Somatosensory Evoked Potentials

Somatosensory evoked potentials are one of the most frequently used monitoring modality in neurosurgery and rely on the ability to record the transmission of a peripheral electric stimulation along somatosensory afferent pathways to the primary somatosensory cortex. Perturbations of the standard characteristics of this signal transmission can be exploited to detect early and reversible neuronal compromise.

Dawson first described SSEPs in 1947.¹¹ Initially used in the laboratory and clinical setting, further technical developments in the late 1970s and 1980s allowed for their use for intraoperative monitoring. Currently the acquisition of SSEPs is done according to international standards^12,13 following stimulation of peripheral afferent nerves, generally the median or ulnar nerve for upper extremities and posterior tibial or peroneal nerve for lower extremities. Recording from the surgically exposed cortex with electrode grids or from the scalp usually with needle electrodes allows recording of a phase reversal signal between the somatosensory cortex and the motor cortex induced by the peripheral nerve stimulation. This allows the reliable identification of the central sulcus. The monitoring of somatosensory pathways throughout a procedure requires continuous monitoring by repeated stimulation. To reduce the signal-to-noise ratio induced by the significant background EEG activity of the brain, it is necessary to average about 200 to 500 stimulation trials and to filter SSEP signals with a predetermined bandwidth acquisition. The stimulation of peripheral nerves is done with needle or surface electrodes generating a bipolar signal and a single ground electrode is used for all limbs (Fig. 21-1). Stimulus intensity between 20 and 30 mA and duration between 200 and 500 µs at approximately 5 Hz is adequate to obtain satisfactory responses. This setup allows for the continuous updating of information, and any changes can be transmitted to the surgical team at less than 1-minute intervals. To differentiate SSEP changes from central or peripheral origins, cranial as well as cervical and peripheral electrodes are routinely used. The stimulation wave can be recorded at several acquisition points with a specific latency (Table 21-1), and the amplitude and morphology of the responses are compared with the baseline preoperative recording of the patient after induction of anesthesia and before positioning and with the intraoperative recordings of the contralateral side that serve as an internal control.

FIGURE 21-1 Neurophysiologic monitoring technique used at Toronto Western Hospital. A, Subdermal needle electrode (Rochester Electro-Medical): used for all recording of SSEPs (including their stimulation) and recording of EMG. B, Monopolar Kartush dissecting stimulator (Medtronic) for evoked EMG. Standard electrical parameters of evoked EMG at Toronto Western Hospital: (1) constant current ranging from 0.05 to 0.50 mA (milliamperes); (2) 3.0-Hz stimulation rate; and (c) 100-μsec pulse width.

Although SSEP principally monitors the integrity of the sensory afferent pathways, it can indirectly monitor the nearby motor corticospinal tracts and motor cortex. Thus, there is a recognized utility of SSEPs for detecting disturbance not exclusively of somatosensory pathways but also of more general hemispheric changes. In this regard, there is a well established relationship between cortical cerebral blood flow and SSEP changes, which make it possible to identify vascular compromise in regions not only related to vascular supply of the somatosensory cortex. Importantly, SSEPs are also subject to change with general physiologic parameters such as blood pressure,¹⁴ body temperature, or anesthetic regimen, which have to be taken into consideration when interpreting changes seen during surgery.

The utility of SSEP monitoring has been repeatedly reported for a variety^8,15,16 of spinal^17–19 and cranial vascular²⁰ as well as tumoral^21–24 procedures. Based on our own experience with SSEP monitoring in more than several thousand cases of cranial procedures, we have found that reductions in amplitude of the cortical wave exceeding 50% are significant in predicting neuronal compromise and warrant alerting and reaction from the surgical team. Minor changes in amplitude and latency, which are not uncommon, are more likely related to changes in anesthesia or general physiologic parameters such as blood pressure, blood gases, and temperature. Changes in the morphology of the SSEP waveforms alone are more difficult to interpret and have no specific value in common clinical practice.

The utility of SSEP in tumor surgery including meningiomas has been assessed by several studies. Romstöck and colleagues reported a successful identification of the central sulcus with phase reversal in more than 90% out of 230 cases of central or paracentral tumors including meningiomas, but this was dependent of tumor location.²⁴ In large and centrally located tumors causing significant displacement and deforming the central cortex it was more difficult to obtain a reliable phase reversal response. Wiedermayer and colleagues have specifically assessed the false-negative findings during SSEP intraoperative monitoring and reported 4% of 658 vascular and tumor cases in which patients with uneventful monitoring developed neurologic deficits.¹⁵ They found that the overall sensitivity was 79% and negative predictive value 96%. Not surprisingly, in their experience SSEPs were less likely to predict deficits in infratentorial lesions compressing the brain stem, small cortical lesions of the motor cortex, and small vessel injury during aneurysm surgery, situations in which motor pathways could selectively be injured without disturbing sensory pathway or causing wider hemispheric damage. In summary, these and other studies confirm the usefulness of SSEP as a routine monitoring modality in intracranial procedures, but also prompt awareness of its limitations, especially for lesion in which selective impairment of motor pathways is the major surgical risk.

The value of SSEP in meningioma surgery in general has been assessed along with other tumors,²⁴ and cranial base meningiomas were included in several series describing the use of SSEPs in skull-base lesions.^5,10,25 Bejjani and colleagues have described several types of SSEP changes in the setting of skull-base tumor surgery.²⁵ The proposed classification, which we have also adopted, is as follows: type I, no significant waveform change from baseline; type II, waveform changes that returned to baseline (>10% increase in latency from baseline and/or >50% decrease of amplitude from baseline); type III, waveform changes that recovered partially but incompletely; type IV, complete flattening of the SSEP waveform without improvement; type V, flat waveform from the beginning of surgery. In a series of 244 patients, these authors reported a very good correlation between SSEP type and postoperative deficits with a positive predictive value of 100%, and a negative predictive value of 90%. In our own unpublished series of 122 procedures for skull-base tumors (excluding vestibular schwannomas) during a 4-year period, we found similar results with a positive predicting value of 100% and a negative predicting value of 89%. Another important aspect in skull-base lesions compressing the brain stem, especially near the craniovertebral junction and in mass lesions of the cervical spine, is the critical value of SSEP monitoring during the positioning of the patient: passive mobilization of the spine and craniocervical junction in anesthetized and relaxed patients carries a real risk of neurologic damage.²⁶ Further benefits to SSEP monitoring include detection of peripheral nerve injury by traction and/or compression (especially in the ulnar SSEP) attributed to improper positioning of the patient: in this situation early detection can prevent permanent nerve damage during a long procedure with complex positioning, such as in the park bench position (Fig. 21-2).

FIGURE 21-2 SSEPs of the left arm in a right cerebellopontine meningioma. Shortly after positioning a patient with a recurrent right cerebellopontine angle meningioma in the lateral, park-bench position a change in the cortical and cervical SSEPs of the left arm was noticed (A, in red, as compared to the green baseline). The left arm was repositioned and the neck position checked, with a subsequent recovery of SSEPs back to baseline (B). Postoperatively the patient experienced a transient left upper brachial plexus deficit. The Erb’s electrode did not document the deficit probably because it had been positioned in an atypical position, inferior instead of superior to the clavicle.

Motor Evoked Potentials

Because SSEPs offer a valuable but only indirect mean of monitoring motor pathways within the limits discussed in the preceding text, many efforts have been made to develop direct surveillance of motor pathway integrity during intracranial and spinal surgery. The identification of the motor cortex is possible through direct bipolar stimulation and observation of muscle contraction or muscle activity recording with EMG in either awake or anesthetized patients. However, as anesthesia has an impact on the stimulation threshold required to induce recognizable muscle contraction, and as awake surgery is not generally practical in most of brain and spine tumors including meningiomas, more refined monitoring modalities were developed for continuous intraoperative motor monitoring. The first attempts to explore the human cerebral cortex by evoking a motor response from a cortical electrical stimulation during surgery were described by Sir Victor Horsley and further refined by Penfield²⁷ and others. However, the scientific groundwork for motor evoked potentials (MEPs) stems from the work on monkeys of Patton and Amassian in the mid-1950s.^28,29 By applying direct cortical stimulation they were able to elicit two distinct recordable discharges from of the corticospinal tracts. The first one is a nonsynaptic early response named the D-wave translating a direct stimulation of the corticospinal tract and the second is a late, synaptic response called the I wave resulting from the action potential firing of stimulated cortical motoneurons. MEP was subsequently applied in humans either by direct cortical stimulation³⁰ or by transcranial electrical stimulation (TES) in awake patients.³¹ Recording of MEPs is possible either from electrodes placed directly in the spinal cord epidural space or currently more often by recording muscular EMG. Volatile anesthetic agents suppress the I response, and intraoperative MEPs in the 1980s relied on the more robust D wave recorded from the spinal epidural space. Transcranial magnetic stimulation²⁹ alleviating scalp electrodes but mainly inducing I waves is valuable in awake patients but does not offer a significant advantage in anesthetized patients because of the suppression of I waves induced by anesthetic agents. The major breakthrough for the reliable use of MEPs in anesthetized patients was the demonstration that the summation of short-train impulses at high frequency by direct cortical stimulation allowed to record MEPs (both the I and D waves) in the presence of most anesthetics.³² This was followed by further work establishing the feasibility of pulse-train TES.³³ Moreover, the wider adoption of intravenous anesthetic agents such as propofol further facilitated the intraoperative use of MEPs that are less suppressed with intravenous substances such as remifentanyl and propofol than with volatile anesthetics.^34–36

The value of current modern MEP’s techniques has been demonstrated by several authors for intra-axial and extra-axial lesions in the perirolandic area,^32,37–41 for intracranial vascular lesions,^39,42–45 and for spinal deformity and spine tumor surgery.^39,46–48 More recently, its use with transcranial stimulation has been suggested to monitor cranial nerves,^38,39,49 as discussed later. In clinical practice it has been established that more than a 50% amplitude reduction of the D wave during intramedullary spinal surgery is associated with injury of the corticospinal tract and long-term motor deficit.^50,51 In perirolandic surgery, more than a 30% to 50% reduction in the D waves is predictive of permanent deficit.³⁷ Preservation of the D waves is a good predictor of the absence of motor deficit although probably not with 100% predictive value. Fujiki and colleagues compared the value of corticospinal D versus I wave in brain tumors. They report that I waves recorded from the muscles are more sensitive than corticospinal MEPs recorded from the spinal epidural space that depend more on D waves.³⁷

Although the use of MEPs for intracranial meningiomas have been reported by some authors,^37,39 there is no convincing evidence of a benefit for the use of MEPs in surgery for these lesions. In our own practice we would recommend the use of MEPs in selective cases of meningiomas in or adjacent to the motor cortex, in parasagittal meningiomas with encasement of the anterior cerebral arteries or invasion of the superior sagittal sinus in its posterior two thirds, in spinal meningiomas, as well as in skull-base meningiomas with suspected encasement of major arteries of the circle of Willis (Fig. 21-3).

FIGURE 21-3 Indications for neurophysiologic monitoring in meningioma surgery. Schematic rendering of the suggested neuromonitoring technique according to the meningioma location: SSEPs are always used; MEPs are indicated when the motor pathways are particularly at risk (e.g., anterior spinal meningioma; anterior foramen magnum lesion); electrical cortical stimulation is useful in mapping the cortical areas involved or in the proximity of the tumor; VEP and extraocular muscles monitoring could be useful in parasellar meningiomas, but are still to be considered experimental; trigeminal nerve monitoring can be useful in middle and posterior fossa meningiomas; monitoring of the facial nerve, BAEP, and lower cranial nerves is essential in most posterior fossa meningiomas; evoked EMG can be useful in spinal meningiomas that involve specific functional neural elements at the level of the cervical and lumbar spinal cord enlargements.