Clinical Characteristics and Neuroimaging Findings in 12 ... - J-Stage

ORIGINAL ARTICLE Neurol Med Chir (Tokyo) 53, 474¿481, 2013

Clinical Characteristics and Neuroimaging Findings in 12 Cases of Recurrent Glioblastoma With Communicating Hydrocephalus Kuniyuki ONUMA,1 Eiichi ISHIKAWA,1 Masahide MATSUDA,1 Koji HIRATA,1 Satoru OSUKA,1 Tetsuya YAMAMOTO,1 Tomohiko MASUMOTO,2 Alexander ZABORONOK,1 and Akira MATSUMURA1 Departments of 1Neurosurgery and 2Radiology, Faculty of Medicine, University of Tsukuba, Tsukuba, Ibaraki

Abstract Clinically, recurrent glioblastoma multiforme (GBM) is often associated with communicating hydrocephalus. We hypothesized that there are specific magnetic resonance (MR) imaging findings at the diagnosis of recurrent GBM that predict subsequent hydrocephalus. Various clinical characteristics were investigated including outcome and MR imaging findings in 12 patients with recurrent GBM followed by hydrocephalus (Hydro group) and 21 patients with recurrent GBM without hydrocephalus (Non-hydro group). Patient age and presence of communicating hydrocephalus were significantly associated with poor outcome. Median survival with recurrent GBM was longer in the Non-hydro group than in the Hydro group. Low Karnofsky performance status (KPS) and poor recursive partitioning analysis (RPA) class (RPA class 3, 5, 6, or 7) at the diagnosis of recurrent GBM were associated with the presence of hydrocephalus. The incidence of leptomeningeal dissemination after recurrent GBM was higher in the Hydro group than in the Non-hydro group. Evans index and fractional anisotropy value showed no difference at the diagnosis of recurrent GBM, but some MR imaging findings indicated that lesion attached to the basal cistern and/or ventricle was closely associated with subsequent hydrocephalus. We recommend careful monitoring of the ventricle size and leptomeningeal dissemination, especially in patients with low KPS and/or poor RPA class, if MR imaging indicates that the lesion is attached to the basal cistern and/or ventricle at recurrence of GBM. Key words: glioma, hydrocephalus, leptomeningeal dissemination, diffusion tensor magnetic resonance imaging, overall survival

Introduction

sequent hydrocephalus. If so, magnetic resonance (MR) imaging should indicate a specific sign such as ventricular dilation at recurrence, even before the definitive diagnosis of subsequent hydrocephalus. Therefore, we hypothesize that ventricle size and brain compression in patients with hydrocephalus at the diagnosis of recurrent GBM are more prominent than in patients without hydrocephalus, and specific MR imaging findings are present at the detection of recurrent GBM that indicate subsequent hydrocephalus. The present study compared the imaging findings of 12 patients with recurrent GBM followed by hydrocephalus, with those of 21 recurrent GBM patients without hydrocephalus, to identify any specific signs and predictive factors.

Primary glioblastoma multiforme (GBM) frequently relapses after initial treatment. In addition, secondary GBM derived from certain primary World Health Organization (WHO) grade II–III gliomas also relapses with more malignant behavior than the primary tumor. Despite intensive treatment, the median overall survival of a patient with recurrent GBM does not exceed 1 year.3,4,7,9) Clinically, some patients suffer progressive ventricular dilation, mainly caused by communicating hydrocephalus, after the diagnosis of recurrent GBM.6,8,10,13) We and others have speculated that undetectable tumor dissemination at the initial stage is the cause of the subReceived 28, 2013

November 30, 2012;

Accepted

January

474

Recurrent Glioblastoma and Communicating Hydrocephalus

Patients and Methods This retrospective study included 12 patients (4.1% of 295 operative cases of WHO grade II–IV gliomas) with recurrent supratentorial GBM with progressive ventricular dilation after diagnosis (Hydro group). All patients were admitted to Tsukuba University Hospital between 2007 and 2011, and followed up with frequent neuroimaging. Patients had progressive ventricular dilation due to communicating hydrocephalus (10 cases) and combined hydrocephalus (2 cases). The study also included 21 agematched patients with recurrent supratentorial GBM without progressive ventricular dilation admitted to our hospital during the same period of time (Non-hydro group). Twenty-three patients had recurrence of primary GBM and 10 patients had recurrence from secondary GBM. The recursive partitioning analysis (RPA) classification3) defined 2, 1, 7, 4, 3, 6, and 10 cases from class 1 to 7, respectively. Communicating hydrocephalus was defined as obvious ventricular dilation without intraventricular mass or severe obstruction of the ventricle system, with corresponding clinical symptoms such as headache, vomiting, gait disturbance, and/or decreased activities of daily living. In two patients with combined hydrocephalus, ventricular dilation was mainly caused by communicating hydrocephalus but partially by the obstructive component. Evans index was measured as the ratio of the maximum width of the anterior horns to the maximum width of the inner table of the skull on the same axial MR image or computed tomography scan. If the size of the unilateral horn could not be measured because of the presence of the tumor or edema, the index was measured as the ratio of the anterior horn to the midline distance and the midline to the inner table distance on the unaffected side. Modified posterior ventricular hemisphere ratio (mPVHR), which was originally based on echography, was also evaluated on the same axial slice as the Evans index, measured as the ratio of the maximum width of the posterior horns to the maximum width of the inner table or as the ratio of the posterior horn to the midline distance and the midline to the inner table distance on the unaffected side. Age, sex, Karnofsky performance status (KPS) score, RPA class, MR imaging findings of tumor location, size, Evans index, mPVHR, enhancement of the ventricle wall, leptomeningeal dissemination, and other factors, and use of steroids, re-operation at the diagnosis of recurrent GBM, ventricle opening at the initial surgery and at re-operation, clinical diagnosis of hydrocephalus after the diagnosis of recurrent GBM, time to recurrence after the initial

Neurol Med Chir (Tokyo) 53, July, 2013

475

diagnosis, time to the shunt surgery after recurrence, and the outcome were investigated. Additionally, to identify any difference in the intra-ventricular pressure at the diagnosis of recurrent GBM between the Hydro and Non-hydro groups, diffusion tensor imaging (DTI) was performed in 3 patients in the Hydro group and 4 patients in the Non-hydro group, and the fractional anisotropy (FA) value in caudate head, which is known as an indicator of intra-ventricle pressure,12) was analyzed. MR imaging used a 1.5-tesla scanner (Achieva; Philips, Best, the Netherlands). DTI was performed using a multi-section single-shot spin-echo echoplanar imaging sequence (repetition time [TR]/echo time [TE] 10885/60 msec, field of view 230 mm, matrix 128 × 128, slice thickness 3 mm without gap, number of acquisitions 2, b values 0 and 1000 sec/mm2 in 15 different directions). The sequences consisted of spin-echo T1-weighted imaging (TR/TE 678/12.5 msec), fast spin-echo T2-weighted imaging (TR/TE 3000/100 msec), and fluid-attenuated inversion recovery imaging (TR/TE 11000/140 msec, inversion time 2800 msec). DTI analysis was performed as described previously2,12) using the software programs Volume-One and diffusion TENSOR visualizer (Department of Radiology, University of Tokyo, Tokyo). Volumes of interest (VOIs) of uniform size were set anatomically in several locations including the head of the caudate nucleus. FA and mean diffusivity at each point were concurrently calculated using the software. VOIs were selected bilaterally in at least 3 sections in which anatomic regions could be identified. The measured volume at each point was a sphere of approximately 3.6 mm in diameter, which was optimal for each region. The diffusion property of each position was obtained by averaging the data from the 3 VOIs. Statistical analysis used Student's t-test (paired and unpaired) for continuous variables, to determine the differences of mean values between the groups and within each group. To assess the significant difference between the groups, including intermittent variables, we used the Mann-Whitney U (MWU) test (unpaired) or Wilcoxon test (paired). The Fisher direct method was used for categorical data. The logrank test was used to determine whether the differences in survival between groups were significant. Value of p º 0.05 was considered significant in each analysis. SPSS 18 (IBM Corp., Armonk, New York, USA) was used for statistical analysis.

Results Representative cases of the Hydro and Non-hydro

476

K. Onuma et al.

Fig. 1 Neuroradiological findings of a patient with recurrent glioblastoma multiforme (GBM), with subsequent progressive ventricle dilation due to communicating hydrocephalus (Hydro group). A: T2-weighted magnetic resonance (MR) image at diagnosis of recurrent GBM showing a high intensity lesion in the left temporal lobe. B–E: Axial and coronal T1-weighted MR images with gadolinium showing the enhanced lesion attached to the basal cistern and ventricle wall, with slight enhancement in the nearby ventricle wall (arrowheads). F: Computed tomography scan 3 months after the diagnosis showing marked ventricular dilation.

groups are shown in Figs. 1 and 2. To evaluate whether ventricular changes in the Hydro group were different from those in the Non-hydro group, the values of Evans index and mPVHR at the diagnosis of recurrent GBM and the maximum values between 20 days and 99 days after the diagnosis were measured in both groups. The median timing of the maximum values was 74.0 days after the diagnosis in the Hydro group and 56.5 days in the Non-hydro group, with no statistical significance (p ＝ 0.208, MWU test). Figure 3 shows that both Evans index and mPVHR were significantly higher after the diagnosis in the Hydro group (p º 0.05, Wilcoxon test), and mPVHR was significantly higher after the diagnosis in the Hydro group than in the Non-hydro group (p º 0.05, MWU test). Median overall survival (OS) after the diagnosis of recurrent GBM was 12.3 months (95% confidence interval [CI] 7.0–17.6 months). Median OS of patients with recurrent GBM from primary GBM (RPA classes 4–7) was 15.4 months (95% CI 4.0–26.8 months) from the recurrence and 24.2 months (95% CI 0–49.0 months) from the initial diagnosis. Median OS of patients with recurrent GBM from WHO grade II/III gliomas (RPA classes 1–3) was 12.3 months (95% CI 5.1–19.5 months) from the recurrence and 79.2 months (95% CI 9.8–148.6 months)

Fig. 2 Neuroradiological findings of a patient with recurrent glioblastoma multiforme (GBM), without subsequent progressive ventricle dilation (Non-hydro group). A, B: T2-weighted and T1-weighted with gadolinium magnetic resonance (MR) images at diagnosis of recurrent GBM showing no changes near the basal cistern. C, D: Axial T1-weighted MR images with gadolinium showing that the enhanced lesion is located in the frontal lobe. E: Coronal T1-weighted MR image with gadolinium showing that the enhanced lesion partially attached to the ventricle wall. There is no enhancement of the ventricle wall excluding enhancement of the veins. F: T1-weighted MR image with gadolinium 3 months after the diagnosis showing no ventricular dilation.

Fig. 3 Evans index (A) and modified posterior ventricular hemisphere ratio (mPVHR) (B) at the diagnosis of recurrent glioblastoma (Pre) and maximum values between 20 and 99 days after the diagnosis (Post) in 11 cases of the Hydro group and 16 cases of the Non-hydro group. Six cases with insufficient neuroimaging data were excluded. Mann-Whitney U test was used for analysis of unpaired data (Non-hydro versus Hydro) and Wilcoxon test for paired data (Pre versus Post). ns: not significant.

from the initial diagnosis. To identify the factors associated with poor out-



Fig. 4 Kaplan-Meier curves of outcome from diagnosis of recurrent glioblastoma multiforme in the Hydro (black solid line) and Non-hydro (black dash line) groups. Significant difference in the outcome from the point of recurrence between the two groups (p ＝ 0.0253, logrank test).

477

come in patients with recurrent GBM, various parameters were evaluated using the hazards model. Patient age (50 years or older, p ＝ 0.0446, exp 2.85) and presence of communicating hydrocephalus (p ＝ 0.0317, exp 2.79) were statistically associated with poor outcome, whereas all other factors, including KPS (p ＝ 0.8059), PRA class (p ＝ 0.4505), re-operation (p ＝ 0.1747), pathological diagnosis of GBM at initial diagnosis (p ＝ 0.4579), tumor location (p ＝ 0.6246), and the use of steroids (p ＝ 0.2490) were not associated. The Kaplan-Meier curves show significant differences in the survival between the Hydro and Non-hydro groups (Fig. 4). Table 1 shows treatments for 12 cases in the Hydro group and for 21 cases in the Non-hydro group, excluding treatments for hydrocephalus. Ten patients (83%) in the Hydro group and 16 cases (76%) in the Non-hydro group underwent conventional radiotherapy before the diagnosis of recurrent GBM. Temozolomide with or without other special therapies was used for 9 patients (75%) in the Hydro

Table 1 Various treatments for recurrent glioblastoma multiforme (GBM) patients in the Hydro and Non-hydro groups excluding treatments for hydrocephalus Hydro group Primary GBM Number of patients Before the diagnosis of recurrent GBM: Surgical removal (yes/no) Any chemotherapy/immunotherapy (yes/no) TMZ and other special therapies* TMZ only others Any radiotherapy (yes/no) HDT CRT (60.0 Gy of total dose) At/after the diagnosis of recurrent GBM: Surgical removal (yes/no) Any chemotherapy/immunotherapy (yes/no) TMZ and other special therapies* TMZ only others Any radiotherapy (yes/no) SRT (8 cases)*** or GKRT (2 cases) CRT whole spine irradiation

Non-hydro group

Secondary GBM

Primary GBM

Secondary GBM

10

2

13

8

10/0 10/0 3 6 1 10/0 1 9**

2/0 0/2 0 0 2 1/1 0 1

13/0 12/1 4 6 2 13/0 4 9

8/0 4/4 0 2 2 7/1 0 7

7/3 8/2 2 3 3 4/6 3 0 1

2/0 1/1 1 0 0 1/1 0 1 0

11/2 12/1 5 5 2 5/8 5 0 0

8/0 8/0 5 3 0 3/5 2 1 0

*Special therapies consist of autologous formalin-fixed tumor vaccine, interferon beta, and/or humanized anti-vascular endothelial growth factor-A monoclonal antibody (bevacizumab) treatments with various schedules, as the combination of temozolomide (TMZ) therapy or as the second/third line therapies after completion (suspension) of TMZ therapy. **Total dose is 64.8 Gy only for 1 case. ***Planning target volume for fractionated stereotactic radiotherapy (SRT) is defined as an enhanced lesion on T1-weighted imaging after gadolinium and a circumferential margin of 0–5 mm. Total dose was 39 Gy given in 13 fractions of 3.0 Gy for 5 cases, and 24, 30, and 40 Gy for the other cases. CRT: conventional radiotherapy, GKRT: gamma knife radiation treatment, HDT: high-dose particle radiotherapies consisting of proton beam radiotherapy (96.6 Gy of total dose) and boron neutron capture therapy.


478 Table 2

K. Onuma et al. Clinical characteristics and outcome in the Hydro and Non-hydro groups Hydro Non-hydro p Value group group

Number of patients At initial diagnosis of glioma: Histological diagnosis of grade IV GBM (yes/no) Histological diagnosis of grade II GBM (yes/no) Ventricle opening during surgery (yes/n.d./no) At the diagnosis of recurrent GBM: Age (years) Median KPS (%) RPA class (1, 2, or 4/others) Steroid use (yes/no) Median number of operation procedures before the diagnosis Tumor location (frontal/others) Mean tumor size (cm) MR imaging findings A: Distant leptomeningeal dissemination suspected before or at the diagnosis of recurrent GBM on MR imaging (yes ＝ 1/no ＝ 0) B: Enhanced lesion attached to the basal cistern (yes ＝ 1/no ＝ 0) C: Enhanced lesion attached to the ventricle wall or invading the ventricle space (yes ＝ 1/no ＝ 0) D: Enhancement of the ventricle wall (yes ＝ 1/no ＝ 0) A ＋ B ＋ C ＋ D: MR imaging score (mean; median) Ventricle opening during surgery for recurrent GBM (yes/no or not performed) Median time to the recurrence of GBM from initial diagnosis (months) After the diagnosis of recurrent GBM: Leptomeningeal dissemination after recurrent GBM on MR imaging (yes/no) Median survival from recurrent GBM diagnosis (months) Median survival from shunt surgery (months) Median survival from initial diagnosis (months)

12

21

10/2 1/11 6/1/5

13/8 5/16 4/8/9

0.259* 0.379* 0.348**

55.9 47.5 0/12 4/8 1 4/8 50.2

52.2 80 7/14 13/8 1 13/8 45.4

0.506*** 0.028** 0.028** 0.157* 0.195** 0.157* 0.681**

3/9

0/21

0.125*

5/7 11/1 7/5 1.9; 2 4/8 8.0

1/20 12/9 2/19 0.7; 1 11/10 10.0

0.016* 0.055* 0.005* 0.001** 0.469* 0.538#

10/2 6.6 2.7 17.0

7/14 16.3 — 75.3

0.010* 0.025# — 0.122#

*Fisher direct method; **Mann-Whitney U test; ***Student t test; #logrank test. GBM: glioblastoma multiforme, KPS: Karnofsky performance status, MR: magnetic resonance, n.d.: not done, RPA: recursive partitioning analysis.

group and 12 patients (57%) in the Non-hydro group at the time. After recurrence, 5 patients (42%) in the Hydro group and 8 patients (38%) in the Non-hydro underwent radiotherapies. There was no statistical difference between any treatment data, including type of radiotherapy, in 2 groups (p À 0.05, Fisher direct method). Table 2 compares the patient characteristics and outcomes in the Hydro and Non-hydro groups, excluding the treatment data described in Table 1. There was no difference between the two groups in pathological etiology, number of operation procedures before the diagnosis, tumor region, or time to recurrence after the initial diagnosis. Lower KPS and poor RPA class (class 3, 5, 6, or 7) at the diagnosis of recurrent GBM were associated with the presence of subsequent hydrocephalus. Two MR imaging findings (B, enhanced lesion attached to the basal cistern; D, enhancement of ventricle wall) were also associated with the presence of hydrocephalus, and two other findings (A, distant leptomeningeal dissemination suspected before or at the diagnosis of recurrent GBM; C, enhanced lesion attached to the ventricle wall or invading the ventricle space) also tended to be associated with the

presence of hydrocephalus. Including all 4 factors (A–D), the total score was closely related with the presence of hydrocephalus (p ＝ 0.001, MWU test). The incidence of leptomeningeal dissemination (multiple leptomeningeal enhancement clearly detected by MR imaging) after recurrent GBM was higher in the Hydro group than in the Non-hydro group (p ＝ 0.0104, Fisher direct method). Nine of 12 patients in the Hydro group underwent shunt surgery. Only one patient underwent spinal drainage, and the other died before treatment for cerebrospinal fluid circulation disturbance. In the remaining 7 patients, mean cell count was 52.5 ± 128.0 (range 1–368) cells/mm3, mean protein level was 493.0 ± 1050.4 (range 40–3084) mg/dl, and mean cerebrospinal fluid pressure was 20.0 ± 7.7 (range 15–35) cmH2O. The disseminating lesions were detected after the diagnosis of hydrocephalus in only 2 cases, and were identified before or at the time of diagnosis of hydrocephalus in the other cases. Although there was no difference in KPS before and after the shunt surgery (median KPS of 30 versus 40, p ＝ 0.1936, Wilcoxon test), hydrocephalus-associated symptoms including unconsciousness, headache, and nausea clearly improved in 6 of 10 patients.



479

º Fig. 5 Median fractional anisotropy (FA) values of the gray matter (caudate nucleus, thalamus, and putamen) and the white matter (genu and splenium of the corpus callosum [CC], posterior peduncle of the internal capsule [Post IC], and corona radiata [CR]) at the diagnosis of recurrent glioblastoma multiforme (GBM). To estimate the intra-ventricle pressure at diagnosis of recurrent GBM in the Hydro and Non-hydro groups, diffusion tensor imaging was performed in 3 cases (2 volumes of interest [VOIs] per each case) in the Hydro group and in 4 cases (2 VOIs per each case) in the Nonhydro group. Some VOIs directly affected by GBM were excluded from the analysis. There is no statistical difference between median FA values of each VOI between the 2 groups (p À 0.05, Mann-Whitney U test). ns: not significant.

Median survival with recurrent GBM was longer in the Non-hydro group than in the Hydro group (16.3 versus 6.6 months, p ＝ 0.0253, logrank test). Median survival of 9 patients with shunt surgery and 3 patients without shunt surgery in the Hydro group was 10.4 and 4.3 months, but without statistically significant difference (p ＝ 0.2408, logrank test). To estimate the intra-ventricle pressure at the diagnosis of recurrent GBM in the Hydro and Nonhydro groups, median FA values of the gray matter including the caudate nucleus at the diagnosis of recurrent GBM were evaluated in 3 patients in the Hydro group and 4 patients in the Non-hydro group, but no difference was found (p À 0.05, MWU test) (Fig. 5).

Discussion The frequency of hydrocephalus in this study was 4.1% (12 cases), including 10 cases of pure communicating hydrocephalus, which was compatible with the findings of other studies. In previous series, hydrocephalus or ventricle dilation developed in 3.4–26% of patients with primary high-grade gliomas or GBM.8,10,13) In one study, all 10 patients underwent multiple craniotomies, and the frequency of hydrocephalus in patients with multiple craniotomies was 22%.10) In our experience, communicating hydrocephalus usually developed after the diagnosis of recurrent GBM, but was very rarely associated with primary GBM or WHO grade II–III gliomas at


the recurrent stage (detailed data are not shown). Therefore, this study evaluated hydrocephalus only at the recurrent stage of GBM. As limitations of this study, there were many biases in inclusion criteria. For example, young or middle-age patients with recurrent GBM tended to be admitted to our hospital for treatment including surgery. Accordingly, elderly patients were excluded from this study, and the median OS of patients with recurrence from primary GBM was 15.4 months from the recurrence and 24.2 months from the initial diagnosis. Such long survival is probably due to adequate local control after aggressive removal and irradiation of the tumor, but might be associated with leptomeningeal dissemination and hydrocephalus after the late stage. Hydrocephalus after tumor recurrence as well as patient age was clearly associated with poor outcome in this study. Survival time from the diagnosis of recurrent GBM was statistically shorter in the Hydro group than in the Non-hydro group. Ventricle opening (after radio-chemotherapy) is a significant predictor of communicating hydrocephalus in patients with high-grade gliomas.8,10) On the other hand, any association between ventricle opening and leptomeningeal dissemination is controversial.1,5) In our study, ventricle opening at the initial surgery or surgery for recurrent GBM did not predict hydrocephalus, although ventricle opening tended to be more frequent in the Hydro group than in the Non-hydro group. On the other hand, communicating hydrocephalus was associated with low

K. Onuma et al.

480

KPS, poor RPA class, and MR imaging findings at the diagnosis of recurrent GBM indicating attachment of the lesion to the ventricle wall and/or basal cistern. We speculate that these factors were related to undetectable or potential cerebrospinal fluid dissemination of tumor cells. In our study, the incidence of hydrocephalus was closely associated with the occurrence of leptomeningeal dissemination before (or after) the diagnosis of hydrocephalus, as pointed out previously.6) Therefore, we hypothesized that ventricular dilation is present before or at the diagnosis of recurrent GBM due to dissemination undetectable by MR imaging. However, both Evans index and mPVHR were similar in the Hydro and Non-hydro groups at the diagnosis of recurrent GBM in this study, which indicates that these indexes cannot predict subsequent hydrocephalus. Moreover, FA values of the gray matter, which is increased in the hydrocephalic state or high intracranial pressure,11,12) showed no difference between the Hydro and Non-hydro groups at the time of recurrence. This finding indicates that FA value at the diagnosis of recurrent GBM could not predict subsequent hydrocephalus, probably because the hydrocephalic state is not present at the time of recurrence but occurs thereafter. On the other hand, MR imaging findings that indicate lesion attachment to the basal cistern and/or ventricle, for example, enhanced lesion attached to the basal cistern and enhancement of the nearby ventricle wall, were closely associated with subsequent hydrocephalus. In particular, we think that obscure (or obvious) enhancement of the nearby ventricle wall may be a key indicator of subsequent hydrocephalus. In the present study, enhancement of the nearby ventricle wall on MR imaging at the diagnosis of recurrent GBM was detected in 7 of 12 cases in the Hydro group, including retrospective analysis. These findings support our hypothesis that the potential for hydrocephalus is present at the diagnosis of recurrent GBM, although marked ventricular dilation only occurs after the diagnosis. The present study suggests that some MR imaging findings at the diagnosis of recurrent GBM may predict subsequent hydrocephalus. We should carefully monitor ventricle size and dissemination, especially in patients with low KPS and/or poor RPA class, if MR imaging indicates lesion attachment to the basal cistern and/or ventricle at the diagnosis of recurrent GBM.

Acknowledgments The authors would like to express their gratitude to Drs. Yosuke Masuda, Kei Nakai, and Shingo Takano

of the Department of Neurosurgery, Graduate School of Comprehensive Human Sciences, University of Tsukuba, for the treatment of patients included in this study.

Conflicts of Interest Disclosure No conflict of interest exists. All authors of The Japan Neurosurgical Society (JNS) members have registered Self-reported COI Disclosure Statement Forms through the website for JNS members.

References 1)

Bae JS, Yang SH, Yoon WS, Kang SG, Hong YK, Jeun SS: The clinical features of spinal leptomeningeal dissemination from malignant gliomas. J Korean Neurosurg Soc 49: 334–338, 2011 2) Basser PJ, Pierpaoli C: Microstructural and physiological features of tissues elucidated by quantitativediffusion-tensor MRI. J Magn Reson B 111: 209–219, 1996 3) Carson KA, Grossman SA, Fisher JD, Shaw EG: Prognostic factors for survival in adult patients with recurrent glioma enrolled onto the new approaches to brain tumor therapy CNS consortium phase I and II clinical trials. J Clin Oncol 25: 2601–2606, 2007 4) Desjardins A, Reardon DA, Coan A, Marcello J, Herndon JE 2nd, Bailey L, Peters KB, Friedman HS, Vredenburgh JJ: Bevacizumab and daily temozolomide for recurrent glioblastoma. Cancer 118: 1302–1312, 2012 5) Elliott JP, Keles GE, Waite M, Temkin N, Berger MS: Ventricular entry during resection of malignant gliomas: effect on intracranial cerebrospinal fluid tumor dissemination. J Neurosurg 80: 834–839, 1994 6) Inamasu J, Nakamura Y, Saito R, Kuroshima Y, Mayanagi K, Orii M, Ichikizaki K: Postoperative communicating hydrocephalus in patients with supratentorial malignant glioma. Clin Neurol Neurosurg 106: 9–15, 2003 7) Ishikawa E, Tsuboi K, Yamamoto T, Muroi A, Takano S, Enomoto T, Matsumura A, Ohno T: Clinical trial of autologous formalin-fixed tumor vaccine for glioblastoma multiforme patients. Cancer Sci 98: 1226–1233, 2007 8) Marquardt G, Setzer M, Lang J, Seifert V: Delayed hydrocephalus after resection of supratentorial malignant gliomas. Acta Neurochir (Wien) 144: 227–231, 2002 9) Miyatake S, Kawabata S, Yokoyama K, Kuroiwa T, Michiue H, Sakurai Y, Kumada H, Suzuki M, Maruhashi A, Kirihata M, Ono K: Survival benefit of boron neutron capture therapy for recurrent malignant gliomas. Appl Radiat Isot 67(7–8 Suppl): S22–24, 2009 10) Montano N, D'Alessandris QG, Bianchi F, Lauretti L, Doglietto F, Fernandez E, Maira G, Pallini R: Communicating hydrocephalus following surgery and ad-


Recurrent Glioblastoma and Communicating Hydrocephalus juvant radiochemotherapy for glioblastoma. J Neurosurg 115: 1126–1130, 2011 11) Osuka S, Matsushita A, Ishikawa E, Saotome K, Yamamoto T, Marushima A, Satou N, Zaboronok A, Masumoto T, Matsumura A: Elevated diffusion anisotropy in gray matter and the degree of brain compression. J Neurosurg 117: 363–371, 2012 12) Osuka S, Matsushita A, Yamamoto T, Saotome K, Isobe T, Nagatomo Y, Masumoto T, Komatsu Y, Ishikawa E, Matsumura A: Evaluation of ventriculomegaly using diffusion tensor imaging: correlations with chronic hydrocephalus and atrophy. J Neurosurg 112: 832–839, 2010 13) Smith KA, Ashby LS, Gonzalez LF, Brachman DG, Thomas T, Coons SW, Battaglia M, Scheck A: Prospective trial of gross-total resection with Gliadel


481

wafers followed by early postoperative Gamma Knife radiosurgery and conformal fractionated radiotherapy as the initial treatment for patients with radiographically suspected, newly diagnosed glioblastoma multiforme. J Neurosurg 109 Suppl: 106–117, 2008

Address reprint requests to: Eiichi Ishikawa, MD, PhD, Department of Neurosurgery, Graduate School of Comprehensive Human Sciences, University of Tsukuba, 1–1–1 Tennodai, Tsukuba, Ibaraki 305–8575, Japan. e-mail: e-ishikawa＠md.tsukuba.ac.jp