Ependymal Tumors

General

Ependymomas are glial tumors that display morphologic, immunophenotypic, and ultrastructural similarities to the ependymal cells lining the cerebral ventricular system. As a group, ependymomas have recently undergone extensive genetic and epigenetic examination which has subcategorized them into nine distinct genetic groups (1). Although four of these groups involve WHO grade I myxopapillary ependymomas and subependymomas and are not necessarily relevant to clinical management, the other five groups represent clinically meaningful entities among classic ependymomas, or what previously was in essence one category with two histologic grades. The 2016 WHO classification is a snapshot of the as yet incomplete transition from histologically defined classic ependymomas to a more genetically based scheme, in which the future relevance of histologic grading is unclear. It retains the general morphologic categories and histologic grading of the previous system while adding a category of supratentorial ependymoma defined by the presence of RELA gene fusions in the context of ependymoma morphology (2). Although the grade I entities are unlikely to change, the grade II and III ependymomas will likely undergo further subcategorization by genetic features in future editions. Here I present the morphologic categories of ependymoma and include the separate, genetically defined WHO diagnostic entity of “Ependymoma, RELA fusion–positive” among classic (grade II and III) ependymomas.

Ependymoma (WHO Grade II), Anaplastic Ependymoma (WHO Grade III), and Ependymoma, RELA Fusion–Positive

Clinical Context

Classic ependymomas occur primarily in children and young adults, and may arise anywhere along the neuraxis within the ventricles, spinal cord, or the cerebral hemispheres (3,4,5,6). As a whole, ependymal tumors peak in incidence in children ages 1 to 4 years where the rate is about a third higher than in infancy and twice as high as from 5 to 18 years (7). Most of those cases are in the posterior fossa and the rest are supratentorial or, rarely, spinal (1). Ependymoma incidence decreases with increasing age,
decreasing to rare events after the age of 35 (8). A second, smaller peak of ependymomas occurs in older children and adolescents, in whom the supratentorial compartment in the lateral ventricles and hemispheres is the most common site. Spinal cord ependymomas occur most often in adults (30 to 40 years) and are most frequent in the cervical and cervicothoracic regions.

The clinical symptoms of ependymomas necessarily depend on patient age and tumor location (9,10). Posterior fossa masses in infants may present with an enlarging head due to increased intracranial pressure with incompletely fused cranial sutures, whereas slightly older children will often present with increased intracranial pressure and hydrocephalus with headaches, nausea, vomiting, dizziness, and ataxia. Spinal ependymomas present clinically with sensory and motor deficits that correspond to the spinal level involved.

By neuroimaging, ependymomas are generally solid, well circumscribed, and contrast enhancing (Figure 5-1). They push aside adjacent nervous system rather than diffusely invading it. In the posterior fossa, ependymomas typically fill the fourth ventricle and displace the cerebellum posteriorly. It is often difficult to determine if a posterior fossa tumor arises from the cerebellum and extends into the fourth ventricle (as expected for medulloblastoma) or if the tumor arises in the ventricle and extends into the cerebellum (typical of ependymoma). Ependymomas in this location may extend into the adjacent subarachnoid space by exiting through the foramina of Luschka. Although ependymomas are contrast-enhancing tumors, the degree of enhancement is variable and the pattern is often heterogeneous.

FIGURE 5-1 Sagittal, T2-weighted, noncontrast magnetic resonance imaging of a typical fourth ventricular ependymoma with cystic spaces.

The only known consistent risk factor for ependymomas is neurofibromatosis type 2 (NF2), in which patients are also predisposed to developing vestibular schwannomas and meningiomas. The ependymomas in NF2 usually occur in the spinal cord, particularly the cervical segment, and have an indolent clinical course compared to intracranial examples.

Prognostic Factors

Age and Location

Patient age is one of the major prognostic features for ependymomas (11,12,13). The effect of age, though, is difficult to separate from location and extent of resection, because the vast majority of ependymomas occurring before 4 years are in the posterior fossa where they are many times impossible to dissect free of the floor of the fourth ventricle without severely harming the patient. Nevertheless, these tumors appear to have a poorer prognosis compared with ependymomas that occur in adults when adjusting for those factors. In contrast, ependymomas arising in adults, most of which are in the spinal cord, have a relatively good prognosis following resection.

Extent of Resection

Gross total resection has been shown to be an important and often the statistically strongest prognostic indicator for ependymomas, and numerous studies have shown a survival benefit from total versus subtotal resection. Given their generally circumscribed margins, at least in comparison to the diffuse gliomas, this makes some intuitive sense. However, extension of fourth ventricular tumors into the subarachnoid space to involve brainstem structures and cranial nerves may preclude complete resection.

Genetic Classification

Evidence from DNA methylation profiling and copy number studies has supported the idea that tumor location is a major determinant of ependymoma biology and that there are subgroups within each anatomic segment that have significant prognostic impacts. In brief, posterior fossa ependymomas can be divided into two groups; the majority are “group A,” and the rest are “group B.” Group A, posterior fossa ependymomas occur in younger children, mostly <4 years, and have significantly lower survival (Table 5-1). Group B tumors have better outcomes and occur in older children and young adults. Similarly, supratentorial ependymomas can be placed in two clusters along genetic lines, one group with fusions of the RELA gene and poor survival and others with YAP1 gene fusions and better survival (Table 5-1). Spinal cord ependymomas form their own genetic group among classic ependymomas and have a much better outcome than intracranial cases. These five groups are discussed further below.

Gains of 1q

Gain of the long arm of chromosome 1 may be seen in any posterior fossa ependymoma or in RELA fusion–positive supratentorial ependymomas and is associated with decreased survival and/or increased recurrence rates when assessed in aggregate with all intracranial
ependymomas (14,15). However, when separated by molecular groups, decreases in overall survival were only noted among group A posterior fossa tumors, not RELA fusion or group B cases, although the latter did show increased recurrence (1).

TABLE 5-1 Intracranial Classic Ependymomas: Genetic Groups

Location	Genetic Group	Characteristics
Posterior fossa	Group A	Age: peak 1–4 yrs, rare after 18 Genetic feature: balanced genome Prognosis: good, 5-yr OS: ∼70%
	Group B	Age: Mostly adults, some older children Genetic feature: unstable genome, chromosome 6 del Prognosis: good, 5-yr OS: ∼100%
Supratentorial	RELA fusion	Age: broad range, peaks 4–18 yrs Genetic feature: RELA fusion, chromothripsis on 11q Prognosis: good, 5-yr OS: ∼75%
	YAP1 fusion	Age: peak 1–4 yrs, some in older adults Genetic feature: YAP1 gene fusion Prognosis: good, 5-yr OS: ∼100%

Tumor Grade

Much effort has been made to construct a prognostically relevant, reproducible histologic grading scheme for classic ependymomas, and the results have been mixed. DNA methylation profiling shows no meaningful biological differences between grade II and anaplastic grade III ependymomas (1,16). Large series conducted by experienced neuropathologists have not shown significant differences in overall survival based on histologic grade, although there does appear to be differences in recurrence rates (17,18). The current consensus statement from an international group of pediatric brain tumor researchers recommends histologic grading not be used alone for stratifying clinical trials or clinical decision making, rather that it be integrated with molecular subgrouping routinely (19).

Treatment

Because they have shown consistent positive effects, both surgical resection and postoperative local irradiation of the tumor bed are the mainstays of current therapy for intracranial ependymomas. Chemotherapy has shown inconsistent results and does not have an established role in the treatment of ependymoma. Spinal cord ependymomas are usually treated with surgical resection alone.

FIGURE 5-2 The perivascular pseudorosette is the hallmark of ependymomas, although similar structures can be seen in several other lesions.

Histopathology

FIGURE 5-3 In smear preparations, the pseudorosettes of ependymomas form fibrillar pseudopapillary structures.

Classic ependymomas are characterized by tumor cells that form pseudorosettes around central blood vessels and “true” ependymal rosettes. Pseudorosettes consist of ependymal tumor cells oriented around a central blood vessel with long fibrillar processes that extend radially from the vessel and give a pinwheel appearance (Figures 5-2 and 5-3). True rosettes
(or canals), on the other hand, form a lumen centrally, reminiscent of the central canal of the spinal cord (Figure 5-4). One may also see epithelioid surfaces lined by ependymal tumor cells recapitulating their function of lining the ventricular system (Figure 5-5). Ependymomas show a wide range of cellularity and degrees of fibrillarity. Other features that can be seen in ependymoma include intratumoral hemorrhage, foci of necrosis,
cartilaginous and osseous metaplasia (20), calcification, and hyalinization of blood vessels (Figure 5-6).