Volumetric Extent of Resection and Seizure Freedom in Pediatric Supratentorial Gangliogliomas.

BACKGROUND AND OBJECTIVES: Gangliogliomas in children and adults are often associated with refractory seizures. Gross total resection (GTR) has consistently been associated with higher rates of seizure freedom as compared with subtotal resection (STR); however, it is unknown if there is an optimum extent of resection (EOR) threshold above which seizure outcomes are favorable. METHODS: An institutional cohort of 47 supratentorial gangliogliomas in children and young adults presenting with seizures was identified. Surgical resection was limited to tumor margins, and no patient underwent extended lesionectomy. Volumetric analysis of pre- and early postoperative images was performed to quantitatively determine the EOR. The optimal threshold was identified using a receiver operating characteristic analysis. The primary outcome was seizure freedom at 12 months, defined by an Engel Class of I. RESULTS: The mean age at surgery was 13.1 years, and 48.9% of tumors involved the medial temporal lobe. GTR (100% resection of all fluid attenuated inversion recovery-abnormal tissue) was achieved in 48.9%. Engel Class I was achieved in 76.6% of patients at 12 months after surgery. Patients were followed for a mean duration of 5.0 years after surgery, and 5 patients had late-onset worsening of seizure control associated with recurrence. EOR was significantly associated with seizure freedom at 12 months (area under curve 0.72; 95% CI 0.532-0.912; P = .027); the optimum EOR threshold was ≥84% (sensitivity 80.6%, specificity 63.6%). Further dichotomization of EOR based on this threshold demonstrated significant association with seizure control on multivariate analysis (odds ratio 6.403; 95% CI 1.418-28.913; P = .016). CONCLUSION: In this study, volumetric EOR ≥84% was significantly associated with seizure freedom at 1 year after resection of supratentorial gangliogliomas. While GTR should be attempted where feasible, volumetric EOR may be used as a more nuanced predictor of seizure outcome in patients undergoing STR.