We assessed the relationship between temporal lobe metabolism measured quantitatively and qualitatively with PET using [18F]-fluorodeoxyglucose (FDG) and postoperative seizure frequency after anterior temporal lobectomy. Forty-three patients with refractory partial epilepsy had anterior temporal lobectomy and preoperative assessment with PET-FDG. Qualitative PET analysis was performed visually by two blinded observers, and quantitative PET analysis was performed using an anatomic template for six control and six temporal lobe subregions, deriving an asymmetry index for each region. Seizure outcome was assessed 1 year after surgery; patients were classified as being seizure-free or as having persistent seizures. Qualitative data were analyzed using Fisher's exact test and the t test, and quantitative data were analyzed using a repeated-measures ANOVA. Thirty-two patients (74%) were seizure-free at follow-up, and 11 had persistent seizures, although most improved. Twenty-nine of 35 patients (83%) with restricted temporal lobe hypometabolism by visual analysis were seizure-free, compared with three of eight patients (37.5%) with normal scans or multilobar hypometabolism. Quantitative analysis revealed that an asymmetry of mesial temporal lobe glucose consumption (uncal region) correlated with improved surgical outcome (p < 0.02). We developed a logistic regression model to predict individual outcome based on the asymmetry in uncal metabolism. Lateral temporal metabolism did not correlate with outcome. We conclude that both visual PET analysis and quantitative PET analysis predict outcome after temporal lobectomy, although quantitative measures offer more precise information.