The Feasibility of Diffusion Tensor Imaging in Magnetic Resonance- guided Linear Accelerator for a Glioblastoma

Abstract

Background and aim: Glioblastoma (GBM) is the most malignant primary brain malignancy in adults, and radiotherapy (RT) can be one of the most crucial treatment options. However, RT has the potential to damage white matter (WM), resulting in cognitive and functional deficits. Although the introduction of the magnetic resonance-guided linear accelerator (MR-linac) promises high-precision RT coupled with real-time imaging, there is still a significant need for advanced imaging sequences to ensure the protection of WM tracts. Diffusion tensor imaging (DTI) provides insights into the microscopic characteristics of tissue cellularity and anisotropy, offering clear visualization of WM tracts. While DTI is promising, the integration of DTI within the MR-linac system for GBM treatment remains unexplored. This study aims to investigate the feasibility of DTI in MR- linac for GBM. Methods: This longitudinal prospective case study examined a GBM patient treated with radiotherapy via conventional linac. DTI images were acquired by the MR-linac imaging component at two predetermined timepoints after 8Gy and 20Gy radiation delivery. Using the 3D Slicer open-source platform, a directionally encoded colour map, fractional anisotropy (FA) scalar measurement, and tractography were generated to compare WM tracts in direct relation to the tumour with WM tracts on the unaffected contralateral side, along with their changes across second imaging sessions. Additionally, mean diffusivity (MD) scalar measurements were processed to evaluate tumour response between imaging sessions. Results: Analysis of DT images yielded an overall low image quality. However, these images can illuminate WM pathways. The unaffected contralateral hemispheres consistently exhibit normal WM tracts, whereas the affected side revealed anatomical alterations due to the tumour. Directionally encoded colour maps post-8Gy and 20Gy revealed a significant asymmetry between the cerebral hemispheres, with enhanced visibility post- 20Gy. Fractional anisotropy (FA) map images indicated compromised WM structural integrity near the tumour, with decreased FA values, notably after 20Gy of delivery. The mean diffusivity (MD) map images show no significant changes between the two time points, and the MD value for the tumour indicated a similar response. Conclusion: The incorporation of DTI into the MR-linac system could be demonstrated as feasible approach for GBM patients and other cerebral malignancies. This integration might be useful in obtaining more precise and targeted radiotherapy, with the potential to enhance treatment outcomes and reduce complications associated with RT.