SAFETY AND PRELIMINARY EFFICACY OF THE RNS™ RESPONSIVE NEUROSTIMULATOR FOR THE TREATMENT OF INTRACTABLE EPILEPSY IN ADULTS
Gregory L. Barkley, Brien Smith, Gregory Bergey, Gregory Worrell, David Chabolla, Joseph Drazkowski, Douglas Labar, Robert Duckrow, Anthony Murro, Michael Smith, Ryder Gwinn, Bruce Fisch, Lawrence Hirsch, and Martha Morrell
Rationale: A multi-center feasibility investigation assessed safety and possible efficacy of the cranially based implantable, programmable Responsive Neurostimulator (RNS™) system.
Methods: Subjects were 18-65 years with intractable partial-onset seizures and localized epileptogenic onset region(s). Subjects with >12 simple partial (SP) sensory or motor seizures, complex partial seizures (CPS) or generalized tonic-clonic (GTC) seizures over an 84-day baseline period qualified for implant. The RNS was connected to up to 2 leads (subdural and/or depth), which were targeted to the seizure focus. Adverse events (AEs) were monitored throughout the trial. Efficacy was assessed during two time periods, the 84-day period beginning 28 days post-implant (primary) and the most recent 84 days for which a subject could have received therapy (secondary).
Results: During the primary evaluation period, the responder rate (>50% reduction in seizures) in 50 subjects (excluding 1 subject with no disabling seizures at baseline and 14 subjects blinded off) was 32% for CPS (n=44), 63% for GTC (n=16), and 26% (n=50) for total disabling seizures (TDS) (SP motor, CPS and GTC). The median percentage reduction in seizure frequency was: CPS 27%, GTC 59% and TDS 29%; seizure reduction was significant for CPS (p<0.05) and TDS (p<0.001) (Wilcoxon signed-rank test).
For the secondary evaluation period, as of 4/28/06, the responder rate for 62 subjects (excluding 1 subject with no disabling seizures at baseline and 2 subjects previously blinded off with fewer than 84-days of therapy on) was 40% for CPS (n=52), 55% for GTC (n=22), and 41% for TDS (n=61). The median percentage reduction was: CPS 34%, GTC 66% and TDS 35%, and seizure reduction was significant for CPS (p<0.05), GTC (p<0.005), and TDS (p<0.001).
In 65 implanted subjects (including 17 device replacements) representing 65 patient years, there were no serious unanticipated device-related AEs, and responsive neurostimulation was well tolerated.
Conclusion: An investigation of the RNS™ system demonstrated safety and a sustained reduction in CPS, GTC and TDS events. Preliminary results indicated that the RNS™ system may provide a safe and effective treatment for adults with intractable partial-onset epilepsy.
Supported by NeuroPace, Inc.
 Return to Top
TREATMENT OF MESIAL TEMPORAL LOBE EPILEPSY WITH RESPONSIVE HIPPOCAMPAL STIMULATION BY THE RNS™ NEUROSTIMULATOR
David R. Chabolla, Anthony M. Murro, Robert R. Goodman, Gregory L. Barkley, Gregory A. Worrell, Joseph F. Drazkowski, Douglas R. Labar, Michael C. Smith, Ryder Gwinn, Felice Sun, Rosana Esteller, and Martha J. Morrell
Rationale: Responsive neurostimulation with the RNS™ system is being investigated as a treatment for medically intractable partial onset seizures in a feasibility trial. Intractable epilepsy of unilateral or bilateral mesial temporal origin is a common syndrome. This analysis assessed safety and whether there is preliminary evidence for efficacy of treatment with the RNS™ neurostimulator in treating MTLE.
Methods: Subjects were adults with unilateral or bilateral (MTLE) who did not have a suspected extrahippocampal seizure focus and no prior resective epilepsy surgery. Analyses included efficacy as a function of unilateral vs. bilateral hippocampal onset (using a permutation test), and correlation of demographic variables such as age at implant, onset age, gender, and etiology with efficacy (by determining the corresponding correlation coefficients and their significance using a t-statistic under the null hypothesis of no correlation). Efficacy was evaluated by measuring the percent change in total disabling seizures (TDS; consisting of SP motor, CPS, and GTC) with respect to baseline during two post-implant periods; the 84-days period beginning 28-days post-implant (EVAL), and the most recent 84-day period during which a patient could receive therapy (MR84D).
Results: For the 18 patients (7 male), the average age at implant was 32 yrs (19-56 yrs), and the mean onset age was 18.6 yrs (0-46yrs). The mean follow-up time was 481 days (210-681 days). One anticipated serious adverse event was reported (depression) in a subject with bilateral leads; relationship to the RNS was uncertain. There were no device related serious adverse events. Subjects with bilateral leads experienced a 19% reduction in TDS during the EVAL period and 53% reduction during the MR84D period. Subjects with unilateral leads experienced a 31% reduction in TDS during the EVAL period and 43% reduction during the MR84D period. The difference in seizure response between unilateral and bilateral subjects was not significant (p =0.25). No correlation was observed between seizure response and subject age (p=0.96, r=0.014), onset age (p=0.56, r=-0.15), gender (p=0.52, r=0.16), or etiology (p=0.58, r=0.14).
Conclusion: Preliminary results suggest that responsive stimulation is safe and may have efficacy in treating seizures in patients with MTLE. Results also suggest that the RNS system may be a treatment option for patients with intractable bilateral MTLE for whom resective surgery is not an option.
Supported by NeuroPace, Inc.
Return to Top
SENSITIVITY OF DEPTH VERSUS STRIP ELECTRODES IN THE DETECTION OF HIPPOCAMPAL ICTAL ONSETS BY THE RNS™ RESPONSIVE NEUROSTIMULATOR
Robert R. Goodman, Gregory Bergey, Gregory Worrell, David Chabolla, Joseph Drazkowski, Douglas Labar, Robert Duckrow, Anthony Murro, Michael Smith, Ryder Gwinn, Gregory L. Barkley, and Martha Morrell
Rationale: Patients diagnosed with mesial temporal lobe epilepsy (MTLE) were implanted with the RNS™ responsive neurostimulator as part of a larger multicenter feasibility study to assess safety and potential efficacy. This analysis examines the onset patterns and seizure frequency associated with lead type.
Methods: 20 of 65 subjects enrolled in an RNS Investigation were diagnosed with intractable MTLE and are evaluated here. Hippocampal depth and/or temporal subdural leads were implanted chronically to monitor electrographic activity from and deliver stimulation to the presumed seizure onset region(s). Onset patterns were assessed via review of electrographic activity stored by the RNS; clinical response was evaluated by comparing seizure count over the most recent 84 days (post-implant) to an 84-day pre-implant baseline.
Results: Follow-up was 6 months to 2 years. There were no serious, unanticipated device-related adverse events.
| Lead Configuration and Efficacy* |
| Unilateral |
Mean % change (range) |
N |
| Hippocampal depth only |
-14.3% |
1 |
| Hippocampal depth + Subtemporal strip |
-56.7% (-42.0%,-68.2%) |
5 |
| Hippocampal depth + Lateral temp strip |
-17.4% |
1 |
| Subtemporal strip + Lateral temp strip |
+31.3% |
1 |
| BLANK |
BLANK |
BLANK |
| Bilateral |
Mean % change (range) |
N |
| Hippocampal depths only |
-58.8% (-14.7%,-100%) |
7 |
| Hippocampal depth + Subtemporal strip |
-33.3% |
1 |
| Subtemporal strips only |
-40.7% (-31.4%,-50.0%) |
2 |
|
*Excludes 2 subjects with fewer than 84 days of therapy ON.
In most subjects implanted unilaterally, the earliest onsets were on the depth lead except in 2 cases, the onset appeared earlier in the subtemporal strip with rapid spread to the depth. In the bilateral cases, earliest onsets were observed on both leads.
Eighty percent of subjects had at least one depth lead. The most common onset patterns were alpha/beta (48% of depth onsets, 69% of strip onsets), 1-2 Hz spiking (44% of depth and 23% of strip onsets) and low-voltage fast activity (8% of depth and strip onsets).
Conclusion: Although the sample size of subjects with intractable MTLE is small, detection and stimulation provided by depth lead(s) was associated with a greater reduction in seizures than with strip electrodes alone. Depth electrodes may also be more sensitive than strip electrodes in detecting ictal spiking.
Supported by NeuroPace, Inc.
Return to Top
HEALTH-RELATED QUALITY OF LIFE RESULTS IN ADULTS WITH INTRACTABLE EPILEPSY IMPLANTED WITH THE CRANIALLY BASED, PROGRAMMABLE RNS™ RESPONSIVE NEUROSTIMULATOR
Joseph Drazkowski, Gregory Barkley, Gregory Bergey, Gregory Worrell, David Chabolla, Douglas Labar, Robert Duckrow, Anthony Murro, Michael Smith, Ryder Gwinn, Lawrence Hirsch, Bruce Fisch, Janet Greenwood, and Martha Morrell.
Rationale: Health-related quality of life (HRQOL) is adversely impacted for many persons with medically intractable epilepsy. A validated inventory (QOLIE-89) has been developed to assess HRQOL in people with epilepsy and is incorporated into many new epilepsy therapy trials as an important measure of outcome.
Methods: 65 subjects participated in a feasibility investigation of the RNS™ neurostimulator. The QOLIE-89 was administered to each subject before implantation of the device and at one year and two years after implantation. Nine self-report domains of health concepts were evaluated. Data for 41 subjects were available for the one-year analysis. For this preliminary analysis, a two-sided paired t-test using a significance threshold of p<0.1 was performed. The correlation between QOLIE scores and reduction in seizure frequency was assessed using the Pearson correlation coefficient and corresponding t-statistic.
Results: QOLIE scores for the visit prior to the RNS implant and at one year post-implant were compared. Eight of the nine domains did not change. The energy/fatigue measure showed a trend towards improvement (p<0.09). There was a trend of correspondence (p<0.1) between a decrease in seizures at 12 months with an increase in QOLIE scores for physical function and social function.
Conclusion: In this preliminary analysis, self reported HRQOL shows a trend towards improvement at one year in those with reduced seizures. HRQOL scores in this cohort are similar to those reported in patients with medically intractable partial epilepsy1. Longer follow up may be required to demonstrate HRQOL improvements in association with seizure reduction.
1Vickery B, Hays RD, Rausch R et al. Quality of life of epilepsy surgery patients compared with outpatients with hypertension, diabetes, heart disease and/or depressive symptoms. Epilepsia 1994; 35(3): 597.
Supported by NeuroPace, Inc.
Return to Top
CIRCADIAN VARIATION IN AN ELECTROCORTICOGRAPHIC FEATURE DETECTED BY A RESPONSIVE NEUROSTIMULATOR SYSTEM
Robert B. Duckrow, and Tom Tcheng
Rationale: An implanted responsive neurostimulator system (RNS™, NeuroPace, Inc.) is being developed to treat localization-related epilepsy. The recording and analysis functions of this device detect localized epileptiform activity and trigger responsive electrical stimulation to abort a developing seizure. Algorithmic feature detectors in this device provide measures of ongoing electrocorticographic (ECoG) activity. Based on the observation that epileptic seizures can occur at specific times of the day, we hypothesized that one such feature of the baseline ECoG, the line length, would show circadian variation and that this variation could be interpreted in light of what is known about cortical activity or seizure susceptibility.
Methods: Data were collected during intracranial monitoring of patients with medically intractable epilepsy as part of a safety and feasibility trial of an external responsive neurostimulator system (eRNS). Using bipolar recordings, the eRNS sampled a total of 87 brain regions from 24 consecutive patients. The ECoG was characterized by calculating line length, a feature sensitive to high frequency content or signal complexity and defined for a uniformly sampled time series as the normalized sum of the absolute value of the difference between consecutive elements. The feature was derived from 32.7 second-long baseline segments that preceded interictal activity during 2 to 54 day monitoring sessions. When collapsed to a 24-hour cycle, a median of 494 values per recording site were available to be fit with a cosine function. These functions were analyzed with respect to time of peak and cerebral location.
Results: Line length values were represented by a cosine function (F<0.05) in 41 of 54 valid data sets (76%). The average daily variation was [plusmn]7% of the mean. The time of the peak had a bimodal distribution with local maxima at 0530 and 1500 hours. There was no segregation of the peak to day or night based on recordings from cerebral lobes. However, all data sets obtained from lateral or inferior temporal lobe had a peak during the day while none occurred at night (Fisher's exact test: p=0.03). Peak times from mesial temporal sites were evenly distributed over day and night.
Conclusion: The line length feature of the ECoG shows circadian (nychthemeral) variation. The segregation of peak amplitude to either 0530 or 1500 hours suggests a complex modulation of the ECoG that is independent of the sleep cycle but possibly relates to known fluctuations in spectral energy measured from the scalp during wakefulness. Speculative correlations can be made with the propensity of certain seizures types to occur at specific times of the day. This study shows how implanted devices could be used to provide chronic outpatient recording of similar features to monitor seizure susceptibility.
Supported by NeuroPace, Inc.
Return to Top
EXTERNAL RESPONSIVE NEUROSTIMULATION SYSTEM (eRNS) EFFECTS ON POWER AND COHERENCE SPECTRA OF INTRACRANIAL ELECTROENCEPHALOGRAPHY
Erik J. Kobylarz, Nicholas D. Schiff, Theodore H. Schwartz, and Douglas R. Labar
Rationale: A novel implantable closed-loop Responsive Neurostimulation System (RNS, NeuroPace, Inc., Mountain View, CA) consisting of pulse generator, subdural strip/depth electrodes and programmer is undergoing a prospective multicenter clinical trial. RNS continuously analyzes electrocorticograms (ECoGs) and automatically applies electrical stimulation when specific epileptiform activity is detected. This study has shown that RNS is safe for patients and can reduce partial seizure frequency and severity. However, the exact mechanism of action of RNS is unknown.
There have been no published frequency analyses of EEG with RNS. Power spectra demonstrate relative power of regional cortical activity at a broad range of frequencies. Coherence, an index of functional connectivity, is the cross-correlation of inter-regional frequency content. We analyze acute effects of RNS on interictal ECoG power spectra and coherence.
Methods: Three adult patients with refractory seizures were admitted to the epilepsy monitoring unit for intracranial EEG monitoring for epilepsy surgery evaluation. ECoG was recorded with subdural electrodes placed over the most active cortical regions. An externalized RNS (eRNS) device was connected to the most active intracranial electrodes for epileptiform activity detection. Subjects received neurostimulation in response to epileptiform discharges. Continuous video-ECoG was recorded, prior to turning RNS ON, then for 12-24 hours afterwards. ECoG during awake periods was analyzed, immediately before and after the RNS ON period. Power and coherence spectra during these epochs were compared for regions with greatest interictal epileptiform activity.
Results: In all patients a broad reduction in ECoG power occurred in the most active electrodes after RNS ON. In 2/3 patients coherence between adjacent electrode pairs was decreased for lower frequencies (1-5, 1-20 Hz) after RNS ON and increased at higher frequencies (45-100 Hz). In the 3rd patient the coherence was increased for lower frequencies (1-5 Hz) and decreased for higher frequencies (15-100 Hz).
Conclusion:
1. RNS had similar effects on ECoG power spectra for all patients.
2. ECoG coherence changed after RNS at most frequencies for all patients, although direction was different for 1/3. This may reflect different mechanisms of action for RNS.
3. The change in intra- and inter-regional ECoG frequency content could reflect neuromodulatory effects from neurostimulation.
4. Further ECoG frequency analyses to determine different patterns of change with RNS in regions adjacent to and remote from epileptic foci could be useful. These results will be
presented.
5. Further analyses could elucidate what response patterns can predict RNS efficacy.
Supported by NeuroPace, Inc.
Return to Top
PREOPERATIVE DEPTH LEAD PLACEMENT PLANNING TO ACTIVATE DISTANT CORTEX REPLICATES VISUAL AURA SEMIOLOGY DURING RESPONSIVE NEUROSTIMULATION (RNS)
Marvin A. Rossi, Thomas J. Hoeppner, Travis Stoub, Andres M. Kanner, Richard Byrne, Donna Bergen, and Michael C. Smith
Rationale: A novel approach for implanting investigational RNS (NeuroPace, Inc) electrodes is presented for optimizing overlap of the epileptic circuit. White matter pathways are targeted such that the biophysical properties of myelinated axons are used to propagate electrical current distant from the source of stimulation. This hypothesis was tested post-RNS implant by reproducing a patient's typical visual aura when stimulating depth contacts implanted in parahippocampal white matter. RNS current was 'injected' during the aura onset while capturing resulting transient blood flow changes using subtracted activated SPECT (SAS).
Methods: RK (age: 38 years) was stereotactically implanted with a 4-contact depth lead placed completely within the left parahippocampal white matter. In addition, a 4-contact subdural strip was placed in the left mesial occipital region followed by securing a skull-based RNS device. Multiple, presurgically acquired subtracted ictal SPECT studies aided localization for lead placement. An SAS study was performed at 18 months post-implant following delivery of stimulation current without producing an afterdischarge. Pre-implant diffusion tensor imaging was used to compare white matter orientation and distribution with SAS findings.
Results: Preliminary results obtained from RK during stimulation of the depth electrode suggest distant neuronal spread of current. That is, upon repetitive bipolar stimulation of posterior mesial temporal white matter depth contacts, propagation to primary visual cortex was observed by semiology and blood flow measures. Specifically, a reproducible elementary visual response was reported by the patient in the right upper visual quadrant. This sensory phenomenon overlapped the patient's typical visual aura in distribution, color, form, and movement. An unexpected finding of hypoperfusion was shown near the stimulated depth contacts and in the mesial occipital region. Distant focal blood flow changes following RNS lead stimulation in a different patient were previously reported by our center1.
Conclusion: Clinical manifestations of direct cortical stimulation are shown to relate to transient hypoperfusion at distant white matter propagation pathways. This information ostensibly represents the extent of cortical modulation for a given set of focal stimulation parameters passed through a specific electrode contact shape, orientation and location in white matter. Presurgical planning using axonal pathways to direct the spread of RNS current to distant epileptic sources may amplify the efficacy of the available intracranial electrode set and simplify the surgical approach.
1. Rossi, M.A., et al. (2005). Noninvasive presurgical estimation of cortical activation for optimizing intracranial electrode placement for responsive neurostimulation in refractory epilepsy, AES 3.169
Supported by NeuroPace, Inc.
Return to Top
CHRONIC MEASUREMENT OF INCREASED EPILEPTIFORM ACTIVITY OVER MULTIPLE MENSTRUAL CYCLES IN TWO PATIENTS USING THE RESPONSIVE NEUROSTIMULATOR SYSTEM (RNS)
Marianna V. Spanaki, David Greene, Brien J. Smith, David Burdette, Martha Morrell, and Gregory L. Barkley
Rationale: A significant percentage of women with epilepsy (WWE) have reported increased seizure frequency during the perimenstrual period (catamenial epilepsy). We first documented catamenial exacerbation of epileptiform activity using the Responsive Neurostimulator System (RNS) in 2005 [M V Spanaki, et al. Epilepsia 2005;46(Suppl 8):222]. In this study we continuously measured epileptiform activity in two WWE over a total of 17 menstrual cycles.
Methods: As part of a FDA-approved multicenter feasibility study investigating the safety and potential efficacy of the RNS for epilepsy, we have implanted two WWE who were not good surgical candidates. The patients kept diaries of their menstrual periods. After seizure onset patterns were identified for both patients, detectors were programmed to identify electrocorticogram (ECOG) patterns that occurred prior to clinical seizures. These patterns appear epileptiform and occur frequently, but the vast majority of these bursts do not result in clinical seizures. Detection data from the RNS were then analyzed in multi-day windows around the date of menses onset and compared to detection rates on days exclusive of these perimenstrual windows to determine if the daily detection rate differed between the menstrual and non-menstrual intervals.
Results: Data was gathered for 9 and 8 menstrual cycles for patients H and M, respectively. The detection rates for the different windows are shown in the graph [figure1]. The detection data show statistical difference for both patients at window sizes of +/- 3 days relative to the onset of the patient's menstrual cycle at the alpha = 0.05 level (95% confidence).
Conclusion: The RNS has detection and diagnostic capabilities that allow objective measurement of epileptiform activity over prolonged periods of time in ambulatory patients. Use of RNS has confirmed the existence of perimenstrually exacerbated seizure activity.
Supported by NeuroPace, Inc.
Return to Top
OPTIMIZATION OF POST-OPERATIVE COMPUTERIZED TOMOGRAPHIC IMAGING IN PATIENTS WITH THE IMPLANTED RESPONSIVE NEUROSTIMULATOR SYSTEM
Allen Waziri, Hasit Mehta, and Robert R. Goodman
Rationale: The Responsive Neurostimulator System (RNS™ NeuroPace, Inc.) is currently under investigation as a neurosurgical option for medically refractory epilepsy. Following surgical positioning within a craniectomy, this pacemaker-like device allows for chronic cortical EEG monitoring (via two intracranial electrodes) and the delivery of focused electrical pulses with the goal of terminating seizure activity. Patients with implanted RNS™ devices may require post-operative imaging, either for the assessment of lead placement or, more importantly, for the purpose of emergent radiographic evaluation in cases of perioperative neurological decline. As MRI is currently contraindicated in implanted patients, CT remains the imaging modality of choice. Unfortunately, the metallic nature, large surface area, and proximity of the device to the underlying cortex produce overwhelming artifact on conventional axial CT images, resulting in limited diagnostic utility. We sought to develop a strategy that would overcome this technical difficulty and allow for the generation of diagnostically useful CT images in patients with implanted RNS™ devices.
Methods: Two individuals with previously implanted RNS™ devices, admitted for battery replacement, were included in the study. Following acquisition of conventional CT scans oriented in the standard axial plane, the patients were repositioned in the lateral decubitus position with their heads rotated and extended away from the side of the device. The scanner gantry was then appropriately angled to allow for the acquisition of images in the parasagittal plane parallel to the device axis. Contiguous fine-cut images were obtained and subsequently reconstructed using commercially available imaging software (Siemens). The resulting data were rendered in the true axial plane, for comparison to corresponding images from the conventional protocol, and were also used to generate three-dimensional reconstructions.
Results: Significant decreases were noted in the severity of device-related artifact on reconstructed axial images when compared to conventionally acquired images. The minimization of artifact allowed for diagnostic visualization of the subdural space and underlying cortex immediately adjacent to the device. In addition, artifact-free fine-cut imaging data were successfully used to generate detailed three-dimensional reconstructions of the cranial vault, providing for visualization of the stimulator and the intracranial position of attached electrodes.
Conclusion: This technique provides a straightforward method for the acquisition of diagnostic-quality CT images in patients with the implanted RNS™ device. We predict that similar strategies may be employed to optimize post-operative imaging in patients harboring a broad spectrum of intracranial hardware.
Supported by NeuroPace, Inc.
Return to Top
|
