Dose-Dependent Cortical Thinning After Partial Brain Irradiation in High-Grade Glioma

Abstract

Purpose: Radiation-induced cognitive deficits may be mediated by tissue damage to cortical regions. Volumetric changes in cortex can be reliably measured using high-resolution magnetic resonance imaging (MRI). We used these methods to study the association between radiation therapy (RT) dose and change in cortical thickness in high-grade glioma (HGG) patients.

Methods and materials: We performed a voxel-wise analysis of MRI from 15 HGG patients who underwent fractionated partial brain RT. Three-dimensional MRI was acquired pre- and 1 year post RT. Cortex was parceled with well-validated segmentation software. Surgical cavities were censored. Each cortical voxel was assigned a change in cortical thickness between time points, RT dose value, and neuroanatomic label by lobe. Effects of dose, neuroanatomic location, age, and chemotherapy on cortical thickness were tested using linear mixed effects (LME) modeling.

Results: Cortical atrophy was seen after 1 year post RT with greater effects at higher doses. Estimates from LME modeling showed that cortical thickness decreased by -0.0033 mm (P<.001) for every 1-Gy increase in RT dose. Temporal and limbic cortex exhibited the largest changes in cortical thickness per Gy compared to that in other regions (P<.001). Age and chemotherapy were not significantly associated with change in cortical thickness.

Conclusions: We found dose-dependent thinning of the cerebral cortex, with varying neuroanatomical regional sensitivity, 1 year after fractionated partial brain RT. The magnitude of thinning parallels 1-year atrophy rates seen in neurodegenerative diseases and may contribute to cognitive decline following high-dose RT.

Figure 1

A) Cortical segmentation of left hemisphere overlaid on T1-weighted pre-contrast MR sagittal, axial, and coronal images. B) Neuroanatomical labels are shown in color on the left pial surface

Figure 2

Patient-specific (gray lines) and patient-averaged (blue diamond data points) 1-year change in cortical thickness for radiation doses from 0 to 60 Gy.

Figure 3

Relationship between 1-year change in cortical thickness and radiation dose for A) frontal, B) temporal, C) occipital, D) parietal, and E) limbic lobes. Error bars are standard errors. The solid line represents the result of the univariable regression. The limbic, temporal, and parietal lobes exhibited the steepest rate of change of cortical thickness with dose.

Figure 4

Cortical surface representation of a) radiation dose in Gy and b) cortical thinning at one year observed in an example patient. Regions receiving higher dose show greater degree of cortical thinning at 1-year post-RT.