Otology & Neurotology 32:291Y296 2011, Otology & Neurotology, Inc.

Hearing Results After Hypotympanotomy for Glomus Tympanicum Tumors

*Konstantinos Papaspyrou, *Torsten Mewes, *Miklos Toth, Irene Schmidtmann, Ronald G. Amedee, and *Wolf J. Mann
*Department of Otorhinolaryngology, Head and Neck Surgery, and Institute for Medical Biostatistics, Epidemiology and Informatics, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany; and Ochsner Health System, Department of OtolaryngologyYHead and Neck Surgery, New Orleans, Louisiana, U.S.A.

Objective: We postulate, that glomus tympanicum tumors (GTTs) may be safely removed without interference with the ossicular chain via a hypotympanotomy approach. Study Design: Prospective, nonrandomized anatomic and clinical study. Setting: Tertiary referral center. Patients: All 17 patients between 1989 and 2009 with GTTs without involvement of the lumen of the jugular bulb. Interventions: We used a modified hypotympanotomy approach. Our technique is a modification of the one first published by Shambaugh (1955). Pure-tone audiograms were performed in all patients. Preoperative and postoperative audiograms were modeled in a linear mixed model evaluating hearing threshold for air and bone conduction and air-bone gap at 500, 1,000, 2,000, and 3,000 Hz. In an effort to preserve the normal sound conducting apparatus and hearing, we used a retroauricular approach, exposing widely the jugular bulb, the

carotid artery, the protympanum, and even the bony part of the Eustachian tube via a hypotympanotomy. Three formalin-fixed and one macerated temporal bones were dissected step by step under the operating microscope to demonstrate the approach in cadaver dissections. Main Outcome Measure: To evaluate if GTTs can be completely resected without interference with the ossicular chain to improve conductive hearing loss. Results: We found a substantial improvement of hearing threshold after surgery at all frequencies in air conduction. For bone conduction, there was only a slight gain within random variation. The air-bone gap decreased significantly after surgery. Conclusion: Our approach demonstrated a safe avenue for complete tumor removal without interference with the continuity of the ossicular chain. Key Words: Glomus tympanicum tumorVHearingVHypotympanotomyVOssicular chain. Otol Neurotol 32:291Y296, 2011.

In 1955, Shambaugh (1) first described the technique of hypotympanotomy for removal of early glomus jugulare tumors confined to the hypotympanum and tympanic cavity in an effort to gain adequate exposure of the tumor without sacrificing the hearing. This approach to the tympanic cavity had not been described before, although Rosen (2) suggested that Bin cases where the tumor arises from the promontory, it might be possible to excise the entire tumor by simply elevating the tympanic membrane.[ Brown (3) detailed that Bit may be possible to

remove a small hypotympanic tumor through an incision in the floor of the external auditory canal, whereby the lower half of the drum membrane could be turned up and the tympanum exposed.[ MATERIALS AND METHODS Temporal Bones
One macerated temporal bone and 3 formalin-fixed bones were dissected step by step under the operating microscope to demonstrate the hypotympanotomy approach for glomus tympanicum tumor removal (Fig. 1).

Address correspondence and reprint requests to Konstantinos Papaspyrou, M.D., Department of Otorhinolaryngology, Head and Neck Surgery, University Medical Center of the Johannes Gutenberg University Mainz, Langenbeckstr. 1, D-55101 Mainz, Germany; E-mail: papaspyrou@hno.klinik.uni-mainz.de K. P. and T. M. contributed equally to this work.

Patients
We prospectively evaluated, in a nonrandomized study, 17 patients who had a hypotympanotomy approach for glomus tympanicum tumor (GTT) removal between 1989 and 2009; there were 12 women and 5 men. All patients underwent preoperative

291

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

292

K. PAPASPYROU ET AL.
the ossicular chain into the aditus ad antrum were candidates for this approach, extended upward or downward according to the individual situation. The therapeutic concept included no angiography or embolization for these isolated tympanic GTTs (4,5). Pure-tone audiograms were performed in all patients. This study was reviewed and approved by the institutional review board. In an effort to preserve the normal sound conducting mechanism and hearing, the following technique was used to expose the tympanic cavity widely via the hypotympanum, once hypotympanic involvement was demonstrated by preoperative imaging (Fig. 1): after a retroauricular incision, the skin of the posterior meatal wall was incised. A supplementary incision through the skin of the inferior and anterior meatal wall along the outer edge of the osseous meatus follows. The skin of the anterior, inferior, and posterior osseous meatal wall was elevated as a tube to the sulcus tympanicus, thus exposing almost the entire extent of the tympanic bone. After elevation of the tympanomeatal flap that ensures the visualization of the superior part of the tumor and identifies the position of the auditory ossicles, the surgical procedure should include circumferential enlargement of the bony external canal by drilling inferiorly as far as the tympanic sulcus. The inferior osseous meatal wall was thinned down with a burr so as to widen and enlarge inferiorly and anteriorly the osseous meatus. During further drilling inferiorly and medially, the vertical portion of the carotid canal should be identified. This represents the anterior border of the surgical field. The carotid canal makes a sharp angle with the dome of the jugular bulb. After identification, the jugular dome is progressively exposed as far as the styloid prominence. This structure always lies in front of the facial canal; that is, the preservation of the styloid prominence, which is the cranial end of the styloid process, impedes injury to the FN during drilling. The position of the FN in the posterior meatal wall where it emerges from the stylomastoid foramen must be kept in mind. By removing the floor of the meatus anterior to, but not beyond,

FIG. 1. A, Macerated temporal bone after exposure of the hypotympanic cells (right side, lateral view) using a transcanal approach. B, The same anatomic situation in a formalin-fixed specimen with the tympanic membrane in its normal position but with the hypotympanon drilled down, so exposing the jugular bulb and the internal carotid artery if needed. C, Same anatomic situation with the tympanic membrane now reflected after complete tumor removal. The remaining bony defect is filled with fascia and muscle. ICA indicates internal carotid artery; JB, jugular bulb; P, promontory; Sp, styloid process; TM, tympanic membrane; white arrow, umbo; black arrow, round window niche.

computed tomography (CT) (Fig. 2) and/or magnetic resonance imaging (MRI) and/or magnetic resonance angiography. This way, a glomus jugulare tumor could be ruled out. Once a glomus jugulare tumor was excluded based on imaging findings, patients with glomus tympanicum tumors even extending around
Otology & Neurotology, Vol. 32, No. 2, 2011

FIG. 2. Preoperative CT of a patient with a left-sided glomus tympanicum tumor. (This figure is courtesy of Prof. W. Muller Forell, Department of Neuroradiology, Mainz.)

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

HYPOTYMPANOTOMY FOR GLOMUS TYMPANICUM

a vertical line through the posterior edge of the sulcus tympanicus, the facial nerve (FN) is safely avoided. After correct removal of all bony overhangs around the styloid prominence, there is direct visualization into the round window niche. The superior limit of the surgical field is the promontory just below the semicanal of the tensor tympani muscle. For sufficient exposure of the hypotympanum, it also is necessary to lower the anterior wall of the external canal formed by the tympanic bone. It allows not only the exposure of the carotid canal but also the identification of the caroticotympanic artery. The complete exenteration of the angle between the carotid canal and the jugular dome permits removal of the infralabyrinthine cells in the event of tumor involvement. The anulus tympanicus is elevated from its sulcus anteriorly, inferiorly, and posteriorly. This permits the pars tensa to be folded upward upon itself, exposing all of the tympanic cavity and the tumor within it. After complete tumor removal, the defect in the hypotympanon is filled with muscle and fascia, and the tympanic membrane and meatal skin are folded back into position.

293

conduction, frequency, and time were categorical covariates. Duration of follow-up was modeled as continuous covariate. In the second model, the air-bone gap was the dependent variable, whereas frequency and time were categorical covariates. p values for F tests of main effects and interaction terms are given. Furthermore, we present p values of t tests testing for specific differences of interest. These tests were performed in an explorative fashion; therefore, p values are descriptive. Descriptive statistics were obtained using SPSS 15.0 (SPSS, Inc., Chicago, IL, USA). Modeling was performed using PROC MIXED from SAS 9.2 (SAS Institute Inc., Cary, NC, USA).

RESULTS Follow-up time, including audiologic data, ranged from 1 to 236 months (mean, 43.3 mo). Patients most recent audiograms were selected. Hearing threshold depended on type of conduction ( p G 0.0001), frequency ( p G 0.0001), and time of audiometry ( p G 0.0001); duration of follow-up was not associated with hearing threshold ( p = 0.3976). The hearing threshold improved, on average, by 3.6 dB at 500 Hz ( p = 0.2138), by 4.7 dB at 1,000 Hz ( p = 0.1091), by 5.9 dB at 2,000 Hz ( p = 0.0434), and by 8.5 dB at 3,000 Hz ( p = 0.0039). For bone conduction, we found only a slight improvement in hearing threshold that is within random variation with a mean increase of 1.2 dB at 500 Hz ( p = 0.6733), of 1.6 dB at 1,000 Hz ( p = 0.5740), 3.5 dB at 2,000 Hz ( p = 0.2367), and a marginally significant improvement in hearing threshold of 5.4 dB at 3,000 Hz ( p = 0.0686). A composite preoperative and postoperative audiogram is depicted in Figure 3. The air-bone gap decreased, on average, after

Statistical Methods
Mean and standard deviations for the hearing thresholds observed in air conduction and in bone conduction at the main speech frequencies of 500, 1,000, 2,000, and 3,000 Hz both before and after surgery were computed. This included the airbone gap between air conduction and bone conduction at each frequency before and after surgery. We assumed that hearing threshold and air-bone gap were normally distributed. The dependence of hearing threshold and air-bone gap on type of conduction, frequency, time (before/ after surgery), and duration of follow-up was modeled in a linear mixed model, thereby taking into account possible dependencies because of repeated measurements. In the first model, hearing threshold was the dependent variable whereas the type of

FIG. 3.

Hearing threshold before and after surgery for air and bone conduction. Otology & Neurotology, Vol. 32, No. 2, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

294

K. PAPASPYROU ET AL.

FIG. 4.

Air-bone gap before and after surgery.

surgery by 2.4 dB at 500 Hz ( p = 0.1645), by 3.1 dB at 1,000 Hz (p = 0.0787), by 2.5 dB at 2,000 Hz ( p = 0.1546), and 3.2 at 3,000 Hz ( p = 0.0680) (Fig. 4).

DISCUSSION In patients with glomus tympanicum tumors, we routinely use a retroauricular approach because these tumors often are not adequately demonstrated on CT or MRI; thus, their real extent often is determined intraoperatively. Of course, as Shambaugh (1)(p597) points out, the preservation of hearing should not be at the risk of incomplete removal of the tumor. This speaks in favor of a retroauricular instead of a transcanal approach. In addition, if the surgeon finds that hypotympanotomy does not fully and adequately expose the tumor or enters the jugular bulb, there is the possibility to proceed with exposure and occlusion of the jugular bulb via a radical mastoidectomy approach. In 1967, Farrior (6) described a postauricular hypotympanoplasty technique for the removal of glomus tumors similar to the one we used in this study. House and Glasscock (7) maintained that virtually all tympanicum tumors could be safely removed surgically either by transcanal exposure or through an extended facial recess approach, preserving the posterior external auditory canal. Glasscock et al. (8) state that, when the tumor is limited to the promontory and all margins can be clearly defined through the intact tympanic membrane, a transcanal (tympanomeatal flap) procedure is the method of choice. A large glomus tympanicum lesion filling the middle ear
Otology & Neurotology, Vol. 32, No. 2, 2011

and hypotympanum would best be removed by a postauricular approach using an extended facial recess exposure of the middle ear and hypotympanum. However, this often poses a greater risk to the FN. We prefer the postauricular transcanal approach without facial recess exposure. With this technique, the FN is safely avoided. Glasscock et al. reported approximately 35 tympanicum tumor removals, 30 of which were removed by the transcanal or extended facial recess approach, thereby preserving the posterior canal wall in 30 cases, the tympanic membrane in 24, and the ossicular chain in 24. Six patients in this series, however, required tympanic membrane grafting and ossicular reconstruction. Five patients underwent radical mastoidectomy. There were no Bdead[ ears in this group, and in all cases, preoperative hearing was maintained or improved. FN function was spared in all 35 cases, and to date, no patient has demonstrated any sign of recurrent disease. In our study, the tympanic membrane, the ossicular chain, and the FN function were preserved in all cases. Pensak and Jackler (9) describe the creation of a fallopian bridge with preservation of the FN while removing tumors that arise within or juxtaposed to the jugular fossa reporting on 35 patients, 13 of which had glomus jugulare tumors. The vertical segment of the FN was skeletonized from the stylomastoid foramen to the second genu. An extended facial-recess exposure was performed. Bone lying anterior to the FN was burred down, hugging the anterior surface of the fallopian canal to optimize visualization of the posterior mesotympanum and hypotympanum. Perilabyrinthine cells were widely exenterated in both the infralabyrinthine space and the retrofacial

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

HYPOTYMPANOTOMY FOR GLOMUS TYMPANICUM air-cell system. In the authors_ experience, in selected cases, leaving the FN in situ has neither compromised the extent of tumor resection nor increased the liability toward neurologic injury. OLeary et al. (10) reported on 64 patients. Surgical approaches were classified as either transcanal or transmastoid. One fifth of the cases were small enough for transcanal removal. Approximately 70% of these tumors were removed transmeatally and 30% via a postauricular incision. Lesions in this latter group usually required increased access anteriorly. The remaining 80% of cases required a transmastoid/facial recess approach. A radical mastoidectomy with removal of the posterior canal wall was performed in 3 patients. The preoperative mean airbone gap of 10 dB was reduced to a mean gap of 4 dB postoperatively. This improvement in conductive hearing was accompanied by a slight worsening of bone conduction postoperatively, with a rise in the mean bone PTA from 24 to 32 dB. Our method, on the contrary, reached substantial hearing improvement at 2,000 and 3,000 Hz in air conduction, a slight gain within random variation for bone conduction with a marginally significant improvement at 3,000 Hz, and a significant decrease of the air-bone gap after surgery at 3,000 Hz. Three tympanic membrane perforations failed to heal and required secondary tympanoplasty. A cholesteatoma developed postoperatively in 1 patient and was treated by revision mastoidectomy. In another case, an FN weakness noted postoperatively prompted reexploration and FN decompression. The patient experienced a mild (IINI) permanent weakness. In our study, as already mentioned, such complications did not occur. Gjuric et al. (11) report on 11 glomus tympanicum tumors. All patients were treated surgically with apparent complete removal of the tumor. The surgical approach was either the transcanal (endaural or retroauricular) or the classic retroauricular transmastoid approach. The transcanal approach (performed on 7 patients) was reserved for tumors confined to the tympanic cavity, and it included resection of the floor of the external auditory meatus for visualization of the tumor margins. If necessary, the FN was exposed in its mastoid portion, and the dome of the jugular bulb and the internal carotid artery were inspected. The evaluation of hearing was available for 10 patients. In 4 instances, the ossicular chain was temporarily disarticulated, for exposure purposes, and in 1 patient, a canal wall down procedure was performed. Three patients required a second-stage tympanoplasty. The air-bone gap closed postoperatively to within 10 dB in 3 patients, to within 20 dB in 6 patients, and to more than 30 dB in 1 patient (average of 0.5, 1, 2, and 4 kHz). Two minor complications occurred. One patient had a temporary facial palsy, and another patient had a persistent defect of the external meatal wall, requiring closure. Comparing with our study, additionally to the 2 minor complications, there are no detailed, statistically analyzed results regarding air and bone conduction and air-bone gap, and the 4 kHz frequency was preferred instead of 3 kHz.

295

There are 4 reports in the literature of using lasers for GTT removal totaling 13 patients. Durvasula et al. (12) used a Diode or KTP laser. Postoperatively, none of the 9 patients had sensorineural hearing loss or neural deficit; there was no recurrence but an increase in the air-bone gap of 10 dB was observed in 2 patients. Molony et al. (13) also used the KTP laser, whereas Robinson et al. (14) used the Nd YAG laser, and Kouzaki et al. (15) used a potassium titanyl phosphate laser, the feeding vessels having been embolized the previous day. In the authors_ experience, laser excision can be performed with minimal bleeding and morbidity, and it provided excellent tumor control in glomus tympanicum surgery to shrink and coagulate the tumor progressively with minimal hemorrhage, avoiding ossicular disarticulation and, sometimes, the need for extended facial recess or hypotympanotomy surgery. According to Robinson et al. (14), care should be taken to avoid accidental energy transmission to the cochlea. However, the same authors report sensorineural hearing loss of 20 dB after the excision of the tumor. Salami et al. (16) used a piezoelectric device in 10 patients affected by type A glomus tympanicum tumors (Fisch classification). The piezoelectric device confirmed its safety on soft tissues; during unintentional contact with the mass, no side effects were detected. Postoperatively, all 10 patients had no evidence of sensorineural hearing loss, neural deficit, or recurrence. In all the patients, 1 month, 6 months, and 1 year after surgery and at the end point, pure tone audiometry showed a mean hearing improvement of at least 16.4 dB of audibility threshold. Blood loss was not an issue, neither in the present study nor in any of the other series referred to above. Only OLeary et al. (10)(p1039) reported a significant intraoperative blood loss in their series of 64 surgically removed tumors. The remaining air-bone gap between 9 and 3 dB in our patients was possibly related to distortion of the drum or enlargement of the tympanic ring inferiorly. As the middle ear ventilation was not affected in none of the cases, Eustachian tube function was not impaired.

CONCLUSION We have demonstrated that our hypotympanotomy approach for removal of glomus tympanicum tumors modified from Shambaugh_s description provided for substantial hearing improvement at 2,000 and 3,000 Hz in air conduction, a slight gain within random variation for bone conduction with a marginally significant improvement at 3,000 Hz, and a significant decrease of the airbone gap after surgery at 3,000 Hz. This approach allowed for a safe approach for complete tumor removal without interference of the continuity of the ossicular chain while providing excellent exposure of the jugular bulb, the carotid artery, the protympanum, and even the bony part of the Eustachian tube.
Otology & Neurotology, Vol. 32, No. 2, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

296 REFERENCES

K. PAPASPYROU ET AL.
9. Pensak ML, Jackler RK. Removal of jugular foramen tumors: the fallopian bridge technique. Otolaryngol Head Neck Surg 1997; 117:586Y91. 10. OLeary MJ, Shelton C, Giddings NA, Kwartler J, Brackmann DE. Glomus tympanicum tumors: a clinical perspective. Laryngoscope 1991;101:1038Y43. 11. Gjuric M, Seidinger L, Wigand ME. Long-term results of surgery for temporal bone paraganglioma. Skull Base Surg 1996;6: 147Y52. 12. Durvasula VS, De R, Baguley DM, Moffat DA. Laser excision of glomus tympanicum tumours: long-term results. Eur Arch Otorhinolaryngol 2005;262:325Y7. 13. Molony NC, Salto-Tellez M, Grant WE. KTP laser assisted excision of glomus tympanicum. J Laryngol Otol 1998;112:956Y8. 14. Robinson PJ, Grant HR, Bown SG. Nd YAG laser treatment of a glomus tympanicum tumour. J Laryngol Otol 1993;107:236Y7. 15. Kouzaki H, Fukui J, Shimizu T. Management of a catecholaminesecreting tympanicum glomus tumour: case report. J Laryngol Otol 2008;122:1377Y80. 16. Salami A, Dellepiane M, Proto E, Mora R. Piezosurgery in otologic surgery: four years of experience. Otolaryngol Head Neck Surg 2009;140:412Y8.

1. Shambaugh GE Jr. Surgical approach for so-called glomus jugulare tumors of the middle ear. Laryngoscope 1955;65:185Y98. 2. Rosen S. Glomus jugulare tumor of middle ear with normal drum; improved biopsy technique. Ann Otol Rhinol Laryngol 1952;61: 448Y51. 3. Brown LA. Glomus Jugulare of the Middle Ear, quoted in Shambaugh GE Jr. Surgical approach for so-called glomus jugulare tumors of the middle ear. Laryngoscope 1955;65:185Y98. 4. Papaspyrou K, Rossmann H, Fottner C, et al. Malignant paraganglioma caused by a novel germline mutation of the succinate dehydrogenase D-gene-a case report. Head Neck 2008;30:964Y9. 5. Papaspyrou K, Mann WJ, Amedee RG. Management of head and neck paragangliomas: review of 120 patients. Head Neck 2009;31:381Y7. 6. Farrior JB. Glomus tumors. Postauricular hypotympanotomy and hypotympanoplasty. Arch Otolaryngol 1967;86:367Y73. 7. House WF, Glasscock ME 3rd. Glomus tympanicum tumors. Arch Otolaryngol 1968;87:550Y4. 8. Glasscock ME 3rd, Jackson CG, Dickins JR, Wiet RJ. Panel discussion: glomus jugulare tumors of the temporal bone. The surgical management of glomus tumors. Laryngoscope 1979;89:1640Y54.

Otology & Neurotology, Vol. 32, No. 2, 2011

Copyright 2011 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.