Introduction

Neurofibromatosis type 2-related schwannomatosis (NF2) is usually characterized by bilateral vestibular schwannomas (VS), but also schwannomas on other cranial, spinal, peripheral/cutaneous nerves, meningiomas, and intramedullary ependymomas [4, 22]. NF2 is a heterogeneous disease, with variable tumor burdens and tumor growth rates, resulting in both severe, life-threatening forms, and very mild, late-onset forms characterized by bilateral VS only. However, regardless of NF2 severity, all patients are at risk of severe hearing impairment or even total deafness. Hearing function is difficult to manage, particularly in patients with many comorbid conditions and a low capacity for lip-reading due to ophthalmic features, such as lenticular opacities, retinal hamartoma or epiretinal membranes [4]. Auditory rehabilitation strategies must be considered when planning tumor treatment, as surgery or radiosurgery can worsen hearing function [21]. Cochlear implants are an effective and well-known method of hearing rehabilitation but may not be sufficiently beneficial in NF2 patients, due to a loss of cochlear nerve anatomical or electrical integrity linked to VS growth or surgery. In such cases, direct electric stimulation of the cochlear nucleus in the brainstem via an auditory brainstem implant (ABI) is a possible solution [3, 18]. Following ABI implantation, most patients recognize environmental sounds well and have some improvement in lip-reading skills, but only a few achieve significant speech perception with the ABI alone [27].

In recent years, the indications for ABIs have decreased due to the emergence of bevacizumab as an alternative treatment for VS that is growing or that compromises hearing, regardless of tumor size [11, 24] and due to the intraoperative use of electrically evoked auditory brainstem response audiometry to test cochlear nerve function after VS resection for the selection of suitable candidates for cochlear implantation [5, 10]. It is therefore difficult to determine whether there is still some place for ABIs in the treatment of NF2 patients, and to identify the patients most likely to benefit from such treatment or to suffer nonauditory adverse effects.

The aim of this study was to assess long-term (after at least five years of follow-up) functional results in NF2 patients fitted with ABIs and to investigate prognostic factors of ABI use.

Methods

Patient population

A monocentric retrospective cohort study was performed at the national reference center for the management of NF2 and implantation. The medical charts of 53 NF2 patients who underwent ABI surgery from September 1997 to December 2022 were reviewed (Fig. 1 and 2). Patients for whom the medical file or hearing outcomes were untraceable or with less than five years of follow-up were excluded from the study (Fig. 2). In total, 27 patients (32 ABIs) were included, implanted from September 1997 to February 2016.

Fig. 1
figure 1

Number of ABI implantations at our center from 1997 to 2022

Full size image
Fig. 2
figure 2

Flow chart for ABI inclusion. Abbreviations: ABI, auditory brainstem implant; CPA, cerebellopontine angle; NF2, neurofibromatosis type 2-related schwannomatosis

Full size image

For each patient, the data collected included demographic characteristics, NF2 genetic severity score [9] and clinical symptoms (age at diagnosis, other tumors, swallowing problems and facial nerve function), initial characteristics of the VS on the ABI side, surgical approach, adjuvant treatment for VS (surgery, radiosurgery, bevacizumab), sizes of the ipsi- and contralateral tumors in the cerebellopontine angle (CPA) after surgery.

The surgical indications for ABIs in these patients were discussed by a multidisciplinary team composed of otologists, neurosurgeons, audiologists, radiologists, speech therapists and psychologists from the national reference center for this rare disease. The VS was excised and the ABI inserted via a retrosigmoid, transotic or translabyrinthine approach, depending on the extent of the tumor and changes in practice over time [6, 12]. A 22-channel implant (Nucleus 22, Cochlear Inc., Lane Cove, Australia) was used from 1996 to 2011, and a 12-channel implant (Synchrony, Med-El, Innsbruck, Austria) was used from 2003 to 2016. The implant was inserted into the lateral recess of the fourth ventricle in contact with the cochlear nucleus. Intraoperative electrically evoked auditory brainstem responses were used to control the position of the electrode array. The ABI was inserted either immediately after VS removal (one-step surgery) or during a second procedure (two-step surgery).

All ABI devices were activated six weeks after surgery and evaluated every three months during the first year, and annually thereafter [6]. Electrodes inducing nonauditory sensations were deactivated. The number of electrodes activated was noted at the time of activation and at each check-up.

Auditory performance

Auditory performance was classified as: “ABI-use” or “ABI non-use”, according to whether or not the patients were using regularly their implants during follow-up.

In cases of ABI-use, performance was evaluated in quiet conditions, with disyllabic words (Fournier list) and sentences (Marginal Benefit for Acoustic Amplification, MBAA) at 3, 6, and 12 months after implantation, and annually thereafter. The results are expressed as the percentage of correct words for disyllabic lists, and as the percentage of correct sentences for the MBAA sentences. All tests were administered in three modes: ABI only, lip-reading only, and ABI + lip-reading. We classified performance as follows: awareness of sound (no improvement of lip-reading performance with ABI), lip-reading improvement (improvement of lip-reading performance of at least 20% with use of the ABI) or intelligibility (intelligibility of at least 40% with the ABI alone, without lip-reading). In cases of ABI use with at least an improvement in lip-reading, hearing improvement was calculated as \(\Delta\) ABI = “ABI + lip-reading” – “lip-reading” scores, one and five years post-implantation, and at last follow-up.

Tumor size and progression

The VS was classified into four stages (1: intracanalicular; 2: < 15 mm; 3: < 30 mm; 4: ≥ 30 mm [1]) according to the initial size measured on gadolinium-enhanced magnetic resonance images (MRI) by the neuro-otologists. Tumor size for stages 2 to 4 was measured as the maximal extrameatal diameter. After implantation, MRI follow-up was not possible for most patients, mostly due to the magnet-related artifact. CT scans were therefore performed annually to assess the sizes of ipsi- and contralateral tumors.

Statistical analysis

Statistical analysis was performed with GraphPad Prism® (version 8.2.0, San Diego, CA, USA) and R 3.0.2 statistical software (www.r-project.org). Results are expressed as the mean ± standard deviation (SD) [range]. Wilcoxon tests were used to compare auditory performance. Logistic regression was used to identify potential prognostic factors for ABI-use or non-use one and five years post-implantation. A p value less than 0.05 was considered significant.

Ethical statement

The patients or both their parents (for minors) gave written informed consent for the use of medical records. The study was registered with the French Data Protection Authority (Commission nationale de l'informatique et des libertés # 2,040,854).

Results

Population

The demographic characteristics of the 27 patients included are shown in Table 1. Five patients underwent sequential bilateral implantation; 22 patients had bilateral profound hearing loss, and five patients had profound hearing loss and a contralateral conventional hearing aid for severe to profound hearing loss. During follow-up (13 ± 5.8 years), 16 (59%) patients suffered a decrease in visual acuity, 24 (89%) had unilateral facial palsy (ipsi- or contralateral), and 8 (29%) had a swallowing disorder. Seven (26%) patients died, after 12 ± 3.6 [9–16] years of follow-up. None of the patients had a cochlear implant.

Table 1 General characteristics of the population
Full size table

For each ABI (n = 32), the general characteristics, ABI surgery data, neoadjuvant, and adjuvant treatments for the ipsilateral VS are provided in Table 2. Implantation was performed at the age of 32 ± 14.3 [11–64] years, and mean follow-up was 12 ± 5.6 [5–24] years.

Table 2 General characteristics of the ABIs (n = 32, 27 patients) for total population
Full size table

ABI-related complications

The ABI was displaced in two cases, probably due to the growth of the contralateral VS (extrameatal sizes of 40 and 41 mm), necessitating replacement surgery one and five years after initial implantation. The auditory performance of these patients was unchanged following implant replacement, with “lip-reading improvement” throughout the follow-up period (5 and 8 years). The implant of one patient was removed one year after implantation due to skin necrosis around the magnet, with reimplantation four years later, leading to stable “lip-reading improvement”.

ABI use and hearing outcomes

ABI use during follow-up

One year after implantation, 20 (74%) patients were users of at least one ABI (for patients with bilateral implants): 5 patients (18.5%) had good speech intelligibility with the ABI alone, 10 (37%) displayed an improvement of lip-reading, and 5 (18.5%) were able to detect sounds (Fig. 3). Considering each ABI separately, 21 (66%) were used: 6 (19%) enabled speech intelligibility with the ABI alone, 10 (31%) facilitated improvements in lip-reading, and 5 (16%) improved sound detection (Fig. 4). For ABIs for which both speech intelligibility and lip-reading improved, lip-reading was improved by the ABI (Δ ABI) in disyllabic word tests (+ 32 ± 27.3, p = 0.002, Wilcoxon test) or MBAA sentence tests (+ 28 ± 30.4, p = 0.002, Wilcoxon test, Table 3).

Table 3 Hearing performance one and five years post-implantation and at last follow-up (14 ± 6.9 years), for ABI-use, with lip-reading improvement and speech intelligibility
Full size table
Fig. 3
figure 3

For the 27 patients, changes in hearing performance status one and five years post-implantation and at last follow-up (14±5.2 [7–24] years). Patients with bilateral implants were sorted according to the ABI with the highest category

Full size image
Fig. 4
figure 4

For the 32 ABIs, change in hearing performance one and five years post-implantation and at last follow-up (14±5.2 [7–24] years)

Full size image

Five years after implantation, two of the patients initially classified as users were no longer using their implants. In total, 18 (67%) patients were users (of at least one ABI in patients with bilateral implants), and 19 (59%) ABIs were used (Fig. 3 and 4). Hearing performance remained stable between the one and five years post-implantation for both disyllabic word tests (p = 0.6, Wilcoxon test) and MBAA sentence tests (p = 0.2, Wilcoxon test, Table 3).

Follow-up did not exceed five years for three users patients, with improvements in speech intelligibility for one and of lip-reading for the other two. For the other 24 patients, at last follow-up (14 ± 5.2 years), one had the implant removed, one had become a non-user, and 13 (54%) were still using their ABIs (Fig. 3). Considering each ABI separately, 13 (46%) implants were being used at last follow-up (Fig. 4). For ABIs resulting in improvements of both lip-reading and speech intelligibility, hearing performance decreased only slightly, by 12 ± 14.7% in disyllabic word tests (n = 10, p = 0.05, Wilcoxon test), and by 15 ± 23.8% in MBAA sentence tests (n = 10, p = 0.08, Wilcoxon test, Table 3) between five years post-implantation and last evaluation.

Changes in hearing performance with ABIs according to tumor size

For the ABIs that were used and yielded improvements in speech intelligibility and lip-reading one year post-implantation (n = 16, Fig. 5), hearing performance decreased in four cases and was stable in 12. The presence of a large, growing, residual ipsilateral tumor appeared to be related to a decrease in performance.

Fig. 5
figure 5

Change in auditory performance of the ABI users (ABI + lip-reading performance, %) for intelligibility and lip-reading improvement one and five years post-implantation (n = 16), with declining (a) or stable (b) performance

Full size image

Three of the four cases of decreasing performance were associated with the growth of the residual ipsilateral tumor, which extended 30 mm or more into the CPA (Fig. 5A). The decreases in performance for these ABIs were detected 1 (O.A.), 6 (S.L.) and 10 (G.I.) years after implantation. Two patients underwent surgical excision of the residual tumor (S.L. and G.I.) with no recovery of auditory performance. The third patient (O.A.) received bevacizumab to treat ipsi- and contralateral tumor growth. For the fourth ABI for which a decrease in performance was registered (C.B., right side), the benefits of the implant were rapidly lost, for “intelligibility” at one year and “sound detection” at five years, despite the absence of a residual tumor, and non-use of this implant was recorded at last follow-up (16 years post-implantation). The reasons for the declining performance of this ABI remain unknown. However, it should be noted that the contralateral ABI of this patient was not used one year post-implantation.

In 12 cases, auditory performance remained stable: 11 (92%) patients had no ipsilateral residual tumor, and the remaining patient (8%) had an ipsilateral residual tumor measuring 25 mm (M.L.) (Fig. 5B). Three of these patients had a large, growing, residual contralateral tumor extending 30 mm or more into the CPA. Two were treated with bevacizumab (4 and 11 years after implantation, K.B.Y. and B.C.), and one underwent surgical removal of the implant five years after implantation (M.L.) with stable performance over time.

ABI non-use

One year post-implantation, 7 (26%) patients were non-users of the implant (Fig. 3). Four patients with bilateral implantation used only one of the implants, with 11/32 (34%) ABIs non-used.

Five years post-implantation, two of the patients classified as users one year after implantation were no longer using their implants (Fig. 3). One patient had stopped using his ABI five years after implantation, but began using it again after the birth of his child, resulting in renewed lip-reading improvement. The other had “lip-reading improvement” and the loss of hearing benefit coincided with the growth of a large ipsilateral residual tumor (see Fig. 5A, O.A.).

At last follow-up, four of the ABIs that had previously been used were no longer used (Fig. 4). Two ABIs were used only for sound detection five years after implantation, and were no longer used after nine and 15 years. For the other two ABIs, the cessation of use was related to growth of the residual ipsilateral tumor (see Fig. 5A, S.L. and G.I.).

In this study, only one patient (with bilateral implants) underwent explantation, 11 and 15 years post-implantation, to facilitate MRI follow-up: she did not use either of her ABIs (Fig. 4).

Prognostic factors of ABI use

Univariate analysis was performed to identify prognostic factors for ABI use one and five years post-implantation. No differences in general characteristics (age at implantation or at diagnosis of e NF2, duration of deafness) or initial tumor characteristics were found between users and non-users of ABIs. Surprisingly, the number of electrodes activated at ABI activation was also similar in the two groups. We observed no differences between patients with and without VS excision before ABI insertion, or between patients with neoadjuvant or adjuvant treatment (radiosurgery, bevacizumab). However, all but one of the five patients with a useful contralateral ear (with conventional hearing aid) did not use the ABI (p = 0.06, Chi2 test).

Discussion

NF2 management has undergone a profound transformation in recent years, with the widespread use of bevacizumab, limiting the need for surgery on the second side and preserving hearing. The population of patients has continued to grow, but with less severe presentations following the introduction of bevacizumab. This shift in treatment strategy has resulted in a transition from the predominant reliance on ABI placed during the second side VS removal to the use of CIs. However, ABIs remain valuable for the treatment of specific NF2 patients. This is the first study to evaluate auditory performance in the long term after ABI insertion, and changes in auditory performance as a function of ipsi- and contralateral tumor progression. We found that 59% of ABIs were used five years after implantation. Performance remained stable over time for all but four of the ABIs, three of which were associated with growth of the residual ipsilateral tumor.

We chose to study patients with at least five years of follow-up as Taslimi et al. reported a significant improvement in hearing outcomes over time, particularly five years post-implantation [26]. The mean duration of follow-up in our study was 12 [5–24] years, for 32 ABIs. This is one of the longest follow-up periods reported (mean follow-up: 1–7 years in previous studies [2, 13,14,15, 18, 19]), adding credibility to our analysis of changes in auditory performance over the years.

Despite the long duration of follow-up after ABI placement in this series of NF2 patients, this study is subject to a main limitation: the retrospective nature of the study and the number of missing data, with only 32 ABIs of the 60 implanted at our center included in this study. The chart could be outdated, with insufficient data, or some patients did not return regularly to our reference center. Furthermore, some important data could be missing, such as the number of active electrodes at each session.

At its most basic level of functioning, the ABI makes it possible to detect sounds [17]. Performance declines with the complexity of the tests. In this study, five years post-implantation, only 12.5% of the ABIs were able to render speech intelligible when used on their own. Nevertheless, about half the ABIs were used in the long term and, in most cases, they yielded a mean additional increase in speech intelligibility of 28–42% relative to lip-reading alone. Our functional results are similar to those for other reported series, with 40–70% of patients claiming to use their ABIs [2, 6, 15, 16, 18, 25]. Nevertheless, patients should be warned that they will probably be unable to rely on the implant alone, but with the ABI enhancing their lip-reading performance to help them follow basic conversation [17].

Both ABIs and CIs are used for auditory rehabilitation but direct comparisons of these two types of implant remain challenging due to differences in the technologies used and different clinical applications within the NF2 population [2, 7]. The tonotopic organization of the cochlear nucleus is variable and displays a more intricate pattern than the cochlea. It progresses from depth to the surface and may be influenced by changes in the size of CPA tumors and surgical interventions, with probable effects on tonotopic mapping. Unlike the slender and flexible electrodes of CIs, which adapt to the curved arrangement of frequency detection along the tonotopic axis of the cochlea, ABI electrodes are inserted into an inflexible silastic paddle, and are placed along the highly curved surface of the cochlear nucleus [8]. Advances in CI electrode-arrays, becoming increasingly thin, have been made, whereas the basic design of the ABI multichannel array has remained largely unchanged for several decades [28].

The factors potentially predictive of ABI use remain unclear, probably due to the complexity of NF2 disease. Decisions concerning the implantation of ABIs are therefore often highly personalized. A few prognostic factors have been identified in other studies but have not been standardized. The initial size of the ipsilateral tumor has been identified as a possible prognosis factor, as it can affect anatomic landmarks leading to the lateral recess and also result in distorted brainstem after tumor removal, potentially causing ABI migration. However, this factor is probably no longer prognostic as, with changes in treatment, all candidates for ABI implantation now have large tumors, as reported in this study. A large number of active electrodes and a short duration of total loss have also been found to be correlated with auditory performance [6]. In our study, the only factor influencing ABI performance was the presence of a useful contralateral ear. These findings strongly suggest that efforts should be made to avoid “sleeper” ABIs [6, 23], most of which (4/5) were never used.

Finally, the management of NF2 patients with ABIs should also take into account tumor burden and genetic subtype. Tumor size, location, and genetic factors, such as NF2 gene mutations, guide treatment decisions and affect long-term outcomes [9, 11, 20].

Conclusion

There are still indications for ABIs for hearing rehabilitation in NF2 patients: bilateral deafness with a non-functional cochlear nerve, ruling out the use of a CI. It is difficult to identify the candidates most likely to use the implant, as no clear prognostic factor for ABI use was identified. Most of the ABIs used made it possible to improve lip-reading, with performance remaining stable over time. Ipsilateral tumors should be controlled pharmacologically (bevacizumab) or by radiosurgery/radiotherapy to prevent declining performance.

Authors'contributions

I.M., E.A.D, A.L.C, M.S., Y.N., O.S., G.L. and M.K. participated to the patient inclusion.

I.M., O.S. and M.K. did the study conception and design.

H.D. did the data collection.

H.D., R.T., I.M., O.S., M.K. and E.F. analyzed data.

H.D. and R.T. did the statistical analysis.

H.D. did the radiological acquisition.

H.D., R.T., I.M. and E.F. wrote the main manuscript.

R.T., I.M., E.A.D, A.L.C, M.S., Y.N., O.S., G.L. and M.K. reviewed the manuscript.

All authors gave their final approval and agreed with the work.