J Clin Pathol 1983;36:68-73

Neuroblastoma arising in a mature cystic teratoma oftheovary

HAS REID, JD VAN DER WALT, H FOX*

From the Departments ofHistopathology, Enfield Distric t Hospital, Enflield, and *Reproductive Pathology, UniversityofManchester, Manchester

SUMMARY A case is described of a neuroblastoma which appeared to arise in a mature cystic teratoma ofthe ovary. The literature concerning neural tumours of the ovary is briefly reviewed and the grounds forbelieving that the development ot such neoplasms is an indication of the presence of immatureneuroepithelial components in a teratoma, rather than a result of malignant change in a fully matureteratoma, are discussed.

We describe here the clinical and pathological features of aneuroblastoma which arose in a mature cystic teratoma ofthe ovary, this being, we believe, the first authenticexample of malignant change of this type occurring in anotherwise fully benign ovarian teratoma.

Case report

An 18-year-old girl presented with complaints of cramp-like lower abdominal pain for eight months, urinaryfrequency for six months and bleeding per vaginam for 10days. Eighteen months previously she had born afull-term normal infant and since then had been on oralcontraceptives. Menstruation had been previously regularand her only previous illness was acute appendicitis as achild. On examination, a pelvic mass the size of an18-week gravid uterus was palpable. The uterus itself wasof normal size and retroverted. There were no otherabnormal physical findings and sonography showed alarge left sided multilocular mass. At laparotomy anapparently cystic mass measuring approximately 10 cm indiameter was found replacing the left ovary and a leftsalpingo-oophorectomy was performed. A "dermoidcyst," measuring 5 cm in diameter, was found arising inthe right ovary and this was enucleated with conservationof the right ovary and tube. The patient made a goodrecovery from surgery and was well on discharge 12 dayslater.

PATHOLOGICAL FINDINGS

Macroscopic appearancesThe left ovarian tumour was a grey-brown, fluctuant massmeasuring 10 x 10 x 8 cm. The outer surface was smooth

Accepted for publication 2 August 1982

and free of adhesions and the capsule appeared intact. Oncross-sectioning, it was apparent that the tumour waspartially formed of a typical mature cystic teratoma,measuring 5 cm in diameter, which contained hair andsebaceous material. The remainder of the tumourconsisted of a multilocular, partially cystic structurewhich enveloped the cystic teratoma.The central part of this was largely necrotic but the

peripheral portion consisted of locules, mostly measuring1 to 2 cm in diameter, which were- separated and lined byfriable pale brown tissue. The lesion from the right ovarywas a macroscopically typical small mature cysticteratoma.

Light microscopyHistological sections from the left ovarian mass showedtwo distinct elements. The mature cystic teratomaidentified macroscopically was lined by keratinisingstratified squamous epithelium with associatedpilosebaceous follicles and occasional apocrine glands.Foci of mucus-secreting columnar epithelium, bone and atooth were also present. The second element, whichclosely invested the cystic teratoma, was a densely cellularneoplasm with central areas of cystic degeneration (Fig.1). The tumour was surrounded by a thick fibrous capsulefrom which a few fibrous trabeculae entered theneoplastic mass, but generally the large masses of cellswere pooly supported by stroma. The tumour wasmarkedly vascular and the thin-walled vessels weresupported by delicate fibrous septa which tended toseparate the tumour cells into alveolar masses. Areas ofhaemorrhage were numerous, particularly towards thecentre of the tumour. The predominant tumour cell wassmall with a round or polygonal, darkly staining nucleuswith closely clumped chromatin and barely discernibleeosinophilic cytoplasm. These cells were interpreted as

68

copyright. on F

ebruary 3, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.1.68 on 1 January 1983. D

ownloaded from

Fig. I Low-powerphotomicrographshowing at the top ofthe picture thetypical squamous epithelium withattached pilosebaceousfollicles ofthedermoid cyst and at the bottom ofthepicture the densely cellularneuroblastoma which completelyinvested the dermoid cyst.Haematoxvlin and eosin x 23

Fig. 2 High power detail ofrosetteshowing sympathogonia. Glycolmethacrylate, haematoxylin andeosin x 576

copyright. on F

ebruary 3, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.1.68 on 1 January 1983. D

ownloaded from

Reid, van der Walt, Fox

Fig. 3 Neuroblastoma showingvarying size ofcells. A clump ofsympathoblasts is present in themiddle ofthefield. Glycolmethacrylate, haematoxvlin anodeosin x 576

Fig. 4 One ofthe very fewt areas inwhich ganglionic differentiation ispresent. Glycol methacrylate,haematoxYlin and eosin x 225

70

copyright. on F

ebruary 3, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.1.68 on 1 January 1983. D

ownloaded from

Neuroblastoma arising in a mature cystic teratoma ofthe ovarvRlL. RlY, _ s , ,~~~~~I I=Kwd:.t Ew.,.

Fig.5 General lowpower view oftumour showing rosettes and

m- neurofibrillarv substance.Haematoxvlin and eosin X 225

neoplastic homologues of sympathogonia (Fig. 2). Insome parts of the tumour clumps of cells with larger.lighter staining, vesicular nuclei and more substantialeosinophilic cytoplasm were seen. The cell borders wereindistinct, producing a pseudosyncytial appearance.These cells were interpreted as being homologous withsympathoblasts (Fig. 3). In a very few areas, large cellswith voluminous eosinophilic cytoplasm and a single,round, pale staining, vesicular nucleus containing aprominent nucleolus were present and these wereinterpreted as evidence of minimal ganglionic differentia-tion (Fig. 4).The arrangement of tumour cells was generally

haphazard but rosette formation was obvious in mostparts of the neoplasm. The rosettes consisted of a ring ofround or oval cells with tail-like cytoplasmic processesprojecting inwards. In a few areas neurofibrillary materialwas abundant and formed small eosinophilic strands andislets (Fig. 5). The tumour was confined within thecapsule with no evidence of capsular invasion. Tumouremboli were, however, seen in hilar lymphatics.The right ovarian tumour was a mature cystic teratoma

which did not contain any neural elements.

Fig. 6 Electron micrograph oftumour cells showvingneurosecretorx granules. x 153 000

Electron microscopyElectron microscopic examination was made on formalin-

A.6 a. '

71

copyright. on F

ebruary 3, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.1.68 on 1 January 1983. D

ownloaded from

Reid, van der Walt, Fox

Postoperative urinary biochemical markers

VMA HVA NMA MA(pmol/24h)(p mol/24h) (p mol/24h) (p moll24h)

Normal range 9-36 10-40 0 55-273 0 25-1*787 weeks post-op 25 21 3-1 0 712weekspost-op 48 26 5 4 1-07 months post-op - - 7-1 1.1II months post-op (245 105 - -

(3 consecutive days) . 310 1301 188 88 8.7 0.8

fixed tissue. There was considerable loss of cytologicaland organelle detail but neurosecretory granules were

seen in a small proportion of the tumour cells (Fig. 6).

CLINICAL FOLLOW-UP

The patient has so far been followed up for some 12months as an outpatient. Twenty-four hour urine sampleswere collected postoperatively and the urinary excretionconcentrations of vanillyl methoxymandelic acid (VMA),homovanillic acid (HVA), free and conjugated normetad-renaline (NMA) and free and conjugated metadrenaline(MA) were measured. The results are shown in the Table.The rising concentrations of VMA, HVA and NMA are

thought to indicate the presence of metastatic tumour, andat time of writing further diagnostic procedures are

planned. Indefinite follow-up will be pursued.

Discussion

Malignant change occurs in between I and 2% of maturecystic teratomata of the ovary' and whilst in theory any

type of malignant neoplasm can arise in a previouslybenign teratoma, the vast majority which do occur are inpractice squamous cell carcinomata, most of the remain-der being adenocarcinomata. Neural tissue occurs withconsiderable frequency in mature cystic teratomata butthe development of a malignant neural neoplasm in an

ovarian tumour of this type is an event of exceptionalrarity. There has, however, been one indisputable exampledescribed of a grade 4 astrocytoma arising in a maturecystic teratoma of the ovary2 whilst there have been fourprevious reports of a neuroblastoma developing in an

ovarian mature cystic teratoma3-6. The validity of theselatter cases does not, however, withstand critical scrutinyfor three of the reported tumours3-5 would more correctlybe interpreted as teratomata containing immatureneuroepithelial elements, some of which had a neuroblas-tomatous appearance whilst the fourth5 was an immature(malignant) teratoma which contained, inter alia, imma-ture neuroepithelial tissue and gave rise to metastasescontaining multiple tissue elements, some of which were

neural with a neuroblastomatous component. In none ofthese cases was there any real evidence of malignantchange having occurred in a previously fully benignteratoma and in none was a definite tumour mass,

showing the histological features of a neuroblastoma,present.

Recently, however, Aguirre and Scully7 have reportedfive malignant neural tumours of the ovary, four of whichappeared to be primary ovarian neoplasms and one whichwas possibly a metastasis from an adrenal neuroblastoma.Of the four definitely primary ovarian tumours, one wasclassified as a primitive neuroectodermal neoplasm withependymal, neuroblastic and glial differentiation and asecond as a primitive neuroectodermal tumour withependymal and glial differentiation and medulloepi-thelioma. The remaining two tumours resembled aglioblastoma multiforme. In one of these cases there wasno evidence of any other teratomatous elements whilst inanother the ovarian tumour was devoid of non-neuraltissue components but gave rise to metastases containingcartilage. The remaining two cases, both classed asglioblastoma, were associated with mature, but relativelyinconspicuous, non-neural tissue elements and appearedto have developed in mature cystic teratomata. It is ofinterest that these latter two cases had a macroscopicappearance similar to that of the tumour reported here inso far as the malignant neural tumour appeared to envelopthe cystic teratoma rather than being confined within it.Aguirre and Scully7 comment on the diagnostic problemposed by a pure, or largely pure, malignant neuralneoplasm arising in the ovary and upon the possibleconfusion with a granulosa cell tumour.

It would appear therefore that the tumour describedhere is the first definite example of a neuroblastomawhich has arisen in a mature cystic teratoma of the ovary.The purist, however, would maintain, with considerablevalidity, that a neuroblastoma cannot arise in a fullymature teratoma and the development of a malignantneural tumour indicates that immature neuroepithelialelements must have been present in the otherwise matureteratoma. This view would accord with the age distribu-tion of malignant neural neoplasms of the ovary. Ourpatient was, as were all those in other acceptable cases2 7less than 20 years of age, this being the typical age groupin which immature, and hence malignant, ovarianteratomata occur. By contrast, malignant change in a fullymature ovarian teratoma occurs most commonly duringthe fifth and sixth decades, ' this age-related patternlending support to the view that a malignant neuralneoplasm can only arise from immature neuroepithelialtissue in an ovarian teratoma.The prognosis for patients with a malignant neural

tumour of the ovary is gloomy, three of the four patientsreported by Aguirre and Scully7 died, all within four yearsof surgical removal of the primary neoplasm, whilst thepatient described by Shirley, Piro and Crocker2 died 11months after oophorectomy. Ganglionic differentiation inneuroblastomas elsewhere in the body is associated with arelatively better prognosis but the very minimal degree ofganglionic differentiation present in this case is not

72

copyright. on F

ebruary 3, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.1.68 on 1 January 1983. D

ownloaded from

Neuroblastoma arising in a mature cyvstic teratoma ofthe ovary

thought to be significant. The patient described in thisreport is currently well but biochemical tests indicate thatmetastatic tumour is probably present.

The authors wish to thank Mr AJ Minchin for permissionto publish this case, Mr M Blaxland for technicalassistance and Mrs EA Vasey for preparing the manu-

script.

References

Fox H. Langley FA. Tunmours (of 1he OvorY: London: HeinemannMedical. 1981. ,

Shirley RL. Piro AJ. Crocker DW. Malignant neural elements in a

benign cystic teratoma: a case report. Obstet Gvnecol 1971:37:402-7.

McCullough CD, Hardart F. Neuroblastomatous transformation in a

benign cystic teratoma. Obstet Gynecol 1963;21:259-61.4Mayberger HW, Carlson AS, Lim S. Immature neural elements

(neuroblastomatous change) in benign cystic teratoma: malignant or

not? Obstet Gynecol 1969;33:114-7.Theppisai R, Shuangshoti S, Amatyakul A. Neuroblastoma arising in

metastasizing ovarian teratoma. J MedAssoc Thai 1977;60:396-404.6 Shuangshoti S, Tharavej A. Neoplasms of neuroepithelial origin in a

cystic teratoma of ovary. J Med Assoc Thai 1980;63:284-9.7Aguirre F, Scully RE. Malignant neuroectodermal tumor of the ovary. A

distinctive form of monodermal teratoma: report of five cases. Am JSurg Pathol 1982;6:283-92.

Requests for reprints to: Dr HAS Reid, Department ofHistopathology, Chase Farm Hospital, The Ridgeway, Enfield,EN2 8JL, England.

73

copyright. on F

ebruary 3, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.1.68 on 1 January 1983. D

ownloaded from