Optimization of epilepsy surgery through virtual resections on individual structural brain networks

Abstract

BackgroundThe success of epilepsy surgery in patients with refractory epilepsy depends upon correct identification of the epileptogenic zone (EZ) and an optimal choice of the resection area. In this study we developed individualized computational models based upon structural brain networks to explore the impact of different virtual resections on the propagation of seizures.MethodsThe propagation of seizures was modelled as an epidemic process (susceptible-infected-recovered (SIR) model) on individual structural networks derived from presurgical diffusion tensor imaging (DTI) in 19 patients. The candidate connections for the virtual resection were all connections from the clinically hypothesized EZ, from which the seizures were modelled to start, to other brain areas. As a computationally feasible surrogate for the SIR model, we also removed the connections that maximally reduced the Eigenvector Centrality (EC) (large values indicate network hubs) of the hypothesized EZ, with a large reduction meaning a large effect. The optimal combination of connections to be removed for a maximal effect were found using simulated annealing. For comparison, the same number of connections were removed randomly, or based on measures that quantify the importance of a node or connection within the network.ResultsWe found that 90% of the effect (defined as reduction of EC of the hypothesized EZ) could already be obtained by removing substantially less than 90% of the connections. Thus, a smaller, optimized, virtual resection achieved almost the same effect as the actual surgery yet at a considerably smaller cost, sparing on average 27.49% (standard deviation: 4.65%) of the connections. Furthermore, the maximally effective connections linked the hypothesized EZ to hubs. Finally, the optimized resection was more effective than random removal of the same number of connections, and equally or more effective than removal based on structural network characteristics.ConclusionThe approach of using reduced EC as a surrogate for simulating seizure propagation can suggest more restrictive resection strategies, whilst obtaining an almost optimal effect on reducing seizure propagation, by taking into account the unique topology of individual structural brain networks of patients.

Epidemic models characterize seizure propagation and the effects of epilepsy surgery in individualized brain networks based on MEG and invasive EEG recordings

“…We also found that patient specific connectivity reproduces seizure propagation better than fully connected networks, and marginally (although not significantly) better than the average connectivity network (see figure 7). This result is in line with previous studies [36,41] in which possible benefits of using patient-specific connectivity were suggested, but could not be corroborated by a significant difference in the model. Likely, larger data sets would be necessary to unravel how the models benefit from considering patient-specific connectivity.…”

“…As opposed to previous studies which considered highly detailed, non-linear, stochastic models to simulate the activity of each brain region in detail [35,37,40,43,[79][80][81], here we considered an abstract model of epidemic spreading, the SI model, as a proxy for seizure propagation dynamics (see figures 3 and 4). Epidemic models capture the basic mechanisms of processes that diffuse on networked systems, and have been used, for example, to study the propagation of pathological proteins on brain networks [52] and of ictal activity [41].…”

“…Going beyond topological network analysis, the simulation of ictal activity on top of brain networks can aid the identification of the EZ and prediction of surgery outcome, as well as predict possible side-effects. Such computational models can be used to identify epileptogenic areas [33,34] or analyze different resection strategies [26,32,[35][36][37][38], such that patient-specific resection strategies, that may lead to a better outcome or fewer side-effects than the standard surgery, can be tested [33,[39][40][41]. Validation of the models is usually attempted by looking for differences in the model predictions between seizure free and non-seizure free patients [42,43] or by correlating seizure propagation on the model with the empirical data [44].…”

“…Data of outbreak patterns can also be used retroactively to find the location of the origin of an epidemic [57,58]. Within this framework, in a previous study [41] we modelled seizure propagation as an epidemic spreading process and, using the eigenvector centrality as a surrogate measure, found that the size of the resection area could be largely reduced with only a small decrease in the efficacy of the virtual surgery.…”

“…The temporal and spatial resolution of the resulting network-based model, and the interpretation of the connections between regions, will depend on the modality that was used to define the network structure. Studies on epileptogenic networks have considered both functional [32,35,37,42,43,45] and structural [33,36,40,41] networks, as they are both affected in patients with epilepsy [40][41][42][43]. However, functional networks can capture abnormalities in brain activity even in the absence of structural abnormalities [46].…”