Abstracts

Rationale: Positron emission tomography (PET) and single-photon emission computed tomography (SPECT) are important diagnostic tools in the evaluation for epilepsy surgery. Both techniques have advantages and limitations. SPECT allows visualization of ictal activity, but has poor temporal resolution and often detects seizure propagation rather than onset. Interical PET is less sensitive to the timing of injection, but may occasionally have less spatial resolution. In this study we correlated PET and SPECT findings in children with focal intractable epilepsy with MRI and EEG, comparing between lesional and non-lesional epilepsy, and between temporal and extra-temporal epilepsy. Methods: We retrospectively reviewed data on noninvasive presurgical evaluations in children with intractable pediatric epilepsy, over a 2 year period (January 2012-December 2013) at Boston Children's Hospital. This study was approved by the institutional review board. The data collection included patient demographics and clinical information, cause of epilepsy, and EEG, MRI, PET, ictal and inter-ictal SPECT results. Results: 171 patients underwent non-invasive presurgical evaluation during the study period (mean age 10.3 years, 4 weeks to 25 years). PET was obtained on 117 patients, and SPECTs (both ictal and interictal) on 131 patients. 97 patients had both PET and SPECTs. Concordance was categorized as 1) concordant (same laterality as MRI and/or EEG), 2) discordant, or 3) non-localizing / lateralizing (generalized, multifocal). PET was concordant in 58.9%, discordant in 9.4%, non-localizing /lateralizing in 31.6% of the cases. SPECT was concordant in 46.2%, discordant in 19.7%, and non-localizing in 34.1%. Comparing for lesional vs non-lesional epilepsies, PET was concordant with MRI/EEG in 65.5% v/s 39.4%, discordant in 4.8% v/s 27.3% and non-localizing /lateralizing in 29.8% v/s 33.3% of cases. SPECT was concordant with MRI /EEG in 50.5% v/s 27.3%, discordant in 16.8% v/s 27.3%, and non-localizing /lateralizing in 32.6% v/s 45.5%. Comparing for temporal v/s extratemporal epilepsy, PET was concordant in 87.5% v/s 50%, discordant in 2.5% v/s 9.3%, and non-localizing /lateralizing in 10% v/s 40.6%. SPECT concordance was found in 75.6% v/s 36.6%, discordance in 4.7% v/s 25.4%, and localizing /lateralizing n 23.3% v/s 38% respectively. Among concordant SPECT studies in temporal epilepsy, 30/31 showed increased uptake within specific regions in the temporal lobe, while 8/26 of the extra-temporal SPECTS were lobe specific. 97 patients underwent both PET and SPECT studies; both were concordant with MRI/EEG in 36% of cases, PET alone in 16.5%, and SPECT alone in 6.2%. . Conclusions: PET had higher concordance and lower discordance with MRI/EEG compared to ictal SPECT. Rate of non-localizing /lateralizing studies was similar in both. SPECT was more helpful for localization in temporal lobe epilepsy. Combination of both modalities adds information to non-invasive pre-surgical evaluation for seizure localization, as each modality has strengths and weaknesses. Further studies will be necessary to confirm the findings.