Abstracts

Rationale: Decisions to proceed with epilepsy surgery is a challenge in children with medically refractory epilepsy and nonlocalizing MRI findings (normal, unilateral or bilateral multilobar abnormalities). We investigate whether subtraction ictal SPECT co-registered to MRI (SISCOM) is useful in localizing seizure onset and predicting post-surgical outcome in children with nonlocalizing MRI. Methods: A retrospective review was performed in all children with medically refractory epilepsy and nonlocalizing MRI scans, who underwent SISCOM prior to epilepsy surgery at the Mayo Clinic Rochester between 1996 and 2009. Patients with tuberous sclerosis complex were excluded. Favorable surgical outcome was defined by rare or seizure freedom (Engel s class I and II). Results: Of 50 children (male 68%) with nonlocalizing MRI findings, 19 patients had normal MRI, 21 had unilateral multilobar abnormalities, and 10 had bi-hemispheric lesions. The median age of seizure onset was 4.8 years; the median age at surgery was 13.8 years. Focal SISCOM abnormality was present in 72% of cases. SISCOM was concordant to scalp EEG findings in 17 (61%) of 28 patients who had a dominant ictal onset. SISCOM provided further localizing information in 14 (64%) of 22 patients who had no dominant ictal scalp EEG onset. Extratemporal resection was performed in 78% of patients, including hemispherectomy 6, multilobar resection 1, single lobar resection with or without additional corticectomy 7, and focal cortical resection in 25 patients. The median duration of post-surgical follow-up was 3.2 years, with overall favorable outcome in 46% of patients. Favorable outcome was seen in 7 (37%) patients with normal MRI, 3 (30%) patients with bi-hemispheric lesions, and 12 (57%) patients with unilateral multilobar abnormalities. Post-operative MRIs were available in 17 patients, and favorable outcomes were present in 6/12 (50%) patients with completely or partially resected versus 0/5 patients with non-resected SISCOM abnormality (one-sided Fisher s exact test p=0.07). Conclusions: Despite the limited overall favorable outcome among children with nonlocalizing MRI findings, surgical benefit may be achieved in about a third of patients with normal MRI or bi-hemispheric lesions, in more than half of patients with unilateral multilobar abnormalities. SISCOM may provide further localizing information for planning resective epilepsy surgery in which other noninvasive evaluations are nonlocalizing. Furthermore, resection of SISCOM abnormality appears predictive of favorable surgical outcome.