INTRODUCTION — Ependymomas are a group of glial tumors that typically arise within or adjacent to the ependymal lining of the ventricular system and are thought to be derived from the radial glial cells in the subventricular zone [1,2]. Ependymomas most commonly occur in the posterior fossa, in contact with the fourth ventricle, or in the intramedullary spinal cord; they occasionally occur in the brain parenchyma outside the posterior fossa, and very rarely outside the central nervous system (CNS).
Ependymomas account for less than 10 percent of tumors arising in the CNS and 25 percent of primary tumors originating in the spinal cord.
The clinical presentation and management of intracranial ependymomas will be reviewed here. Ependymomas arising in the spinal cord are discussed separately. (See "Spinal cord tumors".)
EPIDEMIOLOGY — Intracranial ependymomas have a peak incidence in early childhood but can occur at any age [3]. There is a slight male predominance. The median age at diagnosis is five years, and 25 to 40 percent of patients are less than two years old [4-6]. Intracranial ependymomas are uncommon in adults and mostly occur before the age of 40 years. Risk factors for intracranial ependymoma are not well defined.
Spinal ependymomas are rare in children and more common in adults, with a median age at diagnosis of 30 to 40 years. There is an increased incidence of intramedullary spinal cord ependymoma in patients with NF2-related schwannomatosis (NF2). (See "Spinal cord tumors", section on 'Ependymomas' and "NF2-related schwannomatosis (formerly neurofibromatosis type 2)", section on 'Clinical features'.)
The incidence of subependymoma is not well characterized, and many are encountered incidentally or at the time of autopsy. They mainly occur in middle-aged and older adults and are more common in males [3].
PATHOLOGY
Histology and classification — In the World Health Organization (WHO) classification of central nervous system (CNS) tumors, ependymal tumors are classified based on anatomic location, histology, and molecular features. The 2021 revision of the WHO classification (5th edition, CNS5) recognizes several new entities and subgroups based on molecular genetic features (table 1) [7,8]. (See "Classification and pathologic diagnosis of gliomas, glioneuronal tumors, and neuronal tumors", section on 'Ependymal tumors'.)
●Ependymoma – Ependymomas are classified by anatomic site (supratentorial, posterior fossa, and spinal) as well as histology ("classic ependymoma," or WHO grade 2, and "anaplastic ependymoma," or WHO grade 3) and molecular features (see 'Molecular subgroups' below). These tumors can occur anywhere in the ventricular system or spinal canal; they are most common in the fourth ventricle and spinal cord.
Ependymomas are usually well demarcated with frequent areas of calcification, hemorrhage, and cysts. The tumor cells often form ependymal rosettes, which are diagnostic but not always present. Papillary, clear cell, and tanycytic histologic phenotypes are observed but do not have particular clinical significance.
There is uncertainty regarding the prognostic significance of ependymoma grade, particularly grade 2 versus grade 3 tumors. While some reports have indicated a poorer prognosis for anaplastic ependymoma (grade 3), most notably the Children's Oncology Group (COG) ACNS0121 phase II trial [9], this has not been confirmed in other studies [10], the majority of which were performed retrospectively. Tumor heterogeneity and interobserver variability in grading among neuropathologists may account for the lack of strong correlation with outcomes [11]. Specific genetic alterations and molecular subgroups likely provide better prognostic differentiation and may ultimately replace histologic grading in ependymoma [12-16]. (See 'Molecular subgroups' below.)
●Subependymoma – Subependymomas are rare tumors found in the fourth or lateral ventricles of adults [3]. They have a benign histologic appearance and are classified as WHO grade 1. They consist of a coarse fibrillar matrix with clusters of uniform nuclei and microcysts. Calcifications and hemorrhage can be seen.
●Myxopapillary ependymoma – Myxopapillary ependymomas arise almost exclusively in the conus medullaris and filum terminale of the spinal cord (image 1). Molecular analysis suggests that adult and pediatric tumors differ [17]. As of WHO CNS5, myxopapillary ependymomas are designated WHO grade 2, recognizing that the risk of recurrence is similar to that of conventional spinal ependymomas [7,8,18-21]. (See "Spinal cord tumors", section on 'Myxopapillary ependymoma'.)
Molecular subgroups — Genomic alterations are common in ependymomas and may vary based on anatomic location [14,16,22-34]. A number of molecular subgroups are recognized based on comprehensive tumor genetic sequencing and epigenetic profiling [35-37]. Based on these and other data, the following subgroups have emerged but require prospective evaluation to better understand prognostic and therapeutic implications [38]:
●Posterior fossa ependymoma group A (PF-EPN-A) – PF-EPN-A is a poor prognostic subgroup, found predominantly in infants and young children [39,40]. Tumors exhibit a CpG island methylator phenotype and transcriptional silencing of the polycomb repressive complex 2 (PRC2), leading to repressed expression of differentiation genes [41]. Chromosome 1q gain and possibly 6q loss in PF-EPN-A tumors are prognostic of poor outcomes [9,10,13,14,16,22,26,42]. Such changes are present in approximately 15 to 20 percent of newly diagnosed tumors and up to 60 percent of recurrent tumors [43].
Loss of nuclear expression of trimethylation of the lysine residue at position 27 on the histone protein (H3K27me3) occurs in PF-EPN-A tumor cells [30], a phenotype initially described in H3 K27-altered diffuse midline gliomas [44]. Mechanistic studies have suggested a shared oncogenic program between PF-EPN-A ependymomas and H3 K27-altered midline gliomas to inhibit the PRC2 complex through the action of peptidyl PRC2 inhibitors [45,46]. The absence of nuclear H3K27me3 immunohistochemical (IHC) staining in the tumor cells on histopathology has clinical relevance as an accurate and reliable way to identify PF-EPN-A subgroup tumors by IHC (picture 1) [29,47].
●Posterior fossa ependymoma group B (PF-EPN-B) – PF-EPN-B tumors occur in older children and adults and are associated with a better prognosis [36,39].
●Supratentorial ependymoma with ZFTA fusion (ST-EPN-ZFTA) – A majority (70 to 80 percent) of supratentorial ependymomas harbor an oncogenic fusion between ZFTA (previously known as C11orf95) on chromosome 11 and another gene, most commonly RELA, the principal effector of NF-κB signaling [48,49]. The ZFTA gene fusion is most commonly detected using fluorescence in situ hybridization (FISH) [3]. (See "Classification and pathologic diagnosis of gliomas, glioneuronal tumors, and neuronal tumors", section on 'ZFTA fusion'.)
In retrospective studies, ZFTA fusion has been reported to confer a poor prognosis [48]. This finding has been challenged in prospective studies, however [9,10,50]. Hence, we exercise caution when interpreting and discussing this finding with affected patients and caregivers.
Other prognostic markers in this subgroup are being investigated. In one retrospective study, the presence of a cyclin-dependent kinase inhibitor 2A (CDKN2A) biallelic deletion conferred a very poor prognosis, but this finding needs to be confirmed in larger prospective studies [51].
●Supratentorial ependymoma with YAP1 fusion (ST-EPN-YAP) – YAP1 fusion tumors constitute a small percentage of supratentorial ependymomas and are enriched in infants. This subgroup may have a better prognosis compared with other subgroups of ependymoma [10], although this will need prospective validation.
●Supratentorial ependymoma without ZFTA or YAP1 fusion – A small number of supratentorial ependymomas lack either the ZFTA or YAP fusion [52,53]. These tumors appear to be clinically and molecularly heterogenous, and larger studies are needed to understand the pathogenesis and natural history.
Epigenomic and transcriptomic data suggest that ependymomas are composed of at least nine molecularly, demographically, and clinically distinct disease entities [31]. Still other "omic-based" studies have suggested further heterogeneity within PF-EPN-A [32] and PF-EPN-B [33], while detailed analysis of gene enhancer landscapes of ependymoma provides a framework for drug discovery across and within these molecular subtypes [34].
CLINICAL FEATURES
Presenting signs and symptoms — The clinical presentation of patients with ependymoma depends upon the location of the tumor:
●Increased intracranial pressure – Most patients with posterior fossa ependymomas have evidence of increased intracranial pressure due to obstructive hydrocephalus. As a result, headache, nausea and vomiting, ataxia, vertigo, and papilledema are common at presentation. Cranial nerve palsies are also common, especially involving cranial nerves VI to X. Brainstem invasion may occur.
●Seizures and focal deficits – Seizures or focal neurologic deficits (eg, hemiparesis) are common presenting signs of supratentorial tumors, which exert mass effect and may have surrounding edema.
●Myelopathy and radiculopathy – Spinal cord ependymomas present with deficits due to involvement of ascending or descending nerve tracts or the exiting peripheral nerves. Specific abnormalities are related to the anatomic level of the tumor. (See "Spinal cord tumors".)
●Leptomeningeal disease – Metastatic dissemination of tumor through the cerebrospinal fluid (CSF) is observed in less than 5 percent of patients at the time of initial diagnosis [10,50]. Dissemination can occur with both infratentorial and supratentorial ependymomas. Patients may be minimally symptomatic or have multifocal signs and symptoms referable to cranial nerves or cauda equina disease.
Although historical data suggested a correlation between high-grade histology (classic grade 2 versus anaplastic grade 3) and spinal seeding, more recent data have not confirmed this relationship, and prognosis may depend more on molecular subgroup. (See 'Histology and classification' above.)
Anatomic location — The fourth ventricle is the most common site of intracranial ependymomas, and extension into the subarachnoid space occurs frequently, sometimes with encasement of the medulla and upper cervical cord. Supratentorial lesions can be either intraventricular, typically in the lateral ventricles (image 2), or parenchymal.
The typical locations in which ependymomas arise differ depending upon patient age:
●Children – In children, approximately 90 percent of ependymomas are intracranial, including 75 percent located in the posterior fossa, while the remaining 10 percent arise in the spinal cord [4,5].
●Adults – By contrast, approximately 65 percent of ependymomas in adults are spinal, 25 percent are infratentorial, and 10 percent are supratentorial [54]. (See "Spinal cord tumors", section on 'Ependymomas'.)
DIAGNOSIS — The diagnosis of ependymoma requires histologic confirmation but may be suspected preoperatively based on a combination of imaging characteristics, tumor location, and age of the patient. Because gross total resection is so important for intracranial ependymoma as well as most other tumors on the differential diagnosis in children, the diagnosis of ependymoma is typically obtained at the time of open surgery and resection. Biopsy is indicated in selected patients when there is diagnostic uncertainty or prohibitive surgical risk.
Neuroimaging appearance — On magnetic resonance imaging (MRI), ependymomas have a hypointense appearance on T1 and are hyperintense on T2 or proton density images; gadolinium enhancement is usually prominent (image 3). Extension into the foramen of Luschka is commonly observed. Ependymomas often result in hydrocephalus due to obstruction of the fourth ventricle or supratentorial ventricular system by the growing mass, but peritumoral edema is rare. On diffusion-weighted imaging sequences, ependymomas may demonstrate restricted diffusion.
The computed tomography (CT) appearance of ependymoma is often hyperdense with homogeneous enhancement; cysts and calcifications are common. The presence of calcifications within a tumor located in the fourth ventricle is highly suggestive but not diagnostic of ependymoma.
Subependymomas are nonenhancing, well-demarcated, nodular intraventricular tumors that are isodense on CT. On MRI, they typically appear isointense on T1-weighted images and hyperintense in T2-weighted images [55].
Differential diagnosis — The radiologic differential diagnosis for posterior fossa tumors in children includes medulloblastoma, pilocytic astrocytoma, embryonal tumors, and choroid plexus tumors. Although histopathology is ultimately required to distinguish among these tumors, certain radiographic features can help inform preoperative suspicion for one over another.
For example, ependymomas most often arise from the fourth ventricle and may be accompanied by restricted diffusion on diffusion-weighted images. This feature sometimes differentiates an ependymoma from a pilocytic astrocytoma in the posterior fossa, although it is not sufficiently reliable to guide clinical decisions. In addition, in some patients, ependymoma extends down from the posterior fossa to the cranio-cervical junction, a finding that is rare in other posterior fossa tumors. Posterior fossa ependymomas also often have pronounced extension through the foramina of Luschka, which is less commonly observed in medulloblastoma.
In the supratentorial location, glial tumors, embryonal tumors, and choroid plexus carcinoma or papilloma should be considered.
Extent of disease evaluation — All patients with suspected or confirmed ependymoma should have an MRI of the brain and entire spine with contrast and a lumbar puncture. Up to 10 percent of patients in some studies show evidence of spinal or leptomeningeal seeding, while other studies report a lower incidence of dissemination at diagnosis [10,50,56].
Cerebrospinal fluid (CSF) cytology is important for staging for patients with classic or anaplastic ependymomas. CSF is also important in the staging evaluation of patients with myxopapillary ependymoma if there is concern for disseminated disease on imaging.
Ideally, CSF should be obtained preoperatively from the lumbar site, as CSF obtained postoperatively can be more difficult to interpret due to cellular debris from surgery. However, lumbar puncture is often contraindicated at the time of presentation due to obstructive hydrocephalus. CSF sampling postoperatively should be delayed for 10 to 14 days after surgery to allow for surgical debris to clear [57].
Although uncommon, the presence of disseminated disease significantly affects both treatment and prognosis. One-third of patients with disseminated disease may be identified only on the basis of positive CSF cytology [50,58]. Due to the rarity of dissemination at diagnosis and because ependymoma cells have a cytologic appearance that overlaps with nonneoplastic or reactive cells, we typically suggest a second CSF sample for confirmation in patients with a single positive cytology.
SURGICAL RESECTION — The initial treatment for patients with a suspected ependymoma consists of maximal safe resection.
Extent of resection — Ependymomas are often located in the posterior fossa, in close proximity to cranial nerves and the brainstem, and surgery carries significant risk. Extent of resection is strongly correlated with oncologic outcomes and survival, however, and patients stand to gain significant benefit from an optimal initial resection. Patients should therefore be referred to a specialized center with pediatric oncologic neurosurgical expertise whenever possible.
Although terminology varies, the degree of resection is generally defined as follows:
●Gross total resection – Patients are considered to have a gross total (complete) resection if postoperative MRI shows no residual enhancing or nonenhancing tumor and the surgeon feels that a complete resection was achieved at the time of the operation.
●Near total resection – Near total resection is often used to describe patients who either have a small amount of residual tumor on postoperative MRI (eg, <5 mm) or have no macroscopic tumor visible on postoperative MRI but were noted under the operating microscope to have tumor adherent to a critical structure, such as the brainstem or a cranial nerve, that could not be fully removed safely. For the purposes of treatment and prognosis, these patients are usually considered similarly to those with gross total resection.
●Subtotal resection – Subtotal resection includes the remaining patients with residual tumor on postoperative MRI. Other terms that may be used interchangeably include partial resection and incomplete resection.
There are no randomized trials evaluating the extent of surgery. Observational data have consistently indicated that patients undergoing gross total resection have lower rates of local recurrence and better prognosis for long-term survival than those in whom only a partial resection is possible [59-62]. Local recurrences are often more difficult to treat, and the majority of deaths in prospective studies of intracranial ependymoma are due to recurrent ependymoma. (See 'Recurrent disease' below and 'Prognosis' below.)
Lesions invading the brainstem can be particularly challenging. Patients with incomplete resections do less well than those with gross total resection. As a result, many centers consider preradiation chemotherapy and a second-look operation if there is residual macroscopic disease that is felt to be safely resectable. (See 'Subtotal resection, age 1 to 21 years' below.)
Complications — The most common potential complications of posterior fossa ependymoma resections are:
●Cerebellar ataxia – Patients may have new or increased ataxia after resection of tumors from the posterior fossa. Lateral hemisphere cerebellar injury can manifest with limb dysmetria, and midline cerebellar injury is more likely to cause gait ataxia.
●Lower cranial nerve damage – Cerebellopontine angle lesions can often result in significant lower cranial nerve damage resulting in hemifacial weakness, dysarthria, dysphagia, or hearing loss. At least in infants, recovery of function may follow successful resection.
●Posterior fossa syndrome – Cerebellar mutism, also referred to as posterior fossa syndrome, is a recognized complication of posterior fossa surgery, and emerging data suggest that this is most common when the superior or middle cerebellar peduncles are involved [63-65]. (See "Treatment and prognosis of medulloblastoma", section on 'Posterior fossa syndrome'.)
POSTOPERATIVE THERAPY
Intracranial ependymoma, grade 2 or 3 — Ependymomas are invasive primary brain tumors with the potential to disseminate throughout the central nervous system (CNS). Maximal safe resection, adjuvant radiation therapy (RT), and chemotherapy all play a role, depending on the age of the patient, location of the tumor, and extent of resection (algorithm 1) [37,66,67]. At present, the treatment approach is not based on the molecular subgroup of the tumor.
Because of the complexity involved in these cases, patients with ependymomas should be referred to highly specialized centers whenever possible, particularly for infants with subtotal resections and children with positive cerebrospinal fluid (CSF) cytology.
Gross total resection, age 1 to 21 years — Postoperative management of children >1 year of age who undergo gross or near total resection of an intracranial grade 2 or 3 ependymoma is evolving. RT alone is the standard of care for most such children [66-69]. Although preliminary results from the Children's Oncology Group (COG) ACNS0831 trial suggested a benefit of post-RT maintenance chemotherapy in some patients [70], the data are not yet mature, and final results from the trial are awaited. (See 'Post-RT chemotherapy' below.)
The main exception to postoperative RT in this age group is in patients with supratentorial grade 2 ependymoma who undergo gross total resection. These patients appear to represent a favorable group prognostically and may be considered for observation rather than immediate postoperative RT and chemotherapy. This approach was evaluated in one arm of the completed ACNS0831 cooperative group trial, and the final results are awaited. (See 'Areas of uncertainty and ongoing trials' below.)
Adjuvant focal RT — Following surgery, most children >1 year of age with ependymoma grade 2 or 3 who undergo gross total or near total resection should proceed directly to conformal focal RT.
The amount of normal tissue to be included in the RT treatment volume of posterior fossa ependymoma and supratentorial ependymoma has been a source of controversy in the past. While prophylactic cranial irradiation (PCI) was once used in all patients with ependymoma, data from observational series have now defined focal RT, using contemporary techniques, as the standard of care. Local recurrence is the predominant site of failure, and PCI does not improve survival [4,5,71,72]. More extensive radiation fields are only indicated if there is evidence of tumor dissemination based upon neuroimaging and/or CSF cytology. The reduction in irradiated brain volume is significant in preserving patient quality of life during treatment and long-term, with expected marked decrease in risk of neurocognitive deficits, stroke, and second tumor formation. (See 'Patients with disseminated disease' below.)
Conformal techniques allow irradiation of the tumor bed and an adequate margin of adjacent normal tissue while sparing much of the brain. The current standard dose range to the target for intracranial ependymoma is 54 to 59.4 Gy, and higher doses may be recommended for areas with residual macroscopic disease. The extent of the target volume for focal RT continues to be studied, with shrinking clinical target margins used in the completed ACNS0121 (1 cm) and ACNS0831 (0.5 cm) trials.
The outcomes of adjuvant RT in patients with gross or near total resection are illustrated by the COG ACNS0121 phase II trial, which included 118 patients with near total or macroscopic gross total resection of a grade 2 or 3 ependymoma (stratum 3) and 163 patients with microscopic gross total resection of a grade 2 or 3 ependymoma (stratum 4), excluding supratentorial grade 2 tumors, all of whom received postoperative focal RT [9]. With a median follow-up of 7.9 years, 5-year event-free survival (EFS) and overall survival (OS) for stratum 3 were 67 and 83 percent, respectively. Similar EFS and OS results were observed for stratum 4 (70 and 88 percent, respectively). For the combined cohort of 281 children with gross or near total resection of a grade 2 or 3 ependymoma (excluding completely resected grade 2 supratentorial tumors), five-year EFS and OS were 69 and 86 percent, respectively, after focal radiation alone.
Post-RT chemotherapy — Post-RT chemotherapy is under investigation in patients with ependymoma but has not yet been adopted, pending mature results of the completed COG ACNS0831 trial.
The phase III COG ACNS0831 trial enrolled children from 1 to 21 years of age with gross total or near total resection of a newly diagnosed grade 2 or 3 ependymoma and randomly assigned them to receive focal RT alone or focal RT plus maintenance chemotherapy (vincristine, cisplatin, cyclophosphamide, and etoposide). Preliminary results were presented at the 19th International Symposium on Pediatric Neuro-Oncology in December 2020 [70]. The intent-to-treat population consisted of 325 eligible randomized patients with a median age of 4.9 years. With a median follow-up of 3.5 years, there was a trend towards improved EFS in patients randomly assigned to RT plus maintenance chemotherapy (78 versus 72 percent; hazard ratio [HR] 0.73, 90% CI 0.51-1.06). In the prespecified stratum of patients with gross or near total initial resection (286 patients), the improvement in EFS was statistically significant (81 versus 71 percent, p = 0.03). OS data have not yet been presented.
Of note, adherence to maintenance chemotherapy was lower than expected, with 27 percent of the chemotherapy arm treated with RT alone, mostly due to patient refusal. In an as-treated analysis excluding patients who did not receive any chemotherapy as well as several in the RT arm who did not receive RT, EFS was greater with chemotherapy (80 versus 71 percent; HR 0.58, 95% CI 0.36-0.94). However, this analysis is subject to bias and therefore of uncertain significance.
The final results of ACNS0831, which will include outcomes by molecular subgroups of ependymoma, are not yet available. The ongoing phase III study through the International Society of Paediatric Oncology (SIOP; SIOP-EP-II, NCT02265770), which randomizes patients to adjuvant chemotherapy or observation after surgical resection and RT, will provide additional evidence that could further modify the role of adjuvant chemotherapy in children with newly diagnosed ependymoma. Hence, the role of adjuvant chemotherapy following RT remains unclear and is not recommended in this population.
Subtotal resection, age 1 to 21 years — Patients with incompletely resected ependymomas of either grade (2 or 3) have consistently worse progression-free survival (PFS) and OS compared with patients who undergo gross total resection. In these high-risk patients, a short course of chemotherapy is suggested postoperatively, followed by second-look surgery if eligible, followed by conformal RT [4,73].
Cisplatin, carboplatin, cyclophosphamide, and etoposide appear to be the most active agents for ependymoma, and higher response rates have been observed with more intense combination regimens [74-79]. Patients are typically treated with one to four cycles of multidrug chemotherapy after surgery, and MRI is reviewed to determine whether there is residual disease and, if present, whether the residual disease is felt to be safely resectable. Patients with potentially resectable residual disease then undergo a second surgical debulking prior to proceeding with adjuvant focal RT.
RT considerations and dosing are similar to those reviewed above, except that higher doses may be recommended for areas with residual macroscopic disease. (See 'Gross total resection, age 1 to 21 years' above.)
Supporting evidence for this approach includes data from both single-center and multicenter series [4,9,69,73,80]. The largest prospective experience is that of stratum 2 of the COG ACNS0121 phase II trial, which included 64 patients with incompletely resected grade 2 or 3 ependymoma treated with chemotherapy, second surgery (if eligible), and focal RT [9]. With a median follow-up of 7.9 years, EFS and OS were 37 and 70 percent, respectively. The rate of second surgery completion was 39 percent. Chemotherapy consisted of two cycles of a combination of carboplatin, cyclophosphamide, etoposide, and vincristine over approximately seven weeks.
There does not appear to be a benefit of additional post-RT chemotherapy in this group of patients. In patients enrolled in the ACNS0831 trial who had an initial subtotal resection and achieved a complete response with induction chemotherapy or gross total resection with second-look surgery, preliminary results indicate that patients randomly assigned to RT plus maintenance chemotherapy had similar EFS compared with those assigned to RT alone [70]. Patients who failed to achieve a complete response after induction chemotherapy also did not appear to benefit from post-RT chemotherapy.
Children <1 year of age — Although adjuvant RT is the standard treatment for most patients with ependymoma and offers the best chance of durable local control, concerns about potential adverse developmental effects limit its use in very young children [40]. For children less than one year of age, we therefore suggest adjuvant chemotherapy as a bridge to delay the use of adjuvant RT.
Those with residual tumor after chemotherapy may be considered for second-look surgery, similar to older children who have undergone subtotal resection, and these patients should be managed at a specialized tertiary care center whenever possible.
Above the age of one, the potential benefits of focal RT on local control and long-term survival are generally felt to outweigh the risks, and chemotherapy in place of RT for children age one to three should only be in the context of a formal protocol. (See 'Areas of uncertainty and ongoing trials' below.)
There are no randomized trials comparing immediate RT following surgery with a strategy in which surgery is followed by chemotherapy, and RT is deferred. The potential role of this approach in children up to the age of three to five years is illustrated by three cooperative group series:
●In a series of 41 children younger than three years of age who presented with nondisseminated disease with the exception of one child, adjuvant multiagent chemotherapy was administered following surgical resection and included either vincristine, methotrexate, and cyclophosphamide alternating with cisplatin and etoposide for 14 months, or vincristine, etoposide, and cyclophosphamide for six months. Twenty-nine patients progressed locally after a median of nine months, yet of the 13 survivors, 6 never received RT. The five-year PFS, EFS, and OS rates were 27, 26, and 37 percent, respectively. Among survivors with median follow-up of eight years, there were no significant differences in cognitive outcomes for those receiving or not receiving RT [81].
●In a series of 73 children under five years of age who presented without disseminated disease, the initial treatment following maximal surgical resection consisted of seven cycles of chemotherapy [75]. At a median follow-up of five years, the four-year OS was 59 percent, and the five-year EFS was 22 percent. Overall, 36 of 73 children (49 percent) required RT for recurrent disease at a median of 15 months after diagnosis.
●In another study, 89 children aged three years or less were treated with combination chemotherapy after maximal surgical resection [82]. Of the 80 patients without disseminated disease at presentation, 50 eventually relapsed. At a median follow-up of six years, the five-year EFS and OS rates were 42 and 63 percent, respectively, and the five-year incidence of avoiding RT was 42 percent.
High-dose chemotherapy with autologous stem cell transplantation is considered experimental for ependymoma. In a pilot study, five children with anaplastic ependymoma, all under the age of three, were treated with high-dose chemotherapy and stem cell rescue [83]. Four patients had bulk residual and three presented with disseminated disease. All patients made it to age three without RT, and only one progression has been noted at a median follow-up of 45 months.
Patients with disseminated disease — Patients with leptomeningeal dissemination and/or spinal deposits of an intracranial ependymoma generally have a poor prognosis, although survival can be prolonged in some patients with ZFTA fusion-positive tumors [50]. Treatment should be individualized. Given challenges in the interpretation of CSF cytology in ependymoma, cytology should be repeated 10 to 14 days postoperatively to confirm results (see 'Extent of disease evaluation' above), and patients should be referred to a tertiary care center whenever possible.
Craniospinal irradiation (CSI) is typically indicated after surgery, but concerns about developmental adverse effects are greater for CSI compared with focal radiation. CSI is usually avoided in children younger than three years of age; in these patients, therapy usually consists of chemotherapy with or without focal RT.
Areas of uncertainty and ongoing trials
●Observation alone after gross total resection of supratentorial differentiated/grade 2 tumors – Retrospective data indicate that patients with completely resected supratentorial grade 2 ependymomas may do well without adjuvant RT, and the role of adjuvant RT in this setting remains controversial [62,84-86]. Because most recurrences are local, the rationale of delayed RT is that patients may be spared the potential toxicities of RT upfront, and those who recur after surgery alone can be salvaged with further surgery and postoperative RT.
In the COG ACNS0121 prospective phase II trial, the five-year PFS was 61 percent in 11 pediatric patients with gross totally resected grade 2 supratentorial ependymoma who were observed after surgery, and all patients were alive at five years [9]. Low patient enrollment in this treatment stratum limited the analysis, and a similar approach was evaluated in the completed ACNS0831 trial. Final results of this strata are not yet available, including the outcomes by molecularly subgroups of supratentorial ependymoma.
●Chemotherapy alone in children >1 year of age – Additional studies are needed to establish the role of chemotherapy as a component of primary treatment in young children with ependymoma, particularly with the emergence of conformal radiation techniques that have reduced concerns for radiation-associated toxicities. At present, chemotherapy in place of RT for patients older than one year of age is advised to be used within the context of a formal protocol.
Adults, age >21 years — Intracranial ependymoma is uncommon in adults, and no randomized trials have been performed. The treatment approach is based on retrospective data in adults and extrapolation from the experience in children. As in children, extent of initial resection is strongly associated with survival, and maximal safe resection is recommended in all patients [86-89]. (See 'Surgical resection' above.)
Postoperatively, focal RT is suggested for patients with localized grade 3 (anaplastic) ependymoma as well as those with residual disease after resection of a grade 2 ependymoma [66,67,90,91]. Craniospinal RT is reserved for those with evidence of disseminated disease by imaging or CSF cytology [90]. Adjuvant chemotherapy is not known to be beneficial in adults with ependymoma.
The role of adjuvant focal RT after gross total resection of a grade 2 ependymoma in adults is uncertain. Consensus guidelines from both the National Comprehensive Cancer Network (NCCN) and the European Association for Neuro-Oncology (EANO) suggest observation after gross total resection of a grade 2 tumor in adults [67,90]. This is based on evidence from retrospective studies of adults with grade 2 ependymoma, in which RT has not been associated with improved PFS or OS for patients with completely resected tumors [92,93]. Others, however, prefer to give postoperative focal RT even after complete resection, particularly if there is any concern about microscopic residual disease at the time of surgery, with the rationale that focal RT is well tolerated, salvage strategies are not proven to be beneficial, and recurrences are associated with morbidity, particularly in the posterior fossa.
Other ependymal tumors
Subependymoma — Subependymomas are often discovered incidentally in older adults, and many presumed subependymomas do not require intervention unless they grow over time and cause symptoms.
Patients with large, symptomatic subependymomas often do well after complete resection alone, underscoring the different natural history of these lesions compared with other brain ependymomas [94]. RT is typically reserved for symptomatic or enlarging tumors that cannot be resected.
Spinal ependymomas — Treatment of spinal cord ependymal tumors is reviewed elsewhere. (See "Spinal cord tumors", section on 'Ependymomas'.)
FOLLOW-UP AND MONITORING — Although the majority of treatment failures in children with ependymoma develop within five years from diagnosis, late failures and recurrence or progression of the disease beyond five years can be seen, warranting prolonged surveillance beyond the traditional five-year period [95].
Surveillance — While the exact frequency and methods of surveillance imaging may vary according to institutional policies, in general, we recommend the following strategy, most of which is consistent with the ongoing Children's Oncology Group (COG) ACNS0831 trial.
●MRI of the brain:
•Every three to four months for the first three years after completion of therapy.
•Every six months from three to five years after completion of therapy.
•In most centers, patients are transitioned to a survivorship clinic beyond five years after completion of therapy. We recommend annual MRI of the brain for another two to five years from then to detect late recurrences in these children.
●An MRI of the spine should be considered under the following circumstances:
•Annually along with an MRI of the brain.
•In the presence of symptoms suggestive of spinal cord involvement by the tumor.
•At the time of progression of disease noticed on MRI of the brain.
Survivorship — Children who are long-term survivors after treatment for central nervous system (CNS) tumors are at risk for a number of problems, including neurocognitive deficits, focal neurologic deficits, sensorineural hearing loss, growth abnormalities, endocrine abnormalities, radiation necrosis, vasculopathy, and second malignancies [40]. These problems can be due to the damage originally done by the tumor or the treatment (surgery, radiation therapy [RT], and/or chemotherapy) [96-103]. The nature and extent of these problems will be influenced by the location of the tumor and the specific treatment given. (See "Delayed complications of cranial irradiation" and "Overview of the management of central nervous system tumors in children", section on 'Long-term morbidity'.)
Adult patients can experience significant long-term morbidity including fatigue, numbness, pain, and altered sleep [104].
Survivorship guidelines with recommendations for long-term follow-up after treatment of childhood cancer are available from COG [105].
RECURRENT DISEASE — The long-term prognosis for patients with recurrent disease is poor, and while patients can be palliated for extended periods of time, most will eventually succumb within years of relapse. First recurrence is local in approximately two-thirds of patients, metastatic (within the central nervous system) in approximately 20 percent, or a combination in the remaining 10 to 15 percent [106,107].
While multiple treatment options are available, it is important that patients and caregivers understand the poor long-term prognosis of patients with recurrent or progressive ependymoma after surgical resection and radiation therapy (RT). Careful selection of treatment options for these patients can provide excellent palliation and good quality of life; no single option is considered standard or proven effective [108].
Outside of a formal clinical trial, each patient should be managed on a case-by-case basis, taking into account their age, location of the original and recurrent disease, presence of metastases, prior treatment, and functional status. In addition to experimental therapy, multiple approaches have been tried in this population:
●Surgery – Aggressive surgical resection may provide effective palliation in some patients. Chemotherapy with or without RT may have a role in reducing residual or recurrent tumor prior to attempted reresection, all aimed at prolonging the time to further disease progression and death [74,77,109].
●Radiation – Reirradiation of recurrent tumors may be beneficial and has yielded a salvage rate in carefully selected cases [110-114]. Options include stereotactic radiosurgery (SRS), focal fractionated reirradiation, and craniospinal irradiation (CSI), depending upon the location, patient age, and extent of recurrence [110,115]. Patients treated with focal reirradiation remain at risk for the development of disseminated metastases, and recurrence at the primary site eventually occurs in most patients.
In one of the larger retrospective series of patients treated with aggressive repeat surgical resection and reirradiation at a median of 27 months after first RT, the median progression-free survival (PFS) and overall survival (OS) from the first day of the second RT were 27 and 75 months, respectively [114]. Of the 101 patients in this series, 57 patients (56 percent) experienced subsequent tumor progression following either focal RT or craniospinal and boost reirradiation, and local failure was a component of progression in 35 of these patients (61 percent). Outcomes were most favorable in those patients who experienced a distant-only failure after the first RT course without anaplasia at recurrence and who went on to receive CSI and boost re-RT. The completed phase II RERTEP trial explored this treatment option, with results pending (NCT02125786).
●Chemotherapy – Conventional chemotherapy may provide symptom palliation. There is no single preferred drug or regimen for recurrent ependymoma in children [90]. As described above, cisplatin, carboplatin, cyclophosphamide, and etoposide appear to be the most active agents for ependymoma, based on use in the adjuvant or neoadjuvant setting (see 'Subtotal resection, age 1 to 21 years' above). In a series of 28 adults with recurrent or progressive intracranial ependymoma, six had an objective response to chemotherapy; the response rate was higher with platinum-based regimens [116]. In addition to platinum-based therapy, oral etoposide, nitrosoureas, temozolomide, and fluorouracil have some activity as single agents in patients with recurrent disease [90,117-121]. Responses are usually short lived, and tumor progression after chemotherapy alone usually occurs quickly.
●Molecular targets – There is a growing interest in potential molecular targets for ependymoma, including the epidermal growth factor receptor (EGFR) [16] and vascular endothelial growth factor (VEGF), as well as epigenetic and metabolic regulators [122-126].
In adults, the combination of temozolomide and lapatinib, an oral human epidermal growth factor receptor 2 (HER2) and EGFR inhibitor, is an active regimen for recurrent disease. In a phase II study in 50 adults with recurrent intracranial and spinal ependymoma, median PFS and OS were 7.8 months and 2.25 years, respectively [127]. Two complete and six partial responses were seen, and symptoms stabilized or improved in many patients. In a series of eight adults, six had a partial response to bevacizumab, with a median time to progression of six months [128], although other retrospective series have not demonstrated significant efficacy [129].
In children, prospective trials are in progress, and all approaches are considered investigational. A clinical trial of bevacizumab and irinotecan in 13 children with recurrent or progressive ependymoma failed to show significant activity (no objective responses, median time to progression 2.2 months) [130]. A small phase II study of bevacizumab plus lapatinib also failed to show benefit in 24 children with recurrent or refractory ependymoma [131].
PROGNOSIS
Children — Intracranial ependymomas in children are associated with poor long-term survival, even after complete resection and adjuvant therapy. With long-term follow-up, 10-year overall survival (OS) is approximately 50 to 70 percent [9,69,95].
A number of factors affect the likelihood of disease-free survival after treatment in children [132]:
●Extent of resection – Approximately 80 percent of recurrences are local, underscoring the importance of maximal safe resection [59,91,133-135]. Although the reported degree of benefit has varied, all studies have indicated that patients who can be treated with gross total resection have a better prognosis [4,59,91,133]. Even with gross total resection, however, ependymomas can recur and lead to death in approximately 40 percent of children by 10 years after diagnosis [95].
●Tumor location – Comparisons of survival based on tumor location have yielded conflicting results. While several studies have suggested that supratentorial tumors have a better survival [4,136], others have found that patients with infratentorial lesions do better [91]. Involvement of the cervical spinal cord has been associated with decreased survival due to failures outside the primary site [137].
●Age – In children, older age is associated with better prognosis [135]. Compared with older children, infants may do less well, partly because they have a higher incidence of infratentorial tumors, and partly because they tend not to receive timely adjuvant radiotherapy due to concerns of toxic effects on brain development.
●Histologic grade – The prognostic significance of histology grade (2 versus 3) in ependymoma remains controversial. While some studies have found that grade 3 tumors are associated with worse outcome compared with grade 2 tumors [9,138], other studies have been less conclusive given the difficulty in arriving at a definitive diagnosis of the grade of tumor [10]. (See 'Histology and classification' above.)
●Molecular genetic subgroups – A number of molecular subgroups have been identified based on comprehensive genomic sequencing of series of ependymoma samples, which ultimately may prove more powerful than histology for prognostic differentiation. (See 'Molecular subgroups' above.)
●Additional molecular alterations – Copy number gain of chromosome arm 1q in posterior fossa ependymoma of childhood has been associated with unfavorable outcome in a number of retrospective series [13,16,22,26,42]. Subsequent data from prospective clinical trials have supported the association of 1q gain in posterior fossa ependymoma group A (PF-EPN-A) tumors in children, while the prognostic implication of 1q gain remains to be fully defined in posterior fossa ependymoma group B (PF-EPN-B) tumors [9,10,22,33]. In tumors with ZFTA fusion, CDKN2A deletion may identify a subset with a particularly poor prognosis, but further studies are needed [51].
Adults — Intracranial ependymoma is less common in adults, and outcomes data are more limited. Two large European multi-institution series that included a total of 222 patients and 123 patients from a single United States institution analyzed the outcomes and prognostic factors for adults with intracranial ependymomas [87,88,139]. In these series, the five-year survival rate ranged from 67 to 85 percent, and the 10-year survival varied from 50 to 77 percent. On multivariate analysis, factors associated with a poor prognosis included high histologic grade, incomplete surgical resection [140], location [139,141], and a Karnofsky Performance Status ≤80 (table 2) [88]. The prognostic importance of incomplete resection has also been observed in a population-based study of both adults and children with ependymoma using the Surveillance Epidemiology and End Results (SEER) database [61].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Primary brain tumors".)
SUMMARY AND RECOMMENDATIONS
●Epidemiology – Intracranial ependymomas have a peak incidence in early childhood, with a median age at diagnosis of five years. In adults, ependymomas more commonly occur in the spinal cord. (See 'Epidemiology' above and "Spinal cord tumors", section on 'Ependymomas'.)
●Classification – The three major types of ependymoma are subependymoma (World Health Organization [WHO] grade 1), myxopapillary ependymoma (WHO grade 2), and ependymoma (WHO grade 2 or 3, may be supratentorial, posterior fossa, or spinal) (table 1). For intracranial ependymomas, the distinction between grade 2 and grade 3 is of uncertain prognostic significance and may be replaced by molecular stratification in the future. (See 'Pathology' above.)
●Clinical presentation – Signs and symptoms depend on tumor location. For posterior fossa tumors, children often present with headache, nausea and vomiting, ataxia, vertigo, and papilledema due to obstructive hydrocephalus and tumor mass effect on the cerebellum. (See 'Presenting signs and symptoms' above.)
●Diagnosis and staging – The diagnosis of intracranial ependymoma requires histologic confirmation. Ependymoma may be suspected preoperatively based on imaging characteristics, tumor location, and patient age. Staging with complete spine MRI and cerebrospinal fluid (CSF) testing should be performed, as up to 10 percent of patients have disseminated disease at the time of diagnosis. (See 'Diagnosis' above.)
●Surgical resection – The initial treatment for patients with a suspected ependymoma is maximal safe resection. Gross total resection is the goal but is not always feasible due to brainstem involvement or proximity to other critical structures. Patients should be referred to a specialized center with pediatric oncologic neurosurgical expertise whenever possible. (See 'Surgical resection' above.)
●Postoperative therapy – Intracranial ependymomas have the potential to recur and disseminate throughout the nervous system. Postoperative radiation therapy (RT) and/or chemotherapy is therefore required in most patients, even in the presence of a negative staging evaluation (algorithm 1).
•For children 1 to 21 years of age with an intracranial ependymoma grade 2 or 3 who undergo gross total or near total resection upon initial surgical resection, we recommend postoperative focal RT (Grade 1B) (see 'Gross total resection, age 1 to 21 years' above). Potential exceptions include the following:
-For children between one and three years of age who undergo gross total resection, chemotherapy may be an alternative to immediate RT in an effort to avoid the neurologic complications of RT in young children. However, this is advised in the context of a formal clinical trial. (See 'Children <1 year of age' above and 'Areas of uncertainty and ongoing trials' above.)
-For children with a supratentorial grade 2 ependymoma who undergo gross total resection, observation may be an alternative to adjuvant therapy. (See 'Areas of uncertainty and ongoing trials' above.)
•For most children 1 to 21 years of age who undergo subtotal resection of an intracranial ependymoma (WHO grade 2 or 3), we suggest postoperative chemotherapy followed by second-look surgery, if eligible, followed by adjuvant focal RT, rather than immediate postoperative RT (Grade 2C). (See 'Subtotal resection, age 1 to 21 years' above.)
•For children less than one year of age with an intracranial ependymoma, we suggest adjuvant chemotherapy as a bridge to delay the use of adjuvant RT (Grade 2C). (See 'Children <1 year of age' above.)
•Children with leptomeningeal dissemination and/or spinal deposits at the time of diagnosis have a poor prognosis. Treatment typically consists of craniospinal RT or multiagent chemotherapy, depending on the age of the patient. (See 'Patients with disseminated disease' above.)
•For adults with a grade 3 intracranial ependymoma and those with an incompletely resected grade 2 ependymoma, we suggest immediate postoperative focal RT, rather than no RT or RT plus chemotherapy (Grade 2C). Decisions about postoperative RT after gross total resection of a grade 2 ependymoma are individualized. (See 'Adults, age >21 years' above.)
●Follow-up – Patients with intracranial ependymoma remain at risk for local and distant recurrence after completion of initial therapy. In children, we conduct regular imaging and clinical surveillance for 7 to 10 years after completion of therapy. (See 'Surveillance' above.)
●Prognosis – Even with maximal therapy, intracranial ependymomas are associated with significant risk of recurrence and decreased long-term survival. With long-term follow-up in children, 10-year overall survival (OS) is approximately 50 to 70 percent. (See 'Prognosis' above and 'Recurrent disease' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Mark Kieran, MD, PhD, who contributed to an earlier version of this topic review.
Do you want to add Medilib to your home screen?