Clinical Neurology and Neurosurgery 133 (2015) 83–89

Contents lists available at ScienceDirect

Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro

Intraoperative neuromonitoring for removal of large vestibular schwannoma: Facial nerve outcome and predictive factors Sheng-wen Liu a , Wei Jiang a,∗ , Hua-qiu Zhang a , Xiao-peng Li a , Xue-yan Wan a , Bosomah Emmanuel b , Kai Shu a , Jing-cao Chen a , Jian Chen a , Ting Lei a a b

Department of Neurosurgery, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China

a r t i c l e

i n f o

Article history: Received 23 October 2014 Received in revised form 27 February 2015 Accepted 21 March 2015 Available online 28 March 2015 Keywords: Facial nerve Intraoperative neuromonitoring Large vestibular schwannoma Microsurgery Predictive factors

a b s t r a c t Background: Preservation of facial nerve (FN) function is one of the major goals for resection of large vestibular schwannoma (VS) (≥30 mm). Little is known about the FN outcome and its predictive factors due to limited data. Objective: To explore the predictive factors affecting FN outcome following resection of large VS. Methods: 106 Large VS patients underwent surgical resection from 2010 to 2012 via intraoperative neuromonitoring for FN preservation approach. Postoperative FN function evaluation was conducted at the time points of 3–7th day, 3rd month and at the end of the 2nd year. Correlation between tumor size, intraoperative parameters and FN function were examined. Results: The ratios of total and subtotal resection were 82.1% and 14.2%, respectively. Acceptable FN function was achieved in 78% patients. Patients with good FN function showed much smaller (P < 0.01) VS size than those of poor-FN function patients at 3–7th day, 3rd month and 2nd year. There was a significant correlation between facial motor evoked potential (FMEP) ratios and postoperative FN function at 3–7th day (r = −0.709, P < 0.001) 3rd month (r = −0.709, P < 0.001) and 2nd year (r = −0.750, P < 0.001). Maximal response amplitude (MRA) ratio was a supplementary indicator for train time in predicting both immediate and long-term FN function in patients with large VS. Conclusion: Indicative factors of both immediate and long-term postoperative FN function in large VSs include tumor size, intraoperative train time, start to final FMEP ratios and proximal to distal MRA ratios. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Maintaining a functionally favorable facial nerve (FN) is one of the main objectives in management of vestibular schwannoma (VS) since the mortality rate is no longer a major concern for neurosurgeons because of significant reduction of this figure [1]. Most previously reported studies documenting acceptable facial nerve outcome mainly centered on patients with small tumors (diameter 0.05) of postoperative FN function was found between cystic (n = 6) and non-cystic VS (n = 100) (A), total (n = 87) and subtotal/partial (n = 19) removal VS (B). Size of vestibular schwannoma was larger in poor FN function group (C) (** P < 0.01).

analysis, train time was recorded and analyzed offline. Longer train time was an indicator of poor FN function at 3–7th day, 3rd month, and 2nd year (Fig. 3A). Spearman correlation analysis also proved the predictive effect of train time on FN function at 3–7th day (P < 0.01), 3rd month (P < 0.05), but not significant at 2nd year (Table 2).

The baseline FMEP amplitude was 109.1 ± 42.5 ␮V and the latency was 14.2 ± 1.1 ms. Pre to post-resection ratios of FMEP were higher in patients presenting with good FN function outcome than in those with poor (Grade 3–6) result at 3–7th day, 3rd month and 2nd year (Fig. 3B). Furthermore, a strong correlation (P < 0.001) between FMEP ratios and FN function was testified by Spearman correlation analysis at 3–7th day, 3rd month, and 2nd year (Table 2). The baseline MRA was 228.0 ± 39.0 ␮V. The proximal to distal ratios of the maximal response amplitudes were higher for patients with good FN outcome than those with poor outcome at 3–7th day, 3rd month, and 2nd year (Fig. 3C). However, the predictive effect was significant only at 2nd year in Spearman correlation analysis (Table 2).

S.-w. Liu et al. / Clinical Neurology and Neurosurgery 133 (2015) 83–89

3.4. Postoperative complications No death occurred among patients included in this study. Thirty patients complained of transient headache (n = 25) and dizziness (n = 14), all of which were relieved completely in 3-months follow up. No patient suffered from persistent hydrocephalus. Three patients had to undergo a second operation due to extradural hemorrhage (n = 1) or hemorrhage of the operation site where the tumor was located (n = 2). Five patients suffered from cerebrospinal rhinorrhea. Two of them resolved after one week of lumbar spinal fluid drainage. In three patients, the fistula lasted for more than 1 month, so muscle, fascia and bone wax was required to seal the opened posterior wall of the IAC for a second time. However, the HB grade of FN function did not change before or after the second surgery either for hemorrhage or rhinorrhea. 16 cases (15.1%) suffered from persistent tinnitus at last time of follow-up. 27 cases (25.5%) had unsteady gait and 54 cases (50.9%) had vertigo at 3rd month but all relieved by rehabilitation and medication 2 years after operation. 4. Discussion Factors as FN function, hearing and postoperative vertigo were found to be correlated with the long-term quality of life for patients with VS. Preservation of the anatomical and functional integrity of FN is one of the primary aims in microsurgical management of VS [2,14,15]. It is generally easy to maintain an intact facial nerve in the case of small VS. However, it is quite difficult to keep the facial nerve intact in the case of large VS, for which microsurgery still remains the main treatment modality [16]. Advancement in surgical techniques as well as the introduction of intraoperative FN monitoring has raised postoperative FN function preservation rate to 80% or even higher in small VS [2,17,18]. The major function of intraoperative facial EMG monitoring is to locate and ensure an intact facial nerve during VS dissection [19]. To our knowledge, most researches based either on cases with VS smaller than 30 mm or on mixed groups of patients with VS of any size. Only a few studies referring to the surgical result of FN function in exclusively large-VS patient population have been reported (Table 3). But the number of cases in these studies is limited. From this point of view, we reviewed the FN outcome of a relatively greater number of patients with large VS and discussed the related prognostic factors. Previous reports indicated that there was a significant correlation between size of VS and FN grade at both immediate and long term periods of postoperative follow-up [17,20]. Lalwani et al. [5] revealed that HB grade 1 or 2 FN function results was 100% for tumors ≤30 mm but 79% for tumors larger than 30 mm in 1 or more years after surgery. The predictive function was further studied by Fenton et al. [19] and demonstrated that the long-term FN function was strongly correlated with tumor size by a correlation factor of 0.47. In another study, Samii and colleagues made a comparison of FN function between tumors larger than 4 cm (giant) and tumors

87

smaller than 3.9 cm (2.6 cm in average). The result indicated that patients with giant tumor were less likely to achieve FN function of HB grade 1. No significant difference was found between two groups with grade 2 to 6 [21]. For exclusively large tumors, whether there is FN outcome difference between relatively large (3–4 cm) and giant (>4 cm) tumor is still unclear. In this series, size of VS for good-postoperative FN function patients was around 3–4 cm and poor function was about 4 cm or even larger (Fig. 2). Analysis revealed that for exclusively large tumors, size was also an indicator of poor FN function with a correlation factor even reaching 0.678. Anatomically, FN is slightly displaced by small VS due to slow enlargement of the tumor capsule, so it is reasonable that the FN function is correlated with the tumor size. For large VS, FN is often displaced to the ventral side of the tumor with strong adhesion or encapsulation by the tumor tissue. For large VS resection, one of the primary goals is to maintain FN integrity, which is thought to be the premise of good FN function [5]. In this study, the FNs of 98.1% patients were intact after tumor removal. Acceptable FN function was achieved in 37.8% of total cases 3–7 days after surgery and 78% remained good FN function after 2 years. The advantage of INOM is to lessen intraoperative FN damage and to get a higher total resection rate. One INOM parameter as prognostic factor for postoperative FN function is A-train activity, which is a distinct EMG waveform classified as a sinusoidal symmetric sequence of high-frequency and low-amplitude signal during intraoperative EMG monitoring [16]. The overall appearance of its pattern might vary in different cases. By calculating the train time, Prell et al. [22] found that train time exceeding 10 s was correlated with deterioration of post-surgical FN function by two or more grades immediately after surgery. In this study, long train time predicted poor facial function. We further analyzed and proved that the train time was also correlated with postoperative HB grade. But the correlation was not obvious at 2nd year. This may be due to limited number of cases but we can see that a short term of A-train activity is a sign of FN irritation and the FN deficit is reversible in early postoperative period, while persistent pattern reflects mild to moderate FN damage. Another factor is FMEP, which was recently documented to be correlated with postoperative FN function [17,23]. Because of its advantage of less invasivity, FMEP monitoring can be performed before and after operation to detect nerve integrity without direct visualization and it is thought to be a useful supplement for FN monitoring [17]. FMEP amplitude at the start of surgery was a good predictive factor for postoperative FN function. Besides, the amplitude ratio of beginning to end was correlated with long-term FN outcome [24]. Fukuda et al. [20,25] and Acioly et al. [17] found that a start-to-final baseline FMEP amplitude ratio of 50% was a predictive threshold for a poor FN outcome (HB Grade 3–6) both in short and long-term follow-up. In our series, final-to-start FMEP ratio in tumors ≥30 mm was strongly correlated with postoperative FN function. However, technique of FMEP recording is more complicated during dissection. Identification of its waveform is based on

Table 3 Literature review of studies on facial nerve function for large VSa . Author and Year

No. of cases

Size of VS (mm)

Rate of total resection (%)

Months of follow up

Final HB grade 1–2 (%)

IONM

Lalwani, 1994 [5] Fenton, 1999 [19] Raftopoulos, 2005 [8] Liu, 2007 [6] Chen, 2009 [4]

24 11 16 19 46 39 50 15

>30 >30 30–70 42–67 25–39 ≥40 40–65 >30

NA 100.0 68.7 63.2 100.0 94.9 100.0 NA

12 24 12 3–10 3

79.0 72.7 93.8 63.2 89.1 69.2 90.0 20.0

EMG EMG EMG TCMEP EMG

Samii, 2010 [21] Bernat, 2010 [27]

5–62 1

EMG EMG

a VS, vestibular schwannoma; HB grade, House–Brackmann grade; IONM: intraoperative neuromonitoring; TCMEP: motor evoked potentials monitoring with transcranial electrical stimulation; NA: not available.

88

S.-w. Liu et al. / Clinical Neurology and Neurosurgery 133 (2015) 83–89

its shape and latency, which is largely dependent on the surgeon’s experience. Contralateral side is also recorded to serve as a control throughout all stages of surgery. Patterns of FN eletromyographic stimulation and the corresponding facial muscle response during VS dissection have been considered as predictors of FN outcome both in short and long-term follow-up. Direct intraoperative stimulation of FN is most commonly performed at the brain stem and internal acoustic canal after VS resection. As the reflection of FN conductance, post EMG stimulation response amplitude was paid more attention to in predicting of FN outcome [2,16,23]. Lin et al. [23] found that the percentage based on the response amplitude of direct facial nerve stimulation with increasing intensities at the pontomedullary junction and the maximum amplitude of transcutaneous stimulation at the stylomastoid foramen was a good predictive factor for immediate postoperative FN function. The positive predictive values reached 93% with sensitivity of 61% when the FN stimulus intensity was increased to 0.3 mA. Our group calculated amplitude ratios of the exit from the brain stem (proximal) to the fundus of the internal auditory canal (distal) and this was proved to be a predictive factor for long-term FN outcome in large VS. Other studies including smaller tumors demonstrated that proximal-distal FMEP ratio was more powerful than absolute values of EMG amplitude in predicting both immediate and long-term FN outcomes [2,26]. Consequently, the role of MRA ratio can be regarded as a supplementary indicator for train time in predicting both immediate and long-term FN function in patients with VS ≥30 mm. For effective prediction of postsurgical FN function, two or more factors were combined to enhance the predictive value in some studies. Neff et al. [16] demonstrated that both a stimulus threshold ≤0.05 mA and a response amplitude ≥240 mV resulted in a HB Grade 1 to 2 outcomes with a 98% probability 2 years after surgery. When the three factors, a stimulation threshold ≤0.04 mA at proximal site, responses amplitude ≥800 ␮V and a proximal/distal post ratio ≥0.6 were taken into consideration simultaneously, 93% of patients presented with an acceptable FN function [27]. There are two major limitations for EMG in monitoring FN response: one is numerous false-positive responses caused by electrical artifacts during microsurgical process. The other is invasivsity that carries a possible risk of developing subcutaneous hematoma or facial skin burned by the electrodes [28]. Video monitoring was carried out by placing a mask equipped with video cameras on the face to directly detect the motion of the facial muscles. Reports showed that video monitoring was more specific but less sensitive than EMG monitoring in predicting FN outcome [29]. However, inclusion criteria for these studies were patients with small VS (