Clinico-Radiological Manifestations of Cochlear Schwannomas
Article information
Abstract
Background and Objectives
Cochlear schwannomas, which are categorized into intracochlear and intravestibulocochlear schwannomas (ICs and IVCs, respectively) are rare and may cause hearing loss (HL). The affected region is invariably correlated with tumor location, which can be detected on magnetic resonance imaging (MRI). We describe the cochleovestibular manifestations of ICs and IVCs.
Subjects and Methods
The study included 31 patients with ICs or IVCs. Tumor extent and exact locations were delineated using MRI. Types of HL were subcategorized into the low-to-mid frequency (250 Hz to 1 kHz), mid-to-high frequency (>1 kHz), and all-frequency (universal) HL groups.
Results
The tumors involved the entire cochlear turn (two ICs) or extended beyond the cochleae (nine IVCs) in 11 patients, and 20 ICs were located in specific locations as follows: 14 in the basal, 3 in the middle, and 3 in the middle and apical turns. No patient showed tumor invasion of the internal auditory canal or middle ear. The pattern of HL usually reflects the location or extent of a tumor. We observed HL at all frequencies, at low-to-mid frequencies, and at mid-to-high frequencies in 13, 4, and 14 patients, respectively. Dizziness or tinnitus was observed in >50% of patients. Surgical tumor removal was performed in 10 patients, and the remaining patients are undergoing annual monitoring.
Conclusions
Cochlear schwannomas may be associated with HL, which may worsen over time and reflect tumor location. Therefore, these lesions should be considered in the differential diagnosis in patients who present with idiopathic, fluctuating, progressive or sudden HL.
Introduction
Inner ear schwannomas can be classified based on their location and extent of organ involvement. Schwannomas that involve the cochlea, but not the internal auditory canal or middle ear, can be classified as intracochlear or intravestibulocochlear schwannomas. An intracochlear schwannoma (IC) originates from the terminal Schwann cells of the membranous labyrinth and most commonly occurs in the modiolus and the basal turn of the cochlea [1].
An IC may be a solitary growth or an extension of a (more common) vestibular schwannoma arising in the internal auditory canal. A benign, slow-growing IC is usually discovered much later, only when symptoms (hearing loss [HL], tinnitus, or dizziness) develop. Unilateral HL (either progressive or sudden) usually precedes other presentations. Prior to recent advances in imaging, such tumors were incidentally found either during labyrinthectomy to treat Meniere’s disease, or autopsy [2,3]. Given the recent rapid advances in magnetic resonance imaging (MRI), ICs have become more frequently (and more easily) detected during evaluation of patients with HL in daily practice. Several authors have developed MRI-guided anatomically based staging systems for similar tumors [4-6]. Some ICs are associated with intralabyrinthine tumors. These tumors, which may involve both the cochlea and vestibule, are called intravestibulocochlear schwannomas (IVCs) [6]. At initial presentation, the HL may be restricted to certain frequencies, suggesting the possibility of a focal cochlear lesion. The indolent growth pattern may explain why ICs were often misdiagnosed as Meniere’s disease (even after several follow-up investigations); the isolated HL persisted [2,3]. However, progression of HL may reflect tumor growth. Neff, et al. [2] suggested that patients with serviceable hearing (better than 50 dB HL and 50% discrimination) and tolerable (infrequent) vertigo should be treated conservatively (i.e., a wait-and-scan strategy) [2,4,6,7]. The cited authors considered that surgical tumor removal was indicated only if medications or vestibular rehabilitation did not relieve the vertigo, despite the fact that hearing may deteriorate as disease progresses. Here, we explore the characteristics of ICs, the relationships between HL and tumor locations, and provide more information on these diseases.
Subjects and Methods
Patient selection
We retrospectively reviewed patient records and included 31 patients with MRI-evident schwannomas involving the cochlea, seen from October 2006 to January 2023. Patients with schwannomas involving the IAC were excluded. The clinical history, physical examination data, and hearing and vestibular assessments were studied. Tumor location was defined using the modified Kennedy system [6]. Tumors with extracochlear extensions into the vestibule were included and termed intravestibulocochlear schwannomas (IVCs). The study was approved by the Institutional Review Board of Severance Hospital (approval no. 1-2021-0302). All patients gave written informed consent.
Audiologic and vestibular evaluation
The hearing thresholds at 0.25, 0.5, 1, 2, 4, and 8 kHz were determined by pure-tone audiometry. We categorized these frequencies as low (0.25 and 0.5 kHz), intermediate (1 and 2 kHz), and high (4 and 8 kHz) [8]. HL was categorized into low-to-mid-frequency HL (0.25 to 1 kHz), mid-to-high-frequency HL (>1 kHz), and all-frequency (universal) loss. The initial hearing data were matched with the concurrent MRI findings. Word recognition test results were collected; however, they were not used in this study because of their low discrimination.
Vestibular functions were measured objectively using, first, the caloric test; a canal paresis (CP) value >26% (derived using Jongkee’s formula) is considered abnormal. Also, the video head impulse test (v-HIT) was performed; a low lateral canal gain (<0.8) and a low vertical canal gain (<0.7) accompanied by a catch-up saccades amplitude >90°/s [9,10] was considered abnormal. Finally, absent or reduced cervical vestibular-evoked myogenic potentials (VEMPs) on the affected (compared to the unaffected) side reflected an abnormality.
MRI
MRI employed a 3.0-tesla (Achieva; Philips Medical Systems, Best, the Netherlands) system fitted with an eight-channel sensitivity-encoding (SENSE) head coil. Conventional temporal MRI was performed prior to contrast injection. The protocol included axial spin-echo (SE) T1-weighted imaging (repetition time [TR]=514 ms, echo time [TE]=8 ms, number of signals acquired [NSA]=1, section thickness=2 mm, intersection gap=0.2 mm, matrix 256×205, field of view [FOV]=18×18 cm), and axial isotropic 3D fast SE T2-weighted imaging with variable, flip angle refocusing pulses (TR=2,000 ms, TEeff=228 ms, NSA=2, echo train length [ETL]=70, matrix 248×250, FOV=18×18 cm, slab number=1, slab thickness=40 mm, partition number=50, partition thickness=0.6 mm, acquisition time=6 min 40 s). Then, gadopentetate dimeglumine (Magnevist; Bayer Schering Pharma, Berlin, Germany) was administered intravenously at 0.2 mL/kg body weight at 2 mL/s. Forty seconds later, fat-saturated axial isotropic 3D T1 fast-field echo (FFE) sequences (TR=25 ms, TE=4.6 ms, flip angle=30°, NSA=1, matrix 512×512, FOV=18×18 cm, partition thickness=0.6 mm, acquisition time=3 min 57 s) were acquired and both isotropic 3D MR sequences reformatted into the coronal plane.
One experienced head-and-neck radiologist, blinded to patient histories and audiological results, retrospectively reviewed all MR images. IC locations were determined by comparing the tumor signal intensities to those of the adjacent membranous labyrinths evident on fast SE T2-weighted and contrast-enhanced 3D T1 FFE images. Tumor locations were categorized into the basal, middle, and apical cochlear turns; all cochlear turns; and extra-cochlear extensions into the vestibule. Surveillance MRI was performed once a year if the symptoms were stable. Tumor size and extension were thus monitored.
Statistical analysis
Correlations between symptoms and locations were assessed using a two-tailed Fischer’s exact test, and p-values less than 0.05 were considered statistically significant. Categorical variables were evaluated using a two-tailed Fisher’s exact test. Because the variables did not show normality, we adopted non-parametric statistical methods and established categorical variables. The cases were subdivided into two groups based on either partial or total tumor involvement, whereas hearing was subdivided into partially preserved or totally deaf.
Results
Thirty-one patients (12 men and 19 women [61.3%]) aged 20 to 73 (mean 46.58±11.13) years were included. Overall, 21/31 patients had left-sided tumors. All presented with progressive, unilateral hearing impairment, except one with sudden-onset unilateral HL. Seven patients (22.6%) reported only HL, four (12.9%) reported HL with both tinnitus and dizziness, and 20 (64.5%) reported HL with either tinnitus or dizziness (Table 1). MRI revealed that 22 patients had ICs confined to the cochlea, and the remaining nine tumors extended into the vestibule. Of the ICs, 14 were in the basal cochlear turn, three were in the middle turn, three were in both the middle and apical turns, and two involved all turns. Two of the nine patients with IVCs had tumors in the basal turns and vestibules; in the remaining seven patients, the tumors involved all cochlear turns. However, no tumor invaded the internal auditory canal or middle ear during serial follow-up.
Of the 14 patients with tumors confined to the basal cochlear turn, 13 presented with mid-to-high frequency HL (Case 3; Fig. 1A-D) and one with universal HL (mean 64.5±13.6 dB HL). Patients with ICs confined to the middle cochlear turn (n=3) (mean 47.9±17.3 dB HL) or middle and apical turns (n=3) (mean 55.0±19.9 dB HL) had mostly low-to-mid frequency HL (Case 17; Fig. 1E-H). IC involving all cochlear turns was associated with all-frequency HL (Case 22; Fig. 1I-L) (mean 86.3±18.6 dB HL). Patients with IVCs presented with universal HL regardless of the cochlear tumor location (Case 25; Fig. 1M-P) (mean 114.7±9.1 dB HL). The HL pattern reflected the tumor location or extent. Hearing remained in 90% (18/20) of patients with ICs that partially involved the cochlea, whereas hearing was not preserved in patients with total cochlear involvement (mean 114.7±9.1 dB HL) (p=0.0001).
In four cases, serial hearing and imaging monitoring revealed that hearing deteriorated as the tumor grew. In one case, the symptoms worsened despite no evidence of tumor growth on imaging. The duration of follow-up ranged 13 months to more than 10 years with our conservative wait-and-scan policy. Seven patients with ICs and three patients with IVCs underwent surgical tumor removal after the symptoms worsened; most tumors were removed through an endoscopic transcanal rout (Table 1 and Supplementary Table 1 in the online-only Data Supplement) [11]. One patient with extreme anxiety about the tumor repeatedly requested tumor removal, and the senior author agreed. Three patients with IVCs underwent contemporaneous cochlear implantation (CI). All patients with IVCs reported dizziness compared to only 23% of IC patients (n=5/22). After tumor removal, the dizziness and anxiety resolved in seven patients, and hearing was rehabilitated in three patients (Supplementary Table 1 in the online-only Data Supplement).
Discussion
ICs can be difficult to image; the tumors are very small and vary in terms of location and the clinical presentations. An IC is the second most common intralabyrinthine schwannoma of the seven types identified by Kennedy, et al. [5]. It is challenging to finalize an IC diagnosis; the clinical symptoms and imaging findings are similar to those of other more common conditions including sudden HL, labyrinthitis ossificans, and Meniere’s disease [4,9]. A tumor in the vestibule in conjunction with a cochlear schwannoma is termed an IVC. Inner ear schwannomas are frequently missed; they are minute. Van Abel, et al. [6] reported that the average time to diagnosis after the initial symptoms appear may be up to 7 years; only then does the tumor become visible. Our patients were heterogeneous, but the tumor locations and HL patterns were consistent; hearing worsened in three cases as tumor growth progressed. In one case, symptoms worsened in the absence of any imaging evidence of tumor growth, indicating that anatomical detection does not always precede physiological changes. A prior isolated HL may gradually progress to a global HL attributable to tumor growth or pathological deterioration.
In this study, the pattern of HL reflected the tumor location or extent. Some HL was confined to specific frequencies at the initial presentation. High-frequency HL was observed in patients with a tumor localized to the basal turn. However, no residual hearing was detected in cases with involvement of all turns or the vestibule. Up-sloping hearing configurations involving the low-to-mid frequencies were audiographically evident in four cases with tumors in the middle and apical turns, attributable to the tonotopic organization of human hearing (which is linked to specific locations along the cochlear turns) [10,12]. Earlier reports suggested that tumor location predicted the symptoms [6,13-15] but no work has focused on specific (frequency-based) HL. However, the differences may be apparent only in the early stages of tumor growth because devascularization or chemical destruction of the basilar membrane throughout the cochlear turns can yield “flat loss” audiographs [5,16,17]. Such (continuous and subtle) pathological progression may also explain why, sometimes, early hearing frequency patterns do not anatomically correlate with tumor locations on surveillance images. Similarly, the correlation between vestibular symptoms and tumor location was stronger. All nine IVC patients experienced dizziness, which was probably the result of the tumor extending into the vestibule and affecting vestibular function. Van Abel, et al. [6] found that vertigo and imbalance were common when tumors extended into the vestibular labyrinth, but only 36% of IC patients reported these symptoms. However, the ubiquitous presence of dizziness has also been documented in isolated intracochlear tumors. Jerin, et al. [18] reported that tumor compression of the endolymph flow leading to endolymphatic hydrops contributed to dizziness in ICs without vestibular extension [2]. One systematic review suggested that vertigo was the second-most common complaint of patients with intralabyrinthine schwannomas [19]. Perhaps the tumors damage the vestibular system after direct cochlear tumor compression or induction of vascular insufficiency; sparing of Scarpa’s ganglion may trigger (continuous) faulty neural transmission and endolymphatic hydrops formation in the inner ear [7,20,21]. Tumor compression along the endolymph flow path (triggering endolymphatic hydrops) may cause dizziness even in the absence of vestibular extension [18,21,22]. This sequela may explain why patients with intravestibular tumors experienced much worse HL, including total deafness, despite the variable cochlear extension apparent in our case series [5,7,21]. This sequela can also explain the mismatch of vestibular function tests, such as a normal head impulse test and an abnormal caloric result. We noticed that patients with dizziness almost always exhibited basal turn involvement, including those with only cochlear schwannomas.
There have been many reports in which an abnormal caloric test was correlated with an abnormal vHIT test in a patient with intralabyrinthine schwannoma [18-21,23]. Lee, et al. [24] found that an abnormal caloric test result correlated with an abnormal vHIT test result. By contrast, we observed normal vHIT data in IVC cases with CP. A difference in the results of the two tests is commonly seen in patients with chronically injured vestibular systems or compensated vestibular function (such as those with Meniere’s disease [25,26]). Tumors of vestibular origin (such as IVCs) may infiltrate and damage the balance system more so than ICs. The VEMP test is more sensitive to vestibular nerve distension or compression than the caloric test. Thus, a reduced VEMP figure but a normal caloric test result may indicate nerve infiltration by a schwannoma rather than compression by an external tumor [27]. Fröhlich, et al. [28] recently investigated the VEMP responses of patients with (only) ICs; the VEMP amplitudes were enhanced, as was also observed in patients with third window syndrome and Meniere’s disease [29,30]. Both ICs and IVCs may trigger endolymphatic hydrops via direct tumor compression or obstruction of the labyrinthine fluid. The VEMP test thus aids clinicians in counseling patients and selecting an optimal surgical approach, particularly for patients with small tumors and residual hearing [28]. Our findings suggest that the worsening of hearing and vestibular symptoms in IC patients can be predicted. It is absolutely essential to use MRI to screen and monitor ICs in patients with unilateral HL or vestibular dysfunction. Recent MRI developments render it possible to detect vestibuloacoustic tumors.
Clinicians must also be aware that an IC or IVC must be diligently sought in the absence of a (more common) vestibular schwannoma [2,13]. Most IC/IVC patients may evidence some residual hearing or complain of (only) tolerable dizziness. Thus, management typically features serial hearing monitoring, vestibular tests, imaging, and optimization of medical support [6,7]. Comprehensive counseling on the benefits and risks of surgery is essential. Given recent advancements in endoscopic surgery, removal of an IC or IVC with concurrent CI may be possible. A consensus classification would aid definitive patient management. A wait-and-scan policy is often recommended; we adapted this in earlier years. However, we are now of the view that young patients with refractory symptoms, tumors confined to the basal turn, and generous residual hearing, should be offered surgical removal. For patients with universal HL caused by extensive intracochlear tumor extension, or those with vestibular involvement, hearing rehabilitation combined with CI is optimal. Plontke, et al. [31] reported good hearing outcomes after CI even after extensive drilling over the cochlea, and tumor removal, if the modiolus (especially the basal turn) was preserved.
In conclusion, cochlear schwannoma can cause distinct patterns of HL depending on its location, which can worsen over time after the initial presentation. Therefore, even in an incomplete HL, the possibility of cochlear schwannoma should be considered.
Supplementary Materials
The online-only Data Supplement is available with this article at https://doi.org/10.7874/jao.2024.00115.
Notes
Conflicts of Interest
The authors have no financial conflicts of interest.
Author Contributions
Conceptualization: In Seok Moon. Data curation: Jinna Kim, Seo Jin Moon, Hae Eun Noh. Formal analysis: Noor Dina Hashim, Seo Jin Moon. Funding acquisition: In Seok Moon. Investigation: Seo Jin Moon, Hae Eun Noh. Methodology: In Seok Moon, Noor Dina Hashim. Project administration: In Seok Moon. Resources: In Seok Moon. Software: Khairunnisak Misron. Supervision: In Seok Moon. Validation: Jinna Kim. Visualization: Khairunnisak Misron, Noor Dina Hashim. Writing—original draft: Noor Dina Hashim. Writing—review & editing: Noor Dina Hashim, Khairunnisak Misron, In Seok Moon. Approval of final manuscript: all authors.
Funding Statement
This study was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Grant No. RS-2023-00253371) to I.S.M., Republic of Korea
Acknowledgements
None