Abstracts

Surgical outcome for patients with refractory partial epilepsy secondary to head trauma has been previously reported to be less favorable when compared to other risk factors. Localization of an epileptogenic zone for focal resection in this population has been postulated to be difficult because of the diffuse nature of the initial brain injury and the potential for multiple sites of epileptogenicity. This study analyzes the localizing value of MRI and SPECT when determining surgical candidacy in this population. All patients in whom head injury (HI) was the sole risk factor for refractory partial epilepsy and who underwent epilepsy surgery at Henry Ford Hospital between 6/93 and 12/02 were identified. Patients were stratified into two subgroups: mild/moderate head injury ([lt]30 minutes alteration in consciousness), and severe head trauma ([gt]30 minutes alteration in consciousness). To determine accuracy of MRI, ictal SPECT, and interictal SPECT, only patients with well localized intracranial ictal patterns and Class I or Class II surgical outcomes were included in the final analysis. Localizing accuracy of imaging studies was based on determining concordance of imaging abnormalities with either the focal ictal onset identified by intracranial recording, and/or the focal resection. MRI abnormalities (gliosis, obvious atrophy, or encephalomalacia) either matched or overlapped the epileptogenic zone in 13/25 (52%) of patients. Accuracy of MRI was 46% in patients with mild/moderate head injury, and 60% in patients with severe head injury. Ictal SPECT showed an area of hyperperfusion in the region of the epileptogenic zone when compared to the interictal SPECT in 14/16 (87.5%) patients. Both incorrect studies were in patients with severe head injury. Interictal SPECT demonstrated a region of hypoperfusion matching the epileptogenic zone in 14/23 (61%) of subjects. Accuracy of interictal SPECT was 87% in severe HI, and 38% in mild/moderate HI. Although MRI and interictal SPECT studies in refractory posttraumatic epilepsy patients may show abnormalities concordant with the epileptogenic zone, localizing value may be limited or incorrect in some cases due to the multiplicity of the abnormalities. The maximum area of brain unjury determined by neuroimaging, or by reported neurologic deficits at the time of injury, may not correlate with the epileptogenic zone.