Customize your JAMA Network experience by selecting one or more topics from the list below.
The differential diagnosis of malignant melanoma of the uveal tract
continues to challenge clinicians. In recent decades, we have come to expect
a 95% or better accuracy in the clinical diagnosis of malignant melanoma,
and in some centers such as the Wills Eye Institute, Baltimore, Md,1 and the Mayo Clinic, Rochester, Minn,2
96.3% and 97.7% accuracies, respectively, have been reported. In 1990, a North
American prospective randomized multicenter study group, the Collaborative
Ocular Melanoma Study (COMS) Group, reported a 99.5% accuracy in the diagnosis
of choroidal melanoma.3 By 1998, the COMS
had enrolled more than 2300 eyes into the study, and among the 1532 eyes randomized
and enucleated, 1527 were found to have a choroidal melanoma, resulting in
a diagnostic accuracy of 99.67%.4-6
The 5 misdiagnosed cases included 4 adenocarcinomas that metastasized to the
choroid and 1 choroidal hemangioma. The low misdiagnosis rate in the COMS
is a reflection of the level of expertise of the ophthalmologists participating
in this study, but it also results from the study aim, which was to include
only patients whose tumors could be adequately visualized and were believed
to clearly fit the criteria of malignant choroidal melanoma. The COMS reported
2 additional misdiagnosed cases that underwent enucleation after original
treatment with iodine I 125 (125I) brachytherapy. Both were melanocytomas
discovered in eyes that initially had been randomized to 125I brachytherapy
for presumed melanoma, and both were later enucleated because of tumor growth.3 Similarly, we report an additional case of a choroidal
tumor that received 125I brachytherapy as primary treatment of
presumed malignant melanoma and that was identified as melanocytoma after
histopathologic examination of the enucleated specimen.
Report of a Case
A 60-year-old woman was initially examined at the Mayo Clinic in 1990
and was found to have a pigmented 5 × 4.5-mm choroidal lesion exhibiting
lipofuscin and a collar-button configuration. Subretinal fluid associated
with the tumor extended into the macula, reducing visual acuity (Figure 1). Results of fluorescein angiography
demonstrated patchy hypofluorescence and late tissue staining within the tumor.
Results of ultrasonography demonstrated a solid lesion with choroidal excavation
and low levels of internal reflectivity (A-scan [Figure 2]). The tumor was 2.9 mm thick. In accordance with the patient's
strong preference to have the tumor treated with brachytherapy, she was not
enrolled in the COMS, and 125I brachytherapy was administered.
Three years after brachytherapy, extensive radiation retinopathy was observed.
Five years after brachytherapy, in 1996, the eye became blind and painful
owing to neovascular glaucoma. The tumor measured 3.2 mm in thickness by ultrasonography.
The eye was enucleated.
Fundus photographs showing tumor
with collar-button configuration, lipofuscin, and subretinal fluid.
Results of ultrasonography demonstrating
solid lesion with choroidal excavation (A) and low internal reflectivity on
On gross pathological examination, findings included a black mushroom-shaped
mass immediately temporal to the optic nerve, with a measurement of 7 ×
7 × 3 mm (Figure 3). Histologically,
it corresponded to a heavily pigmented choroidal tumor that had broken through
Bruch's membrane to infiltrate the overlying retina (Figure 4). Bleached sections showed a uniform population of large
cells with abundant cytoplasm and small, bland nuclei consistent with the
diagnosis of melanocytoma (Figure 5).
Occasional inconspicuous nucleoli were present. Necrosis and mitotic figures
were not seen. The histocytic marker CD68 did not reveal foamy macrophages.
Results of electron microscopy showed type II melanocytoma cells with small
melanosomes and premelanosomes measuring less than 1 µm in diameter
Gross specimen showing the collar-button
configuration of heavily pigmented tumor that had broken through Bruch's membrane.
Medium-power photomicrograph demonstrating
the collar-button configuration of the tumor, resulting from the invasion
through Bruch's membrane into the retina. Retinal detachment, cystic degeneration,
and exudates are observed (hematoxylin-eosin, original magnification ×100).
Bleached sections showing neoplastic
cells with small, bland-appearing nuclei and abundant cytoplasm. Necrosis,
mitosis, and nucleoli are absent (hematoxylin-eosin, original magnification
Results of electron microscopy
showing type II melanocytoma cells with small melanosomes of less than 1 µm
in diameter (lead citrate, original magnification ×26 000).
Melanocytomas of the eye may occur in any location within the uveal
tract, but most commonly they arise from pigmented cells that are incorporated
within the optic disc.7 Melanocytomas of
the optic nerve head are generally heavily pigmented and frequently have a
feathered black edge that extends into the adjacent retina.8
In this typical location, the clinical diagnosis of melanocytoma is usually
not a problem. However, the diagnosis of this tumor elsewhere in the uveal
tract may prove to be more of a challenge. In the ciliary body or iris, melanocytomas
tend to be heavily pigmented, and the correct diagnosis is suspected on the
basis of this clinical feature. However, melanocytomas involving the choroid
are unusual.9 Choroidal melanocytomas have
been reported as part of paraneoplastic syndromes (super nevus syndrome or
bilateral diffuse uveal melanocytic proliferation syndrome10-12),
and in those cases, they tend to be very dark, multiple, and bilateral, and
the correct diagnosis may be suspected clinically.
Several clinical features in our case, eg, the color, the presence of
lipofuscin and subretinal fluid, and the collar-button configuration, suggested
the clinical diagnosis of choroidal melanoma. In fact, the presence of a collar-button
configuration is unusual in any tumor other than malignant melanoma, and its
presence alone strongly suggests that diagnosis. However, this finding is
not pathognomonic of melanoma; it has been reported in association with melanocytoma13 as well as a tumor of the retinal pigment epithelium.14 Tumor growth in this case also supported the diagnosis
of malignant melanoma, as melanocytomas often remain clinically stable except
when m occurring with the super nevus syndrome.10,12,15
In our case, results of fluorescein angiography and ultrasonography
also supported the diagnosis of melanoma. The fluorescein study showed patchy
hyperfluorescence and late tissue staining of the tumor. A typical melanocytoma
would be expected to obscure background fluorescein significantly, but this
was not observed in our case. The ultrasonographic features in our case included
a relatively low level ofinternal reflectivity with acoustic hollowing and
choroidal excavation, findings characteristic of a melanoma rather than a
melanocytoma, which has been observed to have high levels of internal reflectivity.16
The classic histological appearance of a melanocytoma, such as that
seen in our case, is a heavily pigmented neoplasm with abundant cytoplasmic
pigment that obscures the nuclear features. Bleached sections disclose large,
uniform cells with a distinct cytoplasmic membrane, abundant cytoplasm, and
a rounded, small, bland-appearing nucleus that may or may not contain an inconspicuous
nucleolus.7 The difficulties in differentiating
the histological features of benign and malignant pigmented lesions have long
been recognized and have been the subject of individual publications.17,18
Is there an indication to perform a fine-needle aspiration (FNA) biopsy
in this case? Fine-needle aspiration biopsy is generally reserved for cases
with atypical clinical features, and since this case presented with clinical
characteristics typical of melanoma, FNA biopsy was not considered. Even if
an FNA biopsy had been performed, we have no guarantee that a conclusive diagnosis
could be achieved because of the difficulties in obtaining diagnostic material
that properly samples the tumor and because of difficulties interpreting the
cells. Shetlar and colleagues19 reported
a case of choroidal neoplasm diagnosed by means of FNA biopsy. In that case,
a diagnosis of melanoma was rendered, but at enucleation the tumor proved
to be a composite tumor consisting of 2 different cell populations. Melanocytoma
occupied one side of the tumor, and malignant melanoma occupied the other.
Melanocytomas are generally easily recognized by ophthalmic pathologists
familiar with this entity. Irradiation may alter the histological features
of an intraocular tumor, confounding interpretation of the cell type. In case
of extensive tumor necrosis, one may not be able to recognize tumor type with
certainty. Radiation therapy may produce, in addition to tumor necrosis, a
macrophage response with numerous pigment-laden histiocytes. The interpretation
of such cells may prove to be a challenge to the pathologist during light
microscopy. However, histiocytic markers, eg, CD68, are helpful in this differential
diagnosis. Foamy macrophages were not seen in our case, and histological examination
of the tumor showed little evidence of necrosis 5 years after brachytherapy.
Ultrastructural studies may help support the diagnosis of melanocytoma and
differentiate between 2 types of melanocytoma cells. The type II melanocytoma
cells containing small melanosomes were seen in our case (Figure 6). These cells are considered more metabolically active
than type I cells.20
In conclusion, this case illustrates the difficulty in differentiating
a choroidal melanocytoma from a choroidal melanoma on the basis of clinical
features alone. It remains uncertain whether the histological features in
this case accurately portrayed the tumor characteristics that existed before
radiation; however, the absence of necrosis and fibrosis support our impression
that this neoplasm was always a melanocytoma that mimicked a malignant melanoma.
This study was supported in part by a grant from Research to Prevent
Blindness, Inc, New York, NY, and in part by the Mayo Foundation, Rochester,
We thank Bonnie Ronken for secretarial support and Cheryl Hann, MS,
for support with the electron microscopy studies.
Corresponding author and reprints: Dennis M. Robertson, MD, Department
of Ophthalmology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (e-mail: firstname.lastname@example.org)
Robertson DM, Campbell RJ, Salomão DR. Mushroom-Shaped Choroidal Melanocytoma Mimicking Malignant Melanoma. Arch Ophthalmol. 2002;120(1):82–85. doi:
Monkeypox Resource Center