Session Item

Radiotherapy (RT)-induced volume reduction in gray matter is associated with deterioration of brain function. We performed a large-scale quantitative measurement of volumetric change in gray matter and identified potential predictors of severe volume reduction after RT.

A single institution cohort of 461 magnetic resonance imaging (MRI) in 105 patients with glioma treated with postoperative RT was retrospectively analyzed. Study patients had MRIs at five timepoints (before and 1 month, 6 months, 1 year and 2 years after RT) and some were missing due to follow-up loss or death. Using FastSurfer – a deep learning based neuroimaging pipeline, a total of 81 subsites of brain were automatically segmented from longitudinal MRIs and their volumetric changes were calculated. Each subsite was classified into 10 functional fields. A linear mixed effects model was performed to identify potential predictors of volumetric changes, including gender, tumor grade, tumor location, RT dose, IDH-mutation, MGMT promoter methylation, 1p/19q co-deletion, progressive disease after RT, adjuvant chemotherapy, and interaction between patients’ age at RT and time interval after RT. Statistical significance was evaluated at a p-value <0.05.

The median volumetric change after RT was -5.1% (range, -37.9% to +4.1%) in overall subsites and the ipsilateral side with irradiated field showed greater volumetric change than contralateral side (ipsilateral side: -6.9% [range, -37.8% to +5.3%]; contralateral side: -2.8% [range, -33.3% to +2.6%], p<0.001). The median volume of hippocampus, caudate, and precentral gyrus were reduced by 5.5%, 12.0% and 4.5%, respectively. In the multivariate analysis, female (vs. male), longer time interval after RT, older age at RT, and progressive disease after RT were the best predictors for volume reduction. The interaction between age at RT and time interval after RT showed significant correlation with volumetric change, suggesting that the volumetric change over the time after RT was significant only if the age at RT was older than a specific subsite-dependent cutoff age (e.g., 45 years for ipsilateral hippocampus and 55 years for ipsilateral precentral area). Adjuvant chemotherapy and RT dose had little effect on volumetric change, affecting 4% of subsites. Regarding the functional fields, the areas of cognitive/emotional, language, and execution of movement showed the largest volume reduction and the coordinating area (cerebellum) showed the smallest volume reduction after RT.

Volume reduction in gray matter after RT was observed in most subsites and all functional groups. For the feasible and safe radiotherapy in brain, the powerful predictors presented in the study should be considered.