Lateralizing and localizing value of seizure semiology: Comparison with scalp EEG, MRI and PET in patients successfully treated with resective epilepsy surgery

Purpose: To evaluate the lateralizing and localizing value of seizure semiology in patients who became completely seizure free after resective epilepsy surgery. Methods: We analyzed seizures of patients who were seizure free after focal resection limited to the temporal lobe (30 patients), frontal lobe (27 patients), parietal lobe (8 patients) and occipital lobe (8 patients). Three investigators independently analyzed video segments of seizures blinded to clinical information and attempted to lateralize and localize the seizure focus, based on pre-defined criteria. Inter-observer agreement measured and positive predictive value were calculated and compared with scalp interictal, ictal EEG, MRI and PET. Results: Seizure semiology correctly lateralized 19/30 (63%) temporal lobe epilepsy patients (kappa = 0.64); lobar localization was correct in 27/30 patients (90%) (Kappa = 0.6). Twenty of 27 (74%) frontal lobe patients were correctly lateralized (kappa = 0.55) and localized (Kappa = 0.41). Seven of 8 (87%) of parietal lobe patients were correctly lateralized (kappa = 0.83) and lobar localization was correct in 3/8 patients (37%) (Kappa = 0.5). Seven of 8 (87%) occipital lobe patients were correctly lateralized (kappa = 0.67) and lobar localization was correct in 7/8 patients (87%) (Kappa = 0.67). Collectively, seizure semiology was correctly lateralizing in 74% and correctly localizing in 77% of patients. Lateralization in secondarily generalized seizures was significantly higher compared to seizures that did not generalize (p = 0.0001). In non-lesional patients, lateralization was correct in 81% while localization was correct in 88% of patients. Conclusion: Seizure semiology has high lateralizing and localizing value, on par with interictal/ictal scalp EEG and MRI.