0902

Diffusion and quantitative MR changes in normal appearing brain following radiotherapy.
Felix Raschke1, Tim Wesemann2, Hannes Wahl2, Steffen Appold3, Mechthild Krause1,3,4,5,6, Jennifer Linn2, and Esther G. C. Troost1,3,4,5,6

1Institute of Radiooncology - OncoRay, Helmholtz-Zentrum Dresden-Rossendorf, Rossendorf, Germany, 2Institute of Neuroradiology, University Hospital Carl Gustav Carus and Medical Faculty of Technische Universität Dresden, Dresden, Germany, 3Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany, 4OncoRay - National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany, 5National Center for Tumor Diseases (NCT), Partner Site Dresden, Germany: German Cancer Research Center (DKFZ), Heidelberg, Germany; Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany, and; Helmholtz Association / Helmholtz-Zentrum Dresden - Rossendorf (HZDR), Dresden, Germany, 6German Cancer Consortium (DKTK), Partner Site Dresden, and German Cancer Research Center (DKFZ), Heidelberg, Germany

Synopsis

Irradiation of gliomas inevitably involves irradiation of surrounding normal appearing brain. We analysed longitudinal, quantitative MR data of 24 glioma patients before and at 3, 6, 9 and 12 months after radiotherapy and found significant reductions in mean-, axial- and radial diffusivity as well as in T2* in normal appearing white matter. These changes are greater the higher the received dose and progress over time. The diffusion reductions point towards axonal swelling. T2* reductions indicate either increased tissue heterogeneity, e.g. due to microglial activation or changes in tissue oxygenation, e.g. due to vascular alterations.

Introduction

Radiotherapy (RTx) is part of the standard treatment of gliomas. Radiation dose to surrounding normal appearing brain (NAB) is inevitably delivered to account for microscopic tumour extension of often infiltrative and heterogeneous gliomas and to compensate for systematic and random positioning uncertainties. The aim of this study was to use diffusion tensor imaging (DTI) and quantitative MRI to investigate dose and time dependent changes in NAB following RTx.

Methods

Data was collected as part of an ongoing, longitudinal study in accordance with local ethics procedures. All patients underwent primary gross tumour resection followed by RTx. 24 glioma patients (3 grade II, 13 grade III, 8 grade IV, age 48.4y ± 13.4y) had a pre-RTx MRI and at least one follow-up MRI after 3, 6, 9 and/or 12 months available. 13 patients received adjuvant chemotherapy during or after RTx. Four and 19 patients were treated with photons and protons, respectively, whereas one patient received a mixed treatment.

All MRI data was acquired on a 3T Philips Ingenuity PET/MR scanner (Philips, Eindhoven, The Netherlands) using an 8 channel head coil and included: DTI (TR/TE=6500/66ms, 2×2×2mm³, 32 directions, b=1000mm/s²), T1 and quantitative proton density (PD) mapping (3D-GRE, TR/TE=10/3.7ms, FA=3°/20°, 1×1×1mm³)1, T2* mapping (2D-GRE TR=1355ms, TEs=5.8/9.1/12.4/15.8/19.1ms, 2×2×2mm³), clinical 3D FLAIR (TR/TE/TI=4800/293/1650ms). DTI data was processed with FSL top-up2,3 to correct for geometric distortions and fractional anisotropy (FA), mean diffusivity (MD), q, axial diffusivity (AD) and radial diffusivity (RD) were subsequently calculated4.

Corresponding CTs used for RTx planning, dose maps and clinical target volume (CTV) contours were coregistered to the MRIs using ANTs5 (Figure 1). The T1 map was segmented into grey matter (GM), white matter (WM) and cerebrospinal fluid (CSF) using SPM126. The resulting tissue probability maps were rigidly coregistered to the DTI images and T2* map (Figure 2).

Mean signal intensities were extracted in GM (T2*, MD, T1, PD) and WM (T2*, T1, PD, MD, FA, q, AD, RD) for each dose region [0-3Gy, 3-10Gy, 10-20Gy, 20-30Gy, 30-40Gy, 40-50Gy, >50Gy] using a tissue probability threshold of 0.95. Only voxels outside the CTV and any additional abnormality seen on FLAIR imaging were considered. The relative signal change between follow-up and baseline was calculated and evaluated using a paired t-test with a Bonferroni correction factor of 12.

To rule out instrumental bias, six healthy controls (age 39.4y ± 8.7y) underwent two identical MRI scans with an interval of 3.6±0.5 months. Dose maps from the patients were warped to the healthy individuals’ MRIs and relative signal changes were calculated as for the patients.

Results

Relative mean MR changes within the different dose bins in WM are shown in Figure 3A. MD, AD, RD and T2* show signal reductions which increase with dose and over time. FA shows a trend for increase, whereas the anisotropy measure q demonstrates no dose dependent changes. Figure 3C shows that relative signal changes progress over time with no sign of recovery within the observed timeframe. No significant changes were seen in GM (Figure 4). No dose dependant changes were observed in the control cohort (Figure 3B/4B).

Discussion

Previous studies have both found decrease7,8 and increase8 in FA in NAB following RTx. Similarly both increase8,9 and decrease8,10 in mean-, radial- and axial diffusion were previously reported.

We found a trend for FA increase, but this is likely driven by the reduction of MD rather than a real increase in tissue anisotropy. This is supported by the fact that the anisotropy measure q, which isn’t normalised by the mean diffusion4, does not increase. It is therefore unlikely that demyelination, at least at an advanced stage, is a major process within our patient cohort. This is supported by non-significant changes of T1, a sensitive marker for myelin11.

Instead, the observed reductions in MD, RD and AD point towards a reduction in extra cellular space. This may be due to axonal swelling, as commonly seen in stroke12. Resolving oedema can be ruled out as a cause due to stable PD values.

Decrease in T2* can indicate increased tissue heterogeneity, e.g. caused by microglial activation, or changes in tissue oxygenation related to vascular changes.

Conclusion

Radiation effects on NAB can be quantified in vivo using quantitative MRI and the magnitude of change increases with dose and time. Further work is needed to understand the underlying microstructural changes using perfusion imaging and MRS data to investigate vascular and/or inflammatory alterations. A substantially larger dataset is compulsory for multivariate analysis of the effect of photon versus proton radiotherapy and of chemotherapy.

Acknowledgements

We grateful to all patients for participating in this study. This work was partly funded by the National Center for Tumor Diseases (NCT) and the National Center for Radiation Research in Oncology (OncoRay).

References

1. Volz S, et al. Correction of systematic errors in quantitative proton density mapping. Magn Reson Med 2012;68:74-85

2. Andersson J L R, et al. How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging. Neuroimage 2003;20:870-888

3. Smith S M, et al. Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage 2004;23 Suppl 1:S208-S219

4. Peña A, et al. Enhanced visualization and quantification of magnetic resonance diffusion tensor imaging using the p:q tensor decomposition. Br J Radiol 2006;79:101-109

5. Avants B B, et al. The Insight ToolKit image registration framework. Frontiers in Neuroinformatics 2014;8:44

6. Statistical Parametric Mapping: https://www.fil.ion.ucl.ac.uk/spm/

7. Chapman C H, et al. Regional variation in brain white matter diffusion index changes following chemoradiotherapy: a prospective study using tract-based spatial statistics. PloS one 2013;8:e57768

8. Connor M, et al. Regional susceptibility to dose-dependent white matter damage after brain radiotherapy. Radiology and Oncology 2017;123:209-217

9. Nagesh V et al. Radiation-induced changes in normal-appearing white matter in patients with cerebral tumors: a diffusion tensor imaging study. Int J Radiat Oncol Biol Phys 2008;70:1002-1010

10. Zhu T et al. Effect of the Maximum Dose on White Matter Fiber Bundles Using Longitudinal Diffusion Tensor Imaging. International Journal of Radiation Oncology, Biology, Physics 2016;96:696-705

11. Harkins K D et al. The microstructural correlates of T1 in white matter. Magn Reson Med 2016;75:1341-1345

12. Fung S H, et al. MR diffusion imaging in ischemic stroke. Neuroimaging Clin N Am. 2011;21(2):345-77

Figures

Figure 1: Planning CT and corresponding T1w images for a patient with resected grade III anaplastic astrocytoma. The planning CT, corresponding dose map and CTV contour (green line) were rigidly coregistered to the pre-RTx T1w MRI and nonlinearly warped to the post-RTx time points to compensate for tissue relaxation and deformation following RTx as seen by the expansion of the lateral ventricles.

Figure 2: Complete set of example images from a patient with a resected grade III anaplastic astrocytoma before the start of radiotherapy. GM and WM masks were obtained by SPM12 segmentation of the T1 map. The GM and WM masks were coregistered to the DTI b0 images and T2* map using a rigid body transformation and nearest neighbour interpolation. A subsequent probability threshold of 0.95 is chosen to binarize the GM and WM maps. Abnormal tissue has been manually masked out prior to segmentation.

Figure 3: (A) Mean relative signal changes and standard deviation in WM after 3, 6, 9 and 12 months after RTx compared to baseline. Blue bars indicate no significant change, orange bars correspond to p<0.05 and red bars correspond to a Bonferroni corrected p-value p<0.0042. (B) Relative signal changes in HCs. (C) Dose dependent time course of the mean relative signal changes in the patient cohort.

Figure 4: As Figure 3 but for GM.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
0902