Hemispherotomy in children: A retrospective analysis of 152 surgeries at a single center and predictors for long-term seizure outcome.

ABSTRACT

OBJECTIVE: Completeness as a predictor of seizure freedom is broadly accepted in epilepsy surgery. We focused on the requirements for a complete hemispherotomy and hypothesized that the disconnection of the insula contributes to a favorable postoperative seizure outcome. We analyzed surgical and nonsurgical predictors influencing long-term seizure outcome before and after a modification of our hemispherotomy technique. METHODS: We retrospectively studied surgical procedures, electroclinical parameters, magnetic resonance imaging (MRI) results, and follow-up data in all children who had undergone hemispherotomy between 2001 and 2018 at our institution. We used logistic regression models to analyze the influence of different factors on seizure outcome. RESULTS: A total of 152 patients were eligible for seizure outcome analysis only. Of these, 140 cases had complete follow-up data for ≥24 months and provide the basis for the following results. The median age at surgery was 4.3 years (range = .3-17.9 years). Complete disconnection (including the insular tissue) was achieved in 63.6% (89/140). At 2-year follow-up, seizure freedom (Engel class IA) was observed in 34.8% (8/23) with incomplete insular disconnection, whereas this was achieved in 88.8% (79/89) with complete surgical disconnection (p < .001, odds ratio [OR] = 10.41). In the latter group (n = 89), a potentially epileptogenic contralateral MRI lesion was the strongest predictor for postoperative seizure recurrence (OR = 22.20). SIGNIFICANCE: Complete surgical disconnection is the most important predictor of seizure freedom following hemispherotomy and requires disconnection of the insular tissue at the basal ganglia level. Even if the hemispherotomy is performed surgically completely, a potentially epileptogenic contralateral lesion on preoperative MRI significantly reduces the chances of postoperative seizure freedom.