Long-term Treatment Response and Patient Outcomes for Vestibular Schwannoma Patients Treated with Hypofractionated Stereotactic Radiotherapy

Purpose The aim of this study is to evaluate long-term treatment outcome and toxicities among vestibular schwannoma (VS) patients treated with hypofractionated stereotactic radiotherapy (HSRT). Methods 383 patients with unilateral VS treated with HSRT (25 Gy, five fractions) between 1995 and 2007 were retrospectively reviewed. Treatment failure was defined as requiring salvage microsurgery. Posttreatment new/progressive clinical symptoms or increases in baseline tumor volume (BTV) due to treatment effect or progression were noted. Symptom outcomes were reported as baseline and posttreatment ± improvement, respectively. Symptoms were grouped by cranial nerve (CN) VII or CNVIII. Audiometry was assessed baseline and posttreatment hearing. Patients were grouped as having greater than serviceable hearing [Gardner Robertson (GR) score 1–2] or less than non-serviceable hearing (GR score 3–5) by audiometry. Results Median follow-up was 72.0 months. Nine (2.3%) experienced treatment failure. At last follow-up, 74 (19.3%) had new/progressive symptoms and were categorized as radiologic non-responders, whereas 300 (78.3%) had no tumor progression and were grouped as radiologic responders. Average pretreatment BTV for treatment failures, radiologic non-responders, and radiologic responders was 2.11, 0.44, and 1.87 cm3, respectively. Pretreatment CNVII and CNVIII symptoms were present in 9.4 and 93.4% of patients, respectively. Eight (24%) with pre-HSRT CNVII and 37 (10%) with pre-HSRT CNVIII symptoms recovered CN function post-HSRT. Thirty-five (9%) and 36 (9.4%) experienced new CNVII and CNVIII deficit, respectively, after HSRT. Of these, 20 (57%) and 18 (50%) recovered CNVII and CNVIII function, respectively, after HSRT. Evaluable audiograms were available in 199 patients. At baseline and at last follow-up, 65.8 and 36.2% had serviceable hearing, respectively. Fifty-one percent had preservation of serviceable hearing at last follow-up. Conclusion Treatment of VS with HSRT is effective with treatment success in 97.7% and an acceptable toxicity profile. Less than one-third of patients experience any new CNVII or CNVIII deficit posttreatment. Greater than 50% of patients with serviceable hearing at baseline maintained hearing function. Improved methods to differentiate treatment effect and tumor progression are needed.

Purpose: The aim of this study is to evaluate long-term treatment outcome and toxicities among vestibular schwannoma (VS) patients treated with hypofractionated stereotactic radiotherapy (HSRT).
Methods: 383 patients with unilateral VS treated with HSRT (25 Gy, five fractions) between 1995 and 2007 were retrospectively reviewed. Treatment failure was defined as requiring salvage microsurgery. Posttreatment new/progressive clinical symptoms or increases in baseline tumor volume (BTV) due to treatment effect or progression were noted. Symptom outcomes were reported as baseline and posttreatment ± improvement, respectively. Symptoms were grouped by cranial nerve (CN) VII or CNVIII. Audiometry was assessed baseline and posttreatment hearing. Patients were grouped as having greater than serviceable hearing [Gardner Robertson (GR) score [1][2] or less than non-serviceable hearing (GR score 3-5) by audiometry.
conclusion: Treatment of VS with HSRT is effective with treatment success in 97.7% and an acceptable toxicity profile. Less than one-third of patients experience any new VS HSRT Outcomes Frontiers in Oncology | www.frontiersin.org September 2017 | Volume 7 | Article 200 CNVII or CNVIII deficit posttreatment. Greater than 50% of patients with serviceable hearing at baseline maintained hearing function. Improved methods to differentiate treatment effect and tumor progression are needed.
Keywords: vestibular schwannoma, acoustic neuroma, hypofractionated radiotherapy, hearing loss, cranial nerves inTrODUcTiOn Vestibular schwannomas (VS) are benign skull base tumors that are managed by observation, surgical resection, or radiotherapy. The optimal approach to radiotherapy is unknown and includes single-dose stereotactic radiosurgery, hypofractionated stereotactic radiotherapy (HSRT) delivered over 2-5 treatments, or conventionally fractionated stereotactic radiotherapy generally administered over 25-30 sessions . In our practice, we preferentially utilized HSRT, 25 Gy in five fractions, based on a hypothesis that this would maintain the excellent tumor control achieved with high daily dose treatment but with reduced toxicities compared to single-dose stereotactic radiosurgery. The long-term impact of HSRT on local control, hearing, and treatment-related toxicity remains uncertain, although data from a few small cohorts suggest that the majority of patients maintain preserved serviceable hearing and low rates of facial nerve dysfunction after HSRT (1,(3)(4)(5)(6)(7)(8)(9)(10). In addition, the association of imaging outcome with clinical endpoints has not been rigorously explored. We present long-term tumor volume progression, cranial nerve (CN) function, and audiometric data for a cohort of patients treated with HSRT.

Patient selection and stratification
After receiving institutional review board approval, the electronic medical record at a single, high volume institution was retrospectively reviewed for patients with unilateral VS receiving radiation therapy alone between 1995 and 2007. Only those who received HSRT (25 Gy in five fractions) were included. Data were collected on patient demographic information and comorbidities, tumor characteristics including laterality and baseline tumor volume (BTV), median follow-up, pretreatment CN VII or VIII deficits, posttreatment CN VII or VIII deficits, other posttreatment complications, Gardner Robertson (GR) Score by audiometric assessment at baseline and after treatment completion, and changes in tumor volume after treatment and at last follow-up (23). Patients were stratified by degree of tumor progression: patients were identified as radiologic responders if their imaging abnormality volume at last follow-up was <20% larger than their BTV, as radiologic non-responders if their imaging volume at last follow-up was ≥20% larger than their BTV, and as treatment failures if they required any kind of additional salvage therapy including microsurgery.

Treatment regimen
Patients received highly conformal HSRT via the BrainLAB Treatment Planning System (BrainLAB, Feldkirchen, Germany) and Brain Scan software, v. 5.3. Stereotactic radiation was delivered using the BrainLAB mask and frame immobilization system. The dose utilized was 500 cGy × 5 prescribed to the 80% isodose line. Contouring of treatment volume was performed using CT fused and co-registered with gadolinium contrast-enhanced T1-weighted MRI. During treatment, all patients received daily localization.

radiologic evaluation
Each patient received a baseline MRI scan prior to treatment and serial MRI scans at clinically appropriate time intervals thereafter to document changes in tumor volume, measured in cubic centimeter. Tumor volume measurements were performed using the cranial-caudal, transverse, and anterior-posterior dimensions. Imaging progression or regression was documented as a percent change in volume relative to BTV.

landmark analysis of Outcome
To evaluate treatment failures (need for further therapeutic intervention) as a function of time based on their tumor volume, we performed a landmark analysis. Postradiosurgery imaging outcome was assessed. Tumor volume/imaging abnormality recorded at intervals of 0-3.99, 4.0-5.99, and >6.0 years after baseline imaging. Each patient's imaging volume was considered separately within each time interval, and they were assigned response status relative to BTV at each time point. As such, response status of one patient could change over time depending on the growth characteristics of their tumor. Patients with <20% increase in tumor volume relative to the BTV at each time interval were considered radiologic responders, those with a ≥20% increase relative to their BTV were radiologic non-responders, and those who required salvage microsurgery were treatment failures. Given the paucity of evidence regarding measures of volumetric response, we reviewed the literature, and on the basis of our clinical judgment, we decided upon a volumetric threshold of 20% above BTVs, which we felt was a conservative estimate to distinguish between progression, the error inherent in estimating volumetric changes using simple tri-dimensional measurements, and pseudoprogression (24,25). If a patient did not have imaging within a particular time interval, they were excluded from the analysis within that time interval.

clinical evaluation of symptomatology and hearing
The presence of CN VII or VIII deficit or other symptoms unrelated to CN VII or VIII were documented by clinical report at baseline prior to treatment and at each clinic visit posttreatment. CN VIII outcomes post-HSRT were reported separately as non-hearing and hearing outcomes with non-hearing outcomes  including imbalance, tinnitus, and dizziness. Hearing outcomes were reported as a function of the Gardener-Robertson scale. Audiometric evaluation occurred prior to treatment initiation and post-HSRT. Only those patients who received audiometric evaluation prior to and after treatment were considered evaluable. Patients were grouped as having greater than serviceable hearing (GR score 1-2) or less than non-serviceable hearing (GR score 3-5) by audiometry.

radiation-induced Morbidity
Treatment-related complications during follow-up including acute RT-induced neoplasm, hydrocephalus, pituitary treatment failure, and RT-related neurologic deficits were assessed and documented.

statistical analysis
Categorical data were analyzed using summary statistics and presented as frequency (%). Continuous variables were summarized as mean (SD). Two-tailed t-test was performed to compare radiologic responders vs. radiologic non-responders with regard to BTV and percentage change in tumor volume at last follow-up ( Table 2). Chi-squared test was performed to compare radiologic responders vs. radiologic non-responders with regard to posttreatment toxicities (

Tumor Volume Progression
Based on our criteria, there were 300 radiologic responders, 74 radiologic non-responders, and 9 treatment failures within his cohort. BTV for radiologic responders, radiologic nonresponders, and treatment failures was 1.87, 0.44, and 2.11 cm 3 , respectively (Table 2; Figure 1). Radiologic responders had significantly higher BTV relative to radiologic non-responders (p < 0.001; Table 2). While radiologic responders had an average decline in tumor volume of 45.5% of their BTV at last follow-up, radiologic non-responders and treatment failures had a 165.9 and 169.0% increase, respectively (  Landmark analysis of tumor response status revealed that there was a 6% increase in the number of radiologic responders at >6 years of follow-up relative to BTV, that there were half as many radiologic non-responders at >6 years of follow-up relative  to BTV, and that there were no treatment failures after 6 years ( Table 3). Notably, after 6 years of follow-up, patients continued to undergo surveillance imaging every 1-2 years. Moreover, of the radiologic non-responders with imaging between 0 and 3.99 years after baseline imaging, 30 ultimately became radiologic responders at last follow-up; of the radiologic responders with imaging between 0 and 3.99 years after baseline imaging, 7 ultimately became radiologic non-responders at last follow-up, and there were 5 treatment failures who initially responded to treatment within 1.5 years after baseline imaging. There were 50 radiologic responders and 19 radiologic non-responders without follow-up within 3.99 years who were excluded in the first interval. There were 15 radiologic responders and 7 radiologic non-responders without follow-up within 5.99 years who were excluded from the first 2 time intervals.

Treatment-related cn symptoms and complications
The most common overall posttreatment complication was hydrocephalus (five patients, 1.3%), for which one individual required ventricular shunt placement.

DiscUssiOn
Our findings demonstrate that HSRT is an appropriate treatment modality for unilateral VS, with an uncommon need for further  intervention, low posttreatment CNVII and CNVIII toxicity, and hearing preservation in greater than half of treated patients. Recent studies regarding the efficacy of HSRT for the treatment of VS demonstrate local control rates of 92-100% (5-7, 9, 10), with control defined as achieving the objective of no further treatment requirements. Similarly, in our cohort, we had a local control rate of 98%-as defined by the lack of need for surgical intervention-with radiologic responders having an average decrease in tumor volume of 45.5% at last follow-up. Relative to stereotactic radiosurgery (SRS) for VS, HSRT results in comparable intervention-free local tumor control (26)(27)(28)(29). In a large series of 839 patients treated with gamma knife radiosurgery that established the role of single-dose stereotactic therapy as an important treatment for this illness, control rate was 97% defined as no need for further intervention (30). Furthermore, similar to studies of SRS treatment of VS, in this cohort, there were no new post-HSRT treatment failures after 4-6 years (15).
The long-term durability of control after radiosurgery remains an important question, especially as the alternative of surgery provides durable control albeit with a different risk profile. We did not identify any treatment failures later than the first 6 years of follow-up, suggesting that HSRT may result in durable control. This is emphasized by the fact that patients continued to have imaging follow-up every 1-2 years even 6 years after baseline imaging. This is similar to the long-term results reported by other investigators that suggest that late treatment failure is uncommon after single-dose radiosurgery (15,29). Hasegawa et al. reported results of 440 patients with median follow-up of 12.5 years and observed 12 treatment failures within 3 years of therapy, 8 treatment failures >3 years after therapy, and no treatment failures after 10 years (29). In another large series from the University of Pittsburgh, no treatment failures were observed after 4 years (15). Interestingly, we observed that transient volumetric enlargement of tumors in the initial years after treatment did not necessarily predict treatment failure, as a number of these individuals ultimately did not demonstrate further imaging progression but actually became radiologic responders.
Indeed, imaging outcome may not be a clear measure of treatment success given radiation-related imaging changes and the difficulties of accurate determination of volume by bidimensional measurement of relatively small tumors. For this reason, we propose that long-term avoidance of need for further tumor therapy, although subject to some bias, and not imaging stability is the ultimate goal and should be the metric for treatment success. The imaging outcome reaffirms that enlargement of abnormality after radiosurgery for VS may only represent a progressive process that will require intervention in a small proportion of patients. Reports of outcome emphasize either lack of need for further intervention or imaging control as the outcome measure, without standard definitions (31). Of note, 44% of responders in our cohort had transient tumor volumetric enlargement after HSRT, which may have resulted in temporary CNVII and CNVIII symptoms that ultimately resolved with tumor regression. In addition, although 21.7% of patients had enlarged imaging abnormality at last followup, only 1.8% of patients required intervention for a progressive process suggestive of tumor growth. This has been demonstrated in other studies reporting imaging outcome where the majority of patients with imaging enlargement did not require further intervention for recurrent tumor, and treatment failure was defined as progressive growth without stabilization of tumor volume (15,27). Boari et al. in reviewing the literature emphasized that for 15-30% of cases, there is a transient increase in tumor volume; for 5-10% of cases, there is stabilization after increase in tumor volume; and for <5% of cases, continuous growth is considered to be true progression (27). Although very long-term follow-up is required to confirm the significance of enlargement in this patients, we hypothesize that there may be both temporary and permanent radiosurgery-related imaging changes. The majority of patients in recent studies experience no permanent damage to CN function after HSRT (5,9,32,33). In one study, all patients with VS treated with HSRT had total preservation of long-term CN function; in another, no permanent trigeminal nerve or facial nerve toxicity was observed in any patient as a result of treatment, and in a third study of 60 patients with VS treated with HSRT, there were no severe treatment-related CN complications, and less than 15% of patients experienced any new CN symptom after treatment (5,9,34). Only 1.7% demonstrated a facial nerve palsy or paralysis, similar to other reports ( Table 6). Indeed, Meijer et al. demonstrate that the 5-year preservation probability of facial nerve function after HSRT is near 100% (6). Similarly, we found that in the 9% who had new CNVII and 9.4% who had new CNVIII deficits after HSRT, 57 and 50% of those individuals, respectively, recovered CN function ( Table 4). In this retrospective study, we could not assess the severity of the deficits.
Hearing preservation is a topic of great concern after radiation therapy for VS, particularly given the close proximity of the structures of hearing to the planning target volume of radiation. The use of hypofractionated stereotactic treatment as an alternative to single-dose stereotactic radiosurgery was driven by the goal of taking advantage of the greater therapeutic ratio observed with fractionation with regard to hearing preservation. Fractionated SRT provides the benefit of delivering a lower dose of radiation during each treatment visit to such critical structures, rather than the high-dose radiation delivered by SRS. Thus far, the majority of studies of hearing preservation after HSRT demonstrate that greater than half of patients have preserved serviceable hearing (42-81%; Table 6) (3,9,10,32,33,(35)(36)(37), although gradient deterioration of pure tone averages has been observed even in those with hearing preservation (7). Hayden Gephart et al. comment that a higher radiation dose and larger cochlear volume within the radiation field result in poorer hearing outcomes posttreatment (37). Following this logic, the delivery of a high total dose of radiation divided into few low-dose fractions via HSRT would improve hearing outcomes relative to SRS. Indeed, HSRT has demonstrated improved hearing preservation relative to SRS in a number of studies (42-81.5% preserved hearing vs. 36-60%; Table 6) (27-29, 32, 33, 35, 36). Our results corroborate previous findings, as we observed preservation of serviceable hearing in 51% of patients after HSRT (Tables 5 and 6). This result is similar to the hearing preservation observed with stereotactic radiosurgery and does not support superiority of hypofractionated stereotactic treatment ( Table 6). Overall, less than 5% of patients in our cohort continued to experience treatment-related CNVII and CNVIII deficit beyond our follow-up time of 6 years, which is consistent with the current literature regarding functional CN outcomes after HSRT and is comparable to CN outcomes after SRS (15).
A weakness of this study is that it is a retrospective analysis of patients from a wide geographic base such that a significant proportion were lost to follow-up as they ceased care and followup at the treating facility. Indeed, large series reporting results of therapeutic options for VS are retrospective, and comparing results should be done with caution and understanding of the limitations. Nevertheless, this report of outcome for a large cohort of patients treated with HSRT does not provide evidence to support the hypothesis that control or toxicity is superior to the outcome achieved with standard single-dose radiosurgery. This study represents results with this treatment approach of hypofractionated stereotactic radiation, and as such there is no direct comparison group; the results must be interpreted in the context of results at other institutions with this limitation in mind.
We show that HSRT is an appropriate treatment modality for unilateral VS, with a high tumor control rate, low CN toxicity, and acceptable hearing preservation relative to prior studies of SRS. However, the results do not appear to confirm that hearing preservation and toxicity could be superior to the outcome with standard single-dose radiosurgery, and we have reverted to that approach for lesions meeting appropriate size criteria. Continued research to document long-term outcomes and to reduce treatment morbidity is required. Research directions may be directed at better selection of those who require treatment rather than