Longitudinal Performance of Cochlear Implants in Neurofibromatosis Type 2

Abstract

Objective:

Cochlear implants (CI) are a well-established treatment modality for hearing loss due to neurofibromatosis type 2 (NF2). Our aim is to investigate variables that affect longitudinal performance of CIs among patients with NF2.

Study Design:

Retrospective review at a single academic institution consisting of patients who have received cochlear implants following hearing loss due to NF2.

Methods:

The primary outcome examined was CI disuse or explantation. Associated clinical and surgical variables were analyzed using descriptive statistics. These included post-operative pure tone average (PTA) at 500, 1000, and 2000 Hz, tumor size, previous surgery, and comorbid depression.

Results:

A total of 12 patients and 14 cochlear implants received at our institution from 2001 to 2022 were included. 35.7% of CIs (5 out of 14 cases) resulted in disuse or explantation. The average interval until explant was 9.4 years (range 3–14 years). In explanted CI cases, 20% had previous surgery and 80% had a diagnosis of comorbid depression as compared to 22.2% and 22.2% respectively in intact CI cases. Maximum tumor diameter was the only variable found to impact CI usage outcome (p=0.028). Long-term data showed that on average, patients benefit from13.85 years of CI utility and a maximum PTA improvement of 45.0 ± 29.0 dB.

Conclusion:

Despite the recurrent nature of NF2, patients continue to receive audiological benefit from cochlear implants. We found that larger tumor size may be associated with longitudinal CI failure.

Summary:

We found that larger tumor size is significantly correlated with the odds of becoming a non-user of a cochlear implant. While the failure of implants and profound hearing loss are frequent in Neurofibromatosis 2, patients may still derive long-term successful usage of the device.

Introduction

Neurofibromatosis type 2 (NF2) is a rare autosomal dominant condition that results in numerous benign tumors of the nervous system. Schwannomas along the vestibulocochlear nerve are the most common tumor in NF2, which also manifests as other central and peripheral nervous system tumors, peripheral neuropathy, and ocular pathology.¹ While benign, the unpredictable course of vestibular schwannomas (VS) in NF2 patients can cause compressive neural loss and cochlear dysfunction which have devastating consequences for hearing loss, tinnitus, and imbalance.² Patients often begin experiencing symptoms at an early age, and 90–95% eventually suffer from progressive bilateral VS and total deafness.³ As two of the most impactful factors for quality of life in NF2 patients, hearing ability and anatomical preservation warrant attention within care management strategies for these patients.⁴

Management of hearing loss in NF2 patients has traditionally involved hearing aids and auditory brain implants (ABI). Concurrent treatment options include VS resection, stereotactic radiosurgery (SRS), and medications such Bevacizumab (Avastin). With improvements in microsurgical techniques, cochlear implantation (CI) even after VS tumor removal became possible and has been regarded as another hearing preservation strategy for NF2 patients since 1992.^5–7 However, NF2 cases and reports of CI in this patient group remain scarce which leaves many uncertainties about the standard of care for VS and hearing management. As of 2021, only 132 unique cases of CI in NF2 existed in literature.⁸

While prior literature has focused on short-term CI performance, there is little data on longitudinal outcomes within this patient population. Anecdotally, patient disuse or device failure has often been reported within the first decade of implant, likely due to the high rates of schwannoma regrowth. However, more quantitative data is needed to help guide treatment decisions and establish patient expectations. Our aim is to investigate the longitudinal utilization of CIs in NF2-affected ears, and examine patient and disease factors associated with utility versus failure in these patients.

Materials and Methods

This study was approved by the Johns Hopkins Institutional Review Board. A retrospective cohort study was conducted on all NF2 patients who received CI at our tertiary care center between 2001 and 2022. A review of electronic medical records was performed to collect relevant clinical, imaging, and demographic data. The primary outcome examined was CI disuse which was defined as reported disuse from the patient or explantation of the CI device or the magnet.

The assessed clinical variables were maximum tumor diameter, history of tumor resection or stereotactic radiosurgery (SRS), medications, implant model, and electrode array type. Radiological assessments were reviewed chronologically to obtain VS tumor(s) location and maximum diameter. The medications recorded were chemotherapy drugs, brigatinib, and bevacizumab (Avastin). Post-operative pure tone average (PTA) was also collected through audiological testing and measured in decibel hearing level (dB HL) at 500, 1000, and 2000 Hz, with substitution of the 2000–4000 Hz average if necessary.³ Demographic data collected were sex, race, history of depression, age at implant, and age at explantation or disuse. All demographic data were collected from the patient medical record. Depression was diagnosed according to DSM-5 guidelines.

Non-parametric Mann-Whitney U and Fischer’s exact tests were used to perform comparison testing for continuous and categorical variables respectively. Pearson’s correlation and regression analyses were used to examine the relationship between variables and CI disuse. Kaplan-Meier survival curve modeling was performed to display aggregate failure rates of CIs and losses to follow-up. Patients were right-censored at the time of loss to follow-up. Statistical analysis was performed on STATA version 14.0 (Stata Crop LP, College Station, TX, USA) and a significance level of p<0.05 was use for all analyses.

Results

Demographic and Clinical Findings

Twelve NF2 patients received fourteen CIs (ten unilateral implant and two bilateral implant patients) at the Johns Hopkins Hospital from 2001 to 2022 (Table 1). Four male and eight female patients were included in the case series. No intraoperative or postoperative complications occurred in any of the patients. All patients had history of VS with recorded mass(es) in the cerebellopontine angle (CPA), internal auditory canal (IAC), adjacent structures, or a combination of locations.

Table 1:

NF2 case series 2001–2022; baseline characteristics and outcome details

Case	Sex	Race	Age at CI	Size of VS (diameter mm)	Location of VS	Management of VS	Year of CI; device	Electrode array type	Baseline PTA (dB)	Best PTA (dB)	Max PTA Improvement (dB)	User vs. Non-user; Explant Date
1	F	White	18	60	CPA with jugular foramen extension	SRS	2007; Medel	Straight LW	90	26.67	63.33	Non-user; 5/4/2021
2	F	White	40	17	IAC	SRS	2003; Medel	Straight LW	60	42.5	17.5	User
3	F	White	57	27	CPA with mininal IAC extension	SRS	2020; Cochlear	Pre-curved MH	95.83	20	75.83	User
4	M	White	50	20	IAC with extension toward Meckel’s cave	Resection	2001; Cochlear	Pre-curved MH	NA	18.33	NA	Non-user; 12/17/2015
5	F	White	31	35	CPA with IAC extension	Avastin	2011; Medel	Straight LW	75	40	35	Non-user; 9/12/2014
6	F	Black	36	30	CPA with IAC extension	Observation	2021; Cochlear	Pre-curved MH	73.3	67.5	5.8	User
7	M	White	15	16	IAC	Resection, Avastin	2002; Cochlear	Pre-curved MH	NA	18.33	NA	User
8	F	Black	39	12	CPA with IAC extension	SRS, Avastin	2018; Medel	Straight LW	61	40	21	User
9	M	Black	26	14	IAC	Resection, Avastin	2009; Cochlear	Pre-curved MH	NA	30	NA	User
10	F	Black	34	7	IAC with extension into basal turn of cochlea	Observation	2022; Medel	Straight LW	103.33	16.67	86.66	User
11	F	Black	32	9	IAC	Observation	2020; Medel	Straight LW	116.67	85	31.67	User
12	M	Black	41	17	IAC	Avastin	2020; Bionics	Pre-curved MH	105	36.67	68.33	User
13	F	Black	32	25	CPA with IAC extension	Resection, SRS	2008; Cochlear	Pre-curved MH	NA	30	NA	Non-user; 6/21/2013
14	F	Black	26	22	CPA with IAC extension	SRS	2002; Cochlear	Pre-curved MH	NA	25	NA	Non-user; 6/21/2013

Variables to Performance

Demographic variables

Sex, race, and depression diagnosis were not found to be significantly associated with CI failure or non-usage (p=1.00, p=0.580, and p=0.091 respectively) (Table 2). Non-users of CI underwent cochlear implant surgery earlier than users at 31.9 ± 11.8 years compared to 36.1 ± 11.3 years respectively, although the difference was not statistically significant (p=0.26). Non-users benefitted from a mean of 9.3 ± 5.3 years of CI utility before disuse. The mean age at explant surgery or reported disuse was 41.2 ± 13.6 years.

Table 2:

Demographic Variables

	Users (n=9)	Non-users (n=5)	p-value
Sex			1
male	3 (33.3)	1 (20.0)
female	6 (66.6)	4 (80.0)
Race			0.58
white	3 (33.3)	3 (60.0)
black	6 (66.6)	2 (40.0)
Depression	2 (22.2)	4 (80.0)	0.091
Age at implant	36.1 ± 11.3 years	31.9 ± 11.8 years	0.257
Age at explant/disuse	NA	41.2 ± 13.6 years
Implant to explant time	NA	9.3 ± 5.3 years

Tumor size

The outcome of CI explant or disuse was associated with a larger VS tumor size (Table 3). The average maximum VS diameter of users were smaller at an average of 16.6 ± 7.6 mm compared to non-users whose tumors averaged 32.4 ± 16.5 mm (p=0.028). Categorizing maximum tumor size into larger or smaller than 10 mm did not result in significant differences between the usage cohorts. Linear regression analysis was performed to determine the correlation between age at implant and tumor size with CI usage. A weak negative correlation between increasing implant age and tumor size was noted (r² = −0.23), but this was not statistically significant (p = 0.43) (Figure 1). Univariate and multivariate analyses ultimately did not show a significant relationship between age at implant, tumor size, and CI failure (Table 4).

Table 3:

Clinical and device variables to CI performance

	Users (n=9)	Non-users (n=5)	p-value
Maximum tumor diameter	16.6 ± 7.6 mm	32.4 ± 16.5 mm	0.028
Threshold diameter, n (%)			0.505
<10mm	2 (22.2)	0 (0.0)
>10mm	7 (77.8)	5 (100.0)
Past surgery/SRS, n (%)	5 (55.6)	4 (80.0)	0.580
Implant model, n (%)			1
Cochlear	4 (44.4)	3 (60.0)
Medel	4 (44.4)	2 (40.0)
Bionics	1 (11.1)	0 (0.0)
Electrode array, n (%)			1
Straight/Lateral wall	4 (44.4)	2 (40.0)
Pre-curved/Modiolar hugging	5 (55.6)	3 (60.0)
Medications, n (%)
Avastin (bevacizumab)	4 (44.4)	1 (20.0)	0.580
other therapies	0 (0.0)	1 (20.0)	0.357

Figure 1:

Pearson Correlation; Age at implant versus maximum tumor diameter

Table 4:

Linear regression analyses

	Univariate Analysis		Multivariate Analysis
	OR (95% CI)	p-value	OR (95% CI)	p-value
Age at implant	0.96 (0.87–1.07)	0.495	0.97 (0.86–1.09)	0.599
Tumor Size	1.19 (0.97–1.46)	0.094	1.20 (0.97–1.50)	0.097

Other variables

The fourteen CIs were categorized into three different implant device models, as well as two electrode array patterns (Table 3). Neither device variable was found to be associated with cochlear implant performance (p=1.0 for both). Nine ears were found to have previous surgical history, with four VS surgical resections and six with radiosurgery treatment. One ear received both surgery and radiosurgery and became a non-user at 14.4 years since implant (Table 1). History of past surgery and Avastin treatment were not found to be predictors of CI performance (p=0.58 and p=0.58)

Audiometric Outcomes

Short-term PTA scores are reported for nine patients with audiologic data available before and after CI (Figure 2). Short-term was defined as spanning one year pre- and post-implant. Two patients were lost to follow-up within the first year after implant surgery. Loss to follow-up was defined as an absence of audiometric measurements for longer than twelve months. As patients undergo follow-up every six to twelve months, the endpoint of a trendline indicates last known follow-up appointment but does not necessarily indicate loss to follow-up. At baseline, the mean PTA was 86.7 ± 20.3 dB, falling within the severe hearing loss category (70 to 90 dB). Available short-term PTA trends across all nine patients show that in the short-term, hearing thresholds remain stable or decrease which signify effective hearing and CI utility. PTA scores for all CIs over a timespan of twenty years is shown in Figure 3. Patients were followed-up for a mean of 5.5 ± 6.4 years. After undergoing CI, patients benefitted from a mean maximum improvement of 45.0 ± 29.0 dB. There were no statistically significant differences between user and non-user ears for all PTA score measurements.

Figure 2:

Short-term PTA score trends across the baseline. As patients undergo follow-up every six to twelve months, the endpoint of a trendline indicates last known follow-up appointment but does not necessarily indicate loss to follow-up.

Figure 3:

Long-term PTA scores of all patients

Long-term Device Use

A Kaplan-Meier survival curve is shown for the outcome of CI explant or disuse (Figure 4). Six patients (first number denoted on the curve) were right-censored before the time of the first explant at 2.71 years; of these, two patients were lost to follow-up and four patients had not yet had their CI for longer than this time interval. Five CIs in four patients underwent CI or magnet explant surgery or reported disuse by the time of data collection. The total time at risk was 88.35 years and the median survival time was 13.85 years.

Figure 4:

Twenty-year survival analysis of CI usage; numbers denoted on the curve represent patients lost to follow-up and/or right-censored.

Discussion

The hallmark presentation of Neurofibromatosis type 2 is bilateral vestibular schwannomas and, as a result, progressive hearing loss. Traditionally, cochlear implants have not been used in NF2 patients due to both numerous, recurring tumors as part of the disease’s clinical course and the implant’s incompatibility with MRIs which are especially necessary for tumor surveillance in these patients. Owing to advances in microsurgery and imaging techniques, cochlear implants are becoming a well-established treatment modality for NF2-related sensorineural hearing loss. Rising interest in CI application in NF2 is compounded by poor speech recognition outcomes found in patients with ABI.^9,10 However, CI outcomes in NF2 patients have remained difficult to study due to low case numbers and the challenge of long-term follow-up.

Of the few existing studies on CI outcomes in NF2 patients, some concomitant findings are associations between larger tumors and history of previous tumor resections with worse hearing outcomes.^8,11,12 In ears with tumors ranging from 1 to 10 mm in size, 78% achieved high performance in open-set speech recognition tasks while only 39% of those with 21 to 30 mm tumors achieved the same score.⁸ On the other hand, authors have reported conflicting results on the impact of preoperative contralateral hearing scores and radiosurgery history on hearing ability. For tumors larger than 15 mm, Deep et. al observed better outcomes in VS treated with radiosurgery or observation over surgical resection.¹³ A different study found greater usage trends in NF2 patients treated with concurrent tumor resection.^3,14 A common finding is that patients who undergo SRS before CI require more current to provide auditory input, perhaps due to radiation-associated neural tissue damage.^15,16 Factors often found to be irrelevant are length of deafness in the ipsilateral ear and other demographic characteristics such as age. One notable problem in existing literature is the high variability in hearing measurements which include PTA, consonant-nucleus-consonant, word recognition, and open-set speech recognition scores which complicates a standardized comparison of cases.^8,17

In our study, the only variable found to impact CI usage outcome was VS tumor size. Tumor size was captured as lifetime maximum tumor diameter throughout the full timeline of existing electronic medical records for each patient. Therefore, it was measured as an independent variable separate from resection surgery which was not found to significantly impact maximum tumor size or the primary outcome. The average maximum VS diameter of non-users were significantly larger at 32.4 ± 16.5 mm compared to 16.6 ± 7.6 mm in CI users (p=0.028). Sorting tumor size into <10 mm and >10 mm did not result in significant differences between the usage cohorts (Table 3). A Pearson correlation test and regression analyses were performed for age at implant and maximum tumor size, based on the rationale that more aggressive phenotypic penetrance of NF2 may result in larger, faster growing tumors and thus a need for CI at a younger age (Figure 1 and Table 4). The results were found to be non-significant which confirmed tumor size to be a true independent variable inversely correlated with CI utility in non-parametric analyses. At this time, there is not enough data to conclude whether other related variables such as tumor decompression surgery and auditory nerve preservation are also associated with favorable CI outcomes.

We opted to use CI disuse/explantation as the primary outcome due to the small study size and inconsistent availability of other objective measures such as consonant-nucleus-consonant (CNC) or AzBio within our electronic medical records system. The lack of speech recognition data leaves room for future studies and is addressed with the rest of our study’s limitations. Of the three patients with reported AzBio scores, one patient underwent explantation while the two patients who did not maintained consistently functional scores across the study period. Our utilization of binary outcome values for CI use is not as robust as speech recognition measurements but offers better reliability across our current cohorts.

Our longitudinal PTA data shows successful, long-term retention of normal hearing capacity (−10 to 25 dB) in some patients, suggesting a beneficial role for CI in a subset of future NF2 patients. Figure 3 shows excellent long term hearing scores for a proportion of implanted NF2 patients (n=4 at 10 years). The Kaplan-Meier curve for trends in CI disuse time does not show a meaningful temporal threshold but indicates a gradual loss of CI use across 20 years. At ten years, over 50% of CIs were still in use (Figure 4). Thus, while eventual failure of CI in NF2 patients is not uncommon, we have shown that patients may still derive an average of 13.85 years of CI benefit before experiencing a decrease in device function. These results are promising for managing progressing hearing loss in the disease course of NF2 which may be inherently unpredictable but eventually affects nearly all patients.

Some limitations to our study are a small sample size of twelve patients and fourteen CI cases and losses to follow-up. Approximately 130 new and existing NF2 patients are managed at our institution per year, of whom twelve patients over the study period of 22 years received CIs. As mentioned above, the conservative utilization of CIs in the management of NF2 may have been due to incompatibility of older implant models with MRI, patient history of prior surgery and concern for functional acoustic nerve, and clinician concern for the outcomes and longevity of CIs. While all studies are limited by the low incidence rate of NF2, our cohort is comparable to previously reported cohorts from a single institution. There were also non-negligible losses to follow-up; two of the twelve patients were lost during the first follow-up year, with three more lost in subsequent years. While this makes the generalizability of our findings more difficult, these individuals were right-censored in statistical analyses and the available data still allowed for meaningful conclusions to be drawn through short-term and long-term analyses. One reason for patient attrition, as well as inconsistently available speech recognition scores, may be the progressively debilitating nature of the disease which is associated with both a higher burden of competing priorities for medical management and overall difficulty with completing timely audiologic follow-ups. Another reason may be our institution’s status as a specialty center treating patients from a wide geographical breadth which could pose difficulties for long-term follow-up. Future studies will be needed for a better understanding of which patients and to what extent they may benefit from CI in the setting of NF2.

Conclusion

While the typical course of neurofibromatosis type 2 leads to significant, progressive hearing deficits associated with recurring vestibular schwannomas, our findings show promising long-term utility of cochlear implants in the ipsilateral ear. We found that larger lifetime VS tumor size is significantly correlated with the odds of becoming a non-user of CI (p=0.028). While failure of CI function and profound hearing loss are frequent in the course of NF2, patients may still experience long-term benefits from cochlear implants.