Authors :
Presenting Author: Lise Johnson, PhD – NeuroPace, Inc.
Thomas Tcheng, PhD – NeuroPace, Inc.
Jacob Norman, PhD – NeuroPace, Inc.
Muhammad Furqan Afzal, PhD – NeuroPace, Inc.
David Greene, BS – NeuroPace, Inc.
Aaron Warren, PhD – Brigham and Women's Hospital, Harvard Medical School
Zoé Dary, PhD – Brigham and Women's hospital, Harvard Medical School
John Rolston, MD, PhD – Brigham and Women's hospital, Harvard Medical School
Christopher Butson, PhD – University of Florida
Daniel Friedman, MD – Department of Neurology, New York University Grossman School of Medicine, NYU Langone Health
Vikram Rao, MD – Department of Neurology and Weill Institute for Neurosciences, University of California, San Francisco
Ji Yeoun Yoo, MD – Icahn School of Medicine at Mount Sinai, New York City
Katie Bullinger, MD, PhD – Emory University School of Medicine
Sydney Cash, MD, PhD – Massachusetts General Hospital
Jerzy Szaflarski, MD, PhD – University of Alabama at Birmingham
Andrew Cole, MD – Massachusetts General Hospital and Harvard Medical School
Mark Richardson, MD, PhD – Massachusetts General Hospital
Saadi Ghatan, MD – Mount Sinai - New York
Edward Chang, MD – University of California, San Francisco
Martha Morrell, MD – NeuroPace
Rationale: Seizure diary data is the gold standard for measuring clinical response in clinical trials, but these data are often incomplete and subject to bias. In a clinical trial evaluating thalamocortical responsive neurostimulation in Lennox-Gastaut Syndrome (LGS, NCT05339126), two bilaterally implanted neurostimulators continuously monitor brain data and record detections of epileptiform activity on thalamic and cortical leads. These device-recorded metrics may be predictive of clinical seizure frequency and could contribute to development of a complete and unbiased proxy for clinical outcomes.
Methods:
A correlation analysis was performed between numbers of clinical drop seizures (DS) recorded in the diary and device-recorded metrics over time. The absolute number of event detections is affected by device programming, therefore, the correlation analysis was conducted piece-wise during periods of time when device programming was constant.
The number of DS, long episodes (LEs), total detections, delta detections, and gamma detections on both thalamic and cortical leads were summed each day. A 3-day moving average was applied to all data and Pearson’s correlation was performed for 30-day epochs; significance was determined by building a distribution of correlations using randomly shuffled data (α=0.01).
The most valuable predictor for each participant was determined by selecting the variable which was significantly positively correlated for the largest percentage of 30-day epochs.
Results: The absolute number and percentage of epochs with significant correlations, and the particular variables demonstrating the most significant correlations, varied across participants. In some, many or most device detection metrics were generally positively correlated with DS frequency. In others, significant correlations were observed in only a small percentage of epochs. In some, there were variables that were significantly negatively correlated with DS frequency.
Conclusions:
This analysis demonstrates that device detections and LEs can function as biomarkers of changes in DS frequency. LEs were not always the best correlate of clinical outcomes; different detectors on different leads performed better or worse for different participants. This supports the use of personalized biomarkers in this population.
Differences between participants in consistency and degree of correlation between device recorded variables and diary recorded DS likely reflects a combination of variability in underlying physiology, variability in electrode placement, and variability in consistency and quality of diary data.
Observed correlations with device detections could allow evaluation of response to therapy without reliance on diary data and may facilitate discovery of additional biomarkers.
Funding: Research reported in this publication was supported by the National Institute Of Neurological Disorders And Stroke of the National Institutes of Health under Award Number UH3NS109557. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.