Purpose/Objective(s) Radiation and chemotherapy are highly effective treatments for patients with nasopharyngeal carcinoma (NPC), with a cure rate of 80% or higher. Long-term survivors, despite having normal MRIs, commonly suffer from memory impairment and neurocognitive dysfunction. The purpose of the study was to quantify the long-term impact of chemoradiation on brain function using general linear model (GLM) and brain connectivity and to identify brain regions which show significant functional changes after radiation. Materials/Methods Between 2002 and 2020, 26 NPC patients underwent IMRT or proton at our institution. The median age was 51 years (17 male). For all patients, high-resolution T1-W MRIs were obtained before radiation (pre-RT) and at least 5 years after (post-RT) (median 8 years). Open-source software for processing and analyzing brain MRI images was used to perform automated cortical region segmentation. Cortical area information was extracted from the 34 brain regions from both left and right brain to analyze. Age and gender were used as covariates. GLM was used to assess change in cortical regions. Simulations were used to correct for multiple comparisons. Graph theory-based brain connectome was evaluated to assess the connectivity between different regions of brain. Local efficiency and clustering coefficient were used to determine local region connectivity. Global efficiency, transitivity, and modularity were used to assess global connectivity. Hub regions (by regions of unusually high degree of connectivity) were identified. Results Five clusters were generated in the right brain: Cluster 1 and 3 (superiortemporal: p=0.05), Cluster 2 (supramarginal: p=0.001), Cluster 4 (precentral: p=0.003) and Cluster 5 (caudalanteriorcingulate: p= 0.04). Brain connectome was evaluated to analyze the brain region connectivity. After RT, patients presented an abnormal global network as reflected by increased modularity (p<0.001) and decreased global efficiency (p=0.004), clustering coefficient (p=0.002), and local efficiency (p=0.002). Reorganization of hubs was observed in post-RT. Pre-RT hubs: left (rostralanteriorcingulate, posteriorcingulate, parstriangularis, rostralmiddlefrontal, temporalpole, caudalanteriorcingulate, and pericalcarine) and right entorhinal, pericalcarine, and inferiorparietal) brain. Post-RT hubs: left (superiorfrontal, frontalpole, Lateralorbitofrontal, superiorparietal, parsopercularis, insula, posteriorcingulate, pericalcarine, and precuneus) and right (rostralmiddlefrontal, parsopercularis, and lingual) brain. Conclusion Global and permanent change in cortical areas can be seen after chemoradiation in long-term survivors. Postcentral, supramarginal, superiotemporal, and precentral regions showed the most change in cortical areas and the most disrupted connectivity. Reorganization of hubs was observed after chemoradiation. Radiation and chemotherapy are highly effective treatments for patients with nasopharyngeal carcinoma (NPC), with a cure rate of 80% or higher. Long-term survivors, despite having normal MRIs, commonly suffer from memory impairment and neurocognitive dysfunction. The purpose of the study was to quantify the long-term impact of chemoradiation on brain function using general linear model (GLM) and brain connectivity and to identify brain regions which show significant functional changes after radiation. Between 2002 and 2020, 26 NPC patients underwent IMRT or proton at our institution. The median age was 51 years (17 male). For all patients, high-resolution T1-W MRIs were obtained before radiation (pre-RT) and at least 5 years after (post-RT) (median 8 years). Open-source software for processing and analyzing brain MRI images was used to perform automated cortical region segmentation. Cortical area information was extracted from the 34 brain regions from both left and right brain to analyze. Age and gender were used as covariates. GLM was used to assess change in cortical regions. Simulations were used to correct for multiple comparisons. Graph theory-based brain connectome was evaluated to assess the connectivity between different regions of brain. Local efficiency and clustering coefficient were used to determine local region connectivity. Global efficiency, transitivity, and modularity were used to assess global connectivity. Hub regions (by regions of unusually high degree of connectivity) were identified. Five clusters were generated in the right brain: Cluster 1 and 3 (superiortemporal: p=0.05), Cluster 2 (supramarginal: p=0.001), Cluster 4 (precentral: p=0.003) and Cluster 5 (caudalanteriorcingulate: p= 0.04). Brain connectome was evaluated to analyze the brain region connectivity. After RT, patients presented an abnormal global network as reflected by increased modularity (p<0.001) and decreased global efficiency (p=0.004), clustering coefficient (p=0.002), and local efficiency (p=0.002). Reorganization of hubs was observed in post-RT. Pre-RT hubs: left (rostralanteriorcingulate, posteriorcingulate, parstriangularis, rostralmiddlefrontal, temporalpole, caudalanteriorcingulate, and pericalcarine) and right entorhinal, pericalcarine, and inferiorparietal) brain. Post-RT hubs: left (superiorfrontal, frontalpole, Lateralorbitofrontal, superiorparietal, parsopercularis, insula, posteriorcingulate, pericalcarine, and precuneus) and right (rostralmiddlefrontal, parsopercularis, and lingual) brain. Global and permanent change in cortical areas can be seen after chemoradiation in long-term survivors. Postcentral, supramarginal, superiotemporal, and precentral regions showed the most change in cortical areas and the most disrupted connectivity. Reorganization of hubs was observed after chemoradiation.
Read full abstract