Abstracts

Rationale: It is still largely unknown why some patients with medication refractory Temporal Lobe Epilepsy (TLE) continue to experience disabling seizures after Anterior Temporal Lobectomy (ATL). In this study, we employ new imaging tools to quantify personalized measures of neural architecture (the brain connectome) to test whether the topography of abnormal networks can be used to predict individual chances of surgical success. Methods: We studied 35 consecutive patients with TLE who underwent ATL. The brain connectome was reconstructed from all patients based on pre-surgical diffusion magnetic resonance imaging (dMRI) and contrasted with normative connectomes reconstructed from similar dMRI obtained from 18 healthy controls. The topography of abnormalities in link-wise elements of the connectome was assessed on sub-networks linking temporal with extra-temporal regions. Predictive models were constructed from different subnetworks and contrasted with the predictive values from conventional clinical data. Results: Patients were more likely to achieve post-surgical seizure-freedom if they exhibited fewer abnormalities within a sub-network composed of the ipsilateral hippocampus, amygdala, thalamus, superior frontal region, lateral temporal gyri, insula, orbitofrontal cortex, cingulate and lateral occipital gyrus (Figure 1, Table 1). A model composed of this sub-network achieved a cross validated predictive value of 0.89 towards seizure freedom (Figure 1). Conclusions: These results suggest that the absence of abnormal networks extending beyond medial and anterior temporal regions is likely a crucial determinant of the individual chances of surgical success. New technological advancements in brain mapping applied to conventional pre-surgical dMRI may improve epilepsy treatment outcome prediction based on personalized measures of neural network architecture.