Observation for VS

Progressive growth of tumors in the cerebellopontine angle carries a high rate of morbidity, with cranial nerve problems (including hearing loss), brainstem or cerebellar compression, and eventually hydrocephalus from compression of the fourth ventricle. However, the natural history of VS is unpredictable, especially for sporadic VS. Tumors may grow at a steady rate, plateau at a certain size, or rarely even shrink. Additionally, tumors may recur after treatment. Given the increasing availability of cranial imaging and the high resolution now available in routine brain MRIs, diagnosis of small VS has now become commonplace.

Several studies have attempted to elucidate the natural history of VS. The largest to date is a prospective evaluation of tumor growth data from more than 500 patients in the Danish national registry, followed for a median of 3.6 years. During the follow-up period, 17% of tumors confined to the internal acoustic meatus grew, compared with 29% of extrameatal tumors. No relationship between tumor growth and patient gender or age could be determined. The same registry was used to assess changes in hearing over time in patients allocated to expectant management strategies. Patients received annual audiologic and MRI examinations. At the outset of the study, 53% of patients had speech discrimination of greater than 70%; at the end of the 10-year observation period, this rate had fallen to 31%. Patients with hearing that was initially good maintained their hearing at higher rates than patients with impaired hearing at baseline.

These results seem to confirm that primary observation is justified for smaller, asymptomatic tumors. Indeed, a “watchful waiting” approach is common in clinical practice, particularly given the relatively slow growth of most VS and high risk of treatment-related morbidity including complete hearing loss. Larger, symptomatic tumors, or those with substantial cystic components, are often removed because they are more likely to cause morbidity related to progressive growth, and surgery becomes more challenging and risky with larger tumors. Thus, the chosen therapeutic approach may depend on the clinician’s interpretation of the evidence for ongoing tumor growth. Consensus recommendations vary, but commonly use a tumor size threshold of 20 mm and a growth rate of greater than 2 mm per year (based on increase in the largest dimension) for recommendation of treatment. At this size, further growth may compromise cranial nerve function, and may also make future surgery more difficult. Additionally, an actively growing tumor is more likely to result in progressive hearing decline over time, especially if the tumor is already large at diagnosis.

Surgery for VS

The decision to manage VS surgically is often straightforward for large, symptomatic tumors and those causing brainstem compression or hydrocephalus. However, surgery is frequently performed for smaller, asymptomatic tumors. The most compelling argument for surgery is the theoretical risk of progressive hearing deterioration over time. Larger tumors are particularly problematic to manage conservatively. Based on what is known about the natural history of VS, progressive hearing loss is more likely with rapidly growing tumors, especially those that are large at baseline. The morbidity of surgery increases with tumor size, because of the difficulties of exposure including cerebellar retraction and facial nerve damage. These issues arise frequently in surgical treatment of NF2-related VS, because they have a markedly faster growth rate than seen in sporadic VS and may present at a larger size. Additionally, NF2 patients with bilateral VS may already have impaired hearing in one or both ears, raising the stakes for a high-risk surgery. Although hearing may stabilize with either surgery or radiation therapy, it usually does not improve.

As is the case for many complex, high-risk surgical procedures, outcomes for surgical treatment of VS are superior at larger institutions with more experienced, specialized surgeons. A retrospective cohort study of 2643 operations selected from a large national administrative database demonstrated a significant volume-outcome relationship for VS resection when adjusting for potential confounders such as age, gender, race, payor (type of insurance), geographic region, procedure timing, admission type and source, medical comorbidities, and NF status. Postoperative complications including neurologic deficits, prolonged mechanical ventilation, facial nerve injury, and need for blood transfusion were less frequent at higher-volume hospitals or with higher-volume surgeons. Furthermore, length of stay was shorter, hospital charges were less, and perioperative mortality was reduced. The potentially high risks of short- and long-term morbidity from VS surgery and the relatively small number of tertiary care centers that have the requisite caseload and surgical experience to manage these risks further complicate the delivery of care for patients with VS, not all of whom have access to these settings.

Radiation Therapy for VS

Radiation is often used as adjuvant therapy for treatment of sporadic brain tumors. As the other mainstay of treatment for VS, radiation therapy can help avoid some of the morbidities associated with surgery. Although many larger VS are believed to be poor candidates for radiotherapy because of the concomitant risks of swelling near the brainstem, recent technical advances including fractionated radiation and radiosurgery have been used with some success to treat smaller tumors. However, outcomes for NF2-related VS are generally worse than those for sporadic VS.

In particular, fractionated stereotactic radiotherapy (FSRT) has been advocated to minimize the risk of hearing loss. A recent series of 106 patients treated with FSRT demonstrated excellent local tumor control (94% at 3 years and 93% at 5 years). Notably, the “useful” hearing preservation rate was 98% at 5 years in patients without NF2 but only 64% in patients with NF2. The median irradiated tumor volume was small (3.9 cm ³ ), although the maximum tumor volume was greater than 30 cm ³ . FSRT was generally well tolerated, with a low incidence of cranial neuropathies including radiation-induced toxicity to the trigeminal and facial nerve (<5%).

Another series of 190 previously untreated patients with unilateral VS examined the efficacy and complication rate of gamma knife radiosurgery. The 5-year tumor control rate, as defined by no requirement for surgical intervention, was 97%. Patients had low rates of radiation-induced cranial neuropathies, although increasing marginal radiation dose did lead to higher rates of facial nerve weakness and decreased preservation of speech discrimination. Hearing-level preservation was 71% and preservation of speech discrimination was 91%, although the authors do not stratify by NF2 status; most patients in the sample had unilateral VS, which were more likely to be sporadic. As with surgery, hearing improvement was rare (7%), although many patients preserved their pretreatment hearing status.

In patients with smaller tumors, or in older patients with medical comorbidities, radiosurgery or FSRT may be an attractive alternative to neurosurgery, with lower rates of treatment-associated cranial neuropathies. Additionally, radiosurgery avoids problematic perioperative complications such as infection and cerebrospinal fluid leak. Patients successfully treated with radiotherapy may be able to avoid surgery in the future. However, for larger tumors, radiotherapy is often impractical or unsafe. In the NF2 population, the risk of secondary malignancy after radiation is a particular concern. These patients are usually younger than sporadic patients at the time of treatment (with more time to develop a malignancy) and have a genetic predisposition to tumor formation.

Hearing End Points

Hearing preservation is of crucial importance to patients, particularly those with NF2 who have bilateral tumors and frequently experience the loss of all usable hearing within their lifetimes as a result of tumor growth or treatment. Because most NF2 patients are diagnosed well after their acquisition of language, this hearing loss has a profound effect on their quality of life, social skills, and ability to function at home and in the workplace. Subjective measures of hearing, although they may be highly relevant to the patient’s perceived symptoms and quality of life, are problematic for use as a primary outcome in a clinical trial.

Several methods have been suggested for the objective measure of changes in hearing, although no single method has been adopted for routine use in drug trials. The standard audiogram commonly incorporates the ability to detect sound (pure tone threshold) and speech discrimination (word recognition score based on standardized lists of words). For example, the American Academy of Otolaryngology recommends testing pure tone thresholds at levels of 0.5, 1, 2, and 3 kHz, and testing speech discrimination scores at up to 40 dB above the speech discrimination threshold using a 50-word recorded list in the subject’s native language. A composite hearing scale, the Gardner-Robertson classification, is sometimes used to report changes in hearing after treatment with open surgery or radiosurgery, combining pure tone threshold and word recognition into a four-level scale. However, this scale lacks sensitivity because of “floor effects” for those in lower categories (classes III, IV, and V), in which NF2 patients are more likely to fall. Time to hearing loss or failure has been proposed as an end point for clinical trials, but given the difficulty of assessing subjects with severely impaired hearing at baseline and the long follow-up times required, it may be impractical to use as a primary outcome.

The authors suggest that maximum word recognition scores be used as the primary hearing end point in phase II clinical trials for anti-VS drugs. Word recognition affects basic communication more profoundly than does pure tone threshold. VS have a propensity to cause auditory nerve dysfunction because of direct pressure and secondary degeneration of the cochlea. Indeed, hearing aids, although they decrease the pure tone threshold, are often of little benefit for patients with NF2 and bilateral VS, because sound amplification alone does not address the underlying auditory nerve dysfunction.

Finally, recent evidence on drug therapies for VS suggests that patients treated with targeted agents can experience measurable and clinically meaningful improvement in their hearing even in the absence of a radiographic response or with progressive disease as measured by MRI scan (discussed next). This may be caused by the amelioration of the auditory nerve dysfunction caused by tumor growth and tumor-associated edema. Intuitively, a clinically meaningful improvement in hearing is the most important outcome from the perspective of a patient agreeing to participate in a clinical trial, and probably the most objective and reliable way to achieve this is by measuring the word recognition score.

Radiographic End Points

Traditionally, changes in contrast enhancement on MRI or CT have been used to define objective radiographic tumor response to treatment. The neuro-oncology literature includes several novel modalities to measure radiographic response of malignant CNS tumors, including diffusion or perfusion changes in MRI imaging. However, as with most benign CNS tumors, these methods have not been validated in either sporadic or NF2-related VS.

Historic methods of measuring tumor size and response, such as the 1981 World Health Organization criteria, revised in 2000 as the RECIST criteria, and the Macdonald criteria for malignant glioma, have all been used in the VS literature. Linear growth (based on largest tumor dimension) is often used as a treatment recommendation threshold, most commonly 2 mm per year. However, NF2-related VS in particular are often larger and have a more irregular shape than sporadic VS, so linear dimension, whether unidimensional or bidimensional, is less likely to be an adequate surrogate for volume. More recently, and with the advent of higher-resolution MRI imaging, volumetric measurement has become the accepted method of defining radiographic response. Using 3-mm axial slices through the internal auditory canal produces the greatest accuracy, with an intrarater coefficient of variation of less than 5% for tumors larger than 1 cm ³ . In a recent paper on suggested response criteria, the authors suggest volumetric methods, with an objective radiographic response defined as a 20% or greater reduction in VS volume based on postcontrast T1-weighted MRI images collected with 3-mm or finer cuts through the internal auditory canal.

Radiographic tumor response has been successfully used in phase II chemotherapy trials of various cancers. The underlying assumptions are that most chemotherapeutic drugs work by a cytotoxic mechanism, resulting in radiographic shrinkage, and that objective decrease in tumor size is related to other end points, such as overall survival or progression-free survival, that are clinically relevant to patients. However, some agents are postulated to have cytostatic, rather than cytotoxic, effects, and may result in radiographic stabilization (or decrease in growth rate) rather than overt shrinkage. For sporadic VS, spontaneous involution occurs in approximately 8% of tumors, usually during a long period of observation given their slow growth rate. However, this rarely occurs in NF2-related tumors; in a recent analysis, only 3 (3.6%) of 84 tumors regressed by 2 mm or more during 9 months to 2 years of follow-up, suggesting that any objective radiographic shrinkage of NF2-related schwannomas would be caused by drug activity and not by spontaneous involution.

It is well known that contrast enhancement may be misleading, reflecting radiation necrosis or “treatment effect” rather than tumor growth. In such cases, time to disease progression, rather than objective radiographic response, may be a more reliable indicator of the drug’s clinical efficacy. Indeed, progression-free survival at 6 months is now a frequently used primary end point in drug trials for recurrent malignant glioma. Although histologically benign, given their problematic location and rapid growth, it is appropriate to consider NF2-related VS in a similar clinical category as malignant CNS tumors. In this setting, determination of progression-free survival at 12, 24, and 60 months would provide meaningful clinical information about drug activity for VS.

A final consideration relevant to NF2-related VS is the presence of multiple lesions. In principle, one target lesion (usually the largest or most symptomatic lesion) should be selected as the lesion of interest for assessing the primary end point of radiographic response. In NF2-related VS, this can be difficult because of the irregularity of the tumors, changes in previously treated tumors, and the presence of multiple or confluent tumors.

Tinnitus and Other End Points

In addition to the possible primary end points described previously, the design of future clinical trials should take into consideration subjective improvement in patient symptoms. Among the most bothersome of these symptoms is tinnitus. As with many subjective symptoms, there is no diagnostic test for tinnitus. Measurement of tinnitus is complicated by the fact that it is an extremely common symptom, with up to a third of the adult population reporting it at some point, although few individuals experience symptoms significant enough to affect their quality of life. An individual’s subjective emotional and psychologic responses to tinnitus can also affect its perceived severity, making an objective measurement more difficult.

Tinnitus can be a sensitive indicator of cochlear dysfunction. It can encompass the clinical spectrum from barely noticeable “white noise” to troublesome or even distressing whistling and ringing that can be heard despite loud background noise. In VS, as in Meniere’s disease, it may correlate with hearing loss, although it may fluctuate less as hearing loss progresses. A clinical grading system for tinnitus has been proposed, assigning a value of 0 to 3, which incorporates degree of severity and continuity versus fluctuation. More recently, a five-point severity grading scale (slight, mild, moderate, severe, and catastrophic) has been proposed, but no one system has been accepted for general use in clinical trials. In their studies, the authors have used a grading scale along with the Tinnitus Reaction Questionnaire, which assesses the psychologic distress associated with tinnitus.

Vertigo and balance problems are common in NF2 patients with bilateral VS, but are notoriously difficult to measure objectively. Although there is a well-validated objective scale for facial nerve palsy, the House-Brackmann scale, facial nerve dysfunction is often a consequence of surgical treatment rather than of the natural history of the disease. The same is true of trigeminal nerve dysfunction. Therefore, these entities are unlikely to be used as primary clinical end points, although they certainly have a profound influence on the patient’s overall quality of life.