A Pilot Trial of Scrambler Therapy for Pain Associated With Pancreas Cancer

Pancreas cancer causes pain in 70% of newly diagnosed patients rising to over 80% in advanced cancer patients, and is the most common serious symptom. Pancreas cancer pain is strongly correlated with worsened survival, and relief of pain is correlated with better survival. Pancreas cancer pain is neuropathic in origin and sensation, arising from direct invasion of local tissues. Pancreas cancer pain is relieved by local nerve blocks in about 75% of cases, but the relief is often incomplete, temporary, and the procedure is invasive with some risks. As the cancer grows to involve more nerves, or the nerves regrow, the pain typically comes back. For the 46,420 newly diagnosed pancreas cancer patients and the 39,500 patients who die of it each year, pain is the one constant symptom.

Scrambler Therapy is a novel treatment increasingly supported by clinical experience and multiple trials. The goal of the device is to provide "non-pain" information through the cutaneous nerves to block the effect of pain information. The device is a cutaneous electrical stimulator that creates waveforms simulating 16 different action potentials, then transmits the action potentials via C fibers. The goal is to reset the brain to perceive "non-pain" from areas previously interpreted as painful, similar to spinal cord stimulation. The multiprocessor apparatus directly stimulates the peripheral nerves by the application of surface electrodes on the skin, similar to EKG electrodes, putting the electrodes above and below the site of pain.

Preliminary data on Scrambler therapy in pancreas cancer Scrambler Therapy appears to work on pain based on results from multiple trials. In fact, the first trial was done on patients with abdominal pain from abdominal cancers. In this first trial, 11 cancer patients (3 pancreas, 4 colon, 4 gastric) suffering from drug resistant visceral pain were studied during the patients' first ten treatment sessions. The pain score before the treatment is shown next to the pain score after treatment, for each daily treatment. The electrodes were placed to surround the pain, following the dermatomes, as shown in Figure 2. The active (yellow) electrodes are always paired across the pain with the passive (black) electrode. Pain was quickly and markedly reduced and maintained until death. As then inventor wrote, "During the applications, all the patients reported a very rapid (in the order of a few seconds) disappearance of the perception of pain. All patients responded fully to the protocol and none reported undesirable side effects. Compliance was excellent." Nine of 11 stopped pain drugs within the first 5 applications. "During the reference period, nine out of eleven patients (81.8%) are seen to have stopped requesting painkillers between the second and the fifth treatment session. The remaining two patients (18.2%) considerably reduced the dosage and undertook mild therapy."

To be blunt, no one really believed these results. The magnitude of pain relief was large, with a simple machine using non-invasive therapy. To this day, the trial has not been replicated. But over the next decade, more and more evidence about Scrambler Therapy emerged from different investigators that led the investigators to believe that there was a signal worth exploring for relief in pancreas cancer. In the second trial, 226 patients with neuropathic pain including failed back surgery, brachial plexus neuropathy, and others were treated. Over 80% of patients responded with > 50% pain relief, 10% responded with pain relief from 25% to 49%, and only 10% had no response (P<0.0001). Based on these results the FDA approved Scrambler Therapy for cancer pain treatment February 25, 2009. The investigators have recently compiled the published reports of Scrambler Therapy for a review.

1. Study Procedures

Study Design and Treatment Plan The investigators propose a straightforward single arm trial of actual Scrambler Therapy to see if pain can be relieved. This study also serves to get preliminary information for planning future, larger, phase III studies that could compare to celiac plexus block, spinal cord stimulation, or sham treatment. The investigators anticipate completing the study in less than 12 months, with 18 completed subjects having up to 10 treatment sessions each.

Recruitment

1. Patients will be recruited through the Kimmel Cancer Center's outpatient oncology clinics, and the inpatient oncology and palliative care consult services. The palliative care service now sees over 1000 new consultations annually, with about half with cancer, and pancreas cancer pain is one of the most common reasons for consultation.

Treatment Days

Day 1/Treatment Initiation:

1. Treatment should be initiated by turning on the stimulus for the first electrode pair. The intensity of the stimulus is increased until the patient can first feel some sensation associated with one or both of the electrodes. Script: "Tell me when something is felt."
2. Then, over a few seconds, the intensity is increased to what is maximally tolerated. "Tell me when tolerance has been reached." In practice, this has been universally understood.
3. Once the intensity is at its maximum setting, the research therapist will evaluate the level of pain. If the pain level is not decreased, the machine will be reset to zero, the electrodes will be repositioned and the machine will be restarted in the manner described above. If the pain is not resolved with one set of electrodes, a second set will be applied in a similar fashion. Once satisfactory electrode placement and stimulus intensity is determined, therapy is maintained for a total of 30 minutes.
4. If a patient develops pain or a burning sensation with any of the electrodes, then the treatment should be interrupted and the electrode should be evaluated. Considerations include subjective patient intolerance, stimulation of a cutaneous nerve branch, or exacerbation of hyperesthesia or allodynia associated with the neuropathic process. These problems can be addressed by moving one or both of the electrodes farther away from the area of pain. For the electrodes that were moved, the intensity should again be increased and maintained at maximum for 30 minutes as described above.

Days 2-10:

1. Treatment will be administered using the same principles (each day evaluated independently and not necessarily reproducing the electrode arrangements and stimulation parameters of the previous day) for 30 minutes on consecutive days (Monday-Friday for 2 weeks).
2. Up to two or three days may be skipped to allow for weekends and/or holidays, if needed. If the participant presents without any pain, then the treatment will be "held" for that day and this information will be recorded.
3. Treatment does not proceed if the patient does not have pain.

Dose and Application

1. Electrodes are applied on the skin surrounding the pain-affected area.
2. The electrodes are never applied directly on the pain area, unless there is no pain free area. In that case, the electrode will be applied to the most pain

Statistical Considerations Overall This is a single-arm, pilot study to evaluate the effect of Scrambler Therapy for patients with pancreas cancer who have an average daily pain rating of ≥4 out of 10 based on the Modified Brief Pain Index, question #3.

Sample Size and Accrual The investigators are anticipating that the starting pain score on average will be ≥4 based on Virginia Commonwealth (VCU), Hopkins, Italian, and Mayo Clinic data. The investigators are conservatively anticipating that a relative reduction in average pain score by Day 28 will be 33%, a level which is considered minimal clinically important, based on the VCU, Italian, Mayo Clinic, and Walter Reed data. It is anticipated that approximately 90% of the patients will agree to be re-evaluated at Day 28 and have paired ratings available for evaluation. Thus, the study will enroll 20 patients expecting 18 to be evaluable for the primary endpoint. Based on prior studies, the investigators assume that the average pain value at baseline is at least 4 on a 0-10 scale with a standard deviation (SD) of the original pain value expected to fall in the range of 1-1.5 in this patient population.

As such, a conservative estimate of the standard deviation of the change across patients from Day 0 to Day 28 would be approximately up to 2. Using a one-sample paired t test with a sample size of 18 patients, this design will provide at least 80% power to detect a minimum of 33% reduction in average pain score at Day 28 compared to Day 0 with a two-sided type I error of 5%. The table below shows statistical power under different scenarios of a range of baseline pain scores and intra-patient correlation for a 33% relative reduction at Day 28. In the situation when variation of the starting pain score is larger than expected (e.g., SD=2.0), the sample size remains sufficient for this endpoint if the patients have moderate to severe pain to start with (i.e., 5 or above).

Analysis of Primary Endpoint

1. Baseline and change at Day 28 in average daily pain score will be treated as continuous variables and summarized with descriptive statistics. Each will be explored to determine if transformations (e.g. log or square-root) are necessary to achieve normality.
2. Exploratory plots will be created and means will be estimated along with 95% confidence intervals.
3. The investigators will use the paired t-test and Wilcoxon sign rank test as appropriate to determine whether or not the data shows evidence of change from baseline.
4. The primary analysis will include all patients who have received at least one session of treatment; the investigators' experience is that >90% will complete treatment.
5. A sensitivity analysis may be performed with patients who have completed 10 sessions of treatment, but the investigators anticipate that this highly motivated group will have >90% completion rates.

Analysis of Secondary Endpoints

1. All of the secondary endpoints will be summarized by item and by type of questionnaire, respectively, using descriptive statistics. A summary score of Brief Pain Inventory (BPI) will also be calculated by adding the total number of points across items.
2. The normality of distributions will be assessed and transformations be made as appropriate. Mean changes will be estimated with 95% confidence intervals. Changes from baseline will be evaluated using paired t tests or Wilcoxon signed rank tests.
3. A repeated measure of analysis of variance (ANOVA) will be used to test if there are any changes over time on pain scores.
4. Adjustment for multiple comparisons will not be made due to the pilot nature of study seeking to generate hypotheses to be tested in future larger randomized trials.