The majority of schwannomas and neurofibromas are benign, slow-growing tumors with little or no capacity for infiltration of surrounding neural tissues. Therefore, the focus of initial management has traditionally been aggressive surgical resection (Jackler and Pitts 1990; Macfarlane and King 1995; Samii et al 1995; Seppala et al 1995a; 1995b; Strauss and Post 1995). This approach is appropriate for most intracranial, spinal, and peripheral nerve schwannomas and neurofibromas because complete surgical extirpation is often curative. However, in certain cases of vestibular schwannoma, it is not always clear whether to proceed directly with surgical intervention or to carefully observe the patient (Fucci et al 1999; Mirz et al 1999; Rosenberg 2000; Nutik and Babb 2001; Al Sanosi et al 2006). In the MRI era, many vestibular tumors are detected at an early stage as small, intracanalicular or cisternal lesions less than 2.0 cm in diameter. When treated conservatively, approximately 20% to 60% of these small tumors do not enlarge during the period of observation (Wazen et al 1985; Thomsen and Tos 1990; Bederson et al 1991; Nedzelski et al 1992; Macfarlane and King 1995; Fucci et al 1999; Mirz et al 1999; Rosenberg 2000; Flint et al 2005; Al Sanosi et al 2006; Ferri et al 2008; Martin et al 2008; Bakkouri et al 2009; Sughrue et al 2010). Of those tumors that do enlarge, 75% to 80% grow slowly, at a rate of 1 mm to 2 mm per year. The growth rate may be even slower in elderly patients. For example, in a series of 70 patients over 65 years of age followed by Rosenberg over a mean follow-up period of 4.8 years, only 4 patients (5.7%) required surgical intervention (Rosenberg 2000). The growth rate does not seem to correlate significantly with patient age, initial tumor size, or duration of symptoms (Bederson et al 1991; Nedzelski et al 1992; Fucci et al 1999; Mirz et al 1999; Bakkouri et al 2009). In more than 383 patients treated in an expectant, conservative manner, tumor behavior during the initial follow-up period of 12 to 58 months was predictive of further growth (Bederson et al 1991; Nedzelski et al 1992; Fucci et al 1999; Mirz et al 1999; Al Sanosi et al 2006; Ferri et al 2008). Tumors were likely to remain quiescent if they demonstrated little or no growth during the observation period. Conversely, tumors that subsequently required surgery grew steadily during the observation period, often with an accelerated growth rate (greater than 2 mm per year). These data have led many authors to adopt a conservative, observational approach for the following patients: any patient with generally poor health, elderly patients with tumors less than 10 mm in size, elderly patients reluctant to proceed with surgery, and any patient with significant hearing loss in the opposite ear (Jackler and Pitts 1990; Bederson et al 1991; Nedzelski et al 1992; Macfarlane and King 1995; Fucci et al 1999; Mirz et al 1999; Bakkouri et al 2009). Tumors with a growth rate equal to or greater than 2 mm per year should be considered for surgical removal. In general, the delay in surgical intervention does not appear to result in more morbidity for the patient (Flint et al 2005; Ferri et al 2008). In a recent comparison of conservative versus primary surgical management of acoustic schwannomas, Martin and colleagues noted that facial nerve preservation (p < 0.001) and hearing preservation (p < 0.000) were both superior in the conservatively treated cohort of patients (Martin et al 2008). Conservative approaches are unjustified in most young patients, whose tumors generally have accelerated growth rates, and in patients with tumors exceeding 2.5 cm in diameter. A recent meta-analysis of the conservative approach literature reviewed the data on 982 patients to assess hearing outcomes (Sughrue et al 2010). Patients with lower rates of tumor growth (≤ 2.5 mm/year) had significantly higher rates of hearing preservation in comparison to those with higher growth rates (75% vs. 32%, p < 0.0001). The authors concluded that a rapid growth rate (ie, > 2.5 mm/year) was a better predictor of hearing loss than the initial size of the tumor in patients with tumors less than 25 mm in diameter.
Surgical resection of vestibular schwannoma. Surgical resection is the treatment of choice for most patients with a vestibular schwannoma. Complete removal is usually attempted in all patients, except for selected cases (eg, elderly patients) wherein a shortened procedure may significantly reduce potential morbidity and mortality. Three surgical approaches are used for vestibular schwannoma removal: (1) suboccipital (retrosigmoid), (2) translabyrinthine (anterosigmoid), and (3) middle fossa (subtemporal) (King and Morrison 1980; Bentivoglio et al 1988; Jackler and Pitts 1990; Jackler and Pitts 1992; Macfarlane and King 1995). Each has specific advantages to offer, based on selection criteria that include tumor size, depth of internal auditory canal penetration by tumor, hearing status, exposure of the facial nerve, and patient age (Jackler and Pitts 1992). In general, the ability to perform a complete resection and preserve cranial nerve function correlates strongly with tumor size and is most favorable in tumors less than 1 cm in diameter.
The suboccipital or retrosigmoid approach uses a craniectomy just posterior to the sigmoid sinus; exposure of the tumor and cerebellopontine angle is excellent (Bentivoglio et al 1988; Jackler and Pitts 1990; 1992; Samii and Matthies 1997a; Ojemann 2001). The major advantages of this approach are the level of exposure and the possibility of preserving hearing. Disadvantages include the need for cerebellar retraction (with an associated increased incidence of postoperative dysmetria) and frequent postoperative headaches. Many surgeons use this approach for tumors less than 2 cm in patients with good hearing, and for large tumors that extend toward the jugular foramen (Jackler and Pitts 1990; 1992; Macfarlane and King 1995; Samii and Matthies 1997a; Ojemann 2001). A retrosigmoid transmeatal approach is particularly good for “extra-large” tumors (greater than 4 cm in diameter) that require extensive exposure for tumor resection and preservation of cranial nerve function (Jung et al 2000).
The translabyrinthine or retrosigmoid approach uses a craniotomy of the lateral portion of the temporal bone, including the mastoid air cells and semicircular canals, to gain exposure of the internal auditory canal and cerebellopontine angle (King and Morrison 1980; Jackler and Pitts 1992; Lanman et al 1999; Sluyter et al 2001). The advantages of this approach are that minimal cerebellar retraction is necessary, cerebellopontine angle exposure is good, all drilling of the temporal bone is completed before the dura is opened, the entire intratemporal course of the facial nerve is accessible, and the incidence of postoperative headache and cerebrospinal fluid leakage is reduced. The major disadvantages of the translabyrinthine approach are that hearing is irrevocably abolished and exposure may be limited inferiorly by the jugular foramen (Jackler and Pitts 1990; 1992; Macfarlane and King 1995). Most surgeons recommend this approach for all intracanalicular and medium-sized cisternal tumors associated with poor hearing, selected tumors with deep internal auditory canal penetration associated with good hearing, and most large cisternal tumors exceeding 3 cm (Jackler and Pitts 1992; Lanman et al 1999; Sluyter et al 2001).
The middle fossa, or subtemporal approach, uses a small temporal craniotomy anterosuperior to the external auditory canal. The facial nerve is usually draped over the top surface of the tumor with this exposure, and must be manipulated for tumor removal. The middle fossa approach is most often used for intracanalicular tumors or small cisternal tumors (less than 5 mm beyond the porus acusticus); exposure of the cerebellopontine angle is minimal. The advantages of this approach are the possibility of sparing hearing and the low incidence of postoperative headaches as much of the procedure is extradural (Jackler and Pitts 1990; 1992; Irving et al 1998; Brackman et al 2000; Gonzalez et al 2000). In a direct comparison of hearing preservation after use of the retrosigmoid or middle fossa approaches, the middle fossa method demonstrated significantly improved results (52% vs. 14% of patients with Class B or better hearing; p = 0.009) (Irving et al 1998). The middle fossa approach is more likely to achieve hearing preservation (1) when the patient has adequate preoperative hearing, (2) if the auditory evoked brainstem response shows shorter latency value, and (3) when the tumor arises from the superior vestibular branch of the nerve (Brackman et al 2000). Disadvantages of the middle fossa approach include restricted cerebellopontine angle exposure, increased potential for facial nerve trauma, and complications from retraction of the temporal lobe (ie, epilepsy, dysphasia, cerebral hematoma).
Reports by several authors document the potential benefits of using endoscopy during surgical resection of vestibular schwannomas (Goksu et al 1999; Wackym et al 1999). The major benefits are higher magnification (to better define neurovascular anatomy of the posterior fossa), improved visualization of the tumor (especially within the internal auditory canal), and a reduced risk of postoperative cerebrospinal fluid leakage.
Intraoperative cranial nerve monitoring has emerged as an excellent method for potentially reducing the morbidity of vestibular schwannoma resection (Jackler and Pitts 1990; Harper et al 1992; Yingling and Gardi 1992; Glasscock et al 1993; Macfarlane and King 1995). This technique assists in identifying cranial nerves in the operative field (using intracranial electrical stimulation), and facilitates their dissection from tumor while minimizing significant trauma. Commonly monitored motor nerves include the facial, trigeminal, and spinal accessory. Postoperative function of cranial nerves V, VII, and XI is significantly improved when monitoring is used for tumor resection (Wiet et al 1992; Yingling and Gardi 1992; Strauss 2002). This is especially important for large medially placed tumors; in 36% of patients in 1 series, the facial nerve was split in its passage through and around the mass (Strauss 2002). The auditory nerve is monitored using auditory evoked brainstem response only during hearing preservation operations (Ojemann et al 1984; Harper et al 1992; Glasscock et al 1993; Matthies and Samii 1997b). The benefits of operative monitoring of cranial nerve 8 are controversial, but procedural morbidity is probably reduced in some patients with tumors less than 2 cm (Jackler and Pitts 1990; Harper et al 1992; Yingling and Gardi 1992; Matthies and Samii 1997b). Furthermore, some authors contend that the quality of the preoperative auditory brainstem response positively correlates with postoperative cranial nerve VIII function and extent of hearing preservation (Matthies and Samii 1997c).
Overall, the mortality of vestibular schwannoma removal is 1% to 2% for an experienced surgeon. Most of the fatalities occur in elderly patients or patients with large tumors. Almost 50% of these deaths are caused by medical complications such as pulmonary embolism, myocardial infarction, and pneumonia. Potential complications are numerous and include vascular injury, hemorrhage (0.5% to 2%), cerebellar injury (1% to 2%), cranial nerve injury, headache (10% to 30% last 4 weeks or more), vertigo, pneumocephalus, meningitis (2% to 6%), aseptic meningitis (5% to 10%), cerebrospinal fluid leakage (10% to 30%), hydrocephalus, and various medical conditions (King and Morrison 1980; Bentivoglio et al 1988; Jackler and Pitts 1990; 1992; Wiet et al 1992; Macfarlane and King 1995; Samii and Matthies 1997a; 1997b; Pirouzmand et al 2001). There is no difference in the rate of postoperative cerebrospinal fluid leak between the rectosigmoid and translabyrinthine approaches (Brennan et al 2001). However, leaks from translabyrinthine approaches more often need surgical repair.
After resection of a vestibular schwannoma, follow-up imaging with MRI is recommended. A recent study suggests that for uncomplicated cases with a complete resection, initial follow-up imaging should be at 1 year (Bennett et al 2008). Patients with residual enhancement and/or subtotal resections, or underlying NF2, should undergo follow-up MRI on a more frequent and consistent basis.
Surgical resection of other cranial nerve, spinal, and peripheral nerve schwannomas. The basic principles of surgery for schwannomas from other sites are similar to those for vestibular tumors. In most cases, a complete resection is curative; subtotal removal often results in eventual recurrence of tumor. The most common operation for resection of a trigeminal schwannoma uses a subtemporal-intradural approach (McCormick et al 1988; Miller 1988; Pollack et al 1989; Samii et al 1995; Strauss and Post 1995; Sarma et al 2002; Moffat et al 2006; Zhang et al 2009). This technique is best for tumors in the middle fossa, Meckel cave region. Tumors that have significant extension toward the cavernous sinus and superior orbital fissure may require a frontotemporal-transsylvian approach. Tumors confined mainly to the posterior fossa are resected using a suboccipital approach in most cases, although some authors prefer the orbitozygomatic extradural or the subtemporal-infratemporal approaches instead, depending on the branch of the nerve involved (Krishnamurthy et al 1998). A combined approach is necessary for dumbbell-shaped tumors; for instance, the extradural zygomatic middle fossa approach. With this technique, Al-Mefty and coworkers were able to perform gross total resections with excellent preservation of cranial nerve function (Al-Mefty et al 2002). Intraoperative monitoring may be helpful in some cases (Strauss and Post 1995). Using conventional techniques, the mortality rate is 2% to 3%, with total or near-total removal of tumor in 70% to 75% of cases (Samii et al 1995; Strauss and Post 1995; Zhang et al 2009). However, the recurrence rate approaches 50% in many series. Some authors recommend more extensive skull base approaches for resection, claiming that improved tumor exposure and ease of dissection allow for more frequent complete and improved clinical outcome (Taha et al 1995; Yoshida and Kawase 1999; Moffat et al 2006; Fukaya et al 2010). For example, Fukaya and colleagues recently reported the surgical results of a series of 57 trigeminal schwannomas, most of which underwent resection using a skull base approach (Fukaya et al 2010). Complete resection was accomplished in 42 of 45 patients (90%), with no surgery-related mortalities. The most common surgical complications of skull base surgery are cranial nerve injury, cerebrospinal fluid leak, meningitis, and hydrocephalus. Facial nerve schwannomas are resected using an approach based on the extent and location of the tumor: middle fossa, suboccipital, or translabyrinthine (Rocchi et al 1991; Strauss and Post 1995; Sarma et al 2002; Kim et al 2003). If hearing is poor, the translabyrinthine approach is recommended. If a hearing preservation operation is attempted, the preferred approach is either suboccipital or middle fossa. Complications are similar to those of trigeminal tumors. Schwannomas of the jugular complex (IX, X, XI) are resected most often using a suboccipital approach, occasionally in combination with a mastoidectomy (Sweasey et al 1991; Strauss and Post 1995; Rapana et al 1997; Wilson et al 2005; Bulsara et al 2008). Schwannomas of the ocular motor cranial nerves (III, IV, VI) are generally resected using frontotemporal or subtemporal approaches (Miller 1988; Mehta et al 1990; Tung et al 1991; Jackowski et al 1994; Strauss and Post 1995; Mariniello et al 1999; Sarma et al 2002). Tumors located mainly within the posterior fossa can be removed with a suboccipital technique. The complete resection rate for ocular motor nerve schwannomas is approximately 50% (Strauss and Post 1995). In selected cases, preservation of oculomotor nerve function can be attempted by sural nerve grafting (Mariniello et al 1999). Hypoglossal schwannomas are most often resected using a lateral suboccipital approach (Miller 1988; Strauss and Post 1995; Tucker et al 2007). The operative mortality in these patients is relatively high (6% to 7%), usually because of respiratory complications such as aspiration and pneumonia. Surgical resection of intrasellar schwannomas can be accomplished using a trans-sphenoidal approach (Honegger et al 2005). A gross total resection is often possible.
Surgical treatment of spinal schwannomas and neurofibromas is similar to that for tumors of the cranial nerves; complete resection is curative in most cases. The operative approach is usually a midline partial or total laminectomy (Seppala et al 1995a; 1995b; Conti et al 2004). The tumors are always attached to at least 1 nerve root. Schwannomas can be completely resected in 85% to 90% of cases without sacrifice of the parent nerve root (Seppala et al 1995b; Conti et al 2004; Safavi-Abbasi et al 2008). Radical resection is also possible for large, invasive tumors that extend into surrounding bones and soft tissues (Sridhar et al 2001). For tumors of the thoracic spine, some authors recommend resection via thoracoscopy (Dickman and Apfelbaum 1998). This technique is an excellent alternative to thoracotomy because of the smaller incision used, better cosmetic results, reduced postoperative pain, and earlier return to activity. Neurofibromas can be completely resected in 90% of cases (Seppala et al 1995a). However, in 80% to 90% of these patients the parent nerve root must be sacrificed during tumor removal. Complications most often consist of hemorrhage, wound infection, deep venous thrombosis, wound dehiscence, and pulmonary compromise (eg, embolism, pneumonia).
Schwannomas and neurofibromas of the peripheral nerves are usually solitary lesions in patients without neurofibromatosis. Surgical resection of schwannomas, when they develop from distal nerve branches, consists of simple dissection of the parent nerve from the tumor capsule and en bloc removal of the mass (Ariel 1988). Neurofibromas are often more difficult to dissect away from the parent nerve, as it may become encased within the mass. Schwannomas of the large nerve trunks (ie, brachial or lumbosacral plexus) can also be completely resected in most cases with careful microdissection of the parent nerve away from the tumor capsule (Lusk et al 1987). In contrast, complete resection of neurofibromas often requires sacrifice of fascicles of the parent nerve and nearby nerves that have become encased within the mass. Adherent tumor must often be dissected from remaining nerve fascicles. Some authors recommend an interfascicular approach in combination with nerve conduction testing to maximize the chance for complete resection (Kim et al 2005). With the recent improvements in microneurosurgical techniques, some authors are recommending a "nerve sparing" approach to surgical resection of benign peripheral nerve schwannomas and neurofibromas (Russell 2007). Localized surgical approaches are not adequate for treatment of malignant schwannomas or neurofibromas (Sordillo et al 1981; Ariel 1988; Baehring et al 2003; Carli et al 2005; Gupta and Maniker 2007; Widemann 2009). These tumors generally require either radical local excision or amputation. Radical local excision involves en bloc removal of the tumor and any attached nerve, bone, muscle, and blood vessels. Normal tissue planes must be attained on all surfaces, or amputation is necessary.
Radiation therapy of vestibular schwannomas. For a small subgroup of patients with vestibular schwannomas, conventional external beam radiation therapy or stereotactic radiosurgery may be adjunctive or alternative forms of therapy (Noren et al 1983; Wallner et al 1987; Hirsch and Noren 1988; Pollack et al 1989; Linskey et al 1990; 1992; Flickinger et al 1991; Lunsford and Linskey 1992; Macfarlane and King 1995; Murphy and Suh 2011). Conventional radiation therapy is not indicated for patients after a complete or near-total resection. It should be considered in patients with substantial residual tumor after surgery, for recurrent tumors in advanced stages of disease, and for those patients with large tumors who are poor surgical candidates. In a review of 124 patients with vestibular tumors that received radiation therapy after subtotal tumor removal, Wallner and colleagues concluded that irradiation significantly reduced the possibility of tumor progression (Wallner et al 1987). Irradiation with doses of 5000 to 5500 cGy, administered in 180 cGy fractions, decreased the recurrence rate from 46% to 6% (p = 0.01). Preoperative irradiation can also be used to reduce the risk of hemorrhage during resection for patients with highly vascular schwannomas (Wallner et al 1987; Ikeda et al 1988). Doses of approximately 3000 cGy can reduce vascularity after a period of 6 to 8 weeks.
Stereotactic radiosurgery is a method of delivering focused irradiation within the boundaries of a tumor in a single fraction, using great precision (Loeffler and Alexander 1990; Battista 2009). The treatment is most often administered using a gamma knife; however, linear accelerator and proton beam units are also used and demonstrate comparable local control and complication rates (Mendenhall et al 1996; Suh et al 2000; Spiegelmann et al 2001; Friedman et al 2006). Similar to conventional external beam irradiation, radiosurgery is considered an alternative or adjunctive therapy in carefully selected patients with vestibular schwannomas (Linskey et al 1990; 1992; Flickinger et al 1991; Lunsford and Linskey 1992; Shetter 1997; Kondziolka et al 1998; Pollock et al 1998b; Prasad et al 2000; Suh et al 2000; Lunsford et al 2005; Likhterov et al 2007; Battista 2009; Murphy and Suh 2011). Patients most appropriate for radiosurgical therapy include those who are medically unstable, are elderly (greater than 65 years old), are contralaterally deaf, have failed an initial surgical resection, or refuse surgical intervention (Linskey et al 1992; Lunsford and Linskey 1992). Lesions less than 3 cm in diameter are most suitable for radiosurgery. The most common treatment plan involves a margin dose of 16 to 18 Gy (in a single fraction) at or above the 50% isodose line, depending on the estimated tumor volume (Flickinger et al 1991; Linskey et al 1992; Lunsford and Linskey 1992; Kondziolka et al 1998). Depending on the configuration of the tumor, 1 or more isocenters are used to cover the treatment volume (average 2.5 isocenters). Clinical response as measured by CT or MRI demonstrates tumor shrinkage in 20% to 30%, stable tumor in 60% to 75%, and tumor progression in 3% to 5% (Flickinger et al 1991; Linskey et al 1992; Lunsford and Linskey 1992; Shetter 1997). More recent papers with long term follow-up after radiosurgery have noted reduction in tumor size ranging from 62% to 81% (Kondziolka et al 1998; Prasad et al 2000; Lunsford et al 2005; Battista 2009). The estimated overall local control rate for radiosurgery is approximately 92% to 98% (Lunsford and Linskey 1992; Mendenhall et al 1996; Shetter 1997; Kondziolka et al 1998; Pollock et al 1998a; 1998b; Lunsford et al 2005; Friedman et al 2006; Likhterov et al 2007). A recent review of 208 consecutive patients, with median follow-up time of 56 months, noted growth of tumors 30 patients (14%) (Pollock 2006). Of this cohort, only 6 had progressive tumor enlargement that required further treatment with surgical resection or radiosurgery. The conclusion of the author was that initial enlargement after radiosurgery did not necessarily denote failure of the procedure. This has been corroborated in a recent study by Nagano and colleagues in which a 25% to 47% increase in tumor size was noted in over half their cohort of 100 consecutive vestibular schwannoma patients treated with radiosurgery. Peak tumor expansion was usually noted by 6 months after treatment and normalized within 12 months. High-dose treatment (ie, 3.5 Gy/min) was marginally correlated with the likelihood of tumor expansion (Nagano et al 2008). In a review of schwannomas that recurred after 1 or more attempts at surgical resection, radiosurgery was able to achieve growth control in 73 of 78 tumors (94%) (Pollock et al 1998a). The median interval from time of treatment to objective tumor shrinkage is approximately 12 months. Loss of central tumor enhancement is observed in 75% to 80% of cases after a median interval of 6 months. This is postulated to occur by radiation-induced vascular injury, thrombosis, and occlusion. The rate of useful preservation of hearing is 50% at 6 months and 38% at 1 year (Linskey et al 1992; Lunsford and Linskey 1992; Shetter 1997). Hearing begins to decline at a median of 6 months and is not correlated with loss of tumor contrast enhancement, tumor margin dose, or tumor margin isodose. Several recent reviews of hearing preservation after radiosurgery for vestibular schwannoma suggest an overall rate of preservation of approximately 51% at 3 to 4 years after treatment, with doses of less than 13 Gy being more likely to maintain hearing (Kano et al 2009a; Timmer et al 2009; Yang et al 2010). The degree of hearing preservation appears to be correlated to the maximal radiation dose at the cochlea, emphasizing the need for meticulous radiation planning. Some authors have also used gamma knife radiosurgery in patients with vestibular schwannomas that have failed initial radiosurgery with a different platform (ie, LINAC) (Dewan and Noren 2008). In 8 of the tumors, shrinkage was noted on follow-up MRI. Acute complications include nausea or vomiting (21%) and headaches (10% to 12%). Although rare, hearing loss and facial weakness within 24 to 48 hours of radiosurgical treatment have also been reported (Chang et al 1998; Tago et al 2000). Delayed facial neuropathy occurs in 34% of patients with normal preoperative function and 27% of patients with abnormal preoperative function. The facial weakness improves or recovers by 6 months in most patients. Radiosurgical doses of greater than or equal to 18 Gy appear to cause permanent facial neuropathy (Miller et al 1999). Patients receiving doses of less than or equal to 16 Gy were significantly less likely to develop facial neuropathy. Longer follow-up is required before conclusions can be drawn regarding efficacy of local growth control using the reduced dose protocol. Delayed trigeminal neuropathy occurs in 32% of patients with normal preoperative function and 46% of patients with abnormal preoperative function. Trigeminal dysfunction improves, but does not resolve within 6 months in most patients. The mechanism of cranial nerve injury remains unclear, but is probably a combination of direct radiation injury, localized edema, demyelination, and vascular compromise. Although data by Linskey and colleagues suggest that the length of cranial nerve irradiated correlates with the risk of delayed injury, additional studies indicate that the dose to the brainstem may be a more important predictor of post-treatment cranial neuropathy (Linskey et al 1993; Foote et al 2001). In a review of complication rates in a cohort of 190 patients treated with a median of 13 Gy, Flickinger and colleagues noted similar local control rates to other studies using higher doses; however, there were lower rates of hearing loss, facial numbness, and facial weakness (Flickinger et al 2001). Other potential complications include worsened balance (31%), vertigo (4%), and hydrocephalus. Another radiosurgical approach is to use protons, which confer a radiobiological advantage for beam accuracy. Weber and colleagues have used proton beam radiosurgery in a series of 88 patients, with 93.6%, 5-year tumor control rates and comparatively low rates of facial and trigeminal neuropathy (Weber et al 2003).
Some authors feel fractionated radiosurgery may be superior to standard radiosurgery, due to the radiobiological advantages inherent to fractionation, such as reduced risk of cranial nerve injury (Szumacher et al 2002; Sawamura et al 2003; Combs et al 2005; Likhterov et al 2007). In a series of 39 patients with vestibular schwannomas, Szumacher and colleagues noted local tumor control in 95% of the cohort at 21.8 months median follow-up (Szumacher et al 2002). At 22 months median follow-up, no new cases of cranial nerve dysfunction were noted. After fractionated radiosurgery, of those patients with functional hearing before treatment, 68% were able to maintain a similar level of function. Similar results have been reported by several groups in large cohorts of more than 100 patients (Sawamura et al 2003; Combs et al 2005). In contrast, other authors report that single-fraction linear accelerator-based radiosurgery may be as effective as fractionated methods in terms of tumor control rates (Meijer et al 2003). The Stanford group recently reported their experience using the Cyberknife, with minimal fractionation over 3 days (Sakamoto et al 2009). A total of 61 patients with vestibular schwannomas were treated, with a mean maximal tumor dimension of 18.5 mm. The local control rate was 98%; tumor shrinkage was noted in 29 tumors. None of the patients developed new facial weakness or trigeminal deficits. Some authors have also suggested that hypofractionated stereotactic radiotherapy can be effective (Sakanaka et al 2011). In a series of 27 patients with vestibular schwannoma, linear accelerator radiosurgery was administered to a total dose of either 30 to 39 Gy over 10 to 13 fractions, or 20 to 24 Gy over 5 to 6 fractions. Local control rates were 100% and 92% for the high- and low-dose cohorts, respectively. In addition, with this approach, minimal facial and trigeminal nerve morbidity was noted.
Currently, 20% to 25% of vestibular schwannomas are treated with radiosurgery either as initial therapy or at recurrence (Pollock et al 1998b). Pollock and colleagues predict that over the next 10 years to 20 years, radiosurgery will replace surgical resection as the initial treatment of choice for these tumors (Pollock et al 1998b). Nonrandomized comparisons between series of patients treated operatively and by gamma knife suggest similar tumor growth control and a reduced risk of hearing loss, facial palsy, facial hypesthesia, and feeding problems in the radiosurgical cohort (Regis et al 2002; Karpinos et al 2003; Pollock et al 2006). In a recent prospective, nonrandomized comparison of stereotactic radiosurgery (N = 46) and surgical resection (N = 36), Pollock and colleagues concluded that radiosurgery was more effective in terms of facial function and hearing preservation (p < 0.001), as well as on several subscales of the Health Status Questionnaire (eg, physical functioning, energy/fatigue) (Pollock et al 2006). The improved results of radiosurgery were only noted for patients with small and medium sized tumors. A similar study was recently published out of Norway by Myrseth and colleagues in which a cohort of 91 prospective patients with small- to medium-sized vestibular schwannomas were treated with gamma knife radiosurgery (N = 63) or open microsurgery (N = 28) (Myrseth et al 2009). The results were again in favor of the radiosurgery cohort, with better preserved facial nerve function and hearing (P < 0.001) than the surgical group. In addition, quality of life measures were significantly better in the radiosurgery cohort.
Radiation therapy of nonvestibular schwannomas. Several authors have used radiosurgical approaches to treat nonvestibular schwannomas with excellent results, similar to the data for vestibular tumors. Pollock and colleagues treated 23 patients with tumors of the trochlear, trigeminal, jugular, and hypoglossal nerves (Pollock et al 2002). Local control rates were 96%, with a 17% rate of cranial nerve morbidity. Similar results are reported by Zhang and colleagues in a series of patients with jugular foramen schwannomas (Zhang et al 2002). Gamma knife radiosurgery has been applied to a series of 21 patients with trigeminal schwannomas (Phi et al 2007). Tumor growth control was achieved in 95% of cases. In 6 patients (27%), new or worsening cranial neuropathies were noted after treatment. In a more recent series of 33 patients with trigeminal schwannomas, progression-free survival rates at 5 and 10 years were both 82.0% after gamma knife radiosurgery (Kano et al 2009b). In a similar study of jugular foramen schwannomas, stereotactic radiosurgery was applied to 35 tumors (Martin et al 2007). Tumors regressed in 17 patients and remained stable in 16 patients. The 5- and 10-year actuarial control rates were 97% and 94%, respectively. Preexisting cranial neuropathies improved in 20% of cases and remained stable in 77%. Recent data from several groups suggest that gamma knife radiosurgery can also be of benefit for patients with facial nerve schwannomas (Litre et al 2008; Madhok et al 2009). In a series of 11 patients, 10 remained stable or regressed after treatment, with satisfactory preservation of facial nerve function (Litre et al 2008). Similar results were noted from Madhok and colleagues in a series of 6 patients (Madhok et al 2009). A small series of 8 patients with schwannomas of cranial nerves III, IV, and VI have been reported by Kim and colleagues (Kim et al 2008). All of the tumors had shrinkage on follow-up MRI, and several patients had improvement in diplopia.
Some authors have applied fractionated radiosurgical techniques to nonvestibular schwannomas. Zabel and coworkers used a median dose of 57.6 Gy with 1.8 Gy fractions on 13 patients (Zabel et al 2001). Local control rates were 100%, with 4 tumors decreasing in size. No new cranial nerve or brainstem deficits were noted. After maximal surgical resection, external beam radiotherapy should also be considered in selected patients with malignant peripheral nerve sheath tumors, especially those with gross residual disease (Baehring et al 2003; Carli et al 2005; Gupta and Maniker 2007; Widemann 2009).
Chemotherapy. In general, chemotherapy has not been applied to patients with sporadic schwannomas or neurofibromas. A small phase I trial of thalidomide, an angiogenesis inhibitor, has been completed in patients with neurofibromatosis type 1 and plexiform neurofibromas (Gupta et al 2003). The drug was well tolerated up to doses of 200 mg/day. Several patients were noted to have minor responses (less than 25% reduction in size) and stabilized disease. Plotkin and colleagues recently tested a series of 43 patients with unresectable neurofibromatosis type 2-related and sporadic vestibular schwannomas for expression of vascular endothelial growth factor (VEGF) and VEGF receptors and attempted treatment with bevacizumab, a humanized monoclonal antibody against VEGF (Plotkin et al 2009). VEGF was expressed in 100% of the tumors, with expression of VEGFR-2 noted in 32% of all tumor vessels. Ten patients were treated with bevacizumab (5 mg/kg intravenously every 2 weeks). Tumor shrinkage was noted in 9 patients on follow-up MRI, with a median best response to treatment volume reduction of 26%. Several patients had durable responses that were maintained over 11 to 14 months. In addition, 4 of 7 evaluable patients had some improvement in hearing. These results are consistent with subsequent case reports of patients treated with bevacizumab. Mautner and colleagues reported 2 patients with NF2-related vestibular schwannomas treated with bevacizumab, both of which had tumor regression of 40% or more (Mautner et al 2010). One of the patients was treated for 6 months and also had improved hearing. Other authors have evaluated anti-VEGF therapy in an animal model (Wong et al 2010). They treated rats implanted with HE1193 or murine NF2-/- tumors with either bevacizumab (10 mg/kg/week) or vandetanib (50 mg/kg/day). There were improvements in tumor vessel diameter, length/surface area density, and permeability in both types of tumors, with both drugs. An increase in necrosis was noted in HE1193 tumors, whereas the NF2-/- tumors were noted to have increased apoptosis. In addition, the tumor growth rate was decreased by 50%, and the survival time was increased by 50% in both drug groups. Although chemotherapy is not generally considered to be of benefit for patients with malignant peripheral nerve sheath tumors, some authors have noted modest benefit in pediatric patients using ifosfamide-based regimens (Carli et al 2005; Gupta and Maniker 2007).
In vitro and animal model experiments with molecular chemotherapy drugs are now underway. Using an in vitro model system of schwannoma, Ammoun and colleagues have shown that sorafenib (BAY 43-9006), an inhibitor of PDGFR and c-Raf, can inhibit PDGFR-beta-mediated ERK1/2 and Akt activity and can reduce cell proliferation in schwannoma cells (Ammoun et al 2008). Another study investigated the activity of AZD6244, an inhibitor of MEK1/2 that is involved in the activation of the extracellular signal-regulated kinase pathways and cell proliferation in schwannoma cells (Ammoun et al 2010b). At low concentrations, AZD6244 was able to abolish platelet-derived growth factor-mediated activation and cell proliferation in schwannoma cell lines. Another follow-up study by Ammoun and colleagues used receptor tyrosine kinase arrays to screen tumor samples from NF2-related and sporadic vestibular schwannomas (Ammoun et al 2010a). Eleven patient samples and 2 control samples were analyzed; all of the tumors were positive for activated epidermal growth factor receptor, and more than half were also positive for activated ErbB2 and ErbB3. Activated ERK1/2 was also noted in all of the tumor samples. Based on these results, the small molecule EGFR/ErbB2 inhibitor, lapatinib, was used to treat in vitro human schwannoma cells. Lapatinib was able to inhibit ErbB2 phosphorylation and survivin upregulation, as well as downstream ERK1/2 and Akt activation. Proliferation of schwannoma cells was inhibited. OSU-0312, a newly developed drug that selectively targets the Akt pathway via inhibition of PDK1, has been shown to suppress schwannoma cell proliferation and decrease phospho-Akt in culture systems (Jacob et al 2008). Recent follow-up studies from the same authors have noted that OSU-03012 has activity against typical and malignant schwannoma cells (Lee et al 2009). The drug was able to inhibit cell proliferation in culture, with an IC50 of 2.6 to 3.1 μM. OSU-03012 was able to induce apoptosis in regular and malignant cell lines while markedly reducing Akt phosphorylation. In xenograft models with the malignant cell line, OSU-03012 was able to inhibit growth by 55% after 9 weeks of oral treatment. Using a mouse schwannoma xenograft model, Clark and colleagues treated animals with trastuzumab, erlotinib, or saline. Both trastuzumab and erlotinib significantly reduced the growth of schwannoma xenografts in comparison with controls (p < 0.05). Erlotinib, but not trastuzumab, was noted to induce a significantly higher rate of apoptosis in schwannoma cells (p < 0.01) (Clark et al 2008). Erlotinib (150 mg/day) was recently tested in a series of 11 patients with NF2 who had progressive vestibular schwannomas (Plotkin et al 2010). There were no patients with objective MRI responses, although minimal shrinkage was noted in a few. Several patients had stabilization of disease, with a median time to progression of 9.2 months. Hearing responses were not significantly improved in any of the cohort.
Mukherjee and colleagues evaluated sporadic and neurofibromatosis type 2 (NF2)-related schwannomas for the presence of platelet-derived growth factor receptor (PDGFR) and c-kit receptors to assess potential sensitivity to imatinib mesylate, a tyrosine kinase inhibitor of PDGFR and c-kit (Mukherjee et al 2009). Increased expression and activation of PDGFR-alpha, PDGFR-beta, and c-kit receptors was noted. Imatinib mesylate was able to inhibit proliferation and anchorage-dependent growth of NF2-null HEI-193 schwannoma cells. In addition, imatinib was able to induce apoptosis in a dose-dependent manner. Similar work by Altuna and colleagues noted expression of PDGFR in 67.5% of vestibular schwannoma samples (Altuna et al 2011). After treatment with imatinib (5 or 10 μM), there was downregulation of phospho-PDGFR. In addition, the use of imatinib induced a dose-dependent increase in G1 percentage (61.6% to 70.7% and 74%; at 5 or 10 μM, respectively), during cell cycle analysis. Imatinib was also able to induce a dose-dependent growth inhibition in colony formation assays using cell lines and cultures derived from fresh tumor tissue.