Neuroblastoma Within a Congenital Orbital Teratoma

Congenital orbital teratomas are rarely malignant. In fact, a total of 4 prior reports exist in the literature describing malignant orbital teratomas, and only 2 of these describe the nature of the malignant cells. We report a case of a congenital orbital teratoma containing foci of neuroblastomatous elements confirmed by immunohistochemical techniques. Given that the tumor was noninvasive and no systemic involvement was discovered, no further treatment was undertaken. Two years following the operation, the patient remained free of tumor in the orbit and at other sites. A review of the literature concludes that similar cases should be handled by complete excision, and further therapy should be based on the pathological findings of the individual case.

Teratomas are generally benign congenital growths, which by definition consist of derivatives from each of the 3 germ-cell layers. They are most commonly found in the gonads, but they may develop ectopically in cervical, mediastinal, sacrococcygeal, intracranial, and orbital locations.¹ Primitive neuroectodermal tissue has long been known to be associated with teratomatous tissue in both intracranial² and ovarian sites.^3,4 One report found a neuroblastoma within a testicular teratoma.⁵

Primary congenital teratomas of the orbit are uncommon and rarely malignant. Four prior reports describe malignant orbital teratomas.^6-9 One case was primarily an intracranial teratoma involving the optic nerve.⁶ Another report had no description of the nature of the malignant cells.⁷ Garden and McManis⁸ described a case in which the malignant cell population appeared 3 years after the excision of a benign orbital teratoma and consisted of cells believed to be a malignant germ cell tumor. An orbital teratoma with neuroblastic and rhabdomyoblastic elements has also been documented.⁹ We report a case of orbital teratoma with malignant neuroblastomatous features.

The patient was the product of an uncomplicated pregnancy and was born at full term via normal abdominal delivery. She was initially seen at age 5 weeks for the evaluation of congenital and increasing proptosis of her left eye. At that time, grossly she had 4 to 5 mm of proptosis of that eye, and there was positive resistance to retropulsion. Pupils were 3 mm and sluggish bilaterally, without a relative afferent defect. The anterior chamber was quiet and of moderate depth. Biomicroscopy and fundoscopy were unremarkable. Cycloplegic retinoscopy revealed a refraction of +3.50 +.50 × 80 OD and +5.50 +1.50 × 90 OS. A magnetic resonance imaging scan showed a diffuse mass that enhanced with contrast in the left orbit. An orbital ultrasound revealed cystic cavities with fluid levels present within the orbit. Lymphangioma was believed to be the most likely diagnosis, and the mass was observed.

By 6 months of age, the patient had progressive proptosis of her left eye. The extraocular movements remained intact, but the left eye was unable to fix. External examination of the left eye revealed a palpable mass in the inferolateral orbit. No relative afferent pupillary defect was present. Cycloplegic retinoscopy revealed increased hyperopia with a refraction of +4.00 +1.50 × 90 OD and +14.00 +3.00 × 85 OS. The right fundus was unremarkable, while the left showed chronic papilledema characterized by a pale halo of edema around the disc. An orbital computed tomography scan showed a septated intraconal mass with calcifications that had increased in size from the previous magnetic resonance imaging findings, and had extended into the lacrimal fossa (Figure 1). The orbital rim was expanded and thinned, but there was no osseous destruction. The orbital tumor was excised by a left lateral orbitotomy.

The specimen consisted of 3 pieces of tissue, the largest measuring 30 × 19 × 9 mm. The specimen was bisected, and examination revealed areas of a serosalike surface. In addition, there was one cystic area filled with gelatinous material and hair. Another area was firm and gray, suggestive of cartilage.

Microscopic examination revealed a complex cystic structure lined partially by pseudostratified ciliated columnar epithelium and partially by keratinized stratified squamous epithelium. There were adnexal structures in the portion of the wall lined by the squamous epithelium. The cyst was filled with keratinous debris. Adjacent to the cyst was a well-circumscribed cartilaginous structure. In addition, there was mature central nervous system tissue present. In one focal area associated with the central nervous system tissue, there was a focus (measuring 5 × 4 mm) of undifferentiated, small, round cells with modest amounts of eosinophilic cytoplasm (Figure 2 and Figure 3). The chromatin of the cells was moderately condensed, and there were occasional small nucleoli. There was 1 mitosis in 20 high-power fields. There were no rosettes present.

Paraffin-embedded sections were stained using an immunoperoxidase technique. The small, undifferentiated cells displayed negative immunoreactivity with keratin, CD99 (product of MIC2 gene, positive in glioblastoma, ependymoma, Ewing sarcoma, and acute lymphoblastic leukemia), and vimentin. There was positive immunoreactivity with neuron-specific enolase (Figure 4) and synaptophysin. The pattern of reactivity was consistent with neuroblastoma, and would be unlikely in any of the morphologic differentials in this case, which include leukemic infiltrates, Ewing sarcoma, primitive neuroectodermal tumor, and Wilms tumor.

The intriguing aspect of this case was the presence of immature cells consistent with neuroblastoma in an otherwise typical orbital teratoma. The presence of these cells raised interesting questions in the subsequent management of the patient. Some authorities feel that the presence of immature cellular components should not be taken as evidence of malignancy.⁴ However, the prognosis of central nervous system teratomas is worse for the immature teratoma than for the mature teratoma, and dissemination of tumor cells systemically and in cerebrospinal fluid has been documented in immature intracranial teratomas.²

Mayberger et al,⁴ in a report on a benign cystic teratoma of the ovary with neuroblastomatous change, suggested that if the teratoma is noninvasive, without metastasis, then removal alone is adequate. This suggestion relies on the theory that the primitive neuroectodermal tissue within the teratoma is derived from ectodermal germ cells and not from mature neural tissue, which developed into undifferentiated tissue. However, a separate study warned that the clinician must be aware that on occasion, primary neuroectodermal tumors arise ectopically in the ovary, along with teratomatous tissue.³ These latter tumors are known to metastasize, with the risk increasing as the degree of differentiation of the neuroectodermal tissue decreases. The same study suggested that tumors with endodermal and mesodermal features along with primitive neuroectodermal tissue may be classified and treated as benign teratomas, while those with massive, confluent growth of neuroectodermal tissue and little of the other 2 germ layers may be primary malignant tumors.

Ulbright et al⁵ argue that treatment should be individualized depending on the type of non–germ cell tumor involved, speculating that some forms of non–germ cell malignancies (specifically leiomyosarcoma, nephroblastoma, neuroblastoma, and giant cell tumor), when found within teratomas, may not affect the prognosis.⁵ As a result, they argue that such tumors should be called "non-germ cell malignancy in teratoma" and advocate treatment of the growth as a benign teratoma with complete resection. The results of their study further indicate that the role of chemotherapy is questionable in such tumors.

The management of this case was discussed at the Pediatric Tumor Board. Based on the above considerations, this tumor was classified as an immature teratoma containing neuroblastomatous elements. The tumor had been completely excised, and no other treatment was undertaken. The child has been followed up for 2 years and has no evidence of recurrent orbital tumor or neuroblastoma at other sites.

In conclusion, orbital teratomas may rarely contain non–germ cell malignancies. The initial clinical management of these cases should be complete excision. Decisions regarding adjunctive chemotherapy or radiation therapy should be based on the pathologic findings for the individual case.

Corresponding author and reprints: David J. Wilson, MD, 3375 SW Terwilliger Blvd, Casey Eye Institute, Portland, OR 97201-4197 (e-mail: wilsonda@ohsu.edu).