bone tissue. It was very adherent to the right internal carotidartery and optic nerve, therefore, a subtotal excision was per-formed. On pathological examination, the lesion was composedof interconnecting osteoid islands and highly vascularized stroma.The osteoid tissue was rimmed by osteoblasts. Scattered osteo-clasts and small bony trabeculae were seen. There was no evi-dence of malignancy. The lesion was diagnosed as a benignosteoblastoma (Fig. 3).Post-operative neurological examination revealed no additional

deficit and there was no neurological progression over 18 years,other than the right eye lateral gaze palsy present preoperatively.There was no evidence of enlargement at 18 years on cranial CTscan and MRI (Figs. 4 and 5).

DISCUSSION

Benign osteoblastoma is a rare bone tumor with an indolent bio-logical behavior and good clinical prognosis. It is very rare inthe cranium.5 The temporal bone is the most common site of oste-oblastoma in the neocranium.6 In the reported case, the tumororiginated from the petrous temporal bone and extended to theparasellar region and cavernous sinus.It is important to differentiate osteoblastoma from osteoid oste-

oma. The pain from osteoid osteoma generally increases noctur-nally and responds to salicylates, while the pain fromosteoblastoma is continuous and does not respond to salicylates.Osteoblastoma has a greater potential to enlarge than osteoid oste-oma and presents by mass effect rather than pain.5,6 In this case,paralysis of the sixth cranial nerve aided localization of the pathol-ogy. Although imaging is insufficient to determine the nature ofthe lesion, it may be highly suggestive.2 However, low-gradechordoma, chondrosarcoma and giant cell tumors also have simi-lar radiological features.2 They present as cortical or medullary,diaphyseal or metaphyseal, prominently radiolucent lesions withvariable calcification and clearly defined, non-lobular margins.2

Histological diagnosis is important. On microscopic examination,early lesions contain actively propagated connective tissue isletsand older lesions frequently show large foci of ossification. Inall lesions, irregular fibrous stroma and osteoid deposits are seen.Stroma tends to be well vascularized, particularly in early tumorsand rarely exhibits extravasation. Rarely minor atypical changesmay be observed in osteoblast mitochondria.3

Optimal primary treatment of benign osteoblastoma is totalsurgical excision.7 If, as reported here, the tumor is not suitablefor total surgical excision due to its location, subtotal excision isadequate as the probability of recurrence is low. No progressionhas been observed during 18 years post-operative follow up inthis patient. Therefore, radical surgical excision, which may havesignificant post-operative morbidity and mortality, should beavoided and safer palliative excisions considered. If there isany evidence of malignancy on pathological examination, radio-therapy and chemotherapy may be administered. However, it hasbeen reported that radiotherapy may induce or promote malig-nant degeneration.1,2,3 Hence the role of the radiotherapy iscontroversial.

REFERENCES

1. Batay F, Savas A, Ugur HC, Kanpolat Y. Benign osteoblastoma of the orbitalpart of the frontal bone: case report. Acta Neurochir 1998; 140: 729–730.

2. Clutter DJ, Leopold LV. Benign osteoblastoma: report of a case and review ofthe literature. Arch Otolaryngol 1984; 110: 334–336.

3. Lucas DR, Unni KK, McLeod RA, O’Connor MI, Sim FH. Osteoblastoma:clinicopathologic study of 306 cases. Hum Pathol 1994; 25: 117–134.

4. Berciano J, Perez JL, Fernandez F. Voluminous benign osteoblastoma of theskull. Surg Neurol 1983; 20: 383–386.

5. Chen K, Weinberg R, Simpson P, Tschang T. Osteoblastoma of the nasal cavity.Laryngol Otol 1993; 107: 737–739.

6. Khasaba A, Donata GD, Vassalo G. Benign osteoblastoma of the mastoid part ofthe temporal bone: case report. J Laryngol Otol 1995; 109: 565–568.

7. Ciappetta P, Salvati M, Raco A, Artico M. Benign osteoblastoma of the sphenoidbone. Neurochirurgica 1991; 34: 97–100.

Primitive supratentorialneuroectodermal tumor in anadult

AT Kouyialis1 MDMD, EI Boviatsis1 MDMD, IK Karampelas1 MDMD,

S Korfias1 MDMD, P Korkolopoulou2MDMD, DE Sakas1 MDMD

1Department of Neurosurgery, University of Athens Medical School,

‘‘Evangelismos’’ Hospital; 2Department of Pathology, University of Athens

Medical School; Athens, Greece

Summary We report the case of a 32-year-old female with a diag-

nosis of supratentorial tumour. Total removal of the tumour was

achieved in a two-stage procedure. Histopathology revealed a primi-

tive neuroectodermal tumour (PNET), an unusual and highly malig-

nant, mainly infratentorial tumour of childhood that is uncommonly

described in the supratentorial compartment of adults. We review the

literature and describe the existing knowledge of these tumours.

ª 2005 Elsevier Ltd. All rights reserved.

Journal of Clinical Neuroscience (2005) 12(4), 492–495

0967-5868/$ - see front matter ª 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jocn.2004.07.014

Keywords: primitive neuroectodermal tumour, supratentorial,

medulloblastoma, adulthood

Received 11 February 2004

Accepted 13 July 2004

Correspondence to: Efstathios J Boviatsis MD, Department of Neurosurgery,

Evangelismos General Hospital, 61 ipsilantou Street, Athens 115 21, Greece.

Tel.: +30 210 7201654; Fax: +30 210 7215281;

E-mail: eboviats@med.uoa.gr

INTRODUCTION

The term primitive neuroectodermal tumour (PNET) refers to agroup of neoplasms consisting of highly undifferentiated malignantcells that present mainly in children and histologically resemblecerebellar medulloblastoma. They are chiefly characterised bysmall, round, undifferentiated cells with scanty cytoplasm. Oneof the outstanding features of the malignant character of PNET istheir ability to disseminate within the central nervous system(CNS) via the cerebrospinal fluid (CSF) pathways. Thus, at the timeof diagnosis, multiple foci of disease are frequently discovered.Although this tumour is relatively uncommon, it is well

recognised, particularly in children, yet there is still ongoingdebate and considerable controversy concerning the histogenesisand classification of PNET. An extensive literature review hasrevealed few reports of PNET in adults.1 In this report, an adultfemale with a supratentorial PNET is described and the existingliterature regarding the unusual features of these tumoursreviewed.

Journal of Clinical Neuroscience (2005) 12(4) ª 2005 Elsevier Ltd. All rights reserved.

492 Kouyialis et al.

CASE REPORT

A 32-year-old woman was admitted complaining of severe morn-ing headaches, recent gait disturbance and fatigue with walkingdue to progressive decline in function of both lower limbs. Neuro-logical examination revealed increased tendon reflexes in theupper and lower limbs with normal muscle power. She had bilat-eral papilledema but no other neurological abnormality. CT andMRI of the brain revealed a large space-occupying lesion in themid- and parasagittal convexity of both hemispheres, just beneaththe calvarium (Fig. 1). The mass consists of three components; thelargest, measuring 5 · 7 · 7 cm, was directly under the bone;the other two sections, measuring 4 · 3 · 3 cm each, lay withinthe parenchyma, yet had clearly defined borders. Digital substrac-tion angiography (DSA) showed blood supply deriving frombranches of the left external carotid artery and a considerabledownward displacement of the superior sagittal sinus (Fig. 2).Due to the large size of the tumour, a two-stage procedure was

advised. The goal of the first procedure was to remove the centraltumour segment. This would allow the two smaller components toascend closer to the surface and enable better visualisation at thesecond stage, thus decreasing the risk of damage to eloquent areas.The first operation revealed that the tumour had extensively in-vaded and destroyed the overlying bone. This was removed untila healthy margin was reached. The central part of the tumour wasa whitish-grey colour, was highly vascular and friable and wastotally removed.Postoperatively, the patient had motor weakness on her left side

(2/5) that was treated conservatively with corticosteroids andphysical therapy and progressively recovered. Histopathologicalexamination showed neoplastic tissue consisting of a homogenouspopulation of small cells with deeply stained oval nuclei, scantycytoplasm and strongly positive for synaptophysin CD99 (Fig.3). The cells were growing within a fibrillary network and amongthem some larger cells with neuronal morphology were also iden-tified. The gross and microscopic morphology of the tissue and theresults of the immunocytochemical analysis were consistent withthe diagnosis of PNET with glial and cartilaginous differentiation.Due to propensity of PNET to metastasise via the CSF path-

ways, an MRI of the whole neuraxis with contrast enhancementwas obtained but no other lesion was demonstrated. The secondoperation was performed 4 months later in another hospital andcomplete tumour removal was achieved with no neurologic defi-cits. The patient refused irradiation or chemotherapy. She is cur-

rently under clinical and radiological follow up and 24 monthsafter the first operation there are no signs of recurrence.

DISCUSSION

Bailey and Cushing first reported PNET in 1924. They described aseries of 25 patients with tumours that seemed to arise from the4th ventricle and protrude into the cerebellum. Initially, theynamed this tumour cerebellar spongioblastoma. However, in1925, they extended their series to a total of 29 patients and re-named the tumour cerebellar medulloblastoma. Under this head-ing, low-grade differentiated tumours of the cerebellum werealso included, originating around the 4th ventricle and occurringmainly in children. The term cerebellar medulloblastoma was se-lected due to the assumption that the tumour originated from theprimitive myeloblast, a neuronal stem cell able, in theory, to dif-ferentiate to all neuronal cell types.2

In 1973, Hart and Earl3 introduced the term primitive neuroec-todermal tumour/PNET to describe embryonal neoplasms arisingoutside the cerebellum but morphologically similar to medullo-blastoma. In 1983, Rorke suggested the inclusion of the term‘‘medulloblastoma’’ within the broader category of PNET4 onthe basis of their histological similarity to other CNS tumoursoriginating from the malignant transformation of primitive neuro-epithelial cells. This classification was accepted by the World

Fig. 1 T1-weighted coronal MRI showing a large supratentorial lesion

located on the mid- and parasagittal convexity of both hemispheres.

Fig. 2 Lateral digital subtraction angiogram showing considerable

downward displacement of the superior sagittal sinus by tumour (arrows).

Fig. 3 Microscopy of the tumour, showing a PNET with glial and

cartilaginous differentiation, consisting of small cells with oval nuclei and

scanty cytoplasm. The tumour stained strongly positive for CD99.

ª 2005 Elsevier Ltd. All rights reserved. Journal of Clinical Neuroscience (2005) 12(4)

493Adult supratentorial primitive neuroectodermal tumor

Health Organization (WHO) in 19855 and in 1993, in the WHOclassification of brain tumours, the term PNET was assignedto both medulloblastoma and other histologically similar CNStumours that did not necessarily originate from the cerebellum(PNET/medulloblastoma if within the cerebellum, supratentorialPNET if outside the cerebellum).6

The precise relationship of PNET to cerebellar medulloblas-toma, however, remains speculative largely due to the fact thattheir true cellular origin and histogenesis are not yet clarified. Re-cent studies by Gyure et al.7 and Russo et al.8 (1999), revealedconsiderable differences between medulloblastoma and otherPNET, both at the genetic and immunohistochemical level andtheir studies suggested these tumours to be distinct biological enti-ties. The final classification of PNETs thus remains ill defined.With their histogenesis incompletely understood, different groupsof investigators have offered their own interpretations and haveused the term in different ways, seldom defining their pathologiccriteria in a clear and specific manner.The underlying uncertainty over pathologic definitions under-

mines the validity of the clinical characteristics and comparisonof cases reported as PNET. The classification evolved from theoriginal work of Bailey and Cushing, was elaborated upon byRubinstein9 in 1972 and is based on the hypothesis that the pri-mordial neural tube cells constitute the cellular series of originfor medulloblastoma. Given the variety of the histopathologiccharacteristics of PNET, inherent in this tenet is also the presump-tion that neural tube embryonic cells are capable of differentiatingto all cell types in the nervous system, including neuroglia, neu-rons and ependymal cells. This hypothesis, though popular, re-mains unproven. According to the scheme of Rubinstein9

(Table 1), six different types of PNETs are recognised. Addition-ally, the retinoblastoma is sometimes incorporated in this schemedue to the primitive morphology of its constituting cells and theirpotential ability to differentiate towards either neuroblasts or pho-toreceptor cells.10

The uncertainty that prevails regarding the histogenesis andclassification of PNETs has led to equal uncertainty in definingand comparing the various clinical characteristics of these tu-mours within reported series.10 Furthermore, while these tumoursare uncommon even in children, in whom they constitute about 3–7% of primary brain tumours;11 reports in the literature of suchneoplasms occurring in the supratentorial compartment of adultsare rare.The clinical presentation is generally non-specific and no char-

acteristic symptom or sign has been described.1 It includes mostlysymptoms and signs of increased intracranial pressure of recentonset as a consequence of rapid tumour growth. Patients usuallycomplain of morning headaches frequently accompanied by vom-iting; the headache is sometimes localised in the occipital area, asa result of a postulated small degree of cerebellar tonsillar herni-ation. Papilledema is frequent. Focal neurological signs appear asa result of local pressure on the brain parenchyma.Radiological characteristics vary considerably. Commonly,

these are sizeable heterogeneous lesions on CT and MRI, with cys-tic and necrotic regions (the latter implying rapid growth). Focal

intratumoural calcification is not uncommon, thus CT is importantbecause calcification is helpful in the differential diagnosis frommetastatic brain tumour and high-grade astrocytoma.1 Haemor-rhage into the tumour bed is not infrequent, while in some casesa large degree of peritumoural oedema is also present. DSA usuallyreveals a rich vascular supply and is indicated in all cases of sus-pected PNETs since pre-operative embolisation of feeding vessels,if feasible, may assist operative resection. In our reported case, thepattern of the tumoural vascular supply at the time of DSA provednot to be amenable to embolisation. Finally, preoperative radiolog-ical investigation of such patients should always include imagingof the spine to detect dissemination along the CSF pathways.Histologically, the tumour cell population consists of undiffer-

entiated round or oval cells with hyperchromatic nuclei andscanty, ill-defined cytoplasm. These cells exhibit substantial sim-ilarity to the neuronal blast cells of the embryonic neural tube. Themalignant cells commonly show a great number of mitoses butmitotic activity is variable. Microscopic calcifications, necrosesand Homer–Wright rosettes are also observed in a number ofcases. Thin-walled blood vessels are found within the tumour tovarying degrees, while endothelial cell proliferation within thevessel wall is also regularly observed.Therapy of PNETs is surgical removal of as much tumour as

possible without damaging the underlying parenchyma. Extensivesurgical resection has been associated with long-term survival inseveral series although this association has not reached statisticalsignificance and the series were small.12,13 Kim et al.1 on the otherhand have suggested that the degree of resection may be not a ma-jor factor in the prognosis. The large size, poor demarcation fromsurrounding parenchyma, proximity to eloquent areas and highvascularity of these tumours often dictates a two-stage procedure,as in our case. It is common practice to proceed with the resectionof PNET from the periphery inward, that is, contrary to internaldecompression used for other intrinsic neoplasms, so as to avoidsubstantial hemorrhage from tumour vessels.14

Radiotherapy is an important element of the management ofthese tumours. Sensitivity to radiation was established as earlyas 1919 by Bailey and Cushing.2 Due to the ability of PNET todisseminate via the CSF pathways, radiotherapy should includethe entire neuraxis even if the initial investigation for metastaticdisease has proved negative. This may lead to a higher incidenceof cognitive and neurological deficits and seizures due to concen-trated doses of radiation to specific segments of the CNS over ashort time period.2 The results of chemotherapy in the manage-ment of PNET are controversial and in some cases disappointing.Agents such as cisplatinum, cyclophosphamide and vincristinehave been reported to show measurable, although transient, ther-apeutic effects and extend survival.15

The prognosis for PNET is generally poor. Even after macro-scopic total removal at operation, local recurrences are common.Additionally, leptomeningeal metastases are common and systemicdissemination of disease, most commonly to the lungs and lymphnodes is also encountered.16 Overall survival rarely extends beyond24 months with very few patients surviving more than 5 years.

CONCLUSION

PNETs constitute an ill-defined pathological entity with many oftheir biological and clinical parameters still incompletely under-stood. Although occurring rarely in the supratentorial compart-ment of adults, they should be included in the differentialdiagnosis of supratentorial masses The absence of a clear defini-tion of the disease and the small number of reported cases of adultPNET in the literature makes difficult a comprehensive compari-son of true clinical data on these tumours. Carefully documented

Table 1 Classification of PNET, Rubinstein, 19729

1. Medulloblastoma

2. Myeloepithelioma

3. Cerebral neuroblastoma

4. Polar Spongioblastoma

5. Ependymoblastoma

6. Epiphysioblastoma

Journal of Clinical Neuroscience (2005) 12(4) ª 2005 Elsevier Ltd. All rights reserved.

494 Kouyialis et al.

histological assessment from multiple centres would assist in thebetter understanding of the histogenesis, classification, and treat-ment of these neoplasms.

REFERENCES

1. Kim DG, Lee DY, Paek SH, Chi JG, Choe G, Jung HW.Supratentorial primitive neuroectodermal tumors in adults. J Neurooncol2002; 60: 43–52.

2. Tomita T. Medulloblastomas. In: Youmans JR, editor. Youmans NeurologicalSurgery. fourth edn.. W.B. Saunders Company, 1996.

3. Hart MN, Earle KM. Primitive neuroectodermal tumors of the brain in children.Cancer 1973; 32: 890–897.

4. Rorke LB. The cerebellar medulloblastoma and its relationship to primitiveneuroectodermal tumors. J Neuropathol Exp Neurol 1983; 42: 1–15.

5. Rorke LB, Gilles FH, Davis RL, Becker LE. Revision of World HealthOrganization classification of brain tumors for childhood brain tumors. Cancer1985; 56: 1869–1886.

6. Kleihues P, Burger P, Scheithauer BW. The new WHO classification of braintumors. Brain Pathol 1993; 3: 255–268.

7. Gyure KA, Prayson RA, Estes ML. Extracerebellar primitive neuroectodermaltumors: a clinicopathologic study with bcl-2 and CD99 immunohistochemistry.Ann Diagn Pathol 1999; 3: 276–280.

8. Russo C, Pellarin M, Tingby O, et al. Comparative genomic hybridization inpatients with supratentorial and infratentorial primitive neuroectodermaltumors. Cancer 1999; 86: 331–339.

9. Rubinstein LJ. Cytogenesis and differentiation of primitive centralneuroepithelial tumors. J Neuro Exp Neurol 1972; 31: 7–26.

10. Crain BJ. Primitive neuroectodermal tumors. In: Wilkins R,Rengachary S, editors. Neurosurgery. 2nd ed. USA: McGraw-HillCompanies; 1996. p. 1707–1712.

11. Becker LE, Hinton D. Primitive neuroectodermal tumors of the central nervoussystem. Hum Pathol 1983; 14: 538–550.

12. Dirks PB, Harris L, Hoffman HJ, Humphreys RP, Drake JM, Rutka JT.Supratentorial primitive neuroectodermal tumors in children. J Neurooncol1996; 29: 75–84.

13. Tomita T, McLone D, Yasue M. Cerebral primitive neuroectodermal tumors inchildhood. J Neurooncol 1988; 6: 233–243.

14. Goumnerova L. Primary cerebral hemisphere tumors in children. In: Kaye AH,Black PM, editors. Operative Neurosurgery, Vol. 1. London: HarcourtPublisher Limited; 2000. p. 371–379.

15. Bertolone SJ, Baum ES, Krivit W, Hammond GD. A phase II study of cisplatintherapy in recurrent childhood brain tumors: a report from the Children’sCancer Study Group. J Neurooncol 1989; 7: 5–11.

16. Pickuth D, Leutloff U. Computed tomography and magnetic resonance imagingfindings in primitive neuroectodermal tumors in adults. Br J Radiol 1996; 69:1–5.

Recurrent tumefactivedemyelination in a child

Vinod Puri1 MD DMMD DM, Neera Chaudhry1 MDMD,

Parveen Gulati2 MDMD, Medha Tatke3 MDMD,

Daljit Singh4MS MCHMS MCH

1Department of Neurology, G.B.Pant Hospital, 2Organ Imaging Research

Centre, 3Department of Pathology 4Department of Neurosurgery, G.B.Pant

Hospital; New Delhi, India

Summary A 13-year-old female presented with two episodes of

hemiplegia and hemianopia involving opposite sides, each time

associated with seizures. On both occasions, the magnetic reso-

nance (MR) scan showed a giant demyelinating, peripherally

enhancing lesion with mass effect. MR spectroscopy (MRS) was

indistinguishable from a tumor. At the first episode, she had

undergone tumor decompression but the histopathology revealed

an acute demyelinating lesion with no evidence of tumor. Each

time there was partial clinical recovery and resolution of the

radiological lesion, the patient having received corticosteroids

during both of the episodes. She also developed hemiballismus

postoperatively which resolved on withdrawing phenytoin. It is

suggested that a trial of corticosteroids be given in suspected

tumefactive demyelinating lesions, although they may be indistin-

guishable from a tumor.

ª 2004 Elsevier Ltd. All rights reserved.

Journal of Clinical Neuroscience (2005) 12(4), 495–500

0967-5868/$ - see front matter ª 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jocn.2004.07.001

Keywords: tumefactive demyelination, multiple sclerosis variants,

MRS, histopathology, hemiballismus

Received 28 December 2003

Accepted 2 July 2004

Correspondence to: Dr. Vinod Puri MD DM, 16, Type V, MAMC Campus, New

Delhi 110 002, India. Tel.: +91-11-2323-5237; Fax: +91-11-23238695;

E-mail: vpuri@bol.net.in

INTRODUCTION

Multiple sclerosis is a common demyelinating disease presentingwith a relapsing and remitting course in adults. The usual age ofonset is between 20 and 50 years, though onset in childhood hasalso been reported. About 4.4% of patients with multiple sclerosispresent with their first episode before their sixteenth birthday and0.3% present before 10 years of age.1,2

Several reports in the literature describe the infrequent presen-tation of primary demyelinating disorders with large lesionsmimicking intracranial neoplasms or abscesses, which have beentermed tumefactive demyelination.3–21 There are reports oftumefactive demyelinating lesions in the pediatric agegroup.5,12–21

We present a child with two episodes of demyelinating diseaseeach time with a tumefactive lesion.

CASE

A previously well, 13-year-old female presented with left-sidedsimple partial motor seizures with secondary generalization, with-out Jacksonian march. The seizures had been preceded by head-ache of 2 weeks duration. The headache was dull, holocranial,without diurnal or postural variation and with no associated vom-iting or visual complaints. The seizures were followed by leftsided weakness and left sided hemianopia. There were no com-plaints suggesting any other cranial nerve or bladder or bowelinvolvement. There was no history of any neurological problemin the family. Her general physical examination as well as theother systemic examination was non-contributory. Neurologicalexamination revealed a left hemianopia, normal fundi and grade2/5 left hemiparesis with hyperreflexia and an extensor plantar re-sponse. There was no sensory deficit. Her hematological and bio-chemical profile, including blood glucose, urea, creatinine, serumelectrolytes and liver function tests, were normal. ELISA for HIVand hepatitis B surface antigen was non-reactive. Her chest X-ray,electrocardiogram and echocardiogram were normal. CT headscan showed a non-enhancing area of low attenuation in the rightposterior parietal and parieto-occipital region, primarily sub-corti-cal, involving the white matter and extending to the splenium ofthe corpus callosum, with no midline shift. MR brain imaging re-vealed a large area of altered signal intensity involving the sameregion. The lesion was hypo-intense on T1-weighted and hyperin-tense on T2-weighted images (Fig. 1). There was perifocal edema

495Recurrent tumefactive demyelination in a child

ª 2004 Elsevier Ltd. All rights reserved. Journal of Clinical Neuroscience (2005) 12(4)