Stentrode SWITCH study (first in human)

One-line verdict: The safety result that justified recording from inside a blood vessel: 12 months, four patients, no serious adverse events, no vessel occlusion, no device migration.

Quick tags: Recording · Endovascular · Human · Study start 27 May 2019

Overview

What it is: The Stentrode With Thought-Controlled Digital Switch (SWITCH) study, a single-center, prospective, open-label first in-human study at the Royal Melbourne Hospital in Parkville, Australia. Five patients with severe bilateral upper-limb paralysis were enrolled: four with amyotrophic lateral sclerosis and one with primary lateral sclerosis. One was withdrawn because preimplant MRI showed isolated transverse sinus drainage. Four received the implant, delivered by a neurointerventional procedure into the superior sagittal sinus. The study started on 27 May 2019 and final follow-up was completed on 9 January 2022. For the later US trial, see the Synchron COMMAND entry.

Earlier report: A 2021 Journal of NeuroInterventional Surgery paper described two participants with ALS. It reported unsupervised home use from day 86 for participant 1 and day 71 for participant 2, and typing at 13.81 and 20.10 correct characters per minute with an eye tracker for cursor navigation, with predictive text disabled.

Results in the four-patient JAMA Neurology report (published January 2023):

• Safety: no device-related serious adverse events leading to death or permanent disability through 12 months. There were 8 mild adverse device effects, including headache and bruising at incision sites, that resolved without intervention. Patients 1 and 3 died of ALS complications unrelated to the device, 12 months and 1 month after completing the study.
• Device position: CT venograms at 3 and 12 months found no target vessel occlusion or thrombosis. The mean absolute change in distance between the distal electrode and a fiducial line was 0.45 mm (SD 0.28).
• Signal: mean signal bandwidth was 233 Hz (SD 16; range 206 to 245), stable for 12 months.
• Decoding: between 5 and 7 attempted movements, including rest, could be decoded in all patients. Offline accuracy was 85.2% for 2-class, 68.3% for 3-class and 41.1% for 5 or more classes.
• Use: each patient controlled a computer. With an eye-tracker, one reported analysis gave 93.9% selection accuracy and 16.6 correct characters per minute. Patients did texting, email, shopping and finance tasks. The final patient used the BCI alone, without the eye tracker, to type and hold a simple conversation.

Antiplatelet regimen: The authors used dual antiplatelet therapy for 3 months, with aspirin continued for at least 1 year.

Limits: Four patients, all with motor neuron disease. The authors attribute the lack of serious adverse events in part to selecting participants with adequate lung function and no comorbidities at enrollment.

Spec Card Grid

Trial

• Site: Royal Melbourne Hospital, Parkville, Australia
• Enrolled / implanted: 5 / 4
• Mean age at enrollment: 61 (SD 17)
• Follow-up: 12 months

Results

• Serious adverse events: none device-related
• Signal bandwidth: 233 Hz mean
• Position change: 0.45 mm mean
• Decoded movements: 5 to 7 per patient

Evidence and limits

• Not covered here: stent and electrode dimensions in this source; see the Stentrode entry for the device model

References

• Assessment of Safety of a Fully Implanted Endovascular Brain-Computer Interface for Severe Paralysis in 4 Patients: The Stentrode With Thought-Controlled Digital Switch (SWITCH) Study. JAMA Neurology, January 2023. https://jamanetwork.com/journals/jamaneurology/fullarticle/2799839
• Motor neuroprosthesis implanted with neurointerventional surgery improves capacity for activities of daily living tasks in severe paralysis: first in-human experience. Journal of NeuroInterventional Surgery 13(2):102. https://jnis.bmj.com/content/13/2/102