Metabolic Brain Network and Surgical Outcome in Temporal Lobe Epilepsy: A Graph Theoretical Study Based on 18F-fluorodeoxyglucose PET

Shuhua Ren,Qi Huang,Weiqi Bao,Donglang Jiang,Jianfei Xiao,Junpeng Li,Fang Xie,Yihui Guan,Rui Feng,Fengchun Hua
DOI: https://doi.org/10.1016/j.neuroscience.2021.10.012
IF: 3.708
2021-01-01
Neuroscience
Abstract:Drug-resistant temporal lobe epilepsy (TLE) is a potential candidate for surgery; however, nearly one-third subjects had a poor surgical prognosis. We studied the underlying neuromechanism related to the surgical prognosis using graph theory based on metabolic brain network. Sixty-four unilateral TLE subjects with preoperative 18F-fluorodeoxyglucose (FDG) PET scanning were retrospectively enrolled and divided into Ia (Engel class Ia, n = 32) and non-Ia (Engel class Ib-IV, n = 32) groups according to more than 3-year follow-up after unilateral anterior temporal lobectomy (ATL). The metabolic brain network was constructed and the changed metabolic connectivity of Ia and non-Ia was detected compared with 15 matched healthy controls (HCs). Further, the network properties, including small-worldness and global efficiency, were calculated and hub nodes were also identified for the 3 groups respectively. Non-Ia group exhibited increased connectivity between contralateral fusiform gyrus and contralateral lingual gyrus; while Ia showed decreased connectivity mainly among bilateral frontal, temporal and parietal cortex. Graph theoretical analysis revealed that non-Ia group showed increased small-worldness (35%<s < 55%, P ≤ 0.05) compared to HCs; and elevated global efficiency (P = 0.05) and decreased Lp (P = 0.05) compared to Ia group. Ia group showed reduced Cp (55%<s < 63%, P < 0.05) and increased small-worldness (35%<s < 37%, P < 0.05) compared to HCs; Furthermore, disrupted hub nodes distribution pattern with the midcingulate gyrus disappeared, was also found in non-Ia group compared with the Ia group. All those results revealed that elevated network integration and metabolic connectivity, redistributed hub nodes pattern is associated with ongoing postoperative seizures in subjects with intractable TLE.
What problem does this paper attempt to address?