Characterizing functional connectivity in brain tumor patients.

Abstract

2041 Background: Brain tumors (affecting 25,500 individuals in the US alone) cause unique changes to neural function and connectivity resulting in impairments in various functional domains, including cognition. Methods: Between 2012 and 2018, as part of their preoperative surgical planning, 91 brain tumor patients with tumors in right (R) or left (L) temporal (T), parietal (P) or frontal (F) lobes received resting-state functional magnetic resonance imaging (rs-fMRI) and the COWAT verbal fluency (VF) test. UW Hospital and Clinics uses 1.5 T (112 axial slices, 1.0 × 1.0 × 1.5 mm) and 3 T (136 axial slices, 1.0 × 1.0 × 1.2 mm) GE MRI scanners and includes high-resolution 3D BRAVO T1-weighted imaging, and rs-fMRI scans (eyes closed, 28 axial slices, 3.75×3.75×5.0mm) were also acquired during this imaging protocol. These individuals were compared to each other and to 40 age-matched non-tumor controls (Cs). Results: Groups were similar in age, and sex (p>0.05), but different in education (p=0.028) and VF scores (p=0.001). There were significant differences in post-hoc p-values when comparing VF scores between Cs and RT tumor patients (p=0.039). Qualitative observations indicate Cs are more integrated have greater network strength and connectivity higher transitivity, efficiency, and modularity index compared to patients. There were significant differences in VF scores between Cs, LF (p=0.001), RF (p=0.012), RT (p=0.021), RP (p=0.015). Patients show hypo-frontality of hubs and a unique cerebellar module. LF patients showed correlation to transitivity at 25% (R=0.404, p=0.041), and patients with RP tumors had significant correlations to transitivity (R=0.654, p=0.029) and global efficiency (R=0.607, p=0.048). Cs and L tumors did not show any significant correlation to VF scores, R tumors showed correlation with global efficiency at 25% sparsity (R=0.374, p=0.025). All remained significant after FDR correction. Conclusions: Tumor location impacts rs-fMRI-derived GT network structure and VF scores, suggesting tumors in left hemisphere and frontal areas caused the greatest impact to cognition. These methods may be used in future research and clinical care to map, track, and predict functional connectivity changes resulting from brain tumor and can help inform clinicians and care trajectories.