the support of the Pre-Clinical Research Laboratory (particularly Robyn McCain and Christa Crain) and the Purdue Life Science MRI facility (particularly Xiaopeng Zhou) for their help in performing all the procedures.

References

[1] Greene-Schloesser et al. “Radiation-Induced Brain Injury: A Review.” Frontiers in Oncology 2 (2012). [2] Lee et al. “Radiation-Induced Brain Injury: Retrospective Analysis of Twelve Pathologically Proven Cases.” Radiation Oncology Journal 29, no. 3 (2011). [3] Wang et al. “Evolution of Radiation-Induced Brain Injury: MR Imaging–Based Study.” Radiology 254, no. 1 (2010). [4] Saunders, Dawn E. “MR Spectroscopy in Stroke.” British Medical Bulletin, no. 2 (2000).