|NCI/PDQ® Health professionals: Intraocular (Uveal) Melanoma Treatment (PDQ®)|
|National Cancer Institute|
| Last Modified: November 9, 2012
Melanoma of the uveal tract (iris, ciliary body, and choroid), though rare, is the most common primary intraocular malignancy in adults. The mean age-adjusted incidence of uveal melanoma in the United States is approximately 4.3 new cases per million population, with no clear variation by latitude. Males have a higher incidence than females (4.9 vs. 3.7 per million). 1 The age-adjusted incidence of this cancer has remained stable since at least the early 1970s. 1 2 U.S. incidence rates are low compared with the rates of other reporting countries, which vary from about 5.3 to 10.9 cases per million. Some of the variation may be the result of differences in inclusion criteria and methods of calculation. 1
Uveal melanoma is diagnosed mostly at older ages, with a progressively rising, age-specific, incidence rate that peaks near the age of 70 years. 3
In view of these susceptibility factors, numerous observational studies have attempted to explore the relationship between sunlight exposure and risk of uveal melanoma. To date, these studies have found only weak associations or yielded contradictory results. 3 Similarly, there is no consistent evidence that occupational exposure to UV light or other agents is a risk factor for uveal melanoma. 3 5
Uveal melanomas can arise in the anterior (iris) or the posterior (ciliary body or choroid) uveal tract. Iris melanomas have the best prognosis, whereas melanomas of the ciliary body have the least favorable prognosis. 6 Most uveal tract melanomas originate in the choroid. The ciliary body is less commonly a site of origin, and the iris is the least common. The comparatively low incidence of iris melanomas has been attributed to the characteristic features of these tumors, i.e., they tend to be small, slow growing, and relatively dormant in comparison with their posterior counterparts. Iris melanomas rarely metastasize. 7 Melanomas of the posterior uveal tract generally have a more malignant, histologic appearance; are detected later; and metastasize more frequently than iris melanomas. The typical choroidal melanoma is a brown, elevated, dome-shaped subretinal mass. The degree of pigmentation ranges from dark brown to totally amelanotic.
Most uveal melanomas are initially completely asymptomatic. As the tumor enlarges, it may cause distortion of the pupil (iris melanoma), blurred vision (ciliary body melanoma), or markedly decreased visual acuity caused by secondary retinal detachment (choroidal melanoma). Serous detachment of the retina may occur. If extensive detachment occurs, secondary angle-closure glaucoma occasionally develops. Clinically, several lesions simulate uveal melanoma, including metastatic carcinoma, posterior scleritis, and benign tumors, such as nevi and hemangiomas. 8Anatomy of the eye.
Careful examination by an experienced clinician remains the most important test to establish the presence of intraocular melanoma. It is not possible to distinguish a small uveal melanoma from a nevus. Small uveal lesions are often observed for growth to make a diagnosis of melanoma. Clinical findings that may help to identify melanoma include: 6
Ancillary diagnostic testing, including fluorescein angiography and ultrasonography, can be extremely valuable in establishing and/or confirming the diagnosis. 9 In a large, retrospective, single-center series of 2,514 consecutive patients with choroidal nevi, the progression rates to melanoma at 5, 10, and 15 years were 8.6%, 12.8%, and 17.3%, respectively. 10
Cell type is the most commonly used predictor of outcome following enucleation, with spindle-A cell melanomas carrying the best prognosis and epithelioid cell melanomas carrying the least favorable prognosis. 1 4 9 Nevertheless, most tumors have an admixture of cell types, and there is no clear consensus regarding the proportion of epithelioid cells that constitutes designation of a tumor as mixed or epithelioid. 6
Extraocular extension, recurrence, and metastasis are associated with an extremely poor prognosis, and long-term survival cannot be expected. 11 The 5-year mortality rate associated with metastasis from ciliary body or choroidal melanoma is approximately 30%, compared with a rate of 2% to 3% for iris melanomas. 12
Most primary intraocular melanomas contain variable proportions of epithelioid, spindle-A, and spindle-B cells (mixed-cell melanomas). Pure epithelioid-cell primary melanomas are infrequent (approximately 3% of cases). 1 In the Collaborative Ocular Melanoma Study, mixed-cell type melanomas predominated (86% of cases). 3
Uveal melanoma most often assumes a nodular or dome-shaped configuration, but occasionally tumors can be flat or diffuse and involve extensive areas of the uvea with little elevation.
Tumor size classifications according to boundary lines used in a Collaborative Ocular Melanoma Study (COMS) are as follows: 1
Although most ocular melanomas have a raised configuration, about 5% grow in a diffuse pattern that also may have prognostic significance. The tumors have a horizontal, flat-growth pattern, with the thickness measuring approximately 20% or less than the greatest basal dimension. This uncommon variant of uveal melanoma seems to have a poorer prognosis, particularly when the diameter is large, and the margins are poorly defined. 3
In clinical practice, the tumor base may be estimated in average optic disc diameters (1 dd = 1.5 mm). The average elevation may be estimated in diopters (3 diopters = 1 mm). Other techniques, such as ultrasonography, should be used to provide more accurate measurements.
An important function of ophthalmic ultrasonography is the detection of extrascleral extension. 4 5 Extrascleral extension measuring 2 mm or more in thickness can be demonstrated provided it is located behind the equator where the intraocular tumor, sclera, and adjacent orbital fat are readily imaged. 6 Orbital extraocular extension of choroidal melanoma may be found in eyes with medium and large tumors, but it is very rare in eyes with small melanomas.
Systemic metastases are evident in only 1% to 4% of patients at the time of diagnosis of the primary ocular melanoma. 7 Because the uveal tract is a vascular structure without lymphatic channels, tumor spread occurs principally by local extension and by dissemination through the blood stream. Lymphatic spread is rare but may occur after local extension into the conjunctiva and its lymphatics. 8 Given the rarity of nodal metastases, sentinel node biopsies of nonclinically involved nodes are not done as part of the staging procedure. 7
Systemic metastases are generally hematogenous in origin, and the first site identified is usually the liver. 9 Lung, bone, and subcutaneous sites are also common. 9 In the COMS trials, the liver was the only site of detectable metastasis in 46% of patients with metastases reported during follow-up or at the time of death; 43% had metastases diagnosed in the liver and other sites. 9 In patients with a history of ocular melanoma who present with hepatic metastases of unknown origin, metastatic melanoma should be considered in the differential diagnosis.
An American Joint Committee on Cancer staging system has been developed for melanoma of the uveal tract. 7
There are a variety of other factors that are not part of the AJCC staging system but that may help in refining estimates of prognosis. 7
There are a number of key prognostic features that are important to collect in malignant melanoma of the uvea, even though they are not included in staging algorithms. These include the following: 7
Iris melanomas have relatively good outcomes with a 5-year survival rate of more than 95%. They are predominantly of the spindle-cell type and are usually smaller in size than posterior melanomas because of earlier detection. Conservative management is generally advocated whenever possible, but surgical intervention may be justified with unequivocal tumor growth or with extensive disease at initial examination.
The management of small choroidal melanomas is controversial, and it is not clear whether treatment of small tumors prevents metastasis. 1 The natural history of small choroidal melanoma is poorly understood. Small, pigmented, choroidal lesions cannot always be differentiated reliably on examination. Growth is a presumed indicator of malignant potential. 2 The likelihood of progression from the time of diagnosis to the time when tumor growth warrants treatment has not been well characterized. Some ophthalmologists advocate observation. This has been justified on several grounds, including the difficulty with establishing a correct diagnosis, the lack of any documented efficacy for globe-conserving treatments, and concerns for severe treatment-related morbidity. Others have advocated earlier therapeutic intervention. 1 3 4
Although patients diagnosed with small choroidal tumors were not eligible for participation in the Collaborative Ocular Melanoma Study (COMS), these patients were offered participation in a prospective follow-up study to evaluate the natural history of small lesions. Two-year and 5-year tumor growth estimates of 21% and 31%, respectively, were reported. 5 Clinical risk factors associated with tumor growth included: 3 5
In the past, enucleation (eye removal) was the standard treatment for primary choroidal melanoma, and it is still used for large tumors. However, enucleation has been largely replaced by radiation therapy (i.e., brachytherapy with radioactive plaques; or external-beam, charged-particle radiation therapy) to spare the affected eye. 6 7
In the case of large, choroidal tumors judged to require enucleation, the role of pre-enucleation EBRT was tested in a randomized trial and had no impact on overall survival (OS). 8 9[Level of evidence: 1iiA] In a COMS trial, 1,003 patients with large choroidal melanomas (2 mm in height and 16 mm in diameter, or 10 mm in height irrespective of diameter, or 8 mm in height and border <2 mm from the optic disc) with no known metastases were randomly assigned to receive enucleation alone or after preoperative external photon-beam radiation from cobalt 60 or accelerators (20 Gy in five daily fractions) to the orbit and globe. 8 9 Through 10 years of follow-up, the median survival in both arms was approximately 7 years, and the 10-year all-cause mortality was 61% in both arms (relative risk for death of 1.00; 95% CI, 0.851.18). Metastasis-free survival was also virtually identical in both arms.
Eye-sparing transscleral local resection plays a very limited role in the management of uveal melanoma. It is used in patients with large choroidal and ciliary body tumors who are not candidates for radiation therapy but are highly motivated to retain their eye. 10 11 12 The procedure is technically demanding and is generally performed only in centers with specialized expertise in this surgery. There is a substantial risk of retinal detachment, intraocular bleeding, and complications associated with the anesthesia-induced hypotension used to decrease the risk of bleeding. Adjuvant brachytherapy or neoadjuvant proton-beam therapy is often administered. Experience is limited to retrospective, single-center, case series. 10 11 12[Level of evidence: 3iiiDiv]
Surgical resection of metastases from ocular melanoma has been reported in case series of highly selected patients with occasional favorable outcomes. 13 14 However, the favorable outcomes may be the result of strong patient-selection factors, and the role of resection in this setting is unclear. 13 14[Level of evidence: 3iiiDiv]
Episcleral brachytherapy using plaques containing small radioactive seeds is the most common form of radiation used in the management of intraocular melanoma. Iodine-125 (125I), cobalt-60 (60Co), palladium-103 (103Pd), iridium-192 (192Ir) and ruthenium-106 (106Ru) are examples of radioactive isotopes used in the brachytherapy plaques. Isotopes with relatively low photon and electron emissions (125I, 103Pd, and 106Ru) are more easily shielded to reduce the exposure to adjacent normal tissues, and 125I is probably the most commonly used radioisotope. 15 Although plaque radiation therapy allows preservation of the eye, visual acuity is frequently lost over time.
In a case series of 1,106 patients who were treated with plaque radiation therapy for uveal melanoma and who had an initial acuity of at least 20/100, 68% developed poor acuity (i.e., 20/200 or worse) within 10 years. 16
Factors associated with worse acuity outcomes included the following: 16
125I brachytherapy yields equivalent overall and melanoma metastasis-specific survival rates to enucleation for medium-sized melanomas. 17[Level of evidence: 1iiA] The randomized COMS Medium Tumor Trial compared 125I episcleral-plaque brachytherapy (85 Gy at 0.421.06 Gy/hr) to enucleation in 1,317 patients with medium-sized choroidal tumors (tumor height 2.5 mm10.0 mm and tumor diameter 16.0 mm that were not contiguous with the optic disc). 17 Eighty-five percent of the patients treated with 125I brachytherapy retained their eye for 5 years or more, and 37% of them had visual acuity better than 20/200 in the irradiated eye 5 years after treatment. 17 No statistically significant differences in mortality were observed between the two study arms after 12 years of follow-up, whether considering death from all causes or death with histopathologically confirmed melanoma metastasis. 18 Five- and 10-year all-cause mortality rates were 19% and 35% in both study arms; cumulative all-cause mortality at 12 years was 43% in the 125I arm versus 41% in the enucleation arm (RR, 1.04; 95% CI, 0.861.24). Five-year metastasis-specific mortality rates were 13% in both arms; at 10 years, the rates were 21% and 22% (RR for metastasis-specific mortality, 1.07; 95% CI, 0.811.41 through 12 years).
In a companion study within the COMS, 209 patients were prospectively assessed for quality of life during the first 5 years of follow-up. 19 Both study groups reported increasing difficulty with vision-oriented daily activities and ocular pain as time elapsed. Most measures of visual function were similar between the two groups, but there were statistically significant differences favoring the brachytherapy group in comfort with driving for the first year after therapy and in reported peripheral vision for the first two years after therapy. These differences disappeared by year 5 of follow-up. 19[Level of evidence: 1iiC]
Charged-particle EBRT (using protons, carbon ions, or helium ions) is the other major form of radiation therapy used in the management of ocular melanomas. 20 21 22 23 This form of radiation therapy requires sophisticated equipment available only at selected centers, and charged-particle EBRT involves patient cooperation during treatment (e.g., voluntarily fixating the eye on a particular point so the tumor is positioned properly in the radiation beam). A lower risk of early and late local radiation failures has been reported after charged-particle EBRT than after brachytherapy, possibly resulting from differences in dose distribution of the two techniques. 20[Levels of evidence: 1iiDiv and 3iiiDiv]
In a single-center, single-surgeon study 184 patients with uveal melanomas smaller than 15 mm in diameter and smaller than 10 mm in thickness were randomly assigned to receive 125I brachytherapy versus helium ion radiation (to an estimated dose of 70 Gy equivalents in five fractions over 7 to 11 days in each arm). 24 The local tumor regrowth rate by 4 years was 13.3% in the brachytherapy arm versus 0% in the helium ion arm (P < .001). However, the rates of metastasis, death from metastasis, and overall mortality was very similar in both arms. 24[Level of evidence: 1iiDiv]
Because of its dose distribution, charged-particle irradiation can be used to treat larger tumors and tumors closer to the fovea or optic disc than plaque brachytherapy. A large, single-center, single-surgeon series of 2,069 patients treated with proton-beam therapy had an actuarial local control rate of 95% (95% CI, 93%96%) at 15 years. The cumulative rate of enucleation was 16% (95% CI, 13%20%), most frequently as a result of neovascular glaucoma, blind uncomfortable eyes, or local recurrence (46%, 31%, and 23% of enucleations). As with plaque radiation, risk factors for deterioration in visual acuity after charged-particle radiation were tumor size, location near the fovea or optic disc, baseline acuity, and underlying diabetes. 21
Similarly, another large, single-center, single-surgeon, consecutive series of 886 patients treated with proton-beam irradiation reported a local control rate of 92.1% (95% CI, 89.8%94.6%) and ocular conservation rate of 87.3% (95% CI, 83.9%90.9%) at 10 years. 22][Level of evidence: 3iiDiv] The actuarial OS at 10 years was 64.1% (95% CI, 59.5%69.0%).
In a single-center, phase I/II study of 57 evaluable patients treated with carbon ion-beam irradiation and followed for a median of 26 months, only one patient had a local tumor recurrence, but 26 patients had enucleation caused by the development of neovascular glaucoma or severe eye pain from increased intraocular pressure. 23
In an attempt to lower the complication rate and improve functional outcome, a decreased dose of 50 cobalt Gy equivalents (CGE) has been compared to 70 CGE proton beam (each delivered in 5 fractions, usually within a 7-day period). Patients (n = 188) with tumors smaller than 15 mm in diameter and smaller than 5 mm in height that were located near the optic disc or macula were randomly assigned to the two doses in a double-masked study design. At 5 years, there were no statistically significant differences in local tumor control, rate of metastasis, visual acuity, or complication rates. However, the visual fields were better in the 50 CGE group. 25[Level of evidence: 1iDiv]
As noted above in the section on the Role of Surgery in the Treatment Option Overview section of this summary, the role of pre-enucleation external photon-beam radiation therapy has been tested in a randomized trial and has shown no impact on OS for large choroidal tumors treated with enucleation. 8 9
External-beamphoton-beam (gamma-ray) radiation therapy with gamma-knife stereotactic radiation surgery as a single-fraction 26 or fractionated stereotactic radiation 27 28 is being investigated as an alternative to brachytherapy or charged-beam radiation for posterior uveal melanomas, particularly for tumors too large or too close to the optic disc or macula to treat with brachytherapy. Because the dose rate of radiation delivery is slower than is the case with charged particles, specialized techniques are used to immobilize the eye 26 or to avoid delivery of the photons while the eye is moving or closed. 28 Experience is more limited with external-beamphoton therapy than for either brachytherapy or charged-particle EBRT, and there are no controlled comparisons to either of the other techniques. Early results from single-center series suggest similar levels of local tumor control and eye retention rates, but patient-selection factors may play a role. 28[Level of evidence: 3iiiDiv]
Transpupillary thermotherapy (TTT) directs an infrared laser, usually at a wavelength of 810 nm, through a dilated pupil in one or more sessions to induce heat necrosis of uveal melanomas. This method carries the theoretical advantage of high-precision destruction of tumor tissue under direct visualization. However, TTT has important limitations that confine its use to very restricted circumstances. 1 29 The limited ability of TTT to penetrate thick tumors with sufficient energy restricts its use to small melanomas, or tumors of a size that some ophthalmologists recommend for follow-up without any initial therapy. (Refer to the Role of Observation section in the Treatment Option Overview section of this summary for more information.) When used as the primary therapy, there are relatively high rates of local recurrence and retinal vascular damage. Recurrence rates are particularly high when the tumor abuts the optic nerve and overhangs the optic disc. 1[Level of evidence: 3iiiDiv]
In a single-center study, 95 patients with small choroidal melanomas (diameter <10 mm and thickness <3.5 mm) were randomly assigned to TTT versus 125I brachytherapy (100 Gy). 30 The tumor regression rates in the TTT and 125I arms were 92% and 98%, respectively (P = .4). With a mean follow-up time of 56.2 months, there were four recurrences in the TTT arm and one in the 125I arm. However, the study is too small to provide clear information on efficacy differences.
TTT is also under evaluation as an adjunct to primary therapy with proton-beam radiation. In the setting of large uveal melanomas, proton-beam therapy is associated with exudative, inflammatory, and glaucomatous complications that may require enucleation. In a single-center trial, 151 patients with uveal melanomas at least 7 mm thick or at least 15 mm in diameter were randomly assigned to receive proton-beam radiation (60 CGEs over four daily fractions) with or without TTT (810 nm wavelength at 1, 6, and 12 months after therapy) and followed for a median of 38 months. 31 There were no differences between the two groups in maculopathy, papillopathy, or glaucoma. The enucleation rate was lower in the TTT group (about 2% vs. 18% at 5 years, P = .02). However, the study was not masked, and replication would be important.
There are uncertainties regarding the optimal management of intraocular melanoma at all stages. Physicians should discuss with eligible patients opportunities for entry into ongoing clinical trials. Information about ongoing clinical trials is available from the NCI Web site.