Traditional auditory brainstem responses (ABR) are valuable, noninvasive measures that assess hearing impairment, but their effectiveness in detecting small tumors and diagnosing Meniere's disease have been limited. Relatively new ABR tools in retrocochlear and cochlear assessment can now be used to identify these conditions.
ABRs are electrical potentials recorded from the scalp surface and represent synchronous electrical activity of neural elements in the auditory brainstem in response to acoustic stimulation (for an example, see Figure 1 [PDF]). It can be seen that within the first 15 milliseconds after the sound is presented, there are peaks and valleys of the evoked electrical activity. Seven peak components or waves have been defined and are numbered with Roman numerals. All peaks are not always observed except for the largest peak, which is typically wave V.
ABRs are considered objective measures because no behavioral responses are involved; therefore, ABRs are ideal for screening for the presence and assessing the nature of auditory function impairment in infants and children in whom reliable and accurate behavioral measures are often difficult to obtain. The application in adults, however, is usually different. In adults, the focus is not to detect the presence of hearing loss per se because reliable behavioral measures are available. Rather, the focus is on determining either the presence or the underlying cause of the hearing problem, including any neurological problem affecting the auditory central nervous system. Thus, the question is not whether there is hearing impairment, but rather, is the impairment in the cochlea, auditory nerve, higher nervous system, or some combination of these auditory pathway sites?
Retrocochlear Assessment—Standard ABR Measures
There is a large body of literature demonstrating how ABRs have been used to identify the presence of problems affecting the auditory neural pathway in adults. One of the major applications in the field of otology has been in detecting the presence of tumors that affect the hearing and balance nerves. These tumors typically arise from Schwann cells in the sheath that surrounds the vestibular (balance) portion of the eighth nerve. Thus, these tumors are called vestibular schwannomas but are typically referred to as acoustic tumors because of their most notable affect on the auditory nerve.
Early application of ABRs for identifying the presence of acoustic tumors involved measurements of the change in the peak latency of the largest component of the ABR, wave V. Peak latency is a measure of the time between the onset of the sound presentation and the occurrence of the peak voltage of the wave V component in the ABR. In particular, the difference in the latency measures between the suspected tumor ear and the non-tumor ear is used for the diagnosis. This measure, the interaural latency measure of wave V (IT5), introduced by Selters and Brackmann (1977), had great utility in identifying acoustic tumors from about the latter part of the 1970s through the 1980s. One of these early studies (Eggermont et al., 1980) indicated that this measure would likely identify all large- and medium-size acoustic tumors but would likely miss small (1 cm ≥) tumors.
With the advent of magnetic resonance imaging (MRI), it was discovered in the 1990s by a number of studies (see reviews in Don et al., 1997; 2005a) that this IT5 measure and other latency-based measures missed a significant number of small acoustic tumors as predicted by Eggermont et al. (1980). As a result, many of these studies recommended abandoning the practice of ABR testing and relying on MRIs for screening.
However, the use of MRIs to screen for small tumors has several drawbacks: outside of large metropolitan areas, availability of MRI testing may be an issue; the cost may be quite high; MRI testing can be uncomfortable for some patients (claustrophobia and/or the typical use of venous injection of contrast material); and acoustic tumors are fairly rare, found in less than one in 100 who are imaged (Doyle, 1999). Thus, there appears to be a need for a screening tool that is more widely available, less costly, and more comfortable than the MRI, with the MRI reserved for the final diagnosis after screening. While the ABR method circumvents these MRI drawbacks, the all-too-frequent insensitivity of standard ABR measures to small tumors must be resolved if they are to be used for screening.
Why did the standard ABR latency measures frequently fail to detect the small tumors? Don et al. (1997; 2005a) and Don & Kwong (2002) hypothesized that small tumors frequently did not affect a sufficient number of nerve fibers related to the high frequencies. When the cochlea responds to a broad-spectrum stimulus (one that contains a broad range of frequencies) such as a click, the responses to the high frequencies occurs first, with succeeding lower frequencies responding at succeeding longer delays (cochlear traveling wave delay). Because these high-frequency fibers respond the earliest, they dominate the wave V peak latency of the ABR.
When the tumor compresses these high frequency fibers—delaying or removing their activity—the latency of wave V will increase, because the domination of activity will have been delayed and/or will be dominated by lower frequencies with longer latencies due to the traveling wave delay. If the tumor does not affect these high frequencies sufficiently, they will continue to dominate the wave V latency, and latency measures such as the IT5 will appear normal. As a consequence, the tumor is missed. The smaller the tumor, the more likely this is to be the case. Medium- and large-size tumors nearly always affect a sufficient number of high-frequency fibers such that the standard latency measures are abnormal. Thus, the standard IT5 measure is useful for detecting medium and large tumors. All we need is an additional test that has high sensitivity to the presence of small tumors.
Retrocochlear Assessment—Newer Tools
In response to this quest for an ABR measure that will improve ABR sensitivity to small tumors, Don et al. (1997; 2005a) introduced the application of Stacked ABR (SABR) amplitude to tumor detection in 1997. Unlike the standard ABR tumor detection measures that focus on peak latencies and depend on the subset of high-frequency fibers to be affected, the SABR is an amplitude measure not dependent solely on a subset of fibers. Instead, the SABR essentially represents the sum of neural activity initiated across the whole cochlea. The formation of this measure is detailed in Don et al. (1997; 2005a) and is briefly summarized in Figure 1 [PDF].
A series of ABR waveforms from a normal-hearing adult female is shown in Figure 1 [PDF]. The top trace is the response to clicks presented at 60 dB nHL and represents the sum of the activity initiated across the cochlea. Below this trace are a series of derived-band ABRs representing synchronous activity initiated from successive octave-wide regions across the cochlea with the theoretical center frequency (i.e., 11.3, 5.7, 2.8, 1.4, and 0.7 kHz) noted beside each of the derived-bands.
These derived-band ABRs are obtained using a special technique (not detailed here) that involves presenting clicks in the presence of successive high-pass noise masking and subtracting the successive responses. The technique was first described by Teas, Eldredge, and Davis (1962) in animals and applied to ABRs (Don & Eggermont, 1978; Parker & Thornton, 1978). It can be seen that the latency of wave V (noted by the filled circle symbol) increases as the center frequency of the derived-band response decreases. This delay in time with the lowering of frequencies reflects the well-known traveling-wave delay down the cochlea. If the derived-band ABRs are simply added together, the result is similar to the top trace. In addition, by adding the waveforms together, much of the electrical activity from the lower frequency regions of the cochlea will not be seen due to phase cancellation of the electric fields from the higher bands.
It can also be seen (dashed line) that the latency of the response to the clicks alone (the standard ABR, top trace) is similar to and determined by the latencies of the two highest derived-band ABRs (center frequency =11.3 and 5.7 kHz). Therefore, for standard ABR measures, a prolongation of the latency of wave V of the ABR to clicks presented alone will occur only if the tumor affects the high-frequency fibers. In a significant number of small-tumor cases, it appears that the high-frequency fibers are not sufficiently compromised, and there is little change in the wave V latency in the ABR response to clicks.
A solution to the problem of detecting small tumors is to use a measure that assesses activity from essentially all nerve fibers, not just a subset. The measure proposed is the SABR (Don et al.,1997; Don et al., 2005a). As seen in the last large trace in Figure 2 [PDF], the SABR is formed by shifting and aligning the wave V peaks of the derived-band responses shown in Figure 1 [PDF] (stacking the waveforms) and then adding the waveforms together. The SABR represents the sum of synchronized neural responses across the whole cochlea; thus, reduction of activity even by a small tumor will reduce the SABR amplitude.
Don et al. (2005a) demonstrated that SABR had a sensitivity of 95% and a specificity of 88% relative to non-tumor normal-hearing adults. In contrast, the standard IT5 criterion measure of 0.2 ms delay (Selters & Brackmann, 1977) detected only about half of these small tumors (50% sensitivity) with a high specificity (96%). Adjusting the criterion to 95% sensitivity dropped the IT5 specificity to less than 10% (Don et al., 2005a). Thus, the greater sensitivity of the SABR for small tumors warrants its consideration as a screening tool, especially where MRIs are expensive and not widely available. As the least expensive test, it also helps keep medical costs down and keeps audiologists in the forefront of small-tumor detection.
Assessment of Meniere's Disease/Cochlear Hydrops
Meniere's disease is defined as the idiopathic syndrome of endolymphatic (cochlear) hydrops, an abnormal increase in the volume of cochlear fluid (endolymph) in the inner ear. In the early stages of the disease, the symptoms are similar to those observed in small tumors. The hallmark symptoms of Meniere's disease are episodic vertigo, tinnitus, fluctuating hearing loss, and the sensation of fullness or pressure. Because these symptoms are not always present—especially at the onset of the disease—and many patients initially diagnosed with Meniere's disease actually do not have cochlear hydrops, there are difficulties in its diagnosis.
It is thought that the more prolonged the hydrops condition, the poorer the likelihood of a cure and the greater the chance of permanent cochlear damage and of a life-threatening attack of vertigo. Thus, early detection and diagnosis are important but difficult. As seen in a review of the literature, several tests have been proposed that can statistically distinguish, as a group, between Meniere's and non-Meniere's disease subjects. However, there usually is sufficient overlap between distributions of these test measures that diagnosis is equivocal for a number of individuals.
Recently, we proposed an ABR test that shows excellent promise for clearly distinguishing patients with Meniere's disease/cochlear hydrops (Don et al., 2005b). The test is called Cochlear Hydrops Analysis Masking Procedure (CHAMP) and the basic observations are seen in Figure 3 [PDF]. The waveform traces seen in the left panel of Figure 3 [PDF] are the series of responses to clicks and high-pass pink masking noise with varying cutoff frequencies. It is the successive subtraction of these responses that form the octave-wide derived bands shown earlier in Figure 1 [PDF]. However, instead of forming derived-band responses, we simply examine these high-pass masked responses.
For the non-Meniere's subject, with each successive lowering of the high-pass masking frequency, the latency of the ABR response to the clicks and the noise increases. This is the result of removing the contributions from the higher frequencies so that the latency is dominated by unmasked activity from the lower-frequency regions of the cochlea. However, as seen in the right panel for a patient diagnosed with active Meniere's disease, there is virtually no latency shift with each successive high-pass masking condition. In other words, the masking noise appears to be ineffective in sufficiently masking the high-frequency contributions. Thus, the ABR is still dominated by the high-frequency regions and there is little or no shift in the latency of wave V. The hypothesis is that the cochlear hydrops causes stiffness changes in the basilar membrane leading to ineffective noise masking. Don et al. (2005b) demonstrated that a quantitative measure of the amount of latency shift between the response to the clicks alone and the clicks with 0.5 kHz high-pass masking noise clearly separated the non-Meniere's disease subjects from patients with Meniere's disease.
An important next step is replication of the work by independent researchers. New ABR tools that will allow us to use ABRs to screen for small acoustic tumors and for assessing the presence of Meniere's disease/cochlear hydrops extend the value of ABR measures for assessing otologic pathologies in adults.