Vestibular schwannomas (VSs) are histologically

Transcription

1 CLINICAL ARTICLE J Neurosurg 127: , 2017 Magnetic resonance imaging characteristics and the prediction of outcome of vestibular schwannomas following Gamma Knife radiosurgery Chih-Chun Wu, MD, 1,2 Wan-Yuo Guo, MD, PhD, 1,2 Wen-Yuh Chung, MD, 2,3 Hisu-Mei Wu, MD, 1,2 Chung-Jung Lin, MD, 1,2 Cheng-Chia Lee, MD, 3 Kang-Du Liu, MD, 2,3 and Huai-che Yang, MD 3 1 Department of Radiology and 3 Department of Neurosurgery, Neurological Institute, Taipei Veterans General Hospital; and 2 School of Medicine, National Yang-Ming University, Taipei, Taiwan, Republic of China OBJECTIVE Gamma Knife surgery (GKS) is a promising treatment modality for patients with vestibular schwannomas (VSs), but a small percentage of patients have persistent postradiosurgical tumor growth. The aim of this study was to determine the clinical and quantitative MRI features of VS as predictors of long-term tumor control after GKS. METHODS The authors performed a retrospective study of all patients with VS treated with GKS using the Leksell Gamma Knife Unit between 2005 and 2013 at their institution. A total of 187 patients who had a minimum of 24 months of clinical and radiological assessment after radiosurgery were included in this study. Those who underwent a craniotomy with tumor removal before and after GKS were excluded. Study patients comprised 85 (45.5%) males and 102 (54.5%) females, with a median age of 52.2 years (range years). Tumor volumes, enhancing patterns, and apparent diffusion coefficient (ADC) values were measured by region of interest (ROI) analysis of the whole tumor by serial MRI before and after GKS. RESULTS The median follow-up period was 60.8 months (range months), and the median treated tumor volume was 3.54 cm 3 ( cm 3 ). At last follow-up, imaging studies indicated that 150 tumors (80.2%) showed decreased tumor volume, 20 (10.7%) had stabilized, and 17 (9.1%) continued to grow following radiosurgery. The postradiosurgical outcome was not significantly correlated with pretreatment volumes or postradiosurgical enhancing patterns. Tumors that showed regression within the initial 12 months following radiosurgery were more likely to have a larger volume reduction ratio at last follow-up than those that did not (volume reduction ratio 55% vs 23.6%, respectively; p < 0.001). Compared with solid VSs, cystic VSs were more likely to regress or stabilize in the initial postradiosurgical 6 12-month period and during extended follow-up. Cystic VSs exhibited a greater volume reduction ratio at last follow-up (cystic vs solid: 67.6% ± 24.1% vs 31.8% ± 51.9%; p < 0.001). The mean preradiosurgical maximum ADC (ADC max ) values of all VSs were significantly higher for those with tumor regression or stabilization at last follow-up compared with those with progression (2.391 vs mm 2 /sec; p = 0.010). CONCLUSIONS Loss of central enhancement after radiosurgery was a common phenomenon, but it did not correlate with tumor volume outcome. Preradiosurgical MRI features including cystic components and ADC max values can be helpful as predictors of treatment outcome. KEY WORDS cystic vestibular schwannoma; Gamma Knife radiosurgery; magnetic resonance imaging; vestibular schwannoma; stereotactic radiosurgery Vestibular schwannomas (VSs) are histologically benign tumors of the vestibular division of the eighth cranial nerve. Nowadays, radiosurgery has a well-established role in the treatment of VSs, 38 with reported effective growth control of more than 93%. 3,8, 11, 12, 16, 17, 21, 27,51 The 7-year actuarial clinical tumor control or resection-free survival rate at our institution is 95% of 195 patients treated by Gamma Knife surgery (GKS). 8 However, treatment failure, although it affects only a small percentage of patients, still occurs. The purpose of the current ABBREVIATIONS ADC = apparent diffusion coefficient; GKS = Gamma Knife surgery; GTV = gross target/tumor volume; ROI = region of interest; VS = vestibular schwannoma. SUBMITTED June 9, ACCEPTED September 28, INCLUDE WHEN CITING Published online February 10, 2017; DOI: / JNS J Neurosurg Volume 127 December 2017 AANS, 2017

2 MRI predicts vestibular schwannoma radiosurgery study was to identify imaging predictors of the treatment response of VSs to GKS. Methods Patient Characteristics In this retrospective study, approved by our institutional review board, we reviewed a collected database of all patients with VS treated between January 2005 and June 2013 at our institution using the Leksell Gamma Knife Unit (Elekta Instruments). The database has been collected prospectively since 1992 by a team of neuroradiologists and neurosurgeons. Patients with non-neurofibromatosis Type 2 VS and a minimum of 24 months of clinical and radiological assessment before and after radiosurgery were included in the study. The excluded patients were those who had undergone surgery for tumor removal before GKS and those who had undergone surgery or more than 1 session of radiosurgery within 24 months after their first GKS. In all, a total of 187 patients meeting the abovementioned criteria were enrolled in the study. They comprised 85 (45.5%) males and 102 (54.5%) females, with a median age of 52.2 years (range years). The tumor was located on the right side in 100 patients (53.5%) and on the left side in 87 (46.5%). Twenty-seven tumors (14.4%) were located in the internal acoustic meatus, 9 (4.8%) in the cerebellopontine angle, and 151 (80.7%) in both the internal acoustic meatus and cerebellopontine angle. Radiosurgical Technique Treatments were performed using the Leksell Gamma Knife B model between January 2005 and May 2006 and the Leksell Gamma Knife 4C thereafter. After stereotactic frame application, target detection was performed using high-resolution transaxial and coronal contrast-enhanced 1-mm section T1-weighted imaging and T2-weighted imaging sequences in all cases. The median prescribed dose delivered to the margin of the tumor was 12 Gy (range Gy), and the median maximum dose was Gy (range Gy). The median mean dose was 16.3 Gy (range Gy). The mean number of isocenters was 9.54 (range 1 27). Radiological Characteristics and Quantification MR images were obtained in all patients by using a 1.5- T scanner (Signa Horizon LX2, GE Medical Systems). All patients were followed up with serial contrast-enhanced MRI examinations at 6 and 12 months after radiosurgery and then yearly thereafter. MR enhancing pattern analysis, tumor volume, and diffusion coefficient measurements were performed on a picture archiving and communication system before radiosurgery and at each follow-up time point by experienced neuroradiologists (C.C.W., W.Y.G., and H.M.W.) at our institution. The enhancing pattern on MRI within the first 6 and 12 months after GKS was defined according to constant homogeneous enhancement and initial loss of central enhancement. Cystic VS was defined according to the definition of previous studies regarding cystic VS after radiosurgery, and referred to tumors with cysts (including both intramural and extramural cysts) making up at least one-third of the whole tumor volume. 33,39 The diffusion-weighted imaging parameters were as follows: TE msec, TR msec, matrix size , field of view 300 mm, slice thickness 5 mm, voxel resolution 1.2 mm 1.2 mm 6.0 mm, and number of excitations 1. Diffusion was measured in 3 orthogonal directions by use of 3 b-values (0, 500, and 1000 s/mm 2 ). Postprocessing of diffusion-weighted imaging data with calculation of apparent diffusion coefficient (ADC) maps was performed using OsiriX MD software version 7.5 (Pixmeo SARL) with ADC Map Calculation plugin version 1.9 (Stanford University). In each slice, a region of interest (ROI) was delineated according to the tumor geometry on ADC maps in conjunction with T2- weighted and postcontrast sequences. The pixel-by-pixel mean ADC (ADC mean ), minimum ADC (ADC min ), and maximum ADC (ADC max ) values for the entire volume were calculated using OsiriX MD software. Considering that diffusion-weighted images are more susceptible to field inhomogeneity, we performed ADC measurements with preradiosurgical imaging before stereotactic frame placement. To reduce a partial volume effect from CSF, tumors smaller than 1 slice thickness (3 mm) were excluded from ADC measurements. The gross tumor volume or gross target volume (GTV), was calculated as the sum of the products of the surface area and thickness of slices on all tumor slices on a picture archiving and communication system. The surface area was obtained by neuroradiologists, who manually traced the tumor on each high-resolution Gd-enhanced spin-echo T1-weighted MR image, with a 1 3-mm slice thickness without gaps. The same imaging protocol and volumetric measurement were performed at every followup time point. Absolute changes in volume were calculated by subtracting the GTV at treatment from volume at the last follow-up. The tumor volume reduction ratio was calculated as the absolute change divided by the GTV at diagnosis. Considering the cutoff value of a 10% change in volume as the generally accepted limit of growth/progression, the postradiosurgical volume change was defined as significant enlargement or regression if the tumor volume increased or decreased, respectively, more than 10% compared with baseline GTV at GKS. Otherwise, the case was considered stable. Transient tumor enlargement, in which a transient increase in tumor volume occurred within the first postradiosurgical 12 months followed by either stability or regression, was considered pseudoprogression. Considering treatment outcome, tumor control was defined as volume regression or stable disease, whereas treatment failure was defined as tumor enlargement at last follow-up. Statistical Analysis To determine the independent predictors of the tumor volume reduction ratio and treatment outcome, the following were assessed: sex (male vs female), age, prescribed dosage, maximum dose, mean dose, number of isocenters, tumor volume and volume reduction ratio, status of tumor volume changes (transient enlargement or not), and MRI features including enhancing patterns, enhancing pattern changes, and ADC values. Fisher s exact and chi-square tests were used for cate- J Neurosurg Volume 127 December

3 C. C. Wu et al. TABLE 1. Univariate analysis of variables that influence tumor control Variables Tumor Regression or Stable Tumor Progression p Value Patient age (yrs)* 52 (55, 20 82) 51 (51, 25 72) Female/male ratio 93:77 9: GTV (cm 3 )* 3.66 (2.1, ) 2.38 (1.2, ) Prescribed dose (Gy)* 12 (12, 11 13) 12 (12, 12 13) Maximum dose (Gy)* 21 (21, 17 23) 20 (21, 15 23) No. of isocenters 9.6 (9, 1 27) 9.1 (9, 4 17) Follow-up (mos)* 66 (60, ) 60 (60, ) Cystic VS/solid VS 36:134 0: Mean pre-gk ADC ± SD ( 10-3 mm 2 /sec) Maximum ± ± Mean ± ± Minimum ± ± Maximum (in solid VS) 2.291± ± Mean ADC at 6 mos ± SD ( 10-3 mm 2 /sec) Maximum ± ± Mean ± ± Minimum ± ± Loss enhancement after GKS (present: absent) 6 mos 144:26 15: mos 109:61 12: Enlarge after GKS (present: absent) 6 mos 84:83 7: mos 38:126 9: Mean volume reduction ratio ± SD (%) 50 ± ± 76.9 <0.001 * Values represent mean (median, range). Significant difference at p < gorical data comparisons, and differences in mean values between 2 groups with categorical factors were tested for significance using an independent samples t-test. Linear regression was used to test the volume reduction ratio and continuous variables, with predictive performance and optimal evaluation cutoff values determined using receiver operating characteristic curves and areas under the curve. Statistical significance was set at probability values < All analyses were performed with SPSS (version 22.0, IBM). Results Tumor Characteristics and Control The median follow-up period was 60.8 months (range months, mean 64.4 months), and the mean treated tumor volume was 3.54 cm 3 (median 2.0 cm 3, range cm 3 ). At last follow-up, 150 tumors (80.2%) showed decreased tumor volume, 20 (10.7%) had stabilized, and 17 (9.1%) continued to grow following radiosurgery. The mean GTV of tumors that remained either stable or regressed (mean 3.66 ± 3.74 cm 3, median 2.1 cm 3, range cm 3 ) was not significantly different from the mean GTV of those that progressed (mean 2.38 ± 2.66 cm 3, median 1.2 cm 3, range cm 3 ; p = 0.171; Table 1). The mean last tumor volume was 1.78 cm 3 (median 0.9 cm 3, range cm 3 ) with a mean tumor volume reduction ratio of 38.7% for all tumors. For tumors that shrunk at last follow-up, the mean tumor volume reduction ratio reached 55% (median 59%, range 12% 100%), and for tumors that continued to grow, the mean tumor enlargement ratio was 74% (median 29%, range 10% 249%). Tumor Morphology Versus Treatment Outcome Of all 187 tumors, 36 (19%) were cystic VSs containing cystic components and 151 (81%) were solid VSs. A total of 122 (65%) of 187 tumors showed homogenous enhancement, and 65 (35%) of 187 tumors had various degrees of nonenhancing areas before GKS (Table 2). The enhancing pattern at the time of GKS did not significantly correlate with the volume reduction ratio at last follow-up. After radiosurgery, 162 tumors (87%) showed initial loss of central enhancement, and 97 tumors (52%) showed pseudoprogression on follow-up MRI. Tumors with larger pretreatment volumes were more likely to have initial loss of enhancement (p < 0.001), but there was no significant correlation between loss of enhancement and pseudoprogression after radiosurgery. The postradiosurgical outcome was not significantly correlated with postradiosurgical enhancing patterns, but those tumors without pseudoprogression were more likely to regress or stabilize at last follow-up (p = 0.008). Tumors with pseudoprogression 1386 J Neurosurg Volume 127 December 2017

4 MRI predicts vestibular schwannoma radiosurgery TABLE 2. MRI characteristics of VS and volume reduction ratio Characteristics No. of Patients Volume Reduction Ratio (%) p Value Enhancement at GKS Homogenous Heterogeneous VS morphology Cystic <0.001 Solid Enlargement at 6 12 mos after GKS Present <0.001 Absent Loss enhancement at 6 12 mos Present Absent FIG. 1. Transaxial MR images showing a patient with a left-side cystic VS. A: A T2-weighted image obtained before GKS displays multiple cysts in the tumor. B: An ADC map image illustrating the placement of the ROI (yellow outline) when ADC values were measured; the max ADC value is mm 2 /sec. C: A Gd-enhanced T1-weighted image shows enhancement of the cystic walls, which confirms a cystic tumor instead of arachnoid cysts surrounding the tumor. D: Image shows the shrunken tumor at the 5.5-year follow-up, with a volume reduction ratio of 82.6%. Figure is available in color online only. following radiosurgery were more likely to have a smaller volume reduction ratio at last follow-up than those that did not (volume reduction ratio 24% vs 55%, p < 0.001; Table 2). In this analysis, the mean GTV of cystic VSs was significantly larger than that for solid VSs (6.2 ± 3.8 cm 3 vs 2.9 ± 3.4 cm 3 ; p < 0.001). Compared with solid VSs, cystic VSs were more likely to regress or stabilize in the initial postradiosurgical 6 12 months and at follow-up. Cystic tumors also had greater volume reduction ratios at last follow-up (cystic vs solid: 67.6% ± 24.1% vs 31.8% ± 51.9%; p < 0.001, Table 2; Figs. 1 and 2). The mean preradiosurgical ADC max of all VSs was significantly higher for those with tumor regression or stabilization at the last imaging follow-up compared with those with progression (2.391 ± vs ± mm 2 /sec; p = 0.010). Those that showed tumor regression or stabilization at the last imaging follow-up had significantly higher ADC mean at 6 months postradiosurgery than those without (1.534 ± vs ± mm 2 / sec; p = 0.004). However, there was substantial overlap of mean ADC mean 6-month postradiosurgery values between groups (Table 1). Considering the different tumor components, the mean preradiosurgical ADC max of solid VSs was also significantly higher for those with tumor regression or stabilization at the last imaging follow-up compared with those with progression (2.291 ± vs ± mm 2 /second; p = 0.033). Of those tumors showing enlargement at 12 months after GKS, tumors that regressed at last follow-up had significantly higher mean preradiosurgical ADC mean values than those with persistent tumor enlargement or stabilization (1.357 ± 0.25 vs ± mm 2 /second; p = 0.038). Using the receiver operating characteristic curve, a cutoff value of mm 2 /sec for preradiosurgical ADC mean generated the optimal combination of sensitivity (69.2%) and specificity (70%), and had good discriminating power, with the AUC at 0.68 (95% confidence interval , p = 0.041). The ADC values changed heterogeneously after GKS, and the changes in the ADC did not significantly correlate with the presence of pseudoprogression or final outcome. This was true whether considering all VS cases or just the solid or cystic subtypes. Radiosurgical Parameters The maximum dose, prescribed dose, and mean dose of radiation, and the number of isocenters and mean isodose line, did not significantly influence tumor volume outcome (Table 1). Discussion VSs are heterogeneous tumors, with various different tissue compositions, growth rates, MRI appearances, and responses to treatment. GKS has proven to be a highly effective treatment for VS, 3,29,36,39 but the course after GKS is not uniform, with dynamic volume changes generally observed. While most VSs may shrink after transient enlargement within the first 2 years after GKS, a small percentage of VSs continues to grow. Currently no known predictors of VS response to radiosurgery have been discovered, and a wait-and-see policy is the current strategy. In this study we provide a comprehensive analysis that sought to determine whether any imaging characteristics could serve as predictors of treatment response in the early post-gks follow-up period before persistent tumor growth is evident. Cystic VS Versus Solid VS Outcomes Following GKS One of the main findings of our study is that cystic VSs exhibited a greater volume reduction ratio after GKS. In J Neurosurg Volume 127 December

5 C. C. Wu et al. FIG. 2. Transaxial MR images showing a patient with a left-side solid VS. A: A Gd-enhanced T1-weighted image clearly shows enhancement of the tumor. B: An ADC map image illustrating the location of the ROI (yellow outline) when ADC values were measured; the max ADC value is mm 2 /sec. C: Image at 12 months after GKS shows the tumor enlarged, without obvious loss of central enhancement. D: Image at 7.5 years after GKS shows the tumor relatively larger than it was before GKS. The volume reduction ratio is -11.6%. Figure is available in color online only. previous studies, cystic VSs were reported to grow quickly in nature with rapid symptomatic worsening or an unpredictable biological behavior. 45,52 Compared with small solid VSs, cystic VSs were associated with more difficult surgical procedures, a greater rate of facial nerve dysfunction, and even morbidity in some studies. 2,15,22,28,41,46,52,53 However, the treatment outcomes for cystic VSs after GKS have been sparsely reported. An earlier study by Pendl et al. showed a high incidence (6/6 patients) of early symptomatic enlargement of the cystic component 4 8 months after GKS, and they cautioned against the use of GKS for cystic VS. However, a slow spontaneous size decrease after 2 years in 2 of the 6 patients was also mentioned. 34 Another study of a large patient cohort including 54 cystic VSs, treated with GKS by Delsanti and Régis, 9 showed that the failure rate (defined as a need for a second procedure) was 6.4%, higher than that reported for solid VSs. Nonetheless, these authors also reported that cystic VS was the group in which they observed the most dramatic shrinkage. Correspondingly, a study by Shirato et al. revealed that transient enlargement of tumor size and a transient worsening of trigeminal neuropathy is more frequent in cystic VS after stereotactic radiosurgery, but interestingly, the ultimate tumor-reduction rate for cystic tumors was found to be larger than that for solid tumors by fractionated stereotactic radiosurgery. 40 Our results are in accordance with these findings, demonstrating a larger volume reduction ratio for tumors with cystic VS than tumors with solid VS. Cyst formation in VS is commonly observed in radiological studies, but the incidence differs among imaging modality and the definition of cystic VS, with a reported incidence of 5.7% 20.4% reported in the literature. 13,22,23,41 In our series, tumors were classified as cystic according to MRI appearance on the day of GKS. In the literature, few studies on the mechanism of cyst formation and enlargement in cystic VS have been published, and these had inconclusive findings. From early histological studies, 4 the coalescence of microcysts in the Antoni type B area in VS had been suggested to cause grossly detectable cystic areas in neuroimaging. 4,30,49 Some investigators have reported differences in the distribution of Antoni type cells between cystic and solid VSs, showing that cyst formation is due to the degeneration of tumor tissues. 14,24,30 On the other hand, a study by Park et al. showed no significant relationship between cyst formation and the distribution of Antoni cell types. Instead, these authors reported that abnormal vessel proliferation is significantly more prominent in cystic VSs compared with solid VSs based on histological evidence. 33 Other pathogeneses, including intratumoral hemorrhage, 14 necrosis, 45 and secretion within the type B tissue, 4,5 are proposed mechanisms leading to cyst formation in VS. Moreover, the density of cells positive for Ki-67, a marker of cell proliferation, was found to be higher in cystic VSs than in solid VSs. 4 A heterogeneous enhanced appearance of larger VSs was found more commonly in lesions with a higher ratio of Antoni type B tissue to type A tissue. 14 In the current study, the degree of enhancement was not a predictor of post-gks outcome, but the presence of cysts inside VSs was correlated with greater volume reduction ratios at last follow-up. According to the few histological studies on resections of postradiosurgical residual or growing tumors, the biological effects of radiosurgery on VS are apoptotic pathways and vasculopathy. 25 As compared with the solid VS with an inherently low proliferative index, the high density of Ki-67 positive cells, which represent more radiosensitive dividing cells, and the abnormal vessel proliferation observed in cystic VS may be mechanisms underlying the greater tumor shrinkage exhibited by these tumors following GKS. In addition, the radiation effects on apoptotic pathways and vasculature are usually not immediate and may be attributable to the delayed regression of cystic VSs observed in this and previous studies. ADC Values as Useful Predictors of Early VS Response Following GKS Another main finding of our study is that pre-gks ADC values of VSs were helpful in evaluating treatment results and in distinguishing transient tumor enlargement or persistent tumor growth early in the posttreatment period when other imaging results are not definitive. Diffusion imaging is a sensitive imaging biomarker, representing cellularity and extracellular free water content, and has shown promise as a robust approach for preoperative tumor typing, 47 preoperative glioma grading, 1,44,50 posttreatment glioma responses, 6,42 and meningioma grading and prediction of postoperative recurrence. 20,43 In addition, the ADC value is reported as a practical method for evaluating the efficacy of stereotactic radiosurgery and predicting 1388 J Neurosurg Volume 127 December 2017

6 MRI predicts vestibular schwannoma radiosurgery early intracranial metastatic response to GKS. 26 Cellular density is considered a major factor responsible for changes in ADCs. In the current study, we used whole-tumor analysis, which is the most common technique for assessing therapeutic response. The high preradiosurgical ADC values of VSs may reflect cystic components, the increase in the tumor s fluid component, and lower nucleus-to-cytoplasm ratios or lower cellularity in VS. They also probably reflect the higher density of Antoni Type B tissue. 7 VSs with a higher content of radiographically identified cysts and Antoni type B tissue might represent areas with higher ADC values and a better response to radiosurgery. There was no significant correlation between treatment outcome and the absolute ADC values or the differences of ADC values at 6 months and 12 months after GKS in the current study. These results are in agreement with the results from the study by Chuang et al., who reported that after GKS, the ADC values for solid VSs significantly increased from the first follow-up at 6 months, but ADC values did not significantly differ among tumors, regardless of whether they exhibited increased, decreased, or stable volumes. 7 However, the placement of the ROIs was not clearly described in their study, and the ADC values may vary depending on the selected MRI slice, especially in tumors with mixed cystic and solid components. Moreover, in the current study, we found that among the tumors showing enlargement at the first 12 months after radiosurgery, tumors with higher pre-gks ADC mean (> mm 2 /sec) were more likely to shrink after 12 months. We suggest that preradiosurgical ADC values could be useful for predicting pseudoprogression or treatment failure early in the posttreatment period. Loss of Central Enhancement Is Not a Predictor of Tumor Response Loss of central contrast enhancement is a common morphological change in the early postradiosurgical period, with a reported incidence ranging from 54% by Prasad et al. 37 to 84% by Nakamura et al. 32 Loss of central contrast enhancement is generally transient. The incidence of this loss was 86.7% in our study. The postradiosurgical tumor enhancing patterns are usually classified as a transient loss of enhancement, continuous increase in enhancement, or no change in enhancement. 32 Several possible mechanisms for the loss of central enhancement have been proposed, including tumor necrosis and obvious vascular damage based on histopathological findings, 8 decreased vascularity in Tc99 human serum albumin-dtpa PET/CT, hyperacute tumor ischemia and associated edema, and apoptosis. However, none of these clarifies the actual meaning of the development or absence of central enhancement. In an earlier study by Nakamura et al., changes in contrast enhancement were not found to be predictive of clinical outcome. 32 In the current study, we also found no significant correlation between postradiosurgical outcome and the postradiosurgical MRI enhancement pattern in the initial 6 12 months following radiosurgery. We postulate that early postradiosurgical loss of central enhancement represents an early radiation effect and is not a predictor of tumor volume control. A transient increase in size followed by stability or regression has been recognized as treatment effects rather than treatment failures in the first 24 months after GKS in many studies. 19,31, 32, 35,48 However, the mechanisms of these different volumetric responses to radiosurgery have yet to be fully elucidated. A relationship between alternations in enhancing and growth patterns has also been reported, but the results are inconsistent. Nakamura et al. suggest that no significant correlation exists between changes in tumor volume and tumor enhancement on MRI. 32 On the other hand, studies by Hasegawa et al. 18 and Régis et al. 10 show that initial loss of central enhancement is usually associated with the presence of transient tumor expansion. Our study showed that tumors with larger pre-gks volumes were more likely to exhibit initial loss of enhancement on MRI after radiosurgery, but the loss of enhancement was correlated with neither early tumor enlargement nor treatment outcome. Although pseudoprogression within the initial 6 12 months after radiosurgery was commonly observed, we found that VSs with a larger tumor volume ratio within the first 12 months after GKS were more likely to show a lesser degree of volume reduction at last follow-up. The limitations of the present study include the singleinstitution nature of the study and the relatively small number of patients. The patients in the failure group were limited (a total of 17), and multivariate analysis was not performed due to the small sample size. Furthermore, the application of imaging features and ADC values to modify the radiosurgical dose could not be evaluated thoroughly. Accordingly, further prospective studies on larger cohorts of patients are warranted. Nevertheless, this is still one of the larger series published on this topic, and it identifies specific MRI features, including ADC values, as predictors of the treatment response of VSs to GKS. Conclusions GKS is an effective treatment modality for both solid and cystic VS. The presence of loss of central enhancement after radiosurgery is a common phenomenon, which does not correlate with tumor volume outcomes. Transient enlargement of tumor size is also frequently observed in the initial 6 12 months following radiosurgery, and is associated with a lesser degree of volume reduction at last follow-up. Preradiosurgical MRI features, including cystic components and ADC max values, can be helpful when the results of early posttreatment imaging are ambiguous. Cystic VS ultimately can have a larger volume reduction ratio than that of solid VS in the long term. Acknowledgments We thank Miss Feng-Jiau Lee and Miss Shueh-Jen Huang for their assistance in the preparation of this paper. This paper was part of a research project (project for a junior researcher) of the Ministry of Science and Technology, Taiwan (no B ). References 1. Arevalo-Perez J, Peck KK, Young RJ, Holodny AI, Karimi S, Lyo JK: Dynamic contrast-enhanced perfusion MRI and diffusion-weighted imaging of grading of gliomas. J Neuroimaging 25: , Benech F, Perez R, Fontanella MM, Morra B, Albera R, J Neurosurg Volume 127 December

7 C. C. Wu et al. Ducati A: Cystic versus solid vestibular schwannomas: a series of 80 grade III IV patients. Neurosurg Rev 28: , Boari N, Bailo M, Gagliardi F, Franzin A, Gemma M, del Vecchio A, et al: Gamma Knife radiosurgery for vestibular schwannoma: clinical results at long-term follow-up in a series of 379 patients. J Neurosurg 121 Suppl: , Charabi S, Klinken L, Tos M, Thomsen J: Histopathology and growth pattern of cystic acoustic neuromas. Laryngoscope 104: , Charabi S, Tos M, Thomsen J, Rygaard J, Fundova P, Charabi B: Cystic vestibular schwannoma clinical and experimental studies. Acta Otolaryngol Suppl 543:11 13, Chu HH, Choi SH, Ryoo I, Kim SC, Yeom JA, Shin H, et al: Differentiation of true progression from pseudoprogression in glioblastoma treated with radiation therapy and concomitant temozolomide: comparison study of standard and high-b-value diffusion-weighted imaging. Radiology 269: , Chuang CC, Chang CS, Tyan YS, Chuang KS, Tu HT, Huang CF: Use of apparent diffusion coefficients in evaluating the response of vestibular schwannomas to Gamma Knife surgery. J Neurosurg 117 Suppl:63 68, Chung WY, Liu KD, Shiau CY, Wu HM, Wang LW, Guo WY, et al: Gamma Knife surgery for vestibular schwannoma: 10-year experience of 195 cases. J Neurosurg 102 Suppl:87 96, Delsanti C, Régis J: [Cystic vestibular schwannomas.] Neurochirurgie 50: , 2004 (Fr) 10. Delsanti C, Roche PH, Thomassin JM, Régis J: Morphological changes of vestibular schwannomas after radiosurgical treatment: pitfalls and diagnosis of failure. Prog Neurol Surg 21:93 97, Flickinger JC, Kondziolka D, Niranjan A, Maitz A, Voynov G, Lunsford LD: Acoustic neuroma radiosurgery with marginal tumor doses of 12 to 13 Gy. Int J Radiat Oncol Biol Phys 60: , Foote KD, Friedman WA, Buatti JM, Meeks SL, Bova FJ, Kubilis PS: Analysis of risk factors associated with radiosurgery for vestibular schwannoma. J Neurosurg 95: , Fundová P, Charabi S, Tos M, Thomsen J: Cystic vestibular schwannoma: surgical outcome. J Laryngol Otol 114: , Gomez-Brouchet A, Delisle MB, Cognard C, Bonafe A, Charlet JP, Deguine O, et al: Vestibular schwannomas: correlations between magnetic resonance imaging and histopathologic appearance. Otol Neurotol 22:79 86, Gormley WB, Sekhar LN, Wright DC, Kamerer D, Schessel D: Acoustic neuromas: results of current surgical management. Neurosurgery 41:50 60, Hasegawa T, Kida Y, Kato T, Iizuka H, Kuramitsu S, Yamamoto T: Long-term safety and efficacy of stereotactic radiosurgery for vestibular schwannomas: evaluation of 440 patients more than 10 years after treatment with Gamma Knife surgery. J Neurosurg 118: , Hasegawa T, Kida Y, Kobayashi T, Yoshimoto M, Mori Y, Yoshida J: Long-term outcomes in patients with vestibular schwannomas treated using Gamma Knife surgery: 10-year follow up. J Neurosurg 102:10 16, Hasegawa T, Kida Y, Yoshimoto M, Koike J, Goto K: Evaluation of tumor expansion after stereotactic radiosurgery in patients harboring vestibular schwannomas. Neurosurgery 58: , Hayhurst C, Zadeh G: Tumor pseudoprogression following radiosurgery for vestibular schwannoma. Neuro Oncol 14:87 92, Hwang WL, Marciscano AE, Niemierko A, Kim DW, Stemmer-Rachamimov AO, Curry WT, et al: Imaging and extent of surgical resection predict risk of meningioma recurrence better than WHO histopathological grade. Neuro Oncol 18: , Inoue HK: Low-dose radiosurgery for large vestibular schwannomas: long-term results of functional preservation. J Neurosurg 102 Suppl: , Jian BJ, Sughrue ME, Kaur R, Rutkowski MJ, Kane AJ, Kaur G, et al: Implications of cystic features in vestibular schwannomas of patients undergoing microsurgical resection. Neurosurgery 68: , Jones SE, Baguley DM, Moffat DA: Are facial nerve outcomes worse following surgery for cystic vestibular schwannoma? Skull Base 17: , Kameyama S, Tanaka R, Kawaguchi T, Fukuda M, Oyanagi K: Cystic acoustic neurinomas: studies of 14 cases. Acta Neurochir (Wien) 138: , Kondziolka D, Shin SM, Brunswick A, Kim I, Silverman JS: The biology of radiosurgery and its clinical applications for brain tumors. Neuro Oncol 17:29 44, Lee CC, Wintermark M, Xu Z, Yen CP, Schlesinger D, Sheehan JP: Application of diffusion-weighted magnetic resonance imaging to predict the intracranial metastatic tumor response to Gamma Knife radiosurgery. J Neurooncol 118: , Lunsford LD, Niranjan A, Flickinger JC, Maitz A, Kondziolka D: Radiosurgery of vestibular schwannomas: summary of experience in 829 cases. J Neurosurg 102 Suppl: , Metwali H, Samii M, Samii A, Gerganov V: The peculiar cystic vestibular schwannoma: a single-center experience. World Neurosurg 82: , Milligan BD, Pollock BE, Foote RL, Link MJ: Long-term tumor control and cranial nerve outcomes following Gamma Knife surgery for larger-volume vestibular schwannomas. J Neurosurg 116: , Moon KS, Jung S, Seo SK, Jung TY, Kim IY, Ryu HH, et al: Cystic vestibular schwannomas: a possible role of matrix metalloproteinase-2 in cyst development and unfavorable surgical outcome. J Neurosurg 106: , Nagano O, Higuchi Y, Serizawa T, Ono J, Matsuda S, Yamakami I, et al: Transient expansion of vestibular schwannoma following stereotactic radiosurgery. J Neurosurg 109: , Nakamura H, Jokura H, Takahashi K, Boku N, Akabane A, Yoshimoto T: Serial follow-up MR imaging after Gamma Knife radiosurgery for vestibular schwannoma. AJNR Am J Neuroradiol 21: , Park CK, Kim DC, Park SH, Kim JE, Paek SH, Kim DG, et al: Microhemorrhage, a possible mechanism for cyst formation in vestibular schwannomas. J Neurosurg 105: , Pendl G, Ganz JC, Kitz K, Eustacchio S: Acoustic neurinomas with macrocysts treated with Gamma Knife radiosurgery. Stereotact Funct Neurosurg 66 (Suppl 1): , Pollock BE: Management of vestibular schwannomas that enlarge after stereotactic radiosurgery: treatment recommendations based on a 15 year experience. Neurosurgery 58: , Pollock BE, Driscoll CL, Foote RL, Link MJ, Gorman DA, Bauch CD, et al: Patient outcomes after vestibular schwannoma management: a prospective comparison of microsurgical resection and stereotactic radiosurgery. Neurosurgery 59:77 85, Prasad D, Steiner M, Steiner L: Gamma surgery for vestibular schwannoma. J Neurosurg 92: , Régis J, Carron R, Moucharrafien S, Delsantin C, Porcheron D, Thomassin JM, et al: [Role of radiosurgery and stereotactic radiotherapy in the management of vestibular schwannomas.] Cancer Radiother 16 Suppl:S70 S78, 2012 (Fr) 1390 J Neurosurg Volume 127 December 2017

8 MRI predicts vestibular schwannoma radiosurgery 39. Rutten I, Baumert BG, Seidel L, Kotolenko S, Collignon J, Kaschten B, et al: Long-term follow-up reveals low toxicity of radiosurgery for vestibular schwannoma. Radiother Oncol 82:83 89, Shirato H, Sakamoto T, Takeichi N, Aoyama H, Suzuki K, Kagei K, et al: Fractionated stereotactic radiotherapy for vestibular schwannoma (VS): comparison between cystic-type and solid-type VS. Int J Radiat Oncol Biol Phys 48: , Sinha S, Sharma BS: Cystic acoustic neuromas: surgical outcome in a series of 58 patients. J Clin Neurosci 15: , Smith JS, Cha S, Mayo MC, McDermott MW, Parsa AT, Chang SM, et al: Serial diffusion-weighted magnetic resonance imaging in cases of glioma: distinguishing tumor recurrence from postresection injury. J Neurosurg 103: , Surov A, Gottschling S, Mawrin C, Prell J, Spielmann RP, Wienke A, et al: Diffusion-weighted imaging in meningioma: prediction of tumor grade and association with histopathological parameters. Transl Oncol 8: , Svolos P, Tsolaki E, Kapsalaki E, Theodorou K, Fountas K, Fezoulidis I, et al: Investigating brain tumor differentiation with diffusion and perfusion metrics at 3T MRI using pattern recognition techniques. Magn Reson Imaging 31: , Tali ET, Yuh WT, Nguyen HD, Feng G, Koci TM, Jinkins JR, et al: Cystic acoustic schwannomas: MR characteristics. AJNR Am J Neuroradiol 14: , Thakur JD, Khan IS, Shorter CD, Sonig A, Gardner GL, Guthikonda B, et al: Do cystic vestibular schwannomas have worse surgical outcomes? Systematic analysis of the literature. Neurosurg Focus 33(3):E12, Tsougos I, Svolos P, Kousi E, Fountas K, Theodorou K, Fezoulidis I, et al: Differentiation of glioblastoma multiforme from metastatic brain tumor using proton magnetic resonance spectroscopy, diffusion and perfusion metrics at 3 T. Cancer Imaging 12: , Wang M, Jia D, Shen J, Zhang J, Li G: Facial nerve function after large cystic vestibular schwannomas surgery via the retrosigmoid approach. Turk Neurosurg 23: , Wippold FJ II, Lubner M, Perrin RJ, Lämmle M, Perry A: Neuropathology for the neuroradiologist: Antoni A and Antoni B tissue patterns. AJNR Am J Neuroradiol 28: , Wu CC, Guo WY, Chen MH, Ho DM, Hung AS, Chung HW: Direct measurement of the signal intensity of diffusionweighted magnetic resonance imaging for preoperative grading and treatment guidance for brain gliomas. J Chin Med Assoc 75: , Yang HC, Kano H, Awan NR, Lunsford LD, Niranjan A, Flickinger JC, et al: Gamma Knife radiosurgery for largervolume vestibular schwannomas. Clinical article. J Neurosurg 114: , Yashar P, Zada G, Harris B, Giannotta SL: Extent of resection and early postoperative outcomes following removal of cystic vestibular schwannomas: surgical experience over a decade and review of the literature. Neurosurg Focus 33(3):E13, Zhu W, Mao Y: Cystic vestibular schwannomas: surgical considerations and outcomes. World Neurosurg 83: , 2015 Disclosures The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author Contributions Conception and design: Guo, CC Wu, HM Wu. Acquisition of data: all authors. Analysis and interpretation of data: Guo, CC Wu, HM Wu, Lin, Lee. Drafting the article: CC Wu. Critically revising the article: Guo, CC Wu. Reviewed submitted version of manuscript: all authors. Statistical analysis: CC Wu. Administrative/technical/material support: Guo, CC Wu, Chung, Lin, Lee. Study supervision: Guo. Supplemental Information Previous Presentations The abstract of this paper was presented at the 18th International Leksell Gamma Knife Society Meeting, May 15 19, 2016, Amsterdam, The Netherlands. Correspondence Wan-Yuo Guo, Department of Radiology, Taipei Veterans General Hospital, 201, Section 2, Shih-Pai Rd., Taipei 112, Taiwan, Republic of China. wyguo@vghtpe.gov.tw. J Neurosurg Volume 127 December