Surgical Resection of Endolymphatic Sac Tumors in von Hippel-Lindau Disease: Findings, Results, and Indications

Abstract

Objectives/Hypothesis:

To define the surgical treatment and outcomes of von Hippel-Lindau (VHL) disease-associated endolymphatic sac tumors (ELSTs), we analyzed consecutive VHL patients who underwent ELST resection.

Study Design:

Retrospective investigation of consecutive VHL patients who underwent resection of ELSTs at a clinical research center between 1999 and 2010.

Methods:

Analysis of serial clinical examinations, audiograms, imaging studies, and operative findings were analyzed.

Results:

Thirty-one consecutive patients with ELSTs (15 males, 16 females) underwent resection of 33 tumors (mean follow-up, 49.9 ± 48.0 months; range, 1.0–116 months). One patient had bilateral ELST resections and one patient underwent reoperation for recurrence. Mean age at surgery was 38.2 ± 10.2 years (range, 12–67 years). Whereas 29 ears (88%) had direct radiographic evidence of an ELST, four ears (12%) did not. Mean tumor size was 1.3 ± 1.1 cm (range, 0.2–5.2 cm). Whereas two patients (two ears, 6%) were asymptomatic, 29 patients (31 ears, 94% of ears) had associated audiovestibular symptoms, including sensorineural hearing loss (28 ears, 84%), tinnitus (24 ears,73%), and vertigo (21 patients, 68%). Postoperatively, hearing was stabilized (27) or improved (three) in 97% of 31 ears. Complete tumor resection was achieved in 30 ears (91% of 33 ears). Complications included cerebrospinal fluid leak in two ears (6%) and transient lower cranial nerve palsy in one ear (3%).

Conclusions:

Surgical resection of ELSTs can be performed with hearing preservation and a reduction in audiovestibular dysfunction. Early surgical resection can prevent or decrease disabling audiovestibular symptoms, enhance the opportunity for complete resection, and preserve hearing.

INTRODUCTION

Endolymphatic sac tumors (ELSTs) are neuroectoderm-derived tumors that exhibit a papillary and glandular architecture.^1,2 ELSTs were first histologically classified as a low-grade papillary adenocarcinoma of probable endolymphatic sac origin.³ Recent clinicopathologic evidence indicates that these tumors do arise specifically from the endolymphatic sac/duct tissue contained within the vestibular aqueduct.⁴ Despite their benign categorization, ELSTs are locally destructive and invade the surrounding petrous temporal bone. ELSTs can occur sporadically but are frequently found in von Hippel-Lindau (VHL) disease patients who can develop bilateral tumors.^5,6 ELSTs can cause the audiovestibular morbidity, including sensorineural hearing loss (SNHL), vertigo, tinnitus, otalgia, aural fullness, and cranial nerve deficits.^5,7

The mechanisms that underlie audiovestibular morbidity are often not related to tumor size and include otic capsule invasion, intralabyrinthine hemorrhage, and endolymphatic hydrops.^5,8,9 Although early surgery may potentially prevent ELST-associated audiovestibular dysfunction, the long- term results of surgical resection of ELSTs (especially in VHL) and the benefits/safety of early resection has not been firmly established in a large series. Specifically, hearing preservation and the amelioration of other audiovestibular signs/symptoms, as well as the long-term effectiveness of resection have not been determined. To define the role and timing of ELST resection in VHL, we analyzed the clinical, imaging, genetic, and operative findings in consecutive VHL patients who underwent ELST resection at the National Institutes of Health (NIH).

MATERIALS AND METHODS

Patients

This is retrospective analysis of consecutive VHL patients (with genetic analysis confirming a VHL gene mutation or clinical diagnosis of VHL based on the clinical criteria¹⁰) who underwent resection of an ELST by the senior surgeons (h.j.k. and r.r.l.) at NIH, Bethesda, Maryland. All patients were treated under an institutional review board-approved protocol after informed consent was obtained. The natural history of ELST-associated hearing loss (HL) was previously described in some patients included in this study.^9,11,12 Clinical, audiologic, radiographic, operative, and pathologic findings were analyzed.

Patient Evaluation

Clinical evaluation.

Neurotologic and audiological examinations were carried out at intervals of 6 to 24 months. Additional clinical evaluations were performed in response to subjective changes in audiovestibular symptomatology. Audiovestibular signs and symptoms (i.e., HL, tinnitus, vertigo, aural fullness, and cranial nerve abnormalities) were recorded at each visit. HL was defined as sudden (<3 days) or gradual (weeks to months) by the time course over which it occurred.

Genetic analysis.

VHL mutations were determined from peripheral blood samples using Southern blotting, fluorescence in situ hybridization, and complete gene sequenciong.¹³ Mutations are classified according to the nature of DNA changes.

Audiometric evaluation.

Audiometric evaluations were performed during routine VHL clinical evaluations or when changes in hearing and/or other audiovestibular symptoms were noted. A four-frequency pure-tone average (4F-PTA) based on air conduction thresholds at 0.5, 1, 2, and 4 kHz was used to represent degree of HL as follows: grade 0 (normal, 0–25 dB HL); grade 1 (mild, 26–40 dB HL); grade 2 (moderate, 41–55 dB HL and moderate-severe, 56–70 dB HL); grade 3 (severe; 71–90 dB HL); and grade 4 (profound, 91 dB HL or greater). This grading system was used to score hearing changes after surgery.

The immediate preoperative audiograms and postoperative surveillance audiograms were compared to assess hearing status after surgery. Hearing was categorized as stable if the preoperative grade of HL was maintained at the same grade as that at the final follow-up visit and as improved or worse if the preoperative grade of HL was better or worse, respectively, at the final follow-up visit.

Radiographic evaluation.

Patients were evaluated at clinical visits and preoperatively and postoperatively with serial high-resolution, T1-weighted (with and without contrast), T2-weighted, and fluid attenuated inversion recovery (FLAIR) (with and without contrast) magnetic resonance imaging (MRI). Patients were also evaluated preoperatively and postoperatively with high-resolution computed tomography (CT) imaging of the temporal bones. Largest linear in-plane diameter on T1-weighted postcontrast sequences was used to determine tumor size. The presence of otic capsule invasion and/or intralabyrinthine hemorrhage was recorded.^8,9 Patients were evaluated preoperatively and postoperatively with high-resolution contrast-enhanced temporal bone CT imaging.

Pathologic evaluation.

Surgically resected tumors were analyzed using routine hematoxylin and eosin staining and immunohistochemical analysis.^14–16

Statistical Analysis

Descriptive statistics were used to summarize demographic and surgical outcome data.

RESULTS

Patient Characteristics

Thirty-one consecutive patients (15 males, 16 females), including four patients with bilateral ELSTs, underwent resection of 33 ELSTs (Table I). ELSTs were resected from the left ear in 23 cases (70%) and from the right ear in 10 cases (30%). One patient had bilateral ELST resections, and one patient required reoperation for recurrence. Mean follow-up was 49.9 ± 48.0 months (range, 1.0–116 months). Mean age at initial visit at NIH for ELST symptoms was 32.5 ± 10.7 years (range, 11–67 years). Mean age at surgery was 38.2 ± 10.2 years (range, 12–67 years). Twenty-five (76%) ears underwent resection of de novo tumors. One patient underwent stereotactic radiation therapy for an ELST 2 years before surgery but developed fluctuating HL and dizziness 1 year after treatment. Seven ears (21%) underwent removal of residual or recurrent tumor after prior resection.

TABLE I.

Demographics and Preoperative Findings in 31 Patients with von Hippel-Lindau Disease with 33 Ears Affected by Endolymphatic Sac Tumors That Underwent Surgical Resection.

Finding	Frequency of Patients or Ears
Sex
Female	16 patients (52%)
Male	15 patients (48%)
Mean age at initial audiovestibular symptom, yr	30.1 ± 11.9 (11–63)
Mean age at initial visit, yr	32.5 ± 10.7 (11–67)
Mean age at surgery (range), yr	38.2 ± 10.2 (12–67)
Tumor location
Left	23 ears (70%)
Right	10 ears (30%)
Bilateral	1 patient (3%)
De novo tumor	25 ears (76%)
Recurrent/residual after outside resection	7 ears (21%)
Recurrent after our resection	1 ear (3%)
Imaging evidence of tumor mass
Yes	29 ears (88%)
No	4 ears (12%)
Mean tumor size (range), cm	1.3 ± 1.1 cm (0.2–5.2)
Symptomatic	1.4 ± 1.1 cm (0.2–5.2)
Asymptomatic	0.5 ± 0.2 cm (0.2–0.8)
Signs and symptoms
Asymptomatic	2 ears (6%)
Symptomatic	31 ears (94%)
Hearing status (4F-PTA)
Normal (0–25 dB HL)	8 ears (24%)
Normal	5 ears (15%)
HL at 6,000 and 8,000 Hz	2 ears (6%)
HL at 250 and 500 Hz	1 ear (3%)
Mild (25–40 dB HL)	3 ears (9%)
Moderate/moderate-severe (41–70 dB HL)	9 ears (27%)
Severe (71–90 dB HL)	1 ear (3%)
Profound (>90 dB HL)	12 ears (36%)
SNHL onset
Sudden	15 ears (52%)
Gradual	14 ears (48%)
Tinnitus	24 ears (73%)
Vertigo	21 patients (68%)
Aural fullness	13 ears (39%)
Facial nerve weakness	4 patients (12%)

4F-PTA = four-frequency pure-tone average; SNHL = sensorineural hearing loss.

Genotypes

The genotypes of the VHL gene were available in 30 patients (Table II). The most common VHL mutation was a partial deletion and was found in 12 patients (40%). Three patients (10%) had nonsense mutations, and three patients had a mutation on the splicing acceptor sequences. Missense mutations were found in seven patients (23%). Three of the missense mutations occurred in the a-subunit of the VHL gene, which is involved in the ubiquitin ligase complex formation. Four missense mutations occurred in the β-subunit, which is involved in the binding of the primary substrates (i.e., hypoxia-inducible factors). Three of the four patients with a missense mutation in the β-subunits had bilateral ELSTs.

TABLE II.

Genotype of VHL Gene in 31 Patients.

Finding	No. of Patients
Partial deletion	12
Nonsense mutations	3
c.699C>A p.Cys162Stop, nonsense
c.407C>A p.Trp655Stop, nonsense
c.694C>T p.Arg161Stop, nonsense
Splicing mutations	3
Splice acceptor insert GGT in intron
1 at splice acceptor site 5′ of exon 2
G>A at −1 position of exon 3 splice acceptor
G>C at −1 position in exon 3 splice acceptor
Missense mutations	7
c.694C>G p.Arg161 Gly
c.686T>C p.Leu158Pro (2 related patients)
CT>GG NT 514–515 p.Leu158Pro
c>545G>A, p.Ser111Asn
c.302T>G p.Leu101Arg
c.470C>G p.Pro86Arg
No genotype available	6

Preoperative Clinical Findings

ELSTs caused audiovestibular deficits in 31 ears (94%). Two patients (two ears, 6%) had ELSTs that were present on CT and MRI but had no audiovestibular symptoms. Overall, the most common symptoms associated with ELSTs included SNHL (28 ears, 84%), tinnitus (24 ears, 73%), and vertigo (21 patients, 64%) (Table I). SNHL occurred suddenly (14 ears, 56%) or gradually (11 ears, 44%). Less common ELST-associated symptoms included aural fullness (13 ears, 39%). Four patients (12%) had facial nerve weakness (House-Brackmann [HB] grade 2, 2, 4, and 5).

Radiological Tumor Characteristics

General features.

Although 29 ears (88%) demonstrated radiographic evidence of ELST on CT and/or MRI, there was no obvious imaging evidence of ELSTs in four symptomatic ears (12%) (Table III). Mean tumor size of visible tumors was 1.3 ± 1.1 cm (range, 0.2–5.2 cm). Six ELSTs (21%) eroded into the otic capsule only, two ELSTs (7%) eroded into the internal auditory canal (IAC), and four ELSTs (14%) had eroded into both the otic capsule and IAC (Figs. 1 and 2). On postcontrast T1-weighted and FLAIR MRI, ELSTs demonstrated heterogeneous enhancement and were associated with intralabyrinthine hemorrhage in 11 of 29 ears (38%). Although four ears (12%) did not have evidence on MRI of an enhancing lesion around the endolymphatic sac and duct at the time of surgery, three (75%) of the four ears demonstrated evidence of prior intralabyrinthine hemorrhage on MRI. Two of these ears on MRI and two (50%) on CT had evidence of subtle erosion of the vestibular aqueduct.

TABLE III.

Imaging Findings in Radiographically Evident Endolymphatic Sac Tumors.

Imaging Status	No. of Patients or Ears
Radiologically evident	29 ears (88%)
Enhancing tumor mass	29 ears (100%)
Intralabyrinthine hemorrhage	11 ears (38%)
Otic capsule invasion only	6 ears (21%)
IAC invasion only	2 ears (7%)
Combined otic capsule and IAC invasion	4 ears (14%)

IAC = internal auditory canal.

Fig. 1.

Right endolymphatic sac tumor. (A) Axial computed tomography of temporal bone showing bony erosion (white arrows) in the posterior petrous bone including the posterior semicircular canal (white asterisk) and the vertical segment of the facial nerve (black asterisk). Note presence of bony spicules with the tumor. Characteristic noncontrast heterogeneous T1-weighted magnetic resonance imaging (MRI) with the areas of high T1-weighted MRI signal suggestive of intratumor hemorrhage (black arrow) (B), and diffusely enhancing area in the posterior petrous bone (C).

Fig. 2.

Left endolymphatic sac tumor. (A) Computed tomography of left temporal bone with bony erosion in the posterior petrous bone. (B) Noncontrast T1-weighted magnetic resonance imaging (MRI). (C) Gadolinium-enhancing T1-weighted MRI showing enhancement of the tumor.

Relationship of symptoms to imaging features.

Signs and symptoms occurred with ELSTs of all sizes, including those tumors not evident on imaging. Two asymptomatic ears harbored ELSTs that were evident on CT and MRI (0.2 and 0.8 cm). All four patients who lacked direct radiologic evidence of ELST had severe-profound SNHL, tinnitus, and vertigo. Of the 10 ears (30%) that displayed otic capsule invasion with or without IAC invasion (Fig. 3), eight ears (80%) had SNHL that occurred suddenly (four ears, 50%) or gradually (four ears, 40%), and one patient (one ear, 10%) had otic capsule invasion with no audiovestibular signs or symptoms. Of the 23 ears without otic capsule invasion (70%), four ears (17%) had normal hearing and 19 ears (83%) had SNHL that occurred suddenly (11 ears, 58%) or gradually (eight ears, 42%). Sudden SNHL was found in 10 of 14 patients (71%) with intralabyrinthine hemorrhage.

Fig. 3.

Left large endolymphatic sac tumor (arrows). (A) Computed tomography of left temporal bone. (B) Noncontrast T1-weighted magnetic resonance imaging (MRI) showing heterogeneously increased MRI signal in the posterior petrous bone and internal auditory canal. (C) Gadolinium-enhancing T1-weighted MRI showing heterogeneously enhancing tumor. Note a hemangioblastoma of the cerebellum (asterisk).

Surgical Approach

Surgical approach was selected based on tumor size, hearing status, and anatomic relationship to critical structures. Specific approaches included retrolabyrinthine petrosectomy (24 tumors, 73%; one operation included cochlear implantation), transotic (four tumors, 12%), translabyrinthine or infratemporal/translabyrinthine (two tumors, 6%) and combined presigmoid/retrosigmoid (three tumors, 9%). Of the combined presigmoid/retrosigmoid approaches, two were retrosigmoid/transmastoid and one was retrosigmoid/translabyrinthine.

Preoperative tumor embolization of external carotid feeding arteries was used in two patients (6%) with large tumors.

Intraoperative Findings

The ELSTs grossly had a hypervascular raspberry-like appearance and often eroded the solid petrous bone or the mastoid air cells. Cholesterol granuloma was found around or within the tumor in 16 (48%) cases. All of the 25 previously nonoperated ELSTs involved the endolymphatic duct (one tumor, 4%), the endolymphatic sac (one tumor, 4%), or both (23 tumors, 92%). This intraoperative finding supports the previously proposed theory of ELSTs originating from the endolymphatic sac apparatus. In six out of seven previously operated cases, residual tumor was found in the endolymphatic sac and duct. In 11 cases, including four patients who lacked direct radiologic evidence of ELST, the tumors were small and confined to the endolymphatic sac and duct. However, larger tumors extended medially and inferiorly via the retrofacial route to invade into the jugular bulb (two tumors, 6%) and the hypotympanum (five tumors, 15%). One tumor invaded into the sigmoid sinus. Eight tumors (24%) eroded into the posterior semicircular canal and one tumor (3%) eroded into the superior semicircular canal. The semicircular canals were either removed as a part of a translabyrinthine approach or partially occluded to preserve hearing after tumor resection. Larger tumors often involved the posterior fossa dura and extended posteriorly to the cerebellopontine angle (six tumors, 18%) and anteriorly to the internal auditory canal (four tumors, 12%). Four tumors (12%) extended superiorly to the tegmen mastoideum and the middle fossa dura, and the superior petrosal sinus were partially resected in one case.

Six tumors (18%) invaded or encroached on the vertical segment (two tumors) or the meatal segment (four tumors) of the facial nerve, and four of these patients exhibited facial nerve palsy. Once the vertical segment of facial nerve was dissected off the tumor, facial nerve function improved. In two cases it was difficult to dissect the tumor from the facial nerve in the IAC because no discrete anatomical plane separating the tumor from the facial nerve could be identified, and the residual tumor was left behind in the IAC.

Tumor Control

Complete resection of ELSTs confirmed by postoperative imaging was achieved in 30 ears (91%). Complete ELST resection was not possible in three ears (9%) with large tumors (2.6, 4.1, and 5.2 cm) that encased the carotid artery within the carotid canal and invaded the meatal segment of the facial nerve. Two of these patients underwent focused irradiation. Tumor remained stable in one patient (follow-up, 78 months) but progressed in the other (follow-up, 105 months). Recurrence was identified in one patient 46 months after initial resection. This patient underwent successful complete re-resection without recurrence (last follow-up, 25 months).

Postoperative Clinical Findings

Audiometric results.

ELST resection frequently resulted in stabilization and preservation of preoperative hearing levels. Based on 4F-PTA (31 ears had preoperative and postoperative audiometry beyond 7 days after surgery, two ears lacked audiometry beyond 7 days post-operation and were not included in the analysis), hearing grade remained stable or improved in 30 ears (97%). All 12 ears with serviceable 4F-PTA preoperative hearing <50 dB HL demonstrated stable serviceable hearing postoperatively (39% of all ears) at 40 months follow-up (range, 2.5–87 months). Included in this group were eight ears (67%) with normal hearing, two ears (17%) with mild HL, one ear (8%) that improved from serviceable moderate to mild HL, and one ear (8%) with serviceable moderate HL. Stable hearing was also preserved in three ears (10%) with moderate-severe HL, one ear (3%) with severe HL, and 11 ears (36%) with profound HL. Hearing improved over 6 months by one or two grades in two ears (6%) (from profound to moderate-severe and profound to severe). Two ears (6%) decreased in hearing sensitivity from moderate-severe to profound over 34 months after transtympanic gentamicin injection for controlling vertigo. Another ear (3%) decreased in hearing sensitivity from mild to moderate over 116 months and had stable serviceable moderate HL over 62 months.

Vestibular symptoms.

Vertigo resolved in 18 of 21 patients (86%) and improved in the remaining three (14%) patients after surgery. Two patients (6%) developed delayed hydrops-like vertigo 12 and 16 months after surgery and required transtympanic gentamicin injections for resolution of their vestibular symptoms. Tinnitus reportedly resolved or improved in 23 of 24 ears (96%) and remained stable in one ear (4%) after surgery. Aural fullness resolved in 10 of 13 ears (77%) and remained stable in three ears (23%) after surgery.

Facial nerve weakness.

Preoperative facial nerve weakness (HB grade 4, 2, and 2) improved to grade 3, 1, and 1 in three patients (75%) but remained unchanged in the remaining patient (HB grade 5).

Complications

Two patients (6%) with prior posterior fossa irradiation developed cerebrospinal fluid leak after surgery. Both patients underwent wound revision, and one underwent ventriculoperitoneal shunting without further sequelae. One patient with a large tumor (4.1 cm) involving the jugular bulb and carotid canal had transient glossopharyngeal and vagal nerve palsy that resolved after 6 months.

Histologic Findings

All tumors had histopathology and immunohistochemical staining consistent with an ELST (Fig. 4).

Fig. 4.

(A) Hematoxylin and eosin staining of endolymphatic sac tumor with complex papillary structures lined by cuboidal to columnar cells with fairly uniform round nuclei and pale eosinophilic cytoplasm (magnification 20×). (B) Positive immunostaining for cytokeratin AE1/AE3. (C) Positive immunostaining for epithelial membrane antigen (Courtesy of Martha M. Quezado, MD).

DISCUSSION

VHL-Associated ELSTs

VHL disease (MIM no. 193300) is transmitted in an autosomal dominant manner and is caused by germline mutations of the VHL gene.^17–21 The most common VHL-associated tumors are central nervous system and retinal hemangioblastomas, renal cell carcinomas, pheochromocytomas, pancreatic islet tumors, and ELSTs. Manski et al.⁵ established that ELSTs were part of VHL syndrome in 1997, and Vortmeyer et al.²² confirmed ELSTs had loss of heterozygosity of the VHL gene. ELSTs are found in 11% to 16% of VHL patients, and approximately 30% of ELSTs patients will develop bilateral tumors.^5,8,23 Similar to the findings of the present study, previous reports of audiovestibular findings linked to ELSTs include SNHL (90%–95%), tinnitus (60%–80%), and vertigo (60%–80%).^7,8,17

Because SNHL and other audiovestibular morbidity can occur suddenly in ELST and are not related to tumor size, early surgical intervention may be warranted in VHL patients with ELSTs to prevent associated morbidity if surgical resection can be performed safely with hearing preservation. To address the safety and effectiveness of surgical resection, we examined a consecutive series of VHL patients with ELSTs.

Effectiveness of Resection

Surgical ELST resection is safe, effective, and can be accomplished with minimal risk of recurrence. Complete resection was achieved in 30 ears (91% of ears). There was only one case of recurrence after initial complete gross resection (46 months postresection) that was successfully resected. Complete resection of the ELST was precluded in three cases where there was invasion of the carotid canal and where the tumor could not be dissected off from the meatal segment of the facial nerve. These findings under-score the importance of earlier resection, when tumors are smaller and confined to the posterior petrous region.

Hearing Preservation After Resection

Hearing preservation is a likely outcome after ELST resection. The findings of the current study are consistent with the findings of previous studies describing hearing preservation after ELST resection in a small series of patients.^11,23,24 Pure-tone average hearing grade was stable or improved in 30 of 31 ears postoperatively, including all eight ears with normal preoperative hearing. All 12 ears with serviceable preoperative hearing had comparable hearing postoperatively. Three ears demonstrated an improvement in hearing by one or two hearing grades. Two ears demonstrated worsened hearing. Although these findings indicate that surgical resection can be performed with good preservation of hearing, improvement was noted in only three ears, suggesting that reversal of SNHL is unlikely, and only selective cases may show improvement.

Effect of ELST Resection on Audiovestibular Morbidity

In addition to stabilizing a possible progressive SNHL secondary to ELST pathology, resection was beneficial in ameliorating other disabling audiovestibular morbidities such as vertigo, tinnitus, and aural fullness in this study. However, two patients developed vertigo symptoms similar to Meniere’s disease at 12 and 16 months after surgery and required transtympanic gentamicin injection for vertigo. The residual endolymphatic hydrops-like changes from ELSTs may not be reversible after successful tumor resection, and additional treatment may be required for delayed onset vestibulopathy.

Preoperative Considerations

Imaging.

ELSTs can be present with normal hearing. Therefore, we recommend that all VHL patients undergo comprehensive audiologic evaluations, vestibular testing (if indicated), as well as temporal bone CT and MRI of brain/IAC to detect ELSTs early. Combined modalities of CT and MRI studies may enhance currently available diagnostic capability to identify small ELSTs, as CT allows detection of bony erosion around the vestibular aqueduct and MRI allows detection of intralabyrinthine microhemorrhage that can be associated with extremely small ELSTs.

VHL status.

Once a diagnosis of ELST is suspected in a case of unknown VHL status, a comprehensive evaluation for VHL disease should be initiated. It is important to rule out VHL-associated catecholamine-secreting pheochromocytomas before surgery to avoid a life-threatening, perioperative hypertensive crisis.

Embolization.

ELSTs are hypervascular tumors that are often supplied from branches of the external carotid artery and occasionally from the internal carotid artery and posterior cerebral circulation. Large solid tumors may benefit from preoperative embolization before surgical resection to minimize intraoperative bleeding and serious morbidities and to facilitate complete surgical resection.

Indications for ELST Resection in VHL

The findings of the current study, including reproducible hearing preservation and amelioration of audiovestibular morbidity after ELST resection provides clinical evidence to support early resection of ELSTs. To prevent SNHL and/or to alleviate vestibulopathy, early surgical resection of ELSTs should be considered in asymptomatic and symptomatic VHL patients with hearing with MRI/CT evidence of a tumor and/or tumor-associated intralabyrinthine hemorrhage. In cases of profound SNHL in absence of vestibulopathy, ELSTs may be followed with serial imaging studies. However, if tumors demonstrate growth, a surgical resection is required to prevent further morbidity, such as facial nerve palsy. ELST resection and facial nerve decompression should be considered in patients with early onset of facial nerve palsy as the palsy may improve in selective cases. Finally, the incidence of bilateral ELSTs can be as high as 30%^5,24 in VHL, and cochlear implantation can be successfully utilized for aural rehabilitation in deaf patients with ELSTs.^5,25 Consequently, if cochlear implantation is considered in bilateral ELST cases, surgical approaches should be tailored to keep the otic capsule intact and minimize cochlear trauma to achieve optimal cochlear implant benefit.^10,26

Effects of Radiation Therapy

Although our study suggests that surgical resection is an effective mode therapy for ELST, the effectiveness of radiation therapy is unclear for the management of ELSTs. One patient underwent a successful salvage surgery for maintaining hearing and controlling vertigo after failing primary gamma knife therapy for ELST at an outside institution. Two patients with residual tumor underwent external beam radiation therapy. Serial imaging studies indicated that one tumor is stable and the other has been growing. Therefore, our limited experience with radiation therapy suggests that ELSTs may not be sensitive to radiation therapy. Its role as a postoperative adjunct therapy for residual tumors is uncertain and requires further study.

CONCLUSION

Surgical resection of ELSTs can be performed with hearing preservation. Early surgical resection can prevent or decrease disabling audiovestibular symptoms, enhance the opportunity for complete resection, and preserve hearing.