Introduction

Rather than write a ‘this is how I do it’ chapter, it seems more helpful to get under the skin of surgical techniques used specifically in neuromodulation, so to speak, and try to encapsulate the underlying principles of all surgical aspects in neuromodulation. This could serve as an underlayer, or foundation, upon which details of various techniques, some of which may be delineated here as well, can be built. Moreover, not only practitioners, but also designers and physiologists, can gain an appreciation for some of the related thoughts and concerns of the surgeon as they engage in neuromodulation, however applied. Other chapters will provide details of patient selection, troubleshooting and revisions, limiting morbidity, and more specifics on indications and applications per se. Further, I will lean heavily on non-percutaneous examples – other than some commentary about implantable pulse generator (IPG) placements and anchoring which apply in all cases. To this end, the chapter here is organized into what I consider to be the three most pertinent aspects of surgical technique in neuromodulation, as follows:

• •
  

  emphasizing the physiological target

  
  
• •
  

  being attuned to nuances (tissue, physiology, and the patient)

  
  
• •
  

  assuring that intraoperative resources are adequate for decision making.

Emphasizing the physiological target

Every procedure in neuromodulation involves placement of a device that interfaces with neural tissue to effect a particular physiological change toward benefiting the patient. The physiological aspect of this is the key, as it is an overarching principle helping to guide not only where to place the device, but also in how to place the device, when to place the device, and even in whom to place the device. The physiology available to help in this regard can manifest in various forms: patient feedback, somatosensory evoked potentials (SSEPs), evoked responses, single or multicell microelectrode recordings, electromyography (EMG), motor evoked potentials (MEPs), and electroencephalography (EEG). But it also may simply manifest in the knowledge that the physiology itself is the main concern. For example, one might know the general anatomy of the occipital nerve, the medial and lateral branches, as they course and branch in the suboccipital region, up and through the parieto-occipital area, but it is the knowledge of how the physiological interface of a transverse electrode placement will potentially capture and modify the nerve’s activity that helps more in placing one or two horizontally oriented 1 × 8 leads. And because of this, and because of the discomfort and difficulty in trying to use local anesthesia in the region, thereby altering the ability of the stimulation to obtain patient feedback if awake, one can typically place the leads under general anesthesia without patient feedback, or with local anesthetic (for a trial) and not worry about patient feedback.

This example is purposely chosen to illustrate this point because it is so removed from the typical emphasis placed on the physiological bases of using microelectrode recordings (MER) and stimulation in deep brain stimulation (DBS) for movement disorder targets, or cortical mapping techniques in motor cortex stimulation (MCS), where the reliance on physiology is obvious. In a different context, we have almost completely abandoned awakening patients to place dorsal column stimulators. Using a technique relying on stimulation through the lead and examining the evoked EMG is faster, more comfortable, and correlates specifically with the physiological target of the appropriate dermatomal levels in the patient – stimulating the same fiber tracts that are stimulated when the patient is awakened and asked about parasthesias. This technique avoids the sometimes confounding aspects of waking patients up from sedation, prone, sometimes confused, and occasionally misconstruing that we are asking them about where their pain typically is located rather than where they are feeling parasthesias, not to mention the risk of oversedation without a secured airway . This stimulation-EMG technique is entirely dependent on the underlying physiological target – it does not rely on the fluoroscopic imaging, or the patient’s compliance, or the ability to see the anatomy directly, the surgeon often falsely ‘assured’ the electrode is oriented as desired. None of these approaches alone is reliable, as the cord may be altered in its rotation within the canal, the fluoroscopy image notoriously can have parallax problems or, as mentioned, the patient may not be relating accurate enough details to place the lead because of medication, confusion, discomfort, or any combination thereof.

In emphasizing the physiological target, rather than imaging, or anatomy per se, the surgeon places him/herself within the center of the goal, rather than just outside of it, hoping they capture it. Some have recently done excellent work in exploring the potential of placing DBS leads, for example, using high resolution, high-strength magnetic resonance imaging (MRI) better to visualize the subthalamic nucleus (STN) . Although there is a strong correlation between the anatomical STN as a target and the benefit derived with DBS in Parkinson’s disease, it is still quite unclear whether or not the actual target of the stimulation that provides the benefit is the STN itself, fibers in the H2 field of Forel that course over the dorsal aspect of the STN, the zona incerta, or whether the benefit derives from the nature of the stimulation itself (frequency, pattern, amplitude, pulsewidth) in any of these locations, or in stimulating a combination of them. Because these underlying mechanisms are still not worked out definitively, it is premature to rely exclusively on anatomical targeting – although one can often obtain a very good result using this technique because of the high degree of overlap and leeway in both programming and field spread with present DBS systems. In a related example, even if one can see the globus pallidus pars interna (GPi) extremely well on a preoperative or intraoperative MRI scan, testing with a microelectrode within or near the optic tract is still necessary to determine distance of the final electrode tip from the tract to prevent current spread and visual disturbances postoperatively.

The same principle supports this in the placement of an electrode to perform motor cortex stimulation. We rely on a combination of locating the N20 reversal potential with SSEPs and motor-evoked EMG responses using a ball probe and a train of five pulses (see ), mapping the M1 region without opening the dura, and without looking at the cortex directly or its anatomic orientation. But some groups rely on functional MRI exclusively to place the lead. Again, without direct physiological confirmation, accurate placement of the lead to effect the desired result becomes compromised in a higher percentage of cases .

Being attuned to nuances (tissue, physiology, and the patient)

There are always multiple concerns during surgery that require vigilance on the part of the surgeon, who is continuously filtering his or her environment to determine relevance in the case at hand. As the saying ‘the devil is in the details’ goes, so goes this level of attention to nuances in neuromodulation surgery. Focus on concerns that are widely encountered, no matter what the neuromodulation application, would include details of the following: anchor positioning and suturing, locating IPG placement, how much tissue is dissected or removed for lead placement, the attention to detail of the neurophysiology staff and technicians (do they notice when anesthesia has changed? do they understand what effects it will have on the physiology?), company representatives supplying redundant and alternative devices and accessories, and the degree of reliance on the industry representative’s knowledge of how to assess systems and test stimulation. Some of these are beyond the scope of this text – but some are essential elements of surgical techniques, and will be covered.

Anchors are an extremely important part of preventing lead migration, catheter migration, and lead breakage – all of which potentially require revision surgeries. Interestingly, The Neuromodulation Foundation ( www.neuromodfound.org ), incorporated only in 2008, discusses a review of spinal cord stimulator lead migration as a complication, including issues of encapsulation, dural suturing and anchor issues within the discussion. The following were conclusions and recommendations from their review:

• •
  

  Incidence: surgical plate/paddle electrodes resist migration after encapsulation

  
  
• •
  

  Time to appearance of symptoms: immediate, i.e. before encapsulation

  
  
• •
  

  Treatment: non-invasively reassign contact combination if possible; if ineffective, revise electrode

  
  
• •
  

  Usual resolution and impact on therapy: minor displacement usually can be addressed non-invasively; major displacement requires revision

  
  
• •
  

  Risk reduction: some surgeons suture surgical plate/paddle electrodes directly to the dura, but this requires exposing a larger area, which is problematic, and might add mechanical stress. Some use an anchoring/strain relief sleeve to secure the emerging lead wire to the spine. Using absorbable sutures eliminates the continued focal stress that can be caused by non-absorbable sutures after the electrode becomes encapsulated. During system implantation, avoid increasing mechanical stress by avoiding unnecessary bends of small radius and superfluous connectors. Subject to patient preference and surgical judgment, avoid crossing a mobile joint or segment with subcutaneous lead wire or extension cable; e.g. a thoracic electrode encounters more stress and strain if connected to an upper buttock pulse generator than if connected to a lateral abdominal generator.

This review classifies evidence for these recommendations as level B – using their unique reworking of traditional levels of evidence grading, incorporating practical aspects of care like ‘only option’ and data supporting an advantage of risk-benefit analysis if it exists. Their level B includes well-designed clinical studies (prospective, non-randomized cohort studies, case-control studies, etc.), randomized control trials (RCTs) with design problems, and/or weighing risk versus potential benefit with expert consensus revealing a good likelihood of a favorable outcome. Level B review becomes a ‘recommendation’ from them.

I would add several comments on their assessment. On their first point (and this is addressed somewhat by their second point), it is important to remember that once a paddle lead has become encapsulated (within the first 3–6 weeks), it is very unlikely that it can migrate at all. This is not true for percutaneous leads, which may migrate cephalad to caudad even years after implantation. But further, it is important to realize this fact when discussing problems in pain coverage with patients or other care-givers after surgery. If coverage wanes, often the patient (and others) will ascribe it to a fall or other mishap (e.g. motor vehicle accident) and cannot be assuaged from this conclusion until an x-ray is obtained. This x-ray will almost perforce necessitate the ability to compare it to an original postoperative film, or at least a film of the lead when stimulation was working well. Therefore, it behooves one to have such a comparison film available on every patient. Additionally, though, it is unlikely that there will be any migration of the lead noted on such a film if the lead is a paddle and has been in place long enough – under such circumstances, there is more chance that scar has thickened, a wire has broken, or there is fluid in a connection, than a lead migration.

On their third point, while it is no doubt best to try to reprogram the lead initially, it is often the case that coverage still will not be obtained (though there is still stimulation felt), and the lead may still appear to be perfectly well-positioned. In such cases, it can be difficult in determining how to revise such a lead. One should consider moving it further cephalad or caudad, so that scarring thickness changes on the dura under the lead play less of a role in shunting current differentially, but advising the patient that it can recur and create the same problem again.

On the last point in their recommendation, I would suggest that suturing the lead to the dura is a viable option, but not recommended specifically for preventing migration per se, but rather to make sure the lead stays in direct proximity to the dural surface over its entire length. In some cases where the distal end of the paddle lead continues to divert right or left, perhaps because of a midline keel of sorts under the lamina, suturing the tip of the lead can be helpful in preventing this diversion after closing, until it encapsulates. Some paddle leads are easier to suture than others – Medtronic tripole leads have a nice intrinsic mesh within the silastic insulation and a slight margin within which a suture will hold well (4-0 Neurolon, for example). Finally, it is vital to make note in the chart of the fact that a lead was sutured to the dura, in case the same or another surgeon has to remove it, so it is not pulled out in a way that tears the dura.

Anchors themselves can obviate much of the need to worry about the strain on the lead directly from bending or where the lead wire crosses a flexion position of the body (e.g. neck, waist). Titanium cinching anchors work well and seem to prevent kinking and focal stress on the wire itself. However, traditional silastic anchors can work well to avoid such suture stress points if sutured correctly, and the newer titanium anchors can add a small additional cost. Recently, I had a patient with a medication pump who lost significant weight after the original surgery and could manipulate her pump in her abdomen over and over – a pump ‘twiddler’. Eventually, the catheter had either dislodged or was kinked; somehow the medication was not reaching the intrathecal space. Upon surgical exploration, it was appreciated that the anchor was still sutured to the fascia where I had left it and, surprisingly, the catheter did not move within the anchor when tugged – what happened instead was that the twisting of the catheter had caused the catheter material itself to fail and it sheared off just on the pump side of the anchor.

Anchors for vagal nerve stimulation (VNS) specifically can often lead to anterior neck pain if not sequestered below the platysmal tissue layer and are sutured so that they point into the neck rather than out toward the skin. This principle also applies to spinal cord stimulation (SCS) anchors, either thoracic or cervical, wherein pain or discomfort may occur if the anchors are secured too close to the subdermal layer, outside of the fascia. MCS ‘anchors’ typically can be reliably made by simply tacking the lead wire in a short loop (without kinking it) on the dura before the wire exits a burr hole from under the bone flap. The lead itself is also sutured down to the dura in several locations. DBS has available a device at the burr hole itself to secure the lead wire – these seem to work fine for the most part, although in thin scalped individuals, they can place increased pressure on the underside of the galea and create a higher propensity for scalp erosion.

IPG location is also important for patient comfort, convenience, and for mitigating problems with lead dislodgement or migration. Surprisingly, surgeons place IPGs in many locations of the body, often without confirming with the patient beforehand where the patient might want it, or need it, as patients with limited shoulder range of motion, for example, cannot even reach certain locations to use their controllers, or rechargers. Typically, there are only two appropriate locations to place IPGs thoracic leads, two locations to place IPGs cervical or occipital leads, and one general location to place IPGs DBS, VNS, or MCS leads. For thoracic or lower leads, primary preference should be to place the IPG in the upper buttock region, above the area of the ischial tuberosity where the patient will sit and below the region of the ‘belt line’, always avoiding two important features – the side a male patient may carry a wallet routinely, and the side on which the patient may prefer sleeping which may or may not be a pressure concern. In a patient with some extra subcutaneous tissue, the nearby posterior flank may be very acceptable for leads in this location as well, taking care not to make the position too lateral or imposing on the costal processes or scapula. For cervical leads, DBS, MCS, occipital nerve stimulation (ONS), VNS, and a potential variety of head peripheral nerve stimulation (PNS) leads, the subclavicular region (2 finger-breadths below the clavicle with the incision in line with the clavicle, and the pocket always kept superficial to the pectoralis fascia) is preferred, although the flank region as described just previously is also a possibility. I have avoided bringing these leads, with necessary extension wires, all the way to the buttock area. However, the subclavicular region often forces awkward lateral positioning in the operating room (OR) for lead placement, or the need for closing, undraping, repositioning and redraping laterally to place the IPG. For DBS and MCS this is not an issue, and may not be for certain locations of head PNS (e.g. supraorbital nerve stimulation), but for cervical SCS and ONS it should be thought out carefully.

As a final note on IPG implantation, certain IPGs currently require a particular tissue depth, often of 1–2 cm, no more, and no less, for optimal communications and recharging ability. Consideration should be made for contraction of the tissue during the healing and encapsulation process. Adipose tissue may resorb quickly, the capsule tighten, and within 6 weeks or so the IPG is only the thickness of the dermis away – causing pain and tethering and resulting in impossible recharging. The appropriate depth can be less than simple to achieve in many patients, especially if the patient is obese. Care must be taken as well in hemostasis of the pocket before closure, and in the configuration of the wires or catheter deep to the IPG or pump before closure – preventing kinking and migration of wires over the top of the IPG.

On the issue of tissue dissection and removal in the process of placing electrodes, typically in thoracic laminotomies for SCS, an incision can be made which is often as small or smaller than many made for the anchoring portion of percutaneous-to-permanent incisions – in other words, they can be just as minimally invasive as a percutaneous lead placement, aside from the muscle dissection itself. The removal of bone for the laminotomy, likewise, should be minimized for lead placement such that the lead just has enough room to be aligned appropriately and not have a propensity to migrate from pressure against an edge of bone or ligament. But this amount of bone removal can typically be achieved by only dissecting the muscle free from one laminar side (usually chosen to be the side that the patient may need more coverage on – but this should not be a major factor in deciding the side – for instance, surgeon preference can be just as important in this regard). The laminotomy will in no way lead to instability, even if one entire facet joint is removed, or even if a full laminectomy is needed to place the lead adequately. Because of the muscle dissection, and reclosure of the muscle fascia, these patients usually have more postoperative discomfort than a permanent conversion of a percutaneous lead. However, the entire process should take an hour or so of OR time and most patients should still be able to leave the same day.

Performing cervical laminotomies for paddle lead placements often encounters difficulty placing the lead unless small but adequate openings are made over each of several of the lamina in order to manage alignment of the paddle lead, help the lead progress in the epidural space, and even to pull the lead into position if necessary. In both the thoracic region and cervical region, it should be remembered that retrograde placement is always reasonable if coverage or simply placement of the lead at all requires it. Little has been published regarding cervical lead placement , but both percutaneous and paddle leads may be placed either anterograde or retrograde throughout the cervical spine from the occiput above C1 and below, within the limitations of prior surgeries and scar tissues. Leads do not need to be placed only at the most superior levels possible. In most cases, a paddle lead extending from C4–7 may be more appropriate for coverage of pain over C6–T1 dermatomes than a C1–2 lead.

Some have wondered whether MRI of the spine region to be accessed should always be obtained prior to surgical lead placement. Certainly, in any patient in whom myelopathy is suspected, adequate imaging and evaluation for safety of lead placement in the epidural space should be performed. However, in patients who are asymptomatic, it is very unlikely, even after prior fusion or decompressive surgeries, that there will be a concern. Additionally, one should consider the idea that a decompression may be performed in order to place the lead (in fact it may need to be performed due to prior surgery there) and this can eliminate the canal narrowness and the concern for creating excess stenosis in many cases. Anecdotally, in over 250 paddle lead new placements or revisions, including over 50 cervical leads, we have had no removals or injures related to cord compression, in a 10-year period. Some of these patients already coincidentally had prior MRIs of related regions of the spine, but new scans were obtained for this reason only rarely alone.

Assuring that intraoperative resources are adequate for decision making

This final category extends across several broad but converging areas involved in performing neuromodulation surgeries. Surgical decision making in neuromodulation ranges from the self-evident to the radically uncertain, but can be bolstered by planning. Once a patient has been identified for surgery, the assumption being that an appropriate indication has been found and work-up performed, circumstances can arise wherein appropriate therapy will be compromised unless adequate resources are available.

On the surface, one might consider that ‘resources’ in this case means ‘implantable devices’, but only sometimes is that also the case. Consider, for example, the situation of a placement of an MCS, and the craniotomy is performed after making measurements on the scalp, or integrating a functional MRI (fMRI) into a navigation system, or both, and the time has come for physiological mapping. The strip electrode, typically the one to be used for implantation as well, is moved around on the dura, looking for a reversal of the phase in the waveform consistent with the N20 SSEP. A small grid may be utilized for this as well, although the jackbox and setup will need to account for such a method. In any case, factor in that the pain for the patient extends up into the lower face area, where it is quite severe, but also into the hand and forearm, following a stroke 3 years prior. Initially, no N20 phase reversal is obtained, at least any that is consistent and unequivocal. The surgeon begins already to question the location of the craniotomy – is it too anterior, or posterior? Is it large enough to move the electrode to the central sulcus under the edge of the bone, or should it be extended – but which way? A decision is made to try the motor mapping first. The appropriateness of the anesthetic technique is checked and confirmed, after the surgeon asks the technician looking at the SSEPs to discuss this with the nurse anesthetist who has replaced the attending anesthesiologist in the room for this critical juncture of the case. The ball probe is moved around as amplitudes are adjusted and communication between the physiology team and the surgeon eventually confirms where the face region seems to be, inferiorly, and where the lowest thresholds for the hand region might be located, all on the surface of the dura. While apparently satisfactory signals are rechecked and confirmed, and no seizure has occurred, the strip electrode is not large enough to cover both areas, especially with any leeway in electrode configuration.

What are reasonable solutions? Adequate planning would not only account for placing two leads side by side or juxtaposed in tandem, but the discussion with the patient about two IPGs or a single dual channel IPG with two 2 × 4 leads, instead of the planned single 2 × 8 lead that has a similar footprint, also needs to have occurred before the surgery. Does the industry representative have these alternative devices with them? Is the patient prepped and draped for this possibility? Can this therapy use a rechargeable IPG, or only a non-rechargeable system, presently? These are all important aspects of the case that need to have been worked out ahead of time to render appropriate decision making in the OR.

A further example may suffice. While working with an orthopedic joint surgeon to place a peripheral nerve stimulator along the common peroneal nerve, above a postoperative neuromatous region that followed several prior orthopedic procedures, the plan is to bring the lead wire, or extension wire, up to the ipsilateral lower buttock region in the traditional location. However, once exposure is made, the appropriate region of the nerve dissected free and adequate securing of the lead and a strain relief loop with anchors is accomplished, it appears that there is now a problem – the realization that the patient has another device already implanted in an awkward lower buttock area. This was not appreciated in positioning and prepping the patient because the scar was well healed and a tattoo partially covered the area. A chart review confirmed that the patient had had a bladder stimulator (Interstim) placed in the past but it was unclear whether she still used it or whether it was still connected. Moreover, it was in the way of where the new IPG for the PNS needed to go.

While fluoroscopy was not needed in the original surgery, an x-ray was called for. But the patient was not positioned on the OR table in a manner conducive to obtaining an appropriate view on a cross-table lateral, and an anteroposterior (AP) was impossible because it was not an x-ray compatible table. The patient had to be moved by staff from under the drapes just enough to allow for the x-ray to be performed. The x-ray confirmed that the IPG was no longer connected to any lead. Alternative confirmation could have been made with a simple set of sterile needle electrodes or surface electrodes preoperatively to see if the device was turned on. Also, a decision could have been made preoperatively with the patient fully involved in terms of replacing this old IPG and moving the new IPG into a more appropriate location on the same side, or placing it on the other side if needed. Postoperatively, the patient admitted she had forgotten about the Interstim device. In any case, decisions need to be made, and proper planning ahead of time can eliminate major problems with device selection, location of incisions, appropriate removal or revisions, intraoperative testing and analysis of the physiology, and appropriate coverage for the patient, whether SCS, PNS, MCS, or DBS.

Conclusions

Several groups have addressed details and nuances of DBS, MCS, ONS, VNS, and SCS surgeries. This chapter emphasizes three principles in this regard: emphasizing the physiological target, being attuned to anatomical and patient subtleties, and assuring the availability of intraoperative resources necessary to decision making. The surgeon should have more than passing interest in neuromodulation itself. This interest facilitates the attention to detail that is required in these cases. Managing the integration of MER in DBS surgery is a study in itself of sustained focus and attention to detail, supervising multiple disciplines, understanding the probabilistic facets of the neurophysiology, the patient condition, the mechanics of the stereotactic equipment, the skill sets for using targeting software and appropriate positioning, the finer aspects of the burr hole, cannula, and electrode placements, and the role that medication and anesthesia may play in ultimately making the final decision that the electrode should go here or there in the end. DBS using MER, MCS with the use of SSEPs and motor mapping, and to some degree all other neuromodulation surgical techniques, requires this focus, these principles, and a deep belief in the therapy, along with its nuances and inadequacies, in order to perform the surgery.