High-Grade Glioma is not a Feature of Neurofibromatosis Type 2 in the Unirradiated Patient

Abstract

Background:
The Manchester criteria for neurofibromatosis type 2 (NF2) include a range of tumors, and gliomas were incorporated in the original description. The gliomas are now widely accepted to be predominantly spinal cord ependymomas.

Objective:
To determine whether these gliomas include any cases of malignant glioma (WHO grade III and IV) through a database review.

Methods:
The prospective database consists of 1253 patients with NF2. 1009 are known to be alive at last follow-up.

Results:
There was a single case of glioblastoma multiforme (GBM; World Health Organization grade IV) in the series and no WHO grade III gliomas. The GBM was in a patient who had previously undergone stereotactic radiosurgery for a vestibular schwannoma.

Conclusion:
High-grade gliomas are not a feature of NF2 in the unirradiated patient and should be excluded from the diagnostic criteria.

Full text

RESEARCH—HUMAN—CLINICAL STUDIES
High-Grade Glioma is not a Feature of
Neurofibromatosis Type 2 in the Unirradiated
Patient
Andrew T. King, FRCS∗
Scott A. Rutherford, FRCS∗
Charlotte Hammerbeck-Ward, PhD,
FRCS∗
Simon K. Lloyd, MPhil, FRCS‡
Simon M. Freeman, MPhil, FRCS‡
Omar N. Pathmanaban, PhD, FRCS∗
Monica Rodriguez-Valero, MD‡
Owen M. Thomas, DPhil, FRCR§
Roger D. Laitt, FRCP, FRCR§
Stavros Stivaros, PhD, FRCR¶
Mark Kellett, MD, FRCP||
D. Gareth Evans, MD, FRCP#
∗Department of Neurosurgery, Manche-
ster Academic Health Science Centre, Sal-
ford Royal NHS Foundation Trust, Man-
chester, United Kingdom; ‡Department
of Otolaryngology, Manchester Academic
Health Science Centre, Salford Royal
NHS Foundation Trust, Manchester,
United Kingdom; §Department of
Neuroradiology, Manchester Academic
Health Science Centre, Salford Royal
NHS Foundation Trust, Manchester,
United Kingdom; ¶Department
of Neuroradiology, Manchester
Academic Health Science Centre,
Central Manchester NHS Foundation
Trust, Manchester, United Kingdom;
||Department of Neurology, Manchester
Academic Health Science Centre, Salford
Royal NHS Foundation Trust, Manchester,
United Kingdom; #Department of
Manchester Centre for Genomic
Medicine, Manchester Academic Health
Science Centre, Central Manchester NHS
Foundation Trust, Manchester, United
Kingdom
Correspondence:
Andrew T. King, FRCS,
Department of Neurosurgery,
Manchester Academic Health Science
Centre,
Salford Royal NHS Foundation Trust,
Manchester M6 8HD.
E-mail: andrew.king@manchester.ac.uk
Received, May 1, 2017.
Accepted, June 25, 2017.
C
Crown copyright 2017. This article
contains public sector information
licensed under the Open Government
Licence v3.0
(http://www.nationalarchives.
gov.uk/doc/open-government-
licence/version/3/).
BACKGROUND: The Manchester criteria for neurobromatosis type 2 (NF2) include a range
of tumors, and gliomas were incorporated in the original description. The gliomas arenow
widely accepted to be predominantly spinal cord ependymomas.
OBJECTIVE: To determine whether these gliomas include any cases of malignant glioma
(WHO grade III and IV) through a database review.
METHODS: The prospective database consists of 1253 patients with NF2. 1009 are known
to be alive at last follow-up.
RESULTS: Therewasasinglecaseofglioblastomamultiforme(GBM;WorldHealthOrgani-
zation grade IV) in the series and no WHO grade III gliomas. The GBM was in a patient who
had previously undergone stereotactic radiosurgery for a vestibular schwannoma.
CONCLUSION: High-grade gliomas are not a feature of NF2 in the unirradiated patient and
should be excluded from the diagnostic criteria.
KEY WORDS: Neurobromatosis Type 2, Manchester criteria, Glioblastoma, High-grade glioma, Malignant
glioma, Radiotherapy, Radiosurgery
Neurosurgery 0:1–4, 2017 DOI:10.1093/neuros/nyx374 www.neurosurgery-online.com
Neurofibromatosis type 2 (NF2) is a
genetic condition typified by bilateral
vestibular schwannomas, but also
causing a range of other tumors both in the
central and peripheral nervous system. These
tumors include other cranial, spinal, peripheral
nerve and intradermal schwannomas, cranial and
spinal meningiomas, and intrinsic tumors, which
are predominantly spinal ependymomas. These
are summarized by the Manchester criteria.1In
the original criteria developed from National
Institutes of Health (NIH) consensus from 1987,
the intrinsic tumors were described generically
as gliomas rather than as ependymomas specif-
ically. This reflected a radiological opinion on
tumors from the 1980s rather than from histo-
logical diagnosis. We have been concerned that
this imprecise nomenclature has subsequently
resulted in extrapolation of the glioma criterion
to include malignant glioma,2apresumption
ABBREVIATIONS: GBM, glioblastoma multiforme;
NF1, neurobromatosis type 1; NF2, neurobro-
matosis type 2, NIH, National Institutes of Health;
WHO, World Health Organization
that does not marry with our own clinical
impression.Rather, it was our impression, both
from our own NF2 cohort and our review of the
literature that malignant glioma only appears
in the context of NF2 after either radiotherapy,
including stereotactic radiosurgery of vestibular
schwannomas. The confusion in the literature
over whether high-grade gliomas are a feature of
NF2 and as to whether they should be part of
NF2 diagnostic criteria, or, alternatively, recog-
nized as a potential risk of irradiation, has lead
us to re-examine this issue. We have interrogated
the Manchester NF2 database of 1253 NF2
patients for the presence of histologically proven
high-grade gliomas and reviewed the literature.
METHODS
A review was undertaken of the NF2 database
available to the senior author (DGE), which has
ethical approval from the Central Manchester Ethics
Committee. Patient consent was not necessary because
there is no patient identifiable information. The
purpose was to describe the nature of the intracranial
intra-axial tumors. The NF2 database consists of 1253
patients. This consists of families who fulfill at least
NEUROSURGERY VOLUME 0 | NUMBER 0 | 2017 | 1
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/doi/10.1093/neuros/nyx374/3988116/High-Grade-Glioma-is-not-a-Feature-of
by guest
on 14 October 2017

KING ET AL
Manchester criteria for NF2 or who have a proven pathogenic mutation
in the neurofibromatosis type 2 (NF2) gene. Patients outside Manchester
had clinical information and DNA samples sent by geneticists, neurolo-
gists, neurosurgeons, and ear, nose, and throat surgeons from throughout
the UK and some other European countries, notably Norway, Finland,
and Denmark. Four hundred forty-eight patients have been seen by the
senior author.
RESULTS
Five hundred ninety patients were male and 663 female. Of
the 1253 NF2 patients in the database, the mean age at diagnosis
was 32.46 yr (range 0.2-82). Two hundred forty-four patients had
died, with a median age of 39 yr (range 14.4-84.9 yr). There were
1009 patients known to be alive at last follow-up. The median
age of this group was 41.3 yr (range 1.0-90.5 yr). Eight hundred
twelve of 1168 (69.5%) had known mutations. Six hundred forty-
six were germline (constitutional) mutations with 176 mosaics
whilst 346 had no identified mutation from blood DNA. Eighty-
five patients had not undergone mutation testing.
There were 11 123.1 yr of follow-up from diagnosis in 1148
patients with information on date of diagnosis giving a mean
follow-up of 8.92 yr per patient. Follow-up from birth to last
follow-up was 51 525.2 yr.
In the whole group, there was 1 case of malignant grade
IV glioma (glioblastoma multiforme [GBM]) reported. This
single case occurred after stereotactic radiosurgery for a vestibular
schwannoma.
There were 3 pathologically proven low-grade intracranial
intrinsic tumors, 2 optic nerve pilocytic astrocytomas and a gliofi-
broma. Recognizing that pilocytic astrocytomas are a defining
feature of neurofibromatosis type 1 (NF1), the 2 patients with
pilocytic astrocytomas had no features of NF1 and met all the
inclusion criteria for NF2. There was 1 spinal pilocytic astro-
cytoma that was removed in childhood. There were no cases of
malignant gliomas. One hundred forty-eight of 889 (16.6%) with
data on spinal imaging had radiological evidence of a uni- or
multifocal intrinsic spinal cord or, much less frequently, lower
brainstem lesion consistent with an ependymoma. No patient had
an isolated intrinsic lesion of the brain stem alone. Six of the spinal
tumors have been confirmed as grade 1 ependymoma at surgery.
Of the 10 cases of glioma published by the senior author in
19921none were histologically confirmed as high-grade glioma
and none of these patients subsequently died of their intrinsic
brain/spinal cord tumor, effectively ruling out a high-grade
glioma (4 are still living), although none had undergone a surgical
excision or biopsy.
DISCUSSION
In the 1253 total cases in the database, there was only 1 case
of malignant glioma. This occurred 3 yr after stereotactic radio-
surgery, as reported in the Sheffield series.2Outside of this single
case, the striking feature of this large series is both the rarity of
intracranial intra-axial lesions and the total absence of malignant
glioma in the nonirradiated brain or spine. Only 2 such intra-
axial brain tumors were found, both were benign. By comparison,
intra-axial spinal cord tumors are relatively common, occurring in
up to 31% of NF2 series3(16.6% this series).It is now well recog-
nized that these spinal cord tumors are ependymomas.4They
are, therefore, within the World Health Organization classifi-
cation of gliomas,5and therefore the original NIH criteria and the
subsequent Manchester modification were accurate, if somewhat
imprecise and nonspecific. It would appear that subsequently it
has been assumed that NF2 can be associated with any glial tumor
and therefore potentially a malignant glioma.
The case in our series is described by Rowe et al2in their series
of outcomes from radiosurgery and is of a woman treated for a
1.8 cm3vestibular schwannoma in 2000 with a marginal dose of
14 Gy. She developed a malignant glioma within 3 yr of treatment
and died within 6 mo of diagnosis. Details of where the lesion
arose in relation to the treated schwannoma are not published,
to our knowledge. It was estimated from the treatment plan that
24 cm3of the brain received more than 2 Gy and 54 cm3of
the brain received 1 to 2 Gy. In the absence of details as to the
relationship between the laterality of the treated schwannoma
and the site of the malignant glioma, it is not possible to say
with any confidence whether this lesion meets the Cahan criteria6
for radiation-induced neoplasia. With regard to these criteria,
it is clearly histologically distinct from the original vestibular
schwannoma, and we believe we have shown in this paper that
NF2 does not predispose to malignancy. The latency is no more
than 3 yr, relatively brief by Cahan criteria.6It is also the case that
glioblastoma has been reported after stereotactic radiosurgery to a
vestibular schwannoma in 2 cases, neither of them NF2.7,8In the
non-NF2 situation, where the numbers of irradiated vestibular
schwannomas throughout the world is much higher, the possi-
bility of coincidental development of a GBM is plausible, but
the presence of 1 case of GBM in NF2 only after radiotherapy
and none in unirradiated cases still requires explanation. Rowe
et al2argue that given that 4% of NF2 patients develop gliomas,
it is not clear that there is an increased risk in their material.
This figure of 4% relates to the paper of 1992,1,2subsequently
known as defining the “Manchester criteria” for NF2 and widely
accepted since (Table 1). This paper describes astrocytomas in
4.1% of cases and ependymomas in 2.5% of cases. It specifically
states that “astrocytomas and ependymomas were low grade and
affected the lower brain stem and upper cervical cord, although
one astrocytoma had spread to involve the entire medulla of the
spine.” We now know too that these lesions are almost exclusively
ependymomas4,9not astrocytomas and can occur throughout the
spinal cord from the medulla of the brain stem to the conus of
the spinal cord. We have shown from the data we present here
that malignant glioma is, with the exception of the case described
by Rowe et al,2not a feature of NF2. Therefore, more than
20 yr after the description of the criteria for NF2, there is a need
to clarify that the term “glioma” in the Manchester criteria does
not refer to malignant glioma.
2|VOLUME0|NUMBER0|2017 www.neurosurgery-online.com
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/doi/10.1093/neuros/nyx374/3988116/High-Grade-Glioma-is-not-a-Feature-of
by guest
on 14 October 2017

NF2 AND GBM
TAB LE 1 . Manchester Diagnostic Criteria for NF2 with Proposed Change
Bilateral vestibular schwannomas OR family history of NF2 PLUS
1. Unilateral VS OR
2. Any 2 of: meningioma, glioma, neurobroma, schwannoma, posterior subcapsular lenticular opacities
Additional criteria
Unilateral VS +any two of:ameningioma, glioma (low-grade glial tumor usually ependymoma)b, neurobroma, schwannoma, and posterior
subcapsular opacities
OR
Multiple meningioma (2 or more) +unilateral VS or any 2 of:aglioma (low-grade glial tumour usually ependymoma)b, neurobroma,
schwannoma, and cataract.
These incorporate the original NIH criteria with additional criteria.
aRefers to any 2 tumors such as 2 schwannomas or schwannoma plus meningioma.
bSuggested change to diagnostic criteria.
TAB LE 2 . Reports of Testing for NF2 Mutations in Series of Glioma and Ependymoma
Study Number of gliomas NF2 mutation Chr 22 loss
Hoang–Xuan et al 199512 70 0/70 17/70
Rubio et al 199411 30 0/30 –
Watkins et al 199613 47 0/47 10/47
Alonso et al 200215 40 oligodendrogliomas 0/40
De Vitis 199614 30 0/30
Total for gliomas 0/217 27/117 (23%)
Study Number of ependymomas NF2 mutation Chr 22 loss
Birch et al 199618 75/7
Lamszus et al 2001 21 6/21 12/21
Ebert et al 199916 54 6/62 12/54
6/23 spinal
Alonso et al 200215 71/7
Total for ependymomas 18/97 (19%) 24/75 (32%)
The annual incidence rate for astrocytic tumors in the UK is 5
per 100 000 annually, with 63% being high grade.10 As such, in
11 000 yr of follow-up, one might have seen 1 high-grade glioma
in our NF2 series by chance (expected 0.35), yet none occurred in
nonirradiated patients. From birth, one might have expected 1.62
high-grade gliomas in 51 512 yr of follow-up, yet the only one to
occur was postirradiation. The 4 low-grade (nonependymoma)
gliomas would be in excess of the 24% of 5/100 000 rate with an
expected incidence of 0.13 indicating a relative risk of 23 fold.
The fact that we have failed to find a single case of malignant
glioma outside of an irradiated case, as well as the acceptance that
the spinal cord lesions are ependymomas, would be in keeping
with genetic studies of sporadic gliomas and ependymomas in the
literature (Table 2). Although we did not have information on
spinal imaging in 364/1253 (29%) of cases, all had undergone
cranial magnetic resonance imaging.
Rubio et al11 screened 8 ependymomas and 30 fibrillary astro-
cytomas from non-NF2 cases for the NF2 gene. They found
no mutation in astrocytomas but did find a mutation in spinal
ependymoma. They concluded that NF2 gene mutations are
not important for astrocytoma and oligodendroglioma tumorige-
nesis. The absence of NF2 mutations in glioma is confirmed by
other studies.12,13 ,14 In total, these 4 papers studied 177 gliomas
and failed to find a single NF2 mutation. The same absence
of an NF2 mutation was found in a study of another intrinsic
tumor not found in NF2 (oligodendroglioma).15 In contrast,
NF2 mutations are found in some, but by no means all, sporadic
spinal ependymomas12,15 -17 -18 out 58 (31%) spinal ependy-
momas.
CONCLUSION
In this large series of patients with NF2, we found no evidence
that malignant glioma is a feature of NF2. The description in
the literature of gliomas being present in NF2 relates to low-
grade spinal cord and brain stem lesions rather than supratentorial
tumors such as malignant glioma. The NF2 criteria should now
be clarified to reflect the lack of high-grade glioma link.
NEUROSURGERY VOLUME 0 | NUMBER 0 | 2017 | 3
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/doi/10.1093/neuros/nyx374/3988116/High-Grade-Glioma-is-not-a-Feature-of
by guest
on 14 October 2017

KING ET AL
Disclosures
The author wish to thanks NHS England for their support of the National
NF2 program. The authors have no personal, financial, or institutional interest in
any of the drugs, materials, or devices described in this article.
REFERENCES
1. Evans DG, Huson SM, Donnai D, et al. A clinical study of type 2 neurofibro-
matosis. QJMed. 1992;84(304):603-618.
2. Rowe J, Grainger A, Walton L, Silcocks P, Radatz M, Kemeny A. Risk of malig-
nancy after gamma knife stereotactic radiosurgery. Neurosurgery. 2007;60(1):60-
65; discussion 65-66.
3. Mautner VF, Lindenau M, Baser ME, et al. The neuroimaging and clinical
spectrum of neurofibromatosis 2. Neurosurgery. 1996;38(5):880-885; discussion
885-886.
4. Hagel C, Stemmer-Rachamimov AO, Bornemann A, et al. Clinical presentation,
immunohistochemistry and electron microscopy indicate neurofibromatosis type
2-associated gliomas to be spinal ependymomas. Neuropatho logy. 2012;32(6):
611-616.
5. Louis DN, Cavenee WK, Ohgaki H, Wiestler OD, Cancer IAfRo . WHO Classi-
fication of Tumours of the Central Nervous System. International Agency for Research
on Cancer. Acta Neuropathol. 2007;114(2):97-109. Erratum in: Acta Neuropathol.
2007;114(5):547.
6. Cahan WG, Woodward HQ. Sarcoma arising in irradiated bone, report of 11
cases. Cancer. 1948;1(1):3-29.
7. Shamisa A, Bance M, Nag S, et al. Glioblastoma multiforme occurring in a
patient treated with gamma knife surgery: case report and review of the literature.
JNeurosurg. 2013;119(suppl):816-821.
8. Balasubramaniam A, Shannon P, Hodaie M, Laperriere N, Michaels H,
Guha A. Glioblastoma multiforme after stereotactic radiotherapy for acoustic
neuroma: case report and review of the literature. Neuro Oncol. 2007;9(4):
447-453.
9. Louis DN, Ramesh V, Gusella JF. Neuropathology and molecular
genetics of neurofibromatosis 2 and related tumors. Brain Pathol. 1995;5(2):
163-172.
10. Crocetti E, Trama A, Stiller C, et al. Epidemiology of glial and non-glial brain
tumours in Europe. Eur J Cancer. 2012;48(10):1532-1542.
11. Rubio MP, Correa KM, Ramesh V, et al. Analysis of the neurofibromatosis
2 gene in human ependymomas and astrocytomas. Cancer Res. 1994;54(1):
45-47.
12. Hoang-Xuan K, Merel P, Vega F, et al. Analysis of the NF2 tumor-suppressor
gene and of chromosome 22 deletions in gliomas. Int J Cancer. 1995;60(4):478-
481.
13. Watkins D, Ruttledge MH, Sarrazin J, et al. Loss of heterozygosity on
chromosome 22 in human gliomas does not inactivate the neurofibromatosis type
2 gene. Cancer Genet Cytogenet. 1996;92(1):73-78.
14. De Vitis LR, Tedde A, Vitelli F, et al. Analysis of the neurofibromatosis
type 2 gene in different human tumors of neuroectodermal origin. Hum Genet.
1996;97(5):638-641.
15. Alonso ME, Bello MJ, Arjona D, et al. Analysis of the NF2 gene in oligoden-
drogliomas and ependymomas. Cancer Genet Cytogenet. 2002;134(1):1-5.
16. Ebert C, von Haken M, Meyer-Puttlitz B, et al. Molecular genetic analysis
of ependymal tumors. NF2 mutations and chromosome 22q loss occur prefer-
entially in intramedullary spinal ependymomas. Am J Pathol. 1999;155(2):
627-632.
17. Lamszus K, Lachenmayer L, Heinemann U, et al. Molecular genetic alter-
ations on chromosomes 11 and 22 in ependymomas. Int J Cancer. 2001;91(6):
803-808.
18. Birch BD, Johnson JP, Parsa A, et al. Frequent type 2 neurofibromatosis gene
transcript mutations in sporadic intramedullary spinal cord ependymomas. Neuro -
surgery. 1996;39(1):135-140.
4|VOLUME0|NUMBER0|2017 www.neurosurgery-online.com
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/doi/10.1093/neuros/nyx374/3988116/High-Grade-Glioma-is-not-a-Feature-of
by guest
on 14 October 2017