Exploratory definition of a clinically meaningful change in intracranial EEG features: Analysis of data from a long-term treatment study of the RNS® System.

OBJECTIVE: Novel proof-of-concept designs utilizing biomarkers associated with an established clinically meaningful outcome are needed for antiseizure medication (ASM) efficacy assessment in patients with epilepsy, including drug-resistant focal epilepsy. Changes in long episode (LE) frequency, recorded by the responsive neurostimulator (RNS® System), are associated with changes in clinical seizure (CS) frequency following ASM initiation, suggesting that LEs may serve as a valid biomarker. Here, we aimed to establish the LE reduction magnitude associated with a clinically meaningful CS reduction. METHODS: Post hoc retrospective analyses of CS and LE frequencies, responder rate, and receiver operating characteristics from a long-term RNS open-label treatment study in patients who met predefined criteria were performed to evaluate the association between changes in LE and CS frequency and the LE reduction cutpoint associated with a ≥ 50% CS reduction. RESULTS: Patients (N=45) initiated clobazam (n=15), levetiracetam (n=4), or lacosamide (n=26). Patients achieved a median 30% LE and 50% CS reduction. Patients with <50% and ≥50% CS reductions had a median 5.9% LE increase and 52.4% LE reduction, respectively. A positive association between LE and CS frequency reduction across the entire range of responses (0-100%) was observed. A ≥30% LE reduction was associated with a ≥50% CS reduction; AUC=0.765, N=45. SIGNIFICANCE: If LE reductions may serve as a potential biomarker, a threshold associated with meaningful improvement in CSs is paramount. Results from this hypothesis-generating analysis support the use of a ≥ 30% LE reduction as a biomarker associated with clinically meaningful improvement in focal epilepsy.