104

Ependymoma
Elizabeth R. Gerstner, MD1 Kristian W. Pajtler, MD2,3,4

1 Division of Neuro-oncology, Massachusetts General Hospital Cancer Address for correspondence Elizabeth R. Gerstner, MD,
Center, Boston, Massachusetts Massachusetts General Hospital Cancer Center, 55 Fruit St, Boston,
2 Hopp-Children’s Cancer Center, NCT Heidelberg (KiTZ), Heidelberg, MA 02114 (e-mail: [email protected]).
Germany
3 Division of Pediatric Neurooncology, German Cancer Research
Center (DKFZ), German Cancer Consortium (DKTK),
Heidelberg, Germany
4 Department of Pediatric Hematology and Oncology, Heidelberg
University Hospital, Heidelberg, Germany

Semin Neurol 2018;38:104–111.

Abstract Ependymoma can arise throughout the whole neuraxis. In children, tumors predomi-
nantly occur intracranially, whereas the spine is the most prevalent location in adults.
Significant variance in the grade II versus grade III distinction of ependymomas has led

Downloaded by: Iowa State University. Copyrighted material.

to the acknowledgment that the clinical utility of histopathological classification is
limited. Epigenomic profiling efforts have identified molecularly distinct groups of
ependymomas that adequately reflect the biological, clinical, and histopathological
heterogeneities across anatomical compartments, age groups, and grades. The recent
Keywords update of the World Health Organization classification of central nervous system
► ependymoma tumors has already integrated one of these groups, and molecular classification will be
► molecular part of future clinical trials to improve risk stratification. Clinical management of this
classification rare disease is challenging, making professional experience and intensified multi-
► treatment disciplinary cooperation pivotal factors for treatment success. Novel research strate-
► adult gies are currently applied for target discovery in ependymomas since for most
► children molecular groups, genetic drivers are unknown.

Histopathological Classification
Ependymoma in Children and Adolescents
Subependymomas (SEs; World Health Organization [WHO]
Epidemiology grade I) are mostly located intracranially, with the majority of
Although ependymomas can arise at any time during child- tumors being diagnosed in the fourth or the lateral ventricles.
hood, the highest frequency of occurrence is between 0 and Occasionally, SEs are also diagnosed in the SP.6 Occurrence of
4 years of age.1 While the majority of tumors in adult SEs in children appears to be very rare. SEs are clearly
patients are located in the spine (SP; 46%), the brain is the demarcated from the surrounding brain parenchyma without
predominant location in children and adolescents, with 90% infiltrating growth. Microscopically, there is a typical nodular
of tumors located intracranially.2 Two-thirds of childhood pattern with zones of increased nuclear density or fibrillary
intracranial ependymomas are diagnosed in the posterior stroma. Pseudorosettes or cysts might be observed. Abundant
fossa (PF).3 Ependymomas account for 10% of all brain proliferative activity is not a typical feature of SEs.7 Myxopa-
tumors in children and have a male-to-female ratio of pillary ependymomas (WHO grade I), another rare histopatho-
1.77:1.4 Few familial cancer syndromes have been associated logical subtype, predominantly arises in the lower spinal cord.
with ependymomas, for example,, neurofibromatosis type 2 In children, myxopapillary ependymomas may occasionally
(NF2), which is responsible for an increased incidence of also arise in the soft tissue of the sacrococcygeal region and are
spinal ependymomas.5 associated with more indolent behavior compared with spinal

Issue Theme Neuro-Oncology; Guest Copyright © 2018 by Thieme Medical DOI https://doi.org/
Editors, Tracy Batchelor, MD, and Publishers, Inc., 333 Seventh Avenue, 10.1055/s-0038-1636503.
Dr. med. Wolfgang Wick New York, NY 10001, USA. ISSN 0271-8235.
Tel: +1(212) 584-4662.
 Ependymoma Gerstner, Pajtler 105

cord tumors.8 Typical microscopic appearance includes pseu- are predominantly diagnosed in infants and young children
dopapillary structures, mucoid material surrounding tumor and are associated with high recurrence rates and poor
cells, and accumulation of collagen. Although histopatho- clinical outcome.17–19 In contrast, PF-EPN-B tumors are
logically classified as grade I, myxopapillary ependymomas mainly found in adolescents and young adults and are
might show metastatic potential in pediatric patients.9 associated with a better prognosis. Notably, it has recently
Classic ependymomas (grade II) comprise different histologi- been shown that identification of epigenome-wide reduc-
cally defined variants, designated papillary, clear cell, or tion in H3K27me3 is a robust and cost-effective molecular
tanycytic ependymoma. Histological characteristics of grade surrogate for PF-EPN-A, despite H3.3K27 mutations only
II tumors are perivascular pseudorosettes as well as the occurring very rarely in these tumors.20–22 The majority of
pathognomonic true ependymal rosettes.10 Anaplastic epen- ST ependymomas (more than 70%) are characterized by
dymomas (grade III) are characterized by hypercellularity with fusions between a gene with unknown function, C11orf95,
abundant mitotic activity, pseudopalisading necrosis, and and the nuclear factor kappa B effector RELA, resulting from
microvascular proliferation. The histopathology-based classi- massive chromosomal rearrangement events called chromo-
fication has only limited clinical utility (further elaborated in thripsis. The respective molecular group was designated ST-
the following), and the 2016 update of the WHO classification EPN-RELA.12,16 ST-EPN-RELA tumors occur in both children
of tumors of the central nervous system (CNS), for the first and adults, whereas the remaining molecular subgroups of
time, recognized a recently discovered novel molecular group ST ependymomas, ST-EPN-YAP1, are characterized by recur-
of ependymomas, “ependymoma, RELA fusion-positive.”10,11 rent fusions to the oncogene YAP1 and are mainly diagnosed
“Ependymoma, RELA fusion-positive” defines a distinct subset in childhood.12,16 Multivariate survival analyses demon-
of supratentorial (ST) grade II and grade III ependymomas in strated that molecular subgrouping is superior to the histo-

Downloaded by: Iowa State University. Copyrighted material.

children and young adults (see the following for a detailed pathological grading of ependymal tumors.
description of this molecular group). Positive staining of
L1CAM serves as a surrogate marker to histologically diagnose Diagnostics
these tumors.10,12 There is no formal staging system for ependymomas. However,
the evaluation of the primary tumor, the size of the potential
Molecular Classification residual tumor after surgery, and examination of the cere-
There is a wide variance in the grade II versus grade III brospinal fluid (CSF) are crucial for risk stratification and
distinction of ependymomas, even between experienced therapy planning. One of the key issues is comparability of
neuropathologists, and the prognostic ability of histopatho- pre- and postoperative imaging. The initial work-up should
logical grading criteria to risk-stratify patients is both incon- include a preoperative spinal magnetic resonance imaging
clusive and contentious.11,13 Deoxyribonucleic acid (DNA) (MRI) whenever possible since nonspecific subdural enhance-
methylation patterns in tumors have been shown to repre- ment may be identified on postoperative imaging that might be
sent a very stable molecular memory of the respective cell of misinterpreted as meningeal dissemination. Nonspecific intra-
origin throughout the disease course, making them particu- cranial enhancement is often seen 3 days following surgery,
larly suitable for tumor classification purposes.14,15 A robust and postoperative MRI should be performed preferably 1 or
and uniform (epi)genetic classification system of ependy- 2 days after surgery has been completed.23 Imaging of the
momas may guide researchers, neuropathologists, and clin- complete craniospinal axis at the time of diagnosis is also
icians to a better understanding of the heterogeneity of this indicated for spinal ependymomas.9,24 Due to diagnostic chal-
disease, ultimately resulting in more precise diagnostics, lenges associated with imaging and histopathological diagnos-
better prognostication, and improved therapy. Through close tics (see above), central radiological and histological review
international collaboration, a recent study has established a should be part of clinical trials enrolling patients with epen-
molecular classification scheme of pediatric and adult epen- dymoma.25 In addition, molecular subgrouping should be
dymomas based on DNA methylation data of a large collec- performed in all patients, especially in future clinical trials.
tion (n ¼ 500) of patient-derived primary, treatment naïve To further improve molecular diagnostics and identify new
tumor samples.16 This study demonstrated the existence of prognostic factors and therapeutic targets, optimal tissue
nine clinically, demographically, and molecularly distinct material for ongoing and future biological discovery studies
ependymoma entities, with three occurring in each anato- is required. Especially for rare entities such as ependymomas,
mical compartment of the CNS, SP, PF, and ST region. One of efforts should be made to establish standard operating proce-
the molecular subgroups within each compartment was dures to enable reliable collection and submission of frozen
enriched with WHO grade I SEs, named ST-SE, PF-SE, and tissue and blood. Given the rapid developments in the field of
SP-SE, occurring in adults only (►Fig. 1). The two other molecular genetics and the emergence of increasingly power-
molecular subgroups within the SP predominantly matched ful analytical tools, these collections paired with high-quality
the histopathology based diagnoses, named myxopapillary clinical data will contribute to improve the care of future
ependymoma (SP-MPE) and WHO grade II/III ependymoma ependymoma patients.
(SP-EPN). The remaining two molecular types of ependymo-
mas occurred in the PF, termed PF-EPN-A and PF-EPN-B, or Treatment Options for Intracranial Ependymomas
alternatively PF group A and B, and were independently Clinical management, especially of intracranial ependymomas
identified in retrospective studies.17,18 PF-EPN-A tumors (WHO grade II/III), is challenging, and the optimal treatment

Seminars in Neurology Vol. 38 No. 1/2018

106 Ependymoma Gerstner, Pajtler

Molecular Subgroups of Ependymal Tumors

Anatomic
Compartment
SPINE (SP-) Posterior Fossa (PF-) Supratentorial (ST-)

Molecular
SE MPE EPN SE EPN-A EPN-B SE EPN-YAP1 EPN-RELA
Subgroup

sub- myxopapillary (anaplastic) sub- (anaplastic) (anaplastic) sub- (anaplastic) (anaplastic)

Histopathology ependymoma ependymoma ependymoma ependymoma ependymoma ependymoma ependymoma ependymoma ependymoma
(WHO I) (WHO I) (WHO II/III) (WHO I) (WHO II/III) (WHO II/III) (WHO I) (WHO II/III) (WHO II/III)

6q del. CIN CIN balanced balanced CIN balanced aberr. 11q aberr. 11q
Genetics
Chromo-
Oncogenic thripsis
? ? NF2 ? ? ? ? YAP1-fusion RELA-fusion
Driver

Tumor
Location

Downloaded by: Iowa State University. Copyrighted material.

Age
Distribution
(years)
4 18 60 4 18 60 4 18 60 4 18 60 4 18 60 4 18 60 4 18 60 4 18 60 4 18 60

Gender
Distribution

Patient
Survival
(OS; months)
120 120 120 120 120 120 120 120 120

Fig. 1 Molecular subgroups of ependymal tumors. (Reprinted from Pajtler KW, Witt H, Sill M, et al. Molecular classification of ependymal tumors across all
CNS compartments, histopathological grades, and age groups. Cancer Cell 2015;27(5):728–743, Copyright (2015), with permission from Elsevier).

strategy is contentious. Given these constraints, treatment of with locally invasive growth patterns but low metastatic
affected patients necessitates advanced levels of experience potential (<10% at the time of first diagnosis). Neurosurgical
and extensive cooperation between all specialties involved. intervention plays a primary role for local tumor control, and
Pediatric patients with ependymomas have high rates of the extent of neurosurgical resection has been the most
mortality since ependymomas challenge all currently applied consistent independent prognostic factor reported in numer-
forms of therapy. Treatment plans vary between different ous studies.29,30 Since patients without residual disease have
countries, and sometimes even between treatment centers significantly better outcomes with regard to both event-free
within the same country. Although expert opinions on certain survival (EFS) and overall survival compared with patients
aspects of ependymoma treatment still differ, mainly relating with incomplete resection, today’s therapeutic concepts
to the application and relevance of chemotherapy, there is a include aggressive debulking strategies. In case of incomplete
general consensus that surgery and radiotherapy (RT) are the resection, second-look surgery may be performed immedi-
current therapeutic mainstays.25 The role of chemotherapy in ately following the first intervention or after a short course of
the management of ependymomas remains unproven. Retro- induction chemotherapy. In addition to neurosurgical resec-
spective analyses of pediatric and adult cohorts, in which tion, postoperative involved field RT is considered the standard
chemotherapy was part of the treatment of ependymoma of care for patients with nondisseminated ependymomas.31,32
patients, either failed to demonstrate survival advantages Tumor control rates are superior in patients treated with high-
linked to chemotherapy or showed substantial variation dose RT, and current concepts suggest 59.4 Gy for the volume
between individual patients.26–28 In children, the two inter- at highest risk for local tumor recurrence. To date, only a few
national randomized trials described in the following, Chil- children have been treated by proton beam therapy on proto-
dren’s Oncology Group (COG) ACNS0831 and International col, but owing to the challenging localization of ependymoma,
Society for Pediatric Oncology (SIOP) Ependymoma II, particularly in the case of laterally located infant PF tumors,
are currently comparing postirradiation chemotherapy to this treatment modality might be further explored in the near
observation only. Intracranial ependymomas mainly present future to spare proximal neurologic structures.31,33 Surgery

Seminars in Neurology Vol. 38 No. 1/2018

 Ependymoma Gerstner, Pajtler 107

and RT are principal components in current international some patients. Main therapeutic options at this stage are
treatment optimization trials from both the COG and the again surgery and RT.42,43 However, in contrast to only low or
SIOP. The current COG ACNS0831 study is a randomized phase absent cognitive impairments after RT following initial diag-
III trial with the primary aim to compare the 5-year EFS for nosis, re-irradiation was associated with a decline in patient
patients randomly assigned to adjuvant (maintenance) che- intellectual function.42 In cases of local tumor recurrence,
motherapy or observation following RT (ClinicalTrials.gov; the decision whether focal or craniospinal RT should be
NCT01096368). The SIOP Ependymoma II study (Clinical- administered is often based on the patient’s age. In a current
Trials.gov; NCT02265770) is a multiarm phase II and III trial. St. Jude trial, all patients over 3 years of age receive cra-
Strata comprise a randomized phase III trial to evaluate the niospinal irradiation to reduce the risk of metastatic disease
efficacy of postirradiation maintenance chemotherapy for (ClinicalTrials.gov; NCT02125786).
patients with no evidence of residual tumor, a randomized
frontline phase II chemotherapy and exploration of supple- Clinical Management of Intracranial Ependymomas in
mental RT after second surgery for patients with residual the Context of Molecular Subgroups
disease, and a randomized phase II chemotherapy study, Molecular subgrouping of ependymomas has not yet been
including histone deacetylase inhibitors, for patients younger integrated in any of the current clinical trials, but there is clear
than 12 months or patients who are not eligible for RT. Since consensus that it should be part of all such studies hence-
serial testing of patients enrolled in a large St. Jude study, forth.25 It is recognized that ependymomas from different
which included patients as young as 12 months who were compartments of the CNS are biologically distinct and that
treated by aggressive surgery and immediate postoperative there are phenotypically divergent subgroups within each
high-dose photon RT, did not provide evidence for cognitive anatomical compartment. Accounting for these differences

Downloaded by: Iowa State University. Copyrighted material.

impairments in the first 5 years after irradiation,34–37 both the has the potential to improve future risk stratification, leading
COG and SIOP studies apply RT to patients starting from to administration of aggressive treatment regimens in high-
12 months of age. risk patients and prioritizing low-risk patients for de-escala-
tion strategies. Outside of clinical trials, in patients older than
Therapy Options for Spinal Ependymomas 12 months of age with PF-EPN-A tumors, maximal safe resec-
Spinal cord ependymomas are very rare among children and tion and focal RT should be the standard of care.25 Based on
adolescents. They tend to cause back pain, lower extremity evidence from a large retrospective study (n ¼ 820) demon-
weakness, and/or bowel and bladder dysfunction. Among strating that many patients with completely resected PF-EPN-B
these tumors, the histological subtype myxopapillary ependy- tumors did not have a relapse of their disease,44 even in the
moma is more frequent than WHO grade II or III tumors in this absence of RT, therapy de-escalation strategies should be
age group. Due to the rarity of this disease, current knowledge investigated for this group in the framework of clinical trials.25
regarding treatment strategies is based on case reports and There is not enough evidence for distinct treatment recom-
few retrospective case series.38 Standard treatment options for mendations in ST-EPN-RELA and ST-EPN-YAP1 tumors; how-
newly diagnosed spinal ependymomas include surgery with or ever, molecularly informed prospective as well as retrospective
without adjuvant RT. Although the impact of adjuvant treat- analyses of international trial cohorts are expected to pinpoint
ment on progression and survival is still unclear, the decision promising therapeutic strategies soon.
for nonsurgical treatment is currently primarily based on the
WHO grade of the tumor and the extent of the resection. RT is Surveillance
mostly avoided in children with WHO grade II spinal cord Surveillance of patients following therapy should include
ependymomas, irrespective of the extent of resection, but is frequent neuroimaging coupled with clinical assessments.
often administered in anaplastic spinal cord ependymomas as The frequency and duration have been arbitrarily determined,
well as in cases in which full resection has been achieved. and the utility is uncertain.45 Most studies or surveillance
Historically, therapeutic management of myxopapillary epen- protocols schedule MRI of the brain every 3 months for the first
dymoma included an attempt for gross total resection without 2 years after treatment. From the third year, imaging every
further treatment in case of successful intervention.39,40 6 months for the next 3 years is often undertaken. Since
Although overall survival rates are favorable,16,38 relapse relapses of spinal ependymomas may even occur decades after
incidence is high, particularly in myxopapillary ependymo- initial diagnosis, lifelong follow-up for affected children should
mas. This might legitimize general adjuvant RT for myxopa- be considered.9,46 In addition, surveillance protocols often
pillary ependymomas. However, optimal treatment strategies determine regular assessments of neurocognitive function
have to be further analyzed in larger series for all subtypes of and quality of life (QOL), but time points vary between studies.
spinal ependymomas in children. Since secondary malignancies might occur in the irradiation
field, histomolecular reassessment should be considered at the
Therapy Options at Relapse time of apparent recurrence in order not to miss development
Recurrence is not uncommon, especially for molecular of novel neoplasms, for example, glioblastoma.
groups PF-EPN-A and ST-EPN-REAL, and may even develop
many years after initial treatment.7,41 The tumor mostly Prognostics Factors
recurs locally at the primary tumor site. However, concomi- Apart from the extent of resection and age, overall survival is
tant dissemination along the neuraxis may also be seen in influenced by factors such as gain of chromosomal arm 1q,

Seminars in Neurology Vol. 38 No. 1/2018

108 Ependymoma Gerstner, Pajtler

telomerase reactivation, tumor location gender, and mole- Histopathological and Molecular and Features
cular subgrouping.16–18,47,48 While PF-EPN-A and ST-EPN- The histological features of ependymomas in adults mimic
RELA are associated with overall poor prognosis, PF-EPN-B, that of children, as outlined previously, but a variety of
ST-EPN-YAP1, molecular SE (PF-SE/ST-SE/SP-SE), and spinal molecular abnormalities have been reported specifically in
molecular groups (SP-MPE, SP-EPN) are mainly associated adults with ependymomas. TERT promoter mutations are
with favorable prognosis.16–18 associated with older age and intracranial location, whereas
TERT mutations are not seen in pediatric ependymomas.50,51
Future Directions Chromosome 1q gain, epidermal growth factor receptor
Molecular classification is expected to significantly support (EGFR) expression, and nucleolin expression have been
risk stratification strategies in the near future. The close reported in subsets of adult ependymomas as well as in
interconnection of classic tissue-based information with children, but the clinical significance of this information is
more precise diagnostic aids gained from novel, unbiased, unknown, although there are drugs that can target EGFR,
and observer-independent molecular analyses as well as such as lapatinib, leading to exploration of lapatinib in a
data from imaging methods, including central pathology and clinical trial for adult ependymoma patients.52,53
radiological review, might finally converge at a refined diag-
nostic setting and improve treatment strategies for ependy- Presentation and Diagnosis
momas. Genomic sequencing has driven precision-based Like most CNS tumors, presentation depends on the location
oncology therapy; however, genetic drivers remain unknown of the tumor and can include global symptoms related to
or nontargetable for many molecular groups of ependymomas. obstructive hydrocephalus or focal neurologic symptoms
Thus, alternative approaches are needed to identify therapeu- from local mass effect. MRI will reveal an enhancing mass

Downloaded by: Iowa State University. Copyrighted material.

tic leads. A currently ongoing multicenter collaboration is that is hypointense and hyperintense on T1- and T2-
testing the hypothesis that the more frequent and aggressive weighted sequences, respectively. Cysts and calcifications
group of PF-EPN-A tumors show added molecular heteroge- are often seen and can suggest the diagnosis of ependy-
neity with clinical utility. The study comprises more than 600 moma. Within the spinal cord, the tumors tend to occur
PF-EPN-A tumors at primary diagnosis, which are analyzed for centrally, resulting in an associated syrinx from blockage of
subgroups with distinctive clinical and genetic associations the central canal. Alternative diagnoses depend on the loca-
including potential previously unappreciated recurrent tion of the tumor but can include medulloblastoma (if in the
events. Another large international study is mapping the active PF), high-grade gliomas, meningiomas, or metastasis.
chromatin landscapes of primary ependymomas to identify Diagnostic work-up should include imaging (preferably
regulatory enhancer elements that may pinpoint putative MRI) of the entire neuraxis and a lumbar puncture (if safe) to
oncogenes, molecular targets, and pathways. Resulting frame- look for malignant cells since ependymomas can disseminate
works will be used for target and drug discovery in ependy- through the CSF. Disseminated disease requires craniospinal
momas despite the lack of known genetic drivers. Based on radiation; therefore, it is important to identify.
recent work, restoration of p53 function in ST-EPN-RELA is
currently being tested in a preclinical setting as a potential Treatment
therapeutic approach.49 Novel therapeutic strategies are cur- The mainstay of treatment for ependymomas is maximal safe
rently being explored at time of relapse. These include vacci- surgical resection since the extent of resection is an impor-
nation approaches (ClinicalTrials.gov; NCT01795313 and tant prognostic factor in multiple studies.54,55 For patients
NCT02722512) and intraventricular administration of a hypo- with low-grade ependymomas (WHO grade II), limited dis-
methylating agent (5-azacytidine) in patients with relapsed PF ease, and a gross total resection, observation is reasonable,
ependymomas (ClinicalTrials.gov; NCT02940483). although focal radiation to the surgical site can also be
considered. With WHO grade III ependymomas, radiation
is typically recommended even following gross total resec-
Ependymomas in Adults
tion, although the data derived from retrospective studies
Epidemiology are conflicting about the added benefit of radiation, and
Ependymomas are much less common in adults, represent- there are no prospective trials.54–56
ing 1.9% of all adult primary brain tumors, and include SEs
(WHO grade I), myxopapillary ependymomas (WHO grade I), Survival
and classic ependymomas (WHO grade II or III). The peak age Overall survival estimates vary depending on the study, with
at presentation is the fourth decade, and, as opposed to the the five-year survival rate reported to be 67 to 85% and the
pediatric population, the spinal cord (rather than intracra- 10-year survival to be 50 to 77%.54,55,57 Progression-free
nial) is the most common location. There are no clear risk survival (PFS) at 5 years varies from 43 to 64%, and PFS at
factors for ependymomas in adults, with the exception, like 10 years is 24 to 53%.54,55,57 Specific factors associated with
in children, that patients with NF2 are at increased risk for worse survival in large retrospective series include higher
developing ependymomas, particularly spinal ependymo- histological grade, incomplete resection, ST location, and
mas. NF2-related ependymomas have a slower growth rate Karnofsky performance scale  80.2,28,54,55,57
and more benign behavior and therefore are often treated As expected with their WHO grade I histology, myxo-
less aggressively than sporadic ependymomas. papillary ependymomas have a better prognosis with an

Seminars in Neurology Vol. 38 No. 1/2018

 Ependymoma Gerstner, Pajtler 109

estimated 10-year PFS of 61.2% and 10-year survival of 92%.58 months.66 As with other bevacizumab studies in glioma,
Most of these patients are treated with surgery alone if a the radiographic response rate was high at 75%. Since bev-
gross total resection has been achieved. Radiation is typically acizumab has shown efficacy in NF2-related schwannoma,
reserved for subtotal resections or recurrent disease. two recent studies explored the drug’s impact on ependy-
momas in patients with NF2. NF2 patients with spinal
Quality of Life ependymomas who were receiving bevacizumab for growing
Several studies have evaluated QOL in ependymoma patients schwannomas were found to have a decrease in the size of
but have shown conflicting results with both decreases in their ependymomas as well as associated syrinx.67 In NF2
QOL scores following surgery alone and improvement or at patients with symptomatic ependymomas treated with bev-
least stabilization in QOL scores after surgery for spinal acizumab, improvement in symptoms occurred, albeit with-
ependymomas.59–62 out a corresponding radiographic response.51 These authors
also demonstrated that all eight ependymomas studied
Treatment at Recurrence histologically expressed VEGF receptor 2, a prime mediator
With recurrent adult ependymomas, surgery and radiation of VEGF activity. Currently, CERN has a phase II trial of
are considered based on the location, prior treatments, and bevacizumab with carboplatin in patients with recurrent
functional status of the patient. At least one small study of re- ependymoma (NCT01295944). The trial is still enrolling and
irradiation (stereotactic radiosurgery or fractionated stereo- therefore no results are available.
tactic radiosurgery) did not increase toxicity but resulted in
good local control.63 Future Directions
Other potential treatment options in adults include 5-fluor-

Downloaded by: Iowa State University. Copyrighted material.

Cytotoxic Chemotherapy ouracil, based on a phase I study in children and young
If surgery or radiation is not possible, chemotherapy can be adults (22 years) showing some antitumor activity and
considered, but there are limited data on successful regimens tolerability.68
as most reports are based on retrospective, heterogeneous
series or pediatric data. The Gruppo Italiano Cooperativo di Subependymomas
Neuro-Oncologia retrospectively looked at progressive epen- SEs are rare, benign intraventricular tumors that are typically
dymoma patients previously treated with surgery and RT found incidentally. On MRI, SEs appear as a nonenhancing,
who were then treated with cisplatin- or a noncisplatin- well-demarcated, nodular intraventricular mass that is iso-
based regimen and found that patients treated with cisplatin intense on T1-weighted images and hyperintense on T2-
achieved a higher response rate but there was no progres- weighted images, although occasionally enhancement can
sion-free or overall survival benefit.64 In another retrospec- be seen. Under the microscope, small, uniform nuclei form
tive series of 11 grade II or III recurrent ependymomas, clusters within a fibrillary and cystic background.10 Prolif-
temozolomide, procarbazine/CCNU (lomustine)/vincristine, eration index is characteristically low. Treatment for growing
platinum-based chemotherapy, and epirubicin/ifosfamide SEs is maximal safe resection, and typically additional ther-
were all reported regimens. For the entire cohort of che- apy is not needed. Rarely, SEs can recur and disseminate
motherapy-treated patients, the percentage of patients pro- through CSF; therefore, involved-field radiation may be
gression-free at 6 and 24 months was 52.9 and 23.5%, indicated. Chemotherapy has a limited role in the manage-
respectively. The percentage of patients alive at 6 and ment of SEs.
24 months was 82.4 and 70.1%, respectively, but the retro-
spective approach and small numbers prevent any conclu-
Conclusions
sions favoring one regimen over another.27
Temozolomide has been evaluated in a retrospective Ependymomas are increasingly recognized as a collection of
study of grade II and III ependymomas, and median PFS tumors with distinct molecular makeup, which makes study-
and median OS were reported as 9.69 and 30.55 months, ing these tumors difficult. It is widely acknowledged that
respectively.65 In two small studies, there was no correlation grading (II vs. III) of intracranial ependymomas as a basis for
with MGMT methylation status and response to temozolo- treatment decisions should be used with great cautions
mide .27,65 The Collaborative Ependymoma Research Net- outside of clinical trials. Standard of care for patients with
work (CERN) has completed an uncontrolled phase II trial in PF-EPN-A ependymomas (older than 12 months of age)
unselected, recurrent ependymoma of temozolomide and consists of maximal safe surgical resection followed by
lapatinib (to target the known expression of EGFR), but the adjuvant local RT but probably does not include the routine
final results have not been published (NCT00826241). use of chemotherapy outside the setting of a clinical trial. A
subset of PF-EPN-B ependymoma patients for whom gross
Bevacizumab total resection could be achieved might be cured by surgery
Bevacizumab, a monoclonal antibody targeting vascular only. A clinical trial to test the possibility of avoiding RT in the
endothelial growth factor (VEGF), has been explored in context of complete resection for PF-EPN-B patients is indi-
several small studies. One retrospective study using bevaci- cated. Molecular analyses of ependymomas from clinically
zumab in recurrent ependymoma reported the median time well-annotated international multicenter trial cohorts as
to progression as 6.4 months, and median OS as 9.4 well as from large retrospective cohorts with long-term

Seminars in Neurology Vol. 38 No. 1/2018

110 Ependymoma Gerstner, Pajtler

follow-up are of great importance to clarify questions about 19 Mack SC, Witt H, Piro RM, et al. Epigenomic alterations define
the clinical outcome of the molecular variants of ependy- lethal CIMP-positive ependymomas of infancy. Nature 2014;506
moma and to deduce explicit molecularly informed therapy (7489):445–450
20 Panwalkar P, Clark J, Ramaswamy V, et al. Immunohistochemical
recommendations. Especially, molecular subgroups of ST
analysis of H3K27me3 demonstrates global reduction in group-
ependymomas require additional study before specific treat- A childhood posterior fossa ependymoma and is a powerful
ment recommendations can be made. Increasing success predictor of outcome. Acta Neuropathol 2017;134(05):705–714
rates in treating patients affected by this uncommon and 21 Gessi M, Capper D, Sahm F, et al. Evidence of H3 K27M mutations
heterogeneous entity as well as advancing research in this in posterior fossa ependymomas. Acta Neuropathol 2016;132
(04):635–637
field will necessitate great international cooperation and
22 Ryall S, Guzman M, Elbabaa SK, et al. H3 K27M mutations are
close collaboration of the pediatric and adult neuro-oncol-
extremely rare in posterior fossa group A ependymoma. Childs
ogy communities. Nerv Syst 2017;33(07):1047–1051
23 Warmuth-Metz M. Imaging and Diagnosis in Pediatric Brain
Tumor Studies. Cham, Switzerland: Springer; 2017
References 24 Bagley CA, Kothbauer KF, Wilson S, Bookland MJ, Epstein FJ, Jallo
1 Villano JL, Parker CK, Dolecek TA. Descriptive epidemiology of GI. Resection of myxopapillary ependymomas in children.
ependymal tumours in the United States. Br J Cancer 2013;108 J Neurosurg 2007;106(4, Suppl):261–267
(11):2367–2371 25 Pajtler KW, Mack SC, Ramaswamy V, et al. The current consensus
2 Vera-Bolanos E, Aldape K, Yuan Y, et al; CERN Foundation. Clinical on the clinical management of intracranial ependymoma and its
course and progression-free survival of adult intracranial and distinct molecular variants. Acta Neuropathol 2017;133(01):
spinal ependymoma patients. Neuro-oncol 2015;17(03):440–447 5–12
3 Kilday JP, Rahman R, Dyer S, et al. Pediatric ependymoma: 26 Bouffet E, Foreman N. Chemotherapy for intracranial ependymo-

Downloaded by: Iowa State University. Copyrighted material.

biological perspectives. Mol Cancer Res 2009;7(06):765–786 mas. Childs Nerv Syst 1999;15(10):563–570
4 Purdy E, Johnston DL, Bartels U, et al. Ependymoma in children 27 Gramatzki D, Roth P, Felsberg J, et al. Chemotherapy for intracra-
under the age of 3 years: a report from the Canadian Pediatric nial ependymoma in adults. BMC Cancer 2016;16:287
Brain Tumour Consortium. J Neurooncol 2014;117(02):359–364 28 Nuño M, Yu JJ, Varshneya K, et al. Treatment and survival of
5 Yao Y, Mack SC, Taylor MD. Molecular genetics of ependymoma. supratentorial and posterior fossa ependymomas in adults. J Clin
Chin J Cancer 2011;30(10):669–681 Neurosci 2016;28:24–30
6 Jain A, Amin AG, Jain P, et al. Subependymoma: clinical features 29 Bouffet E, Perilongo G, Canete A, Massimino M. Intracranial
and surgical outcomes. Neurol Res 2012;34(07):677–684 ependymomas in children: a critical review of prognostic factors
7 Wu J, Armstrong TS, Gilbert MR. Biology and management of and a plea for cooperation. Med Pediatr Oncol 1998;30(06):
ependymomas. Neuro-oncol 2016;18(07):902–913 319–329, discussion 329–331
8 Cimino PJ, Agarwal A, Dehner LP. Myxopapillary ependymoma in 30 Cage TA, Clark AJ, Aranda D, et al. A systematic review of
children: a study of 11 cases and a comparison with the adult treatment outcomes in pediatric patients with intracranial epen-
experience. Pediatr Blood Cancer 2014;61(11):1969–1971 dymomas. J Neurosurg Pediatr 2013;11(06):673–681
9 Fassett DR, Pingree J, Kestle JR. The high incidence of tumor 31 Merchant TE. Current clinical challenges in childhood ependy-
dissemination in myxopapillary ependymoma in pediatric moma: a focused review. J Clin Oncol 2017;35(21):2364–2369
patients. Report of five cases and review of the literature. 32 Merchant TE, Li C, Xiong X, Kun LE, Boop FA, Sanford RA. Con-
J Neurosurg 2005;102(1, Suppl):59–64 formal radiotherapy after surgery for paediatric ependymoma: a
10 Louis DN, Ohgaki H, Wiestler OD, et al. WHO Classification of prospective study. Lancet Oncol 2009;10(03):258–266
Tumours of the Central Nervous System. Revised 4th ed. Lyon, 33 Macdonald SM, Sethi R, Lavally B, et al. Proton radiotherapy for
France: International Agency for Research on Cancer; 2016 pediatric central nervous system ependymoma: clinical out-
11 Ellison DW, Kocak M, Figarella-Branger D, et al. Histopathological comes for 70 patients. Neuro-oncol 2013;15(11):1552–1559
grading of pediatric ependymoma: reproducibility and clinical rele- 34 Conklin HM, Li C, Xiong X, Ogg RJ, Merchant TE. Predicting
vance in European trial cohorts. J Negat Results Biomed 2011;10:7 change in academic abilities after conformal radiation therapy
12 Parker M, Mohankumar KM, Punchihewa C, et al. C11orf95-RELA for localized ependymoma. J Clin Oncol 2008;26(24):3965–3970
fusions drive oncogenic NF-κB signalling in ependymoma. Nature 35 Di Pinto M, Conklin HM, Li C, Xiong X, Merchant TE. Investigating
2014;506(7489):451–455 verbal and visual auditory learning after conformal radiation
13 Godfraind C. Classification and controversies in pathology of therapy for childhood ependymoma. Int J Radiat Oncol Biol
ependymomas. Childs Nerv Syst 2009;25(10):1185–1193 Phys 2010;77(04):1002–1008
14 Hovestadt V, Jones DT, Picelli S, et al. Decoding the regulatory 36 Netson KL, Conklin HM, Wu S, Xiong X, Merchant TE. A 5-year
landscape of medulloblastoma using DNA methylation sequen- investigation of children’s adaptive functioning following con-
cing. Nature 2014;510(7506):537–541 formal radiation therapy for localized ependymoma. Int J Radiat
15 Hoadley KA, Yau C, Wolf DM, et al; Cancer Genome Atlas Research Oncol Biol Phys 2012;84(01):217–223.e1
Network. Multiplatform analysis of 12 cancer types reveals 37 Willard VW, Conklin HM, Boop FA, Wu S, Merchant TE. Emotional
molecular classification within and across tissues of origin. Cell and behavioral functioning after conformal radiation therapy for
2014;158(04):929–944 pediatric ependymoma. Int J Radiat Oncol Biol Phys 2014;88(04):
16 Pajtler KW, Witt H, Sill M, et al. Molecular classification of 814–821
ependymal tumors across all CNS compartments, histopatholo- 38 Benesch M, Frappaz D, Massimino M. Spinal cord ependymomas
gical grades, and age groups. Cancer Cell 2015;27(05):728–743 in children and adolescents. Childs Nerv Syst 2012;28(12):2017–
17 Wani K, Armstrong TS, Vera-Bolanos E, et al; Collaborative 2028
Ependymoma Research Network. A prognostic gene expression 39 Akyurek S, Chang EL, Yu TK, et al. Spinal myxopapillary ependy-
signature in infratentorial ependymoma. Acta Neuropathol 2012; moma outcomes in patients treated with surgery and radio-
123(05):727–738 therapy at M.D. Anderson Cancer Center. J Neurooncol 2006;80
18 Witt H, Mack SC, Ryzhova M, et al. Delineation of two clinically (02):177–183
and molecularly distinct subgroups of posterior fossa ependy- 40 Bagley CA, Wilson S, Kothbauer KF, Bookland MJ, Epstein F,
moma. Cancer Cell 2011;20(02):143–157 Jallo GI. Long term outcomes following surgical resection of

Seminars in Neurology Vol. 38 No. 1/2018

 Ependymoma Gerstner, Pajtler 111

myxopapillary ependymomas. Neurosurg Rev 2009;32(03): 54 Reni M, Brandes AA, Vavassori V, et al. A multicenter study of the
321–334, discussion 334 prognosis and treatment of adult brain ependymal tumors.
41 Zacharoulis S, Ashley S, Moreno L, Gentet JC, Massimino M, Cancer 2004;100(06):1221–1229
Frappaz D. Treatment and outcome of children with relapsed 55 Metellus P, Barrie M, Figarella-Branger D, et al. Multicentric
ependymoma: a multi-institutional retrospective analysis. Childs French study on adult intracranial ependymomas: prognostic
Nerv Syst 2010;26(07):905–911 factors analysis and therapeutic considerations from a cohort of
42 Bouffet E, Hawkins CE, Ballourah W, et al. Survival benefit 152 patients. Brain 2007;130(Pt 5):1338–1349
for pediatric patients with recurrent ependymoma treated 56 Burkhardt B, Moericke A, Klapper W, et al. Pediatric precursor T
with reirradiation. Int J Radiat Oncol Biol Phys 2012;83(05): lymphoblastic leukemia and lymphoblastic lymphoma: differ-
1541–1548 ences in the common regions with loss of heterozygosity at
43 Merchant TE, Boop FA, Kun LE, Sanford RA. A retrospective study chromosome 6q and their prognostic impact. Leuk Lymphoma
of surgery and reirradiation for recurrent ependymoma. Int J 2008;49(03):451–461
Radiat Oncol Biol Phys 2008;71(01):87–97 57 Armstrong TS, Vera-Bolanos E, Bekele BN, Aldape K, Gilbert MR.
44 Ramaswamy V, Hielscher T, Mack SC, et al. Therapeutic impact of Adult ependymal tumors: prognosis and the M. D. Anderson
cytoreductive surgery and irradiation of posterior fossa ependy- Cancer Center experience. Neuro-oncol 2010;12(08):862–870
moma in the molecular era: a retrospective multicohort analysis. 58 Weber DC, Wang Y, Miller R, et al. Long-term outcome of patients
J Clin Oncol 2016;34(21):2468–2477 with spinal myxopapillary ependymoma: treatment results from
45 Good CD, Wade AM, Hayward RD, et al. Surveillance neuroima- the MD Anderson Cancer Center and institutions from the Rare
ging in childhood intracranial ependymoma: how effective, how Cancer Network. Neuro-oncol 2015;17(04):588–595
often, and for how long? J Neurosurg 2001;94(01):27–32 59 Armstrong TS, Vera-Bolanos E, Gilbert MR. Clinical course of adult
46 Lundar T, Due-Tønnessen BJ, Scheie D, Brandal P. Pediatric spinal patients with ependymoma: results of the Adult Ependymoma
ependymomas: an unpredictable and puzzling disease. Long- Outcomes Project. Cancer 2011;117(22):5133–5141
term follow-up of a single consecutive institutional series of 60 Walbert T, Mendoza TR, Vera-Bolaños E, Acquaye A, Gilbert MR,

Downloaded by: Iowa State University. Copyrighted material.

ten patients. Childs Nerv Syst 2014;30(12):2083–2088 Armstrong TS. Symptoms and socio-economic impact of ependy-
47 Gojo J, Lötsch D, Spiegl-Kreinecker S, et al. Telomerase activation moma on adult patients: results of the Adult Ependymoma Out-
in posterior fossa group A ependymomas is associated with comes Project 2. J Neurooncol 2015;121(02):341–348
dismal prognosis and chromosome 1q gain. Neuro-oncol 2017; 61 Viereck MJ, Ghobrial GM, Beygi S, Harrop JS. Improved patient
19(09):1183–1194 quality of life following intradural extramedullary spinal tumor
48 Korshunov A, Witt H, Hielscher T, et al. Molecular staging of resection. J Neurosurg Spine 2016;25(05):640–645
intracranial ependymoma in children and adults. J Clin Oncol 62 Xiao R, Miller JA, Abdullah KG, Lubelski D, Mroz TE, Benzel EC.
2010;28(19):3182–3190 Quality of life outcomes following resection of adult intramedul-
49 Tzaridis T, Milde T, Pajtler KW, et al. Low-dose Actinomycin-D lary spinal cord tumors. Neurosurgery 2016;78(06):821–828
treatment re-establishes the tumoursuppressive function of P53 63 Murai T, Sato K, Iwabuchi M, et al. Re-irradiation of recurrent
in RELA-positive ependymoma. Oncotarget 2016;7(38):61860– anaplastic ependymoma using radiosurgery or fractionated
61873 stereotactic radiotherapy. Jpn J Radiol 2016;34(03):211–218
50 Brügger F, Dettmer MS, Neuenschwander M, Perren A, Marinoni I, 64 Brandes AA, Cavallo G, Reni M, et al. A multicenter retrospective
Hewer E. TERT promoter mutations but not the alternative study of chemotherapy for recurrent intracranial ependymal
lengthening of telomeres phenotype are present in a subset of tumors in adults by the Gruppo Italiano Cooperativo di Neuro-
ependymomas and are associated with adult onset and progres- Oncologia. Cancer 2005;104(01):143–148
sion to ependymosarcoma. J Neuropathol Exp Neurol 2017;76 65 Rudà R, Bosa C, Magistrello M, et al. Temozolomide as salvage
(01):61–66 treatment for recurrent intracranial ependymomas of the adult: a
51 Barszczyk M, Buczkowicz P, Castelo-Branco P, et al. Telomerase retrospective study. Neuro-oncol 2016;18(02):261–268
inhibition abolishes the tumorigenicity of pediatric ependymoma 66 Green RM, Cloughesy TF, Stupp R, et al. Bevacizumab for recurrent
tumor-initiating cells. Acta Neuropathol 2014;128(06):863–877 ependymoma. Neurology 2009;73(20):1677–1680
52 Chen C, Chen L, Yao Y, Qin Z, Chen H. Nucleolin overexpression is 67 Morris KA, Afridi SK, Evans DG, et al; on behalf of the UK NF2
associated with an unfavorable outcome for ependymoma: a Research Group. The response of spinal cord ependymomas to
multifactorial analysis of 176 patients. J Neurooncol 2016;127 bevacizumab in patients with neurofibromatosis type 2.
(01):43–52 J Neurosurg Spine 2017;26(04):474–482
53 Rajeshwari M, Sharma MC, Kakkar A, et al. Evaluation of chromo- 68 Wright KD, Daryani VM, Turner DC, et al. Phase I study of 5-
some 1q gain in intracranial ependymomas. J Neurooncol 2016; fluorouracil in children and young adults with recurrent epen-
127(02):271–278 dymoma. Neuro-oncol 2015;17(12):1620–1627

Seminars in Neurology Vol. 38 No. 1/2018