Abstract
Contemporary brain tumor surgery increasingly prioritizes a network-based view of cerebral function, shifting from localizationist models toward a meta-network framework that recognizes the brain as an interconnected system. We hypothesize that tumor resection can improve global network organization by alleviating peritumoral hyperexcitability and microenvironmental disruption. Within this framework, structural and functional connectomics provide a promising framework for surgical planning and postoperative assessment. OBJECTIVE: To assess the feasibility of incorporating AI-based connectomic software into clinical workflows and to quantitatively assess changes in large-scale brain network connectivity following tumor resection. METHODS: We retrospectively analyzed all patients undergoing brain tumor surgery with preoperative connectome imaging processed via Quicktome v2.1.0. Demographic, tumor, and cognitive data were collected. A subcohort of patients with both pre- and postoperative connectome imaging—without intervening adjuvant therapy—was analyzed to assess structural connectomic changes. Quantitative metrics included fractional anisotropy (FA), asymmetry index (AI), and percentage asymmetry (%Asym) across key white matter tracts, including the CST, AF, ILF, IFOF, UF, and SLF. Special focus was placed on the dorsal and ventral semantic pathways due to their relevance in language and executive functions. RESULTS: The full cohort included 200 patients (mean age 57.8 ± 16.0 years), 85.5% with intra-axial tumors, predominantly left-sided (53.4%) and medial (60.1%). Of the 113 gliomas, 64.6% were WHO grade 4. In the quantitative subcohort (n=19), tumor resection resulted in measurable improvements in network asymmetry and FA, particularly in the ventral semantic pathway (IFOF and ILF). These changes were most prominent in left-hemispheric tumors patients. Functional outcomes were preserved, with a median postoperative KPS of 80, and 64.5% of patients discharged home. CONCLUSION: AI-guided connectomics can be feasibly integrated into clinical workflows and may offer quantitative insight into postoperative network reorganization. Improvements in tract-level symmetry and FA—especially in ventral semantic pathways—suggest potential for not only preserving but enhancing global network integrity. This approach supports a personalized, connectome-informed definition of eloquence, advancing the principles of cognitive-preserving neurosurgery.