ESTRO 2025 - Abstract Book

S653

Clinical - CNS

ESTRO 2025

1 Department of Oncology, KU Leuven, Leuven, Belgium. 2 Department of Radiotherapy, AZ Turnhout, Turnhout, Belgium. 3 Department of Imaging and Pathology, KU Leuven, Leuven, Belgium. 4 Department of Neurosciences, KU Leuven, Leuven, Belgium. 5 Department of Neurosurgery, UZ Leuven, Leuven, Belgium. 6 Department of Cognitive Neuropsychology, Tilburg University, Tilburg, Netherlands. 7 Department of Radiotherapy and Oncology, UZ Leuven, Leuven, Belgium Purpose/Objective: Emerging evidence suggests that cognitive dysfunction may result from damage to the brain's functional network. This study explores the dose-dependent susceptibility of functional hubs to radiotherapy (RT), their interaction with structural network hubs, and associations with cognitive outcomes. Material/Methods: Attention, language, memory, motor-, and executive functioning were assessed ≥1-year post-radiotherapy in 39 patients irradiated for an adult WHO grade 2 and 3 gliomas and 50 matched healthy controls. Using multi-shell diffusion-weighted and resting-state functional imaging, weighted structural and functional graphs were constructed for each participant, identifying hubs through graph measures. Linear regression models and Spearman’s rho correlations assessed associations between mean RT dose per region, their clustering, and cognitive domains. Results: Higher RT doses to the left fusiform and inferior temporal gyri were linked to memory impairment (r(37)≥-.565, p FDR ≤.026), while poorer language outcomes were associated with higher doses to the left pars opercularis, rostral middle frontal gyrus and caudate (r(37)≥-.510, p FDR ≤.040). Attention deficits were linked to higher doses to the left posterior cingulate, precentral, supramarginal and postcentral gyrus ((37)≥-.499, p FDR ≤.040), with the left postcentral gyrus also associated with executive dysfunction (r(37)=-.526, p FDR =.029). Significant correlations between RT dose and cognitive outcomes were more frequent in hubs than in non-hubs (50% vs. 12%, p=.005) and exclusively found in left-sided regions. No significant effect of RT dose was observed for clustering coefficient values. In comparing hub hub connections between cognitively impaired and non-impaired patients, we found structural hub-hub connections to be significantly more altered than functional hub-hub connections (10% versus 2%, χ²=15.854, p uncorr <.001).

Figure 1 Functional hubs and non-hubs with a significant dose-response relation on cognitive performance