Abstracts

Rationale: Responsive neurostimulation (RNS) is FDA-approved as an adjunctive therapy for reducing focal-onset seizure frequency in adults with no more than two epileptogenic foci, who are refractory to antiepileptic medication, and who have frequent disabling seizures. Conventionally, RNS targets 1-2 seizure foci when these can be localized. However, some cases have broad lobar onsets, generalized onsets (off-label), or rapid propagation; these features suggest that targeting a broader network with stimulation may be beneficial. Some individuals with intractable epilepsy have been implanted with RNS electrodes in the anterior nucleus (ANT) or centromedian nucleus (CMT) of the thalamus. Thalamic RNS can record epileptiform activity in the thalamus and deliver stimulation to (hopefully) aid in the treatment of intractable epilepsy.

Methods: In this retrospective case series, we reviewed a database of 80 patients across the neurostimulation clinic at the Yale Comprehensive Epilepsy Center to identify all patients with the RNS system with thalamic electrodes. Selected patients underwent further analysis, including review of electrocorticographic (ECoG) brain activity, medical records, and patient reported seizure frequency.

Results: We identified 4 patients with the RNS system with thalamic electrodes. Patient 1 had a broad seizure onset that was maximal in the right frontoparietal region and underwent implantation of a right CMT depth and a parietal cortical strip. Seizure reduction was >50% after 3 years. Patient 2 had generalized onset and underwent bilateral ANT depth implantation. Seizure reduction was >90% after 4 years. Patient 3 had broad left temporal onset and underwent implantation of a left hippocampal depth and bilateral ANT depths (only left side connected). Seizure reduction was >90% for focal aware seizures but no change in focal impaired aware seizures after 20 months. Patient 4 had broad seizure onset that was maximal in the right frontoparietal region and underwent implantation of bilateral ANT depths (only right side connected) and a right parietal cortical strip. Seizure reduction was >50% after 8 months. In the 3 patients with recordings from non-thalamic and thalamic leads, onset was earlier in the hippocampus (1 case) or cortex (2 cases) in all seizures. Among these 3, there was a thalamic correlate of rhythmic activity within 1-2 seconds of seizure onset in 2 patients (either in ANT or CMT) or slow waves at onset (ANT; 1 patient), either of which could be used to detect seizures and trigger responsive stimulation.

Conclusions: RNS is feasible with thalamic electrodes which can be used with or without extrathalamic electrodes. Total seizure reduction >50% was observed in 4/4 patients and disabling seizure reduction >50% in 3/4 patients. Either ANT or CMT electrodes can record ictal patterns near the time of onset, and stimulation can be triggered by either thalamic or extrathalamic pattern detection.

Funding: Please list any funding that was received in support of this abstract.: C.G. Swebilius Trust.