Abstracts

Rationale: To investigate a potential diffusion-based biomarker to confirm lateralization of epileptogenicity in temporal lobe epilepsy (TLE) patients and to introduce an uncertainty-based approach for evaluation of various TLE lateralization methods.Methods: Among the cingulum subregions segmented from diffusion tensor images of 14 nonepileptic subjects, a specific subregion with no significant interhemispheric variation of fractional anisotropy (FA) was identified and the hemispheric variation uncertainty (HVU) level of FA estimated. An analysis of the relationship between preoperative FA asymmetry and the laterality of seizure onset was performed upon a retrospective cohort of 10 TLE patients who had pathology proven mesial temporal sclerosis (MTS) and underwent surgical resection with Engel class I outcomes, two of whom underwent extraoperative electrocorticography (ECoG) to determine lateralization. A conservative estimate of a true variation of FA attributable to a unilateral epileptogenicity for each patient was determined using a 95% confidence interval of the HVU. The lateralization outcome of the proposed diffusion-based method was compared with that of HVUs of manual hippocampal T1(longitudinal relaxation time) volumetry and quantitative FLAIR (Fluid Attenuated Inversion Recovery) intensity methods.Results: The FA value within the posteroinferior region of the cingulum was significantly lower on the side ipsilateral to the temporal epileptogenicity in patients with MTS using an estimated HVU of 0.021 (p=0.05). Hemispheric variation was minor and insignificant for nonepileptic subjects. For each individual MTS patient, we observed a lower FA within the side ipsilateral to seizure onset. However, the hemispheric variation for one patient was not beyond uncertainty. The proposed diffusion-based method was able to confidently and correctly lateralize the seizure onset in 9 out of 10 cases, while the hippocampal T1 volumetry and FLAIR intensity analysis outperformed the proposed method and correctly lateralized all 10 cases, among which two cases had undergone extraoperative ECoG to determine lateralization.Conclusions: The presence of MTS in TLE patients is associated with that of a reduced FA value in the ipsilateral posteroinferior subregion of the cingulum. Identification of this attribute in patients with MTS is simpler and less labor intensive than manual hippocampal volumetry and FLAIR intensity analysis. It holds promise for improving decision-making for surgical resection and might reduce the need for implantation of intracranial monitoring electrodes.