Thalamocortical Responsive Neurostimulation for the Treatment of Lennox-Gastaut Syndrome

点击翻译按钮获取中文摘要

Project Summary / Abstract Lennox-Gastaut Syndrome (LGS) is a devastating form of childhood onset epilepsy with cognitive dysfunction and very frequent generalized onset seizures (GOS) often leading to injury. Driven by the lack of effective therapies and the demonstrated safety and efficacy of brain-responsive stimulation for medically intractable focal onset seizures (FOS), this study will test whether brain-responsive neurostimulation of thalamocortical networks (RNS-TCN) is a feasible strategy to treat LGS. Specifically, the project aims to: (1) acquire preliminary evidence for safety and efficacy in treating GOS of LGS with RNS-TCN; (2) create an interactive therapy-decision support system based on patient-specific computational network models and machine learning to identify optimal lead placement and stimulation parameters. Using the RNS® System, which is FDA approved for FOS, an early feasibility IDE study will be conducted at 6 epilepsy centers in 20 patients with LGS and medically intractable GOS. The patients will be enrolled in two cohorts of 10, with safety and efficacy milestones in the first cohort governing the enrollment of the second cohort. Patients will have two depth leads placed bilaterally in the centromedian nucleus of the thalamus and two subdural strip leads placed bilaterally on the medial prefrontal cortex. These targets are selected because they are implicated in the onset and spread of GOS in LGS. Leads will be located within each target such that stimulation maximally engages the thalamocortical network, guided by finite-element biophysical models created from structural magnetic resonance imaging and diffusion weighted imaging. The finite-element biophysical models will also be used to identify the initial RNS-TCN stimulation pathway and current amplitude. During the blinded evaluation period patients will be randomized to receive either high-frequency short burst (HFSB) or low-frequency long burst (LFLB) RNS-TCN, then enter a washout period before crossing over to receive the other treatment condition. During the open label period stimulation parameters can be modified at the discretion of the physician. Parameter adjustments will be informed by using a Bayesian optimization model developed for specifically for each patient. All clinical and electrophysiological data collected during the study will be used to identify a biomarker of clinical response; if found, these will aid future epilepsy research and clinical practice. If safety is favorable and there is preliminary evidence for efficacy, then this early experience will inform the design of a future larger feasibility study. In addition to treating a population in need, this project engages in fundamental discovery of biomarkers in generalized network epilepsies, and develops novel automated therapy selection policies that have the potential to improve the lives of patients with LGS and other seizure disorders.