A Real-World Evidence Study to Support Label Expansion for the NeuroPace RNS System for Drug-Resistant Focal Epilepsy in Adolescent Patients in the United States

AURORA is a retrospective, observational, real-world evidence study run through OM1's Central Site model rather than a prospective trial infrastructure. NeuroPace identifies potentially eligible patients from its existing commercial implant records, stratified by age band (12-14 and 15-17 years) to ensure balanced representation, and nurse navigators conduct an initial prescreen before referring interested patients to OM1 for full informed consent, assent, and HIPAA authorization. EHR data are retrieved from each patient's own treating institutions into the OM1 Origin data warehouse, supplemented by patient-reported outcomes captured through OM1's EDC platform.

Eligible patients are 12 through 17 years old at the time of a single RNS Model 320 implant for focal-onset epilepsy, implanted on or before December 31, 2024, with no concurrent therapeutic surgery. Patients with primary or symmetric generalized epilepsy, developmental and epileptic encephalopathies, non-epileptic seizures, progressive CNS disease, concurrent investigational drug/device trial enrollment, or status epilepticus within three months of implant are excluded. Only sites with at least three patients at least two years post-implant may contribute patients.

Beyond the co-primary endpoints, secondary objectives include the 12-month RR50 under the same 31% performance goal, annual rates of pre-specified serious adverse events of interest, rates of seizure-related adverse events versus a 365-day pre-implant baseline, and SUDEP incidence at 12 and 24 months. Supportive/exploratory endpoints include a study-specific Seizure Frequency Burden Scale (for patients whose seizure counts can't be reliably abstracted from a diary) and a clinician-assessed global impression of seizure change modeled on the CGI-I construct.

The primary effectiveness analysis assumes a true RR50 of 46%, giving ~87% power to show the lower bound of the two-sided 95% exact CI exceeds the 31% performance goal - derived from NeuroPace's Pivotal Clinical Investigation in patients aged 18-21, the closest available comparator given no adolescent-specific external control exists. Because the data are retrospective, every seizure count is graded for reliability before analysis, and sensitivity analyses include multiple imputation (MAR) and inverse probability of censoring weighting for informative attrition. Adverse events are identified solely through a pre-specified ICD-10 code list applied to each patient's EHR, not prospective reporting, consistent with FDA RWE guidance. NeuroPace sponsors the study and leads governance and regulatory submission; OM1 runs Central Site operations and data management.