Treatment Applications of Cortical Stimulation

Published on 14/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 2779 times

CHAPTER 93 Treatment Applications of Cortical Stimulation

Since initial reports in the early 1990s, motor cortical stimulation (MCS) of the primary motor cortex (M1) has been used to treat chronic refractory pain conditions and a variety of movement disorders. In this chapter we report several of our own cases in which MCS was used to treat poststroke and non-poststroke pain syndromes and movement disorders, Parkinson’s disease (PD), essential tremor, and corticobasal degeneration. In addition, we cover the essential history of MCS and details on how the basic aspects of the procedure can be achieved.

Although some attempts to use cortical stimulation for the treatment of neurological disorders were made in the latter half of the 20th century, only recently has chronic MCS been investigated seriously as a surgical treatment of both chronic pain and medically refractory movement disorders. Before the advent of viable chronically implanted stimulation systems, disorders of central nervous system function were treated by ablation or extirpation of the offending tissue. In 1909, Horsley reported suppression of athetosis by resection of the motor cortex.¹ Interestingly, he localized the upper limb region of M1 by direct cortical stimulation. In the 1930s and 1940s, Bucy extirpated the motor cortex to treat PD and other tremor disorders.¹^,² In 1954, Penfield and Jasper resected a patient’s primary sensory cortex (S1) to treat burning pain but did not achieve complete resolution of symptoms until they also resected the precentral gyrus.³ Later that same decade, White and Sweet reported minimal success treating refractory pain with postcentral resection.⁴

The contemporary history of nervous system stimulation as a treatment modality perhaps begins with Heath, who in 1954 stimulated the septal area with the goal of activating pleasure centers and alleviating pain.⁵ Twenty-five years later, Woolsey and colleagues demonstrated inhibition of tremor and rigidity in PD patients via precentral stimulation,⁶ although permanent implantation of electrodes was not attempted. It was not until the 1970s, when Mundinger stimulated the sensory thalamic nuclei to treat spasticity and athetosis, that chronic stimulation of the brain was used to treat movement disorders.⁷^,⁸ Since then, deep brain stimulation (DBS) at several sites, including the thalamus,⁹ cingulum,¹⁰ periventricular and periaqueductal gray areas,¹¹ and more recently the globus pallidus pars interna¹² and subthalamic nucleus,¹³ has been attempted to treat a variety of functional brain disorders. During the mid-1980s, Hosobuchi implanted electrodes subcortically in the somatosensory region in 76 chronic deafferentation pain patients and noted beneficial results in 44.¹⁴ In 1985 and 1991, Tsubokawa and colleagues reported their results with chronic stimulation of the M1 region for the treatment of post-thalamic stroke pain.¹⁵^,¹⁶ In 1998, Ngyuen and coauthors reported using stimulation of M1 to treat PD.¹⁷ A MEDLINE review of articles from 1991 to 2008 revealed 512 cases in which MCS was used for the treatment of pain (422 cases),^15,¹⁸^–⁵⁸ poststroke rehabilitation (9 cases),^19,^59,⁶⁰ or movement disorders (84 cases), 22 of which were combined with DBS.^17,⁶¹^–⁷³ Some studies included previously reported cases from earlier references that the authors tried to differentiate but may have missed.

As of May 2008, Tsubokawa and associates at Nihon University had the largest published experience with MCS, 88 cases in all.^29,^33,^65,⁶⁶ Their work included comparisons of DBS and spinal cord stimulation (SCS) to MCS, predominantly for the treatment of chronic pain and movement disorders. They reported that 59% of patients who underwent DBS or MCS for involuntary movements resulting from stroke experienced some benefit, mostly relief of hemichorea and resting tremor.³⁰ Nineteen percent of patients who underwent MCS primarily for pain control also showed improvement in their movement problems, which consisted of hemiparesis predominantly. The results for treating phantom limb pain were less promising.³³ Only 6 of 19 patients had good pain relief from SCS alone. Ten of those who failed SCS eventually underwent ventrocaudal thalamic DBS. Of these, 6 patients had good long-term pain control. Of another 5 patients who failed SCS, only 1 experienced greater pain relief when subsequently treated with MCS. Nevertheless, this patient still experienced only moderate pain reduction. In another study of 12 patients with various brain lesions (6 with thalamic stroke, 3 with stroke in the posterior limb of the internal capsule, 1 with pontine stroke, 1 with multiple sclerosis, and 1 with postrhizotomy pain), 6 achieved “excellent” outcomes (defined as 100% pain improvement) at 1 year, with 5 of 11 patients maintaining this response after 2 years.¹⁵^,⁵¹ Three patients whose responses were initially categorized as “good” (60% to 80% improvement) were recategorized as either “fair” (40% to 60% improvement) or “poor” (<40% improvement) at 1 year. In 1998, Katayama and coauthors reported that although satisfactory pain control could be achieved in 23 of 31 poststroke patients initially, at 2 years pain control was still satisfactory in only 8 of these initial 23 patients.³¹

Several groups in Italy have developed a moderate experience with the use of MCS primarily for the treatment of movement disorders.^24,^39,^61,^62,^64,^68,⁷¹ Early results from these studies were promising, but as time progressed the benefits diminished. In a study sponsored by the Italian Neurosurgical Society,⁶⁸ 10 patients were monitored for 3 to 30 months after MCS for PD. All but 1 patient were ineligible for DBS because of age, findings on magnetic resonance imaging (MRI), or neurocognitive difficulties. The mean duration of disease was 12.4 years. Three patients improved less than 25%, 6 improved 25% to 50%, and 1 improved more than 50% as measured with a global rating scale. The stimulation parameters were as follows: 2.5 to 6 V, 150 to 180 µsec, and 25 to 40 Hz. Over time, these beneficial effects were lost, as reported in later presentations.⁷⁴^,⁷⁵

Programming and Complications

Historically, stimulation parameters for the treatment of chronic pain have varied widely. Stimulation amplitudes have ranged from a minimum of 0.5 V to the device’s maximal output of 10.5 V, stimulation frequencies from 5 to 210 Hz, and pulse widths from 1 to 500 µsec.* The most common adverse effect is induction of seizures, the majority of which occur during the initial testing phases. Hardware failure and infection are the second and third most likely complications, respectively. Both Myerson and colleagues³⁵ and Nguyen and associates⁷⁷ reported postoperative hematomas; one was asymptomatic and one required surgical intervention. Three groups have described subdural instead of epidural electrode placement.^47,^53,⁵⁴ In one of these reports two of nine patients suffered hemorrhages, one of which was fatal.⁴⁷ The other patient was left in a persistent vegetative state.

Physiologic and Clinical Rationale for Motor Cortical Stimulation

M1 is the final common link between the deeper circuitry that coordinates movement and the part of the spinal cord through which the movement is realized. Moreover, M1 is one of the few brain regions where the pyramidal and extrapyramidal systems interact directly. Because the symptoms of extrapyramidal disorders require a functioning pyramidal system to become manifest, it is reasonable to hypothesize that movement disorders such as PD, tremor, and dystonia may respond to therapeutic electrical stimulation at this target, particularly in light of the clinical precedents noted earlier.

On a practical level, MCS is potentially safer than DBS in that penetration of the brain parenchyma (and in the case of epidural stimulation, even exposure of the brain) is avoided. It may be performed with general anesthesia, does not require frame-based stereotactic guidance or technically advanced neurophysiologic mapping, and virtually eliminates the possibility of intraparenchymal hemorrhage. Consequently, it is important to consider MCS as an alternative or even preliminary procedure to DBS, particularly in patients with PD and refractory pain syndromes involving the face or upper extremity because these conditions are represented on the more accessible lateral convexity of the M1 cortex.

Methods

Surgical and Cortical Mapping Technique

Although the M1 region can be accessed via bur hole,^24,^39,^40,^64,^68,^74,⁷⁵ we prefer a small craniotomy, which allows us to perform detailed electrophysiologic mapping of the region and to suture the stimulating electrode to the dura so that it remains in place. The operation is performed with the patient supine. The anesthetic technique must be tailored so that neurophysiologic mapping can be performed. We prefer a complete total intravenous anesthetic protocol using a continuous infusion of propofol combined with either fentanyl or remifentanil. Standard doses are in the range of 75 to 150 µg/kg/min of propofol and 0.05 to 0.5 µg/kg/min of fentanyl or remifentanil. Inhaled agents are avoided, particularly nitrous oxide, because of their detrimental effects on motor evoked potentials. No muscle relaxants are used except during induction/intubation.

We do not use three-point cranial fixation because rigid fixation is unnecessary and could cause injury to the patient if a seizure is induced during cortical stimulation/mapping. To center the incision over the M1 region, the midpoint from nasion to inion is determined,* and a curved incision extending from approximately 1 cm behind this midpoint toward the anterior margin of the tragus is marked on the scalp before preparing and draping the surgical field (Fig. 93-1). This incision allows a 5- to 6-cm craniotomy, which can be adjusted superiorly or inferiorly along the convexity to map the hand or face, respectively. Even though the craniotomy lies mainly above the superior temporal line, the superior edge of the temporalis muscle may be detached when access to the facial region of M1 is required. One may leave a cuff of temporalis fascia for reattachment of the muscle during closure.

FIGURE 93-1 Schematic drawing denoting the external landmarks and measurements for planning craniotomy for motor cortical stimulation. Nasion-inion (N-I line) measurements are often about 36 cm. The surgeon then makes an imaginary line between a point approximately 1 to 2 cm behind the midpoint of the N-I line and the outer canthus. A curvilinear incision may then be made over this imaginary line, more superior for the hand and arm areas and more inferior along the convexity to access the face region. This should allow a small craniotomy approximately 5 to 6 cm in diameter centered over the M1 region of interest. The lead may be manipulated for mapping and then secured to the dura as shown. The lead wire may then exit through the posterior and inferior bur hole.

Some surgical teams advocate using functional MRI (fMRI) to localize M1 preoperatively.^37,^40,^43,⁵⁰ In our experience, these studies provide no information that cannot be obtained by intraoperative mapping; however, validation of fMRI might one day obviate the need for a small craniotomy or intraoperative mapping, or both. Nevertheless, the potential for target inaccuracies as a result of brain shifting, the desire to better secure the electrode to the dura, and the potential for misrepresentation of functional activity on fMRI have kept us from adopting this technique to date.

Intraoperative Neurophysiologic Mapping

We use both somatosensory and motor mapping to determine the course of the central sulcus and the position of the M1 region. Somatosensory testing consists of placing the paddle electrode (Resume, Medtronic, Inc., Minneapolis, MN) on the dura in a variety of orientations, mostly perpendicular to the suspected precentral gyrus. Median or ulnar nerve somatosensory evoked potentials (SSEPs), or both, are then performed by using a 20- to 50-mA, 100-µsec monopolar square pulse at a rate of 4.32 Hz. The SSEPs are recorded from the lead in both a bipolar (e.g., contacts 0-1, 1-2, and 2-3 for a 4-contact lead) and monopolar (all referenced to the 10-20 location of Fz) recording montage. The central sulcus is defined as the point at which the N20 response phase reverses polarity (Fig. 93-2). Mapping is performed epidurally over the entire exposed area and for short distances under the bone if needed. Typically, 7 to 10 recordings are obtained, which takes no more than 10 minutes of operative time. Early experience with this technique was gained with use of the 4-contact Resume lead (Medtronics, Minneapolis, MN); however, recently, we have used a 16-contact lead connected to a rechargeable implantable pulse generator (IPG) system (Precision Plus, Boston Scientific, Niatick, MA), a constant-current system that in our initial experience has yielded quicker onset of pain relief. Long-term results with this system still need to be evaluated.

FIGURE 93-2 Use of the SSEP phase reversal method to localize M1 extradurally. Contacts 0 to 3 of a typical 4-contact paddle-type electrode are shown with placement across the underlying central sulcus. The upper left waveforms show the SSEP in each contact. Reversal of the phase occurs between contacts 1 and 2 in this example. The inset at the lower left is the intraoperative photo of this technique being used, which reveals the relative size of the lead to the craniotomy. By placing the lead in different positions, the path of the sulcus can be mapped epidurally.

Motor mapping is performed with a 5-mm ball probe (Model E1564, Valleylab, Gosport, UK) that is applied on the dura over the suspected M1 area. The probe functions as the anode and is referenced to a cathode placed at Fz. Electromyographic recording needles are placed in bipolar fashion in the orbicularis oculi, orbicularis oris, trapezius, deltoid, biceps, triceps, flexor carpi ulnaris, abductor pollicis brevis, first dorsal interosseus, quadriceps, tibialis anterior, and abductor hallucis muscles. Stimulation consists of short trains of five to nine stimuli at a rate of one train per second, a pulse width of 500 µsec, and a 4-msec interspike interval. Stimulation is begun at each location starting at 5 mA and increased in 1-mA steps until a motor response is elicited or the maximal stimulus of 25 mA is reached. Ice-cold saline is kept ready in the field to irrigate the dura should a seizure occur. Barbiturates may be given intravenously as well if necessary.

In our experience, the mean stimulation amplitude required to generate an evoked response is 13.0 ± 4.2 mA (range, 7 to 22.4 mA). Stimulation is halted when the first electromyographic response is noted. Figure 93-3 displays electromyographic recordings taken from the extensor carpi ulnaris and abductor pollicis brevis muscles with this technique. We reiterate the importance of anesthetic technique for successful performance of physiologic mapping.

FIGURE 93-3 Use of cortical mapping to determine the location of M1 subregions. The electromyogram from muscles in the upper extremity is shown after stimulation with a ball-tipped electrode in three locations on the dura. The tracing at the far left shows activation of the extensor muscles in the forearm, with the other two tracings showing activation of the abductor pollicis brevis (APB). This technique corroborates the SSEP method in Figure 93-2 and helps determine more precisely the underlying thresholds for individual muscle groups. Extensor, extensor carpi ulnaris; FDI, first dorsal interosseus; Flexor, flexor carpi ulnaris.

Once the location of the appropriate overlying region of dura is identified, the lead is sutured to the dura with 4-0 Neurolon. The cable exits through one of the bur holes and is tunneled subcutaneously to an infraclavicular pocket fashioned to accommodate the IPG. The incisions are closed in standard fashion.

Programming the Implanted Device

The implanted stimulator is activated within 24 hours of surgery. During the initial programming session we evaluate the effects of cathodal stimulation via each contact in a monopolar configuration (i.e., referenced to the IPG) by using pulses 210 µsec in duration at 130 Hz. The amplitude is raised slowly to 4.0 V while the patient is assessed for adverse motor and sensory phenomena. The use of cathodal stimulation might be questioned given that anodal stimulation administered via transcranial electrical and magnetic stimulation generates D waves and motor evoked potentials at lower intensities.⁷⁸ Our use of cathodal stimulation is dictated by the design of contemporary implantable stimulators. Moreover, work by Hanajima and coworkers has revealed that chronic cathodal stimulation activates neurons in the motor cortex at lower stimulation intensities than does anodal stimulation.⁷⁹

As in the operating room, generation of seizures during this testing phase is of primary concern. Two of our patients have experienced seizures during testing. One occurred 2 days after the initial programming in a patient who also had a seizure during intraoperative mapping. In response, we now stimulate patients at or below 2.5 V for the first 2 weeks after surgery if they experienced a seizure during intraoperative mapping. We do not use antiepileptic medications, however. The second seizure occurred in a patient with a stroke 18 months after surgery when the stimulator voltage was raised from 4.0 to 4.3 V. This patient has since been maintained below 4.0 V with no further seizure activity.

In our four PD patients, we stimulated M1 via contacts 1 and 2 referenced to the IPG at an amplitude of 3.0 V. If the patient could not tolerate these settings, either contact 3 or 0 alone was used instead. In all cases, at least two contacts were activated at the initial programming session. If minimal benefit was noted from these initial settings, additional contacts were activated. If monopolar stimulation failed, one contact (either contact 0 or contact 3) was set as the cathode with the other three contacts set as the anode. We investigated additional parameter adjustments, including increasing the stimulation pulse width to 410 µsec, increasing the amplitude to 4.5 V, or changing the frequency to either 185 or 80 Hz. Table 93-1 shows the stimulation parameters relative to the surgical implant date.

TABLE 93-1 Stimulator Settings at Last Follow-up in Four Parkinson’s Disease Patients Who Underwent Unilateral or Bilateral Motor Cortical Stimulation

Initial device programming in patients with either mixed motor/pain syndromes secondary to stroke or pure neuropathic pain is similar to that described for those with PD. One key difference is the influence of the anatomic location of the patient’s pain and the results of intraoperative mapping on which contacts are activated. For upper limb pain, care is taken to note intraoperatively which contact or contacts overlie the corresponding painful area as determined by motor mapping. The electrode configuration is determined by the contact that generates the greatest motor response at the lowest stimulus intensity intraoperatively. Stimulation amplitude starts at 2.5 V, but in some cases it has been increased to 8.0 V or higher. Pulse widths begin at 210 µsec but may be adjusted up or down (range, 90 to 330 µsec). Stimulation frequency begins at 130 Hz, but in our experience a variety of frequencies have been tried with varied success.

Clinical Results

Through July 2008, we had performed MCS on 16 patients (Table 93-2). Among the 4 PD patients, complications were restricted to one intraoperative seizure and one infection, which required device removal. At the time of explantation this patient had improved approximately 50% in comparison to baseline but regressed significantly when the device was removed. A new device was implanted 1 year later, again with nearly 50% improvement. Two PD patients (MCS1 and MCS2) experienced 20% to 30% reductions in their medication requirement at 3 months, but patient MCS2 required more medication at the 6- and 12-month time points. One patient (MCS4) required 37% more dopaminergic medications postoperatively.

TABLE 93-2 Results of Motor Cortical Stimulation in 16 Patients

The motor subscores of the Unified Parkinson’s Disease Rating Scale (UPDRS-III) for our four PD patients are given in Tables 93-3 and 93-4. A significant improvement in the UPDRS-III scores is discerned up to 6 months postoperatively, after which motor function appears to regress. Two-year follow-up in two patients (MCS2 and MCS3) showed a return to their preoperative baseline as measured with the UPDRS-III while the patients were taking medications. Both patients noted reduced motor function. In contrast, patient MCS3 was wheelchair and bed bound after his generator lost power and is now able to feed and dress himself and walk with assistance with a new generator.

TABLE 93-3 UPDRS Part III (Motor) Scores during ON Medication and ON Stimulation

TABLE 93-4 UPDRS Total Scores during ON Medication and ON Stimulation

In our series of patients with neuropathic pain the results are even less consistent. Among the nine patients in whom pain was the dominant symptom, six experienced poststroke pain, one (P3) had atypical head pain secondary to trauma, one (P9) had a complex regional pain syndrome (CRPS) secondary to brachial plexus avulsion, and one had CRPS of the wrist. Of the six stroke patients, four suffered from both pain and a movement disorder. One of these four (P7) received no benefit from MCS. Two others (P2 and P4) experienced varying degrees of benefit. Patient P2 has good control of her upper extremity and face pain, but less pain control in her leg and minimal control of sensations of a “third limb.” Patient P4 has fair control of both her pain and sensation of an “internal” tremor. The patient with atypical head pain (P3) has experienced no relief in his scalp contralateral to the stimulation but is receiving benefit in his ipsilateral scalp region, which is not easily explained but may be related to stimulation of the trigeminal branches innervating the dura.⁸⁰ In addition, this patient found that his posttraumatic seizures improved with MCS, although it should be noted that he had no EEG seizures during a 24-hour EEG evaluation. A wide range of stimulation parameters have been used in these patients, with amplitudes ranging from 2.2 to 4.2 V, pulse widths from 120 to 330 µsec, and frequencies from 50 to 130 Hz. Stimulation in all patients started with monopolar configurations, but bipolar settings were often used with more anodal contacts.

For our three patients with movement disorders other than PD (M1, M4, and M7), two with notable tremor and one with corticobasal ganglionic degeneration, MCS yielded no benefit.

Despite wide variation in results, some of our patients have experienced modest success. Recently, three patients (one with PD and two with poststroke pain syndromes) contacted us complaining of worsened symptoms. In each case the IPG was depleted and symptoms improved with implantation of a new device. None of the patients were able to check the devices independently and thus had no means to discover that they were depleted.

Conclusion

It is important to compare MCS directly with DBS, even at this early stage of development, because DBS is now recognized as the most effective surgical therapy for medically refractory movement disorders. In light of the documented efficacy of DBS, the primary reason to investigate MCS as a treatment option is the potential to reduce neurological complications by eliminating the need to penetrate the brain parenchyma. Most published series suggest that the risk for hemorrhage related to DBS is less than 2%, with a range of 0.2% to 9.5%.⁸¹ Additional advantages of MCS include the ability to perform the surgery under general anesthesia, elimination of the stereotactic head frame, and the speed with which M1 can be localized intraoperatively. Consequently, if MCS were shown to be as effective as DBS for treating the cardinal symptoms of PD, it would become the preferred surgical option.

In contrast to previous reports,⁶⁸ our results suggest that the benefits of MCS for PD are more modest and short-lived than the responses generated by DBS. Nevertheless, 2 of our patients (M3 and M4) still derive benefit from MCS years after implantation. One other preliminary study of MCS for PD in which the stimulating electrodes were placed subdurally found benefit only for tremor.⁸² Akinesia, rigidity, and gait were not improved. It must be remembered that to date, only 27 patients with movement disorders have undergone MCS.^62,^68,⁸² Therefore, more data are required to determine whether MCS is a viable option for patients with movement disorders who cannot undergo DBS, whether for general medical or neurocognitive reasons. Moreover, the optimal technique for performing MCS remains to be determined because electrode orientation (perpendicular or parallel to M1) and stimulation parameters (electrode polarity, frequency, pulse width, and so on) are matters of great debate.

Another disadvantage of MCS is a delay in efficacy, which in some cases has been many months. Drouot and colleagues hypothesized that secondary messengers or long-term potentiation may account for the delayed response.⁸³ One may also hypothesize that chronic MCS alters not only firing patterns in the basal ganglia but also interactions between the pyramidal and extrapyramidal systems, which requires both the thalamocorticobasal ganglionic and corticocortical loops to adjust to these new signal patterns.

At this juncture it remains unclear whether MCS represents a failed therapy or a fledgling approach awaiting further refinement. Our experience, although quite limited, suggests that in certain cases there is long-lasting benefit that might be enhanced with additional knowledge. Therefore, further work will be required to define the proper indications for MCS, to refine the surgical technique, and to optimize stimulation parameters before the full potential of this therapy is realized.