Authors :
Presenting Author: Ann Mertens, MD – University of Gent
Ryan Verner, PhD – LivaNova PLC (or a subsidiary)
Giovanni Ranuzzi, PhD – LivaNova PLC
Roshani Patel, PhD – LivaNova PLC
Jason Begnaud, PhD – LivaNova PLC
Jerzy Szaflarski, MD, PhD – University of Alabama
Kristl Vonck, MD – University of Gent
William Tatum, MD – Mayo Clinic Florida
Lesley Kaye, MD – UC Health University of Colorado Hospital
Cornelia Drees, MD – Mayo Clinic
Muhammad Zafar, MD, FACNS – Duke University Hospital
Rebecca O'Dwyer, MD – Rush University Medical Center
Rationale:
Biomarkers of therapeutic engagement have emerged as an attractive approach to streamline the VNS titration process, providing immediate feedback on whether a individualized target dose of therapy has been achieved. In this work, we examine a brain-based biomarker (stimulation-evoked thalamic Brain-Oxygen-Level Dependent (BOLD) response) as a potential approach to personalize titration of investigational high-frequency VNS (“Microburst”).
Methods:
Patients enrolled into an investigational feasibility study of microburst VNS (NCT03446664). As part of this study, participants had their investigational Microburst VNS generator titrated using an exploratory, fMRI-guided titration approach that involved receiving VNS during the fMRI scan procedure. Participants were exposed to a series of scanning sequences, and each sequence tested a set of 6 different parameter combinations in a traditional fMRI ON-OFF block experimental design. Evoked thalamic BOLD responses to certain microburst VNS parameters were assessed and those parameters that evoked the highest thalamic BOLD response were assigned to the participant’s VNS device. Allowance was made for adjustments based on tolerability. In this exploratory analysis, we examine the impact of this fMRI-guided titration process.
Results:
Thirty-one participants attended the 6-month study visit and at least one prior fMRI study visit. In all participants, the fMRI data collection method and analysis approach revealed significant increases in thalamic BOLD responses. Intrasubject variability in thalamic response between visits was generally low, however, the intrasubject variability was typically larger than the variance between thalamic response to different parameter sets. Participants who were most likely to respond to microburst VNS were those who experienced significant thalamic BOLD response increases among multiple stimulation parameters over multiple visits. While the study outcome doesn’t provide statistically significant justification for prioritization of specific Microburst parameter sets at the population level, the parameters 300 Hz signal frequency and 0.5 second interburst intervals were most commonly selected as “winning” features across the population. Generally, participants responded to microburst VNS using thalamic blood flow as a surrogate biomarker, with a median reduction in seizure frequency of 42.6% at 6 months. There were no MRI-related safety issues associated with active VNS during scanning.
Conclusions:
Titration of microburst VNS by adjusting parameters to maximize the thalamic BOLD response was safe and feasible in our investigational paradigm. While intrasubject variance was too high to determine a specific optimal microburst setting for each patient, a more reliable predictor of outcome was whether a thalamic BOLD signal increase occurred across multiple microburst parameters, rather than a single parameter.
Funding: LivaNova PLC