Abstracts

RATIONALE: Surgical resection is a treatment option among patients with intractable focal epilepsy. Seizure freedom is between 60 to 80% for temporal lobe epilepsy patients undergoing surgical resection. To avoid negative outcome, extensive clinical testing is required to pre-surgically evaluate and select patients. There is no pre-surgical scoring model developed to predict the success of temporal lobe resective surgery. OBJECTIVE: To develop a pre-operative scoring model that is at least 80% accurate in predicting seizure freedom (Class I Engel's Classification) 6 and 12 months post-surgery. METHODS: Records of patients who underwent temporal lobe surgery, with intent to cure medically intractable epilepsy, at the Cleveland Clinic Foundation Epilepsy Center between January 1, 1997 to February 28, 1999 were reviewed. Patients with redo operations, less 6 months follow-up or no imaging data were excluded. Demographic data, video-EEG monitoring report, imaging results and memory asymmetry (WADA test) were evaluated. The odds ratios of the variables were determined univariably and considered signficant if p values were less than 0.05. The odds ratio were converted into relative weights for the scoring model. A cut-off score was selected to maximize the correct classification rate. Sensitivity, specificity and predictive values were computed. MAIN OUTCOME VARIABLE: seizure free (Class I) and not seizure free (Class II,III,IV) 6 and 12 months post-surgery. RESULTS: One hundred sixteen temporal lobe surgeries were performed and 102 patients fulfilled the inclusion criteria.The percent of patients seizure free was 91% and 80% at 6 and 12 months, respectively. The significant variables of the scoring model were the unitemporal interictal(p=0.02) and ictal EEG discharges (p<0.03), and presence of an MRI lesion (p=0.02). The scoring model ranged from 11 to 85. The cut-off score for a good outcome was 60 with a sensitivity, specificity, and positive predictive value of 88, 55, and 95, respectively. CONCLUSIONS: The pre-operative scoring model for predicting the outcome of temporal lobe resective surgery was at least 80% accurate at 6 and 12 months post-surgery.