Object: To explore the post-hemispherotomy seizure outcome and its prognostic predictors in children with refractory epilepsy. Methods: We reviewed 83 consecutive child patients with refractory epilepsy who underwent a hemispherectomy from June 2014 to January 2017 at our Pediatric Epilepsy Center. Demographic, clinical, EEG, neuroimaging, and surgical data were collected. Seizure outcome data were collected via outpatient clinics as well as telephone visits and were graded according to Engel criteria. Logistic regression model and Cox proportional hazard regression model were, respectively, applied to explore the related factors predicting the seizure outcomes of children after a hemispherotomy. Results: Of the 83 patients, 55 (63.2%) were male. The mean seizure onset age was 1.9 years (0-8.7 years), and the mean surgery age was 5 years (0.8-14 years). At a mean follow-up of 3 years, 69 children (83.1%) were seizure free, and 14 (16.9%) exhibited seizure recurrence. In a univariate analysis, whether or not considering follow-up time, a non-lateralized interictal EEG pattern, bilateral PET abnormalities and acute postoperative seizures (APOS) could all predict poor seizure outcomes post-hemispherotomy. Bilateral PET abnormalities were independently correlated with unfavorable seizure outcomes in the multivariate Logistic regression analysis (Odds ratio(OR) = 13.05, 95%CI = 1.52-112.29, P = 0.019) and in the multivariate Cox proportional hazard analysis(OR = 13.99, 95%CI = 2.75-71.17, P = 0.001). Conclusions: Child epileptic patients with bilateral PET abnormalities may have poor seizure outcomes after a hemispherotomy procedure. This study will facilitate better candidate selection for hemispherotomies and early identification of unfavorable seizure outcomes.
Keywords: children; hemispherotomy; predictor; refractory epilepsy; seizure outcome.