Background  Glioma is the most common intracranial malignant tumor, which often shows strong invasiveness. Previous studies have shown that the existence of gliomas will not only cause damage around the lesions, but also distant sites outside the scope of the tumor. The study of structural brain network can help surgeons to make better judgment about the surgical expectation and prognosis for glioma patients.

  Objective  To investigate the changes of topological properties of brain network in patients with gliomas at different sites, and reveal the structural network remodeling related to brain function changes caused by tumor lesions.

  Methods  Totally 113 glioma patients were enrolled from 2018 to 2020 at the Department of Neurosurgery, Chinese PLA General Hospital. The diffusion tensor imaging data from MRI were processed by Mrtrix3 software and the brain structure network was constructed. The topological properties of the network were calculated by GRETNA software using graph theory. The effects of tumors in different locations on brain structural network were observed.

  Results  There were significant differences in global shortest path length (Lp), clustering coefficient (Cp), global network efficiency (Eg) and local efficiency (Eloc) among frontal lobe, temporal lobe and parietal lobe gliomas. The global topological property of parietal lobe glioma group was higher than that of frontal lobe and temporal lobe glioma group (P<0.05). Compared with other brain region node network attribute indexes, betweenness centrality (NBc), degree centrality (NDc), nodular global efficiency (NEg) increased and nodular shortest path length (NLp) decreased in the brain region where the tumor was located, and the difference was statistically significant (P< 0.05).

  Conclusion  The network topology attributes of glioma at different lobe locations have changed, and both the global attributes and local node attributes show changes in the connectivity of the brain structure network.