Medulloblastoma - Translating discoveries from the bench to the bedside

Abstract

Medulloblastoma is a form of brain cancer that mainly arises during infancy and childhood. Our understanding of this disease has transitioned rapidly; what was once thought of as a single disease entity is now known to be a compendium comprising at least four distinct subtypes of tumour (Wnt, sonic hedgehog [SHH], group 3, and group 4 medulloblastomas) that have characteristic molecular signatures, distinctive clinical features, and are associated with different outcomes. Importantly, medulloblastomas occurring in infants (aged up to 3 years) and adults have unique characteristics, which distinguish the disease from that seen in children aged >3 years. Accordingly, modern treatment approaches in medulloblastoma integrate the molecular and clinical features of the disease to enable provision of the most-effective therapies for each patient, and to reduce long-term sequelae. This Review discusses our current knowledge of medulloblastoma. In particular, we present the genetic and histological features, patient demographics, prognosis, and therapeutic options for each the four molecular tumour subtypes that comprise this disease entity. In addition, the unique features of medulloblastoma in infants and in adults, as compared with childhood and/or adolescent forms, are described.

Full text

714
|
DECEMBER 2014
|
VOLUM E 11 www.nature.com/nrclinonc
Department of
Oncology, Division
ofNeuro-Oncology,
StJude Children’s
Research Hospital,
262Danny Thomas
Place, Memphis,
TN38105, USA (A.J.G.,
G.W.R.).
Correspondence to:
A.J.G.
amar.gajjar@stjude.org
Medulloblastoma—translating discoveries
from the bench to the bedside
Amar J.Gajjar and Giles W.Robinson
Abstract | Medulloblastoma is a form of brain cancer that mainly arises during infancy and childhood. Our
understanding of this disease has transitioned rapidly; what was once thought of as a single disease entity
is now known to be a compendium comprising at least four distinct subtypes of tumour (Wnt, sonic hedgehog
[SHH], group3, and group4 medulloblastomas) that have characteristic molecular signatures, distinctive clinical
features, and are associated with different outcomes. Importantly, medulloblastomas occurring in infants (aged
up to 3years) and adults have unique characteristics, which distinguish the disease from that seen in children
aged >3years. Accordingly, modern treatment approaches in medulloblastoma integrate the molecular and
clinical features of the disease to enable provision of the most-effective therapies for each patient, and to
reduce long-term sequelae. This Review discusses our current knowledge of medulloblastoma. In particular,
we present the genetic and histological features, patient demographics, prognosis, and therapeutic options for
each the four molecular tumour subtypes that comprise this disease entity. In addition, the unique features of
medulloblastoma in infants and in adults, as compared with childhood and/or adolescent forms, are described.
Gajjar, A. J. & Robinson, G. W. Nat. Rev. Clin. Oncol. 11, 714–722 (2014); published online 28 October 2014; doi:10.1038/nrclinonc.2014.181
Introduction
Medulloblastoma is a disease that predominantly occurs
in infants and children, and is the most-common type of
paedi atric malignant brain tumour, accounting for about
20% of all childhood brain cancers.1 Contempor ary therapy
for this disease consists of surgical resection, cranio spinal
irradiation, and chemotherapy. The use of these modali-
ties in modern therapeutic protocols has res ulted in a
cure rate of approximately 70–75% among children aged
≥3years.2–5 Current treatment protocols stratify patients
into high-risk and average-risk groups according to the
presence or absence of metastasis at diagnosis or of post-
operative residual disease. This approach has effectively
improved cure rates for patients with high-risk disease,
and enabled treatment exposure (for example, craniospinal
radiation doses) to be reduced in patients with average-risk
disease, resulting in decreased toxicity.6 How ever, medullo-
blastoma survivors continue to pay a high price, in terms
of long-term adverse sequelae, for cure. Deficits in neuro-
cognitive and neuroendocrine function, hearing, fertil-
ity, cardiopulmonary fitness, and physical performance
are some of the common effects of therapy.7–12 Further-
more, among survivors, rates of academic failure and
unemployment are high, and quality of life is reduced.13,14
A major drawback of the current risk-stratification
scheme is that it fails to recognize the heterogeneity of
medulloblastoma. Detection and categorization of this
heterogeneity is of fundamental importance because a
multitude of studies have shown that intrinsic differences
in the molecular profiles of medulloblastomas result in
substantial differences in disease manifestation and clini-
cal outcome;15–17 therefore, to improve cure rates among
patients with molecularly aggressive medulloblastoma and
reduce treatment-related morbidities, a therapeutic strat-
egy in which treatment is tailored to match these distinc-
tions is needed. The intensity of therapy could be reduced
in subpopulations of patients predicted to have good out-
comes, resulting in a low prevalence of treatment- related
morbidities, whereas patients with adverse prognostic
features that were not apparent at clinical presentation
could be allocated more-intensive therapy, which might
improve overall survival. Patient groups for whom current
maximal therapy is not sufficient could be treated with
novel therapies, and patients harbouring highly specific
genetic abnormalities could be treated with targeted
agents specific to the aberrations detected.
The aim of this Review is to discuss our current know-
ledge of medulloblastoma. We outline the contemporary
understanding of the molecular subtypes of medulloblas-
toma, with regard to genetic and histological features,
patient demographics, probabilities of survival, and treat-
ment. We also outline the unique features of medullo-
blastoma in infants (aged <3years) and those in adults,
versus the phenotype of medulloblastomas p resenting in
children and adolescents.
Medulloblastoma subtypes
Histological classification
Pathologists have long described medulloblastoma as
a heterogeneous disease, on the basis of multiple con-
sistently identi fied histological variants. The WHO
Competing interests
A.J.G. and G.W.R. are investigators on a clinical protocol that
isfunded, in part, by Genentech.
REVIEWS
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS
|
CLINICAL ONCOLOGY VOLUME 11
|
DECEMBER 2014
|
715
classification of nervous system tumours lists ‘classic’
medulloblastoma and four histological variants of the
disease: desmoplastic nodular (D/N), medulloblastoma
with extensive nodularity (MBEN), anaplastic, and large
cell.18 The four histological medullo blastoma variants can
be grouped into two pairs with overlapping morpholo-
gies: desmoplastic tumours comprising D/N and MBEN
variants; and large cell and anaplastic (LCA) tumours
(Figure1). The identification of medullo blastoma vari-
ants has clinical utility, as desmo plastic tumours are
associated with a better outcome than classic or LCA
medulloblastomas in infants, and LCA tumours are associ-
ated with a poorer outcome than classic or desmoplastic
tumours in children aged ≥3years.19–22 On the basis of
these findings, two ongoing clinical trials23,24 are assign-
ing therapy according to histological variants, in addition
to clinical risk. This strategy recognizes the importance
of t horoughly characterizing each tumour before initi-
ating treatment, owing to intertumour variability, but it
does not reflect the differences that exist within the same
h istological variants of medulloblastoma.
The current molecular classification
Our understanding of medulloblastoma biology has been
substantially enhanced by high-throughput genomic and
proteomic methods, such as transcriptomic and methyl-
omic analyses. An early genomics study from 2006,25
demonstrated that medulloblastoma might consist of
biologically distinct subgroups of tumours. The groups,
which were distinguished by transcriptomic differences,
exhibited intragroup similarities in mutation profiles,
structural chromosomal alterations, histology, demo-
graphics, and clinical outcome.25 These initial data have
been validated in several laboratories worldwide using
large cohorts of patients with tumours and technologi-
cally more-advanced transcriptomic techniques than
those used in the 2006 study.16,26,27 The result of these
efforts is a consensus that medulloblastoma consists of
four clinical and molecular subtypes of disease: the Wnt
subtype, in which canonical Wnt signalling is upregu-
lated; the sonic hedgehog (SHH) subtype, with hallmark
activation of the SHH-signalling cascade; and ‘group3’
and ‘group4’ medulloblastomas.28
The demographic, transcriptional, genetic, and clinical
differences among these four disease subtypes have impor-
tant clinical implications. The most clinically relevant
finding is that prognosis differs markedly across tumour
subgroups. The Wnt subtype has an amazingly high 5-year
overall survival rate that can exceed 90% with the current
standard therapy, which consists of maximal safe surgical
resection of tumour, risk-adapted radiation therapy, and
adjuvant chemotherapy.17,29,30 By contrast, group3 tumours
have a substantially worse prognosis, with 5-year overall
survival ranging from 40–60%.16,17,26 The other two sub-
groups of medulloblastoma, the SHH subtype and group4
tumours, have an intermediate overall survival rate at 5years
after treatment of around 75%, which varies according to
the presence or absence of metastatic disease, molecular
abnormalities, and histological category.17,21,26,28,30 Apoten-
tial caveat of these data, is that the outcomes werederived
from study protocols in which the patients were treated
mainly according to clinical-risk features, rather than on
the basis of molecular characteristics; indeed, these findings
have led to the obvious conclusion that molecular subtyping
should be used clinically to tailor therapy.
Wnt-subtype medulloblastoma
The rarest subtype of medulloblastoma is the Wnt subtype,
which makes up only about 10% of all medulloblastoma
diagnoses.17,30 Patients with this tumour subtype have
the best prognosis of all the subtypes.3,5 Wnt-subtype
tumours are typically uniform in their genetic aberrations,
h istological pattern, and clinical presentation.31
Genetics
All medulloblastomas with nuclear accumulation of
β-catenin are categorized as Wnt-subtype tumours;5,29,30
nuclear β-catenin interacts with members of the transcrip-
tion factor/lymphoid enhancer-binding factor (TCF/LEF)
family of transcription factors to activate the canonical
Wnt-signalling pathway.32,33 More than 90% of the Wnt-
subtype medulloblastomas harbour mutations in CTNNB1,
the gene that encodes β-catenin (Box1);31,34–36 the result-
ing mutant β-catenin protein is resistant to degra dation,
leading to its accumulation in the cell nucleus.37 Wnt-
subtype medulloblastomas also frequently have deletions
of one copy of chromosome 6 (monosomy6; Box1),25,29
although some Wnt tumours retain two copies of this
chromosome. Other than monosomy6, Wnt-subtype
medullo blastoma is associated with limited occurrence
of gains and/or losses of chromosomal regions across the
genome.31 Thus, monosomy 6, in conjunction with nuclear
β-catenin accumulation, serves as a sensitive and highly
specific marker for this subtype of disease.30
Whole-genome sequencing (WGS) studies have identi-
fied recurrent mutations specific to this medulloblastoma
subgroup. The most prevalently mutated genes, in addi-
tion to CTNNB1, are DDX3X, SMARCA4, TP53, KMT2D,
CSNK2B, and CREBBP (Box1). Many of these genes (that
is, CREBBP, SMARCA4, and KMT2D) encode proteins
that interact with nuclear β-catenin and remodel chro-
matin, suggesting that cooperative mutations occur in the
development of this tumour subtype.34
Key points
■Medulloblastoma is a malignant brain tumour that occurs predominantly in
childhood, but is also seen in infancy and throughout adulthood
■Although the prognosis of medulloblastoma is favourable with current
therapeutic regimens, the heterogeneous nature of this cancer has confounded
efforts to substantially improve survival and reduce therapy-related toxicity
■Advancements in technology and its accessibility have led, through molecular
interrogation, to the recognition that medulloblastoma heterogeneity is broadly
explained by the existence of four main molecular tumour subtypes
■Each molecular medulloblastoma subtype, termed Wnt, SHH, group 3, and
group 4 medulloblastoma, has unique clinical and molecular characteristics,
which influence nearly every facet of the disease, including survival
■Armed with this knowledge, paediatric oncologists find themselves at an
opportune moment to capitalize on these newly elucidated characteristics
to improve survival and reduce morbidity by tailoring therapy towards the
individualsubtypes
FOCUS ON PAEDIATRIC CANCER
© 2014 Macmillan Publishers Limited. All rights reserved

716
|
DECEMBER 2014
|
VOLUM E 11 www.nature.com/nrclinonc
Histology
Most Wnt-subtype tumours have classic histological
features of medulloblastoma. However, rare examples
of Wnt-subtype medulloblastoma with LCA-variant
h istology have been reported (Box1).30
Patient demographics and outcomes
The male to female ratio for Wnt-subtype medulloblas-
toma is almost 1:1, with a slight female predominance.28
These tumours are most commonly found in older chil-
dren and teenagers, and are rarely observed in infants
(Figure2; Box1).17,31 The tumours are typically located
in the midline of the brain, occupying the fourth ven-
tricle and infiltrating the brain stem. A mouse model that
mimics the human disease has demonstrated that the cell
of origin for Wnt medulloblastoma is located in the lower
rhombic lip and fails to migrate normally after accumu-
lating an oncogenic mutation.38 As introduced previously,
Wnt-subtype medulloblastomas have a highly favourable
outcome, with 5-year overall survival exceeding 90% in
most studies in patients with average-risk disease (gross
total resection and no metastaticdisease).3,5
Therapeutic options
On the basis of the good prognosis of Wnt-subtype
medulloblastoma, this form of the disease clearly lends
itself to a judicious reduction in therapy. Nevertheless,
reduced-dose craniospinal radiation and/or reduced-
intensity chemotherapy is warranted only for patients
without metastatic disease, which accounts for around
90–95% of Wnt-subtype tumour diagnoses (Box1).
Furthermore, testing of these alterations to standard
treatment is recommended only in patients who have
been carefully subtyped in a central reference laboratory
and enrolled on approved clinical protocols, with close
monitoring to document treatment outcome.
SHH-subtype medulloblastoma
SHH-subtype medulloblastomas constitute about
30% of all medulloblastoma diagnoses.17 Patients with
this tumour type have a 5-year overall survival rate of
approximately 75% when treated with current standard
therapy.28,32 Unlike Wnt-subtype medulloblastomas,
tumours characterized by activation of SHH signalling
are associated with a variety of genetic aberrations, histo-
logical features, and clinical presentations (Box2).31 The
heterogeneity within the SHH medulloblastoma subtype
is related to patient age at diagnosis and the underlying
genetic alterations. Consequently, risk classification of
these tumours is essential to determining prognosis and,
therefore, treatment strategies.
Genetics
Activation of the SHH-signalling pathway was first
linked with medulloblastoma as a result of the finding
that patients with Gorlin syndrome were found to have
a strong predisposition to the disease.39 These patients
have a high prevalence of basal-cell carcinoma and
develop mental anomalies that are caused by germline
mutations in the PTCH1 tumour-suppressor gene.39
This gene encodes protein patched homologue 1 (PTC1),
which is a receptor for SHH and other hedgehog homo-
logues. In the absence of ligand occupancy, PTC1 inter-
acts with and prevents smoothened homologue (SMO)
function, thus acting as a negative regulator of the SHH-
signalling cascade, a pathway that has many functions
in normal development.40 Normal cerebellar develop-
ment is highly reliant on SHH signal ling; however,
unrestrained SHH activity can lead to neo plasia. Indeed,
the integral role of this signalling pathway in cerebel-
lar development became apparent when engineered
a
c
d e
b
Classic medulloblastoma
D/N MBEN
Anaplastic Large cell
N
IN
IN
N
N
IN
N
Figure 1 | Histological variants of medulloblastoma. a | Microscopy image showing
the classic medulloblastoma histology, which is characterized by sheets of small
cells with a high nuclear-to-cytoplasmic ratio and mild nuclear polymorphism.
b|Photomicrograph depicting the D/N medulloblastoma histological pattern. The
tumour tissue contains nodules (N) of differentiated neurocytic cells that express
neuronal proteins. Internodular (IN) regions are characterized by undifferentiated
embryonal cells and reticulin-positive strands of collagen. c | Photomicrograph of
MBEN tissue. Similar in appearance to D/N histology, this variant is also
characterized by nodular (N) and internodular (IN) regions; however, as illustrated
these nodular regions are more abundant and often filled with streaming cells
(arrow) along a fibrillary background. Both D/N and MBEN histologies are nearly
exclusive to the SHH subtype of medulloblastoma. d | Histological section from an
anaplastic medulloblastoma. This histological variant is characterized by marked
nuclear pleomorphism, cell wrapping, a high mitotic count, and abundant apoptotic
bodies. To qualify as an anaplastic tumour, a medulloblastoma must contain
extensive regions with this histological phenotype. e|Histological sections from a
large cell medulloblastoma displaying the characteristic features of monomorphic
cells with large nuclei and prominent nucleoli. This histological pattern is frequently
found together with regions of anaplasia within the same tumour and thus both the
anaplastic and the large cell variant are grouped into an LCA category.
Abbreviations: D/N, desmoplastic nodular; LCA, large cell/anaplastic; MBEN,
medulloblastoma with extensive nodularity. Permission obtained from Springer ©
Ellison, D. W. Acta Neuropathol. 120, 305–316 (2010).20
REVIEWS
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS
|
CLINICAL ONCOLOGY VOLUME 11
|
DECEMBER 2014
|
717
Ptch1 deletions in mice were demonstrated to result
in medullo blastoma;41 this model represents a poten-
tially value tool for i nvestigating m ultiple facets of
SHH-subtypemedulloblastoma.
PTCH1 mutations have been reported in 25–30%
of SHH-subtype medulloblastomas, making it the
most-prevalent mutation of the subgroup (Box2).31
Interestingly, most of these mutations occur sporadi-
cally in patients without Gorlin syndrome. Other genetic
anomalies in the SHH pathway that have been detected
in this medulloblastoma subtype include mutations in
SMO and SUFU (encoding suppressor of fused homo-
logue, another suppressor of SHH signalling), as well as
amplification of SHH, and the transcription factors GLI2
and MYCN,34–36 confirming the link between this type
of medulloblastoma and the SHH pathway. The preva-
lence of these genetic anomalies is associated with age at
diagnosis. Most infant medulloblastomas carry PTCH1
or SUFU mutations, which are present in the patient’s
germline in a number of cases.42 Adult SHH-subtype
medulloblastomas are characterized by PTCH1 and SMO
mutations.42 In children, however, SHH-subtype tumours
display broader molecular heterogeneity, with amplifica-
tions of SHH, GLI2, and MYCN, as well as somatic and
germline TP53 mutations being observed together with
PTCH1 mutations.42 This pattern of genetic anomalies in
the SHH pathway is important as it might predict disease
response to treatment with SMO inhibitors.36,42,43
In addition, WGS studies of SHH-subtype medullo-
blastoma have reported recurrent somatic mutations
in genes that are seemingly unrelated to the SHH
pathway. These include KMT2D, TP53, BCOR, DDX3X,
LDB1, and GABRG1.34–36 Thus, similar to Wnt-subtype
medullo blastoma, recurrent mutation of genes encoding
chromatin remodelling proteins—KMT2D and BCOR
—features strongly in the SHH medulloblastoma subtype
(Box2); this finding warrants further investigation.
Copy-number variation studies have described some
intriguing chromosomal structure alterations, as well
as focal aberrations, that are characteristic of SHH-
subtype medulloblastoma. In particular, broad chromo-
somal losses of 9q, 10q, and 17p have been described
(Box2).44,45 Focal events that include loss of PTEN and
amplification of genes involved in insulin-like growth
factor signalling implicate increased PI3K activity in
development of some tumours within this subtype.31,44
Histology
Several case series have shown that D/N medullo-
blastomas belong predominantly to the SHH-subtype
of disease; similarly, MBENs are almost exclusive classi-
fied as this subtype.17,30 However, not all SHH-subtype
medulloblastomas have D/N or MBEN histology; the
remaining tumours have either classic medulloblastoma
or LCA histological features. Indeed, the SHH subtype is
the only subgroup of medulloblastoma with a consider-
able representation of all medulloblastoma histological
phenotypes (classic, D/N, MBEN and LCA; Box2).28,32
Patient demographics and outcomes
SHH and Wnt subtypes of medulloblastoma are similar
in that the disease is observed at comparably fre-
quencyin males and females, although SHH tumours
have a slight male predominance among infants
(Box2).28 Cerebellar hemispheric tumour location is
almost exclusively predominant to the SHH subtype,
although some SHH-type tumours do arise in the
Box 1 | Clinical and genomic features of Wnt-subtype medulloblastoma*
Clinical features
■Proportion of all medulloblastomas: ~10%
■Gender ratio (male:female): ~1:1
■Incidence: most common in older children and adolescents (median age
~10years); rare in infants (aged <3 years)
■Histology: classic; very rare cases of large cell/anaplastic
■Proposed cell of origin: lower rhombic lip progenitor cells
■MRI location: 4th ventricular; infiltrating brainstem
■Prevalence of metastasis at diagnosis: ~5–10%
■5-year overall survival rate: ~95%
Genomic features
■Expression signature: Wnt signalling
■Chromosomal gains and losses: monosomy of chromosome 6
■Driver genes‡: CTNNB1 (90.6%); DDX3X (50%); SMARCA4 (26.3%); TP53 (13.5%);
KMT2D (12.4%)
*Subgroup frequency, demographics, clinical features and cytogenetic aberrations were
derived from a cohort of 827 medulloblastomas distributed into subgroups described by
Northcott et al.26 ‡Driver genes are determined by the relative frequency of mutations or
significant copy-number aberrations in medulloblastomas that were distributed into either Wnt,
SHH, group 3 or group 4 molecular subtypes, as described in recent medulloblastoma
sequencing and copy number studies.26,34–36
SHH
Group 3
Group 4
SHH
Group 4
Wnt
Group 3
SHH
Group 4
Wnt
Infant
Child
Adult
Figure 2 | Schematic distribution of the prevalence of the molecular
medulloblastoma subtypes among different age groups. The approximate frequency
of each of the four molecular subtypes of medulloblastoma (Wnt, SHH, group 3 and
group 4) according to age at diagnosis is shown schematically as a proportion of
total medulloblastoma occurrence in each age group, based on data from 827
medulloblastomas that were distributed into these molecular subgroups by
Northcottet al.26 The data are stratified into medulloblastomas occurring in infants
(aged<3years), children (aged >3 years) and adults (aged >16 years). Wnt-subtype
tumours are rarely observed in infants, but make up a similar proportion of
medulloblastoma cases in children and adults. SHH-subtype tumours have a unique
bimodal distribution, and occur most commonly in infants and adults, whereas
group4 medulloblastomas are most frequently seen in children—and are the
predominant form of medulloblastoma in this age group. The peak age of onset for
group 3 tumours is in late infancy/early childhood; this medulloblastoma subtype
isalmost never observed in adults. Abbreviation: SHH, sonic hedgehog (protein).
FOCUS ON PAEDIATRIC CANCER
© 2014 Macmillan Publishers Limited. All rights reserved

718
|
DECEMBER 2014
|
VOLUM E 11 www.nature.com/nrclinonc
midline of the brain.38,46 Metastatic disease at diagno-
sis is also relatively uncommon, occurring in less than
25% of cases.17,31,32 In contrast with other medulloblas-
toma subtypes, SHH-subtype tumours have a unique
bimodal pattern in incidence, with most cases involv-
ing either infants younger than 3years of age or older
adolescents and adults (Figure2; Box2).47 In general,
the 5-year overall survival rate in patients with SHH
medulloblastoma is about 75%. However, this outcome
probably reflects the heterogeneity within the subgroup;
more-detailed analyses of risk factors have documented
that outcomes differ according to patient age, histo-
logical subtype, presence of metastasis at diagnosis, and
u nderlying molecularabnormalities.17,30,47
Therapeutic options
Although a good outcome is achieved for most patients
with SHH-subtype medulloblastoma using current
therapies, improved and refined therapeutic strategies
are needed for this subtype. In this regard, the chal-
lenge lies in our ability to identify diversity within this
subgroup of patients and tailor therapy accordingly.
Treatment of infants with D/N disease has demon-
strated that certain SHH tumours can be cured without
radiotherapy,4 whereas other forms of the disease, such
as tumours with MYCN and GLI2 amplifications, carry
such a dismal prognosis that treatment even with high-
dose craniospinal radiation and adjuvant chemotherapy
is rarely curative.45
SMO inhibitor therapy (with agents such as vismo-
degib and erismodegib) have been associated with a
marked but relatively short-lived response in patients
with recurrent SHH-subtype medulloblastoma.48,49
Prolonged exposure to these drugs induces a muta-
tion in SMO that changes the structure of the protein
and prevents binding of the drug.50 Unfortunately, only
tumours with either PTCH1 or certain SMO mutations
are predicted to respond to this form of intervention,
and although these mutations have been documented
in the majority of adult patients with medulloblastoma,
they occur in only about half of paediatric patients with
thedisease.42 Trials are currently underway to evalu-
atethe use of SMO inhibitors after surgery, irradiation,
and modified adjuvant chemotherapy.51
Owing to the limited applicability of SMO inhibitors
in the treatment of medulloblastomas with activation of
the SHH pathway, novel strategies are needed for patients
who have SUFU mutations, deletion of PTEN, and ampli-
fications of GLI2 and MYCN; these strat egies could
poten tially include epigenetic modifiers and PI3K–AKT
inhibitors. A small group of patients with SHH-subtype
medulloblastoma have large cell histo logy, MYCN ampli-
fication, and germline TP53 mutations. Such patients
need to be identified and potentially offered novel
therapeutic strategies, as they have been documented to
have poor prognosis and often have Li–Fraumeni syn-
drome.42 International collaboration will be imperative
to driving therapeutic advances through well-planned
clinical protocols.
Group3 medulloblastoma
Group3 medulloblastoma accounts for around 25% of
all medulloblastoma diagnoses.17 Patients with group3
disease have the worst outcome among all the medullo-
blastoma subgroups due to the high frequency of adverse
prognostic features, namely younger age at diagnosis,
high prevalence of metastatic disease at diagnosis, LCA
histology, and MYC amplification.28
Genetics
No germline mutations that predispose children to
group3 medulloblastoma have been described, and
recur rent somatic genomic aberrations have only recently
been reported.32,34–36,44 These aberrations include ampli-
fications of MYC and OTX2 (Box3); mutations in the
chromatin remodelling proteins encoded by SMARCA4,
KMT2D, and CHD7; and a variety of mutations in the
lysine-specific demethylase (KDM) gene family (KDM6A,
KDM3A, KDM4C, KDM5B, and KDM7A).31,34 MYC
amplicons are highly prognostic and mutually exclusive
from OTX2 amplicons, suggest ing similar but prognosti-
cally distinct pathways to neo plasia in group3 medullo-
blastoma.44 Frustratingly, >50% of group3 tumours do
not harbour any of these genetic aberrations, but rather
have widespread chromo somal structural alterations
(such as copy-number gains in 1q and 7, and losses in 10q
and 16q).31 Although the role of these structural varia-
tions in the pathogenesis of medulloblastoma remains
poorly understood, a study published in 2014 provided
a potential solution to this conundrum. Therein, recur-
rent structural variations common to both group3 and
group4 tumours were identified to reposition active
Box 2 | Clinical and genomic features of SHH-subtype medulloblastoma*
Clinical features
■Proportion of all medulloblastomas: ~30%
■Gender ratio (male:female): ~1.5:1
■Incidence: bimodal—first peak in infants and young children (<5 years of age);
second peak in older adolescents and adults (aged >16 years); less common
in children aged 5–16 years
■Histology: classic > desmoplastic nodular > LCA > medulloblastoma with
extensive nodularity
■Proposed cell of origin: cerebellar granule-neuron precursor cells of the
external granule-cell layer and cochlear nucleus; neural stem cells of
thesubventricularzone
■MRI location: cerebellar hemispheres and rarely midline
■Prevalence of metastasis at diagnosis: ~15–20%
■5-year overall survival rate: ~75%
Genomic features
■Expression signature: SHH signalling
■Chromosomal gains and losses: frequent loss of 9q, 10q, and 17p; gain of 3q
and 9p
■Driver genes‡: PTCH1 (28%); TP53 (13.6%); KMT2D (12.9%); DDX3X (11.7%);
MYCN amplification (8.2%); BCOR (8%); LDB1 (6.9%); TCF4 (5.5%); GLI2
amplification (5.2%)
*Subgroup frequency, demographics, clinical features and cytogenetic aberrations were
derived from a cohort of 827 medulloblastomas distributed into subgroups described by
Northcott et al.26 ‡Driver genes are determined by the relative frequency of mutations or
significant copy-number aberrations in medulloblastomas that were distributed into either
Wnt, SHH, group 3 or group 4 molecular subtypes, as described in recent medulloblastoma
sequencing and copy number studies.26,34–36 Abbreviations: LCA, large cell/anaplastic;
SHH,sonic hedgehog (protein).
REVIEWS
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS
|
CLINICAL ONCOLOGY VOLUME 11
|
DECEMBER 2014
|
719
enhancer regions in close proximity to the known onco-
genes GFI1 or GFI1B, thus serving to aberrantly activate
oncogene expression within these tumours.52
Histology
Only two histological subclasses of medulloblastoma
are seen in group3 tumours: classic and LCA (Box3).
Group3 tumours harbour the highest prevalence of the
LCA histology of all medulloblastoma subtypes—as high
as 40%.17,31 D/N histology, on the other hand, is almost
never seen in group3 tumours.17,31
Patient demographics and outcomes
Group 3 tumours have a male predominance, being
generally twice as common in males than in females
(Box3). Group3 tumours present predominantly
ininfants and children, rarely in teenagers and not in
adults (Figure2).32 Metastatic disease is present at diag-
nosis in approximately 40–45% of patients. The higher
prevalence of MYC amplification, metastatic disease,
and LCA histology impart the extremely poor prognosis
in group3 medulloblastoma. In fact, in a retrospective
meta- analysis published in 2012,17 group3 tumours were
associated with the worst outcomes of all the medullo-
blastoma subtypes among all patient age groups: the
5-year and 10-year overall survival rates in infants with
group3 tumours were 45% and 39%, respectively; in
c hildren, the rates were 58% and 50%, respectively.17,31
Therapeutic options
Group3 medulloblastoma is the most challenging form
of the disease to treat, and as is evident from the out-
comes presented, the current therapeutic strat egies
are often ineffective. Several groups are conducting
preclinical studies in mouse models that mimic the
human disease in order to discover effective therapies
for this aggressive subtype.53,54 Two FDA-approved
compounds, gemcitabine and pemetrexed, were identi-
fied by a high-throughput screen as potentially having
efficacy against MYC-overexpressing or MYC-amplified
medulloblastoma.55 Another therapeutic strategy with
promise in group3 medulloblastoma is the use of BET
bromodomain inhibitors, which interfere with MYC-
associated transcriptional activity.56 Carefully planned
prospective clinical studies will be needed to determine
if these or other agents identi fied in preclinical models
in the future improve the outcome of patients with
group3medulloblastoma.
Group4 medulloblastoma
Group4 medulloblastoma is the most-prevalent subtype
of this disease and accounts for as many as 35% of
medullo blastoma diagnoses (Box4).17 Group4 tumours
are seen in all age groups and chromosome 17 abnormal-
ities, although not exclusive to group4 tumours, are the
molecular hallmark of this subgroup. Overall, patients
with group4 medulloblastoma have an intermediate
prognosis among the medulloblastoma subtypes when
treated with standard therapy.28
Genetics
Although group4 medulloblastoma is the most common
of the medulloblastoma subgroups, the underlying
biology of this form of the disease is not well under-
stood.32 No familial syndromes predispose an individual
to group4 medulloblastoma, and no murine model of
the disease has been generated. Genes that are recur-
rently mutated or altered in copy number overlap with
those associated with group3 medulloblastomas. These
are mutations affecting the KDM family members,
OTX2 amplicons, DDX31 deletions, CHD7 mutations,
activation of GFI1/GFI1B expression, and KMT2D and
KMT2C mutations. Unlike group3 tumours, however,
preferential amplification of MYCN rather than MYC is
observed in group4 tumours.31,34–36,44
Box 3 | Clinical and genomic features of group 3 medulloblastoma*
Clinical features
■Proportion of all medulloblastomas: ~25%
■Gender ratio (male:female): ~2:1
■Incidence: occurs predominantly in infants and young children
■Histology: classic > large cell/anaplastic
■Proposed cell of origin: prominin 1+/lineage– neural stem cells; cerebellar
granule-neuron precursor cells of the external granule-cell layer
■MRI location: 4th ventricular; midline
■Prevalence of metastasis at diagnosis: ~40–45%
■5-year overall survival rate: ~50%
Genomic features
■Expression signature: MYC signature; retinal signature
■Chromosomal gains and losses: frequent loss of 10q, 16q, and 17p; gain of
1q, 7, 17q, and 18
■Driver genes‡: MYC amplification (16.7%); PVT1 amplification (11.9%);
SMARCA4 (10.5%); OTX2 amplification (7.7%); CTDNEP1 (4.6%); LRP1B (4.6%);
KMT2D (4%)
*Subgroup frequency, demographics, clinical features and cytogenetic aberrations were
derived from a cohort of 827 medulloblastomas distributed into subgroups described by
Northcott et al.26 ‡Driver genes are determined by the relative frequency of mutations or
significant copy-number aberrations in medulloblastomas that were distributed into either Wnt,
SHH, group 3 or group 4 molecular subtypes, as described in recent medulloblastoma
sequencing and copy number studies.26,34–36
Box 4 | Clinical and genomic features of group 4 medulloblastoma*
Clinical features
■Proportion of all medulloblastomas: ~35%
■Gender ratio (male:female): ~3:1
■Incidence: children (median age ~9 years)
■Histology: classic; large cell/anaplastic in rare cases
■Proposed cell of origin: unknown
■MRI location: 4th ventricular; midline
■Prevalence of metastasis at diagnosis: ~35–40%
■5-year overall survival rate: ~75%
Genomic features
■Expression signature: neuronal signature; glutamatergic signature
■Chromosomal gains and losses: loss of 8, 10, 11, and 17p; gain of 4,7, 17q,
and 18
■Driver genes‡: KDM6A (13%); SNCAIP gain (10.4%); MYCN amplification (6.3%);
KMT2C (5.3%); CDK6 amplification (4.7%); ZMYM3 (3.7%)
*Subgroup frequency, demographics, clinical features and cytogenetic aberrations were
derived from a cohort of 827 medulloblastomas distributed into subgroups described by
Northcott et al.26 ‡Driver genes are determined by the relative frequency of mutations or
significant copy-number aberrations in medulloblastomas that were distributed into either Wnt,
SHH, group 3 or group 4 molecular subtypes, as described in recent medulloblastoma
sequencing and copy number studies.26,34–36
FOCUS ON PAEDIATRIC CANCER
© 2014 Macmillan Publishers Limited. All rights reserved

720
|
DECEMBER 2014
|
VOLUM E 11 www.nature.com/nrclinonc
The most frequent focal somatic copy-number aber-
ration, which occurs in 10% of patients with group4
tumours, is a single-copy gain on chromosome 5q23.2,
focused on the SNCAIP gene (Box4).31,44 SNCAIP
encodes synphillin-1, which binds to α-synuclein to
promote the formation of Lewy bodies in the brains of
patients with Parkinson disease; however, a connection
between SNCAIP function and medulloblastoma has
notbeen established.31,44
The most frequent mutation seen in group4 medullo-
blastomas occurs in the KDM6A gene (Box4).31,34–36
KDM6A is a demethylase enzyme that regulates the
methyl ation of lysine-27 of histone H3 (H3K27). H3K27
represents a histone lysine residue that character istically
retains a trimethylated state in stem cells. Therefore, by
preventing H3K27 demethylation, this mutation might
preserve or initiate a stem-cell-like state in tumour cells.
All the more intriguing is that KDM6A is found on
the Xchromosome and, although a homologue exists
on the Ychromosome, KDM6A seems to be more-
frequently mutated in boys with medulloblastoma,
compared with girls with the disease. These mutations
and others on the Xchromosome (such as ZMYM3)
might explain the observed male predominance of this
medulloblastomasubtype.34
Another commonly identified aberration in group4
tumours is isochromosome 17q, which is formed when
the p-arm of chromosome 17 is lost and is replaced
by the q-arm of the same chromosome—generating
a chromo some with two 17q arms. Although some
group3 tumours also exhibit this genetic abnormality,
isochromosome 17q occurs at a much higher frequency
in group4 medulloblastomas, and is rarely seen in
tumours of the Wnt or SHH subtypes. Similar to group3
tumours, many group4 tumours display none of the
aforementioned mutations or focal copy-number altera-
tions, and continued studies are needed to elucidate the
p athogenesis of this medulloblastoma subtype.17,26,31,34,44
Histology
As in group3 tumours, the classic and LCA histology are
the only histological subclasses regularly seen in group4
medulloblastoma.17 Furthermore, the frequency of LCA
is substantially lower in group4 tumours than among
group3 medulloblastomas.17 A D/N histology has been
documented for some group4 tumours;17 however, this
finding probably reflects incorrect categorization rather
than actual histological features associated with the
group4 subtype of medulloblastoma.
Patient demographics and outcomes
Group4 medulloblastomas have the most skewed distri-
bution, according to gender, with the disease occurring
three times more often in males than in females. This ratio
is seen across all age groups, although group4 tumours
rarely arise in infants (Figure2; Box4).17,28,32 This sub-
group accounts for 40–45% of childhood medulloblasto-
mas and 25% of adult medulloblastomas.17,26,47 Metastatic
disease at diagnosis is reported in a pproximately 35–40%
of patients with group4 tumours.17,31
Patients with average-risk group4 medulloblastoma,
as determined by gross-total surgical resection of the
tumour and the absence of metastatic disease, have a
5-year overall survival rate that exceeds 80%.17,45 Patients
with high-risk disease (that is, metastatic disease or
tumours with a LCA histology) have an inferior progno-
sis: 5-year overall survival is observed in approximately
60% of these patients. A retrospective study identified
a molecularly defined subgroup of patients within the
average-risk group4 population in whom a cure rate
>90% might be achievable;45 however, these data await
confirmation in a prospectively treated patient cohort.
Therapeutic options
As described, current standard therapy cures a high pro-
portion of patients with average-risk group4 medullo-
blastoma. Whether there is an opportunity to decrease
the intensity of therapy for a subgroup of these patients
who might have a particular high cure rate will emerge
as findings from the next generation of clinical trials are
published. As the stratification of patients according to
molecular characteristics in the newest clinical cohorts
will more-clearly identify the tumour subtypes and their
intricacies, we anticipate that outcomes among these
subgroups will be better determined and thereby enable
rational reductions in therapy in the future.
The absence of a preclinical murine model that mimics
the human group4 medulloblastoma has hamperedthe
development of novel approaches to treat patientswith
high-risk disease who experience poor outcomes
withstandard therapy. The molecular overlap between
group3 and group4 medulloblastomas implies that
drugs that are effective in group3 medulloblastoma
might also be effective in group4 tumours that have
adverse clinical and molecular features.
Unique aspects of medulloblastoma
Infant medulloblastomas
The SHH subgroup is the most-common form of the
disease in infants, accounting for more than half of all
medulloblastomas observed in this age group;17 group3
tumours, which constitute around 30–40% of infantile
medulloblastoma, are the next most-frequent form.17
Group4 tumours comprise the small remaining frac-
tion of cases, as Wnt-subtype medulloblastomas are
virtually nonexistent in infants.17 Most infants with
SHH-subtype medulloblastoma have tumours witha
D/N or MBEN histology and favourable molecular
characteristics (that is, somatic or germline mutations
in PTCH1), suggesting that the disease might be curable
with chemotherapy alone. Very few infantile medullo-
blastomas harbour the MYCN or GLI2 amplification or
germline TP53 mutations that are characteristic of more-
aggressive SHH tumours and are observed more fre-
quently in SHH-subtype medulloblastomas in children
aged >3years.42 Indeed, an international meta-analysisof
prognostic factors in 270 patients younger than 5years
of age demon strated that infant medulloblastoma with
desmo plastic (D/N or MBEN) histology has an excellent
prognosis and can be usually cured with postoperative
REVIEWS
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS
|
CLINICAL ONCOLOGY VOLUME 11
|
DECEMBER 2014
|
721
chemotherapy alone.57 Moreover, several prospective
studies have reported superior survival for infants with
this subtype of medulloblastoma.4
To protect infants from the toxicity of whole-brain
irradiation, investigators have used several treatment
modalities in attempts to improve the survival of infants
with adverse prognostic features, such as tumours with
classic or LCA histology, metastatic disease at diagno-
sis, and non-SHH-subtype medulloblastoma. These
approaches include using focal radiation therapy,58 and
high-dose myeloablative chemotherapy with stem-cell
rescue,59 but the cure rate among these infants remains
vastly inferior to that achieved in infants with SHH
tumours. Nevertheless, biological-risk features (such as
histology and molecular phenotypes) were not integrated
into the staging criteria used in the previous generation
of clinical trial protocols; hence, clinically average-
risk patients with high-risk biological (histological or
molecu lar) characteristics might have been undertreated
and account for the inferior outcomes observed in these
studies. Introduction of biologically and clinically staged
patients in study protocols that include novel chemo-
therapeutic strategies that are not toxic to the develop-
ing brain and organs will be necessary to improve the
outcome of infants with medulloblastoma.
Adult medulloblastomas
The lack of prospective clinical trials in adult patients with
medulloblastoma has limited the documentation of the
clinical features, molecular characteristics, and outcome
of the disease in this age group. However, the available lit-
erature indicates that SHH-subtype tumours account for
57% of adult medulloblastoma diagnoses.60 The remain-
ing adult medulloblastomas consist of the Wnt-subtype
(13%) and group4 (28%) tumours; group3 medulloblas-
toma is particularly rare in adults (comprising around
2% of adult medulloblastoma diagnoses).60 Both Wnt-
subtype and group4 tumours in adults are associated
with worse outcomes than their paediatric counterparts,
whereas the SHH-subgroup adult tumours have a similar
outcome.60 Interestingly, the molecular characteristics of
SHH-subtype medulloblastoma in adults are transcrip-
tionally distinct from those in children (a group in which
this form of medulloblastoma is less common; Figure2),
but have similar molecular characteristics to those seen
in infant SHH-subtype tumours—the exception being
tumours with SUFU mutations, which are comparatively
rare in adult medulloblastoma when compared with
the infant disease.47,42 On the basis of these molecular
findings, SMO antagonists should be an effective thera-
peutic option, at least in the short term, in a large subset
of adults with SHH-subtype medulloblastoma. Initiating
prospective clinical trials that direct therapy according
to the molecular subgroups will increase our knowledge
of the outcomes and unique toxicities of such agents in
adults with medulloblastoma.
The next generation of clinical trials
Translating the newly acquired knowledge of the molecu-
lar underpinnings of medulloblastoma to tailor alloca-
tion of therapy according to prognosis has already been
implemented in ongoing clinical trials: by the National
Cancer Institute (NCI)-funded paediatric clinical con-
sortium (Children’s Oncology Group ACNS122161)and
the St Jude Children’s Research Hospital (SJMB1262
andSJYC0763) in the USA; and by the International
Society of Paediatric Oncology (PNET564) in Europe.
The study designs for each of these trials vary from pilot
phaseII to single-arm phaseIII studies. Larger random-
ized phaseIII studies will be considered at a later stage,
after the data from current studies have matured.
Conclusions
The rapid progress in identifying the molecular and
clinical characteristics of medulloblastomas has ushered
in a new era in basic and clinical research in paediatric
neuro-oncology. Disease heterogeneity, once shrouded in
ambiguity, is now clearly related to inherent molecular
differences, which can be ascertained through the appli-
cation of modern technology to tumour interrogation.
This progress has already led to the generation of mouse
models that more-accurately mirror the human disease.
These tools and future models, alongside molecularly
driven prospective clinical trials that are designed to
answer questions regarding the outcomes of specific
tumour subtypes, will predictably facilitate the discovery
of novel therapies that are more effective and less toxic,
even to the youngest children with medulloblastoma.
Review criteria
We searched PubMed for English-language full-text
manuscripts and abstracts published between 1980
and 2014. The search terms used, alone and in various
combinations, were “medulloblastoma”, “genomics”,
“targeted therapy” and “mouse models”. The reference
lists of the articles identified were also searched
foradditional relevant publications.
1. Pui, C.H., Gajjar, A.J., Kane, J.R.,
Qaddoumi,I.A. & Pappo, A.S. Challenging
issues in pediatric oncology. Nat. Rev. Clin.
Oncol. 8, 540–549 (2011).
2. Packer, R.J. etal. PhaseIII study of craniospinal
radiation therapy followed by adjuvant
chemotherapy for newly diagnosed average-risk
medulloblastoma. J. Clin. Oncol. 24, 4202–4208
(2006).
3. Gajjar, A. etal. Risk-adapted craniospinal
radiotherapy followed by high-dose
chemotherapy and stem-cell rescue in children
with newly diagnosed medulloblastoma (St Jude
Medulloblastoma-96): long-term results from
aprospective, multicentre trial. Lancet Oncol. 7,
813–820 (2006).
4. Rutkowski, S. etal. Treatment of early childhood
medulloblastoma by postoperative chemotherapy
alone. N. Engl. J. Med. 352, 978–986 (2005).
5. Ellison, D.W. etal. β-catenin status predicts a
favorable outcome in childhood
medulloblastoma: the United Kingdom Children’s
Cancer Study Group Brain Tumour Committee.
J.Clin. Oncol. 23, 7951–7957 (2005).
6. Packer, R.J. etal. Treatment of children
withmedulloblastomas with reduced-dose
craniospinal radiation therapy and adjuvant
chemotherapy: a Children’s Cancer Group Study.
J. Clin. Oncol. 17, 2127–2136 (1999).
7. Silber, J.H. etal. Whole-brain irradiation and
decline in intelligence: the influence of dose
andage on IQ score. J. Clin. Oncol. 10,
1390–1396 (1992).
FOCUS ON PAEDIATRIC CANCER
© 2014 Macmillan Publishers Limited. All rights reserved

722
|
DECEMBER 2014
|
VOLUM E 11 www.nature.com/nrclinonc
8. Mulhern, R.K., Merchant, T.E., Gajjar, A.
Reddick, W.E. & Kun, L.E. Late neurocognitive
sequelae in survivors of brain tumours in
childhood. Lancet Oncol. 5, 399–408 (2004).
9. Hoppe-Hirsch, E. etal. Medulloblastoma in
childhood: progressive intellectual deterioration.
Childs Nerv. Syst. 6, 60–65 (1990).
10. Laughton, S.J. etal. Endocrine outcomes for
children with embryonal brain tumors after risk-
adapted craniospinal and conformal primary-site
irradiation and high-dose chemotherapy with
stem-cell rescue on the SJMB-96 trial. J. Clin.
Oncol. 26, 1112–1118 (2008).
11. Wolfe, K.R. etal. Cardiorespiratory fitness in
survivors of pediatric posterior fossa tumor. J.
Pediatr. Hematol. Oncol. 34, e222–e227 (2012).
12. Ness, K.K., Wall, M.M., Oakes, J.M.,
Robison,L.L. & Gurney, J.G. Physical
performance limitations and participation
restrictions among cancer survivors:
apopulation-based study. Ann. Epidemiol. 16,
197–205 (2006).
13. Armstrong, G.T. etal. Long-term outcomes
among adult survivors of childhood central
nervous system malignancies in the Childhood
Cancer Survivor Study. J. Natl Cancer Inst. 101,
946–958 (2009).
14. Mabbott, D.J. etal. Serial evaluation of
academic and behavioral outcome after
treatment with cranial radiation in childhood.
J.Clin. Oncol. 23, 2256–2263 (2005).
15. Pomeroy, S.L. etal. Prediction of central nervous
system embryonal tumour outcome based on
gene expression. Nature 415, 436–442 (2002).
16. Cho, Y.J. etal. Integrative genomic analysis
ofmedulloblastoma identifies a molecular
subgroup that drives poor clinical outcome.
J.Clin. Oncol. 29, 1424–1430 (2011).
17. Kool, M. etal. Molecular subgroups of
medulloblastoma: an international meta-
analysis of transcriptome, genetic aberrations,
and clinical data of, WNT, SHH, Group 3, and
Group 4 medulloblastomas. Acta Neuropathol.
123, 473–484 (2012).
18. Louis, D.N. etal. The 2007 WHO classification
of tumours of the central nervous system.
ActaNeuropathol. 114, 97–109 (2007).
19. Eberhart, C.G. etal. Histopathologic grading of
medulloblastomas: a Pediatric Oncology Group
study. Cancer 94, 552–560 (2002).
20. Ellison, D.W. Childhood medulloblastoma:
novelapproaches to the classification of a
heterogeneous disease. Acta Neuropathol. 120,
305–316 (2010).
21. Ellison, D.W. etal. Definition of disease-risk
stratification groups in childhood
medulloblastoma using combined clinical,
pathologic, and molecular variables. J. Clin.
Oncol. 29, 1400–1407 (2011).
22. McManamy, C.S. etal. Morphophenotypic
variation predicts clinical behavior in childhood
non-desmoplastic medulloblastomas. J.
Neuropathol. Exp. Neurol. 62, 627–632 (2003).
23. US National Library of Medicine. ClinicalTrials.gov
[online], http://www.clinicaltrials.gov/ct2/
show/NCT00085735?term=NCT00085735&
rank=1 (2014).
24. US National Library of Medicine. ClinicalTrials.gov
[online], https://www.clinicaltrials.gov/ct2/
show/NCT00392327?term=NCT00392327&
rank=1 (2014).
25. Thompson, M.C. etal. Genomics identifies
medulloblastoma subgroups that are enriched
for specific genetic alterations. J. Clin. Oncol. 24,
1924–1931 (2006).
26. Northcott, P.A. etal. Medulloblastoma
comprises four distinct molecular variants.
J.Clin. Oncol. 29, 1408–1414 (2011).
27. Kool, M. etal. Integrated genomics identifies five
medulloblastoma subtypes with distinct genetic
profiles, pathway signatures and
clinicopathological features. PLoS ONE 3, e3088
(2008).
28. Taylor, M.D. etal. Molecular subgroups of
medulloblastoma: the current consensus.
ActaNeuropathol. 123, 465–472 (2012).
29. Clifford, S.C. etal. Wnt/Wingless pathway
activation and chromosome 6 loss
characterize a distinct molecular sub-group
of medulloblastomas associated with a
favorable prognosis. Cell Cycle 5, 2666–2670
(2006).
30. Ellison, D.W. etal. Medulloblastoma:
clinicopathological correlates of, SHH, WNT,
andnon-SHH/WNT molecular subgroups.
ActaNeuropathol. 121, 381–396 (2011).
31. Northcott, P.A. etal. Medulloblastomics: the end
of the beginning. Nat. Rev. Cancer 12, 818–834
(2012).
32. Northcott, P.A., Korshunov, A., Pfister, S.M.
&Taylor, M.D. The clinical implications of
medulloblastoma subgroups. Nat. Rev. Neurol.
8, 340–351 (2012).
33. Mosimann, C., Hausmann, G. & Basler, K.
β-catenin hits chromatin: regulation of Wnt target
gene activation. Nat. Rev. Mol. Cell Biol. 10,
276–286 (2009).
34. Robinson, G. etal. Novel mutations target
distinct subgroups of medulloblastoma. Nature
488, 43–48 (2012).
35. Jones, D.T. etal. Dissecting the genomic
complexity underlying medulloblastoma. Nature
488, 100–105 (2012).
36. Pugh, T.J. etal. Medulloblastoma exome
sequencing uncovers subtype-specific somatic
mutations. Nature 488, 106–110 (2012).
37. Gilbertson, R. J. Medulloblastoma: signalling a
change in treatment. Lancet Oncol. 5, 209–218
(2004).
38. Gibson, P. etal. Subtypes of medulloblastoma
have distinct developmental origins. Nature 468,
1095–1099 (2010).
39. Fujii, K. & Miyashita, T. Gorlin syndrome (nevoid
basal cell carcinoma syndrome)—an update
andliterature review. Pediatr. Int. http://
dx.doi.org/10.1111/ped.12461.
40. Huangfu, D. & Anderson, K.V. Signaling from
Smo to Ci/Gli: conservation and diveregence
ofHedgehog pathways from Drosophilia to
vertebrates. Development 133, 3–14 (2006).
41. Goodrich, L.V., Milenkovic´, L. Higgins, K.M.
&Scott, M.P. Altered neural cell fates and
medulloblastoma in mouse patched mutants.
Science 277, 1109–1113 (1997).
42. Kool, M. etal. Genome sequencing of SHH
medulloblastoma predicts genotype-related
response to smoothened inhibition. Cancer Cell
25, 393–405 (2014).
43. Taylor, M.D. etal. Mutations in SUFU predispose
to medulloblastoma. Nat. Genet. 31, 306–310
(2002).
44. Northcott, P.A. etal. Subgroup-specific
structural variation across 1,000
medulloblastoma genomes. Nature 488, 49–56
(2012).
45. Shih, D.J. etal. Cytogenetic prognostication
within medulloblastoma subgroups. J. Clin.
Oncol. 32, 886–896 (2014).
46. Grammel, D. etal. Sonic hedgehog-associated
medulloblastoma arising from the cochlear
nuclei of the brainstem. Acta Neuropathol. 123,
601–614 (2012).
47. Northcott, P.A. etal. Pediatric and adult sonic
hedgehog medulloblastomas are clinically and
molecularly distinct. Acta Neuropathol. 122,
231–240 (2011).
48. Rudin, C.M. etal. Treatment of medulloblastoma
with hedgehog pathway inhibitor GDC-0449.
N.Engl. J. Med. 361, 1173–1178 (2009).
49. Metcalfe, C. & de Sauvage, F.J. Hedgehog fights
back: mechanisms of acquired resistance
against Smoothened antagonists. Cancer Res.
71, 5057–5061 (2011).
50. Yauch, R.L. etal. Smoothened mutation confers
resistance to a Hedgehog pathway inhibitor in
medulloblastoma. Science 326, 572–574
(2009).
51. US National Library of Medicine. ClinicalTrials.gov
[online], http://www.clinicaltrials.gov/ct2/
show/NCT01878617 (2014).
52. Northcott, P.A. etal. Enhancer hijacking
activates GFI1 family oncogenes in
medulloblastoma. Nature 511, 428–434
(2014).
53. Kawauchi, D. etal. A mouse model of the most
aggressive subgroup of human
medulloblastoma. Cancer Cell 21, 168–180
(2012).
54. Pei, Y. etal. An animal model of MYC-driven
medulloblastoma. Cancer Cell 21, 155–167
(2012).
55. Morfouace, M. etal. Pemetrexed and
gemcitabine as combination therapy for
thetreatment of Group3 medulloblastoma.
CancerCell 25, 516–529 (2014).
56. Bandopadhayay, P. etal. BET bromodomain
inhibition of MYC-amplified medulloblastoma.
Clin. Cancer Res. 20, 912–925 (2014).
57. Rutkowski. S. etal. Survival and prognostic
factors of early childhood medulloblastoma:
aninternational meta-analysis. J. Clin. Oncol. 28,
4961–4968 (2010).
58. Ashley, D.M. etal. Induction chemotherapy
andconformal radiation therapy for very young
children with nonmetastatic medulloblastoma:
Children’s Oncology Group study P9934. J. Clin.
Oncol. 30, 3181–3186 (2012).
59. Chi, S.N. etal. Feasibility and response to
induction chemotherapy intensified with high-
dose methotrexate for young children with newly
diagnosed high-risk disseminated
medulloblastoma. J. Clin. Oncol. 22, 4881–4887
(2004).
60. Kool, M., Korshunov, A. & Pfister, S.M. Update
on molecular and genetic alterations in adult
medulloblastoma. Memo 5, 228–232 (2012).
61. US National Library of Medicine. ClinicalTrials.gov
[online], http://www.clinicaltrials.gov/ct2/
show/NCT02017964?term=NCT02017964&r
ank=1 (2014).
62. US National Library of Medicine. ClinicalTrials.gov
[online], http://www.clinicaltrials.gov/ct2/
show/NCT01878617?term=SJMB12&rank=1
(2014).
63. US National Library of Medicine. ClinicalTrials.gov
[online], http://www.clinicaltrials.gov/ct2/
show/NCT00602667?term=NCT00602667&r
ank=1 (2014).
64. US National Library of Medicine. ClinicalTrials.gov
[online], http://www.clinicaltrials.gov/ct2/
show/NCT02066220?term=NCT02066220&r
ank=1 (2014).
Acknowledgements
The work of the authors is supported, in part, by
Cancer Centre CORE Grant CA 21765, the Noyes
Brain Tumour Foundation, Musicians Against
Childhood Cancer (MACC), and the American
Lebanese Syrian Associated Charities (ALSAC).
Author contributions
Both authors researched the data, contributed
todiscussions of content, wrote the article, and
reviewed/edited the manuscript before submission.
REVIEWS
© 2014 Macmillan Publishers Limited. All rights reserved