Abstracts

Rationale: Temporal lobe epilepsy (TLE) is the most common form of focal and refractory epilepsy. Seizures in patients with TLE often originate in the hippocampus and propagate to the rest of the brain via the main hippocampal output, the fornix. Fewer than 1.5% of refractory patients receive invasive surgery to resect the seizure focus, despite its success, with 62% of surgery patients seeing seizure-freedom. A modern noninvasive technology, focused ultrasound (FUS), when used to lesion the fornix, may provide patients with a surgical solution that addresses barriers to invasive surgeries and provides a similar benefit. Our long-term goal is to quantify the efficacy of lesioning the fornix via MR-guided FUS (MRgFUS) on seizures in a preclinical rodent model of TLE, prior to its use in humans. One challenge with transcranial FUS ablation studies in rodents is that the shape of the skull necessitates the use of small aperture transducers, which results in high acoustic power density on the skull. This, in turn, can result in substantial skull heating. In this study we overcame this obstacle by developing a craniectomy-approach in rats. For accurate and non-invasive FUS targeting, we developed a mechanical positioning system with sub-mm accuracy and used acoustic radiation force imaging to locate the focal spot without causing tissue damage. With the craniectomy, a high frequency transducer (3 MHz) with smaller focal size for precise targeting can be used without causing near-field heating. Proof of principle data in a naïve rodent cohort is provided. Future studies to test the hypothesis that MRgFUS-induced fornix lesions will prevent seizure propagation out of the temporal lobe are planned.

Methods: Male Sprague-Dawley rats (Charles River, n=11) were anesthetized with isoflurane and a 10 x 11 mm midline craniectomy, centered around bregma was performed directly prior to MRgFUS. Rats were head-fixed, mounted in a custom MRgFUS system with a 3 MHz annular transducer array (Image Guided Therapy (IGT), Bordeaux, France) to create lesions of the fornix (-0.60 A/P, 0 M/L, 5 D/V). Pre- and post-ablation MRI (3T Siemens Prisma) was performed (T2-weighted SPACE) to identify lesion size and location, and Magnetic Resonance Thermal Imaging (Thermoguide^TM, IGT) was used to monitor temperature rise in real-time. Rats were sacrificed immediately (n=7) or allowed to recover for 72 hours (n=4) for histological verification of lesion.

Results: Initial experiments (n=9), resulted in successful targeting and ablation of the fornix, but lesions were too large and resulted in collateral damage seen through histology. Thus, MRgFUS system upgrades were completed to optimize targeting, and sonication parameters were adjusted to create smaller lesions. Upgrades were recently tested (n=2) and allowed us to make sub-mm adjustments to the focal position and lesions were smaller with less collateral damage (Figure 1).

Conclusions: Following optimization, future studies will investigate the ability of precise fornix lesions to prevent seizures in a rat model of TLE.

Funding: Please list any funding that was received in support of this abstract.: NSF GRFP; Richard L. Stimson Endowed Chair; Research Incentive Seed Grant, University of Utah.