Abstracts

RATIONALE: From 1993 to 1999, 73 patients underwent temporal resection at UCLA based on MRI, PET, neuropsychological testing, Wada, and ictal video-EEG recording using scalp electrodes. Patients were required to have either a temporal MRI abnormality or PET hypometabolism, but not both, in order to proceed to surgery without depth electrode monitoring. An analysis was performed to determine seizure free status and the relative values of MRI, PET and pathology in predicting outcome. METHODS : Retrospective chart review was performed to obtain EEG, MRI, PET and surgical pathology data. Seizure free status was determined by phone or clinic interview. RESULTS: Of 69 patients with follow up, 61(88%) have been seizure free since surgery (duration 14-60 mo.) Temporal MRI abnormalities were present in 61 patients, of these 54 (89%)are seizure free. Surgery was performed in 8 patients with a normal MRI and 7 are seizure free. Temporal PET hypometabolism was evident in 64 patients. Of these 59 are seizure free. The 1 patient with a normal PET had surgery based on MRI and EEG findings and is also seizure free. Of 57 patients with both temporal MRI and PET abnormalities, 51(89%) are seizure free. Pathology included hippocampal sclerosis (HS) alone (42), dual pathology (HS+ another pathology n=5), and other (tumor, cortical dyspalsya, trauma and AVM n=9.) In patients with HS alone, 92% are seizure free. There was no pathology detected in 9 patients, and 8 are seizure free. CONCLUSIONS: Temporal lobe surgery based on concordant ictal scalp EEG with MRI and/or PET findings has an excellent outcome. Abnormal MRI or PET scan when used with ictal EEG predict a seizure free outcome at the same rate (89% vs 88%.) Having both abnormal PET and MRI did predict a better outcome than either test alone. Neither a normal PET nor MRI ruled out a good outcome. Surgical candidates with a normal MRI should receive a PET scan as the additional data identified 8 patients who proceeded to surgery without invasive monitoring and with good outcome. While HS is a good predictor of outcome, the absence of detectable pathology does not indicate poor outcome.