03.04 - intraventricular tumours

INTRAVENTRICULAR TUMORS

CONTENTS

Preface ixAndrew T. Parsa and Mitchel S. Berger

Epidemiology and Pathology of Intraventricular TumorsJames S. Waldron and Tarik Tihan 469

Intraventricular tumors present a diagnostic challenge to the clinician because of a broaddifferential diagnosis with significant variability in tumor type between adult andpediatric populations. This expansive differential diagnosis includes choroid plexuspapillomas and carcinomas, ependymomas, subependymomas, subependymal giant cellastrocytomas, central neurocytomas, meningiomas, and metastases as well as a numberof cysts, inflammatory lesions, and other rare neoplasms. Posterior fossa ependymomas,subependymal giant cell astrocytomas, and choroid plexus tumors are more likely toappear in childhood, whereas subependymomas, central neurocytomas, intraventricularmeningiomas, and metastases are more frequent in adults. This article reviews theepidemiology, the pathologic characteristics, and the primary diagnostic considerationsof each tumor type.

Intraventricular Neurocytomas 483Janet Lee, Susan M. Chang, Michael W. McDermott, and Andrew T. Parsa

Central neurocytomas (CNCs) are World Health Organization II benign central nervoussystem (CNS) neoplasms first described in 1982 by Hassoun and his colleagues. Hall-mark features of CNC include (1) occurrence in the lateral ventricle of young adults,(2) a well-circumscribed isodense to hyperdense mass with contrast enhancement onCT and isointense to hyperintense compared with normal brain parenchyma on T1-and T2-weighted MRI, (3) resemblance to oligodendroglioma on light microscopy, (4)neuronal origin seen in electron microscopy and immunohistochemistry, and (5) favor-able prognosis with benign biologic behavior. CNCs comprise 0.1% to 0.5% of all CNSneoplasms based on pathologic review at several neurosurgery centers. A population-based incidence has not been established, in part because of the paucity of cases. Givenits recent distinction as a unique tumor and its low incidence, most reports of CNC arefrom the pathologic literature with little data regarding its management. Furthermore,many early cases of CNC were misdiagnosed, and treatment was based on the presumeddiagnosis of oligodendroglioma or ependymoma. Accordingly, this article presents acomprehensive review of the literature and proposes a management paradigm for thetreatment of CNC.

VOLUME 14 NUMBER 4 OCTOBER 2003 v

Surgical Approaches to Tumors of the Lateral Ventricle 509Richard C.E. Anderson, Saadi Ghatan, and Neil A. Feldstein

Tumors of the lateral ventricles comprise a relatively rare heterogeneous group of lesionsin children and adults. They arise from the ependyma and subependyma that line theventricles, from the choroid plexus arachnoid and epithelium, or from ectopic tissuerests that have become trapped within the ventricle or its lining. Although the lateralventricles are among the most surgically inaccessible areas of the brain, numerous op-erative approaches to the ventricles have been developed. This article first discusses theclinical manifestations and differential diagnosis of lateral ventricular tumors. Relevantregional anatomy and general operative strategies for these lesions are then discussed,with particular focus on the following approaches: frontal, temporal, and parietal trans-cortical approaches and anterior and posterior interhemispheric approaches.

Surgical Approaches to Posterior Third Ventricular Tumors 527Alan P. Lozier and Jeffrey N. Bruce

Advanced microsurgical techniques combined with improved neuroanesthetic and post-operative critical care have made aggressive surgical resection a mainstay in the manage-ment of posterior third ventricular and pineal region tumors. Although a variety ofapproaches to the posterior third ventricle have been designed, three are in commonuse. The infratentorial-supracerebellar approach takes advantage of a natural corridorbetween the cerebellum and the tentorium. Supratentorial approaches include theinterhemispheric-transcallosal and occipital-transtentorial approaches. Refinements insurgical technique have led to a more favorable outlook for patients with these un-common tumors.

Endoscopic Adjuncts to Intraventricular Surgery 547Sandeep Kunwar

Recently, endoscopic intraventricular surgery has been performed successfully in severalclinical series. Although the therapeutic results must be compared with conventionalsurgery, neuroendoscopy seems to be a safe surgical technique when performed by sur-geons with appropriate experience and refined endoscopic tools. Rigid or flexible endo-scopes equipped with various-sized working channels should be selected depending onthe nature of the pathologic findings. The well-proven tenets of microsurgery must notbe sacrificed for the sake of more rapid surgical time and noninvasiveness; thus, endo-scopic surgery must adhere to the principles of microsurgery. The improved visualiza-tion and lower morbidity have established neuroendoscopy in the management ofspecific disease processes, such as obstructive hydrocephalus. Its further use in the man-agement of intraventricular cysts and tumors is dependent on long-term follow-up andthe development of even better instrumentation.

Intraventricular Meningiomas 559Michael W. McDermott

Meningiomas arising in the ventricular system are rare; yet, when they do presentclinically, they are often large, most often within the atrium, and most frequently onthe left. For all these reasons, they are tumors for which it is difficult to achieve theperfect surgical result: complete removal of a benign tumor without complicationsor new neurologic morbidity. With a thorough understanding of the anatomy of struc-tures around the ventricle, selection of the proper surgical approach, and use of mod-ern neurosurgical techniques, however, modern-day surgical results should be superiorto those of the past.

vi CONTENTS

Intraventricular Gliomas 571Aaron S. Dumont, Elana Farace, David Schiff, and Mark E. Shaffrey

Significant progress has been realized in the contemporary understanding and treatmentof intraventricular gliomas. However, there remains a substantial need for continued ad-vancement in the clinical management of patients harboring these lesions, particularlyependymomas. This article addresses the specific types of intraventricular gliomas withemphasis on each tumors defining characteristics and the specific nuances of manage-ment in each variant.

Surgical Resection of Metastatic Intraventricular Tumors 593Giacomo G. Vecil and Frederick F. Lang

Intraventricular metastases are a unique challenge for neurosurgical oncologists. Thispaper describes the clinical features and surgical management strategies of intraventric-ular metastases based on a review of the literature and an analysis of 35 patients treatedin the Department of Neurosurgery at The University of Texas M.D. Anderson CancerCenter over the last 10 years. Intraventricular metastases comprise 0.95% of intrapar-enchymal metastases. Renal cell carcinoma has the highest propensity of all primary tu-mors to metastasize to the ventricle. The trigone of the lateral ventricular is the mostcommon location with the ventricle for metastases to occur, presumably due to the highconcentration of choroid plexus in the region. Despite the deep location, surgical resec-tion can be achieved safely in most cases. The survival of surgically treated patients iscomparable to that of patients with intraparenchymal metastases.

Intraventricular Congenital Lesions and Colloid Cysts 607Aurelia Peraud, Anna Illner, and James T. Rutka

Intraventricular congenital lesions and colloid cysts comprise a rather large spectrum ofdifferent pathologic conditions. In most cases, treatment in not warranted unless there isprogressive ventricular obstruction with hydrocephalus or growth of the lesion itself,making tissue biopsy and histopathologic diagnosis necessary. Accordingly, a preciseneuroradiologic evaluation is of the utmost importance, because most lesions, if notsymptomatic, only require clinical and radiologic follow-up.

Choroid Plexus Tumors in Children 621Nalin Gupta

Choroid plexus tumors represent a well-defined subset of brain tumors that occurmainly in young children. Surgical resection for papilloma is usually curative, althoughcareful surgical planning is required to minimize the potential risks. Although adjunctivetherapy for carcinoma includes chemotherapy or radiation, the long-term survival forcarcinoma remains poor.

Cumulative Index 2003 633

CONTENTS vii

FORTHCOMING ISSUES

January 2004

Endoscopy

Rick Abbott, MD, Guest Editor

April 2004

Traumatic Neurovascular Surgery

J. Paul Elliot, MD, Guest Editor

July 2004

Pain Treatment

Gary Heit, MD, Guest Editor

RECENT ISSUES

July 2003

Neuroaugmentation forChronic Pain

Jaimie M. Henderson, MD, Guest Editor

April 2003

Surgery for Psychiatric Disorders

Ali R. Rezai, MD, Steven A. Rasmussen, MD,Benjamin D. Greenberg, MD, PhDGuest Editors

January 2003

Pituitary Surgery

Martin H. Weiss, MD, andWilliam T. Couldwell, MD, PhD, Guest Editors

THE CLINICS ARE NOW AVAILABLE ONLINE!

Access your subscription athttp://www.TheClinics.com
http://www.TheClinics.com

Neurosurg Clin N Am 14 (2003) ix

Preface

Intraventricular tumors

Guest Editors

Mitchel S. Berger, MDAndrew T. Parsa, MD, PhD

Neurosurgical oncology is an evolving disci-

pline that continually benefits from the translationof scientific advances into clinical treatment para-digms. For the past 30 years at the University

of CaliforniaSan Francisco, basic scientists andclinicians have been working together to rapidlyimplement new discoveries for the benefit of ourpatients. Several examples of these collaborations

can be found here and at other premier neurosur-gery departments around the world. A refinedunderstanding of the molecular pathways that

contribute to tumor development has yieldednew targets for chemotherapy, while our increas-ing experience with radiosurgery has broadened

treatment options for patients. The advent of sur-gical adjuncts such as functional mapping tech-niques, computerized frameless stereotaxy, and

endoscopy has significantly decreased surgicalmorbidity. In addition, surgical procedures arenow in place to facilitate local delivery of che-motherapeutic agents with unprecedented tumor

specificity.Intraventricular tumors epitomize the chal-

lenges faced by neurosurgical oncologists in the

1042-3680/03/$ - see front matter 2003 Elsevier Inc. All rigdoi:10.1016/S1042-3680(03)00059-7

twenty-first century. These lesions have a complex

biology and require significant skill to excisewithout attendant morbidity. A comprehensiveunderstanding of ventricular anatomy, surgical

approaches, and nonsurgical treatment optionsis requisite for the neurosurgeon. Successful treat-ment of patients with these lesions requires adedicated team of pathologists, oncologists, and

neurosurgeons. In this issue of the NeurosurgeryClinics of North America, we draw upon theexperience of several colleagues to facilitate a

better understanding of intraventricular tumorsin children and adults.

Andrew T. Parsa, MD, PhD

Mitchel S. Berger, MDDepartment of Neurological Surgery

University of CaliforniaSan Francisco505 Parnassus Avenue M-779

San Francisco, CA 94143, USA

E-mail address: [email protected](A.T. Parsa)

hts reserved.
mailto:[email protected]

Neurosurg Clin N Am 14 (2003) 469482

Epidemiology and pathology of intraventricular tumorsJames S. Waldron, MDa, Tarik Tihan, MD, PhDb,*

aDepartment of Neurological Surgery, University of California at San Francisco, 513 Parnassus, HSW 511,

San Francisco, CA 941430511, USAbNeuropathology Unit, Department of Pathology, University of California at San Francisco, 513 Parnassus,

HSW 408, San Francisco, CA 941430511, USA

Intraventricular tumors present a diagnosticchallenge to the clinician because of a broad

differential diagnosis with significant variability intumor type between adult and pediatric popula-tions. This expansive differential includes choroidplexus papillomas (CPCs) and choroid plexus

carcinomas (CPCs), ependymomas, subependy-momas, subependymal giant cell astrocytomas(SEGAs), central neurocytomas, meningiomas,

and metastases as well as a number of cysts,inflammatory lesions, and other rare neoplasms.Posterior fossa ependymomas, SEGAs, and cho-

roid plexus tumors are more likely to appear inchildhood, whereas subependymomas, centralneurocytomas, intraventricular meningiomas,

and metastases are more frequent in adults. Eachof these tumor types involves the ependymallining and subependymal plate of the ventricularwall, the septum pellucidum, or the highly

vascular choroid plexus. This article reviews theepidemiology, the pathologic characteristics, andthe primary diagnostic considerations of each

tumor type.

Choroid plexus papilloma and carcinoma

Epidemiology

Choroid plexus tumors are epithelial neo-plasms with a prevalence of 0.3 cases per million[1]. In two large series, choroid plexus tumors

accounted for 0.4% [2] and 0.6% [3] of allreported intracranial tumors. The tumor pre-

* Corresponding author.

E-mail address: [email protected] (T. Tihan).


dominantly occurs in childhood, although it canbe seen at any age. The median age of onset is 3.5

years [4], with 20% of patients presenting in thefirst year of life and almost 50% in the first decade[5]. The most common locations for choroidplexus tumors are the lateral and fourth ventricles,

followed by the third ventricle. Cerebellopontineangle examples are less common and are causedby extension of tumor through the foramen of

Luschka [6]. In addition, rare suprasellar caseshave been reported [7]. Tumor location is closelycorrelated with patient age. The most common

location in children is the lateral ventricle,whereas the fourth ventricle is the most frequentsite in adults. Choroid plexus tumors are divided

into the CPP (World Health Organization [WHO]grade I) and the more aggressive CPC (WHOgrade III). CPCs make up a small proportion ofchoroid plexus tumors, primarily present in

children less than 3 years of age, and are mostcommonly found in the lateral ventricles [8]. CPPsand CPCs have been shown to spread through the

cerebrospinal fluid, and rare metastatic cases havebeen documented outside the CNS [4].

Macroscopic and microscopic features

Choroid plexus tumors are often soft to

rubbery and may have a gritty texture becauseof calcification. The tumors are frequently shadesof orange-brown. During surgery, an anchoring

pedicle can be seen attached to the normalchoroid plexus or the ventricular wall. Somepapillomas have a cauliflower-like appearance.

CPPs and CPCs exhibit features akin tomany papillary neoplasms in other organs. CPPshave well-developed fibrovascular cords within

hts reserved.

470 J.S. Waldron, T. Tihan /Neurosurg Clin N Am 14 (2003) 469482

papillary structures and do not exhibit architec-tural or cytologic atypia (Fig. 1A). Epithelial andstromal cells contain many characteristics of the

normal choroid plexus, such as calcifications andxanthomatous change [9,10]. In some cases, thereis a striking nuclear monomorphism withoutother aggressive features, such as mitoses or

vascular proliferation. Rarely, geographic necro-sis without the pseudopalisading that is suggestiveof an infarct can be seen in an otherwise typical

CPP. Osseous or cartilaginous metaplasia andacinar or tubular differentiation are reported inchoroid plexus neoplasms [1113]. In addition,

a number of studies report a pigmented variantthat contains neuromelanin and lipofuscin [14].CPPs with marked oncocytic transformation aswell as glial differentiation are rare [12]. Transi-

tional zones between the normal and neoplasticchoroid plexus can be found in CPPs and CPCs.

Carcinomas of the choroid plexus are tumors

that exhibit all the histologic hallmarks ofaggressiveness (see Fig. 1B). A typical CPC isa neoplasm with increased architectural complex-

ity demonstrating partially solid and partiallynonpapillary growth. Most tumors have markedcytologic atypia, atypical mitotic figures, and

frank necrosis. Some high-grade tumors havecytologic and architectural features that resembleanaplastic oligodendrogliomas. Invasion intoneuropil is characteristic of CPCs, although some

CPPs can occasionally exhibit invasion intosurrounding parenchyma. Rare CPCs resembleundifferentiated carcinomas without any distin-

guishing features [15].

Immunohistochemical features

Cytokeratins and vimentin are expressed byvirtually all CPPs and most CPCs. Glial fibrillary

acidic protein (GFAP) can be found focally inabout 25% to 55% of CPPs and in 20% of CPCs[16]. Most of the GFAP-positive cells are simul-

taneously positive for cytokeratin [17]. S-100protein is present in almost all cases of CPP and,less frequently, in CPCs. The staining for S-100 is

often stronger and more diffuse than GFAPstaining. Synaptophysin has been suggested asa possible marker for choroid plexus epithelium,but staining of tumors with this marker produces

variable results. Epithelial membrane antigen(EMA) is positive in tumor cells only focally, ifat all. A recent study suggested that immunohis-

tochemical staining for prealbumin and carci-noembryonic acid (CEA) is of significant value forthe differentiation of CPPs and CPCs [18].

Staining for insulin-like growth factor-II (IGF-II) is also a potentially useful marker to distin-guish normal choroid plexus and CPP from CPC[19]. Additionally, indirect indices of proliferation,

such as the Ki-67/MIB-1 antibody, have been usedto distinguish CPP from CPC [20,21]. The meanKi-67/MIB-1 labeling index is often less than 2%

in CPPs and greater than 10% in CPCs. Immu-nohistochemical staining for p53 protein is foundmore often in carcinomas than in CPPs [22].

Ultrastructural features

Most CPPs exhibit apical microvilli withscattered cilia, junctional complexes, interdigitat-ing lateral cell borders, basement membrane, and

fenestrated capillaries. Cilia contain the 9 + 0microtubule configuration characteristic of neuro-epithelial cells. Some tumors have irregularly

shaped structures containing lipid droplets, fila-mentous material, and structures that resemblethe silver bodies of Biondi seen in normal

choroid plexus [23]. CPCs are often more varied intheir ultrastructural appearance and can showepithelial features as well as cilia and microvilli,although such findings are focal in many cases.

CPCs can also display immature cellular features,such as polyribosomes, glycogen granules, andhypertrophied rough endoplasmic reticulum [24].

Fig. 1. (a) Choroid plexus papilloma: low magnification showing well-formed papillae composed of uniform small

epithelial type cells. Mitotic figures and necrosis are rare. (b) Choroid plexus carcinoma: a tumor with irregular

architecture, the presence of marked pleomorphic cells with a less prominent papillary pattern, and frequent mitoses and

necrosis. (c) Ependymoma: medium magnification showing uniform cells arranged in a perivascular fashion. (d )

Ependymoma: high magnification of an ependymal pseudorosette, an angiocentric arrangement of cells with fibrillary

processes perpendicular to the luminal axis. (e) Subependymoma: a paucicellular tumor showing a multinodular compact

architecture without mitotic figures. ( f ) Subependymal giant cell astrocytoma: a tumor composed of gemistocytic

astrocytes, scattered inflammatory cells, and dystrophic calcifications. (g) Central neurocytoma: a tumor typically

described as oligodendroglioma-like with clear cells (fried-egg cells) and a delicate vasculature (chicken-wire

vasculature). (h) Meningioma: a typical meningioma in the lateral ventricle. The tumor shows multiple whorl formation as

well as calcifications known as psammoma bodies.

c

471J.S. Waldron, T. Tihan /Neurosurg Clin N Am 14 (2003) 469482


Molecular and genetic features

In CPP, recurrent abnormalities, includingpartial gains of chromosome 7, have beenreported [25]. A comparative genomic hybridiza-

tion study of a large number of choroid plexustumors showed that +5q, +6q, +7q, +9q,+15q, +18q, and 21q were significantly morecommon in CPPs, whereas CPCs were character-

ized by +1, +4q, +10, +14q, +20q, +21q,5q, 9p, 11, 15q, and 18q [26].

Choroid plexus neoplasms have been associ-

ated with the Li-Fraumeni syndrome as well asthe Aicardi syndrome [2729]. Several reportshave identified SV-40 virus genetic material within

tumor cells; however, the contribution of thisvirus to the formation of choroid plexus tumors isunknown [3032].

Pathologic differential diagnosis

The most frequent challenge during pathologicevaluation is the distinction of CPP from normalchoroid plexus. The normal choroid plexus has

regular single-layered cells with hobnail luminalsurfaces, whereas CPP displays a more crowdedepithelium with significant nuclear variability. The

diagnosis of CPP by the pathologist is unreason-able in the absence of clinical and radiologicfindings, especially without a distinct contrast-enhancing intraventricular mass.

The second diagnostic challenge is the exclu-sion of a rare papillary ependymoma [33].Papillary ependymomas can form epithelial sur-

faces but retain a fibrillary background. Ependy-momas with focal papillary change can bedistinguished by their predominantly glial appear-

ance. In cases where the distinction cannot bemade in routine stains, immunohistochemistryand ultrastructural studies are helpful. Largepartially intraventricular tumors in young patients

with poorly differentiated morphology should alsoraise the possibility of an atypical teratoid/rhabdoid tumor (AT-RT). Some AT-RTs have

been misdiagnosed as CPC in the past. Distinctionof AT-RT can be made by using a panel ofimmunohistochemical markers as well as genetic

studies to confirm the presence of characteristicabnormalities.

CPCs are extremely rare in adults, and

metastatic carcinoma should be viewed as a morelikely cause for an intraventricular papillarycarcinoma. Metastases from the pulmonary andgenitourinary systems have been shown to mimic

CPC [3436]. Distinction may be difficult because

of overlapping histologic, ultrastructural, andimmunohistochemical features. Typical CPC im-munohistochemistry reveals positivity for cyto-

keratin cocktail and absent or only faint EMAand CEA immunoreactivity. If positive, synapto-physin can also be used to distinguish CPC. Inaddition, BerEp4 staining is considered a reliable

marker for most metastatic carcinomas, and itspresence may exclude a CPC (Marc K. Rose-nblum, MD, personal communication, 1999).

Ependymomas

Epidemiology

Ependymomas are neoplasms derived from the

ependymal layer lining the ventricular system andcan occur intracranially and in the spine. In-tracranial ependymomas account for 2% to 8% of

all primary CNS neoplasms [37], with more thanhalf presenting in the first two decades of life. Ina series of 467 pediatric intracranial neoplasms

reviewed by Farwell et al [38], ependymomasmade up 9% of all intracranial tumors, making itthe third most common pediatric intracranial

tumor. Within the pediatric population, ependy-momas favor young patients, with more than 50%occurring within the first 3 years of life. Noconsistent gender predilection has been identified.

Intracranial ependymomas can be divided bylocation into those appearing infratentoriallyand those appearing supratentorially. Infratento-

rial ependymomas make up approximately twothirds of all cases [39], comprise most pediatriccases, and most frequently occur in the fourth

ventricle [40]. Supratentorial ependymomas occurmore frequently in older children and adults. Inaddition to the lateral ventricles, approximately50% of supratentorial ependymomas involve the

parenchyma [41].


Ependymomas are often sharply demarcated,

fleshy, hemorrhagic, soft, and sometimes rubberymasses. Rare examples are heavily calcified, givingthe tumor a gritty texture. Intraventricular exam-

ples of ependymomas are often lobulated anddisplay a discrete interface with surroundingbrain. Some tumors may exhibit a delicate over-

lying ependymal layer that gives them a shinytexture.

Ependymomas are glial neoplasms composedof a monomorphous proliferation of neoplastic

cells with typical perivascular pseudorosettes


(see Fig. 1C, D). Some ependymomas are pre-dominantly glial in appearance and may not havedistinct perivascular pseudorosettes, whereasothers may be predominantly epithelial. The latter

may present as a tumor with oval to round nuclei,discrete cytoplasmic borders, frank papillarystructures, and well-formed fibrovascular cores.

Other tumors may show true ependymal ro-settes distinguished by their well-defined luminaand cells forming pseudoglandular structures.

Most ependymomas show a substantial numberof nuclear grooves that can be identified inintraoperative smears and help with the rapid

interpretation of frozen sections [42]. This feature,however, needs to be interpreted in the context ofother histologic findings, because many othertumors, such as meningiomas and other gliomas,

can exhibit nuclear grooves. The tumor nuclei areuniform, round to oval, and often featurea distinct nucleolus.

Clear cell change in ependymoma is a rare butsignificant finding [43]. Intraventricular ependy-momas may exhibit focal or predominant clear

cell change. When clear cell change is predomi-nant, the hematoxylin-eosin appearance of anoligodendroglioma is recapitulated. It is likely

that many tumors previously reported as in-traventricular oligodendroglioma are examplesof clear cell ependymoma [21]. Clear cell ependy-momas are usually higher grade and exhibit

increased mitotic activity and vascular prolifera-tion. The so-called tanycytic ependymoma isremarkably similar to a pilocytic astrocytoma.

This highly fibrillary tumor has moderate celldensity, spindled cells, and a fascicular architec-ture. It has also been described as a piloid tumor

with ependymal nuclei [44]. The tanycyticependymoma often lacks nuclear pleomorphismor aggressive features, such as mitoses or vascularproliferation. Perivascular pseudorosettes are ru-

dimentary and sometimes absent.Ependymomas are commonly calcified and

rarely exhibit cartilaginous and osseous meta-

plasia. Rare ependymomas contain cytoplasmiceosinophilic granules, clear vacuoles, lipid, ormelanin [45,46].

The current WHO classification defines gradeII ependymomas as tumors with mild cellularpleomorphism, pseudorosettes, or true ependymal

rosettes. The tumors can have occasional mitoticfigures and necrosis without pseudopalisading.Occasional foci of hypercellularity and increasedmitoses are allowed. Anaplastic, high-grade,

or grade III ependymomas have moderate to high

cellularity, increased mitotic figures, and vascularproliferation. Necrosis is often present, either inthe form of geographic necrosis or, rarely, in thepseudopalisading form. Perivascular pseudoro-

settes or occasional true ependymal rosettes canbe found in most high-grade ependymomas. Thereis controversy around whether focal atypia or

anaplasia should elevate a lesion to grade IIIanaplastic ependymoma. Some require atypiaand anaplasia to predominate in the tumor tissue,

whereas others report a less favorable prognosiseven for tumors with focal anaplastic features.


Ependymomas are variably positive for

GFAP, which highlights the fibrillary processesaround vessels. Tumors are diffusely positive forvimentin and stain less avidly with S-100 protein

and neurospecific enolase (NSE). Positive stainingfor epithelial markers, such as EMA and cytoker-atins, has been reported in most posterior fossa

and spinal cord ependymomas [47]. Rare tumorcells, true rosettes, and occasional papillarystructures are EMA-positive.

Studies suggest that high Ki-67/MIB-1 and p53protein positivity might be reliable indicators ofhigh-grade ependymomas [48]. Even though thereseems to be a positive correlation between high-

grade features and the Ki-67/MIB-1 index [49],none of the immunohistochemical variables sig-nificantly correlate with tumor grade. Conversely,

Ki-67/MIB-1 and p53 were reported to correlatewith patient survival [50]. Currently, there is noclear evidence for the utility of these markers in

determination of tumor grade or behavior.


The acellular zones around pseudorosettes arecomposed of large numbers of closely packed,

filament-rich, cytoplasmic processes. Microluminaare often present, even though they may not beobserved by light microscopy [51]. These micro-

lumina contain slender curving microvilli anda variable number of cilia. Bordering cells areconnected by unusually long tight junctions. This

triad (cilia, intracytoplasmic intermediate fila-ments, and cell junctions) makes up the typicalultrastructural components. The epithelioid cellsfound in ependymomas and true rosettes are

characterized by intracellular lumina, cilia, andmicrovilli. Clear cell ependymomas reveal denselypacked polyhedral cells with clear cytoplasm and

well-developed intercellular junctions. Abundant


hyaloplasmic lipid vacuoles can also contribute tothe clear appearance of the tumor cells [46].


There is a body of evidence suggesting the

presence of a tumor suppressor gene on the longarm of chromosome 22 that plays a role in thepathogenesis of ependymomas [52]. In one study,

the most frequent copy number abnormality inependymomas was 22q loss, followed by gain ofchromosome 9 and occasional loss of 6q, 3p, 10q,

and 15q [25]. A heterozygous mutation in theMEN1 gene has also been reported in ependymo-mas [53].


Formulation of the differential diagnosis forependymoma is dependent on the location of thelesion. In the posterior fossa, medulloblastoma

needs to be considered first in the differentialdiagnosis, although its architecture is morereminiscent of a small blue round cell tumor thanthat of a glioma. Pilocytic astrocytoma of the

cerebellum or brain stem is a second possibilitybut can be easily excluded when classic features ofpilocytic astrocytomas, such as Rosenthal fibers,

eosinophilic granular bodies, and a fairly pauci-cellular appearance, are present. Infiltratingastrocytomas or the so-called brain stem glio-

mas may have an exophytic quality and mayresemble ependymoma. They are easily distin-guished by their invasive quality, lack of epithelial

features or pseudorosettes, and marked nuclearpleomorphism.

Supratentorial intraventricular ependymomasneed to be distinguished from subependymomas.

Such distinction is often subjective and may notalways translate into a significant change inclinical outcome. Nevertheless, based on the

overall clinical behavior of ependymomas andthe likelihood of supratentorial examples beinghigher grade, one is compelled to make the

distinction. The distinction is usually not difficult,and the differential diagnosis is confounded bylimited tissue sample size. A second yet more

important differential diagnosis is oligodendro-glioma, which can easily be confused with clearcell ependymoma. Clear cell ependymomas arenoninfiltrating, solid, and distinct from the

surrounding brain. Purely intraventricular neo-plasms are not likely to be oligodendrogliomas,but when a question exists, immunohistochemis-

try and electron microscopy readily settle the

issue. Another diagnostic consideration is thecentral neurocytoma. The central neurocytoma isa highly cellular neoplasm that may show

perivascular pseudorosettes. The cells appearmore neurocytic, and the fibrillar areas resembleneuropil. The tumor strongly reacts with synap-tophysin and only weakly (if at all) with GFAP.

Electron microscopy can distinguish the twoentities. Papillary ependymomas may resembleCPP. The overall immunohistochemical profile

and ultrastructural features can be used toseparate the two entities.

Subependymoma

Epidemiology

Subependymomas are slow-growing, benignintraventricular lesions first identified as a separate

entity in 1945 by Scheinker [54]. They originate inthe subependymal glial matrix and typically pro-ject into the ventricular lumen. Intracranial

subependymomas frequently remain asymptom-atic and are documented on autopsy or as anincidental finding on imaging. A prevalence of

0.4% has been reported in a series of 1000necropsies of asymptomatic patients reviewed byMatsumura and colleagues [55]. Subependymo-mas have been reported over a wide age range, but

generally occur in middle-aged to older adults.The fourth ventricle, followed by the lateralventricles, is the most common site of presenta-

tion. Less common locations include the thirdventricle, the septum pellucidum, and the cerebralaqueduct.


Subependymomas are solid nodular tumorsfirmly attached to the ventricular surface. Tumors

are typically soft but can be rubbery and, rarely,cystic and occasionally have a gritty texture.

Subependymomas are typically paucicellular,

fibrillar, and markedly nodular neoplasms (seeFig. 1E). Tumor nuclei cluster within the nodularregions. Supratentorial tumors, especially those

near the foramen of Monro, are predominantlymicrocystic and focally myxoid. The cells areoften spindled with oval nuclei and fibrillaryprocesses. The tumor exhibits an extensive fibril-

lary background on intraoperative smear prepa-rations. Nuclear pleomorphism is rare, andmitoses are typically absent. Tumor vessels show

focal hyalinization with occasional hemosiderindeposition. There is some evidence that the


histologic features in larger and symptomaticsubependymomas may be different. Larger symp-tomatic tumors more frequently demonstrate cystformation, microcalcification, and vessel degener-

ation accompanied by hemorrhage [56]. Subepen-dymomas in the posterior fossa are usually smallerwithout significant microcystic change. The tu-

mors are prominently nodular and show nuclearclustering and calcifications. Mitoses are rare, andvascular proliferation and necrosis are absent.


Subependymomas are strongly GFAP-positivein accord with their high content of intermediate

glial filaments. Vimentin staining is often strong,and S-100 protein stains the cytoplasm and thenuclei. Compared with other ependymal tumors,

subependymomas have the lowest rate of cellproliferation, as evidenced by a Ki-67/MIB-1index of less than 1% [57]. In contrast to

ependymomas, staining with epithelial markers,such as EMA, is usually not observed.


Subependymomas display an abundance ofclosely packed cell processes filled with interme-diate filaments. This meshwork of processes

widely separates small clusters of tumor cells.Larger cells lacking specialized features andresembling ependymal precursor cells are oftenfound. Other cells with transitional forms between

these two types can be identified [58]. Pockets ofmicrovilli are present, but they differ fromependymal type rosettes because of the lack of

tight junctions.


There is limited information on the cytogenet-

ics and molecular genetics of subependymomas. Afew case reports have demonstrated a normalkaryotype in subependymomas investigated with

conventional cytogenetic techniques [59].


The main component of the differential di-

agnosis for intraventricular subependymoma isthe classic ependymoma. It may not be possible todistinguish all cases, especially if the amount of

tissue available for pathologic analysis is limited.Both neoplasms appear remarkably similar, andfoci identical to subependymoma are commonly

seen in ventricular ependymomas. In general,ependymomas can be distinguished by occurrenceprimarily in children, hypercellularity, perivascu-lar pseudorosettes, and true ependymal rosettes.

Most supratentorial ependymomas have a solidcystic appearance and are symptomatic. Subepen-dymomas may also be confused with the tanycytic

variant of ependymomas. Often, tanycytic epen-dymomas are more cellular and resemble pilocyticastrocytomas. Ultrastructural examination of

tanycytic ependymoma reveals characteristicependymal features, including intracytoplasmicintermediate filaments, prominent intercellular

junctions, numerous slender surface microvilli,and microvilli-lined lumina.

Subependymal giant cell astrocytoma

Epidemiology

SEGAs are intimately associated with thetuberous sclerosis complex, an autosomal domi-

nant dysgenetic syndrome that is associated withthe classic triad of seizures, mental retardation,and papular facial lesions. In the CNS, the

complex is characterized by cortical tubers,subependymal nodules, and SEGAs. The inci-dence of tuberous sclerosis is approximately1:10,000 [60] in the general population. Approx-

imately 6% [61] of these patients develop SEGAs.Almost all SEGAs arise near the foramen ofMonro and typically present with hydrocephalus

or increased seizure frequency, most commonlywithin the first two decades of life [61].


SEGAs are well-defined typically pedunculated

intraventricular masses that can be soft to rubberyand often have a broad base on the ventricularsurface. The tumor can be easily removed from its

base. Tumors may be friable, pink as a result ofvascularization, and occasionally gritty fromcalcification.

SEGAs are characteristically solid and havea typical swirling architecture. They exhibitcompact growth with spindled and gemistocytic

cells and are sharply demarcated from theadjacent normal parenchyma (see Fig. 1F).Spindled cells are responsible for the swirlingappearance of the tumor on low magnification.

The gemistocyte-like cells have round vesicularnuclei with distinct nucleoli and an eosinophiliccytoplasm. In addition, they display thick hairlike

processes and have a tendency to form cohesive


clusters and occasional pseudorosettes [62]. Thetumor maintains a compact and uniform appear-ance despite varying tumor cell types. SEGAs

often contain inflammatory cells, including occa-sional mast cells. Even though rare mitotic figuresare occasionally seen, brisk mitotic activity, nec-rosis, or vascular proliferation is typically absent.

Some tumors undergo focal infarction, which canappear ominous and be confused with the necrosischaracteristic of a high-grade astrocytic neoplasm.

Calcification is sometimes present.


The gemistocytic and spindle cells are oftenstrongly positive for GFAP; however, the abso-

lute number of positive cells in each tumor ishighly variable. SEGAs also show strong positiv-ity for S-100 protein. Neurofilament epitopes,

class III b-tubulin, and calbindin 28-kDa areexpressed in some cases [63]. Cytoplasmic stainingfor somatostatin, met-enkephalin, 5-hydroxytryp-

tamine, b-endorphin, and neuropeptide Y has alsobeen noted in more than half of cases of SEGA[63,64]. The divergent glial and neuronal staining

has been shown to colocalize within the same cell.SEGAs are negative for HMB-45 antibody, andthe Ki-67/MIB-1 labeling index is usually lessthan 2% [65].


SEGAs contain numerous intermediate fila-ments, frequent lysosomes, and occasional rectan-

gular or rhomboid membrane-bound crystalloidsthat exhibit lamellar periodicity and structuraltransition to lysosomes. Microtubules and stacksof rough endoplasmic reticulum are common, but

true neuronal differentiation, such as neurosecre-tory granules or synaptic formations, is oftenabsent [63]. Rare tumor cells have features

suggestive of neuronal differentiation, includingstacks of rough endoplasmic reticulum, occasionalmicrotubules, and a few poorly defined dense core

granules. Gemistocytic cells are characterized byabundant intermediate filaments within the cellbody and the processes. Lysosomes are common

and, rarely, may contain distinctive membrane-bound crystalloids.


Cytogenetic analysis of SEGAs within thetuberous sclerosis complex (TSC) reveals clonalchromosomal changes, resulting in the partial loss

of chromosome 22q in some tumors [66]. TSC-associated tumors also demonstrate loss ofheterozygosity in chromosomes 9 and 16, which

are known to harbor TSC genes [67]. One of twosuspected genes, TSC2, was found in chromosome16 by positional cloning. The gene product fromTSC2 has been named tuberin. TSC1 was

discovered earlier in chromosome 9 but has notyet been characterized. Genetic analysis on TSCfamilies reveals mutations in chromosome 9q34

(TSC1) and chromosome 16p13 (TSC2) as theonly common genetic anomalies [68].

The Eker rat, a naturally occurring animal

model of TSC, provides a powerful tool forinvestigations of TSC. In this model, a conservedlinkage group on rat 10q corresponds to human16p13.3 (TSC2 gene) [69]. Currently, it is believed

that the products of TSC1 and TSC2 genesinteract with each other in the cell.


SEGAs are fairly distinct intraventricular neo-plasms that may be confused with gemistocyticastrocytoma or high-grade glioma if the typical

pathologic and radiologic features are overlooked.Small biopsies can also potentially be interpretedas tanycytic ependymoma or subependymoma,

but this is less likely, because SEGAs are in-variably more cellular, less fibrillary, and far moregemistocytic.

Central neurocytoma

Epidemiology

The term central neurocytoma was first used byHassoun et al [70] in 1982 to describe differenti-ated intraventricular neuronal lesions observed in

2 cases. Central neurocytomas are rare neoplasms,with 127 reported cases through 1993 [71].Reported rates in series of pathologically con-firmed primary CNS neoplasms range from 0.1%

to 0.5% [7274]. Central neurocytomas areprimarily tumors of young adults, with 45%occurring in the third decade of life and almost

75% between the ages of 20 and 40 years [71].Gender distribution is equal. Central neuro-cytomas arise predominantly from the septum

pellucidum or, less frequently, from the lateral ven-tricular wall. The anterior lateral ventricle is themost frequent site (77%), followed by lateral andthird ventricle involvement (21%) [71]. Bilateral

lateral ventricular involvement is uncommon.


Rare cases have been reported in the third andfourth ventricles.


The tumor forms a soft to gritty, tan, discrete

mass that may be solid or partly cystic.Central neurocytomas are histologically and

cytologically uniform neoplasms. The cells arestrikingly monomorphous with finely distributed

chromatin and a fine fibrillary matrix. The tumoris one of several neuroepithelial neoplasms withsalt and pepper chromatin. Central neuro-

cytoma joins the list of oligodendroglioma-liketumors because of a striking preponderance ofcells with perinuclear halos that resemble classic

oligodendroglioma (see Fig. 1G). In some cases,there are perivascular fibrillary zones reminiscentof ependymal pseudorosettes. In addition, someexamples resemble nodular medulloblastomas by

exhibiting neurocytic differentiation with cellstreaming and nodular growth.

Intraoperative frozen sections can sometimes

obscure the histologic and cytologic uniformitytypical of central neurocytomas. The processingof frozen tissue also adds a degree of nuclear

pleomorphism that can raise the possibility ofa small blue round cell tumor. This is furtherconfounded in permanent sections of frozen tissue

because of the obscured neuronal/neurocyticbackground. Most central neurocytomas aregrade II lesions with minimal nuclear pleomor-phism and rare mitotic figures. Tumors with

atypical features and transitional characteristicsbetween neurocytoma and neuroblastoma havebeen reported [75].

Central neurocytomas may show ganglioniccell differentiation and have a preponderance ofneuropil with variable numbers of ganglion-like

cells. Such cases have been designated as gan-glioneurocytoma or differentiated neurocy-toma. An intraventricular lesion that combinesthe features of a neurocytoma with ganglion cells

and a malignant small cell component has beenreported but is extremely rare. It has also beensuggested that some central neurocytomas can

express photoreceptor differentiation, potentiallyrelating them to pineocytomas [76]. Rare centralneurocytomas exhibit lipofuscin or neuromelanin

pigment [77].


Central neurocytomas consistently exhibitimmunoreactivity for NSE and synaptophysin,

indicating neuronal differentiation [78]. Synapto-physin antibody stains the fibrillar zones and, toa lesser extent, the perinuclear cytoplasm of tumorcells. Anti-Hu autoantibodies stain neurocyte

nuclei. Tumor cells are also positive for Leu-7 andS-100 protein, whereas staining for GFAP is pre-dominantly negative andvimentin is confined to the

nonneoplastic mesenchymal elements of bloodvessels [79]. Staining for myelin basic protein,chromogranin, andneurofilament is often negative.

Some studies have shown a small subpopulation ofGFAP-positive neoplastic cells, and glial differen-tiation has been suggested in tissue culture. This

mixed phenotype of glial and neuronal markerpositivity in central neurocytoma can be interpre-ted as a glioneuronal neoplasm, with an overwhel-mingly neurocytic component. In rare examples,

a tumormayhave an increasedKi-67/MIB-1 index.Such neoplasms are described as atypical neuro-cytomas and have a significantly elevated in-

cidence of local recurrence [80]. Even though noclear cutoff point exists between classic and atypicalneurocytomas, most authors suggest that tumors

with an MIB-1 index of greater than 2% be placedin the atypical category. Nevertheless, some studiesshow no difference in MIB-1 labeling between

tumors with atypical features and typical centralneurocytomas [81]. CurrentlyMIB-1 labeling is notused to modify grading of central neurocytomas.


Central neurocytoma is readily recognizable asneuronal, with microtubules, terminations, clearvesicles, and dense core granules [79,82,83]. Some

examples may display round cells with abundantcell processes containing microtubules, cellularjunctions, and lysosome-like structures. Otherscontain numerous synaptic vesicles, neuritic pro-

cesses, and neurosecretory granules. In addition,rare tumors contain ganglionic cells with well-developed processes.


Reported recurrent genetic changes in centralneurocytomas include alterations on chromo-

somes 2p, 10q, and 18q. The candidate genes inthese loci are currently unknown [84]. Otherstudies have suggested gain of chromosome 7 asa nonrandom genetic alteration in central neuro-

cytomas [85]. Recent studies have demonstratedthat central neurocytomas are genetically distinctfrom oligodendrogliomas and that chromosomes

1p and 19q probably do not play an important


role in their pathogenesis. In addition, N-myc andepidermal growth factor receptor amplificationsare rare or absent in these tumors [86].


The critical differentiation for central neuro-cytoma is from oligodendroglioma. Some cases ofcentral neurocytoma perfectly recapitulate oligo-

dendroglioma in routine microscopic examina-tion. In such cases, it is important to use a panelof immunohistochemical stains and to perform an

ultrastructural examination to establish the cor-rect diagnosis. Furthermore, radiologic informa-tion should be critically interpreted and thediagnosis of oligodendroglioma challenged in

purely intraventricular tumors. A second entityin the differential diagnosis is the clear cellependymoma, which also exhibits a remarkable

resemblance to oligodendroglioma. The presenceof ependymal features as well as immunohisto-chemical analysis should distinguish a clear cell

ependymoma from central neurocytoma. Smallbiopsies from a dysembryoplastic neuroepithelialtumor may also mimic central neurocytoma, butexclusive intraventricular location, the absence of

floating neurons, and the immunohistochemicalprofile should distinguish between the two. Lastly,the presence of an intraventricular clear cell

neoplasm in older patients should raise thepossibility of a metastatic lesion, especially a renalcell carcinoma. Often, the highly anaplastic

histologic features are sufficient to distinguisha renal carcinoma metastasis from a classic centralneurocytoma. Additional immunohistochemical

studies can be used to provide further support.

Other tumors and tumor-like lesions within the

ventricular system

Other purely intraventricular tumors and tu-mor-like lesions are rare. One example includes theintraventricular meningioma [8789]. The intra-

ventricular location is uncommon, with an ap-proximate incidence of 0.5% to 4.5% among allintracranial meningiomas [89]. Intraventricular

meningiomas are more common in adults becauseof the higher overall frequency of meningiomas butmake up a larger percentage of meningiomas in

the pediatric population [9092]. Meningiomascan arise anywhere in the ventricular system andexhibit the histologic features common to allmeningiomas (see Fig. 1H). Rare cases of intraven-

tricular clear cell meningioma [93] and malignantmeningioma [94] have been reported.

Intraventricular metastases from epithelial

malignancies are extremely rare but can mimica choroid plexus tumor clinically and pathologi-cally. Intraventricular metastases originate froma number of cancers, including renal cell carci-

noma [35,95,96], pulmonary adenocarcinoma [34],gastric carcinoma [97], adrenocortical carcinoma[98], and bladder carcinoma [36]. In such cases,

immunohistochemical analysis, including a cyto-keratin panel, can help to identify the nature ofthe neoplasm and differentiate such tumors from

CPCs [99].Rare cases of perineurioma from the choroid

plexus of the third ventricle, malignant schwan-noma, solitary fibrous tumors, and hemangio-

pericytoma have been reported as purelyintraventricular tumors [100102].

A diverse list of cystic tumor-like lesions can

exist within the ventricular system and can beconfused with a neoplasm [103]. Colloid cysts ofthe third ventricle [104], ependymal or glioepen-

dymal cysts [105], choroid plexus cysts [106],arachnoid cysts [107], and cavernous angioma[108] have been reported as intraventricular

masses. Choroid plexus cysts are more commonin fetuses with chromosomal aneuploidies, partic-ularly trisomy 18.

Inflammatory or infectious processes can also

present as purely intraventricular masses thatresemble tumors. Such a presentation is muchless common than the usual parenchymal or

leptomeningeal forms. Reports of infectious orinflammatory processes that present as masseswithin the ventricular system include cysticercosis

[109], cryptococci [110], and nocardiosis [111]among others.

Summary

Tumors that primarily or exclusively involvethe ventricular system constitute a rare and

heterogeneous group. Certain histologic tumortypes predominantly occur in children, whereasothers are more common in adults. Tumor

location provides additional clues to correctdiagnosis. When used in conjunction with clinicaland radiologic data, histopathologic features can

distinguish among this wide range of possibilitiesto provide the correct diagnosis for optimalpatient management.


Acknowledgement

J.S. Waldron was supported in part by a grantfrom the Khatib Research Foundation as a Khatib

Fellow 20022003.

References

[1] Janisch W, Staneczek W. Primary tumors of the

choroid plexus. Frequency, localization and age

[in German]. Zentralbl Allg Pathol 1989;135(3):

23540.

[2] Schiffer D. Brain tumors. Pathology and biological

correlates. Berlin: Springer Verlag; 1993. p. 21721.

[3] Zulch K. Brain tumors. Their biology and pathol-

ogy. 3rd edition. Berlin: Springer Verlag; 1986.

[4] Wolff JE, et al. Choroid plexus tumours. Br J

Cancer 2002;87(10):108691.

[5] Matson DD, Crofton FDL. Papilloma of the

choroid plexus in childhood. J Neurosurg 1960;

17:100227.

[6] Burger PC, Scheithauer BW. Tumors of the central

nervous system. Washington, DC: Armed Forces

Institute of Pathology; 1994.

[7] Kimura M, et al. Primary choroid plexus papillo-

ma located in the suprasellar region: case report.

Neurosurgery 1992;31(3):5636.

[8] Pierga JY, et al. Carcinoma of the choroid plexus:

a pediatric experience. Med Pediatr Oncol 1993;

21(7):4807.

[9] Kepes JJ. Xanthomatous changes in a papilloma

of the choroid plexus. Acta Neuropathol (Berl)

1970;16(4):3679.

[10] Nakashima N, et al. Choroid plexus papilloma.

Light and electron microscopic study. Virchows

Arch A Pathol Anat Histopathol 1983;400(2):

20111.

[11] Cardozo J, et al. Choroid plexus papilloma

containing bone. Acta Neuropathol (Berl) 1985;

68(1):835.

[12] Bonnin JM, Colon LE, Morawetz RB. Focal glial

differentiation and oncocytic transformation in

choroid plexus papilloma. Acta Neuropathol (Berl)

1987;72(3):27780.

[13] Stefanko SZ, Vuzevski VD. Oncocytic variant of

choroid plexus papilloma. Acta Neuropathol (Berl)

1985;66(2):1602.

[14] Reimund EL, Sitton JE, Harkin JC. Pigmented

choroid plexus papilloma. Arch Pathol Lab Med

1990;114(8):9025.

[15] Paulus W, Janisch W. Clinicopathologic correla-

tions in epithelial choroid plexus neoplasms:

a study of 52 cases. Acta Neuropathol (Berl) 1990;

80(6):63541.

[16] Rickert CH, Paulus W. Tumors of the choroid

plexus. Microsc Res Tech 2001;52(1):10411.

[17] Kouno M, et al. An immunohistochemical study

of cytokeratin and glial fibrillary acidic protein in

choroid plexus papilloma. Acta Neuropathol

(Berl) 1988;75(3):31720.

[18] Kato T, et al. Clinicopathological study of choroid

plexus tumors: immunohistochemical features and

evaluation of proliferative potential by PCNA and

Ki-67 immunostaining. Noshuyo Byori 1996;13(2):

99105.

[19] Kubo S, et al. Immunocytochemical detection of

insulin-like growth factor II (IGF-II) in choroid

plexus papilloma: a possible marker for differential

diagnosis. Clin Neuropathol 1999;18(2):749.

[20] Coons S, et al. Choroid plexus carcinoma in

siblings: a study by light and electron microscopy

with Ki-67 immunocytochemistry. J Neuropathol

Exp Neurol 1989;48(4):48393.

[21] Carlotti CG Jr, et al. Evaluation of proliferative

index and cell cycle protein expression in choroid

plexus tumors in children. Acta Neuropathol (Berl)

2002;103(1):110.

[22] Jay V, et al. P53 expression in choroid plexus

neoplasms: an immunohistochemical study. Arch

Pathol Lab Med 1996;120(11):10615.

[23] Navas JJ, Battifora H. Choroid plexus papilloma:

light and electron microscopic study of three cases.

Acta Neuropathol (Berl) 1978;44(3):2359.

[24] Anguilar D, et al. The fine structure of choroid

plexus carcinoma. Histopathology 1983;7(6):

93946.

[25] Grill J, et al. Comparative genomic hybridization

detects specific cytogenetic abnormalities in pedi-

atric ependymomas and choroid plexus papillo-

mas. Cancer Genet Cytogenet 2002;136(2):1215.

[26] Rickert CH, Wiestler OD, Paulus W. Chromo-

somal imbalances in choroid plexus tumors. Am

J Pathol 2002;160(3):110513.

[27] Yuasa H, Tokito S, Tokunaga M. Primary

carcinoma of the choroid plexus in Li-Fraumeni

syndrome: case report. Neurosurgery 1993;32(1):

1314.

[28] Uchiyama CM, et al. Choroid plexus papilloma

and cysts in the Aicardi syndrome: case reports.

Pediatr Neurosurg 1997;27(2):1004.

[29] Vital A, et al. Astrocytomas and choroid plexus

tumors in two families with identical p53 germline

mutations. J Neuropathol Exp Neurol 1998;

57(11):10619.

[30] Palmiter RD, et al. SV40 enhancer and large-T

antigen are instrumental in development of cho-

roid plexus tumours in transgenic mice. Nature

1985;316(6027):45760.

[31] Van Dyke TA, et al. Relationship between simian

virus 40 large tumor antigen expression and tumor

formation in transgenic mice. J Virol 1987;

61(6):202932.

[32] Chen J, et al. T-antigen mutant activities in vivo:

roles of p53 and pRB binding in tumorigenesis of

the choroid plexus. Oncogene 1992;7(6):116775.

[33] Park SH, Park HR, Chi JG. Papillary ependy-

moma: its differential diagnosis from choroid


plexus papilloma. J Korean Med Sci 1996;11(5):

41521.

[34] Tanimoto M, et al. Choroid plexus metastasis of

lung carcinomacase report. Neurol Med Chir

(Tokyo) 1991;31(3):1525.

[35] Mizuno M, et al. Renal cell carcinoma metasta-

sizing to choroid plexus of lateral ventricle; a case

report [in Japanese]. No Shinkei Geka 1992;

20(4):46974.

[36] Qasho R, et al. Choroid plexus metastasis from

carcinoma of the bladder: case report and review

of the literature. J Neurooncol 1999;45(3):23740.

[37] Duncan JAI, Hoffman HJ. Intracranial ependy-

momas. In: Kaye AH, editor. Brain tumors.

Edinburgh: Churchill Livingstone; 1995. p. 493

504.

[38] Farwell JR, Dohrmann GJ, Flannery JT. Central

nervous system tumors in children. Cancer 1977;

40(6):312332.

[39] Mork SJ, Loken AC. Ependymoma: a follow-up

study of 101 cases. Cancer 1977;40(2):90715.

[40] Schiffer D, et al. Histologic prognostic factors in

ependymoma. Childs Nerv Syst 1991;7(4):17782.

[41] Schwartz TH, et al. Supratentorial ependymomas

in adult patients. Neurosurgery 1999;44(4):72131.

[42] Kumar PV. Nuclear grooves in ependymoma.

Cytologic study of 21 cases. Acta Cytol 1997;

41(6):172631.

[43] Kawano N, Yada K, Yagishita S. Clear cell

ependymoma. A histological variant with diagnos-

tic implications. Virchows Arch A Pathol Anat

Histopathol 1989;415(5):46772.

[44] Langford LA, Barre GM. Tanycytic ependymoma.

Ultrastruct Pathol 1997;21(2):13542.

[45] Rosenblum MK, et al. Melanotic ependymoma

and subependymoma. Am J Surg Pathol 1990;

14(8):72936.

[46] Severi B, et al. Ependymoma of the foramen of

Monro: ultrastructural characterization. Ultra-

struct Pathol 1989;13(1):3542.

[47] Takeuchi H, et al. Epithelial differentiation and

proliferative potential in spinal ependymomas.

J Neurooncol 2002;58(1):139.

[48] Rushing EJ, et al. Correlation of bcl-2, p53, and

MIB-1 expression with ependymoma grade and

subtype. Mod Pathol 1998;11(5):46470.

[49] Ritter AM, et al. Ependymomas: MIB-1 pro-

liferation index and survival. J Neurooncol 1998;

40(1):517.

[50] Verstegen MJ, et al. Proliferation- and apoptosis-

related proteins in intracranial ependymomas: an

immunohistochemical analysis. J Neurooncol

2002;56(1):218.

[51] Sara A, Bruner JM, Mackay B. Ultrastructure

of ependymoma. Ultrastruct Pathol 1994;18(1 2):

3342.

[52] Park JP, et al. Constitutional de novo

t(1;22)(p22;q11.2) and ependymoma. Cancer

Genet Cytogenet 1996;86(2):1502.

[53] Urioste M, et al. Complex cytogenetic abnormal-

ities including telomeric associations and MEN1

mutation in a pediatric ependymoma. Cancer

Genet Cytogenet 2002;138(2):10710.

[54] Scheinker I. Subependymoma: a newly recognized

tumor of subependymal derivation. J Neurosurg

1945;2:23240.

[55] Matsumura A, et al. Intracerebral subependymo-

mas. Clinical and neuropathological analyses with

special reference to the possible existence of a less

benign variant. Acta Neurochir (Wien) 1989;96

(1 2):1525.

[56] Scheithauer BW. Symptomatic subependymoma.

Report of 21 cases with review of the literature.

J Neurosurg 1978;49(5):68996.

[57] Prayson RA, Suh JH. Subependymomas: clinico-

pathologic study of 14 tumors, including compar-

ative MIB-1 immunohistochemical analysis with

other ependymal neoplasms. Arch Pathol Lab

Med 1999;123(4):3069.

[58] Moss TH. Observations on the nature of sub-

ependymoma: an electron microscopic study.

Neuropathol Appl Neurobiol 1984;10(1):6375.

[59] Dal Cin P, et al. Cytogenetic investigation in

subependymoma. Cancer Genet Cytogenet 1999;

108(1):84.

[60] Wiederholt WC, Gomez MR, Kurland LT. In-

cidence and prevalence of tuberous sclerosis in

Rochester, Minnesota, 1950 through 1982. Neu-

rology 1985;35(4):6003.

[61] Shepherd CW, et al. Subependymal giant cell

astrocytoma: a clinical, pathological, and flow

cytometric study. Neurosurgery 1991;28(6):8648.

[62] Altermatt HJ, Scheithauer BW. Cytomorphology

of subependymal giant cell astrocytoma. Acta

Cytol 1992;36(2):1715.

[63] Hirose T, et al. Tuber and subependymal giant cell

astrocytoma associated with tuberous sclerosis: an

immunohistochemical, ultrastructural, and immu-

noelectron and microscopic study. Acta Neuro-

pathol (Berl) 1995;90(4):38799.

[64] Lopes MB, et al. Immunohistochemical character-

ization of subependymal giant cell astrocytomas.

Acta Neuropathol (Berl) 1996;91(4):36875.

[65] Gyure KA, Prayson RA. Subependymal giant cell

astrocytoma: a clinicopathologic study with

HMB45 and MIB-1 immunohistochemical analy-

sis. Mod Pathol 1997;10(4):3137.

[66] Debiec-Rychter M, et al. Cytogenetic changes in

two cases of subependymal giant-cell astro-

cytoma. Cancer Genet Cytogenet 1999;109(1):

2933.

[67] Sampson JR, Harris PC. The molecular genetics

of tuberous sclerosis. Hum Mol Genet 1994;3:

147780.

[68] Janssen B, et al. Refined localization of TSC1 by

combined analysis of 9q34 and 16p13 data in 14

tuberous sclerosis families. Hum Genet 1994;

94(4):43740.


[69] Hino O, et al. The predisposing gene of the Eker

rat inherited cancer syndrome is tightly linked to

the tuberous sclerosis (TSC2) gene. Biochem

Biophys Res Commun 1994;203(2):13028.

[70] Hassoun J, et al. Central neurocytoma. An

electron-microscopic study of two cases. Acta

Neuropathol (Berl) 1982;56(2):1516.

[71] Hassoun J, et al. Central neurocytoma: a synopsis

of clinical and histological features. Brain Pathol

1993;3(3):297306.

[72] Maiuri F, et al. Central neurocytoma: clinico-

pathological study of 5 cases and review of the liter-

ature. Clin Neurol Neurosurg 1995;97(3):21928.

[73] Kim DG, et al. Intraventricular neurocytoma:

clinicopathological analysis of seven cases. J

Neurosurg 1992;76(5):75965.

[74] Yasargil MG, et al. Central neurocytoma: histo-

pathological variants and therapeutic approaches.

J Neurosurg 1992;76(1):327.

[75] Favereaux A, et al. Histopathological variants of

central neurocytoma: report of 10 cases. Ann

Pathol 2000;20(6):55863.

[76] Mena H, et al. Central neurocytomas express

photoreceptor differentiation. Cancer 2001;91(1):

13643.

[77] Ng TH, et al. Pigmented central neurocytoma: case

report and literature review. Am J Surg Pathol

1999;23(9):113640.

[78] von Deimling A, et al. Patterns of differentiation in

central neurocytoma. An immunohistochemical

study of eleven biopsies. Acta Neuropathol (Berl)

1990;79(5):4739.

[79] Kubota T, et al. Central neurocytoma: immuno-

histochemical and ultrastructural study. Acta

Neuropathol (Berl) 1991;81(4):41827.

[80] Mackenzie IR. Central neurocytoma: histologic

atypia, proliferation potential, and clinical out-

come. Cancer 1999;85(7):160610.

[81] SharmaMC, etal.A studyofproliferativemarkers in

central neurocytoma. Pathology 1998;30(4):3559.

[82] Robbins P, et al. Central neurocytoma. A clinico-

pathological, immunohistochemical and ultra-

structural study of 7 cases. Pathol Res Pract

1995;191(2):10011.

[83] Tsuchida T, et al. Neuronal and glial character-

istics of central neurocytoma: electron microscop-

ical analysis of two cases. Acta Neuropathol (Berl)

1996;91(6):5737.

[84] Yin XL, et al. Detection of chromosomal imbalan-

ces in central neurocytomas by using comparative

genomic hybridization. J Neurosurg 2000;93(1):

7781.

[85] Taruscio D, et al. Nonrandom gain of chromo-

some 7 in central neurocytoma: a chromosomal

analysis and fluorescence in situ hybridization

study. Virchows Arch 1997;430(1):4751.

[86] Tong CY, et al. Central neurocytomas are genet-

ically distinct from oligodendrogliomas and neu-

roblastomas. Histopathology 2000;37(2):1605.

[87] Hoffman JC, Jr, Bufkin WJ, Richardson HD.

Primary intraventricular meningiomas of the

fourth ventricle. Am J Roentgenol Radium Ther

Nucl Med 1972;115(1):1004.

[88] Criscuolo GR, Symon L. Intraventricular meningi-

oma. A review of 10 cases of the National Hospital,

Queen Square (19741985) with reference to the lit-

erature.ActaNeurochir (Wien) 1986;83(34):8391.

[89] Imielinski BL, Kloc W. Meningiomas of the lateral

ventricles of the brain. Zentralbl Neurochir 1997;

58(4):17782.

[90] Vassilouthis J, Ambrose JA. Intraventricular me-

ningioma in a child. Surg Neurol 1978;10(2):1057.

[91] Diaz P, et al. Multiple meningiomas of the fourth

ventricle in infancy: case report. Neurosurgery

1990;26(6):105760.

[92] Sgouros S, Walsh AR, Barber P. Intraventricular

malignant meningioma in a 6-year-old child. Surg

Neurol 1994;42(1):415.

[93] Kakita A, et al. Clear cell variants of intracranial

tumors: meningioma and ependymoma. Noshuyo

Byori 1995;12(2):1116.

[94] Kamiya K, Inagawa T, Nagasako R. Malignant

intraventricular meningioma with spinal metasta-

sis through the cerebrospinal fluid. Surg Neurol

1989;32(3):2138.

[95] Killebrew K, et al. Metastatic renal cell carcinoma

mimicking a meningioma. Neurosurgery 1983;

13(4):4304.

[96] Raila FA, Bottoms WT Jr, Fratkin JD. Solitary

choroid plexus metastasis from a renal cell

carcinoma. South Med J 1998;91(12):115962.

[97] Nakabayashi H, et al. Choroid plexus metastasis

from gastric cancercase report. Neurol Med

Chir (Tokyo) 1994;34(3):1836.

[98] Piniella AM, Siatkowski RM. Adrenal cortical

carcinoma metastatic to the brain in a child.

J Neuroophthalmol 2000;20(1):357.

[99] Gyure KA, Morrison AL. Cytokeratin 7 and 20

expression in choroid plexus tumors: utility in

differentiating these neoplasms from metastatic

carcinomas. Mod Pathol 2000;13(6):63843.

[100] Giannini C, et al. Intraventricular perineurioma:

case report. Neurosurgery 1998;43(6):147882.

[101] Jung JM, et al. Malignant intraventricular schwan-

noma. Case report. J Neurosurg 1995;82(1):1214.

[102] Abrahams JM, et al. Hemangiopericytoma of the

third ventricle. Case report. J Neurosurg 1999;

90(2):35962.

[103] Hirano A, Hirano M. Benign cystic lesions in the

central nervous system. Light and electron micro-

scopic observations of cyst walls. Childs Nerv Syst

1988;4(6):32533.

[104] Jeffree RL, Besser M. Colloid cyst of the third

ventricle: a clinical review of 39 cases. J Clin

Neurosci 2001;8(4):32831.

[105] TillichM, et al. Symptomatic neuroepithelial (epen-

dymal) cyst of the fourth ventricle:MR appearance.

AJR Am J Roentgenol 1999;172(2):5534.


[106] GrattonRJ,HoggeWA,AstonCE. Choroid plexus

cysts and trisomy 18: risk modification based on

maternal age and multiple-marker screening. Am J

Obstet Gynecol 1996;175(6):14937.

[107] Lee KS, Bae HG, Yun IG. Intraventri-

cular arachnoid cyst. J Neurosurg 1989;70(1):

1545.

[108] Suess O, Hammersen S, Brock M. Intraventricular

cavernoma: unusual occurrence in the region of

the foramen of Monro. Br J Neurosurg 2002;16(1):

789.

[109] Madrazo I, et al. Intraventricular cysticercosis.


[110] Vender JR, et al. Intraventricular cryptococcal

cysts. AJNR Am J Neuroradiol 1996;17(1):1103.

[111] Mogilner A, et al. Nocardia abscess of the choroid

plexus: clinical and pathological case report.


Neurosurg Clin N Am 14 (2003) 483508

Intraventricular neurocytomasJanet Lee, MS, Susan M. Chang, MD, Michael W. McDermott, MD,

Andrew T. Parsa, MD, PhD*Department of Neurological Surgery, University of California at San Francisco, 505 Parnassus Avenue,

M-779, San Francisco, CA 94143, USA

Background

Epidemiology and clinical presentation

A review of 385 reported neurocytoma cases[196] shows that, in general, central neuro-

cytomas (CNCs) are well-differentiated intraven-tricular tumors that affect young adult men andwomen equally. Most commonly, CNCs occur in

the anterior portion of the lateral ventricle aroundthe foramen of Monro and can attach to either theseptum pellucidum or the lateral wall of the

ventricle [19,58,71,97]. Whereas 75% of casesoccur in patients between the ages of 20 and40 years [15,28,98], CNC occurring in patients18 years of age or younger [3,5,14,19,32,36,42,

44,58,68,69,71,74,76,78,88,94] and in patients 50years of age or older [3,4,7,19,40,44,45,53,69,76,78,93,99] have been reported. The term

extraventricular neurocytoma is used to describehistologically similar tumors not found in in-traventricular locations [43]. Extraventricular

locations include the occipital lobe [8,22,54,72],parietal lobe [3,22,54], frontal lobe [8,22,24,58,96],temporal lobe [8,9,22,58], thalamus [8,72,100],

hypothalamus [8,22,69], cerebellum [7,17], pons[77], spinal cord [3,12,18,44,47,79,82,83], caudaequina [81], retina [52], and pelvis [101] as well asmature cystic teratoma of the ovary [102]. Cases

J. Lee was supported in part by a grant from the

Khatib Research Foundation as a Khatib Scholar

(20032004).

* Corresponding author.

E-mail address: [email protected]

(A.T. Parsa).


of neurocytoma have previously been reported asintraventricular oligodendroglioma, differentiatedcerebral neuroblastoma [34,95], primary cerebral

neuroblastoma [64,103,104], and intraventricularneuroblastoma [105].

A review of the available data from clinical

reports has provided some insight into commonsigns and symptoms associated with CNC. Theclinical presentation usually involves signs

and symptoms of increased intracranial pressure(ICP) of a few weeks to several months as aresult of noncommunicating hydrocephalus. Asshown in Table 1, reported symptoms include

headache, visual disturbance, motor disturbance,altered mental status, sensory disturbance, sei-zure, dizziness, and nausea or vomiting without

associated headache. Not all reported cases ofCNC include a detailed clinical history, however.Table 1 also reports signs elicited on physical

examination; however, this analysis is limited bythe lack of detailed information in some reports.Other authors have reported headache, nausea

and vomiting, and visual disturbance as the mostcommon symptoms, with papilledema present inmost patients [46,69]. Signs like ataxia [32,42,46,58,69,94], altered level of consciousness [16,

46,62,69,87,91], hemiparesis [16,32,40,46,63], andseizures [3,8,22,42,46,58,63,89] were less com-mon. Patients presenting with intraventricular

hemorrhage (IVH) [12,23,37,62,76,84,91] andsudden death [4] have also been reported. Inaddition, many cases have been discovered in-

cidentally in patients undergoing neuroimagingfor unrelated reasons [15,22,32,58,87,91]. Neuro-logic examination often yields no focal neurologicfindings other than those caused by increased

ICP [46].

hts reserved.

484 J. Lee et al /Neurosurg Clin N Am 14 (2003) 483508

Neuroimaging

CT/MRI/angiographyAppropriate diagnostic imaging studies for

patients with CNC may include CT, MRI, orangiography. Relative to the brain parenchyma,CNC appears as a well-circumscribed, isodense,

hyperdense, or mixed isodense/hyperdense masswith slight to moderate contrast enhancement onCT examination. Noncommunicating (obstruc-

tive) hydrocephalus is often present. Calcificationsand cyst-like areas may also be seen on CTimages. Compared with the surrounding whitematter, T1-weighted and proton-weighted images

typically appear isodense. T2-weighted imagesappear heterogeneous with areas of calcificationand cysts appearing hyperintense and the tumor

appearing isointense to hyperintense. Variableenhancement with gadolinium is common, reflect-ing the heterogeneous vascularity of CNC (Fig. 1).

For tumor localization and visualization of the

Table 1

Signs and symptoms

Signs and Symptoms No. %

HA 178/202 88.1

N/V without HA 2/202 1.0

Dizziness 4/202 2.0

Visual disturbance 51/202 25.2

Altered mental status 22/202 10.9

Seizure 9/202 4.5

Motor disturbance 40/202 19.8

Sensory disturbance 14/202 6.9

Papilledema 46/51 90.2

Abbreviations:HA, headache or reported as signs and

symptoms of increased intracranial pressure; N/V,

nausea and/or vomiting without headache. Dizziness

includes dizziness or vertigo; visual disturbance includes

blurring, diplopia, decreased acuity, intermittent vision

loss, photophobia, blindness, or abducens nerve palsy;

altered mental status includes altered level of conscious-

ness, loss of memory, apathy, disorientation to mild

dementia, loss of concentration, mood swings, syncope,

personality change, depression, psychosis, irritability, or

mental change; motor disturbance includes spasticity of

extremities, hemiparesis, hemiplegia, atrophy, clonus,

hypotonia, increased DTR, imbalance, left hemisyn-

drome, gait disturbance, ataxia, loss of balance, weak-

ness, pyramidal signs unsteady gait, or gait dysfunction;

sensory disturbance includes pain, paresthesia, left

hemisyndrome, hemihypaesthesia; seizure includes all

types of seizures and papilledema includes unilateral and

bilateral types. Data from references [18,10,11,14,15,18,

23,26,27,29,31,33,34,39,41,42,45,5658,6266,68,70,71,

7476,82,85,87,89,91,93,95,96,99,104,108,114, and 118].

attachment site, MRI is preferred [15,46,97,106].Angiography has also been used to assessvascularity; however, the results are nonspecific,

ranging from avascular [3,36,40,71] to highlyvascular [40,46,53,85]. Feeding arteries are re-ported to include anterior choroidal [1,15,40,53,71,76], posterior choroidal [1,31,40,53,57,87],

lenticulostriate [40,53], and branches of perical-losal arteries [40,53].

Magnetic resonance spectroscopy/single photonemission computed tomography/positronemission tomography

Magnetic resonance spectroscopy (MRS)[33,35,38,54,94], single photon emission comput-erized tomography (SPECT) [35], and positron

emission tomography (PET) [35,53,85] have allbeen used in preoperative evaluation. The experi-ence using these modalities is limited, however,and there is no consensus on the characteristics of

CNC. Many investigators suggest that prominentglycine and choline with low N-acetyl aspartate(NAA) peaks are characteristic markers of CNC

on MRS [33,38,94]. Others have found an in-creased choline peak and decreased NAA signalbut failed to consistently find the 3.55-ppm peak

characteristic of glycine [35,54]. SPECT analysisshows increased uptake of 201T on delayed images,indicating a high activity of sodium potassiumtriphosphate on cell membranes. PET analysis has

demonstrated decreased O2 extraction fraction,cerebral metabolic rate of O2, and cerebralmetabolic rate of glucose in CNC. Cerebral blood

flow and blood volume were increased in three offour cases, correlating with the angiographicfindings in these patients [53]. The authors suggest

that CNC metabolism is more oxidative than thatof other brain tumors and that a decreased rate ofglucose metabolism may predict a favorable

prognosis [52]. Further study is necessary todetermine whether these combined findings aretruly characteristic of CNC.

Pathologic examination

After reviewing such factors as presentingsymptoms, patient age, and location of the tumor,

a focused differential diagnosis for CNC includessubependymoma, astrocytoma, ependymoma, in-traventricular meningioma, intraventricular oligo-dendroglioma, and subependymal giant cell

astrocytoma [43,46,58,79,97]. Adding informationfrom imaging studies usually narrows the differ-ential diagnosis to intraventricular meningioma,

485J. Lee et al / Neurosurg Clin N Am 14 (2003) 483508

Fig. 1. (A) Axial CT image with contrast. (B) Contrast-enhanced coronal T1-weighted MRI. (C) Axial T1-weighted

MRI. (D) Sagittal T1-weighted MRI.

ependymoma, and CNC. A definitive diagnosisrequires tissue analysis with light microscopy,immunohistochemistry, and, in some cases, ultra-structural examination.

Light microscopyOn light microscopy, CNC appears similar to

oligodendroglioma (Fig. 2). Clinically, they can be

distinguished based on their location. CNCs aretypically intraventricular and centrally located,whereas oligodendrogliomas arise more peripher-

ally. Light microscopy shows a honeycomb archi-tecture with uniform small round cells with centralnuclei and clear cytoplasm dispersed within a

fibrillary stroma. The chromatin typically has asalt and pepper appearance. Microcalcificationsor microcysts may be present, and mitoses, endo-thelial proliferation, and necrosis are rare [15,

43,46,98,106]. The presence of neuroblastic ro-settes and nuclei with a ganglionic appearance issuggestive of neurocytoma; however, immunohis-

tochemistry or electron microscopy is required toconfirm the diagnosis [46].

Immunohistochemistry

The hallmark characteristic of CNC is positiv-ity for synaptophysin, a calcium-binding mem-brane protein of presynaptic vessels. In addition,CNC is usually positive for neuronal specific

enolase (NSE) and negative for glial fibrillaryacidic protein (GFAP) and neurofilament protein(NFP) (Fig. 3A) [15,43,46,51,98,106]. These char-

acteristics differentiate CNC from oligodendro-glioma and ependymoma.

Ultrastructural featuresUltrastructural examination is only required in

the diagnosis of CNC if synaptophysin is lackingor equivocal or if extraventricular neurocytoma issuspected. CNC shows neuronal differentiation,

microtubules, dense core and clear vesicles, andabortive or typical synapses (see Fig. 3B)[15,27,28,46,51,98,106]. Occasional mitochondria,

486 J. Lee et al /Neurosurg Clin N Am 14 (2003) 483508

Fig. 1 (continued )

moderate free ribosomes, and variable endoplasticreticulum may also be present [15,43,46,98].

Histogenesis/geneticsCNC is thought to be derived from bipotential

progenitor cells from the subependymal plate thatare capable of neuronal and glial differentiation[30,89,93,97,107]. Indeed, on cell culture, CNC

differentiates into neuronal and glial cells [30,89].In addition, CNC is capable of ependymal dif-ferentiation [75]. CNCs are genetically distinct

from oligodendrogliomas and neuroblastomas, asevidenced by a lack of association with specific 1pand 19q loss of heterozygosity and rarity of N-myc amplification [75].

Atypical neurocytomaAtypical neurocytomas are a rare variant of

CNC, with cellular pleomorphism, mitotic activ-

ity, necrosis, or vascular proliferation (Fig. 4)[3,19,54,62,71,73,77,92,108111]. Although mostCNCs appear as uniform small round cells on

light microscopy, atypical CNCs may show peri-

Fig. 2. Histological stain of a central neurocytoma

demonstrating sheets of monomorphic pale cells with

small delicate capillaries in the background. The tumor

cells have round uniform nuclei with fine chromatin and

inconspicuous nucleoli. There is no evidence of necrosis

or mitotic activity. (From Anderson RC, Elder JB,

Parsa AT, Issacson SR, Sisti MB. Radiosurgery for the

treatment of recurrent central neurocytomas. Neurosur-

gery 2001;48(6):12318; with permission.)

487J. Lee et al / Neurosurg Clin N Am 14 (2003) 483508

vascular pseudorosettes, neuropil islands, multi-nucleate cells, or ganglion cells. Mitotic activitycan be as high as 30 mitoses per high-power field,and evidence of necrosis may range from focal to

extensive. Although the correlation betweenhistologic atypia and proliferation potential inatypical CNC was poor [110], vascular prolifera-

tion showed a significant correlation with theMIB-1 labeling index (LI) (P = 0.0006) [77]. It isunclear how the histology of atypical CNC relates

to biologic behavior. Although elevated prolifer-

Fig. 3. (A) Central neurocytoma showing immunoreac-

tivity for synaptophysin (synaptophysin immunohisto-

chemistry, original magnification 20). (B) Electronmicroscopy of central neurocytoma. The tumor cells

have round nuclei and clear cytoplasm. The cytoplasm

contains microtubules, dense core vesicles, and synapses

(*) (original magnification 10,500). (From Hara M,Aoyagi M, Yamamoto M, Maehara T, Takada Y, Nojiri

T, et al. Rapid shrinkage of remnant central neuro-

cytoma after gamma knife radiosurgery: a case report.

J Neurooncol 2003;62(3):26973; with permission.)

ation potential correlates with poor outcome,histologic grade does not seem to have prognosticvalue [77,110].

Treatment

Overview

Treatment strategies for CNC are based onretrospective case series (Table 2), case reports, andanalysis of pooled data. There are no randomized

clinical trials and few prospective studies. In manyearlier reported cases, initial management mayhave been based on a diagnosis that was revised on

retrospective review [40,42,46,65,67,68,70,71,89].Most authors agree that, when possible, com-

plete tumor resection for symptomatic CNC is the

treatment of choice [19,40,44,46,60,70,96]. Theaddition of adjuvant radiation therapy (RT) inthe immediate postoperative period is controver-

sial. Some authors routinely use RT after subtotaltumor resection (STR) [1,5,19,39,46,59,89]. Al-though some have used RT after gross totalresection (GTR) as well [3,7,15,19,27,32,40,45,53,

58,70,96], several authors state that RT after GTRis not indicated [3,5,27,41,46,56,57,64,70,71,90].Given the potential for long-term radiation side

effects, some advocate for adjuvant RT only forrecurrent or progressive CNC [24,46,57], becausethe subependymal and subventricular zone is

sensitive to radiation. More recent reports ofstereotactic radiosurgery address the concerns of

Fig. 4. Central neurocytoma with atypical histologic

features, including necrosis, vascular endothelial pro-

liferation, and cellular pleomorphism (hematoxylin-

eosin, original magnification 10). (From MackenzieIR. Central neurocytoma: histologic atypia, prolifera-

tion potential, and clinical outcome. Cancer 1999;85(7):

160610; with permission.)

cal

ntrol

Survival

rate Outcome

Refer-

ence

p initial

rgery,

0% p GKS

100.0% Returned to work with

full fxn, 3/4 neurologically

nl, 1/4 on dilatin for

postoperative seizure

[2]

.30% 91.7% 7 KPS 100, 1 KPS 50, 1

died secondary to

hemorrhage

[3]

0% 100.0% 3 asymptomatic, 1

neurologically intact

[11]

.40% 83.0% 2 dead, tumor recurrence;

7 alive, no tumor

recurrence; 6 alive, tumor

recurrence; dead no

tumor recurrence; 1 dead

4 days after surgery; 1

dead, 2 years after surgery

[19]

0% 100.0% 10/10 no evidence or

recurrence

[20]

100.0% 2 asymptomatic, returned

to work; 1 mild residual

hemiplegia; 1 behavioral,

cognitive, and dexterity

problems; 1 short-term

memory deficit; 1 re-

covering well after surgery

[23]

488

J.Lee

etal/Neurosurg

Clin

NAm

14(2003)483508

Table 2

Larger case series

n

MIBY

labeling

index Location

Primary

treatment

Recur-

rence

Average

months to

recurrence

(range) Radiation

Timing

of RT

Average

FU in

months

(range)

Lo

co

4 Not

reported

3 LLV, 1

LLV/3rd

4 CTR 4/4 17.25

(925)

GKS, 1620

Gy to tumor margin

4 salvage 20.25 p

GKS, 54.5

p CTR

0%

su

10

9 3.6375 9 intraven-

tricular

6 ITR, 3

CTR, 3 RT

2/9 618 55 Gy (2 adjuvant, 1

salvage for

asymptomatic

progression)

2 adjuvant,

1 salvage

45.33

(689)

83

4 Not

reported

3 LLV,

1 BLV

1 CTR with

GKS for

recurrence, 3

ITR/GKS

1/4 53 GKS, 913 Gy (3

adjuvant, 1 salvage)

3 adjuvant,

1 salvage

44

(1299)

10

18 Not

reported

2 3rd, 3 3rd/

LV, 2 CC,

03.04 - intraventricular tumours

Documents

intraventricular neurocytomas

choroid plexus tumors

lateral ventricles

tumor type

pathologic literature

thetreatment of cnc

pathologic review

parsacentral neurocytomas