Fractionated Stereotactic Radiotherapy in the Treatment of Vestibular Schwannoma (Acoustic Neuroma): Predicting the Risk of Hydrocephalus

Int. J. Radiation Oncology Biol. Phys., Vol. 80, No. 4, pp. 1143–1150, 2011 Copyright Ó 2011 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/$–see front matter

doi:10.1016/j.ijrobp.2010.04.019

CLINICAL INVESTIGATION

Brain

FRACTIONATED STEREOTACTIC RADIOTHERAPY IN THE TREATMENT OF VESTIBULAR SCHWANNOMA (ACOUSTIC NEUROMA): PREDICTING THE RISK OF HYDROCEPHALUS CERI POWELL, F.R.C.R.,* CAROLINE MICALLEF, F.R.C.R.,y ADAM GONSALVES, B.SC. (HONS.),* BEV WHARRAM, B.A. (HONS.),* SUE ASHLEY, PH.D.,* AND MICHAEL BRADA, F.R.C.R., F.R.C.P.*z *Neuro-oncology Unit, Royal Marsden NHS Foundation Trust, London, United Kingdom; yNational Hospital for Neurology and Neurosurgery, London, United Kingdom; and zAcademic Unit of Radiotherapy and Oncology, Institute of Cancer Research, London, United Kingdom Purpose: To determine the incidence and predictive factors for the development of hydrocephalus in patients with acoustic neuromas (AN) treated with fractionated stereotactic radiotherapy. Patients and Methods: Seventy-two patients with AN were treated with fractionated stereotactic radiotherapy between 1998 and 2007 (45–50 Gy in 25–30 fractions over 5 to 6 weeks). The pretreatment MRI scan was assessed for tumor characteristics and anatomic distortion independently of subsequent outcome and correlated with the risk of hydrocephalus. Results: At a median follow-up of 49 months (range, 1–120 months), 5-year event-free survival was 95%. Eight patients (11%) developed hydrocephalus within 19 months of radiotherapy, which was successfully treated. On univariate analysis, pretreatment factors predictive of hydrocephalus were maximum diameter (p = 0.005), proximity to midline (p = 0.009), displacement of the fourth ventricle (p = 0.02), partial effacement of the fourth ventricle (p < 0.001), contact with the medulla (p = 0.005), and more brainstem structures (p = 0.004). On multivariate analysis, after adjusting for fourth ventricular effacement, no other variables remained independently associated with hydrocephalus formation. Conclusions: Fractionated stereotactic radiotherapy results in excellent tumor control of AN, albeit with a risk of developing hydrocephalus. Patients at high risk, identified as those with larger tumors with partial effacement of the fourth ventricle before treatment, should be monitored more closely during follow-up. It would also be preferable to offer treatment to patients with progressive AN while the risk of hydrocephalus is low, before the development of marked distortion of fourth ventricle before tumor diameter significantly exceeds 2 cm. Ó 2011 Elsevier Inc. Vestibular schwannoma, Hydrocephalus, Fractionated, Stereotactic radiotherapy.

INTRODUCTION

The management options include observation, surgery, and high-precision radiotherapy (and a combination of these). Surgery and irradiation strategies are associated with tumor control rates in the region of 90% at 10 years (3–18). Tumor size, rate of growth, symptoms, age, comorbidity, and ultimately patient preference influence the choice of management strategy. Surveillance alone is a reasonable alternative in patients with slow-growing tumors and no or minimal symptoms, and the timing of subsequent

Vestibular schwannomas (acoustic neuromas [AN]) are benign tumors of the vestibular nerve and account for 6–8% of intracranial tumors. They are slow growing, typically presenting with progressive unilateral sensorineural hearing loss, tinnitus, and vertigo, although larger tumors may cause mass effect presenting with ataxia, facial nerve or trigeminal neuropathy, and more rarely compression of lower cranial nerves and the brainstem (1, 2).

Reprint requests to: Michael Brada, F.R.C.R., F.R.C.P., The Royal Marsden NHS Foundation Trust, Downs Road, Sutton, Surrey SM2 5PT, United Kingdom. Tel: (+44) (0) 20-8661-3272; Fax: (+44) (0) 20-8661-3127; E-mail: [email protected] This work was undertaken at The Royal Marsden National Health Service (NHS) Foundation Trust, which received a proportion of its funding from the NHS Executive; the views expressed in this publication are those of the authors and not necessarily those of the NHS Executive. This work was supported by the Institute of Cancer Research, The Royal Marsden NHS Foundation Trust, and by Cancer

Research UK Section of Radiotherapy Grant C46/A2131. The NHS provides funding to the National Institute for Health Research Biomedical Research Centre. Conflict of interest: none. Acknowledgment—The authors thank Dr. Robert Huddart for help with the study design and Mrs. Nicola Bixby for her help in manuscript preparation. Received Dec 29, 2009, and in revised form March 30, 2010. Accepted for publication April 1, 2010. 1143

1144

I. J. Radiation Oncology d Biology d Physics

treatment has no apparent influence on long-term outcome, although the preference is for surgery and radiosurgery to be carried out for smaller lesions. Radiation is used as a noninvasive alternative to surgery and can be delivered as singlefraction radiosurgery or as fractionated stereotactic radiotherapy over 2–6 weeks. Although the principal complications of AN and its treatment are damage to the eighth, seventh, and fifth cranial nerves, hydrocephalus is a complication reported in 18– 42% of patients before treatment (19–21), in 0–8% after radiosurgery (6, 16), and in 0–12% after fractionated stereotactic radiotherapy (22, 23) (Table 1). Because unrecognized hydrocephalus, particularly presenting with nonspecific symptoms, may be life threatening, it is important that it is diagnosed and treated early. We assessed the frequency of hydrocephalus in a cohort of patients treated with a radical dose of fractionated stereotactic radiotherapy and attempted to define predictive factors to identify a population at higher risk. The aim was first to offer treatment at the time of low risk to avoid the need for subsequent surgical intervention and second to identify disease and patient characteristics that would predict a group of patients at high risk of developing hydrocephalus suitable for more intensive surveillance, resulting in early detection and treatment of hydrocephalus. PATIENTS AND METHODS Between May 1998 and December 2007, 81 patients with AN were treated with fractionated stereotactic radiotherapy at the Royal Marsden Hospital. Eight patients who received low-dose radiotherapy (#30 Gy) and 1 patient who did not have AN on histology review were excluded from the analysis. The remaining 72 patients are the subject of this review; their disease and patient characteristics are summarized in Table 2. The study was approved by an institutional research and ethics board as an audit of results. The Royal Marsden Hospital is a tertiary referral center, and patients are referred after full surgical assessment. The indication for treatment is tumor progression after a period of observation. Because the tumors considered for treatment are generally of large size, fractionated stereotactic radiotherapy is offered in preference to single-fraction radiosurgery. All patients received fractionated stereotactic radiotherapy either as primary treatment for progressive, histologically unverified tumors (59 patients) or for recurrent disease after previous surgery (13 patients). Patients were immobilized in a Gill-Thomas-Cosman relocatable stereotactic frame (24), and target volume was delineated on MRI with a slice thickness of 3 mm fused with planning CT. Clinical target volume was defined as enhancing tumor, and a circumferential margin of 3 mm was added for planning target volume (25). Four to six noncoplanar static conformal fields were used (26). Thirty-three patients received 45 Gy in 25 fractions (latterly defined as minimum dose), and 39 patients received 50 Gy in 30 fractions to 100% treating daily. The optimal radiotherapy treatment plan selected ensured coverage of the planning target volume by the 45-Gy isodose. Patients were reviewed weekly during treatment for assessment of acute toxicity. After completion of treatment, a baseline MRI scan was performed at 3 months, and subsequently patients were reviewed annually with a repeat MRI scan or as clinically indicated.

Volume 80, Number 4, 2011

Table 1. Incidence of hydrocephalus for different treatment modalities, including pretreatment data First author (reference)

Year of Patients Treatment Hydrocephalus publication (n) modality (%)

Briggs (40) Atlas (36) Samii (4) Pirouzmand (38) Wada (21) Tanaka (35) Pollock (17) Fukuda (19) Mendenhall (14) Kondziolka (12) Noren (41) Unger (42) Suh (43) Spiegelmann (16) Andrews (33) Prasad (44) Regis (45) Iwai (11) Meijer (7) Chung (15) Rogg (20) Okunaga (6) Pollock (17) Roos (46) Roche (32) Poen (29) Shirato (34) Andrews (33) Sawamura (23) Chung (15) Selch (22) Meijer (7)

1993 1996 1997 2001

1152 104 1000 284y

Surgery Surgery Surgery Surgery

4 14 2/1) 33 pts

2003 2003 2006 2007 1994

24 236 36 68 32

— Surgery Surgery Surgery SRS

42 14z 0 24 6

1998

162

SRS

3

1998 1999 2000 2001

669 56 29 44

SRS SRS SRS SRS

9/1x 2 7 0

2001 2001 2002 2003 2003 2004 2005 2005 2006 2006 2008 1999 2000 2001 2003 2004 2004 2006

69 153 97 51 49 45 157 53 46 65 1000 33 65 56 101 27 50 80

SRS SRS SRS SRS SRS SRS SRS SRS SRS SRS SRS FSRT FSRT FSRT FSRT FSRT FSRT FSRT

3 0 4 6x 2 4 18z 21z/8x 4 5 3/1x 0 5 4 12 4 0 0

Abbreviations: SRS = stereotactic radiosurgery; FSRT = fractionated stereotactic radiotherapy. * Symptomatic hydrocephalus requiring shunt insertion. y Proportion of cerebello-pontine angle tumors that were acoustic neuromas not specified. z Includes asymptomatic hydrocephalus diagnosed on MRI. x Symptomatic hydrocephalus after treatment requiring shunt insertion.

Imaging analysis The pretreatment MRI scan was available in 71 patients and was assessed by a neuroradiologist and the primary researcher, and both carried out the review without knowledge of the outcome. The MRI factors assessed included size and site of the tumor (judged by the horizontal and vertical extent and position of the tumor in relation to adjacent normal central nervous system structures), internal tumor characteristics, and distortion of normal central nervous system architecture. In case of disagreement between investigators a consensus view was reached. T1-weighted, gadolinium-enhanced MRI sequences were used to determine homogeneity of tumor enhancement, the presence of a cystic component (excluding peripheral entrapped cerebrospinal fluid [CSF]), and tumor size. Horizontal extent was defined in the

Hydrocephalus after SCRT for acoustic neuroma d C. POWELL et al.

Table 2. Disease and treatment characteristics Age (y), median (range)

Statistical analysis

58 (20–78)

Sex Male Female Radiotherapy dose (Gy) 45 50 Tumor side Left Right Previous surgery Yes No

1145

37 35 33 39 34 38 13 59

Values are number except where noted.

axial plane in relation to midline. Horizontal tumor diameter was measured at the level of the internal auditory canal (DmaxIAC), consistent for all patients, and maximum diameter (Dmax), which varied as tumor position varied, as defined in Table 3. A measurement of 0 mm represents intracanalicular tumors or tumors with no intracranial component within the cerebello-pontine angle. The distance of the medial tumor edge and displacement of the fourth ventricle from midline were determined. The shape of the fourth ventricle was assessed, and effacement (distortion) was recorded as none, minimal, partial, or complete corresponding to Grade 0–3, respectively, as reported by Rogg et al. (20) and illustrated in Fig. 1. Effacement was defined as Grades 2 and 3. The distortion of the brain parenchyma by the AN at the intracranial point of contact was also recorded, with Grade 1 and 2 indicating no distortion and Grades 3 and 4 indicating distortion, according to the Koos grading system (27). The cranio–caudal extent was assessed by vertical length using the number of brainstem points of contact as a surrogate. Tumor position was determined by the intracranial point of contact (cerebellum, middle cerebellar peduncle, medulla, pontomedullary junction [PMJ], pons, brainstem, or fifth cranial nerve). Pretreatment MRI scans were available for 71 of 72 patients (99%). In 1 patient the pretreatment scan did not have a size scale, and the tumor diameter was not measurable. For all patients who developed hydrocephalus a complete data set was available.

The primary endpoint of the study was the actuarial incidence of hydrocephalus determined using the Kaplan-Meier method and measured from the date of the start of fractionated stereotactic radiotherapy. The secondary endpoint of event-free survival was defined as persistent increase in size of the treated lesion or surgical intervention for tumor reduction, usually due to persistent symptoms without necessarily an increase in tumor size, and was measured from the date of the start of stereotactic radiotherapy. Univariate analysis was used to assess the relationship between patient and tumor characteristics and the incidence of hydrocephalus. The individual statistical test for each comparison is shown in Table 4. Binary logistic regression using a step-up procedure was used for multivariate analysis. Variables included in the analysis were those that were significant at the 10% level in the univariate analysis.

RESULTS Seventy-two patients with AN received fractionated stereotactic radiotherapy according to protocol to a dose of 45 Gy in 25 fractions over 5 weeks or 50 Gy in 30 fractions over 6 weeks (Table 2). The median age was 58 years (range, 20–78 years). Thirteen patients had surgical intervention before radiotherapy, ranging from a biopsy to previous surgical resection, and received fractionated stereotactic radiotherapy for recurrent or residual disease. The median follow-up was 49 months (range, 1–120 months). None of the patients had hydrocephalus or complete effacement of the fourth ventricle before treatment. Eight patients (11%) developed hydrocephalus after treatment within 19 months of completion of treatment (median, 8.5 months; range, 1–19 months) (Fig. 2). The actuarial risk of developing hydrocephalus was 7% (95% confidence interval [CI] 3–16%) at 1 year and 12% (95% CI 6–22%) at 2 years. The presenting symptoms included unsteadiness (4 patients), visual deterioration (2 patients), short-term memory loss (2 patients), headache (2 patients), dizziness, and urinary incontinence. All patients underwent insertion of a ventriculo-peritoneal shunt with resolution of hydrocephalus; 1 patient required revision of the shunt 1 week after insertion.

Table 3. MRI definitions Characteristic Homogenous/heterogeneous Cystic component Midline DmaxIAC Dmax Distance of AN from midline Displacement of fourth ventricle Fourth ventricular effacement Parenchymal signal change Distortion at point of contact Intracranial point of contact

Definition Homogenous enhancement after contrast or mixed signal Cystic component present/absent, not including peripheral CSF entrapment A line drawn on axial imaging between the mid-clivus (at the level of the internal carotid arteries) to the internal occipital protuberance Maximum horizontal distance from the IAC to medial edge of AN Greatest horizontal diameter of AN excluding component within IAC Minimum distance from medial edge of AN to midline Distance of center of fourth ventricle from midline Recorded as none, minimal, partial, or complete (see Fig. 1) Signal change at intracranial point of contact on T2-weighted sequence Recorded as present or absent Contact with cerebellum, middle cerebellar peduncle, pons, pontomedullary junction, medulla, midbrain, or fifth cranial nerve

Abbreviations: CSF = cerebrospinal fluid; IAC = internal auditory canal; AN = acoustic neuroma.

1146

I. J. Radiation Oncology d Biology d Physics

Volume 80, Number 4, 2011

Fig. 1. Example of no effacement of the fourth ventricle (A) and partial effacement (B).

Tumor control None of the patients developed permanent/progressive enlargement of the treated AN. One patient required surgical intervention for drainage of a cystic component causing persistent symptoms 8 months after completion of radiother-

apy, and a further patient underwent revision of a ventriculo-peritoneal shunt followed by surgical resection without evidence of progressive tumor enlargement at the time of surgery. The 5-year event-free survival rate was 95% (Fig. 3).

Table 4. Univariate and multivariate analysis of imaging factors as predictors of hydrocephalus p

Proportion with hydrocephalus (%) Factor

Factor

No factor

Univariate

Cystic component (yes/no) Homogeneous vs. heterogeneous Dmax (mm) <20 vs. $ 20 mm DmaxIAC (mm) Distance from midline (mm) <10 mm vs. $10 mm Displacement of fourth ventricle (mm) Present vs. absent Fourth ventricular effacement N/M vs. P/C Distortion at point of contact (no/yes) Intracranial point of contact Pons Cerebellum MCP Medulla CV Midbrain Transverse vs. vertical vs. both (%) No. of brainstem points of contact 0/1/2/3 (%)

2/12 (17) 1/23 (4) 27.4 (20–36) 0/33 19.8 (12.3–26.3) 5.8 (0–8.8) 8/38 (21) 3.9 (0–10.4) 7/35 (20) 0/41 1/34 (3)

6/58 (10) 7/46 (15) 19.7 (3.9–36.3) 8/37 (22) 14.5 (0–30.4) 10.1 (0–23.8) 0/31 0 (0–6.7) 1/34 (3) 8/29 (28) 7/37 (19)

0.6) 0.4) 0.005y 0.006) 0.02y 0.009y 0.007) 0.02y 0.06) <0.001) 0.04)

0/20 0/13

0.1) 0.3) 0.3) 0.005) 1.0) 0.2) 0.2z 0.004x

8/50 (16) 8/58 (14) 8/59 (14) 6/20 (30) 1/10 (10) 1/2 (50)

2/49 (4) 7/61 (11) 7/67 (9) 0 vs. 0 vs. 16 0 vs. 4 vs. 31 vs. 22

Multivariate

<0.001

0.08

0.1

Abbreviations: IAC = internal auditory canal; N/M = none or minimal; P/C = partial or complete; MCP = middle cerebellar peduncle; CV = fifth cranial nerve; transverse = cerebellum or middle cerebellar peduncle; vertical = brainstem structures. Values are number (percentage) or median (range). * Fisher’s exact test. y Mann-Whitney U test. z 2 c test. x Kendall’s Tau-b.

Hydrocephalus after SCRT for acoustic neuroma d C. POWELL et al.

1147

Fig. 4. Distribution of maximum tumor diameter (Dmax) and the development of hydrocephalus. Fig. 2. Actuarial incidence of hydrocephalus.

Cranio–caudal tumor extent, assessed using the number of brainstem points of contact with the tumor, was predictive of the development of hydrocephalus (p = 0.004) (Fig. 6). In terms of tumor position, contact with the medulla (including the PMJ) was predictive of hydrocephalus formation (p = 0.005) (Table 4). Variables significant at the 10% level in the univariate analysis (Table 4) were included in the multivariate analysis, and effacement of the fourth ventricle before treatment was the only independent predictor of developing hydrocephalus (p < 0.001). Although contact with the medulla (including PMJ) failed to reach significance (p = 0.08), 43% (95% CI 18–71%) of patients with both effacement of the fourth ventricle and tumor in contact with the medulla (including PMJ) developed hydrocephalus after radiotherapy.

Imaging predictors of hydrocephalus The predictors of hydrocephalus on univariate analysis are summarized in Table 4. The tumor diameter (Dmax) in patients who developed hydrocephalus was 27 mm (range, 20–36 mm) compared with 20 mm (range, 4–36 mm) in those who did not (p = 0.005); DmaxIAC was larger in patients developing hydrocephalus (median, 20 mm; range, 12–26 mm) than in those with no hydrocephalus (median, 15 mm; range, 0–30 mm) (p = 0.02). All 8 patients who developed hydrocephalus had tumors larger than 20 mm (Fig. 4), and this size cutoff was a significant predictor of developing hydrocephalus on univariate analysis (p = 0.006). Patients who developed hydrocephalus after radiotherapy had tumors closer to or crossing the midline (median distance to midline, 6 mm; range, 0–9 mm) than those who did not (10 mm; range, 0–24 mm) (p = 0.009) (Fig. 5). Using 1 cm as a cutoff was a significant predictor of hydrocephalus (p = 0.007); all patients with hydrocephalus had tumor within 1 cm or over midline before treatment, compared with 30 of 61 in patients without hydrocephalus. Of patients who had partial effacement of the fourth ventricle before treatment, 8 developed hydrocephalus, whereas none of those without effacement developed the complication (p < 0.001). Greater displacement of the fourth ventricle before treatment was also seen in those with hydrocephalus (median, 4 mm; range, 0–10 mm) than in those without (0 mm; range, 0–7 mm) (p = 0.02). Distortion of the brain parenchyma at the intracranial point of contact was also a predictor of developing hydrocephalus (p = 0.04).

In patients with AN, fractionated stereotactic radiotherapy offers an alternative to surgical resection and radiosurgery, and this is particularly suitable for larger tumors. Our results confirm excellent tumor control, with no patient developing progressive enlargement at a median follow-up of 49 months (range, 1–120 months) albeit 2 patients had a further surgical intervention, with a 5-year event-free survival rate of 95%. This compares favorably with other reports (7, 15, 23, 28). The treatment is, however, not without toxicity, with 11% of patients developing hydrocephalus within 19 months (median, 8.5 months) after completion of treatment. Nevertheless all patients who developed

Fig. 3. Actuarial event-free survival.

Fig. 5. Distribution of tumor distance from midline and the development of hydrocephalus; 0 represents midline.

DISCUSSION

1148

I. J. Radiation Oncology d Biology d Physics

Fig. 6. Brainstem contact and development of hydrocephalus.

hydrocephalus were successfully managed, with insertion of a ventriculo-peritoneal shunt leading to a resolution of symptoms. No patient died as a result of hydrocephalus or disease progression. We identified tumor size, proximity to midline, effacement and displacement of the fourth ventricle, distortion at the point of tumor contact, tumor contact with the medulla (including PMJ), and cranio–caudal extent as predictors of subsequent hydrocephalus. On multivariate analysis partial effacement of the fourth ventricle before treatment was the principal independent predictor of developing hydrocephalus, with a risk of 28% (95% CI 13–47%). The highest risk (43% [95% CI 18– 71%]) was seen in patients with both effacement of the fourth ventricle and caudal extension of the tumor as defined by contact with the medulla, including PMJ. Median follow-up of our cohort of 4 years provides reasonably complete information, particularly for an event that occurs principally in the first 2 years, and compares favorably

Volume 80, Number 4, 2011

with the length of follow-up in other reported series (7, 15, 22, 23, 28–31). The development of hydrocephalus after single-fraction radiosurgery occurs up to 31 months after treatment (32) and after fractionated radiotherapy up to 20 months after treatment (23), suggesting a need for careful monitoring of patients at risk for a minimum of 2 years. The evaluation of imaging was carried out by two observers without the knowledge of the subsequent clinical course (blinded), to minimize the bias in assessing the imaging parameters. The unblinding was only carried out after full collection of the data. Because a proportion of patients had only axial sequences available for inspection, it was not possible to reliably assess the tumor extent in all orthogonal planes. The cranio–caudal extent of the tumor was assessed as the number of brainstem points of contact ranging from the midbrain to the medulla. Although this might be less accurate than direct measurement it provides a reasonable assessment of both the cranio–caudal size and the relationship to the normal structures, and it therefore combines both length and position information. Because the axial sequences are most frequently used in clinical evaluation the predictive value of the findings are likely to be clinically useful. Predictors of hydrocephalus The incidence of hydrocephalus in our cohort is within the range reported by others (none to 12%) after fractionated radiotherapy (Table 1) (7, 15, 22, 23, 29, 33, 34). Although hydrocephalus is an important complication of ANs and their treatment, its incidence is not reported in all series and is difficult to determine owing to heterogeneity in patient populations and radiological definition.

Table 5. Studies reporting factors predictive of hydrocephalus First author (reference) Atlas (36) Briggs (37) Fukuda (19) Pirouzmand (38) Roche (32)

Rogg (20) Sawamura (23) Steenerson (39) Tanaka (35) Wada (21) Present study

No. of patients (% with hydrocephalus) 104 (14) 1152 (4) 68 (24) 284 (14) 1000 (4)

157 (18) 101 (12) 3 cases reviewed 236 (14) 24 (42) 72 (11)

Timing of diagnosis of hydrocephalus At diagnosis (presurgery) At diagnosis (presurgery) At diagnosis (presurgery) At diagnosis (presurgery), 1.2% after surgery 3.2% pre-, 1.1% post-Gamma Knife

At diagnosis (pre-Gamma Knife) After FSRT At diagnosis (presurgery) At diagnosis (presurgery) At diagnosis After FSRT

Abbreviation as in Table 1. * In the subset of patients categorized as having noncommunicating hydrocephalus.

Factor associated with hydrocephalus Mean tumor size Tumor diameter Tumor diameter Cerebrospinal fluid protein concentration Tumor diameter Age Tumor volume Previous surgery Bilateral tumors, NF2 Tumor volume) Fourth ventricular distortion) Mean tumor size Tumor size Age Tumor diameter Tumor diameter Fourth ventricular distortion Tumor diameter Proximity to midline Brainstem contact

Hydrocephalus after SCRT for acoustic neuroma d C. POWELL et al.

Previous retrospective studies identified tumor size as a predictor of hydrocephalus after radiosurgery (Table 5) (19–21, 23, 32, 35–39) and fractionated radiotherapy (23), regardless of the method of assessment (tumor volume as a product of three perpendicular diameters  p/6 [20], or using computer modeling [7], maximum tumor diameter [38], or mean tumor size [23]). Increasing age has also been considered a risk factor for developing hydrocephalus (32, 35), although it can be difficult to distinguish true hydrocephalus from age-related ventricular enlargement, especially if asymptomatic. Mechanism of hydrocephalus formation There are two predominant theories relating to the etiology of hydrocephalus in association with ANs. Because size is a predictor, mechanical obstruction of the CSF flow pathway is likely to be an important cause (20). The aqueduct of Sylvius is a potential candidate for the site of CSF outflow obstruction. Tumors in contact with the medulla (including the PMJ) before irradiation were at higher risk of developing hydrocephalus, which raises the possibility that the inferior extent of the tumor may also play a role in the obstruction by impeding CSF flow from the fourth ventricle into the cisterna magna (cerebellomedullary cistern) via the median (foramen of Magendie) or lateral (foramina of Luschka) apertures or alternatively the central canal. Further identification of the potential site of occlusion requires additional studies. It has been suggested that a communicating hydrocephalus can occur independent of tumor size or fourth ventricle distortion (20, 23, 35, 38), although this has not been confirmed in our study. The proposed mechanism is CSF malabsorption due to accumulation of proteinaceous material in the CSF. Cerebrospinal fluid protein concentration measured intraoperatively in the cerebellomedullary cistern correlated with the development of hydrocephalus, as did tumor size (19). Communicating hydrocephalus observed in

1149

patients after fractionated radiotherapy was also considered to be due to release of large molecules including necrotic debris into the CSF, aggravating pre-existing CSF malabsorption (23), although this is not supported by CSF protein measurements, which were normal (data not published). Although scarring and thickening of the arachnoid membranes is commonly seen at surgical resection after failed radiosurgery, these are confined to the area around the tumor and do not affect distant sites (32). Implication for clinical practice and conclusion Our results confirm that fractionated stereotactic radiotherapy achieves excellent tumor control and provides a reasonable alternative to surgery or single-fraction radiosurgery, especially for larger tumors. However, the treatment carries an 11% risk of developing hydrocephalus. Patients with a large AN causing significant distortion of the fourth ventricle have a higher risk of developing hydrocephalus, which requires CSF diversion. Because this is a potentially serious complication the balance of risks should be weighed for each individual patient. In patients with tumor diameter $2 cm a policy of observation should be used with caution and treatment instituted early before a maximum diameter of 2 cm is significantly exceeded. Patients identified on pretreatment MRI with higher risk of developing hydrocephalus, especially those with partial effacement of the fourth ventricle, require closer surveillance after radiotherapy to ensure prompt diagnosis and treatment. The proposed regimen is 3-monthly clinical review and 3–6-monthly MRI scans for the first 2 years after treatment. This should be combined with information about symptoms indicative of hydrocephalus and fast access to informed help. Conversely, the large majority of patients, who do not have the identified risk factors, can continue lessintensive posttreatment surveillance with a low risk of developing hydrocephalus.

REFERENCES 1. Yoshimoto Y. Systematic review of the natural history of vestibular schwannoma. J Neurosurg 2005;103:59–63. 2. Smouha EE, Yoo M, Mohr K, et al. Conservative management of acoustic neuroma: A meta-analysis and proposed treatment algorithm. Laryngoscope 2005;115:450–454. 3. Kaylie DM, Horgan MJ, Delashaw JB, et al. A meta-analysis comparing outcomes of microsurgery and gamma knife radiosurgery. Laryngoscope 2000;110:1850–1856. 4. Samii M, Matthies C. Management of 1000 vestibular schwannomas (acoustic neuromas): Surgical management and results with an emphasis on complications and how to avoid them. Neurosurgery 1997;40:11–21. discussion 21–13. 5. Acousticneuroma. Consens Statement 1991;9:1–24. 6. Okunaga T, Matsuo T, Hayashi N, et al. Linear accelerator radiosurgery for vestibular schwannoma: Measuring tumor volume changes on serial three-dimensional spoiled gradient-echo magnetic resonance images. J Neurosurg 2005;103:53–58. 7. Meijer OW, Vandertop WP, Baayen JC, et al. Single-fraction vs. fractionated linac-based stereotactic radiosurgery for vestib-

8. 9. 10. 11. 12. 13.

ular schwannoma: A single-institution study. Int J Radiat Oncol Biol Phys 2003;56:1390–1396. Flickinger JC, Kondziolka D, Niranjan A, et al. Results of acoustic neuroma radiosurgery: An analysis of 5 years’ experience using current methods. J Neurosurg 2001;94:1–6. Ito K, Shin M, Matsuzaki M, et al. Risk factors for neurological complications after acoustic neurinoma radiosurgery: Refinement from further experiences. Int J Radiat Oncol Biol Phys 2000;48:75–80. Mendenhall WM, Friedman WA, Buatti JM, et al. Preliminary results of linear accelerator radiosurgery for acoustic schwannomas. J Neurosurg 1996;85:1013–1019. Iwai Y, Yamanaka K, Shiotani M, et al. Radiosurgery for acoustic neuromas: Results of low-dose treatment. Neurosurgery 2003;53:282–287. discussion 287–288. Kondziolka D, Lunsford LD, McLaughlin MR, et al. Long-term outcomes after radiosurgery for acoustic neuromas. N Engl J Med 1998;339:1426–1433. Foote KD, Friedman WA, Buatti JM, et al. Analysis of risk factors associated with radiosurgery for vestibular schwannoma. J Neurosurg 2001;95:440–449.

1150

I. J. Radiation Oncology d Biology d Physics

14. Mendenhall WM, Friedman WA, Bova FJ. Linear acceleratorbased stereotactic radiosurgery for acoustic schwannomas. Int J Radiat Oncol Biol Phys 1994;28:803–810. 15. Chung HT, Ma R, Toyota B, et al. Audiologic and treatment outcomes after linear accelerator-based stereotactic irradiation for acoustic neuroma. Int J Radiat Oncol Biol Phys 2004;59: 1116–1121. 16. Spiegelmann R, Lidar Z, Gofman J, et al. Linear accelerator radiosurgery for vestibular schwannoma. J Neurosurg 2001;94:7–13. 17. Pollock BE, Driscoll CL, Foote RL, et al. Patient outcomes after vestibular schwannoma management: A prospective comparison of microsurgical resection and stereotactic radiosurgery. Neurosurgery 2006;59:77–85. discussion 77–85. 18. Pollock BE. Vestibular schwannoma management: An evidence-based comparison of stereotactic radiosurgery and microsurgical resection. Prog Neurol Surg 2008;21:222–227. 19. Fukuda M, Oishi M, Kawaguchi T, et al. Etiopathological factors related to hydrocephalus associated with vestibular schwannoma. Neurosurgery 2007;61:1186–1192. discussion 1192–1183. 20. Rogg JM, Ahn SH, Tung GA, et al. Prevalence of hydrocephalus in 157 patients with vestibular schwannoma. Neuroradiology 2005;47:344–351. 21. Wada K, Nawashiro H, Shimizu A, et al. MRI analysis of hydrocephalus associated with acoustic neurinoma. Acta Neurochir Suppl 2003;86:549–551. 22. Selch MT, Pedroso A, Lee SP, et al. Stereotactic radiotherapy for the treatment of acoustic neuromas. J Neurosurg 2004; 101(Suppl. 3):362–372. 23. Sawamura Y, Shirato H, Sakamoto T, et al. Management of vestibular schwannoma by fractionated stereotactic radiotherapy and associated cerebrospinal fluid malabsorption. J Neurosurg 2003;99:685–692. 24. Graham JD, Warrington AP, Gill SS, et al. A non-invasive, relocatable stereotactic frame for fractionated radiotherapy and multiple imaging. Radiother Oncol 1991;21:60–62. 25. Kumar S, Burke K, Nalder C, et al. Treatment accuracy of fractionated stereotactic radiotherapy. Radiother Oncol 2005;74:53–59. 26. Perks JR, Jalali R, Cosgrove VP, et al. Optimization of stereotactically-guided conformal treatment planning of sellar and parasellar tumors, based on normal brain dose volume histograms. Int J Radiat Oncol Biol Phys 1999;45:507–513. 27. Koos W, Spetzler R, Lang J. Color atlas of microneurosurgery. 2nd ed. Stuttgart: Thieme; 1993. 28. Williams JA. Fractionated stereotactic radiotherapy for acoustic neuromas. Int J Radiat Oncol Biol Phys 2002;54:500–504. 29. Poen JC, Golby AJ, Forster KM, et al. Fractionated stereotactic radiosurgery and preservation of hearing in patients with vestibular schwannoma: A preliminary report. Neurosurgery 1999; 45:1299–1305. discussion 1305–1297. 30. Kalapurakal JA, Silverman CL, Akhtar N, et al. Improved trigeminal and facial nerve tolerance following fractionated stereotactic radiotherapy for large acoustic neuromas. Br J Radiol 1999;72:1202–1207.

Volume 80, Number 4, 2011

31. Combs SE, Volk S, Schulz-Ertner D, et al. Management of acoustic neuromas with fractionated stereotactic radiotherapy (FSRT): Long-term results in 106 patients treated in a single institution. Int J Radiat Oncol Biol Phys 2005;63:75–81. 32. Roche PH, Khalil M, Soumare O, et al. Hydrocephalus and vestibular schwannomas: Considerations about the impact of gamma knife radiosurgery. Prog Neurol Surg 2008;21: 200–206. 33. Andrews DW, Suarez O, Goldman HW, et al. Stereotactic radiosurgery and fractionated stereotactic radiotherapy for the treatment of acoustic schwannomas: Comparative observations of 125 patients treated at one institution. Int J Radiat Oncol Biol Phys 2001;50:1265–1278. 34. Shirato H, Sakamoto T, Takeichi N, et al. Fractionated stereotactic radiotherapy for vestibular schwannoma (VS): Comparison between cystic-type and solid-type VS. Int J Radiat Oncol Biol Phys 2000;48:1395–1401. 35. Tanaka Y, Kobayashi S, Hongo K, et al. Clinical and neuroimaging characteristics of hydrocephalus associated with vestibular schwannoma. J Neurosurg 2003;98:1188–1193. 36. Atlas MD, Perez de Tagle JR, Cook JA, et al. Evolution of the management of hydrocephalus associated with acoustic neuroma. Laryngoscope 1996;106:204–206. 37. Briggs RJ, Fabinyi G, Kaye AH. Current management of acoustic neuromas: Review of surgical approaches and outcomes. J Clin Neurosci 2000;7:521–526. 38. Pirouzmand F, Tator CH, Rutka J. Management of hydrocephalus associated with vestibular schwannoma and other cerebellopontine angle tumors. Neurosurgery 2001;48:1246–1253. discussion 1253–1244. 39. Steenerson RL, Payne N. Hydrocephalus in the patient with acoustic neuroma. Otolaryngol Head Neck Surg 1992;107: 35–39. 40. Briggs RJ, Shelton C, Kwartler JA, et al. Management of hydrocephalus resulting from acoustic neuromas. Otolaryngol Head Neck Surg 1993;109:1020–1024. 41. Noren G. Long-term complications following gamma knife radiosurgery of vestibular schwannomas. Stereotact Funct Neurosurg 1998;70(Suppl. 1):65–73. 42. Unger F, Walch C, Haselsberger K, et al. Radiosurgery of vestibular schwannomas: A minimally invasive alternative to microsurgery. Acta Neurochir (Wien) 1999;141:1281–1285. discussion 1285–1286. 43. Suh JH, Barnett GH, Sohn JW, et al. Results of linear accelerator-based stereotactic radiosurgery for recurrent and newly diagnosed acoustic neuromas. Int J Cancer 2000;90: 145–151. 44. Prasad D, Steiner M, Steiner L. Gamma surgery for vestibular schwannoma. J Neurosurg 2000;92:745–759. 45. Regis J, Pellet W, Delsanti C, et al. Functional outcome after gamma knife surgery or microsurgery for vestibular schwannomas. J Neurosurg 2002;97:1091–1100. 46. Roos DE, Brophy BP, Bhat MK, et al. Update of radiosurgery at the Royal Adelaide Hospital. Australas Radiol 2006;50:158–167.