Abstracts

Rationale: Subtraction ictal SPECT co-registered to MRI (SISCOM) provides functional localization of the probable epileptogenic zone. While the localizing value of SISCOM has been well studied in highly selected surgical patients, its ability to generate a hypothesis for electrode implantation or resective surgery in unselected patients is less studied, with yields of 42-55% in combined adult and pediatric cases. We evaluated the localizing value of focal SISCOM abnormality in all children with nonlocalizing MRI who underwent SISCOM at a single institution. Methods: Children less than 18 years admitted for SISCOM were identified from the Mayo Clinic Rochester Epilepsy Monitoring Units Database between 2000 and 2009. Retrospective review of presurgical SISCOM, scalp ictal EEG, multidisciplinary epilepsy surgery conference consensus, intracranial EEG monitoring, resective surgery, and surgical outcome were abstracted. For children with focal SISCOM who underwent resective surgery, surgical outcome at last follow-up was defined as favorable (Engel class I and II) versus unfavorable (Engel class III and IV). The associations of SISCOM-EEG concordance with multidisciplinary conference consensus and with surgical outcome were assessed using chi-square analysis. Results: 186 patients with nonfocal MRI findings underwent SISCOM. SISCOM revealed a focal abnormality in 87 (47%), a lateralized abnormality in 12 (6%), and was nonlocalizing in 87 patients (47%). Scalp ictal EEG onset was focal in 52 patients (28%), hemispheric in 26 patients (14%), and nonlocalizing in 108 patients (58%). Among the 28 patients with both focal SISCOM and focal ictal EEG, these studies were concordant in 22 and discordant in 6 patients. Focal SISCOM provided further localization in 17 patients with hemispheric ictal EEG, in whom ipsilateral SISCOM-EEG concordance was present in 16 patients. Focal SISCOM abnormality was seen 42/108 patients with nonlocalizing ictal EEG. Among all patients with focal SISCOM, consensus for non-palliative resective surgery or intracranial monitoring was reached in 43 patients, of whom 29 proceeded with consensus recommendation. Resective surgery was deemed unsuitable in 5 patients after nonlocalizing ictal intracranial monitoring results, while the remaining 24 proceeded with resection. Favorable outcome at last follow up was achieved in 13/24 patients, including 6 patients in whom focal SISCOM was present despite nonlocalizing EEG findings. SISCOM-EEG concordance was not significantly correlated with multidisciplinary conference consensus or surgical outcome. Conclusions: In an unselected group of children with medically refractory epilepsy and nonfocal brain MRI findings, SISCOM generated reasonable hypothesis for electrode implantation or resective surgery in about 50% of those patients with focal SISCOM. Concordance was present in three quarters of those patients who had focal SISCOM and focal EEG findings. Although the overall outcome is still suboptimal, presurgical evaluations including SISCOM should be considered and family counselled regarding the likelihood of favorable outcome.