The Best Candidates for Nerve-Sparing Stripping Surgery for Facial Nerve Schwannoma

Transcription

1 The Laryngoscope VC 2014 The American Laryngological, Rhinological and Otological Society, Inc. The Best Candidates for Nerve-Sparing Stripping Surgery for Facial Nerve Schwannoma Soon H. Park, MD; Jin Kim, MD, PhD; In S. Moon, MD, PhD; Won S. Lee, MD, PhD Objectives/Hypothesis: Clinical decision making for facial nerve schwannoma is particularly complicated in patients with good facial nerve function; however, an early nerve-sparing tumor resection stripping technique minimizes facial deficits associated with treatment. The present study characterized the optimal candidate for this nerve-sparing surgical strategy in patients with good facial function. Study Design: Retrospective study. Methods: Nerve-sparing stripping surgery was performed on 28 patients with facial nerve schwannoma. The House- Brackmann grading system was used to assess pre- and postoperative facial function. We retrospectively analyzed pre- and postoperative facial function, duration of facial palsy, tumor size, and location and number of involved segments. The data were analyzed using Fisher exact test and independent t tests. Results: Of the 28 patients, 18 successfully underwent stripping surgery and 16 had a favorable outcome. Favorable postoperative facial function was associated with good preoperative facial function (House-Brackmann grade [HBG] II); small, localized tumors; and tumors located in the geniculate ganglion and/or its proximal portion. Conclusions: Patients with facial nerve schwannoma who have good preoperative facial function (HBG 2), tumor located in the proximal portion of the geniculate ganglion, and small tumors (<2 cm) involving one or two facial nerve segments can be the best candidates for nerve-sparing stripping surgery. Key Words: Facial nerve, schwannoma, facial palsy. Level of Evidence: 4 Laryngoscope, 124: , 2014 INTRODUCTION Facial nerve schwannomas (FNSs) are rare, benign tumors of the facial nerve that originate in the myelinproducing Schwann cell sheath. 1 The estimated prevalence of FNS is approximately 0.15% to 0.8%. 2,3 FNS can involve any segment of the facial nerve from the cerebellopontine angle to the peripheral branches. All FNSs eventually affect facial nerve function and present most frequently as facial weakness due to interruption of the motor component. 4 The management of FNS is a delicate and critical process, because 50% of patients have normal facial function when they are diagnosed. 4 Clinical decision-making is particularly complicated in patients with good facial nerve function. The primary treatment From the Department of Otorhinolaryngology (S.H.P.), Keimyung University College of Medicine, Daegu; Department of Otorhinolaryngology (J.K.), Inje University College of Medicine, Ilsn Paik Hospital, Goyang; and the Department of Otorhinolaryngology (I.S.M., W.S.L.), Yonsei University College of Medicine, Seoul, Korea. Editor s Note: This Manuscript was accepted for publication June 5, This work was supported by a grant of the Korea Health technology R&D Project, Ministry of Health & Welfare, Republic of Korea (A102065). The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Won Sang Lee, MD, Professor, Department of Otorhinolaryngology, Yonsei University College of Medicine, 134 Sinchon-dong, Seodaemun-gu, Seoul , Korea. wsleemd@yuhs.ac DOI: /lary options for the management of FNS include surgical intervention, observation, and radiotherapy. Surgical resection with facial nerve repair is the standard therapy for patients whose facial nerve function is House- Brackmann grade (HBG) III or worse. 5 The majority of clinicians recommend observation with periodic examination and imaging for patients who have good facial nerve function (HBG I or II), 6,7 because resection and reconstruction cause some degree of permanent facial deficit that may never be better than HBG III. 4,8,9 Pulec 10 described the first nerve-preservation surgery for FNS in 1972, and since then, more than 90 cases of facial nerve-preservation surgery have been reported worldwide. However, most were sporadically reported and no standard guidelines exist for nervepreserving surgery. In 1995, we changed our surgical policy from segmental resection to preservation of the neural fascicles by removing the FNS using an advanced microscopic surgical technique referred to as stripping surgery 11 and adopted this nerve-preserving technique as our primary treatment for FNS. Unlike other nerve-sparing techniques such as debulking, partial removal, or decompression, nerve-sparing stripping can completely remove the tumor while preserving the normal facial fascicles. Our results using this technique have been reported previously However, not all surgeries were successful, and it was necessary to sacrifice the facial nerve and perform additional reconstructive surgery in some cases. Moreover, in some cases, the facial nerve was

2 TABLE I. Patients Demographic Data. Favorable Group (n 5 16) Unfavorable Group (n 5 12) P Value Age, yr Gender 11:5 3:9.054 (male:female) Site (left:right) 7:9 4:8.705 Duration of symptoms (mo) Size of tumor (cm) Follow-up period (mo) anatomically preserved, but the functional results were not satisfactory. The present study characterized the optimal candidate for nerve-sparing stripping surgery among patients with FNS and suggested guidelines for surgeons who perform the surgery. MATERIALS AND METHODS With approval from the institutional review board of Yonsei University Medical Center (No ), we investigated the medical records of patients with FNSs managed at Yonsei University Health System, Severance Hospital between 1995 and We routinely use the stripping technique for FNS; however, we performed surgical resection and reconstruction on patients with preoperative HBG IV, V, VI or whose facial nerve could not be dissected from the tumor. We identified 32 patients with FNSs. Of those whose tumors were located at the cerebellopontine angle and preoperatively misdiagnosed as vestibular schwannomas, in seven patients they were confirmed as FNSs intraoperatively. We explained the surgical risk of facial paralysis and provided counseling, and three of the seven underwent stripping surgery with their patrons consent. The patrons of four patients refused invasive surgery, and we were only able to perform minimally invasive surgery without injuring the intact facial fascicles. These four patients were excluded from the study, three in which the tumor was partially removed and one in which the tumor was not removed during surgery. Thus, 28 patients were enrolled in the study. The nerve stripping surgery was performed in 18 patients, and excision and/or reconstruction was done in 10 patients. The HBG system for facial nerve function was used to assess pre- and postoperative function. Final postoperative facial function was evaluated at least 6 months post-surgery. We divided patients into two groups according to their final facial function outcome: the favorable group and unfavorable group. The favorable group comprised patients whose facial function was preserved or deteriorate less than one grade, and the final facial outcome was HBG I or II. The unfavorable group included patients whose facial function deteriorated more than two grades, and the final outcome was HBG III or higher. The statistical analysis included preoperative facial function, duration of facial palsy, tumor size, location, and number of facial nerve segments involved in the lesion. Serial magnetic resonance imaging (MRI) was obtained every 12 months for 2 years and then performed two or three times over a 10-year period to assess tumor recurrence or regrowth. Significant differences between groups were identified using a power analysis, Fisher exact test, and independent t tests. A P value <.05 was deemed statistically significant. Statistical tests were conducted using the Statistical Package for the Social Sciences version 16.0 (SPSS, Inc., Chicago, IL). Nerve-Preserving Stripping Surgery The nerve-preserving stripping surgery technique has been described in detail previously. 12 Briefly, the schwannoma was removed from the remaining facial nerve fascicle under a microscopic surgical field. Facial nerve monitoring (NIM- Response 2.0 Nerve Integrity Monitoring System; Medtronic Xomed Surgical Products, Inc., Jacksonville, FL) was used to identify and confirm facial nerve function intraoperatively. RESULTS The study included 14 males and 14 females, with a mean age of years (range, years). The schwannoma was on the right in 11 cases and on the left in 17 cases. The mean largest tumor diameter was cm (range, cm). Facial palsy was present in 16 (57%) patients, nine (32%) had, five (18%) had vertigo, and tinnitus was present in four (14%) patients. The surgical approach depended on the location of the tumor: the transmastoid approach was used in 14 (50%) patients, the middle cranial fossa, translabyrinthine, and combined approaches were used in four (14%) patients each, and the transparotid and transcochlear approaches were used in one patient each. The final outcome was favorable in 16 patients and unfavorable in 12 (Table I). Age, sex, tumor site, and follow-up period were not significantly different between groups. Patient clinical characteristics according to facial function outcome are shown in Tables II and III. The most common symptom among patients in the favorable group was (6/16, 38%), and facial weakness was the most common symptom in the unfavorable group (10/12, 83%). All patients in the favorable group and two in the unfavorable group underwent nerve-preserving stripping surgery. Nine patients in the unfavorable group underwent facial nerve resection and reconstruction. Various reconstruction methods including end-to-end anastomosis, hypoglossal-facial nerve anastomosis, and nerve graft were conducted after tumor resection. The tumor was excised in one patient with no reconstruction. Preoperative Facial Function At the time of diagnosis, 12 (43%) patients had normal facial function and 16 (57%) had facial paralysis. Significantly more patients in the unfavorable group (10/ 12, 83%) had any grade of facial nerve paralysis prior to surgery than in the favorable group (6/16, 31%) (P 5.023), and the mean duration of facial weakness was significantly longer in the unfavorable ( months) compared with the favorable group ( months; P 5.040). 2611

3 TABLE II. Summary of Patients in the Favorable Group. Facial Function Patient No. Symptoms Size (cm) Approach Surgical Technique Location of Tumor Preoperative Postoperative 1 Vertigo, tinnitus , MCF Stripping IAC, GG, L, T Hearing loss TM Stripping T, M Facial palsy, facial spasm TM Stripping GG, T Hearing loss TM Stripping GG, T Facial palsy MCF 1 TM Stripping GG, T, M EAC mass TM Stripping M Facial palsy TM Stripping GG Facial palsy, TM Stripping GG, T Tinnitus, facial palsy TM Stripping M Infra-auricular mass TM1 TP Stripping M, ET Hearing loss TL Stripping CPA, IAC Hearing loss MCF Stripping CPA, IAC Facial palsy, vertigo MCF Stripping IAC, T , vertigo TL Stripping IAC Tinnitus, vertigo TL Stripping IAC Tinnitus, vertigo MCF Stripping GG, T 1 1 CPA 5 cerebellopontine angle; EAC 5 external auditory canal; ET, extratemporal segment; GG 5 geniculate ganglion; IAC 5 internal auditory canal; L 5 labyrinthine segment; M 5 mastoid segment; MCF 5 middle cranial fossa approach; T 5 tympanic segment; TL 5 translabyrinthine approach; TM 5 transmastoid approach; TP 5 transparotid approach. Overall pre- and postoperative facial function is shown in Figure 1. All patients in the favorable group had preoperative HBG I or II, with the exception of one who had HBG III. All patients in the unfavorable group were HBG III, with the exception of two patients who were HBG I. Significantly more patients in the favorable group (15/16, 94%) had preoperative HBG scores of I and II than did those in the unfavorable group (2/12, 17%; P 5.001). Tumor Size The largest diameter of the tumor, determined using linear and planimetric measurements, was used to TABLE III. Summary of Patients in the Unfavorable Group. Patient No. Symptoms Size (cm) Approach Surgical Technique Location of Tumor Preoperative Facial Function Postoperative 1 Facial palsy TM Excision with sural n graft GG, T, M Facial palsy, TC Incomplete removal with XII-VI anastomosis CPA, IAC, L, GG Facial palsy TM Excision with XII-VI anastomosis GG, T, M Facial palsy TM Stripping M Ear fullness, TM Excision with sural n graft T, M Facial palsy, TL Excision without reconstruction T Parotid mass TP End to end anastomosis ET Facial palsy MCF 1 TM Excision c greater auricular n graft IAC, L, GG, T Facial palsy TM Excision with sural n graft T, M Facial palsy TM Excision with sural n graft GG, T Facial palsy, MCF 1 TM Excision with sural n graft GG, T, M Facial palsy TM Stripping T, M 6 6 CPA 5 cerebellopontine angle; ET 5extratemporal segment; IAC 5 internal auditory canal; GG 5 geniculate ganglion; L 5 labyrinthine segment; M 5 mastoid segment; MCF 5 middle cranial fossa approach; T 5 tympanic segment; TC 5 transcochlear approach; TL 5 translabyrinthine approach; TM 5 transmastoid approach; TP 5 transparotid approach. 2612

4 Tumor Location We divided tumors into three groups according to location within the geniculate ganglion (GG) (Fig. 3): 1) tumors involving the GG and/or its proximal portion (n 5 9; seven [78%] from the favorable and two [22%] from the unfavorable group, 2) tumors that involved the GG and its distal portion (n 5 9; five [56%] in the favorable and four [44%] in the unfavorable group, and 3) tumors involving only the distal GG (n 5 10; four [40%] in the favorable and six [60%] in the unfavorable group. Tumors that involved the GG and/or proximal portion had the highest percentage of favorable outcomes (78%), whereas tumors that involved only the distal portion of the GG had the lowest (40%) (Fig. 3). However, the difference according to location did not reach statistical significance (P >.05). Fig. 1. Overall facial function outcomes in the favorable and unfavorable groups according to the House-Brackmann grade (HBG). A favorable outcome was achieved only in patients with good facial function (HBG 3) prior to surgery, and all of those patients underwent stripping surgery. The preoperative (Preop.) HBG was variable in the unfavorable group, and facial function was normal in two patients; however, their final outcome deteriorated to HBG IV. Postop. 5 postoperative. evaluate the relationship between tumor size and facial function outcome. The mean tumor size in the favorable group was smaller than that of the unfavorable group ( vs ; P >.05; Table I). Tumors <2 cm in diameter were more frequent in the in favorable (12/16; 75%) than in the unfavorable (7/12; 58%) group; however, the difference was not statistically significant (P >.05). Preoperative Diagnosis We were able to identify the origin of the FNSs in most patients according to their facial weakness and location of the tumor. However, the tumor was preoperatively misdiagnosed as vestibular schwannoma in three patients. All had normal facial function and the tumor was located in the cerebellopontine angle or internal auditory canal. These patients had favorable facial function immediately after the surgery; however, it deteriorated from HBG I to HBG II. Final Facial Function Outcome Figure 1 shows pre- and postoperative facial function according to HBG. Favorable facial function was achieved only in patients who underwent nervepreserving stripping surgery. In addition, in the favorable group, all but one patient had HBG I or II preoperatively (one patient was HBG III). However, two patients who underwent the nerve-preserving surgery had an unfavorable outcome: one who was HBG III prior to surgery deteriorated to IV postsurgery, and the other was HBG VI before surgery and remained so after the surgery. Number of Segments Involved We subdivided the facial nerve into the cerebellopontine angle; internal auditory canal; labyrinthine; geniculate ganglion; and tympanic, mastoid, and extratemporal segments. In the favorable group, the lesion involved one facial nerve segment in five (31%) patients, two segments in nine (56%) cases, and more than three segments in two patients (13%). In the unfavorable group, the lesion involved one segment in three (25%) patients, two segments in four (33%) patients, and more than three segments in five (42%) patients. The percentage of patients with involvement of one or two segments was higher in the favorable than in the unfavorable group (87% vs. 58%, respectively); however, the difference did not reach statistical significance (P >.05, Fig. 2). Tumor Recurrence The mean follow-up period was months (range, months). Twenty-six patients underwent postoperative enhanced MRI, and to date, no clinical or radiological evidence of recurrence has been detected, with the exception of one patient in the unfavorable group whose tumor was partially removed (patient 2). This patient underwent additional gamma knife surgery 2 years after the initial surgery. Most of the between-group differences in the present study did not reach statistical significance, with the exception of preoperative facial function and duration of facial palsy. FNS is a rare disease, and although the present study included all the FNS patients we have treated over 18 years, the number of patients was not sufficient to reach statistical significance. DISCUSSION The goals of FNS surgery are to completely remove the tumor and preserve facial nerve function. Because FNS is a slow-growing tumor with benign characteristics, Liu and Fagan 7 recommended observation and serial radiological imaging treatments for FNS with good facial function. Angeli and Brackmann 6 reported 2613

5 Fig. 2. Number of facial nerve segments involved in the lesion according to facial function outcome. The percentage of patients with involvement of one or two segments was higher in the favorable (14/16, 87%) than in the unfavorable (7/12, 58%) group. seg 5 segment. [Color figure can be viewed in the online issue, which is available at com.] that wide decompression is an alternative to surgical resection. Observation and wide decompression are treatment options for patients with FNS who have good facial function. However, the remaining tumor eventually grows and facial weakness may occur. Surgery following facial nerve paralysis generally results in an unsatisfactory outcome, because long-term progressive changes occur. Stripping surgery is the most suitable technique for FNS excision in patients with good facial function because it is the only method that completely removes the tumor and preserves facial function. Our findings suggest that the most important factor for a successful outcome is preoperative facial nerve function. With the exception of one patient whose facial function improved from HBG III to II after surgery, the preoperative facial function of all patients in the favorable group was HBG I or II. The HBG increases with the duration of facial paralysis. It is possible to dissect the tumor from the facial nerve in cases with good facial function because neural compression is minimal. 11 Kim et al. 14 reported that postoperative results were better when preoperative facial nerve function was HBG III or IV than when it was severe (HBG V or VI). Alternatively, preoperative facial paralysis duration has been proposed to be the most important determinant of postoperative facial function. 15,16 Long-duration facial paralysis causes changes in the facial nerve nucleus in the pons and degeneration of the distal axons. 6,17 We found that facial palsy duration was longer in the unfavorable than in the favorable group. Chung et al. 18 argued that tumor size is a predictive factor for postoperative facial nerve function. Bacciu et al. 5 found that in large intraparotid tumors, the nerve fibers were often located within the tumor mass, thereby preventing surgeons from finding a plane between the tumor and the facial nerve, and the authors were unable Fig. 3. Illustration of tumor classifications according to location within the geniculate ganglion. Tumor involving the geniculate ganglion and/ or its proximal portion (A), tumor involving the geniculate ganglion and its distal portion (B), and tumor involving only the distal portion of the geniculate ganglion. Favorable outcomes according to the location of the facial nerve schwannoma. Tumors that involved the geniculate ganglion and/or its proximal portion had the highest rate of favorable outcomes. Lesions that involved only the distal portion of the GG had the most unfavorable outcomes. CPA 5 cerebellopontine angle; ET 5extratemporal segment; GG 5 geniculate ganglion; IAC 5 internal auditory canal; L 5 labyrinthine segment; M 5 mastoid segment; T 5 tympanic segment. *Rate of favorable outcome. [Color figure can be viewed in the online issue, which is available at

6 to separate the tumor from the perineurium of the facial nerve. In our cases, the largest diameter of the tumor was smaller in the favorable than in the unfavorable group; however, the difference was not statistically significant. Furthermore, stripping surgery performed on tumors 50 mm in diameter, which involved the mastoid and extratemporal facial nerve segments, resulted in a favorable outcome (HBG II). Thus, it is unlikely that tumor size is strongly related to the success of stripping surgery. However, the favorable group had a higher percentage of small tumors (<20 mm) than the unfavorable group; thus, we cautiously suggest that tumors <20 mm may be better candidates for nerve stripping surgery than larger tumors. Wilkinson et al. 9 reported that the more facial nerve segments involved in the lesion, the worse the facial nerve grade. Similarly, we found the percentage of lesions involving one or two segments was higher in the favorable than in the unfavorable group (87 vs. 58%, respectively). Moreover, stripping surgery was difficult to perform on tumors involving more than three segments. Our findings suggest that tumor location has an impact on the success of stripping surgery. Bacciu et al. 5 were unable to separate the tumor from the nerve fascicle in FNS with a large intraparotid tumor. Mowry and Ganz 19 investigated 11 cases of FNSs located in the proximal portion of the GG that underwent tumor-debulking surgery. Nine of the eleven patients had a favorable final outcome (HBG I or II). In our study, tumors located in the GG and/or its proximal portion had a higher rate of favorable recovery (78%) than those that involved the distal GG (40%). Two patients in our study had normal facial function; however, the tumor was located in the distal GG, and we were unable to dissect the facial nerve from the tumor. Of the 18 patients who showed good facial function preoperatively, two (11%) had unfavorable outcomes. These patients had tumors located in the distal portion of the GG (from HBG I to IV). This result also suggests that the proximal portion of the facial nerve is a better candidate for the stripping surgery. Patients with distally located tumors with good facial function should be provided a full explanation of the risks before the surgery. Otherwise, watchful waiting or decompression surgery can be recommended for these patients. Captier et al. 20 reported that the proximal portion of the GG lacked a real fascicular organization, and the perineurium and epineurium were absent in this portion of the ganglion. The authors observed the first real fascicular organization in the distal GG, and it increased in a proximal-to-distal manner, and the diameter of the fascicles diminished. The nerve is separated into fascicles distal to the GG. The tumors originating from the proximal portion of the GG are likely to arise from outer portion of the facial nerve bundle, and the relative ease with which the cleavage line can be detected may give rise to better results. In contrast, because the nerve divides into fascicles in the distal portion of the GG, the tumor can originate from any portion of the facial nerve. In tumors originating from the inner portion of the facial nerve bundle, the intact nerve fascicles may encase the tumor, making it more difficult to find the cleavage line and preserve facial function. Based on microanatomical structures, FNSs located in the proximal portion of the GG are likely to be good candidates for stripping surgery. In the present study, the tumor recurred in one patient in the unfavorable group who underwent partial removal and reconstruction. The tumor did not recur in any patient who underwent stripping surgery. The recurrence rate after stripping surgery does not appear to be higher than that for other surgical methods. CONCLUSION Our findings suggest that the factors associated with a favorable outcome for FNS following stripping surgery are good preoperative facial function, tumor located in the proximal portion of the GG, and small tumors (<2 cm) involving one or two facial nerve segments. Patients with these characteristics are good candidates for stripping surgery, and surgeons should consider this technique as the primary treatment for patients meeting these criteria. BIBLIOGRAPHY 1. Shirazi MA, Leonetti JP, Marzo SJ, Anderson DE. Surgical management of facial neuromas: lessons learned. Otol Neurotol 2007;28: Pulec JL. Facial nerve neuroma. Ear Nose Throat J 1994;73: , , Saito H, Baxter A. Undiagnosed intratemporal facial nerve neurilemomas. Arch Otolaryngol 1972;95: McMonagle B, Al-Sanosi A, Croxson G, Fagan P. Facial schwannoma: results of a large case series and review. J Laryngol Otol 2008;122: Bacciu A, Nusier A, Lauda L, Falcioni M, Russo A, Sanna M. Are the current treatment strategies for facial nerve schwannoma appropriate also for complex cases? Audiol Neurootol 2013;18: Angeli SI, Brackmann DE. Is surgical excision of facial nerve schwannomas always indicated? Otolaryngol Head Neck Surg 1997;117:S144 S Liu R, Fagan P. Facial nerve schwannoma: surgical excision versus conservative management. Ann Otol Rhinol Laryngol 2001;110: Falcioni M, Russo A, Taibah A, Sanna M. Facial nerve tumors. Otol Neurotol 2003;24: Wilkinson EP, Hoa M, Slattery WH III, et al. Evolution in the management of facial nerve schwannoma. Laryngoscope 2011;121: Pulec JL. Facial nerve neuroma. Laryngoscope 1972;82: Lee JD, Kim SH, Song MH, Lee HK, Lee WS. Management of facial nerve schwannoma in patients with favorable facial function. Laryngoscope 2007;117: Lee WS, Kim J. Revised surgical strategy to preserve facial function after resection of facial nerve schwannoma. Otol Neurotol 2011;32: Park HY, Kim SH, Son EJ, Lee HK, Lee WS. Intracanalicular facial nerve schwannoma. Otol Neurotol 2007;28: Kim CS, Chang SO, Oh SH, Ahn SH, Hwang CH, Lee HJ. Management of intratemporal facial nerve schwannoma. Otol Neurotol 2003;24: O Donoghue GM, Brackmann DE, House JW, Jackler RK. Neuromas of the facial nerve. Am J Otol 1989;10: Yamaki T, Morimoto S, Ohtaki M, et al. Intracranial facial nerve neurinoma: surgical strategy of tumor removal and functional reconstruction. Surg Neurol 1998;49: King TT, Morrison AW. Primary facial nerve tumors within the skull. J Neurosurg 1990;72: Chung JW, Ahn JH, Kim JH, Nam SY, Kim CJ, Lee KS. Facial nerve schwannomas: different manifestations and outcomes. Surg Neurol 2004;62: ; discussion Mowry S, Hansen M, Gantz B. Surgical management of internal auditory canal and cerebellopontine angle facial nerve schwannoma. Otol Neurotol 2012;33: Captier G, Canovas F, Bonnel F, Seignarbieux F. Organization and microscopic anatomy of the adult human facial nerve: anatomical and histological basis for surgery. Plast Reconstr Surg 2005;115: