Historical perspective

Archeological findings from 2500 BC provide evidence that acoustic nerve tumors have been present since antiquity. These tumors, called acoustic neuromas or, more properly, schwannomas, were diagnosed based on a recognized progression from deafness to death as early as 1810. The first documented successful removal of an acoustic schwannoma was performed by Thomas Annandale in 1896 in Edinburgh, Scotland. The patient, who was pregnant at the time of the operation, was able to go home and give birth to a healthy child .

During the twentieth century, different surgical approaches (suboccipital, translabrynthine, and middle fos) were developed for the resection of eighth-nerve tumors. Each sought to diminish the inherent anatomic issues associated with total removal. In addition to the obvious hearing loss associated with eighth nerve resection, there was also the risk of seventh nerve damage, cerebrospinal fluid leak, and hemorrhage. Because of difficulties with early diagnosis before modern imaging tests, the pattern of slow intermittent growth, and the morbidity of resection, observation instead of immediate intervention became a frequent consideration.

In 1951, Dr. Lars Leksell, a neurosurgeon, and Borje Larsonn, a physicist, using the Uppsala University cyclotron, ion Uppsala, Sweden, developed an approach to treating small brain lesions with multiple proton beams, using a fixed rigid Cartesian coordinate system to locate the target. Their idea was to develop a noninvasive therapy system to deliver ablative doses of radiation to a geometrically defined discrete volume of tissue, using multiple small beams of radiation. This concept evolved into the Gamma Knife (Elekta, Stockholm, Sweden) stereotactic radiosurgery. The Gamma Knife unit consists of 201 cobalt radiation sources placed in a helmet, within which lie shuttered channels directed toward the center of the helmet. The target lesion is placed at that center position of the helmet by using a stereotactic frame affixed to the patient’s skull. Different dose patterns with sharp dose gradients can be obtained using multiple isocenters designed to match the shape of the tumor (Appendix).

Dr. Leksell and his team working at the Karolinska Institute in Stockholm, Sweden were the first to treat acoustic neuromas with stereotactic radiosurgery (SRS). Their initial series (1969–1974) reported initial tumor control in eight of the nine cases; however, hearing loss was reported in the majority of patients treated . In these early studies, high doses (25–35 Gray) were employed; targeting was crude compared with imaging studies available today. The first Gamma Knife unit in the United States was installed at the University of Pittsburgh in Pittsburgh, Pennsylvania under the direction of Dr. Dade Lunsford. Many of the early results involving radiosurgery for acoustic neuromas in North America were published by the University of Pittsburgh group.

An alternate stereotactic delivery platform was developed using linear accelerators (linacs). Linacs were already available in most modern radiation centers. Instead of multiple sources aimed at a central designated target, the linac systems rotated the treatment beam of the unit around the target in a varying number of rotational arcs. Betti and Colombo initially developed this technique in South America. Because the linac systems involve moving sources of radiation, special devices are needed to limit positional variation in beam delivery.

Both of these techniques require some method to limit patient and target movement. Fixation of target position in a geometric coordinate system can be achieved with metal frames screwed into the skull, which can obtain submillimetric accuracy, or with thermoplastic molded mask systems, which allow for more fractionated schedules, whereby treatment can be administered over several days without placing fixation screws into a patient’s skull.

Treatment goals

The goal of treatment is to eradicate the effects of a tumor with the fewest side effects. In surgical series, the means of achieving this goal is the removal of the offending tumor. In most reports of radiation efficacy, the objective is to achieve tumor reduction; however, in slow-growing tumors such as acoustic schwannomas, the achievable goal in studies with shorter duration of follow-up is the prevention of both tumor growth and symptom progression. Tumor shrinkage assessed by radiographic size may take years before a complete response can be evaluated. Additionally, relapses might be late occurrences.

Increased tumor volumes and irregularly shaped tumor volumes are associated with increased integral dose (energy deposition) within the tumor for Gamma Knife and cone-based linac SRS systems. Gamma Knife doses are prescribed to the tumor periphery. These peripheral doses are frequently 50% of the peak dose delivered within the central volume treated. This dose profile with areas of very high dose within the target volume (hot spots) is less a problem with single isocenter treatment plans using a multileaf collimator, where dose is more uniform. Shaped-beam peripheral doses are usually prescribed to the 90% to 95% isodose line. Central hot spots are 10% to 20% higher than the dose prescribed to the tumor periphery; however, the multileaf collimated treatment plans do not produce dose patterns that are as conformal or tightly fitting as the multiple isocenter systems or Gamma Knife systems. Single isocenter treatment techniques typically treat a small rim of normal tissue surrounding an irregularly shaped tumor.

Treatment goals

The goal of treatment is to eradicate the effects of a tumor with the fewest side effects. In surgical series, the means of achieving this goal is the removal of the offending tumor. In most reports of radiation efficacy, the objective is to achieve tumor reduction; however, in slow-growing tumors such as acoustic schwannomas, the achievable goal in studies with shorter duration of follow-up is the prevention of both tumor growth and symptom progression. Tumor shrinkage assessed by radiographic size may take years before a complete response can be evaluated. Additionally, relapses might be late occurrences.

Increased tumor volumes and irregularly shaped tumor volumes are associated with increased integral dose (energy deposition) within the tumor for Gamma Knife and cone-based linac SRS systems. Gamma Knife doses are prescribed to the tumor periphery. These peripheral doses are frequently 50% of the peak dose delivered within the central volume treated. This dose profile with areas of very high dose within the target volume (hot spots) is less a problem with single isocenter treatment plans using a multileaf collimator, where dose is more uniform. Shaped-beam peripheral doses are usually prescribed to the 90% to 95% isodose line. Central hot spots are 10% to 20% higher than the dose prescribed to the tumor periphery; however, the multileaf collimated treatment plans do not produce dose patterns that are as conformal or tightly fitting as the multiple isocenter systems or Gamma Knife systems. Single isocenter treatment techniques typically treat a small rim of normal tissue surrounding an irregularly shaped tumor.

Stereotactic radiosurgery

The initial studies from Pittsburgh demonstrated the ability of stereotactic radiosurgery to achieve control rates of 95% or more with doses of 16 to 20 Gray; however, these doses were associated with an increased risk of trigeminal nerve injury, seventh nerve injury, and decreased hearing. Before 1992, when using these higher doses, reports of treatment results in patients who were in the Pittsburgh patient cohort receiving stereotactic radiation demonstrate that 34% developed facial nerve weakness, although one half of these were transient. Thirty-two percent developed new trigeminal neuropathy. Useful hearing could be preserved in 38% at 1 year . Foote and colleagues reported similar side-effect profiles from the University of Florida in Gainesville, Florida when higher doses were used, with fewer side effects when doses lower than 13 Gray were prescribed to the tumor margin.

Reduced toxicity to the trigeminal and facial nerves was accomplished by decreasing the marginal SRS dose given to the tumor. Rates of morbidity decreased from 29% to 5% for facial neuropathy, and similarly, to 2% or less for trigeminal neuropathy. Tumor control rates did not appear to be compromised until the marginal dose was decreased to 10 Gray or less . Doses to the tumor margin of 12 to 13 Gray were associated with 5-year tumor control rates of 92% to 98% . Maximum doses were 20 to 26 Gray. Useful hearing was noted in 56% to 78% after treatment.

Optimal dose prescription balances tumor kill and normal tissue survival. Radiation oncologists have long recognized the limited tolerance of cranial nerves to radiation. The best-delineated toxicity data for stereotactic radiation relate to the optic nerves and chiasm. Limits of 8 Gray in one fraction and 50 Gray using standard fractionation (1.8 to 2 Gray per day), serve as dose restraints for this nerve. Radiation toxicity must also take into consideration the length of nerve radiated as well as the dose absorbed. The idea of volume at risk may be a function of vascular damage or of repopulation limits. The dose given to the tumor margin and more plausibly, the maximal dose within the volume treated reflect the degree of tissue damage.

A group from Seoul, Korea related hearing loss in relationship to the cochlear dose received during radiosurgery . The mean dose to the cochlea in those maintaining useful hearing was 6.9 Gray. When the mean dose was greater than 11 Gray, hearing declined. Massager and colleagues also found cochlear dose to be lower in patients retaining useful hearing. They found a significant relationship regarding intracanalicular tumor volume (<100 mm ³ versus 100 mm ³ ) as well as intracanalicular integrated dose as determinants of hearing loss. Their paper postulates that “hearing worsening after the gamma knife radiosurgery (GKR) procedure can be attributed to cochlear injury inside the internal acoustic canal caused by the enlargement of the intracanalicular part of the vestibular schwannomas during the inflammatory edema phase after radiosurgery through an increase of the intracanalicular pressure.”

Fractionated stereotactic radiation therapy

Whereas some centers were investigating lowering the marginal peripheral dose with single dose treatment regimens, others investigated using fractionated radiation therapy. The theory behind fractionated or multiple treatment radiation therapy is that multiple smaller doses of radiation can achieve a similar tumor effect (cell death) while allowing normal tissue time to repair between each dose, and thereby limit toxicity. This approach involves a greater total dose of radiation than a single-dose treatment to overcome whatever repairs the tumor has been able to achieve. Standard fractionation involves doses of 1.8 to 2.0 Gray per day given daily for 25 to 30 treatments. A total dose to tumors as measured at their periphery is 45 to 60 Gray. Central portions of the tumor volume receive 5% to 10% more than the periphery. This regimen has been used for decades in the treatment of malignant tumors to maximize soft tissue repair from radiation damage.

Several institutions have reported their results for fractionated radiotherapy using radiosurgical techniques in the treatment of acoustic schwannomas, with excellent tumor control rates and with minimal toxicity. Relocatable, molded face masks have been used for skull immobilization. These treatments have been delivered using linac-based therapy, with tumor doses prescribed to a peripheral dose encompassing the tumor, plus a small margin (2 mm) to account for the small amount of movement that occurs between daily fractions and any movement within the face mask during treatment.

Chan and colleagues , from Massachusetts General Hospital-Harvard in Boston, Massachusetts, report a 5-year tumor control rate of 98% using a regimen of 54 Gray given in 1.8 Gray fractions as prescribed to the 95% isodose line. They note a distinct relationship between tumor sizes and tumor control. Surgical resection was required for three patients with larger tumors and increasing symptoms at a median of 37 months. At 5-year follow-up, freedom from any surgical intervention was 97% for tumors smaller than 8 cm ³ , and 47% for tumors greater than 8 mm ³ .

Selch and colleagues from the University of California, Los Angeles (UCLA), using a similar radiation regimen—54 Gray in 30 treatments as prescribed to the 90% isodose line—reported a local control rate of 100% at 36-month median follow-up in 50 patients. Useful hearing was preserved in 93%, with a median follow-up of 36 months. Facial numbness occurred in 1 patient (2.2%) and 1 patient experienced the new onset of facial palsy. Twelve of their patients experienced tumor growth. In 6 of the 12, the growth was transient, and was felt to represent a treatment effect. The transient type of enlargement shows subsequent shrinkage within 2 years, and is frequently associated with loss of central enhancement on MRI. The phenomenon of transient enlargement has also been a common finding in other institutions for both SRS and fractionated stereotactic radiation therapy (FSRT) .

A Heidelberg, Germany group treated 106 patients who had acoustic neuromas using standard fractionation, given to a total dose of 57.6 Gray. Local control at 5 years was 93%. Trigeminal and facial toxicity were 3.4% and 2.3%, respectively. Useful hearing was preserved in 94% . In a more recent publication by that same group, Combs and colleagues report that hearing preservation in patients who had useful or serviceable hearing before radiation therapy was 55% at 9 years after SRS, compared with 94% showing serviceable hearing 5 years after FSRT.

Attempts to decrease toxicity by decreasing total tumor dose for ractionated stereotactic radiotherapy have also been described. Thomas Jefferson University in Philadelphia, Pennsylvania presented a retrospective analysis showing no tumor control difference in two cohorts of patients treated with either 50.4 Gray or 46.8 Gray. Although tumor control rates were equivalent (98% versus 100%) with a median follow-up of 3 years, hearing preservation was better in the lower dose group. Hearing preservation was measured by pure tone averages and speech discrimination. Corrected for follow-up and initial hearing, the rate of preservation was 93% for the low-dose group versus 67% for the higher-dose cohort. The median follow-up time for the low-dose group was 29 months . A group at Hokkaido, Japan used 40 to 50 Gray in 20 to 25 fractions. Their actuarial tumor control rate at 5 years was 91%, with no new permanent facial weakness. The rate of useful hearing preservation (Gardner-Robertson Class I or II) was 71%. Complications were mild—transient facial nerve palsy was 4%, trigeminal neuropathy was 14%, and balance disturbance occurred in 17% of patients .

In another low-dose FSRT study, Shirato and colleagues matched a group of patients who had vestibular schwannoma who underwent observation only against a cohort of patients treated with fractionated radiotherapy delivering 36 to 44 Gray in 20 to 22 treatments. The conclusion of the study was that there were no differences in the actuarial Gardner/Robertson hearing class preservation curves after the initial presentation. The rate of hearing deterioration in the treated arm was comparable to that of untreated patients. The mean growth of the tumor in the observation arm was 3.87 mm per year, whereas there was tumor reduction in the radiated cohort.

Hypofractionation

In an attempt to maximize hearing preservation rates without the need for several weeks of daily radiation treatments, a third alternative—hypofractionation—has also been studied. Using biological modeling to provide theoretically equivalent results as standard fractionation, hypofractionation gives higher doses per treatment for fewer treatments than standard fractionation schemes, but less dosage per day than single-dose prescriptions. Hypofractionation regimens use doses in the range of 3 to 7 Gray per day for 3 to 10 days, for total doses in the range of 21 to 30 Gray.

Meijer and colleagues from Vrije Universiteit University Medical Center in Amsterdam used a hypofractionation schedule of 4 to 5 Gray for 5 days as measured at the 80% isodose line. The 20 to 25 Gray was delivered in 1 week. Five-year local control in 80 patients was 94%. Facial nerve function was preserved in 97%. The study authors compared these patients to a group of 49 patients treated at the same institution with a single fraction of 10 to 12.5 Gray, and found no significant differences in outcome in regard to tumor control or facial nerve damage. Five-year hearing preservation favored the fractionated group (75% versus 61%). At Johns Hopkins in Baltimore, Maryland, a similar rate (70%) of hearing preservation was also achieved using 5 Gray for 5 days for smaller tumors or 3 Gray for 10 treatments for larger tumors .

Large tumors

Tumor size can affect control. Foster and colleagues showed that tumors larger than 3 cm had a control rate of 33%, whereas tumors of 2 to 3 cm had a control rate of 86%, and tumors of 2 cm or less could be controlled in 89% of their SRS series. Chan also showed a relationship between increasing tumor volume and the need for surgical intervention (shunt or resection) following FSRT.

Park and colleagues reviewed 50 cases of acoustic neuromas measuring over 3 cm on MRI. Microsurgery was performed on all patients. Among eight patients who underwent subtotal resection followed by radiosurgery, all had tumor control with a median follow-up of 113 months (9.4 years). Gross total resection alone resulted in failure in one patient, and subtotal resection without radiation resulted in a 32% recurrence rate. The facial nerve preservation rate was inversely proportional to the extent of tumor removal.