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

Area of Science:

• Neuroscience
• Computational Biology
• Medical Imaging

Background:

• Epilepsy surgery success hinges on precise identification of the epileptogenic zone (EZ) and optimal resection.
• Refractory epilepsy necessitates advanced strategies for surgical planning.

Purpose of the Study:

• To develop individualized computational models for predicting seizure propagation.
• To explore the impact of virtual resections on seizure spread using structural brain networks.
• To identify optimal resection strategies that minimize tissue removal while maximizing seizure control.

Main Methods:

• Developed individualized computational models based on structural brain networks from diffusion tensor imaging (DTI) in 19 epilepsy patients.
• Modeled seizure propagation as a susceptible-infected-recovered (SIR) process on individual brain networks.
• Used simulated annealing to find optimal virtual resections by removing connections that maximally reduced the eigenvector centrality (EC) of the hypothesized EZ.
• Compared optimized resections with random removal and removal based on other network centrality measures.

Main Results:

• Optimized virtual resections achieved approximately 90% of the effect (reduction in EZ's EC) by removing substantially fewer connections than a full resection.
• The optimized strategy spared an average of 27.49% of connections compared to a complete resection.
• Maximally effective connections identified by the model linked the hypothesized EZ to network hubs.
• Optimized resection was as effective or more effective than other methods in reducing EZ's EC and seizure spread.

Conclusions:

• Computational modeling of seizure propagation using network topology can guide more conservative and effective surgical resection strategies.
• Individualized brain network analysis offers a promising approach for optimizing epilepsy surgery outcomes.
• Reduced eigenvector centrality serves as a viable surrogate for simulating seizure propagation, enabling targeted interventions.