Professional Documents
Culture Documents
RESEARCHHUMANCLINICAL STUDIES
TOPIC
DOI: 10.1227/NEU.0000000000000200
NEUROSURGERY
www.neurosurgery-online.com
SARMIENTO ET AL
Variables
We performed a comprehensive retrospective chart review of medical
records and reviewed brain magnetic resonance images to ascertain the
cystic components within tumors and other key variables. The information collected included age at diagnosis, sex, tumor size, perioperative
Karnofsky Performance Status (KPS), the extent of resection at initial
surgery, and postoperative radiation and chemotherapy. All patients
underwent either surgery or biopsy at the time of diagnosis. Tumor size
was categorized as 1 to 4 cm, 4 to 6 cm, and .6 cm. Perioperative KPS was
identified for each patient on a scale ranging from 0 to 100. The extent of
resection at initial surgery was categorized as biopsy, partial, or near
gross/gross total resection. The outcome of interest in this study was
overall survival, which was defined as the time from initial surgery to the
date of death or last follow-up for those patients with censored date of
death. Date of death was confirmed by the California Cancer Registry.14
Statistical Analysis
Descriptive statistics were reported by subgroups of patients with cystic
and noncystic tumors. Continuous variables were described by using means
and medians, whereas categorical variables were reported by using
frequencies. The Kaplan-Meier estimation method was used to obtain
median survival times and survival probabilities. A multivariate survival
analysis that adjusted for multiple covariates including KPS by using the
Cox proportional hazards model was conducted. In a subanalysis, each cystic
patient was matched with a noncystic control from the 324 patient
noncystic GBM cohorts by using propensity score matching; matching
variables included age in years, extent of resection, and KPS. A P value of less
than .05 was considered to be statistically significant. All statistical analyses
were conducted in SAS 9.2 (SAS Institute, Cary, North Carolina).
RESULTS
Demographics
A total of 351 newly diagnosed GBM patients were identified. Of
these patients, 27 (7.7%) had cystic GBMs and 324 (92.3%) had
noncystic GBMs. The average age in years was 55.4 and 60.6 for
patients with cystic and noncystic tumors, respectively (Table 1).
Overall, 39% of patients were female and the majority of patients
with cystic and noncystic tumors had radiation (85.2% and
82.7%) and temozolomide (81.5% and 75.3%), respectively, as
Cystic
Noncystic
(n = 27, 7.7%) (n = 324, 92.3%) P Value
.16
55.4 (14.6)
54 (42-66)
60.6 (14.2)
61 (52-77)
12 (44.4)
15 (55.6)
125 (38.9)
199 (61.4)
81.1 (12.8)
80 (70-90)
78.1 (14.2)
80 (70-90)
.55
.24
.63
3
15
7
2
(12.0)
(60.0)
(28.0)
(7.4)
44
159
53
68
(17.2)
(62.1)
(28.0)
(21.0)
.32
4 (14.8)
5 (18.5)
18 (66.7)
88 (27.2)
63 (19.4)
173 (53.4)
23 (85.2)
4 (14.8)
268 (82.7)
56 (17.3)
22 (81.5)
5 (18.5)
244 (75.3)
80 (24.7)
1.0
.47
www.neurosurgery-online.com
part of the initial treatment for GBM. Both cohorts had an overall
median KPS of 80, with a range of 70 to 90, and near gross or gross
total resection was achieved in 66.7% of patients with cystic
tumors and 53.4% of patients with noncystic tumors at initial
surgery. Overall, there were no significant differences between the
2 cohorts with respect to age, sex, perioperative KPS, tumor size,
extent of resection, or radiation and temozolomide received
postoperatively (Table 1).
IDH1 Molecular Marker
Tumor samples from all 27 patients with cystic GBM were
analyzed for IDH1 mutations with immunohistochemistry. Two
(7.4%) of the 27 tumor samples were identified as IDH1 positive.
Overall Survival
The median overall survival in months for the cystic cohort was
15.0 (95% confidence interval [CI], 6.1-30.8) and 18.2 (95% CI,
15.6-20.1) for patients with noncystic GBMs (log-rank P = .77)
(Figure). The 12-, 24-, and 36-month survival estimates for the
cystic group were 63.0% (95% CI, 42.1-78.1), 38.6% (95% CI,
20.3-56.7), and 22.1% (95% CI, 7.6-41.3), respectively (Table 2).
The equivalent survival estimates for the noncystic cohort were
66.6% (95% CI, 61-71.5), 36.6% (95% CI, 31.0-42.2), and
22.3% (95% CI, 17.3-27.6). Although the matched-control
survival estimates are not presented in Table 2, our results support
the all-case analysis findings of no statistical difference between the
cystic and noncystic cohorts (P = .35).
Among all patients, younger patients (#65 years) survived an
average of 8.4 months longer than older patients (.65 years).
Higher functioning (KPS $80) patients survived 7.6 months
longer than patients with low functioning levels perioperatively.
Patients that underwent a near/gross total resection compared
with those who underwent a biopsy survived an average of 8.9
months longer (Table 3). Furthermore, patients with methylated O6-methylguanine-DNA methyltransferase (MGMT) status had significantly higher survival than unmethylated patients
(24.8 vs 15.0 months, P = .004).
FIGURE. Kaplan-Meier survival curve for 324 patients with noncystic GBM
and 27 patients with cystic GBM. GBM, glioblastoma multiforme; OS, overall
survival.
DISCUSSION
Glioblastomas (WHO grade IV gliomas) may develop de novo
(primary GBMs) or from malignant progression of WHO grade II
or WHO grade III gliomas (secondary GBMs). Patients with
primary GBMs often have a short clinical history, and the
Cystic (n = 27)
Noncystic (n = 324)
15.0 (6.1-30.8)
18.2 (15.6-20.1)
70.4
63.0
47.2
38.6
22.1
(49.4-84.0)
(42.1-78.1)
(27.6-64.5)
(20.3-56.7)
(7.6-41.3)
NEUROSURGERY
81.3
66.6
50.6
36.6
22.3
All
(76.6-85.1)
(61.0-71.5)
(44.8-56.1)
(31.0-42.2)
(17.3-27.6)
a
b
P Value
.77
15.0 (6.1-30.8)
18.2 (6.1-30.8)
,.001
20.4 (18.8-23.4)
12.0 (9.2-15.6)
,.001
12.6 (9.7-14.9)
20.2 (18.8-23.2)
.56
13.2 (9.5-19.5)
19.0 (16.0-20.5)
,.001
12.0 (7.6-14.5)
15.8 (10.1-20.8)
20.9 (19.0-25.0)
.004b
24.8 (19.5-28.6)
15.0 (12.3-19.0)
SARMIENTO ET AL
a definitive diagnostic molecular marker of secondary glioblastomas and more reliable and objective than clinical and/or
pathological criteria.21 Of our 27 patients with cystic GBM,
only 2 (7.4%) showed IDH1 mutations, thereby suggesting that
most cystic GBMs do not develop from malignant transformation
of previously undiagnosed cystic low-grade glioma.
Prognosis of Patients With Cystic GBMs
A case-control study by Maldaun et al9 first demonstrated
a strong trend toward longer survival and a statistically significantly
longer time to recurrence for GBMs with large cysts ($50% of
tumor volume) compared with noncystic GBMs. The median
overall survival for patients with cystic and noncystic GBMs was
18.7 months and 14.4 months (P = .10 log-rank), respectively, and
the median time to recurrence was 7.6 months and 4.2 months
(P = .04 log-rank), respectively.9 Utsuki et al8 also reported
statistically significant longer survival in patients with cystic vs
noncystic GBMs. Patients with cystic GBMs had a median survival
of 19.8 months, whereas patients with noncystic GBMs had
a median survival of 12.8 months (P , .05 log-rank).8 Although
there are relatively few studies evaluating the prognostic significance of cystic GBMs,3,8,9,22-25 clinical outcomes from the largest
case series do not support the theory of most cystic GBMs
developing from low-grade gliomas and having a better prognosis.3,9 This study by Kaur et al3 comprised 37 patients with cystic
GBM and 317 patients with noncystic GBM, and did not show
a statistically significant difference in survival between the cystic
and noncystic GBM groups (17 vs 15.9 months, respectively, P =
.99). After controlling for known GBM prognostic factors, we did
not observe a statistically significant difference in survival between
cystic and noncystic GBMs (15 vs 18.2 months, log-rank P = .77).
Similar findings were confirmed in the matched-control cohort
analysis of 27 patients.
Limitations
It should be noted that, as with any retrospective study, our study
may be limited by selection bias that may distort the estimate of
association between risk factor and disease. Immunohistochemical
staining may not capture all IDH1 mutations, meaning our cystic
cohort may have had a slightly higher incidence of IDH1-positive
tumors than reported in our study. A small percentage of low-grade
gliomas may also be IDH1 negative, in which case it is possible that
a few cystic tumors in our cohort were a subset of cystic tumors
derived from low-grade, IDH1 negative de novo tumors. Whereas
genome sequencing may have provided a more definitive method
for determining IDH mutation status, this type of investigation
was unable to be performed owing in part to the limited size of
tissue samples available. Additionally, there is a gross discrepancy
between the number of patients in the cystic GBM cohort and the
number of patients in the noncystic GBM cohort because of the
uncommon nature of this disease. Despite these limitations, we
were unable to show any clinical evidence that supports the
hypothesis that cystic GBM develops from low-grade gliomas.
www.neurosurgery-online.com
CONCLUSION
Molecular profiling evidence of IDH1 mutation status indicates
that most cystic GBMs do not develop from malignant transformation of a previously undiagnosed cystic low-grade glioma.
There is no difference in median age or survival between patients
with cystic and those with noncystic newly diagnosed GBM.
Disclosure
The authors have no personal, financial, or institutional interest in any of the
drugs, materials, or devices described in this article.
REFERENCES
1. Afra D, Norman D, Levin VA. Cysts in malignant gliomas. Identification by
computerized tomography. J Neurosurg. 1980;53(6):821-825.
2. Keles GE, Anderson B, Berger MS. The effect of extent of resection on time to
tumor progression and survival in patients with glioblastoma multiforme of the
cerebral hemisphere. Surg Neurol. 1999;52(4):371-379.
3. Kaur G, Bloch O, Jian BJ, et al. A critical evaluation of cystic features in primary
glioblastoma as a prognostic factor for survival. J Neurosurg. 2011;115(4):754-759.
4. Keles GE, Lamborn KR, Berger MS. Low-grade hemispheric gliomas in adults:
a critical review of extent of resection as a factor influencing outcome. J Neurosurg.
2001;95(5):735-745.
5. Laws ER, Taylor WF, Clifton MB, Okazaki H. Neurosurgical management
of low-grade astrocytoma of the cerebral hemispheres. J Neurosurg. 1984;61(4):
665-673.
6. Laws ER Jr, Taylor WF, Bergstralh EJ, Okazaki H, Clifton MB. The neurosurgical
management of low-grade astrocytoma. Clin Neurosurg. 1986;33:575-588.
7. Lohle PN, Verhagen IT, Teelken AW, Blaauw EH, Go KG. The pathogenesis of
cerebral gliomatous cysts. Neurosurgery. 1992;30(2):180-185.
8. Utsuki S, Oka H, Suzuki S, et al. Pathological and clinical features of cystic and
noncystic glioblastomas. Brain Tumor Pathol. 2006;23(1):29-34.
9. Maldaun MV, Suki D, Lang FF, et al. Cystic glioblastoma multiforme: survival
outcomes in 22 cases. J Neurosurg. 2004;100(1):61-67.
10. Parsons DW, Jones S, Zhang X, et al. An integrated genomic analysis of human
glioblastoma multiforme. Science. 2008;321(5897):1807-1812.
NEUROSURGERY
11. Masui K, Cloughesy TF, Mischel PS. Review: molecular pathology in adult highgrade gliomas: from molecular diagnostics to target therapies. Neuropathol Appl
Neurobiol. 2012;38(3):271-291.
12. Yan H, Parsons DW, Jin G, et al. IDH1 and IDH2 mutations in gliomas. N Engl J
Med. 2009;360(8):765-773.
13. Juratli TA, Kirsch M, Robel K, et al. IDH mutations as an early and consistent
marker in low-grade astrocytomas WHO grade II and their consecutive secondary
high-grade gliomas. J Neurooncol. 2012;108(3):403-410.
14. American Cancer Society CDPH, California Cancer Registry. California Cancer
facts and Figures. Oakland, CA: 2011.
15. Barker FG, Chang SM, Huhn SL, et al. Age and the risk of anaplasia in magnetic
resonance-nonenhancing supratentorial cerebral tumors. Cancer. 1997;80(5):
936-941.
16. Dropcho EJ, Soong SJ. The prognostic impact of prior low grade histology in patients
with anaplastic gliomas: a case-control study. Neurology. 1996;47(3):684-690.
17. Karcher S, Steiner HH, Ahmadi R, et al. Different angiogenic phenotypes in
primary and secondary glioblastomas. Int J Cancer. 2006;118(9):2182-2189.
18. Kleihues P, Ohgaki H. Primary and secondary glioblastomas: from concept to
clinical diagnosis. Neuro Oncol. 1999;1(1):44-51.
19. Scherer HJ. Cerebral astrocytomas and their derivatives. Am J Cancer. 1940;40:
159-198.
20. Nobusawa S, Watanabe T, Kleihues P, Ohgaki H. IDH1 mutations as molecular
signature and predictive factor of secondary glioblastomas. Clin Cancer Res. 2009;
15(19):6002-6007.
21. Ohgaki H, Kleihues P. The definition of primary and secondary glioblastoma. Clin
Cancer Res. 2013;19(4):764-772.
22. Choi CY, Yee GT, Lee CH, Joo M. Large cystic glioblastoma multiforme: patient
surviving over 7 years without progression. Clin Neuroradiol. 2013;23(2):145-147.
23. Hakan T, Aker FV. Case report on a patient with neurofibromatosis type 1 and
a frontal cystic glioblastoma. Neurol Neurochir Pol. 2008;42(4):362-365.
24. Kanai R, Tasaka M, Sejima H, et al. Brain stem glioblastoma with multiple large
cyst formation and leptomeningeal dissemination in a 4-year-old girl. Brain Dev.
2005;27(1):58-61.
25. Reiche W, Schuchardt V, Hagen T, Ilyasov KA, Billmann P, Weber J. Differential
diagnosis of intracranial ring enhancing cystic mass lesionsrole of diffusionweighted imaging (DWI) and diffusion-tensor imaging (DTI). Clin Neurol
Neurosurg. 2010;112(3):218-225.
26. Poisson M, Philippon J, van Effenterre R, Racadot J, Sichez JP. Cerebral
pseudocysts following chemotherapy of glioblastomas. Acta Neurochir (Wien).
1977;39(3-4):143-149.
27. Lohle PN, van Mameren H, Zwinderman KH, Teepen HL, Go KG, Wilmink JT.
On the pathogenesis of brain tumour cysts: a volumetric study of tumour, oedema
and cyst. Neuroradiology. 2000;42(9):639-642.
28. Kadota O, Kohno K, Ohue S, et al. Discrimination of brain abscess and cystic
tumor by in vivo proton magnetic resonance spectroscopy. Neurol Med Chir
(Tokyo). 2001;41(3):121-126.
29. Tohma Y, Gratas C, Van Meir EG, et al. Necrogenesis and Fas/APO-1 (CD95)
expression in primary (de novo) and secondary glioblastomas. J Neuropathol Exp
Neurol. 1998;57(3):239-245.
COMMENT
SARMIENTO ET AL
1. Ichimura K, Pearson DM, Kocialkowski S, et al. IDH1 mutations are present in the
majority of common adult gliomas but rare in primary glioblastomas. Neuro Oncol.
2009;11:341-347.
www.neurosurgery-online.com