Introduction: High-grade glioma (HGG) is a highly aggressive and proliferative brain cancer. Treatment most often involves maximum safe resection, followed by adjuvant chemotherapy and radiation. However, an incomplete understanding of HGG's impact on brain connectivity limits the prediction of post-HGG resection complications and extent-of-resection protocols. Previous work has primarily focused on analyzing clinical and structural data to understand and predict post-surgical outcomes. These models overlook connectomics - an emerging field focused on the functional mapping of brain neural networks. In this systematic review, the authors 1) summarize the current understanding of major neural networks impacted by HGG resection through both a functional and structural viewpoint, and 2) discuss associated advancements in connectomics-based machine-learning models and application towards predicting post-surgical outcomes.
Methods: A systematic search was performed of peer-reviewed articles before 10/19/2023 according to PRISMA guidelines. No restrictions on publication date were utilized. Search terms included, "connectomics," and "glioma." Articles were included in the review if DTI structural data and/or rs-fMRI functional data involving white matter tracts impacted by HGG were analyzed. Articles were excluded if results did not apply to HGGs, tumor location was not included, or a full-text copy was unavailable.
Results: We reviewed 41 studies which analyzed the impact of HGGs on the brain connectome. HGGs tend to increase structural connectivity (SC) among rich-club nodes and reduce SC among peripheral nodes, though effects on functional connectivity (FC) tend to vary by tumor location. Frontal HGGs elicit bilateral FC changes, including decreased global efficiency (GE), local efficiency (LE), degree centrality, and increased average path length. Similarly, temporal HGGs often result in altered bilateral FC, including decreased LE with preserved GE and small-world properties. Parietal HGGs have mainly local effects and preserved small world organization, apart from observed bilateral FC changes of the default-mode network (DMN) and with precuneus HGGs. Insular HGGs also primarily affect local FC. In application, connectomic metrics including FC, LE, and GE have improved the predictive capabilities of post-HGG resection complications when combined with clinical and structural metrics.
Conclusions: HGG has distinct impacts on the functional connectome based on tumor location. In general, HGGs tend to decrease long-distance, interhemispheric FC and LE while GE and clustering coefficient are preserved. In application, these metrics have been used in connectomic-based models to predict post-HGG resection complications more accurately than previous clinical- or structural-based models.
Keywords: Connectomics; Functional connectivity; Glioma; Machine-learning; Neurooncology; Structural connectivity.
Copyright © 2025. Published by Elsevier Ltd.