Facial Nerve Outcome and Results in Hypervascular Large Vestibular Schwannomas: Our Personal Experience in A Series of 32 Cases.

Thirty-two consecutive patients were operated on for HVVS (Group-A). Results were compared with those of 25 patients (Group-B) operated on for large low-bleeding VS. Tendency to bleed and adherence of capsule to nervous structures were evaluated by reviewing video records. Cisternal facial nerve (FN) position was reported. Microsurgical removal was classied as total, near-total, subtotal or partial and MIB-1 index evaluated in all. FN results were classied according the House-Brackmann scale. FN outcome, especially in cases with tight capsule adhesion.


Introduction
According to Koos' classi cation (21), Grade IV vestibular schwannomas (VS) are large tumors (longitudinal diameter > 3 cm) compressing brainstem, displacing fourth ventricle and compromising the quality of life of patients. During the last decades their incidence has gradually reduced for the broad and earlier access to MRI imaging (48,50,51). In the past, Grade IV VS represented 40% of all tumors, whereas they accounted only a few percent during the last 10 years. (38). Surgical resection represents the treatment of choice in order to achieve a clinically signi cant improvement (38, 53,54).
Microsurgery of Grade IV VS is technically challenging because of adhesions of tumor capsule to the brain stem and facial nerve (FN), unusual displacement of the nerve and, in several cases, tendency of tumor to bleeding. In particular, even if majority of VS are hypovascular tumors, some tumors may have an unusual vascular architecture and are better de ned "hypervascular vestibular schwannoma" (HVVS).
The rate of HVVS increases with size (49) of tumors and its incidence seems to be higher in large and solid VS and in younger patients (56). Blood supply of VS comes from branches of external carotid artery (ECA) and from vertebral-basilar (VB) system (49,56). According to the angiographical analysis of Teranishi et al (49), HVVS have a high concentration of abnormal vessels, with or without arteriovenous (AV) shunts, with tumor stain from external carotid artery (ECA) and/or vertebral-basilar (VB) arteries.
Few Authors (23,49,56) analyzed in detail the behaviour and the outcome of HVVS in comparison to hypovascular ones. In this study, we retrospectively analyzed the clinical and surgical data and the outcome of a consecutive series of 32 patients with Grade IV HVVS consecutively operated on by retrosigmoid approach, highlighting the extent of tumor removal, postoperative FN outcome, and complications, in comparison to a 25 low-bleeding VS, surgically treated during the same period.

Study design
This observational single-center study was carried out at a tertiary care referral hospital. It was conducted after obtaining clearance from the Internal Ethics Committee of our institution and in accordance with the principles set forth in the Helsinki Declaration. A written consent for scienti c treatment of personal data was obtained from any patient before surgery. Cohorts included all patients who underwent surgery for Koos Grade IV VS in the period between Department between December 2010 and December 2019: 32 hypervascular (Group A) and 25 hypovascular/low-bleeding tumors (Group B). All patients were followed up till June 2020. The mean follow-up duration for inclusion was 63 months (median 63,5).

Data Collection
Data was collected from case sheets, operative notes, neuro-imaging archiving and communication systems (PACS) and discharge summaries, after obtaining informed consent from patients.
Subsequently, the patients were followed up on an out-patient basis. The parameters that were studied included demographic pro les, clinical features, duration of symptoms, neurological status, neuroimaging, operative details, histopathological data, recurrence, functional outcome, mortality and morbidity.
We reviewed 32 consecutive cases of Koos Grade IV hypervascular VS (HVVS) out of 220 unilateral VS surgically treated by the rst Author between December 2010 and December 2019. Results were compared with those of 25 patients (Group B) operated on in the same period for hypovascular grade IV VS, presenting low intraoperative bleeding.
General conditions and preoperative risk were assessed according to the American Society of Anesthesiology (ASA) classi cation (30). Patients with neuro bromatosis type 2 were not included. Clinical data such as patients' age, sex, presenting symptoms and tumor size were recorded. Preoperative neuroimaging included temporal bone CT and contrast-enhanced MRI in all patients and allowed to evaluate VS size and possible presence of one or more intratumoral cysts. Tumor size was categorized according to the international criteria, measuring the largest extrameatal tumor diameter on post-contrast axial MRI. (19) The MRI-slices of HVVS consistently showed multiple "serpiginous" ow voids, representing large feeders and draining veins within the mass, especially in T2-weighted images (Fig. 1).
Preoperative audio-vestibular evaluation included pure tone audiometry and speech audiometry. Hearing level was assessed according to the American Academy of Otolaryngology-Head and Neck Surgery (AAO-HNS) classi cation (8): Class A and B were categorized as good hearing.
Microsurgery via retrosigmoid approach was performed in all cases. Location of FN and its adherence to the tumor were evaluated by reviewing radiological images, surgical and video records. The course of the FN was classi ed according to its position in relation to the tumor: anterior (A), anterior-inferior (AI), anterior-superior (AS), and dorsal (D) (26, 41). The extent of tumor removal was classi ed as total (T), near total (NT: tumor residue < 5%), subtotal (ST: residue 5-10%), and partial (P: residue > 10%). Since about 3 years, at the end of microsurgical resection, a 4-mm Flexible Video Endoscope (4-mm x 65-cm, Karl Storz, GmbH, Tuttlingen, Germany) has been inserted in the surgical cavity, handled by the surgeon (9,10), in order to detect eventual tumor residues in the internal auditory canal (IAC). On pathological examination of tumor, MIB-1 index was evaluated in all cases.
We classi ed the extent of tumor removal as total (T), near total (NT: linear millimetric residual tumor), subtotal (ST: residual tumor volume < 10%), and partial (P: residual tumor volume > 10%). Extent of resection was evaluated by means the blind review of postoperative MRI performed by one member of our neurosurgical staff (coauthor) and one neuroradiologist.
FN function was assessed according to House-Brackmann (16) (HB) grading system and was evaluated preoperatively, at discharge and at nal follow-up (≥6 months).

Neurophysiological monitoring
All the surgical procedures were performed with intraoperative FN monitoring system (Nimbus i-Care 100 intraoperative neurophysiologic monitoring; Newmedic Division of Hemodia, Labège, France), with electrodes inserted in orbicularis oris and orbicularis oculi muscles, to detect FN responsivity. The nerve stimulation was performed with monopolar (on tumor surface) or bipolar (close to the nerve) probes to locate the facial nerve and verify its functional state.

Operative technique
The retrosigmoid approach was used in all the surgical procedures. A slightly curved skin incision, 5 to 6 cm long, was performed behind the ear, 1-2 cm posteriorly to the mastoid process. The lateral occipital bone was exposed, including superior and inferior nuchal lines. A retrosigmoid suboccipital craniotomy, 3cm in length and width, was performed, exposing sigmoid sinus anteriorly and transverse sinus superiorly. The retrosigmoid dura was incised in a semicircular fashion, lateral medullary cistern was opened and cerebrospinal uid (CSF) was aspirated to obtain adequate cerebellar decompression. After cutting the dura along the "Tuebingen line" (3), a dural landmark on the IAC inferior limit on the posterior surface of the petrous temporal bone, the canal was opened through a 4mm extra-coarse diamond burr or by ultrasonic aspirator with dedicated bone tips. The tumor surface was exposed and the rare but possible dorsal displacement of FN was investigated by monopolar stimulation. The intracapsular, either subperineural (subcapsular) or subarachnoid plane of surgery (20) was used in any case for debulking, detachment and dissection, respectively, even if in all cases we tried to follow the arachnoid re ection over cranial nerves and brainstem for obtaining the best functional results.
A V-cut was performed on the dorsal surface of tumor with microscissors or hand-held laser thulium and debulking of tumor was obtained with microscissors, microcurettes, bipolar forceps, hand-held laser ber, and ultrasonic aspirator.
In particular, capsule incision and tumor debulking of HVVS has been performed with hand-held 2µ-Thulium exible laser ber (Revolix jr®, Lisa laser USA, Pleasanton, CA, USA), with a range of power setting from 1 to 14W. Standard 0,9% saline solution irrigation has been used for cooling surrounding tissues and the ber, which is not hindered in its function by the presence of water. The ber is used for cutting, vaporizing, and coagulating the capsule and the intracapsular mass, in combination with lowpower bipolar forceps, microscissors and Sonopet Ultrasonic Aspirator (Stryker, Kalamazoo, MI).
Following tumor debulking, the remaining tumor capsule was then dissected from brainstem and cranial nerves during continuous facial nerve monitoring with standard microsurgical tools. In some cases of strong tumor adherence to surrounding structures, a millimetric remnant of tumor capsule was left, thus yielding a near-total resection. In few cases with tight tumor capsule adherences to FN and/or brainstem and high-bleeding rate of tumor a ST removal was performed. Accurate hemostasis, mastoid cells obliteration, and tight dura closure by pericranial graft, hemostatic materials and sealants were performed and the bone operculum or a tted titanium mesh was placed on the craniectomy with miniscrews.

Bleeding rate and Adhesion of tumor capsule
For assessing the tendency of tumor to bleed and the adherence of capsule to FN and brainstem, two independent "blinded" reviewers and coauthors espressed their opinion watching the surgical video, without considering the impression of surgeon wrote in the report.
Bleeding tumors, classi ed as HVVS, had an unusual and redundant super cial and internal vascular architecture. Except for the angiographical classi cation proposed by Teranishi et al (49), there is not in the literature any objective de nition of "hypervascularity". During surgery, in all these cases intra-and extratumoral vascularity was arterialized due to luxurious shuntings and removal was complicated by high-rate bleeding during debulking and dissection, because of the rupture of the multiple vessels present in the capsular and intracapsular portion of tumor. The possible high-bleeding behaviour of tumor was suspected in the preoperative T2-weighted MRI images when multiple "serpiginous" ow voids, representing large feeders and draining veins, within the mass were detected. Reviewing the surgical videos, the "blind" independent observers rated the bleeding amount and classi ed tumors in hypervascular and hypovascular.
VS without a clear perineural dissection plane between nervous structures (in particular, FN and brainstem) and tumor capsule were considered more adherent, in comparison to VS with well recognizable dissection plane (less adherent). Unfortunately, there is no scale for the degree of adhesion: therefore, less strong adherences were de ned as those in which the FN and brainstem could be separated from the tumor comparatively easily with a microdissector, and strong adherences were de ned as those that were di cult to separate, requiring sharp dissection with microscissors. In cases where adhesions where too tight and in presence of brainstem edema, total resection was considered unsafe and a less than total resection was accepted.

Clinical follow-up
Long-term FN outcome evaluation was performed at 6 months after operation and at last follow-up and was categorized according to House-Brackmann grades I-VI. Clinical and radiological follow-up was scheduled at 6 months after operation and then once a year; follow-up period ranged between 6 and 117 months (mean 63, median 63,5). Statistical analysis Statistical analysis was performed by means chi-square test for categorical variables and Student's t-test for continuous ones, using the software MS Excel (Microsoft Corporation, Redmond, WA, USA). Statistical signi cance was set at p < 0,05.
Preoperative patients data are summarized in Table 1.  Data related to surgical and pathological details are summarized in Table 2. Data related to FN outcome are summarized in Table 3. Complications and recurrences are summarized in Table 4. The signi cance and relevance of bloody supply for microsurgery and outcome of VS has been only seldom reported in the literature (1,32). Although preoperative angiography provides characteristic ndings (49), MRI can con rm the diagnosis of a HVVS by showing multiple ow-voids in the contest of the tumor (Fig.1). Teranishi et al (49) proposed a classi cation of HVVS in 5 types in relation to tumor feeders (VB system only or VB system and ECA branches) and to the presence or not of arterio-venous shunts. The presence of shunts resulted to be less frequent but was associated with statistically signi cant higher rate of recurrence, especially those cases with feeders from VB system and ECA branches and with arterio-venous shunts originating from vertebral-basilar arteries (p=0,0476). Some Authors (23,56) proposed a 2-step surgical strategy for the treatment of HVVS, on considering that partial resection performed with the rst surgery could reduce the hypervascularity of tumor, making total removal less risky and more feasible with the second surgical step. However, we agree with Teranishi et al (49) that it is seems to be preferable to attempt T/NT resection of HVVS during the rst surgery, for the overall comfort of patients and for reducing the risk of postoperative hemorrhage possible after the rst step of partial resection.
According to Peris-Celda et al. (35), large tumors are signi cantly more frequent among younger patients at diagnosis (p<0,001), similarly to what happens in patients affected by NF2, suggesting a possible more aggressive tumor biology. In particular, on comparing VS with maximal diameter >4cm (more than 7% of their series) with the rest of the cohort, they observed a statistically signi cant difference in terms of mean age at diagnosis: 52,3 years for smaller versus 42,4 years for larger tumors (p<0,001). (35) Angiogenesis is essential for the enlargement of any solid tumor, including schwannomas: it has been demonstrated that VEGF expression of VS correlates with tumor growth pattern. (5,36) Vascular endothelial growth factor (VEGF) is considered to be a major regulator and VEGF receptor (VEGFR)-1 and VEGFR-2, have been identi ed on the cell surface of vascular endothelial cells. (13,55) In addition, VEGF and matrix metalloproteinases (MMPs) are strong mediators of tumor angiogenesis: Moller et al. (31) observed that tumor concentration of MMP-9 correlates with VS growth rate and adhesion to nerve structures, concluding that this collagenase is strongly involved in the growth of VS. Moreover, a relationship among vascularization, adhesions and tumor size is quite reasonable, probably through the expression of MEK/ERK effectors, oncogenic gene miR-21 and mTOR pathways (57), which regulate several cellular processes.
To date, few studies attempted to pro le genome-wide alterations in sporadic VS. In a series of 23 sporadic VS, Carlson et al. (4) analyzed fresh frozen tumor specimens and matched peripheral blood leukocytes, in order to identify if more clinically aggressive variants possess different genetic alterations compared to the more indolent. Using high-throughput deep sequencing, ''two-hit'' alterations in the NF2 gene were identi ed in every tumor and were not present in peripheral blood supporting that all events were somatic. (4) Type of NF2 gene alteration and accessory mutations outside the NF2 locus may predict phenotypic expression and clinical course.
Surgical dissection for large, vascularized and adherent VS.
The layers we encounter starting from the surface of VS are: (1) Arachnoid folder; (2) FN and cochlear nerve; (3) perineurium/nerve bers of vestibular nerve of origin of VS (5,36,39). Thus, the capsule of VS is the perineurium of the vestibular nerve of origin: in large and giant VS frequently there is no arachnoid separating the tumor capsule from FN and cochlear nerve. (20) According to Kohno et al. (20), there are 3 planes for possible tumor dissections: A. subarachnoid; B. subperineural (subcapsular); C. intracapsular. During the surgical removal of large VS, these Authors (20) suggest that bimanual dissection is an essential component of the technique and that it is necessary to take in account the tumor capsule and arachnoid re ection for obtaining the best functional results. (20,27,34,40,46) Epiarachnoid tumors are de ned by the absence of an arachnoid membrane on the tumor surface after moving the arachnoid fold (double layers of the arachnoid membrane) towards the brainstem. In contrast, subarachnoid VS maintain the arachnoid membrane on the tumor surface after moving the arachnoid fold. (34) Based on this hypothesis, Kohno et al. (20) used intraoperative views and light and electron microscopy to con rm the existence of an arachnoid membrane after the arachnoid fold had been moved: they observed VS are usually subarachnoid tumors, whereas epiarachnoid variant is considerably less common.

Extent of Removal of large VS and markers of tumor cells proliferation
Large tumor size often compromises safe and effective total resection; in the literature the rate of total resection of large VS ranges between 28,6% (61) and 95,5% 24); in two series total resection could be accomplished in all cases (43,47). Furthermore, factors that negatively affect results could associate, such as hypervascularity, which determines high-bleeding intraoperative rate and increased technical di culties (23,49,56).
In order to limit bipolar coagulation and more heat thermal damage during HVVS microsurgical removal, we decide to use 2µ-Thulium exible hand-held laser ber, fot cutting, vaporizing, and coagulating the capsule and the intracapsular mass of the tumor, in combination with low-power bipolar forceps, microscissors and ultrasonic aspirator In a retrospective series on 78 consecutive cases (28), the use of 2µ-Thulium laser ber in VS surgery proved to be safe, even if did not have signi cant in uence on FN outcome, hearing preservation rate nor surgical time. On the other hand, the necessary reduction of tumor volume of HVVS before microsurgical dissection of facial and cochlear nerve appears to be safer and easier with 2µ-Thulium laser ber in association with ultrasonic aspirator and microsurgical dedicated instruments.
Antigen KI-67, tested with the MIB-1 index, is a nuclear protein that is associated with cellular proliferation. VS with MIB-1 index higher than 3% are actively proliferating with theoretical higher risk for regrowth or recurrence. (28) According to Teranashi et al (49), HVVS have a higher MIB-1 index: in their series, hypervascular VS had a mean MIB-1 of 4,3% versus 2,8% of non-hypervascular tumors (p<0,05).
These data are not in accordance with those observed in our series, in which the mean MIB-1 index was 1,25% (range 1-3) in Group A and 1,08% (range 1-2) in Group B (p=NS).

Functional results
Hearing loss is one of the most common signs of VS at presentation (41,7% of cases) (2) and if socially useful hearing is present preoperatively, attempts should be made -when possible-to accomplish its preservation, especially in small-sized tumors. (12). As far as large and giant VS are concerned, in selected series hearing has been preserved in 21,4-50% (12,29,59,60) and 66,7% (43) of reported cases, respectively. In the present series, preoperative severely impaired hearing or deafness was present in all HVVS and in 23 cases (92,0%) of Group B; hearing preservation was not possible in the 2 AAO-HNS hearing Class B cases with low-bleeding VS.
Although great emphasis is currently placed on preserving FN function after VS resection, its injury still represents a relatively common postsurgical complication especially in large tumors. In addition, even if careful dissection is performed, an anatomically intact nerve does not necessarily predict a HBI FN function. According to the literature, preservation of FN functional state in VS surgery is accomplished in 32,9-83,3% of cases (15,17,24,43,47,58,61,62). On considering long-term results, 20 HVVS (62,5%) had HBI-II FN result at last follow-up control versus 24 (96,0%) low-bleeding cases (p<0,01) ( Table 3). Adhesions of tumor capsule to FN had a negative impact on FN outcome in Group A: HBI-II FN results were obtained only in 40,0% (8/20) of adherent HVVS versus 92,8% (13/14) of low-bleeding adherent tumors (p<0,05). It seems to be correct to underline one bias of the present study: any surgeon gains more experience during the years and therefore results could be better in later part of experience.
These lower HBI-II FN outcome in HVVS induced some surgeons to leave more residue. Zhang et al. (61) obtained the best functional outcome in patients who underwent subtotal resection instead of radical extirpation. Even if controversial results have been reported with planned less-than-total resection performed for FN preservation, according to some Authors (7,22,33,52,63) outcome might be improved in selected cases by combined surgical/radiosurgical treatment. Zumofen et al. (63) reported 89% HBI-II postoperative rate, with no need for salvage surgery after Gamma Knife on planned tumor residues. However, Iwai et al. (22) found that optimal FN outcome (95% postoperative HBI-II) could be jeopardized by the need for salvage surgery after Gamma Knife in case of large VS residues (at least 6cm 3 ).
Notwithstanding, even if surgical removal should be attempted with the objective of maximal safe tumor eradication, such ndings underline that SRS is not an enemy of microsurgery (52).
Complications and recurrences/re-growth of residue In our series, mortality rate was zero and permanent complications (diplopia for abducens nerve paralysis) occurred only in one case of HVVS. Transient postoperative complications were observed in 8 patients (22,8%), without correlation with preoperative ASA class. As regard recurrence/re-growth of residue, at a follow-up ranging from 6 to 113 months, it was observed in 4 cases (3 ASA Class 2 and 1 Class 3): a re-operation was performed in 2 HVVS patients and SRS in other two (one patient ASA3 and one who refused second surgery).
These rates are in line with the literature (2,17,22,24,43,45,53,54,58,63) and con rm that retrosigmoid approach is safe and feasible to remove even giant VS (15,24,43,45,53,58). The translabyrynthine approach has been traditionally suggested for this kind of tumors, with good results in terms of extent of resection (rates of total resection around 90%), postoperative facial outcome (HBI-III close to 75%) and perioperative complications (CSF leaks in about 2% of cases) (6,11,25,61). On the other hand, other Authors reported perioperative complication rates as high as 14,3% (14). Even if translabyrinthyne approach is a feasible alternative, the results of our series contribute to support the use the retrosigmoid approach in large and HVVS too.
Surgical resection represents the ideal treatment for large and giant VS, including HVVS. It signi cantly and positively impacts on the patients' quality of life (54) and should be considered even in the case of elderly ones.

Conclusions
Compared to low-bleeding VS, microsurgery of Koos Grade IV HVVS seems to be associated with higher complication rate, higher recurrence/re-growth rate, and poorer FN outcome, especially in cases with tight capsule adhesion to the nerve.

Declarations FUNDING STATEMENT
None (not applicable).

CONFLICT OF INTEREST
The authors declare that they have no known competing nancial interests or personal relationships that could have appeared to in uence the work reported in this paper.

AVAILABILITY OF DATA AND MATERIAL (Data Transparency)
Data available from the rst Author on demand.

CODE AVAILABILITY
None (not applicable).

ETHICS APPROVAL
The study involves human participants: therefore, it has been reviewed and approved by local ethics committee of the Hospital. A written consent for scienti c treatment of personal data was obtained from any patient before surgery. No potentially identi able human images or data are presented in this study.
All procedures performed in this study were in accordance with the ethical standards of the internal institutional ethics committee ("Comitato Etico Lazio 1" Members of ASLRoma1: Dr. Marco Tubaro, Dr. Teresa Calamia, Dr. Francesco Meo). A written consent was obtained from any patient included in the study.

Consent to participate and Consent for publication
All Co-Authors express formally their consent to participate to this study an to publish it, contributing in different ways.
Author's individual contributions L.M.: study design, study conception, data extraction, data analysis, manuscript writing A.C.: data extraction, data analysis C.G.S.: data extraction, data analysis, radiological measurement E.C.: data extraction, data analysis, radiological measurement G.C.: data extraction, data analysis, statistical analysis, R.R.: data analysis, statistical analysis, critical review of the manuscript A.A.S.: critical review of the manuscript, study supervision