Abstracts

Rationale: Noninvasive evaluation and management of patients with intractable focal epilepsy after failed surgery is particularly challenging. Reoperation may offer a second chance for seizure-freedom in up to 50%. Patients with nonlesional MRI are more likely to fail than patients with lesional MRI. MRI post-processing has been demonstrated to be effective in identifying subtle focal cortical dysplasia lesions in nonlesional epilepsy patients. In this study, we aim to investigate whether MRI post-processing could also help provide relevant targets for the presurgical evaluation of patients with prior failed surgery. Methods: We retrospectively identified patients by reviewing our surgical database from 2002 to 2011. Voxel-based MRI post-processing of the T1-weighted volumetric sequence was implemented in a morphometric analysis program (MAP) in Matlab SPM5, consistent with methodology published by Huppertz et al. MAP findings were not previously available and therefore did not influence the surgical decision. Patients were included if they: (1) had a preoperative 1.5T or 3T MRI; (2) were considered MRI negative prior to surgery; (3) had recurring seizures within 12 months of follow up; and (4) had a MAP+ region unresected. The gray-white junction z-score output of MAP was calculated by comparing the patient with a scanner-specific normal database. Regions with z-score > 4 were considered indicative of significant gray-white blurring. These areas were presented to a neuroradiologist (SEJ), who was the final judge of whether they present true lesions (MAP+). The neuroradiologist was blinded to patients' clinical information. The completeness of resection of MAP+ region was determined by post-operative MRI. Concordance between the unresected MAP+ region and all the available functional imaging/electrophysiology data was then evaluated. Results: Ten patients fulfilled the selection criteria. Three had a single MAP+ region that was not/partially resected. The remaining 7 had two MAP+ regions, with one resected and the other left alone. The locations of the unresected MAP+ region were: 3 in the same lobe, 3 in a different lobe on the ipsilateral side, and 1 on the contralateral side. In the overall 10 patients, two had a second surgery including the unresected MAP+ region, and both patients became seizure-free (> 12 months follow up), as exemplified in the Figure. The remaining 8 patients did not undergo further surgery; however, in 7 of the 8 patients, the unresected MAP+ region was concordant with at least one noninvasive modality, i.e., scalp-EEG, SPECT, PET, and/or magnetic source imaging. Concordance between the unresected MAP+ regions and intracranial EEG findings were not clear because the MAP+ regions were not prospectively targeted. Conclusions: Scrutiny of the pre-surgical MRI with guidance from MRI post-processing may reveal relevant targets for reoperation in nonlesional epilepsies, and thus has the potential to improve seizure outcomes in patients with failed surgery. The unresected MAP+ regions should be considered when devising the implantation scheme.