Implantable Devices Project
I. Needs Assessment
Of all clinically treated pain, chronic pain represents up to 8%, and has an estimated cost of at least $150 billion in the United States alone each year [1]. Several nodes on pain pathways have been targeted for treatment of intractable pain, from peripheral nerves through dorsal roots, the spinal cord, midbrain, thalamus and cerebral cortex [1]. Experimental treatments on these nodes have included (but are not limited to) lesioning, perfusion with analgesics/anesthetics, and electrical stimulation [1]. However, lesioning carries risk of irreversible consequences if any error is made in surgical procedures, and it is widely understood that long-term use of analgesics like ibuprofen for treatment of chronic pain can lead to kidney failure, and increased risk of cancer and sclerosis—so this proposal will primarily examine methods and materials for electrical stimulation to treat chronic pain.
Since the 1950s, deep brain stimulation (DBS) has been studied as a method to relieve chronic pain of several kinds such as phantom limb pain, post stroke pain, facial pain, and brachial plexus avulsion [1]. DBS is a neurological treatment technique commonly used for movement disorders such as Parkinson’s disease, and has been reported to improve symptoms of epilepsy, Tourette’s syndrome, obsessive compulsive disorder, depression, cluster headache, and more [1]. Current DBS treatments for neuropathic pain targets several higher structures in pain pathways in and around the Thalamus [1]. This requires chronic electrode implants, for which maintenance can be difficult; especially if the implant is rejected and surgical removal is required. Regardless, DBS implants usually need to be replaced on a decade timescale. These complications may affect other important neural pathways, considering the crucial brain regions surrounding typical DBS targets for chronic pain relief. Figure 1: shows the placement of one such DBS solution for neuropathic pain, by Boccard et. al., for reference. For the specific placement in Figure 1:, surrounding areas of importance that might be impacted from complications could include the hippocampus, the corpus callosum, the Fornix, the thalamus and more. Complications in the hippocampus, for example, could result in consequences like those of the iconic “H.M.,” who suffered short-term memory loss as a result of its removal.
Figure 1: (Boccard, Pereira, & Aziz, 2015) [1] “Bilateral electrodes in the anterior cingulate cortex for deep brain stimulation (to treat chronic pain). A) Coronal T2-weighted MRI marked with the preoperative trajectory planning. B) Coronal view and C) sagittal view merged on a postoperative CT scan.” The surgical target is located 20mm posterior to the anterior tip of the frontal horns of the lateral ventricles; the contacts are positioned mostly in the cingulum bundle, with the deepest contact in the corpus callosum [1].
Assuming the above risks of DBS implants are of concern to patients, it is worth examining alternative solutions. Peripheral nerve stimulation (PNS) is noteworthy because it offers adjustable, reversible, and testable means for managing focused chronic pain as compared to DBS implants [2]. PNS is accomplished by placing a stimulating electrode near a named peripheral nerve (eg, occipital, genitofemoral) to elicit paresthesias up and down its innervated areas [2]. PNS shouldn’t be confused with peripheral nerve field stimulation (PNfS), which is a similar treatment targeted at regions of maximum pain rather than specific nerves [2]. PNfS may cause paresthesias along neurons that are not necessarily connected to the same major nerve branch or dermatome, which could lead to unexpected sensations in patients [2]. Therefore, PNS may be favorable to PNfS, but it should be noted that PNS will not treat all neuropathic pain more effectively than PNfS, and vice versa; some patients may respond optimally to a combination thereof.
In Petersen et. al.’s (2014) PNS/PNfS patients, permanent leads are sutured to the fascia, which is commonly practiced [2]. Like many other stimulation modalities, periodic adjustment is required for optimal pain relief, often with some degree of patient control [2]. Sometimes patients notice significant quality drops in pain relief even if there are no electrical faults in the system, in which case turning the stimulator off for a certain period sometimes restores efficacy of pain relief [2]. The most common complications include infection, lead erosion, migration of the lead, or mechanical/electrical failure in the device—which are correlated with the capability of current technology to accommodate the elasticity of surrounding tissue during patient movement [9]; neural injury is relatively rare [2]. The table in Figure 2: shows ranges of the frequency of the above complications across multiple studies.
Figure 2: (Petersen et. al.)[2]. Note the references in the figure above are Petersen et. al.’s (2014), and are not listed in this paper. For their references, see Petersen et. al.’s paper which is available online.
Thus, while PNS/PNfS may be most appropriate for some patients with clear targeted pain, the lack of efficacy among a significant number of trials suggests that there is a complexity to successful PNS/PNfS yet to be understood. Particularly, patients with larger areas of pain may not be practically solvable using a reasonable number of implanted electrodes, but spinal cord stimulation (SCS) and DBS can address larger areas of neuropathic pain. Like DBS, a number of clinical studies support the efficacy of SCS in treating failed back surgery syndrome (FBSS) and complex regional pain syndrome, as well as peripheral neuropathic pain, postamputation pain, postherpetic neuralgia, and root and spinal cord injury pain [3][4]. This makes the prospect of creating a compact stimulation device to address a large spectrum of chronic pain via SCS worth examining. Therefore, SCS may be optimal because, like DBS, it can address a large variety of neuropathic pain with a compact solution, and like PNS, SCS is low in the afferent pathway, so it can offer more targeted treatments with relatively low risk and easily adjustable/removable implants.
For cases of medial pain: Grider, Manchikanti, Carayannopoulos et. al. (2016) completed a study which found evidence from multiple relevant high quality randomized controlled trials (RCT’s) for the efficacy of SCS in lumbar FBSS [4]. This evidence was based on a best evidence synthesis from 3 RCT’s found across 175 abstracts and 63 full articles [4]. Kapural’s study reflects this evidence. Kapural also found that SCS is effective for Chronic Abdominal Pain such as mesenteric ischemic pain, esophageal dysmotility, gastroparesis, inflammatory bowel syndrome (IBS), chronic pancreatitis, and more. Kapural cited one study that found 30 out of 35 patients had >50% reduction in pain and decreased opioid use for patients suffering from any of the above chronic abdominal pains [3].
Kapural’s study gives insight to the efficacy of SCS for peripheral pain, as well, such as Complex Regional Pain Syndrome (CPRS), Painful Diabetic Peripheral Neuropathy (PDPN), and peripheral vascular disease and angina. SCS has been shown to significantly reduce pain and allodynia, improve limb function and quality of life in patients with CRPS-I [3]. Because PDPN is present in about 50% of long-term diabetic patients, Kapural’s discussion of five studies on PDPN is another valuable case for SCS. These studies showed >50% pain reduction for 11 out of 15 patients in one study and 63% of 25 patients in the other four studies.
All of the above evidence makes a good base case to pursue designs for SCS as opposed to DBS designs. An SCS device has the potential to manage a wide range of neuropathic pain, like DBS, but at a lower risk; and like PNS, SCS is low in the afferent pathway, so it can offer more targeted treatments with easily adjustable/removable implants. However, the use of SCS combined with PNfS has been shown to achieve broader coverage of axial back pain than either SCS or PNfS individually [2]. This technique is referred to as “hybrid” stimulation, spinal-peripheral neurostimulation, and triangular stimulation [2]. Also, there are a number of neuropathic pains that PNS/PNfS can treat, but SCS cannot, such as various facial nerve pains—so PNS/PNfS is a better choice than SCS in some scenarios to avoid the invasiveness of DBS, and it can be used together with SCS for better efficacy. Therefore, the end goal of this paper will be to propose improvements upon current limitations of SCS/PNS & PNfS hybrid solutions.
Current Solutions
Most SCS’s and PNS’s today consist of one or more tubular encapsulated leads with multiple electrode contacts along the length of the implant. This form factor has evolved from a simple bipole lead and electrical generator to a complicated electrode array lead (Figure 3:), and rechargeable, small generator allowing various current settings [3].
Figure 3: (Kapural, 2014)[3] “Proper, midline lead placement as observed during fluoroscopy. Lead is an octrode inserted percutaneously for the SCS trial using an epidural needle. Patient suffers fromlumbosacral radiculopathy and has a post-laminectomy syndrome.”
For SCS solutions it is important to take into consideration the varying size of the spinal canal and spinal cord, the position of the cord within the canal, and the varying amount of cerebro-spinal fluid (CSF) at each vertebral level [3]. The thickness of the CSF layer is typically thinnest at the C6 disc level and thickest at the T6 disc level [3]. These anatomic differences along the spinal cord will affect current to targeted tissues. One of the resulting frequent side-effects in today’s SCS solutions is that the change in the patient’s posture alters the amplitude of stimulation [3]. Body movements also affect the efficacy of PNS/PNfS solutions similarly, depending on implantation depth, placement relative to nerves, joints, muscles, and other actively moving tissues. My search of present literature did not reveal any present solutions that adjust stimulation amplitude relative to movement for more effective pain relief (search terms are listed in section III).
Like PNfS, efficacy of current SCS devices is hindered by paresthesias, or a tingling/tickling sensation along stimulated sensory pathways that apparently has no cause from the patient’s perspective most of the time, but can be triggered by a change in posture. This is certainly an acceptable side-effect if the implant significantly reduces patient pain, but paresthesias are still an unpleasant experience and several options are being explored to reduce them in future SCS’s. Targeted dorsal-root ganglion (DRG) stimulation is one solution currently being explored in SCS; it has shown considerable effectiveness in relieving chronic neuropathic pain and has several advantages over conventional SCS including lack of paresthesia variation with position and discrete and consistent paresthesia coverage in anatomies that are difficult to cover with dorsal column stimulation [5]. Other studies are utilizing burst stimulations of various kinds. There have been mixed results with burst stimulation, but a pilot study found that 40 Hz bursts with 5 spikes at 500Hz per burst produced long-term relief, and 83% of patients tested did not experience any paresthesia [3], [7]. There has also been research into high frequency SCS at 10kHz using subthreshold stimulation which reported 70% of patients as paresthesia-free, having pain relief, and experiencing no shocks during position change [3], [4],[8]. However, high frequency stimulation has been shown to have mixed results based on the application, and on average seems to be less effective than traditional stimulation within the timescale of neural activity [4].
Therefore, I propose a hybrid solution may improve upon present devices by using more flexible stimulators that better conform to tissue movements with incorporated flex or pressure sensors to actively adjust stimulation amplitude as opposed to selecting single burst frequencies to mitigate paresthesias and improve pain relief.
Target Patients and Optimizations for Human Use
To be included in the study, patients should have chronic, severe, disabling neuropathic pain refractory to other treatments such as medications, nerve blocks, trigger point injections, physical therapy, et cetera. Good target patients could be suffering from the following disorders and other similar neuropathic pain (From [2], [4]):
· Neuropathic pain disorders
o Posttraumatic neuralgia, Postsurgical neuropathic pain, Occipital neuralgia or cervicogenic occipital pain, Postherniorrhaphy inguinal neuralgia, Genitofemoral neuralgia, Postherpetic neuralgia, and Coccygodynia
· Complex regional pain syndrome
· Phantom limb pain
· Axial pain syndromes
o Failed back surgery syndrome (FBSS), nerve damage
· Chronic Abdominal Pain
o Mesenteric ischemic pain, esophageal dysmotility, gastroparesis, inflammatory bowel syndrome (IBS), chronic pancreatitis, familial Mediterranean fever, post-traumatic splenectomy, generalized chronic abdominal pain, postgastric bypass chronic epigastric pain, and chronic visceral pain after various intra-abdominal surgeries with evidence of abdominal adhesions [3]
· Emerging indications
o Musculoskeletal pain Fibromyalgia
Disorder-specific criteria may apply on a per-patient basis, if it is relevant to treatment. Depending on the disorder, there may be initial trials with local anesthetic block or transcutaneous electrical nerve stimulation (TENS) to determine affected nerves. If required, further trials could be continued with varying PNfS/PNS placements to determine longer-term efficacy of stimulation on target nerves.
More recent onset of symptoms may be given preference but would not be required for immediate inclusion in the study. Rizvi & Kumar demonstrated the importance of early referral to the efficacy of SCS treatment for chronic pain [6]. Their results showed success rates of over 80% if implantation happened within two years of symptom onset, while patients who received implants after 20 years had only 15% success [6]. For Complex Regional Pain Syndrome (CRPS), the International Association for Study Pain (IASP) recommends that SCS be introduced within 12–16 weeks if conventional therapy fails [3]. These factors may be important when admitting patients.
Exclusion criteria (modified from [2]):
· Patients with active infections or who are severely immunocompromised
· Patients taking anticoagulant medications that cannot be stopped for the perioperative period, or those with bleeding disorders
· Cognitive impairment that limits the patient’s ability to use the device effectively or to cooperate with lead placement
· Patients with untreated psychiatric disease that is deemed dangerous to the study and the patient’s well-being if they were to participate
· Malingering
Neuropsychological evaluation:
A pre-trial psychological screening is recommended by Petersen et. al. to identify any psychosocial issues that may adversely impact the therapy [2]. Symptoms screened for could include cognitive impairment or dementia, self-abuse or substance abuse, untreated anxiety or depression, or unrealistic expectations related to the stimulator.
Trial Stimulation:
Depending on the disorder being treated, there may be trials involving placement of PNS/PNfS electrode leads along target nerves leading up to the dorsal root. The leads would exit through the skin, and the patient would test stimulation parameters for a time period ranging from 2 to 14 days, as in Petersen’s study [2]. If PNS/PNfS electrodes are deemed unnecessary or not useful, then surgical implantation of the SCS would ensue, after which there may be a patient and physician tuning period to minimize pain. PNS/PNfS hybrid solutions may still be tested in tandem with SCS implants. Patients may have the ability to self-tune their devices; however, pressure sensitive electronics will aim to eliminate that need based on patient adjustment and body movement. Automatic stimulation adjustment would require closed loop and open loop training.
Needs Statement
A more biocompatible SCS/PNS hybrid solution for people with treatable neuropathic pain that will reduce paresthesias and relieve at least 50% of discomfort through automatically adjusted stimulation amplitude.
II. Engineering Design
Implants would ideally be fully addressable and adjustable, with stretchable leads, soft electrodes, and movement sensors embedded in a soft substrate. This idea is inspired by the “electronic dura mater implant,” (e-dura) by Minev, Musienko, Hirsch et. al. (2015), which is mechanically similar to the dura mater around the brain and spinal cord for better biocompatibility [9]. The control hardware and power supply should be implanted outside the vertebrae in a hermetically sealed package with all necessary safety considerations. Leads would connect the control hardware to the spinal stimulators and sensors between vertebrae, and to peripheral stimulators outside the spinal column. The following sections will go into detail.
Material Choices and Design for Biocompatibility
This proposal will utilize Minev et. al.’s (2015) foundation for soft SCS implant design to develop a hybrid solution. Minev et. al. (2015) explain that neurual tissues have elastic and shear moduli in the 100- to 1500-kPa (kilopascal) range, making the tissues quite soft in comparison to most current electrode implants, which present high elastic moduli in the gigapascal range [9]. Consequently, the surgical insertion of such inelastic implants triggers both acute and long-term tissue responses [9]. To remedy this, Minev et al. designed their soft “e-dura” implant (Figure 4:), which integrates a transparent silicone substrate (120µm thickness), stretchable gold interconnects (35µm thickness), soft electrodes coated with a platinum-silicone composite (300 mm in diameter), and a compliant fluidic microchannel (100 mm by 50 mm in cross section) [9].
Minev et al.’s design choices are well formulated and useful considerations in designing a Hybrid SCS for chronic pain. Especially considering Minev et al.’s implant vastly outperformed traditional hard implants in degredation over time in locomotion trials with rats, which can support efficacy for other stimulation applications. To quantify this comparison, Minev et al. designed a “stiff,” 25-um thick polymide implant to compare to their e-dura implant [9]. Their assessment of motor performance was based on high-resolution kinematic recordings of body movement fascilitated by the implants [9]. In the sham stages, there was essentially no difference in the performance of the implants, however after 1 to 2 weeks, Minev et al. noticed significant motor deficits in the rats with stiff implants. After 6 weeks, it was found that there was significant inflammatory responses at chronic stages of implantation, causing deformation of spinal segments under stiff implants [9]. In marked contrast, there was no significant difference in inflammatory responses in rats that were implanted with e-dura [9]. These results support the efficacy of their material choices for improving biocompatibility of hybrid implants. Therefore, my design will also utilize such maleable materials in both SCS and PNS/PNfS solutions.
To improve amplitude output for pain treatment and to reduce parasthesias, components to detect body movements will need to be incorporated into the silicone substrate. One obvious solution is to utilize flex sensors. However, while most flex sensors have long form factors that can bend, they have little to no elasticity, which—as described previously—would likely result in significant inflamatory response in spinal implants considering their typical footprint is not smaller than a centimeter scale [9]. Another solution is well-placed pressure sensors in the soft substrate with a small form factor (typical sensors are as small as millimeter scales—see Figure 5). Though current pressure sensor solutions are composed of hard materials, their small form factor and encasement in a soft substrate should minimalize inflamatory response. Their placement should be relatively close to electrode stimulation sites that are expected to experience significant tissue deformation that would alter tissue resistivity between electrodes and stimulation targets. If sensors are placed far enough from stimulation sites, there may be inconsistencies in tissue deformation due to bones, ligaments, or other bodily structures that would hinder accurate amplitude adjustment. Figure 5 is an example of one pressure sensor that fits the requirements stated previously.
Figure 5: MS5611-01BA03 barometric pressure sensor, by TE Connectivity Sensor Solutions. The model is a low power, high resolution design with I2C and SPI output with a 5mm×3mm×1mm package.
Pressure sensor selection will be dependent on the range of pressures actually exerted in the spinal column or peripheral stimulation sites during typical movement. Trials will need to be run to determine typical pressures and to evaluate the change in inflammatory response vs control implants without hard pressure sensors embedded. The testing paradigm can be similar to Minev et. al.’s (2015) 6-week rat trials, which would establish how my proposed soft implant compares to their biocompatibility results to form a stronger case for the efficacy of such implants. Vertical form factor of pressure sensors like those in Figure 5 can be drastically reduced from 1mm to PCB thickness (~500µm) by removing the hard cover placed over electronics for short-circuit protection, considering the sensors will be encapsulated in an insulative substrate that will also protect against shorts. Should an alternative pressure sensor to that in Figure 5 be required, there are several options with similar form-factors and higher operating ranges (up to thousands of kPa) offered by online retailers like Digikey.
Including such pressure sensors evolve current SCS technology into a variable solution that aims minimalize manual tuning. Additionally, integrating hybrid stimulation through PNS/PNfS electrodes will allow for more effective targeting of pain when SCS is not enough. In some cases, trials may be run to determine if PNS/PNfS alone is enough to successfully treat pain. Regardless of the pain treatment modality (SCS/PNS/PNfS or some combination thereof), the power supply and control system will be the same, packaged outside the spinal column with interconnects extending to stimulation sites. Figure 6 is an early design sketch that was developed with all considerations mentioned in this section.
Figure 6: Rough design sketch for hybrid stimulator concept using soft implant technology. Components are not to scale. Number and placement of electrodes and pressure sensors may vary. Note pressure sensors are encapsulated in soft silicone substrate, and would bulge on the side opposite soft electrode exposure as not to alter contact. Interconnect exit points from the power and data package would be hermetically sealed.
Operational State-machine
A basic state diagram is drawn in Figure 7 to convey operational functionality.
Figure 7: Operational state machine for device solution. This state machine would handle electrode stimulus output for both SCS and PNS leads. Users would be able to selectively disable PNS or SCS outputs individually or altogether. Diagnostic mode can be entered from any other node in the state machine if there is a user interrupt for input. The power off state can also be reached from any node in the state machine if the safety circuit trips or some critical error occurs—the system should enter low power mode from any state when battery charge is low, which prevents the need for an automatic power off on critically low battery from any node.
Electrodes would be addressable and independently capable of frequency adjustment. Each electrode should have its own coefficient to weight its stimulus frequency based on readings from the nearest pressure sensor. Stimulation amplitude should have an inverse relationship with pressure caused by varying posture; so more pressure should result in decreased stimulation amplitude because of electrode proximity to stimulation targets would decrease in that scenario. Once a coefficient is assigned to an electrode, it is only cleared when the user resets the device, when a pressure change occurs and it is overwritten, or when the user manually changes stimulation amplitude. Manual user frequency adjustments will result in all coefficents being adjusted by a certain percentage, which may not be linear across all electrodes at all pressures. Experiments will need to be conducted to determine proper adjustment parameters.
The default mode will likely be adjusted per patient, but ranges utilized for programming settings were as follows in a phantom limb pain study conducted by Eldabe, Burger, Moser et al. [5]: frequency: 20–40 Hz; pulse width: 200–420 μs; amplitude: 150–1800 μA. It is likely that implemented stimulation frequencies will be in physiological ranges, like above. However, it would be worth examining the efficacy of high-frequency and burst frequency stimuli for patients on both individual and wide-use bases.
Power Requirements
The power supply would likely be a long-lasting battery, hermetically sealed in a package outside of the spinal column. For biocompatibility, wireless percutaneous charging is preferable to leaving a charging port through the skin or regular battery replacements. Charging can be implemented via magnetic lock or adhesive for MRI safety. A voltage regulator circuit and an optoisolator will be necessary during charging to prevent current leak into stimulators. The implant will draw more power because it is active and routinely executing control logic. However, the low power mode state can be utilized automatically in several conditions or manually set to extend battery life. Low power mode will also be automatically disabled on pressure sensor interrupts unless the user specifies a power off state (perhaps when the user experiences little or manageable pain). Battery size should be maximized within the available space, so a high-density cell, such as LiPO technology, would be highly considered for use. Therefore, safety circuitry will be extremely important to detect over-current, over-voltage, and reverse polarity.
Data Requirements
In order for patients and physicians to adjust stimulation parameters, wireless data telemetry will be necessary. A simple telemetry solution is to transmit data via the charging coil. To drive this telemetry, microcontrollers running embedded C code can be utilized on the sending and receiving sides of the coupled coils. The microcontroller on the sending side should drive a high-power bipolar transistor to gate power delivery from a battery, because most microcontrollers cannot source the current required to transmit through tissue due to sensitive transistors used to switch their pins. There are multiple low-power microcontroller solutions from TI, PIC, and various other companies. Depending on the number of pins on the microcontroller, an external mux may be useful for addressing electrodes. The microcontroller on the external side of the coupled coils would ideally be packaged in an aesthetic encasement, driving a display with the user interface. 8-bit transmissions can be used to encode all the commands necessary for stimulation adjustment. Executions can be stored in a look-up table on the internal microcontroller and selected based on 8-bit codes received via telemetry.
Fabrication and Hermeticity
In my search to make my proposed implant flexible enough, I observed several solutions that pattern extremely thin films into web-like systems, which causes fractal-like meshes to form that can reversibly hinge during mechanical loading [9], (Figure 4). Medical devices prepared with this patterning can conform to curvilinear tissue surfaces [9]. However, this type of interface requires complex, multistep processing and packaging [9].
On the other hand, the fabrication process of Minev et. al.’s e-dura is comparably simple, and better mimics the mechanical properties of the dura that encapsulated spinal tissue. The relevant components for my hybrid stimulation solution are outlined below. I believe it is best that my fabrication process be closely based on Minev et al.’s materials and methods outlined in their supplementary information. Quantities and measurements for Minev et. al.’s solution can be found in their supplementary information (and in Figure S2), and do not necessarily correlate with my proposal because my solution needs to make room for pressure sensors and adds an additional PNS/PNfS fabrication process, which will have a different footprint and may have different dimension requirements.
Interconnects (fig. S2-1)
i) A substrate of polydimethylsiloxane (PDMS) should be spin-coated on a silicon carrier wafer pre-coated with polystyrene sulfonic acid (water soluble release layer) and left to cure.
ii) A shadow mask patterned with the negative of the electrode layout and pressure sensor leads then needs to be laminated on the PDMS substrate. Thermal evaporation of chromium/gold (Cr/Au) metal films through the shadow mask can be used to deposit the leads.
Electrical passivation layer (fig. S2-2)
i) The interconnect layer can be prepared in parallel. A thick slab of PDMS will be functionalized with a 1H, 1H, 2H, 2H perfluorooctyltriethoxysilane (Sigma-Aldrich) release monolayer under weak vacuum. Two thin PDMS layers should next be spin-coated on the thick PDMS slab, individually cured then treated with the debonding monolayer. In cross-section, the final structure should be a triple stack consisting of a thick PDMS slab and two thin PDMS layers. The release coatings should allow for each of the thin PDMS layers to be peeled off independently at a later stage.
ii) Using a hollow glass capillary of a pre-defined tip diameter, both of the thin PDMS layers should be simultaneously punctured at locations corresponding to the sites of the electrodes and pressure sensors.
Encapsulation (fig. S2-3)
i) Both PDMS triple stack and interconnect wafers will be exposed to brief air to plasma activate the silicone surfaces. The triple stack will then be flipped upside down to align the punctured holes with the underlying electrodes and pressure sensors.
ii) The two pieces can then be brought together to form a covalent bond. Next, the thick PDMS slab can be peeled off, leaving behind the two thin PDMS layers on top of the interconnects and pressure sensor leads.
Soft platinum-silicone composite preparation and patterning (fig. S2-4)
i) The electrode sites can then be coated with a conductive blend of platinum nano-micro particles and PDMS to create soft electrodes. Minev et. al. (2015) details their process thoroughly in their supplementary material, which will be followed in my proposed device. Both for the SCS and the PNS electrode sites.
ii) To form the active electrode coating, a bolus of conductive composite paste can be spread and pressed into the holes of the upper encapsulation layer.
iii) The upper encapsulation layer should then be peeled off, leaving bumps of conductive composite at the active electrode sites. The bottom thin PDMS layer is now permanently bonded to the electrode-interconnect substrate, which will serve as insulation. The array should then be allowed to sit in a hot environment to ensure full polymerization of the conductive paste. Minev et. al. (2015) did so at 60°C overnight.
Release (fig. S2-6)
i) The contour of the finished implant was cut out from the wafer using a razor blade. The implant was released from the carrier wafer upon a brief immersion in water.
For this chronic pain hybrid solution, the bonding of the microfluidic channel to the e-dura would be forgone to create space to add electrodes to the array and pressure sensors, and the bond of fast-cure silicone could be used to secure any other necessary components in that layer. There will need to be room for I2C/SPI communications from the pressure sensors in addition to the electrical connector that Minev et al. included. Pressure sensors should be oriented so any bulge would extend on the opposite side of the implant from the soft electrode contacts to ensure reliable stimulation. The fabrication process above is nearly the same for both the SCS and PNS components of my proposed hybrid solution. The only difference will be a different contour footprint for the PNS and SCS extensions. The PNS stimulators will need to branch out (as shown in Figure 6) to reach target nerves. Stimulators will likely be sutured to targets—more on that in section IV.
Figure S2: Soft
neurotechnology for e-dura implants. “(1) Elastomeric substrate and stretchable
interconnects fabrication. Patterning
(2) and bonding (3) of interconnects’ passivation layer. (4) Coating of the electrodes with a
customized platinumsilicone composite screen-printed above the electrode
sites. (5) Integration of the PDMS
microfluidic channel and connector. (6) Release of the e-dura implant in water”
Major Engineering Design and Manufacturing Challenges
The arrangement of the terminal pads of the gold film interconnects and pressure sensors within the PDMS substrate must be carefully planned to match the form factor of the microcontroller for addressing electrodes and pressure sensors. It will be increasingly difficult to accomplish these steps for every soft electrode contact added in denser arrays without damaging the insulation between contacts or creating a short.
The Microcontroller board, coil, safety circuit, and battery will all need to be hermetically sealed as well at the far end of the implant, and will certainly not fit between the spinal cord and vertebrae. Thus, like the electrical connector used by Minev at al., the aforementioned electronics would need to be secured outside the vertebrae. This produces the complication of having a chronic insertion for the substrate between two vertebrae. There didn’t seem to be issues for Minev et al. for an entry point between L3 and L4 (see Figure 8), but it remains to be seen if the same would be true between other vertebrae in human patients. Because of the elasticity of the PDMS substrate solution, it is hopeful that a small flexible lead encased in e-Dura would not cause significant complication.
Furthermore, with relatively bulky components on the end of the e-dura implant, it would need to be secured very well to avoid shifting that could unnecessarily stretch or contort components. Minev et al. used two orthotic screws, but more may be required for this application. Hopefully tissue encapsulation outside the spinal cord will secure the bulk of the battery and electronics in the orthosis, but further data should be collected to determine whether more careful consideration needs to be taken for securing the part of the device outside of the spinal column. The placement of the battery and electronics should also be carefully planned to minimalize risk of infection, inflammation, and pain distinct from patients’ preexisting, chronic conditions.
Figure 8: Minev et al. (2015)[9] “Re- constructed 3D micro–computed tomography scans of the e-dura inserted in the spinal subdural space covering L2 to S1 spinal segments in rats. The scan was obtained in vivo at week 5 after implantation.”
III. Intellectual Property Evaluation and Discussion
Upon searching patentscout.innography.com, there do not appear to be any existing stimulators that use pressure sensors to actively control stimulation frequency or current for optimal relief of neuropathic pain. Search phrases used include:
· “pressure sensitive stimulator/stimulation/stimulate/implant/device/SCS/PNS”
· “pressure stimulator/stimulation/stimulate/implant/device/SCS/PNS”
· “pressure sensor stimulator/stimulation/stimulate/implant/device/SCS/PNS,”
· “automatic current adjusting stimulator/stimulation/stimulate/implant/device/SCS/PNS”
· “automatic stimulator/stimulation/stimulate/implant/device/SCS/PNS”
· “current adjusting stimulator/stimulation/stimulate/implant/device/SCS/PNS”
· “frequency adjusting stimulator/stimulation/stimulate/implant/device/SCS/PNS”
· “automatic frequency stimulator/stimulation/stimulate/implant/device/SCS/PNS”
So it appears an automatic pressure-sensitive stimulator is patentable, and has a real application potential amongst more patient groups than those who require spinal cord or peripheral nerve stimulation.
Needless to say, e-dura is the intellectual property of Minev et al., and I cannot patent any soft substrate that is too similar here. Upon searching for rechargeable implantable devices, there were a number of already patented ideas, so I also cannot patent that design choice. I am aware of no other potentially patentable ideas in my design, so all credit and rights of invention for various aspects of this project will be observed for rightful originators.
IV. Surgical Approach
Anatomical Considerations
For the e-dura PDMS solution, the shape and unusual resilience of the soft implant should greatly facilitate surgical procedures. Minev et al.’s Supplemental Materials outlines their surgical procedures in detail, which would ideally be emulated for my proposed device. I have included their implantation procedure and considerations below, with considerations for my pressure sensors and PNS components to follow:
“Implantation of e-dura into the spinal subdural space (fig. S5A-B):
The e-dura substrate was implanted under Isoflurane/Dorbene anesthesia (in rats). Under sterile conditions, a dorsal midline skin incision was made and the muscles covering the dorsal vertebral column were removed. A partial laminectomy was performed at vertebrae levels L3-L4 and T12-T13 to create entry and exit points for the implant. To access the intrathecal space, a 3mm long mediolateral incision was performed in the dura mater at both laminectomy sites. A loop of surgical suture (Ethilon 4.0) was inserted through the rostral (T12-T13) dura mater incision and pushed horizontally along the subdural space until the loop emerged through the caudal (L3-L4) dura mater incision. The extremity of the implant was then folded around the suture loop. The loop was then retracted gently to position the implant over the spinal cord. A small portion of the implant protruded beyond the rostral dura mater incision and could be manipulated with fine forceps to adjust the mediolateral and rostrocaudal positioning of the implant. Electrophysiological testing was performed intra-operatively to fine-tune positioning of electrodes with respect to lumbar and sacral segments (30). The protruded extremity of the implant became encapsulated within connective tissues, which secured positioning of the implant in the chronic stages.
The soft-to-wires (and microfluidic) connector was secured to the bone using a newly developed vertebral orthosis. The connector was first positioned above the vertebral bone. Four microscrews (Precision Stainless Steel 303 Machine Screw, Binding Head, Slotted Drive, ANSI B18.6.3, #000-120, 0.125) were inserted into the bone of rostral and caudal vertebrae. Surgical suture (Ethilon 4.0) was used to form a cage around the micro-screws and connector. The walls of the cage were plastered using freshly mixed dental cement (ProBase Cold, Ivoclar Vivadent) extruded through a syringe. After solidification of the dental cement, the electrical wires and microfluidic tube were routed sub-cutaneously to the head of the rat, where the Omnetics electrical connector and the microfluidic access port were secured to the skull using dental cement. The same method was used to create the vertebral orthosis for stiff and sham implants in the biocompatibility study.”
Figure S5: “Surgical procedure to slide the e-dura below the dura mater covering lumbar segments. (B) Side view of the engineered vertebral orthosis that secures the e-dura connector, and ensures long-term functionality of embedded electrodes and chemotrode in vivo.”
Additional surgical considerations will relate to the placement of the electrode array for optimal efficacy given the kind of neuropathic pain being addressed. However, the electrode array will be addressable, so slight inaccuracies in placement of the contacts on the spinal cord will be fixable. Furthermore, the PNS solution will need to be sutured to fascia near stimulation targets as well (see Figure 9). Suture placement should ensure a similar degree of freedom for electrodes and paired pressure sensors for best stimulation results. Targets may be in superficial layers or deeper muscle layers, which will be accessible during insertion of the SCS substrate in the steps outlined above. Targets may also include dorsal roots, in which case deeper incisions would need to be made on either side of the vertebrae of interest. The orientation of the device and extending substrate will vary depending on individual patient need. PNS target may be significantly distant from SCS targets in some patients. Not every patient will require both SCS and PNS/PNfS implantation, so the entire surgical procedure outlined in this section may not be followed for every patient.
Addressing Complications
For the final SCS implant, there are several reasons the device may need to be removed. However, hopefully that can be avoided with proper packaging. If the packaging were to fail, then there could be risk of infection, inflammation or (less likely) hematoma that could cause risks for the spinal cord. Fortunately, the soft nature of the e-dura material makes surgical removal fairly simple as compared to hard IMD’s. After removing the screws and orthosis holding the microcontroller and power supply and addressing any encapsulation, the rest of the array should slide freely out from the spinal column. The PNS leads will need to be surgically cut free with a scalpel because they will have been sutured to target stimulation sites. Recovery time should be short and patients should not expect any major inconveniences during the healing process. However, if patients want to replace their failing implant, they may not be able to do so right away depending on the complications experienced. Therefore, patients may have to finish the recovery process before having another implant surgically placed. Ideally, patients would be able to replace the failing IMD with a working model immediately after removal or shortly thereafter if swelling is a challenge.
References
[10] Artan, N. S., Patel, R. C., Ning, C., & Chao, H. J. (2012, August). High-efficiency wireless power delivery for medical implants using hybrid coils. InEngineering in Medicine and Biology Society (EMBC), 2012 Annual International Conference of the IEEE (pp. 1683-1686). IEEE.
