PATIENT SELECTION

Patients originally received the ABI under a protocol monitored by the FDA. The criteria for implantation are listed in Table 58-1. The device originally was designed for patients with NF-2 manifesting bilateral vestibular schwannomas, although others with compromised auditory nerves were considered to be eventual ABI candidates. At least 90 per cent of NF-2 patients exhibit bilateral eighth nerve neuromas.³ An unpublished review of patients with NF-2 seen at the House Ear Clinic revealed that two thirds had bilateral internal auditory canal–cerebellopontine angle (CPA) tumors alone or with one other tumor as the only central nervous system manifestation of their disease. The patients were young (average age, 28 years). With improvements in medical care and surgical techniques, the life span of most of these patients has been significantly prolonged. Restoration of even rudimentary auditory function can enhance their quality of life and ability to function in a hearing world. Our results have shown that the multichannel ABI has the potential of offering even greater benefit.

TABLE 58-1 Criteria for Implantation

The current NF-2 protocol allows implantation at the time of first-or second-side acoustic neuroma removal or in patients whose tumors have previously been removed. Implantation during removal of the first tumor has allowed experience with the device and may enhance performance when the patient loses all hearing. Also, implantation on the first side gives the patient two chances at obtaining an optimally functioning system should the procedure in the first side not be successful.

The management of bilateral acoustic neuromas should be highly individualized.⁴ Hearing preservation remains an ideal goal in the management of these tumors in patients with NF-2, and early identification and treatment have permitted this in a number of cases. An intact auditory system is highly desirable in preference to an artificial means of restoring hearing. Therefore, preserving as much of the patient’s own hearing as possible is paramount. Patients meeting the criteria listed in Table 58-2 may be considered and observed accordingly. The availability of the ABI provides an alternative to a desperate attempt to preserve nonserviceable hearing when large tumors are removed and hearing conservation is unlikely.

TABLE 58-2 Criteria for Observation in Auditory Brainstem Implant Candidacy in Patients with Neurofibromatosis Type 2

The anticipated applications of the ABI and similar devices now has included bilateral temporal bone fractures, cochlear ossification, cochlear nerve avulsion, and anatomical birth anomalies, The ABI also may be beneficial in cases of demyelinating diseases affecting the eighth cranial nerve but sparing at least one cochlear nucleus.

PREOPERATIVE EVALUATION AND COUNSELING

The goal of implantation is to place a safe and stable device that provides the patient with some degree of environmental sound awareness and recognition and also improves communication in conjunction with lipreading without side effects. Prospective patients are apprised of the goals, limitations, and risks of the ABI during two or three preoperative evaluation and counseling sessions. It is important to impress on the potential candidate that, although the ABI is similar to a cochlear implant, it has provided generally lower levels of performance with more gradual improvement over time. The implant candidate’s expectations are carefully evaluated, and informed consent is obtained. The importance of an experienced multidisciplinary implant team including the neurotologist, neurosurgeon, audiologist, neuro/auditory physiologist, anatomist, radiologist, and others cannot be overemphasized.

Several factors contribute to a successful result from implantation. Experience of team members can greatly influence outcomes. Chief among these factors are the correct identification of the implantation site and the achievement of a stable placement of the electrode array. This, of course, is essential to obtain auditory sensations from stimulation and to optimize performance. The overall results of tumor removal and postoperative recovery also play a role. For example, factors such as eye dryness related to postoperative facial nerve function may affect lipreading ability and communication using the ABI. General health, social activity level, and presence of a support group also can affect ABI use and benefit. Patient expectations are important and may be influenced by publicity about cochlear implants. Assessing expectations for the ABI and ensuring informed consent prior to implantation are highly important but may be complicated in candidates overwhelmed by a plethora of preoperative concerns. A thorough and frank appraisal preoperatively of the potential benefits, limitations, and requirements for adjusting to the device will help increase the likelihood of a satisfied user in the long run. At this time, the ABI requires a certain level of acceptance, motivation, and commitment from the recipient to maximize benefit; therefore, the device may not be for everyone.

DEVICE

The ABI hardware has evolved through a number of modifications since the original ball electrode was inserted by Drs. William Hitselberger and William House in 1979.^1,5 Significant design changes have involved transitioning from a percutaneous connector to a transcutaneous coil link to the implant, converting from ribbon electrodes to .7-mm-diameter disk electrodes, and fabrication of a semiflexible silicone electrode carrier (2.5 × 8.5 mm) with a specialized mesh backing to stabilize placement. The first 25 ABI recipients were fitted with a single-channel sound processor and a 2-or 3-electrode array. From 1992 until about 2000, the electrode array used in most patients employed eight platinum disks in a perforated silicone and mesh carrier connected to an implantable receiver/stimulator. From about 2000 to the present, the 21 electrode Nucleus ABI24 (see Fig. 58-1) has been used. The external device in the 8 electrode patients consisted of a post-auricular microphone, a transcutaneous transmitter coil, and a sound processor (Spectra Model, Cochlear Corporation). Signal-processing strategies have evolved in an effort to improve performance.⁶ Present patients receive a 21-electrode array (Fig. 58-1A and B) interfaced with the Freedom post-auricular sound processor and ear-level or body-worn controller (Cochlear Corporation, Englewood, CO).

FIGURE 58-1 A, Auditory brainstem implant 24 receiver/stimulator with 21-electrode array and remote ball ground electrode. B, Freedom auditory brainstem implant processor with ear-level controller and transmitter coil, and the body-worn controller for longer battery life.

(A and B, Courtesy of Cochlear Corporation, Englewood, CO.)

Since 2003, use of a hybrid ABI array consisting of a version of the present surface electrode and a 10-electrode penetrating array also has been studied in an FDA clinical trials. Implementation of this system in laboratory animals originally demonstrated the capability for improved microstimulation of auditory neurons⁷ and possibly improved perceptual performance in humans.

ANATOMIC CONSIDERATIONS

The target of the ABI electrodes is the cochlear nucleus complex—dorsal and ventral cochlear nuclei. In humans, the cerebellar peduncle that forms the base of the pons covers the auditory nuclei. This means that the nuclei are not visible to the surgeon and must be located from surface landmarks. Figure 58-2 illustrates the major structures of the pontomedullary junction region with the translabyrinthine approach surgical field of view within the dashed lines. The terminus of the sleeve-like lateral recess forms the foramen of Luschka. Just inferior to the foramen is the root of the glossopharyngeal (ninth) nerve. Superior to the foramen lie the root entry and exit zones of the vestibulocochlear and facial nerves. This area is frequently distorted by the tumor.

FIGURE 58-2 Schematic of cochlear nuclei region demonstrating relative location of various landmarks. Dashed area represents approximate surgical view. Electrode is fully inserted into proper position.

The cochlear nuclei come closest to the surface of the brainstem within the medial and superior aspect of the lateral recess.^8,9 The main target for stimulation is the ventral cochlear nucleus, which forms the main relay for cochlear nerve input and the greater part of the ascending auditory pathway. Placement of the electrodes within the recess has resulted in the fewest side effects and preserves auditory stimulation even though some part of the electrode array lies adjacent to the dorsal cochlear nucleus.² Also, the disadvantage of lack of exposure is partially offset by positional stability provided to the electrode carrier by the limited space in the lateral recess.

SURGICAL CONSIDERATIONS

The surgical approach for tumor removal in ABI cases at HEI has been exclusively via translabyrinthine craniotomy (see Chapter 50). The translabyrinthine route has been found to provide the most direct access to the lateral recess and surface of the cochlear nuclei.¹⁰ Until the actual placement of the device, the surgery proceeds as in any other translabyrinthine acoustic neuroma excision, with the following exceptions. Electrodes are placed for recording electrically evoked auditory brainstem responses (EABRs) and for monitoring cranial nerves VII and IX. The postauricular incision used is shown in Figure 58-3. It is important that the incision not cross the receiver/stimulator package.

FIGURE 58-3 Location of incision with respect to planned site of receiver/stimulator.

Electrophysiologic monitoring is performed during implantation to help ensure that the electrode array placement is correct for activating the auditory system, and also to detect possible activation of nonauditory brainstem structures. There may be considerable uncertainty about the correct position for the electrode array when a large tumor has distorted the anatomic landmarks at the brainstem. To aid in placing the electrode array, EABRs are recorded. A repeatable EABR indicates that stimulation of the auditory system is occurring. Intraoperative EABRs obtained with electrical stimulation of the cochlear nucleus differ considerably from brainstem responses routinely recorded with acoustic stimulation (ABRs) in normal-hearing individuals.¹¹ An experienced electrophysiologist interprets these waveforms at the time of implantation based on data collected from previously implanted patients (Figure 58-4).

FIGURE 58-4 Electrically evoked auditory brainstem responses. CN, cochlear nucleus.

For recording EABRs, subdermal needle electrodes are inserted at the vertex of the head, over the seventh cervical vertebra in the neck, and at the hairline on the neck prior to the draping of the sterile field. After the receiver/stimulator of the implant has been fastened to the skull and the electrode array has been placed in the brainstem, the transmitter coil is placed over the receiver antenna. The stimuli for evoking responses are biphasic current pulses. Scalp-recorded evoked potentials are sampled and averaged by computer following suitable amplification and filtering.

Electrophysiologic monitoring also helps determine the electrode array position that minimizes nonauditory side effects. In addition to monitoring the facial nerve in standard manner,¹² bipolar electrodes are inserted in the ipsilateral pharyngeal (soft palate) muscles to monitor activation of cranial nerve IX. If the electromyographic recordings reveal activation of nonauditory centers during stimulation through the implant, or if a muscle evoked potential is seen in the averaged waveform, the electrode array is repositioned.

IMPLANTATION TECHNIQUE

Tumor dissection proceeds normally via a translabyrinthine craniotomy. After tumor removal and hemostasis, an area of cortical bone posterior to the mastoid is flattened, and a trough to accept the wires from the electrodes to the receiver/stimulator is created in a manner similar to that of cochlear implantation. Using a replica of the receiver/stimulator as a guide, a circular area of bony cortex posterosuperior to the mastoid defect is drilled with cutting burrs (Fig. 58-5). A specially designed butterfly bit or other cylindrical bits associated with high-speed drills may be employed. Using a replica of the receiver/stimulator as a guide, the surgeon drills four holes into the bone to accept the tiedown suture. The receiver/stimulator is fixed with nylon suture prior to electrode array positioning so that the manipulation of the leads does not alter the electrode placement (Fig. 58-6). Because only bipolar cautery may be used after the electrode array is inserted to minimize the risk of current shunting through the device into the brainstem, meticulous hemostasis of the entire wound and CPA is ensured prior to implantation.

FIGURE 58-5 Surgical view of completed translabyrinthine craniotomy, trough for wires, and receiver/stimulator site.

FIGURE 58-6 ABI receiver/stimulator and electrode array in place.

Anatomic landmarks lead the way to the surface of the cochlear nuclei (ABI placement video clip). Normally intact choroid plexus marks the entrance to the lateral recess (foramen of Luschka), and the taenia obliquely traverses the roof of the lateral recess, marking the surface of the ventral cochlear nucleus. These structures may not be clearly visible, however, when a large tumor has significantly distorted the lateral aspect of the pons and medulla. Following the stump of the eighth cranial nerve usually leads to the opening of the lateral recess in these cases. The ninth cranial nerve can also be used as a reference point for the lateral recess. A concavity sometimes visualized between the eighth and ninth nerves should not be confused with the introitus of the recess. The location of the lateral recess may be confirmed by noting the egress of cerebrospinal fluid as the anesthesiologist induces a Valsalva maneuver in the patient. This technique should be reserved as a final check after the opening to the recess has been located by standard landmarks because cerebrospinal fluid will be drained quickly and the advantage of this technique will be lost with multiple Valsalva maneuvers.

After identifying the foramen of Luschka, a Rosen needle is used to insert the electrode array into the lateral recess with the electrodes facing superiorly (Fig. 58-7). With experience, we have found that the system functions better, with fewer side effects, when the entire electrode array is placed just inside the lateral recess.² After placement, selected electrodes in the array are activated to confirm their position over the nucleus. They are tested for the presence of EABRs, stimulation of adjacent cranial nerves (VII and IX), and changes in vital signs. The position of the electrode array usually needs some adjustment to maximize the EABRs and minimize electromyographic responses from the other nerves. If stimulation of the IX nerve occurs, the electrode is separated from it with teflon felt.

FIGURE 58-7 Surgical view of the electrode array placed into the lateral recess (magnified view of the dashed area in Figure 58-6) and cross-section through the brainstem in the region of the electrode array.

The electrode array is secured by a small piece of Teflon felt packed into the meatus of the lateral recess. Fibrous tissue eventually stabilizes the array in position. The wires are positioned in the mastoid cavity and the bony trough previously drilled (Fig. 58-8). Abdominal fat obliterates the mastoid defect. The incision is closed in three layers, and care is taken not to disturb the wires. The wound is not drained routinely. A large mastoid-type dressing is left in place for 4 days.

FIGURE 58-8 Implant, wires, and fat in place prior to skin closure.

POSTOPERATIVE CARE

The postoperative care after implantation shares many of the features of routine translabyrinthine tumor resections (see Chapter 50). A similar schedule for advancing patient activity and decreasing the level of intensity of nursing care is maintained. A mastoid dressing should remain in place for at least 4 days. Careful attention to any moisture on the bandages allows prompt identification of cerebrospinal fluid leak through the postauricular wound. Intravenous antibiotics are administered prophylactically 1 day preoperatively and continued through the fifth day postoperatively.

While we originally attempted testing of ABI recipients’ devices within days after surgery, we no longer do so. Swelling of the skin flap covering the receiver/stimulator may prevent an adequate signal from reaching the implant and preventing device power-up. Instead, the device is typically activated for the first time about 4 to 8 weeks after implantation.

Because the magnet is typically removed from the receiver/stimulator so that patients may continue to have magnetic resonance imaging (MRI), there may be difficulty in identifying the location of the antenna of the receiver/stimulator at the time of initial stimulation. If the antenna cannot be located, the transmitter coil may not be properly positioned at initial stimulation, and it may mistakenly appear that the device is nonfunctional or the patient is nonstimulable. While normally it is possible to palpate the scalp for the location of the receiver antenna, this may not be possible in patients with thicker skin. Therefore, consideration should be given to this potential difficulty at the time of implantation and appropriate steps taken (such as possible thinning of the skin, or by marking the antenna location with a tattoo) prior to the initial stimulation session. In actual use, ABI patients must shave this area and apply a thin tape and metal disk to which the magnetic transmitter coil can adhere. The patient, or a companion, must be trained to ensure proper and consistent positioning of the transmitter coil over the implant receiver antenna.

POSTOPERATIVE COMPLICATIONS

The most significant complication in the immediate postoperative period is cerebrospinal fluid leak. Unlike routine translabyrinthine surgery, in which the fluid usually takes the nasal route via the eustachian tube, the ABI electrode and wires provide a path along which cerebrospinal fluid can travel beneath the skin flap. We have noted a marked reduction in the rate of leak after transitioning over to the fully implantable receiver from a percutaneous connector used with the early single-channel ABI. Prevention of a leak begins with meticulous dural approximation and packing of the eustachian tube and mastoid cavity with various materials. Although the dural opening cannot be closed in a watertight manner, it should be approximated as closely as possible to minimize the opening. A dumbbell-shaped graft of fat will plug the residual space. Muscle and oxidized cellulose (Surgicel) commonly are employed for eustachian tube closure, and autologous fat works well in the mastoid. Multilayered closure for the wound decreases pathways for cerebrospinal fluid egress.

Despite these precautions, patients with the ABI appear more prone to cerebrospinal fluid leak than do those undergoing translabyrinthine procedures without implantation. Leaks from the nose and wound usually respond to reapplication of mastoid pressure dressing and bed rest. A lumbarsubarachnoid cerebrospinal fluid drain is added for persistent leaks. Finally, surgical exploration and repacking of the wound can be employed for leakage unresponsive to more conservative measures.

Meningitis can occur either spontaneously or as a result of postoperative cerebrospinal fluid leak. This unusual complication, when identified promptly, responds to antibiotics and cessation of the leak.

Normal healing to a stable implant situation usually takes 4 to 6 weeks, after which initial activation of the device occurs. Multichannel ABI recipients now have experienced up to 15 years of trouble-free use of the device. The first patient ever to be implanted with an ABI in 1979 continues to use her single-channel implant with benefit on a daily basis.

RESULTS

Over 200 patients with NF-2 have been implanted with the Nucleus multichannel ABI system at HEI between 1992 and the present. Data from the earliest group of implantees comprised a large portion of a clinical trials submission to the FDA.¹³ In this chapter, we present results from speech and environmental sounds testing of a large group of patients experienced in using their ABIs. Since performance with the ABI improves more gradually than in cochlear implants, results from experienced users are more representative of longer term benefit. Early results on many of these patients have been presented elsewhere.⁶ Even 2 years of experience should not be considered sufficient to reach asymptomatic performance. Although improvements are generally greatest during the first year, many patients have continued to improve even after 10 years of use. All patients used the SPEAK (spectral maxima) speech processing strategy.¹⁴

Figure 58-9 shows mean scores on a portion of the multichannel ABI perceptual test battery. The lowest scores shown are for the CID and City University of New York (CUNY) sentence tests, which are presented in sound only. These tests are difficult for the majority of ABI recipients, and they reflect the generally limited capability of the ABI in the absence of lipreading cues. The CUNY sentence scores in lipreading alone and in sound plus lipreading modes are higher than those administered with sound alone because of the visual cues provided. Obviously lipreading cues are highly important in face-to-face communication, and a large average increase (39 per cent) in sentence recognition occurs when ABI sound is added to lipreading as indicated by the lipreading only and sound plus lipreading CUNY scores.

FIGURE 58-9 Mean speech perception test scores of auditory brainstem implant patients with 2 or more years of experience (N = 25). MTS-W, Monosyllable, Trochee, Spondee Word Score; MTS-S, Monosyllable, Trochee, Spondee Stress Score; SERT, Sound Effects Recognition Test; NUCHIPS, Northwestern University Children’s Perception of Speech Test; CID (S), Central Institute for the Deaf Sentence Test in sound only; CUNY (S), CUNY (V), and CUNY (SV), City University of New York Sentence Tests in sound only, vision only, and sound plus vision, respectively.

The Monosyllable, Trochee, Spondee (MTS) test, Sound Effects Recognition Test (SERT), and Northwestern University Children’s Perception of Speech (NUCHIPS) test are all “closed-set” tests in which the individual has to select the correct answer from a limited set of options. Therefore, these tests are somewhat easier than “open-set” (essentially unlimited set) tests, but they nevertheless represent a challenging auditory discrimination task to ABI recipients. In the MTS Word (MTS-W) test and the NUCHIPS test the listener selects a word from a set of alternatives. In the case of NUCHIPS, they are rhyming words. In the SERT, the patient selects the correct sound from a set of four pictured alternatives. The highest mean score on all the perceptual tests occurs on the MTS Stress (MTS-S) test, which is a derived score from the MTS-W test. The patient does not have to correctly identify the word—only the correct stress pattern of the word presented, making it one of the easiest tasks. Tests such as the MTS-S and SERT provide rather immediate evidence to new ABI recipients that the auditory cues they receive from their implants are indeed useful.

With notable exceptions, these results indicate that ABI performance generally has not reached the high levels typically seen with multichannel cochlear implants. At least five patients have shown high levels (≥50 per cent) of open-set speech recognition ability on sound-only sentence tests, and about 16% of our ABI recipients have shown significant (at least 20 per cent correct) ability in this area. There is reason to be hopeful that ABI performance in general will improve. Many ABI recipients experience electrode-specific pitch sensations similar to cochlear implant recipients, and it may be possible to increase these cues with improved microstimulation systems such as the penetrating ABI (PABI). Capitalizing on these cues by carefully assessing auditory percepts from ABI stimulation can be time consuming, but is a necessary part of programming the speech processor to optimize performance.¹³

Perceptual test scores of ABI recipients presently indicate a significant ability to discriminate many environmental sounds as well as enhance sentence recognition ability over lipreading only. Several of the best performers use the telephone with familiar speakers in controlled conditions. Nevertheless, as is true of hearing aids and cochlear implants, the ABI cannot be expected to be highly beneficial for every potential candidate. Because sound from the device is most effective in combination with lipreading cues, patients with limited vision have generally experienced relatively less communication benefit. Severe visual impairment also greatly complicates the process of accurately testing and programming the ABI sound processor, and special techniques and procedures must be used.

Patients with limited social contact may find fewer occasions to use the ABI, which may inhibit their progress with it. Also, we have noted some difficulties with acceptance and use of the ABI by teenagers who may have special cosmetic concerns in addition to problems related to dealing with NF2. Some patients (about 9 per cent in our hands) have not received auditory sensations at all, instead experiencing only mild or moderate nonauditory side effects when their device was activated. Many of these cases were noted to involve anatomic difficulties at the time of implantation. Preoperative MRIs may signal potential problems leading to nonstimulation, such as a large lateral recess or tumor damage to the cochlear nucleus region.

Speech processing for brainstem stimulation has profited from research in cochlear implants. Interestingly, similar strategies used in cochlear implants also have worked well with the brainstem implant.⁶ Flexibility of the ABI programming system has been essential to accommodate anatomic variations and the range of auditory and nonauditory sensations that can result from stimulation. Proper assessment and use of this information in configuring ABI speech processors can have a significant effect on speech perception performance. Poor selection or misalignment of frequency bands to electrode channels in speech processor programming can limit performance.¹³ Particularly in patients with a greater incidence of nonauditory sensations, experience and flexibility in the clinician’s approach can sometimes mean the difference between use and nonuse of a device. Future studies regarding stimulation rates, frequency assignment of channels, and methods of coding speech cues will contribute to improvements in speech processing strategies for the ABI.

OTHER RESEARCH

Niparko and associates¹⁵ demonstrated the feasibility of implanting and stimulating within the substance of the cochlear nucleus in guinea pigs. In a related study, el-Kashlan and colleagues¹⁶ compared the effectiveness of surface electrodes with those placed into the nucleus. They found lower thresholds and a wider dynamic range in animals with penetrating electrodes than in those with surface placement. These results were influential in the development of the penetrating ABI system that employs an array of needle microelectrodes. McCreery and coworkers⁷ demonstrated the efficacy of such a system for activating discrete populations of tonotopically tuned neurons within the substance of the cochlear nucleus. This was achieved without significant risk to tissue or blood supply in longer term preparations with properly constructed and inserted electrodes. The needle-type electrodes with a somewhat blunt-tip configuration (Fig. 58-10) were atraumatically inserted on-axis with a specialized spring-powered tool.

FIGURE 58-10 Schematic illustration (right) of an early penetrating auditory brainstem implant (PABI) electrodearray and implant site (left) accessing tonotopic gradient of fibers in ventral cochlear nucleus (VCN). DCN, dorsal cochlear nucleus. The present PABI uses 10 penetrating microelectrodes with the depth of the stimulating surfaces ranging from 1 to 2.5 mm

(Courtesy of Huntington Medical Research Institute, Pasadena, CA.)

SUMMARY

A multiple-electrode array for electrical stimulation can be safely and reliably placed on the brainstem of patients and chronically stimulated to produce useful auditory sensations by selective activation of the cochlear nucleus. Few side effects and minimal morbidity typically characterize the clinical course of patients with these implants. Speech perception performance varies and, with some exceptions, does not reach the high levels typically seen with modern cochlear implants. In combination with lipreading cues, however, ABI sound has proven to be highly beneficial for the majority of device recipients.

Further improvements in the hardware (including possibly the PABI, AMI and ICI) and in sound processing methods could give patients improved speech understanding in the future.

Acknowledgments

The authors are grateful to Michael Waring for editorial assistance on electrophysiology, Butch Welch for graphics assistance, and the patients of the House Ear Clinic for their time and effort in laboratory testing.