Primary orbital melanoma is an exceedingly rare tumor1 thatprobably develops from congenital rests of neural crest cells in the orbit.2 It represents less than 1% of primary orbital neoplasmsand usually occurs in the presence of clinical or histological evidence ofocular melanosis or blue nevus syndrome.2 Althoughorbital melanoma in general carries a poor prognosis, our case was unusualfor its extremely aggressive clinical course that resulted in the death ofthe patient 6 months after presentation.
A 40-year-old white man was first seen by us in July 1998 with a painless3-mm proptosis of the left eye of 6 months' duration. There was no clinicalevidence of ocular melanosis or blue nevus syndrome. A computed tomographicscan of the orbits revealed a well-defined intraconal mass that was discretefrom the optic nerve and the horizontal rectus muscles (Figure 1). In the following 6 weeks, the proptosis enlarged, progressingfrom 3 mm to 8 mm. There was associated conjunctival exposure with ulceration,marked global restriction of movement, and transient visual obscurations.Optic nerve function was preserved, with Snellen visual acuity retained at6/5, normal color vision, and intact pupillary reactions. A lateral orbitotomywith excision biopsy was performed in September 1998. A large pigmented masswas removed, histological analysis of which showed the lesion to be an orbitalmelanoma. A cilioretinal artery occlusion occurred postoperatively, reducingthe left visual acuity to light perception.
Axial computed tomographic scanobtained at presentation shows a left intraconal mass, causing a mild leftproptosis.
Histologically, it is not possible to differentiate a primary from ametastatic orbital melanoma; therefore, we performed a thorough systemic evaluationfor a primary source. Detailed clinical examination augmented with liver functiontests; liver ultrasonography; computed tomographic imaging of the brain, chest,and abdomen; and a bone scan showed no evidence of systemic melanoma. OcularB-scan ultrasonography, performed to look for an intraocular primary tumor,suggested a possible small anterior ciliary body lesion with a collar-studappearance and measuring 1.4 mm.
Five weeks after the excision biopsy, there was an alarmingly rapidrecurrence of the left proptosis (Figure 2 and Figure 3). Magneticresonance imaging revealed local tumor regrowth occupying the inferior aspectof the left orbit and extending superiorly as far as the superior rectus muscle.There was no evidence of spread beyond the orbit. A left exenteration withpreservation of the eyelids was carried out in November 1998. Postoperativeradiotherapy, to be followed by chemotherapy and adjunctive interferon treatment,was planned. However, 3 days before the commencement of radiation treatment,a firm swelling was noted at the lateral aspect of the left orbit. Resultsof fine-needle aspiration biopsy confirmed the presence of malignant melanomacells, indicating that surgical clearance had not been achieved.
Resolution of left proptosis afterlateral orbitotomy and excision biopsy.
Preexenteration photograph showsaggressive tumor recurrence only 5 weeks after excision biopsy, with a largemass filling the left orbit and causing massive proptosis and conjunctivalprolapse.
The patient's clinical course rapidly deteriorated, with massive anddisfiguring growth of the melanoma within the exenteration cavity. Magneticresonance images showed that the extensive melanoma deposit was causing hugesoft tissue expansion, with extension into the paranasal sinuses and lefttemporal fossa and destruction of the anterior aspect of the sphenoid wing(Figure 4). In December 1998, therewas evidence of metastatic disease with distant dissemination to the anteriorchest wall and the liver. Treatment with subcutaneous interferon alfa-2a,20 mIU given on alternate days, failed to alter the course of the disease,and the patient died 2 weeks later.
Axial magnetic resonance imagedemonstrates extensive melanoma recurrence, causing huge and disfiguring softtissue expansion with direct invasion of the left temporal fossa and erosionthrough to the left maxillary antrum.
A dark irregular mass measuring 23 mm × 32 mm was excised. Sectionsshowed a partly necrotic, partly pigmented mass consisting of spindle cells(Figure 5). There were frequentmitoses, considerable nuclear pleomorphism, and large numbers of small microvessels,indicating a vascular tumor. The pigment, which was patchy, was negative foriron (Perls test). The tumor cells stained positively for NKIC3 antibody,S100 protein, and melanoma-associated antigen (HMB45), indicating their melanocyticnature.
Excision biopsy specimen showsan area of mainly spindle-shaped melanoma cells (hematoxylin-eosin, originalmagnification ×180).
The exenteration specimen (Figure 6) showed extensive infiltration of the orbital fibrofatty tissueby a high-grade malignant melanoma. The tumor was now formed from clustersof medium to large epithelioid cells (Figure7), but with a staining pattern similar to that of the excisionbiopsy. There was a high mitotic rate (more than 50 mitoses per 10 ×40 objective fields) and geographic areas of necrosis. There was no evidenceof congenital melanosis oculi. Biopsy specimens taken from the margins ofexcision showed that the skin from the left lateral eyelid and inferior orbitalrim was clear, but residual tumor was found at the inferior orbital rim adjacentto the left infraorbital nerve and at the left orbital apex. Ten-micrometerhistological sections were made of the entire globe (with every 10th sectionsubmitted for staining and examination) and demonstrated no pathological evidenceof an intraocular or ciliary-body melanoma as was suggested on ultrasonography.
Macroscopic view of the left orbitalexenteration specimen shows preservation of the normal architecture of theeye but widespread infiltration of the retrobulbar fibrous and adipose tissueby melanoma. No evidence of congenital melanosis oculi is seen.
Exenteration specimen shows anarea composed of mainly epithelioid melanoma cells (hematoxylin-eosin, originalmagnification ×200).
We found no histological evidence of any predisposing disease, eitherin the form of blue nevi or ocular melanosis in the excision biopsy or exenterationspecimen. The overall conclusion was, therefore, of a highly aggressive malignantmelanoma that, in the absence of an intraocular or systemic primary tumor,appeared to have arisen de novo within the orbit.
Primary orbital melanoma is a rare condition that is histopathologicallysimilar to uveal melanoma.2 It is primarilyseen in white adults who seek care most commonly for painless proptosis andhas an average age at onset of 42 years.2 Predisposingocular features are present in most cases. In the series by Tellado et al,2 47.5% of the patients had clinical evidence ofsome form of congenital melanosis. Intraorbital blue nevus was recognizedhistologically in 90%. The case presented herein is remarkable for the absenceof both. Had there been any evidence of ocular melanosis, this patient wouldprobably have proceeded, after orbital computed tomography, to excision biopsysooner than actually occurred, as clinical suspicion would have been greater.
It is possible to miss small foci of pigmentation, both clinically andhistologically, that may represent ocular, oculodermal, or orbital melanocytosis.Detailed inspection of the gross specimens of both the excision biopsy andthe exenteration took place to look for minute areas of pigmentation thatcould have predisposed to the development of melanoma. This was augmentedby a thorough intraoperative examination of the orbital periosteum, with multiplebiopsy specimens taken from suspicious areas of grey, brown, or black pigmentedareas, specifically at the orbital apex and the lateral orbital rim. No evidenceof melanocytosis was found. We accept, however, as no autopsy was performed,that it is possible (although unlikely) that an isolated area of melanocytosisfrom which the melanoma could have arisen may have been overlooked.
B-scan ultrasonographic examination suggested that the orbital massrepresented spread from an anterior ciliary primary melanoma, tumors thatare known to behave more aggressively than more posteriorly located lesions.3 Histological examination confirmed that this wasnot the case. The B-scan had been performed the day after excision biopsy,suggesting that the "lesion" observed represented a postoperative artefact,as it would be unlikely that our method of histopathological sampling couldhave missed a lesion picked up on ultrasonography 5 weeks earlier and estimatedto be at least 1.4 mm.
Although primary orbital melanoma carries a poor prognosis, with a 4.5-yearmortality of 38%,2 our case represents aparticularly aggressive lesion. The best approach to treatment of orbitalmelanoma is not scientifically known.1 Inour case, an active multidisciplinary approach to tumor control was taken.Surgical excision and local radiotherapy to the orbit followed by chemotherapyand adjunctive interferon treatment was considered the combination most likelyto achieve success. However, the very active nature of tumor recurrence, followingboth the excision biopsy and exenteration, was suprising, and the likelihoodof radiotherapy or chemotherapy controlling such macroscopic disease was remote.1,4- 6 Radicalskull base surgery, which entails an intraoperative mortality rate of severalpercent, was considered. However, it was highly improbable that even thiskind of surgery could have successfully eradicated the tumor, and the patientdeclined further intervention.
This unusual case report of a rare but life-threatening condition illustratesthat primary orbital melanoma should be considered in cases of rapidly progressingproptosis in young white patients, even where no predisposing ocular diseaseis present.
Corresponding author: Susan Hague, FRCOphth, Radcliffe Infirmary,Oxford Eye Hospital, Woodstock Road, Oxford OX2 6HE, England (e-mail: firstname.lastname@example.org)
Delaney YM, Hague S, McDonald B. Aggressive Primary Orbital Melanoma in a Young White Man With No PredisposingOcular Features. Arch Ophthalmol. 2004;122(1):118-121. doi:10.1001/archopht.122.1.118