Vestibular schwannoma microsurgery with special reference to facial nerve preservation

1. Introduction 

Although vestibular schwannomas (VSs) are benign tumors characterized by slow growth, they may erode the internal auditory canal and compress adjacent cranial nerves. Total tumor removal and a good quality of life, defined as preservation of facial function and hearing, have become the accepted treatment goals and increased expectations have paralleled improvements in surgical techniques as well as the introduction of intraoperative facial nerve monitoring [1]. Anatomical and functional preservation of the facial nerve have become routine in experienced hands [2]. Although increasing rates of post-operative facial nerve preservation for VSs, paralysis remains a major concern. Advances in neuroimaging, intra-operative nerve monitoring, and microsurgical technique have transformed a once dangerous operation into a technique associated with low morbidity and mortality and have shifted the focus of VS surgery from prolonging life to preserving nerve function. Up-to-date reports cite approximately 90% of VS patients have normal or near-normal post-operative facial nerve function [3].

During the past 4 years, the retrosigmoid transmeatal approach and a standardized microsurgical technique have been employed in the resection of 103 VSs by the authors. In this study, we review our results and discuss various issues, including the different anatomical locations of the facial nerve in relation to the tumor capsule in the cerebellar-pontine angle (CPA), surgical technique, and ways to protect the facial nerve in patients diagnosed with a VS.

2. Clinical materials and methods 

3. Pre-operative examination 

A pre-operative clinical evaluation which included otorhinolaryngological investigations, CTs with bone windows, MRIs (Fig. 1A and B), functional X-rays of the cervical spine, and audiometric tests were performed in all patients. Anatomic relations between the labyrinth, jugular bulb, and the IAM could be evaluated pre- and post-operatively by CT. The location of the labyrinth was classified into three types based on its relationship with the sigmoid-fundus line (S-F line—a virtual line that connects the fundus of the IAM and the medial border of the sigmoid sinus). The lateral type in which both the vestibule and the common crus were lateral to the S-F line, the line type where the vestibule and/or common crus were on the S-F line, or the medial type characterized by the presence of both the vestibule or common crus was medial to the S-F line. If the posterior semi-circular canal and the crus commune were located lateral to S-F line, then there was no risk of injuring them. If the posterior semi-circular canal and the crus commune were medial to the S-F line, they were in danger of being opened. Pre-operative CT scans were used for anatomic landmark orientation. The extent of bone removal from the posterior wall of the IAM was evaluated by post-operative CTs.

View Large Image Download to PowerPoint

Fig. 1. This 42-year-old female presented with a 3-year history of dysaudia on the left side and a 2-month history of dysphagia. A 3.5cm diameter vestibular schwannoma was diagnosed. (A and B) Pre-operative axial and coronal contrast-enhanced T1-weighted MRI scans revealed a left CPA vestibular schwannoma. (C and D) Post-operative MRI confirmed total tumor resection with anatomic preservation of the facial nerve (H-B Grade II 3 months post-surgically).

4. Surgical technique 

4.1. Retrosigmoid craniotomy 

Fifty-eight cases were operated on in semi-seated position and the remaining 45 were in a lateral position. The craniotomoy was initiated via a lateral suboccipital retrosigmoid transmeatal approach. The advantages of the semi-sitting position are that it facilitates control of intra-cranial pressure by better drainage of CSF from the cisterns compared to the lateral position, and exposure is better. A part of the sub-occipital skin was marked by either a vertical, “C”-shape, or small “S”-shaped incision 6–7cm in length. A sub-occipital craniectomy was performed by extending the incision to the posterior margin of the sigmoid sinus and exposing the inferior margin of the transverse sinus. The surgeon could then explore and navigate the area to locate the transverse and retrosigmoid sinuses (Fig. 2B).

View Large Image Download to PowerPoint

Fig. 2. Navigational guidance. (A) Location of IAM during drilling of the posterior wall of the IAM to confirm adequate drilling. (B) Location of the transverse sinus.

5. Statistical analysis 

Mean, maximum, and minimum values were calculated and preservation of facial nerve function and measurement of tumor size were calculated. Statistical evaluation included Student’s t-test. P values less than 0.05 were considered significant.

6. Results 

It is the author’s opinion that facial nerve preservation and total tumor removal are more important than preserving the patient’s hearing. Nonetheless, all three goals (i.e., complete tumor resection and preservation of facial nerve function and hearing) were achieved in 13 of the 103 cases (12.6%). The mortality rate in our series was 0%. Based on the length of the remaining posterior wall of the IAM (determined by post-operative CT) patients were divided into the partially resected group (n=48) and the widely opened IAM group (n=55) where the mean length of the posterior wall at the fundus was reduced to 4.6±1.0 and 1.9±0.5mm, respectively. The location of the labyrinth (determined by pre-operative CT) was lateral in 52 cases, medial in 28 cases, and on the SF-line in 23 cases. Resection of the posterior wall of the IAM did not extend beyond the transverse crest.

7. Extent of tumor resection 

Of the 103 cases, visual inspection and post-operative MR imaging showed that in 101 cases (98.1%) complete resection was achieved without tumor recurrence (Fig. 1C and D). The remaining two patients underwent intentional subtotal tumor removal. One of these patients had NF-2 and subtotal removal was necessary to avoid facial nerve paralysis and preserve hearing in the only functional ear and the other case showed deleterious changes in vital signs during the operation when brain stem decompression was performed. Those two patients were both in Group III and had large tumors that severely compressed the brain stem. As such, we intentionally left a small part of the tumor capsule that was tightly adhered to the facial nerve to preserve facial nerve function. When the tumor adhered tightly to the facial and/or trigeminal nerve, any attempt to dissect the nerve from the tumor capsule led to intense electromyographic changes.

8. Relationship between tumor and nerve/blood vessel 

The most common location of the facial nerve was on the anterior middle third of the tumor capsule, regardless of tumor size. The facial nerve was on the anterior middle third of the tumor in 51 cases (49.5%), anterior superior aspect in 24 cases (23.3%), and the anterior inferior aspect of the tumor in 21 cases (20.4%) as described in Table 1. The facial nerve passed directly through the tumor proper in one patient so that the tumor infiltrated the nerve sheath and enfolded it completely. The most common location of the eighth cranial nerve complex was the anterior inferior portion of the tumor. Not surprisingly, 39 larger tumors (Group III) invaded the 5th cranial nerve in all patients, and the 9–11th cranial nerve complex in 28 cases. The anterior inferior cerebellar artery (AICA) trunk and its branches was the blood vessel most closely associated with VSs. The trunk of the AICA was adherent to the tumor capsule in one patient included in Group I, 13 patients in Group II, and 39 (i.e., 100%) patients in Group III. The AICA and/or its branches traversed through the tumor itself in 5 and 11 patients, in Groups II and III, respectively. The superior cerebellar artery (SCA) was also frequently involved with the tumor capsule in Group III patients (31/39, 79.5%). The main trunk of the posterior inferior cerebellar artery (PICA) was not involved in Group I tumors, but was adherent to tumor capsules in 13 patients in Group II and 31 patients in Group III. The PICA and/or its branches passed into the tumor only in patients with large tumors (i.e., 5 patients in Group III). In 32 patients in Group III, the vertebral artery (VA) was invaded by the tumor.

Table 1. Facial nerve location in the 103 patients with vestibular schwannoma.

	Tumor size	Anterior middle third	Anterior superior	Anterior inferior	Superior pole	Inferior pole	In tumor
	Group I	13	3	2	0	0	0
	Group II	18	13	13	1	1	0
	Group II	20	8	6	0	4	1
	Total	51	24	21	1	5	1

Data are presented as number of patients in each group.

9. Anatomical and functional facial nerve preservation 

As summarized in Table 2, the facial nerve was anatomically preserved in 101 of the 103 patients (98.1%). In both Groups I and II, anatomic preservation of the facial nerve was achieved in 100% of the patients. In Group III, the facial nerve was not preserved in two cases because the tumor destroyed the nerve along the anterior surface of the tumor and just medial to the porus. In both cases, an end-to-end facial nerve anastomosis was performed and satisfactory facial function of H-B Grade III recovery was noted 1 year after reconstruction. The tumors in these two cases both had cystic changes.

Table 2. Facial nerve function outcome in the 103 patients with vestibular schwannoma as determined by the House-Brackmann grading scheme.

	House-Brackmann grade	Immediately post-op.	At discharge	3 months post-op.	1 year post-op.
	I	14	13	15	16
	II	67	70	71	70
	III	15	14	12	13
	IV	6	5	5	4
	V	1	1	0	0
	VI	0	0	0	0

In terms of facial nerve function, most patients displayed adequate facial movements and no patient suffered complete facial nerve paralysis. Overall, at the time of the final follow-up examination, 83.5% patients had normal or near-normal facial nerve function (Grade I or II) 3–12 months after tumor resection. Facial nerve weakness (H-B Grades III–VI) was noted in 17 cases (16.5%). No patient showed deterioration in facial nerve function in the 3–12 months interval between discharge and last follow-up.

10. Tumor size and facial function 

Post-operative facial function was intimately correlated with tumor size. Evaluation by H-B grading revealed excellent results (Grade I or II) in 100% of the patients with small tumors within 3 months of operation. In the 46 patients with medium-sized tumors, 89.1% had Grade I or II function and in the 39 patients with large tumors, 69.2% had Grade I or II function (Table 3).

Table 3. Relationship between facial nerve function preservation and tumor size.

	Tumor size (cm)	Facial nerve function preservation (n)
		Pre-operative		Post-operative
		Grade I or II	Grade I or II	Grade III or IV	Grade VI or V
	Small (≤2.0)	18	18	0	0
	Medium (2.1–3.9)	46	41	5	0
	Large (≥4.0)	39	27	12	0
	Total	103	86	17	0

11. Post-operative complications 

The most frequent post-operative complications were transient headache (n=11) and dizziness (n=4). Cerebellar edema rarely observed (n=3) and in one case was related to petrosal vein sacrifice. A cerebral spinal fistula (CSF) occurred in one case. In one case the caudal cranial nerves exhibited a minimal amount of decreased function manifesting as transient dysphonia and dysphagia.

None of the patients had persisting headaches at the time of the follow-up examinations. Cerebellar edema was treated with diuretics and cortisone until the symptoms resolved. The CSF fistula was transient and resolved with lumbar spinal fluid drainage and direct compression applied to the surgery site.

12. Discussion 

The goal of VS surgery is to achieve complete tumor resection while preserving the function of various cranial nerves. With experienced surgeons, complete tumor removal is achievable in 80–99% of the cases [4]. In the last decade, VS surgery has witnessed a substantial improvement in mortality rate as well as in preservation of the functions of both the auditory and facial nerves. Large tumors remain a clinical challenge as they are associated with a worse outcome compared to smaller tumors, especially regarding facial function [5].

13. Selection of surgical approach in VSs 

Selecting the best surgical approach is primarily dependent on tumor size, its extension into the IAM, status of pre-operative hearing, and the surgeon’s experience and/or preference. Microsurgery and the retrosigmoid, transmeatal approach have become the standard procedure for the removal of VSs. For all VSs, regardless of size and hearing, we routinely select the retrosigmoid, transmeatal approach [6] which has proven particularly valuable in patients with large tumors. With this approach, there is frequently a layer of tissue between the tumor and the facial nerve, which makes dissection along the facial nerve safer than when the facial nerve is more directly exposed to the operative field.

The facial nerve was preserved in 98.1% of our cases. At discharge, 83.5% of the patients had facial nerve function of H-B Grade 1 or 2. After operating on 3200 patients, Gharabaghi suggests that the retrosigmoid approach affords a long-term tumor-free rate while minimizing complications of the facial nerve [7]. One problem associated with this approach, however, is that there is limited access to the IAM (the lateral end of the IAM is often not adequately visualized) which could potentially pose a risk of injury to the cochlear nerve and to the ES and ED during removal of the posterior meatal wall due to their close proximity to the IAM [9]. Ultimately, this limitation in exposure also increases the chance of incomplete tumor removal [8]. Pre-operative radiological evaluation of anatomic relationships is mandatory when preservation of hearing is desired.

14. Contribution of removal of the posterior meatal wall for VS resection 

Removal of the posterior wall of the IAM is an essential step in VS surgery to obtain maximal exposure and minimize morbidity associated with loss of hearing and balance. Drilling the posterior wall is important for obtaining an unobstructed view for sharp dissection and to identify the cochlear and vestibular nerves at the most lateral part of the meatus. The labyrinthine structures are known to exhibit great anatomic variability and opening of the IAM is performed based on the pre-operative CT findings and the experience of the surgeon. Yokoyama et al. [10] attempted to estimate the pre-operative risk associated with opening the inner ear structures via the retrosigmoid approach by introducing the S-F line, which is also known as the “safe lateral line.” As our data indicate, drilling the IAM is usually incomplete and the remaining posterior wall obstructs the view of the lateral aspect of the meatus neessitating blind tumor removal and, inevitably, tumor recurrence from remnant tissue. Our pre-operative evaluation discloses that the location of the labyrinth in relation to this line is variable and can also be classified into three types. When the labyrinth is medial to or on the S-F line, it is most susceptible to injury during drilling of the IAM. The external aperture of the vestibular aqueduct can be identified by high-resolution CT scans of the temporal bones which assists the surgeon to preserve the labyrinth when using the retrosigmoid transmeatal approach [11].

15. Completeness of tumor removal and long-term tumor control 

Total resection of VSs can be achieved in conjunction with low morbidity (associated with facial and cochlear nerve damage) and very low mortality rates [7]. The introduction of microsurgery has resulted in increasing the rate of precise anatomic preservation of the cranial nerves. As tumor size increases, however, facial nerve preservation is correspondingly difficult. In our series, small fragments of tumor remained in two cases each having tumors exceeding 4cm in diameter. We advocate that leaving a small amount of residual tumor on the facial nerve in case of severe adhesion or apparent infiltration; however, tumor recurrence is likely.

Historically, inadequate tumor exposure in the IAC was the most frequent cause of incomplete removal. In this series, the long-term tumor control rate was 98.1% following complete tumor resection. Only 2 of the 103 patients suffered tumor recurrence. According to Samii’s experience [12], there are three groups of patients who have a major risk of tumor recurrence: patients with severe brain stem compression; when the tumors involve the caudal cranial nerves, and if the tumor is cystic and therefore has a potential of causing post-operative hemorrhage.

For larger tumors, it may be desirable to debulk the tumor and stimulate the most proximal end of the facial nerve at 0.05mA to confirm its integrity. We believe that, with a thorough understanding of the pathophysiological mechanisms of cranial nerve injury, the anatomical variations in the location of the facial and cochlear nerves, and the application of microsurgical technique, there is no reason why surgeons cannot achieve preservation of cranial nerve integrity in the vast majority of patients. It is notable that the paucity of consistent landmarks can potentially lead to inadvertent injury of the cranial nerves and other important neurovascular structures [13].

16. Facial nerve preservation 

Exposure of the IAM by the retrosigmoid approach facilitates removal of VSs and preservation of facial nerve function. Anatomical preservation of the facial nerve is expected and is achieved in 93–99% of cases [14]. It is possible to obtain normal to near-normal facial function in up to 80% of the patients with large tumors by using standard operative techniques and approaches [6]. Nerve function, graded by the H-B system, should be assessed for up to 1 year post-operatively to allow for full facial nerve recovery.

Predictive factors for nerve preservation include tumor size and extension, tumor consistency (e.g., cystic vs. non-cystic), previous surgery or radiosurgery, and surgeon’s experience. When the VS is >4cm in diameter, adequate facial nerve function is achieved in 38–58% [15]. Patients who have small tumors with <20mm extension into the posterior fossa have achieved excellent outcomes: anatomical preservation of the facial nerve is 98.5% [16]. More specifically, 2 weeks after surgery, excellent facial nerve function (Grades I and II) is present in 59% of patients, good function (Grade III) in 16%, fair (Grade IV) in 17%, and poor or no function (Grades V and VI) in 8% [13]. Eighty percent of patients obtained normal to near-normal facial function (Grades I and II) in the larger VSs (measuring 3cm or larger in diameter) [6]. With the support of facial nerve monitoring, the rate of nerve preservation continues to increase.

Samii reported an overall anatomical nerve preservation rate of 98.5 and 100% in small- and middle-sized VSs [16]. Koos reported that full nerve function was maintained in 88% of patients (overall) and full (1005) anatomical preservation in patients with smaller tumors [13]. In the study described herein, two patients (corresponding to Grades III and IV) recovered facial motion following plastic surgery. Gormley WB [15] reported a 2% loss in anatomic continuity of the facial nerve. In our experience, an end-to-end facial nerve anastomosis produced an acceptable recovery of facial motion up to one year post-surgically suggesting that immediate end-to-end facial nerve anastomosis is a useful procedure. Together, these studies clearly show that tumor size plays a major role in facial nerve preservation. Evaluation of post-operative facial nerve function within 3 months post-resection revealed Grade I or II function in 100% of small tumors, 89% of medium-sized tumors, and 69.2% of large tumors (Table 3). If the tumor invades the sheath of the facial nerve, it is virtually impossible to dissect the tumor away from the nerve completely without causing considerable post-operative deficits. If the tumor enfolds the facial nerve, careful inspection should identify a plane on which the two surfaces are interposed.

In our series, one patient with a large tumor developed post-operative cerebellar edema related to superior petrosal vein sacrifice. Venous complications following the transsigmoid approach have not been widely addressed in the literature to date. Sacrifice of the petrosal vein during surgery does not appear to greatly impact post-operative auditory function [17]; however, Koerbel reported a venous complication following microvascular decompression for a classical right sided trigeminal neuralgia, characterized by peduncular hallucinosis. The petrosal vein and a transverse pontine vein were sacrificed during the procedure [18]. The importance of preserving the superior petrosal vein has been subject to increasing attention in the surgical treatment of tumors involving the CPA. The best policy in cerebellopontine angle surgery is to preserve the petrosal vein complex, especially the large caliber veins, whenever possible to minimize post-surgical complications and increase the safety of surgery. In cases of petrosal vein obliteration, complete tumor resection is essential to minimize the risk of cerebellar congestion and edema.

In conclusion, given the use of microsurgery and facial EMG monitoring, the basic goal of contemporary VS therapy should be complete tumor resection in conjunction with preservation of facial nerve function. In the majority of cases, VSs can successfully be resected with minimal recurrence, independent of tumor size. Generally, 83.5% of patients can expect normal or near-normal (Grade I or II) facial nerve function. By combining a microsurgical technique with continuous electrophysiological monitoring, we achieved anatomical preservation of the facial nerve in 98.1% of our patients (overall) and in 100% of patients with small- and medium-sized VSs.