Brain connectivity predict surgical outcomes of low-grade epilepsy-associated neuroepithelial tumors.

OBJECTIVE: Low-grade epilepsy-associated neuroepithelial tumors (LEATs) often cause drug-resistant epilepsy. Despite complete resection of these lesions, approximately 20% of patients continue to experience suboptimal seizure control. This study aims to investigate the predictive value of quantitative features in determining the surgical outcomes for LEAT patients. METHODS: We retrospectively analyzed 44 temporal LEAT patients who underwent gross-total lesionectomy. EEG features, including power spectral density (PSD) and weighted phase lag index (wPLI), were compared between patients with good (Engel I) and poor (Engel II-IV) outcomes. Significant EEG features were identified through these analyses. Domain Adversarial Neural Network (DANN) was employed to assess the predictive value of these features for surgical outcomes. RESULTS: No significant PSD differences were found, but patients with good outcomes had higher alpha-band wPLI (p = 0.008). LEATnet, predicted outcomes with an AUC of 0.81and correctly classified 8 of 11 patients in the independent validation cohort. CONCLUSIONS: Alpha-band functional connectivity is a key predictor of surgical outcomes in LEAT patients. SIGNIFICANCE: EEG-based connectivity analysis may improve prognostic accuracy and aid clinical decision-making in LEAT epilepsy.