gms | German Medical Science

Objective: Cerebral tumors, which infiltrate the motor system lead to considerable disabilities. We have investigated how tumors in the motor system affect the patients’ networks. Network-based statistics (NBS) allows to analyze global networks. NBS works on basis of the connectome, which defines a matrix representing all possible pairwise anatomical connections between neural elements of the brain.

Methods: We included 37 left- and right-handed adult patients in this study (15 females, 22 males, average age=48.24, SD=16.47, age range 20-78). Only patients with initial diagnosis of unilateral WHO grade II-IV gliomas were included. We used the NIHSS, a Nine-Hole Peg Test and a hand dynamometer to quantify tumor-related impairments. Probabilistic tractography was performed with Anatomically-Constrained Tractography. We computed 50'000'000 streamlines which were filtered to 10'000'000. We constructed symmetric connectome matrices with the Desikan–Killiany–Tourville node parcellation. Tractography and statistical analysis was performed by MRtrix3 for connectome group-wise edge level statistics using non-parametric permutation testing and control of family-level errors.

Results: NBS revealed significant edges with a structural connectivity decrease in pathological hemisphere matrices compared to healthy hemisphere matrices. Tumor localization-based subgroups showed distinct patterns of affected subnetworks. Especially frontal edges correlated with NIHSS, WHO degree or motor status.

Conclusion: Our findings suggest that we can use network-based statistics and connectome analyses to reveal global impacts of tumors on the network. Furthermore, our findings suggest that a functional decrease can be derived from the structural connectome. This method proved feasible and shows promising results to possibly further improve neurooncological and -surgical analyses.