MRI supersedes ictal EEG when other presurgical data are concordant.

PURPOSE: When ictal EEG is discordant with MRI and other presurgical data, our group has sometimes discounted the ictal findings and proceeded with epilepsy surgical resection based on MRI. We aimed to evaluate the prudence of such practice by comparing the outcome of MRI-lesional epilepsy surgery patients with discordant ictal EEG with those with concordant ictal EEG.

METHOD: We retrospectively studied 115 children with epilepsy who underwent surgical resection of an MRI lesion that was corroborated as the epileptogenic focus by other presurgical findings. Ictal findings on video-EEG were categorized as: "positive ictal EEG" if the ictal onset localization was concordant with MRI and other presurgical data; "negative ictal EEG" if the ictus was discordant with them. Seizure-free outcome at 2 years was compared between the "positive" and the "negative" ictal EEG groups.

RESULTS: Seizure-free outcome did not differ between children with positive ictal EEG (73%) and those with negative ictal EEG (80%). Positive ictal EEG did not result in better outcome regardless of the location of the surgery or the pathology of the lesion. Ictal EEG with 73% positive predictive value provided no added benefit in this cohort whose seizure-free outcome was of 77% irrespective of ictal EEG findings.

CONCLUSIONS: In our selected cohort of pediatric epilepsy surgery patients with an epileptogenic lesion on MRI and concordant other data, ictal EEG had limited predictive value. This calls into question the additive role of ictal recordings in patients with an MRI lesion and concordant other presurgical data.