Clinical and molecular predictors of postoperative seizure outcomes in adult low-grade glioma.

Abstract

e14084 Background: Seizures are a common presenting symptom in patients with low-grade glioma (LGG) and significantly impact quality of life. While surgical resection improves seizure control, the clinical and molecular factors predicting postoperative seizure outcomes remain incompletely defined. We performed a systematic review to evaluate predictors of postoperative seizure control in adult patients with LGG. Methods: A systematic review (PROSPERO: CRD420261295317) was conducted; PubMed, Embase, Scopus, and Cochrane Library were searched for studies evaluating clinical and molecular predictors of postoperative seizure outcomes in adults (≥18 years) with WHO grade II low-grade glioma undergoing surgical resection. Eligible studies reported seizure outcomes using standardized classifications (Engel or ILAE). Extracted data included demographic characteristics, seizure history, extent of resection (EOR), tumor location, molecular markers, follow-up duration, and effect estimates. Risk of bias was assessed qualitatively. Due to heterogeneity in study design and outcome reporting, quantitative synthesis was descriptive without meta-analysis. Results: Twenty-six retrospective studies comprising over 3,000 patients were included, with median or mean follow-up ranging from 3 months to 6 years. Postoperative seizure freedom (Engel class I or ILAE class I) was achieved in 65–83% of patients. Gross or near-total resection was the most consistent predictor of seizure freedom, with odds ratios (OR) ranging from 3.2 to 16.0 compared with subtotal resection (p < 0.05 across most multivariable models). Increasing EOR as a continuous variable independently predicted seizure freedom (adjusted OR per percentage increase 1.02–1.03; p ≤ 0.001). Molecular predictors associated with favorable seizure outcomes included 1p/19q codeletion (OR 3.2–4.7; p ≤ 0.004), while IDH mutation was associated with seizure presentation and postoperative seizure persistence in multicenter cohorts (OR 2.7–3.1; p < 0.001). Elevated Ki-67 expression was consistently associated with poorer seizure control (OR 0.07–0.38; p ≤ 0.02). Higher preoperative seizure frequency, longer seizure duration, and secondary generalized seizures were associated with worse postoperative seizure outcomes (OR 2.0–6.6; p ≤ 0.01). Overall risk of bias across studies was moderate. Conclusions: In adult patients with low-grade glioma, extent of resection is the strongest and most consistent predictor of postoperative seizure freedom. Molecular features, particularly 1p/19q codeletion and proliferative indices, provide additional prognostic value. These findings support aggressive yet safe surgical resection and highlight the role of integrated clinical–molecular risk stratification to optimize seizure outcomes in LGG.