CHAPTER 269 Intraoperative Monitoring of the Spinal Cord and Nerve Roots

Intraoperative neurophysiologic monitoring is an important technique that is essential for improvement of safety in modern complex spine surgery.¹^–³ In its ideal form, a properly engineered intraoperative monitoring environment provides real-time feedback information to the surgical team about the functional status of the spinal cord and nerve roots under surgical manipulation so that preventive or corrective actions can be taken to avoid irreversible injuries.⁴^–⁸ To gain universal acceptance and facilitate utilization, an ideal monitoring environment should include multiple properly designed human-machine interfaces, some of which involve biomedical engineering. It is also imperative that an optimal team of neurophysiologists and neurosurgeons be assembled to properly and efficiently use these technologies. Intraoperative monitoring modalities should have high sensitivity, high specificity, low invasiveness, and ease of use. Although numerous intraoperative monitoring techniques have been investigated since the days of Penfield and Boldrey,⁹ no single monitoring method has been able to satisfy these requirements in clinical spine surgery practice. However, general consensus regarding the use of intraoperative neurophysiologic monitoring is emerging as the evidence for monitoring continues to build,¹⁰ thereby encouraging optimism for the development of an ideal intraoperative monitoring modality.

This chapter aims to provide a concise overview of conventional monitoring techniques commonly encountered in clinical practice, such as somatosensory evoked potentials (SSEPs), electromyography (EMG), motor evoked potentials (MEPs), and compound muscle action potentials (CMAPs), as well as their respective supporting evidence. Modern multimodality comprehensive monitoring includes additional techniques such as external anal sphincter (EAS) and external urethral sphincter (EUS) monitoring. The value of comprehensive monitoring is exemplified in technically challenging spine microsurgeries, such as those involving surgery on a tethered spinal cord, conus medullaris, and cauda equina. Selected case examples are provided to illustrate the clinical scenarios. The effectiveness of intraoperative monitoring is intimately coupled with appropriate choices of anesthetics and intraoperative management of patient temperature and mean arterial pressure (MAP). Therefore, special anesthetic considerations are also reviewed in this chapter.

Finally, although newer techniques and applications of Doppler ultrasound imaging, spinal cord mapping, and other techniques are constantly being developed, we aim to provide a review of the current, state of the art equipment routinely used for intraoperative monitoring. We also look at potential future applications, where advances in quantum physics and biomedical engineering may provide a new direction in the next generation of multimodality monitoring techniques.

Conventional Monitoring Modalities and Evidence for Monitoring

A subset of electrophysiologic monitoring techniques is commonly used in routine clinical practice, such as SSEPs, EMG, and MEPs. Each technique has its own advantages and disadvantages, and the choice of one or a combination of several should be carefully considered in individual patient scenarios as part of the overall surgical management plan. Monitoring plans are made in consultation with the electrophysiologist and anesthetist and should always take into consideration preoperative information such as the clinical examination, previous electrophysiologic testing, structural anatomic imaging, comorbid systemic diseases, and tolerance of the planned anesthesia. All these factors have implications on the methodology and reliability of intraoperative monitoring. The surgical site, positioning, and draping may also interfere with electrode placement and sometimes necessitate the need for combined sterile and nonsterile leads as part of the surgical planning. Increased use of intraoperative imaging such as fluoroscopy, optical tracking–based spinal navigation, and intraoperative C-arm computed tomography increases the complexity of access around the patient. The temporal flow of the preincisional period in modern complex spine surgery is therefore an integral part of the planning of intraoperative monitoring and involves consideration of the presence of the surgical team, the electrophysiologist, the anesthetist, medical imaging technicians, and other support staff. All these individuals must be organized to ensure proper electrical integrity of the monitoring modality used and to obtain appropriate postinduction baseline measurements with reproducibility and redundancy. Confidence in the electrical integrity and baseline measurements and anticipation of movement or repositioning of equipment around the patient during the operation provide the necessary environment to proceed with surgery and maximize the utility of intraoperative monitoring. With these considerations in mind, the next sections provide a focused overview of commonly used individual intraoperative monitoring techniques.

Somatosensory Evoked Potentials

Monitoring of dorsal column integrity with SSEPs is the most commonly used technique in spine surgery.¹¹ Early monitoring solely with SSEPs resulted in false-negative results in multiple reports and thus necessitated the development of multimodality monitoring.¹²^–¹⁴ Nevertheless, SSEPs remain a standard modality with demonstrated improved neurological outcomes in corrective procedures for spinal deformity,^4,^6,⁷ and it serves as a building block for more complex monitoring arrangements. Multiple reports of the use of SSEPs in spine surgery, ranging from the early 1980s to current prospective studies involving larger patient populations, have demonstrated an improvement in outcome.¹⁵^–¹⁷

In SSEP monitoring, stimulation electrodes excite controlled repetitive action potentials propagating from peripheral nerves to dorsal roots, the dorsomedial tracts of the spinal cord, and eventually the contralateral sensory cortex, along which multiple recording sites may be anatomically accessible. Typically, platinum subdermal needle electrodes are used for both purposes. Stimulation sites for the upper extremity include the median and ulnar nerves. For the lower extremity, the posterior tibial nerve is a standard location. Bipolar stimulation (requiring both cathode and anode placement) is typically used. Electrode impedances are key parameters to ensure contact and reduce artifacts. Impedance values below 1 to 5 kΩ are clinically used thresholds for interelectrode or individual electrode measurements, respectively. Typical stimulation waveforms are 250-µsec-duration square-wave pulse trains at 4.7 Hz, with stimulation amplitudes ranging from 20 to 40 mA, depending on patient variables. Recording electrode positioning follows set standards, such as the international 10-20 system (Fig. 269-1).¹⁸ For example, Cv2-Fpz for the N31 subcortical response at the cervicomedullary junction and Cpz-Fpz for P37/N45 responses at the cerebral sensory cortex are commonly used. Other recording sites such as Erb’s point are also used. It is worth noting that after neural stimulation of the lower extremities (e.g., at the posterior tibial nerve), SSEP monitoring also reflects the physiologic integrity of the spinocerebellar tracts, which involve the dorsal nucleus of Clarke’s column between T1 and L2.

FIGURE 269-1 The international 10-20 system seen from the left (A) and above the head (B). A, ear lobe; C, central; F, frontal; F_p, frontal polar; P, parietal; P_g, nasopharyngeal; O, occipital. C, Location and nomenclature of the intermediate 10% electrodes, as standardized by the American Electroencephalographic Society.

(From Jasper HH. The ten-twenty electrode system of the international federation. Electroencephalogr Clin Neurophysiol. 1958;10:371-375. Reprinted with permission.)

Recorded responses require bandpass filters to remove artifacts and continuous averaging over multiple stimulation pulse trains to improve the signal-to-noise ratio. Comparison is made with the postinduction baseline measurements. The resultant changes, such as a 50% decrease in amplitude with an associated 10% increase in latency in comparison to the patient’s baseline values, prompt electrophysiologists to alert the surgical and anesthetic teams. The morphology of the recorded SSEPs may also be associated with intraoperatively detected changes in amplitude and latency, thus requiring a comprehensive assessment and evaluation by the electrophysiologist. Nonsurgical variables such as depth of anesthesia, patient temperature, and MAP have been shown to cause changes in SSEPs, and these variables must be ruled out when changes occur, or corrective actions must be taken. If surgical manipulation is determined to be the cause, timely cessation of additional manipulation, coordinated management to improve spinal cord perfusion, or initiation of intraoperative medical/steroid treatment would be required. Administration of a wake-up test may be necessary in select clinical scenarios.

Subcortical signals have lower signal amplitude and a reduced signal-to-noise ratio within a given amplifier gain and averaging setting but are more resistant to changes in depth of anesthesia. Cortical signals have larger amplitude and a higher signal-to-noise ratio and therefore tend to be more reliable electrically but are sensitive to anesthetic-induced changes. This tradeoff is illustrated in the improvement in reliability, approaching 93% to 98% in selected studies,^4,¹⁹^–²⁶ when using multiple cortical and subcortical recording sites. Recent advances in digitization equipment, filtering algorithms, and signal processing have resulted in the continued reduction of system complexity, which may potentially translate into a reduction in cost or an increase in the number of concurrently monitored channels.

The continuous recording nature and ease of setup are the main advantages of SSEPs, thus explaining its widespread use. The main disadvantage of the SSEP technique lies in the fact that only somatosensory pathways are monitored and, therefore, although SSEPs can be highly sensitive to surgically induced sensory deficits, motor deficits can be missed in a significant proportion of patients. This issue leads to the combination of additional modalities such as MEPs.

Motor Evoked Potentials

Transcranial electrical or magnetic stimulation of the cerebral motor cortex causes muscle activation, which can be detected peripherally to provide additional monitoring coverage specifically designed to detect motor deficits.^27,²⁹ This is particularly important with intramedullary resections and in procedures with a risk for vascular compromise of the anterior spinal artery and its associated vascular territory, which can affect motor function. Initial work^28,²⁹ paved the way for current single-pulse stimulation techniques, in which corticospinal tract recordings of direct waves (D waves) are detected, and multipulse stimulation techniques, in which short pulse trains are used to elicit CMAPs.³⁰

Typically, in patients under general anesthesia, MEP monitoring involves an electrode montage over the scalp based on the international 10-20 electroencephalography system. The standard setup uses C3-4 for upper limb MEPs and C1-2 for the lower limbs, with selection of corkscrew, needle, or cup electrodes being based on the clinical scenario and the patient’s age. Skull thickness, electrode type and impedance, and the need for depth of stimulation determine the required stimulation amplitude. Typically, selective stimulation is preferred, and therefore superficial coverage of a small cortical area is used to provide activation of a single limb. With few exceptions, D waves directly generated by the electrical activation of axons of cortical motor neurons are recorded in patients under general anesthesia. This is due, in part, to the fact that vertically oriented excitatory chains of neurons terminating on the cortical motor neuron are blocked by anesthetics, thus abolishing indirect activation of the motor neuron and the associated indirect (I) waves.¹ For D-wave recording, one or more sterile intraoperative epidural electrodes are required and placed caudally and cranially relative to the surgical lesion. Signal amplification typically requires a gain setting of 10,000× and bandpass filtering between 1.5 and 1700 Hz. Only a few traces are required for averaging, and the stimulation can be repeated at 0.5 to 2 Hz to provide real-time feedback.¹ A reduction in D-wave amplitude of greater than 50% has been shown to correlate with new postoperative motor deficits.³¹ Myogenic MEPs use a short (5 to 7) pulse train with a 4-msec interpulse interval at the transcranial electrodes. Recordings of the resultant CMAPs at sites such as the thenar muscles or tibialis anterior are amplified at 10,000× and bandpass-filtered between 1.5 and 853 Hz. Again, multitrace signal averaging is not frequently required, and real-time feedback is feasible.

Multipulse stimulation and CMAP detection provide the benefit of monitoring the corticospinal tract, in addition to the distal muscle functional units, including the neuromuscular junction. Although the muscle units provide a physiologic nonlinear amplifier for the MEPs, which improves the electrical signal-to-noise ratio, the nonlinear relationship and preoperative muscle atrophy may cloud the interpretation. Thus, myogenic MEPs are typically interpreted as an all-or-none phenomenon with regard to preoperative clinical assessment and electrophysiologic findings. When MEP amplitude is lost, immediate intraoperative intervention can result in complete reversal, which can be sensitive and specific in predicting new postsurgical motor deficits.³² In conditions in which reliable transcranially stimulated MEPs are difficult to obtain, neurogenic MEPs, obtained through percutaneous stimulation, have been investigated.^33,³⁴ When combined with SSEP monitoring of the dorsal column, MEPs provide additional coverage for the ventral corticospinal tract and improve the efficacy of intraoperative monitoring (Fig. 269-2). The intermittent nature of MEP monitoring requires the surgeon to actively request interval measurements. These measurements are generally first taken as a baseline before surgery and then at surgical opening, during exposures approaching the spinal cord, on manipulation of the spinal cord, during different phases of intramedullary resection, and at closure. Close interaction between the neurophysiologist and the surgical team also allows periodic pauses intraoperatively to ensure that reliable recordings are being obtained throughout the procedure.