We used positron emission tomography (PET) and bolus injection of H2(15O) to measure interictal cerebral blood flow (CBF) in 32 patients who had temporal lobectomy for uncontrolled complex partial seizures. Seizure focus localization was confirmed by ictal video-electroencephalographic (video-EEG) telemetry, and patients who had imaging findings suggesting the presence of a tumor were excluded. PET-CBF studies were interpreted using a standard template by raters blinded to EEG and clinical data, and were not used in surgical planning. After surgery, patients were followed for 32 +/- 18 months. Twenty-six patients were seizure free and 6 had persistent seizures. Mean lateral temporal hypoperfusion was 8 +/- 14% in patients who became seizure free and 9 +/- 6% in patients with persistent seizures. Fourteen patients had at least 15%, and 11 at least 20% hypoperfusion, but were not more likely to be seizure free. Three additional patients had 15 to 20% hypoperfusion in the temporal lobe contralateral to their EEG focus. PET measurement of CBF using bolus H2(15O) should not be used to help select patients for temporal lobectomy.