Signal Optimization in Intraoperative Neuromonitoring

Abstract

Signal optimization in neuromonitoring attempts to make neuromonitoring data as close a representation of the neurologic state as possible to improve our ability to accurately identify new deterioration in neurologic function. This process depends on the correct identification of suboptimal signals, identification of underlying causes for suboptimal signals, and the implementation of steps to best counter the identified causes. A step-wise and logical approach to this process allows an efficient process that can meet the time-pressured needs of the typical surgical environment.

Intraoperative neuromonitoring (IONM) provides actionable insight into the state of the nervous system using an array of neurophysiologic tests. If new neural dysfunction occurs during a surgical procedure, we expect to be able to readily identify that impairment via an alteration of the electrical signals obtained. Unfortunately, this simple scenario is often greatly complicated by many factors that may alter a neurophysiologic signal unrelated to the state of the targeted neural pathways being assessed. Signal optimization is the process by which we eliminate or minimize these extraneous factors, thereby providing the most accurate insight possible into the state of the neural structures we hope to assess.

The first step toward signal optimization involves good basic neuromonitoring practices in order to obtain signals. Covering the basic techniques of signal acquisition and correct identification in IONM is a prohibitively expansive topic and hence we will assume such practices are already in place (see Chap. 20, “Monitoring Applications and Evaluating Changes”). Instead, the focus is on subsequent steps toward improving signals that do not meet expectations despite the use of good basic technique.

The next step is to assess the quality of testing results and this is reflected in the signal-to-noise ratio (SNR) of the responses produced. The ability to distinguish the targeted signal from noise is essential to understand the state of the nervous system, as noise may mask or mimic the changes IONM should identify. Moreover, any signals with poor initial quality (poor SNR) are more susceptible to additional impairment from ubiquitous operating room factors such as anesthetic sensitivity, flux in systemic parameters (e.g., blood pressure or oxygenation), or other nonsurgical factors (e.g., new electrical interference). The heightened intrinsic variability of poor signals increases the likelihood of false-positive findings if typical interpretative criteria are applied. In response to this, interpretative criteria will need to be altered, often at the expense of sensitivity, to prevent inevitable false-positive findings. Thus, signal improvement may serve to increase both the sensitivity and specificity of monitoring.

Next, we will seek to understand the reasons for signal impairment. In some cases, the reasons are easily and instantly recognizable to the experienced neuromonitoring person. However, when a root cause is not readily apparent, a successful search for the cause(s) will typically elucidate the solution. Our approach to gaining greater understanding starts by trying to categorize causes into patient-related factors, anesthetic/systemic factors, and factors related to signal acquisition technique (“technical factors”). These three factors encompass all causes of suboptimal signals and each, along with proposed remedies, is discussed in detail below, forming the emphasis of this chapter. Approaching problems with each of these in mind ensures the search is appropriately broad, provides a framework within which we can narrow possibilities, and aids in an efficient approach to the problem.

The final step in signal optimization is to remedy the factors leading to the suboptimal testing results. Some of these are directly addressable, such as noise from a faulty electrical device. Other factors, such as neural dysfunction in the monitored pathways, are not directly addressable and must be “worked around” either by avoiding nonfunctioning elements, choosing alternate tests, or when no further optimization exists, accepting and understanding the monitoring limitations of the available neurophysiologic responses.