MRI-based analysis of volumetric changes of healthy brain tissue in glioma patients after photon radio(chemo)therapy

Motivation and Objective
State-of-the-art Linac-based photon beam irradiation achieves highly conformal target volume coverage in glioma patients, but is also known to cause side-effects to surrounding tissues and organs. Apart from subjective measures (e.g., questionnaires, function tests) objective means to quantify tissue damage, e.g., anatomical or functional magnetic resonance imaging (MRI) are urgently needed to compare different treatment techniques and beam qualities (e.g., protons vs. photons) and to develop predictive measures for optimal sparing of normal brain tissue. As initial part of our program for dose-dependent spatial mapping of structural and functional radiation induced brain damage, we assessed here a retrospectively collected MRI-dataset in order to potentially detect volumetric changes of the healthy brain tissue (gray and white matter) in the non-affected hemisphere of glioma patients treated with photon irradiation.

Material and Methods
Structural MRI-scans (T1-weighted) from 18 glioma patients (grade II and III), who underwent high dose radio(chemo)therapy (54-60 Gy) with curative intent have been analyzed. MRIs were acquired before treatment and at several time intervals thereafter. Because of the individual characteristics of these data e.g., voxel size (0.5…6 mm³) and the field strength (1…3 T) a standardized image processing approach was developed. For bias field correction, registration with atlas data, resampling, and segmentation of different tissue types, image processing methods from the ANTs-, FSL- and SPM-toolbox were used, respectively. Based on these images the volumes of white matter and gray matter have been longitudinally analyzed.

Results
Figure 1 shows the changes of brain tissue volume depicted as box plots with the median values highlighted in red. While the entire brain volume on average remains constant over two years after therapy, in the same time period the volume of gray matter and white matter varies conversely in a wide range. Noteworthy, this work points out the difficulties of retrospectively analyzing clinically acquired data due to differences in acquisition parameters and in investigation intervals.

Outlook
The observed changes over time underpin the importance of exact follow-up protocols in quantitative evaluation of structural brain changes after radiotherapy. Together with the data on interpatient heterogeneity, our findings allowed to design a prospective study in a larger cohort of patients treated by photons vs. protons for assessing the dependence of MRI-detected volumetric changes with delivered dose.