The predictive value of FDG-PET with 3D-SSP for surgical outcomes in patients with temporal lobe epilepsy.

PURPOSE: We retrospectively evaluated the diagnostic value of (18)F-2-fluorodeoxy-d-glucose positron emission tomography (FDG-PET) with statistical analysis for the foci detection and predictive utility for postsurgical seizure outcome of patients with mesial temporal lobe epilepsy (mTLE).

METHOD: We evaluated 40 patients who were diagnosed mTLE and underwent selective amygdalohippocampectomy (SAH) or anterior temporal lobectomy (ATL) in our institute. Preoperative interictal FDG-PET with statistical analysis using three-dimensional stereotactic surface projection (3D-SSP) was detected with several clinical data including seizure semiology, MRI, scalp electroencephalography, surgical procedure with SAH or ATL and postsurgical outcome. The region of interest (ROI) was defined on 'Hippocampus & Amygdala', 'Parahippocampal gyrus & Uncus', 'T1 & T2', and 'T3 & Fusiform gyrus'. We obtained the ratio of hypometabolism difference (RHD) by 3D-SSP, and evaluated the relation among hypometabolic extent, surgical outcome and surgical procedure.

RESULT: The RHD in each ROIs ipsilateral to operative side was significantly higher than that of contralateral side in good outcome group. Hypometabolism of 'Hippocampus & Amygdala' was most reliable prognostic factor. Patients of discordant with presurgical examinations hardly showed obvious lateralized hypometabolism. Nevertheless, when they have significantly high RHD in mesial temporal lobe, good surgical outcome was expected. There was not significant difference of RHD distribution between SAH and ATL in good outcome group.

CONCLUSION: Significant hypometabolism in mesial temporal lobe on FDG-PET with 3D-SSP is useful to predict good surgical outcome for patients with mTLE, particularly in discordant patients with hypometabolism in mesial temporal structure. However, FDG-PET is not indicative of surgical procedure.