[en] Objective: To assess the value of ¹⁸F-fluorodeoxyglucose (FDG) PET imaging in localizing the epileptic foci and its use in surgical therapy. Methods: PET brain imaging was performed in 110 patients with primary epilepsy. The images were analyzed by visual inspection and semi-quantitative measurement. The scalp electroencephalogram (EEG) was examined in all patients, among whom electrocorticogram (ECoG) or depth electroencephalogram (DEEG) was examined in 26 cases and MRI and(or) CT were performed in 66 cases. In 110 cases, temporal lobectomy was performed in 17 patients and X-ray or γ knife therapy was performed in 69 cases. The duration of follow-up lasted at least 1 year. Results: 88.2% cases had abnormal PET imaging among 110 cases including 94.8% hypometabolic foci and 5.2% hypermetabolic foci. PET was more sensitive than EEG and MRI (88.2%, 67.3% and 43.5% respectively, χ² were 13.88 and 24.17, all P <0.01). PET detected solitary lesion well in 60.8% patients, the result was also more sensitive than EEG (60.8% vs 35.1% , χ²=11.08, P<0.01). Compared with ECoG or DEEG, the sensitivity and accuracy of PET to localize the epileptic foci were 92% and 87%. 17 patients treated by temporal lobectomy and 69 patients by X-ray or γ knife therapy guided with PET, all got satisfactory results. Conclusion: ¹⁸F-FDG PET imaging is a sensitive modality in localizing epileptic foci and also useful to guide surgical and orientation radiation therapy. (authors)