Abstracts

Rationale: Mesial temporal sclerosis (MTS) is the most common cause of adult epilepsy. Up to one third of patients with epilepsy will be refractory to medications and potentially require surgery to decrease seizure burden, including open resection or laser ablation. Surgical resection of mesial temporal structures is the current gold-standard of therapy, but it requires an open craniotomy and removal of lateral neocortex. Laser therapy is a less-invasive modality for mesial temporal ablation, however it still requires surgical intervention. MRI-guided focused ultrasound (MRgFUS) has emerged as a non-invasive treatment modality approved for the management of essential tremor. It is capable of creating small, targeted lesions in deep subcortical structures without the need for ionizing radiation or invasive surgical interventions. The use of focused ultrasound to disconnect the mesial temporal structures could represent a novel, non-invasive treatment for MTS. We therefore tested the feasibility of modeling the fornix-fimbriae using diffusion tensor imaging (DTI) and planning theoretical ablation sites using standard MRgFUS software.  Methods: Five patients with available CT and 3.0T brain MRI data were retrospectively reviewed. The study group consisted of a cohort of patients with essential tremor who underwent MRgFUS for their underlying condition.  DTI data associated with the 3T MRI study was processed, and tractograms were generated of the fornix-fimbriae and inferior optic radiations.  Fiber tracking was performed using Brainlab iPlanNet Cranial v3.0 (Munich, Germany).  Fractional anisotropy threshold was set to 0.2.  Minimum fiber length was 90 mm.  Optimal ablation sites were selected based on the patient-specific location of the tractograms to mimic surgical ablation sites targeted in open posterior hippocampal disconnection.  Fiber tracts were displayed three-dimensionally, and an ablation site was selected in consensus between a neurosurgeon and a neuroradiologist (Figure 1).  The MRgFUS procedure was then modeled in the same fashion as a clinical ablation case; a CT-MR fusion was performed and air-filled sinuses and calcifications were blocked using the Insightec® surgical planning software.  The ablation site was then demarcated and a theoretical helmet angulation was prescribed.  The number of available elements from the 1024-array were calculated. Results: The fornix-fimbriae and optic radiations were successfully fiber tracked in all patients. More than 900 elements were available for this target, indicating adequate ultrasound elements above the 800-element minimum recommended by the manufacturer.  Conclusions: MRgFUS may provide a minimally invasive option for seizure tract disruption and avoid the morbidity associated with craniotomy and deep cranial dissection. DTI can reliably identify the fornix-fimbriae to assist localization of the optimal ablation target as well as the critical surrounding structures such as the optic radiations which could potentially reduce the incidence and severity of postoperative visual field deficits. These findings may have utility for both MRgFUS procedure planning in surgically-naive patients and retreatment of patients who have failed prior epilepsy surgery. This theoretical modeling study is significant because it provides the necessary groundwork for future clinical trials on this novel neurosurgical technique applied to patients with refractory mesial temporal lobe epilepsy. The future challenge will be patient selection given the constraints of the limited area targeted by MRgFUS.  Funding: None