Abstracts

Rationale: The aim of this study is to investigate the usefulness of 18-FDG PET scan prior to an epilepsy surgery in non-lesional neocortical pediatric epilepsy.Methods: We performed a retrospective study on 35 intractable childhood epilepsy patients who received epilepsy surgeries from 2003 to 2010 at Severance Children s Hospital. None of the patients had any detectable lesions on MRI. Surgical outcomes were assessed at least 1 year postoperatively. Clinical characteristics, surgical outcomes, pathologic findings, types of surgery, and localizing features on 18-FDG PET, long-term video-EEG, and SPECT were reviewed.Results: Twenty five out of 35 patients (71%) showed abnormal focal localizing features on 18-FDG PET. In 10 epileptic encephalopathy patients, 3 out of 10 patients (70%) showed abnormal focal localizing features on 18-FDG PET. Thirteen patients out of 18 extra-temporal patients (72%) and 5 patients out of 7 patients (71%) showed abnormal focal localizing features on 18-FDG PET. 18-FDG PET yielded a sensitivity of 80.8%, specificity of 55.6% (p=0.040). Lobectomy or cortisectomy was performed according to 18-FDG PET findings concordant with results of long-term video-EEG and SPECT. Single lobectomy was the most common procedure and was performed in 14 patients (40%). Multi-lobar resection was performed in 13 patients (37%), and cortisectomy was done in 8 patients (23%). Cortical dysplasia was the most common finding during pathological examination and was observed in 17 (49%) patients. The mean follow-up duration was 4.1 years (range: 1.0~7.1 years). Twenty three patients (66%) demonstrated an ILAE pediatric class I & II outcomes during the initial 3 month postoperatively. Only 16 patients (46%) had remained in class I until the last follow-up. In epileptic encephalopathy patients, 70% had class I outcomes and 30% had class IV outcomes. Class I outcomes were 28% in temporal lobe epilepsy and 39% in extra-temporal patients. From 25 abnormal focal localized patients in 18-FDG PET, 60% had class I outcomes, while only 10% of other 10 patients with no localization showed class I outcomes. 84% of the group of 13 patients whose resection sites included all of the localized areas shown on PET had class I outcomes. On the other hand, in the other 12 patients, only 36% showed class I outcomes (p=0.001). Conclusions: In children with intractable epilepsy and an MRI demonstrating no abnormal lesions, 18-FDG PET image may be a reliable method for pre-surgical localization.