We investigated the relationship between the presence of extratemporal hypometabolism on fluorine-18 fluorodeoxyglucose positron emission tomography (FDG PET) and seizure outcome after temporal lobectomy in patients with medically intractable temporal lobe epilepsy (TLE). In 47 patients with intractable unilateral mesial TLE, regional metabolic changes on FDG PET images obtained during the 2 months preceding anterior temporal lobectomy were compared with postoperative seizure outcome. Postoperative seizure outcome was evaluated with a mean follow-up period of 6.1+/-0.6 years (range 5.2-7.2 years). Forty-two (89%) of the 47 patients achieved a good postoperative seizure outcome (Engel class I or II). All patients had hypometabolism in the temporal cortex ipsilateral to the epileptogenic region on FDG PET scans. Fourteen (78%) of the 18 patients with hypometabolism only in the ipsilateral temporal cortex were completely seizure free (Engel class Ia) after surgery. In contrast, five (45%) of the 11 patients with extratemporal cortical hypometabolism confined to the ipsilateral cerebral hemisphere and only four (22%) of the 18 patients with hypometabolism in the contralateral cerebral cortex were completely seizure free after surgery. The postoperative seizure-free rates were significantly different across the three groups of patients with different cortical metabolic patterns ( P<0.005). Furthermore, all of the nine patients with a non-class I outcome (Engel class II-IV) had extratemporal (including contralateral temporal) cortical hypometabolism. Thalamic hypometabolism was noted in 20 (43%) of the 47 patients (ipsilateral in 12, bilateral in 8). Sixteen (59%) of the 27 patients with normal thalamic metabolism were completely seizure free after surgery, while only seven (35%) of the 20 patients with thalamic hypometabolism became completely seizure free ( P<0.05). Multivariate analysis revealed that among variables including clinical, EEG, magnetic resonance imaging, pathological and FDG PET metabolic findings, only cortical metabolic pattern was an independent factor for the prediction of postoperative seizure outcome ( P<0.005). It is concluded that extratemporal cortical hypometabolism outside the seizure focus, in particular hypometabolism in the contralateral cerebral cortex, may be associated with a poorer postoperative seizure outcome in TLE and may represent underlying pathology that is potentially epileptogenic. Thalamic hypometabolism, which was associated, but not independently, with a higher likelihood of postoperative seizures, may be secondary to extratemporal or temporal pathology.