Melanoma is a relatively rare tumor and appears when melanocytes, the cells that form the pigment (melanin) in the skin, hair and eyes, undergo a transformation and become malignant. If the melanoma occurs in the uvea or uveal tract, one of the three layers from the eye wall, it is called uveal melanoma. The uvea consists of the iris (the colored area in the front of the eye), the ciliary body (a ring of muscle tissue behind the iris that changes the size of the pupil) and choroid (a layer of blood vessels that brings oxygen and nutrients to the eye). In a study of 8033 patients by Shields et al.(1), the tumor was located in the iris in 285 patients (4% of cases), in the ciliary body in 492 patients (6% of cases), and in the choroid in 7256 patients (90% of cases). Therefore, choroidal malignant melanoma is the most common primary intraocular tumor and the second most common primary malignant melanoma in humans.
1. a) The incidence of primary ocular malignant melanoma in Europe, according to Virgili et al.(2), follows a decreasing gradient from north to south, from a minimum of two cases per million per year in Spain and southern Italy to a maximum of eight cases per million per year in Norway and Denmark. According to data provided by Singh et al.(3), from the National Cancer Institute of the United States of America, for a period of 36 years, between 1973 and 2008, that included 4070 patients with primary uveal melanoma, an adjusted incidence was reported according to age of 5.1 cases per million per year. Uveal melanoma is the most common primary malignant intraocular tumor and the second most common type of malignant melanoma in the body.
1. b) Risk factors in uveal melanoma. Lutz et al.(4) stated in an article that a family history of uveal melanoma and preexisting ocular nevi were considered predisposing risk factors for this ocular neoplasia. A wide variety of risk factors have been investigated, such as factors related to sex, socioeconomic factors or factors related to social status, light skin and hair color, eye color, sun exposure, smoking, viruses, exposure to chemical substances, but no causative agent has been recognized. Following advances in understanding the pathogenesis of cutaneous melanoma, several studies have investigated the role of ultraviolet light and found a positive association with this disease (a relative risk significantly greater than 3). However, these results remain controversial and a recent decline in the incidence of cutaneous melanoma in the United States is further evidence that ultraviolet exposure is not a major risk factor. In addition, dosimetry studies suggest that only 1-2% of incident radiation in type B ultraviolet range reaches the back of the eyeball.
Several studies have addressed the possible importance of genetic factors, following the demonstration of genetically determined predisposition in other cancers(4). Clear evidence of a genetic component in uveal melanoma is the unbalanced racial risk, with a strong predominance among white people and an absence among people with dark skin(4). There is a clear racial predisposition that follows the same pattern as cutaneous melanoma. The familial predisposition in uveal melanoma has been known since 1892, following Silcok’s report on the appearance of the tumor at three successive generations(4).
However, such families are rare and, although many family reports have been published since then, it is clear that genetic factors are only a part of the etiological process(4). A specific genetic mutation, if any, would probably only explain a small proportion of disease, as it has been shown for the BRCA1 and BRCA2 genes in breast cancer(4). Currently, the data therefore suggest that environmental risk factors do exist, but these are yet to be demonstrated(4).
1. c) Screening. Uveal melanoma remains asymptomatic for long periods of time and can be discovered by chance on a routine eye exam. As a result, primary eye tumor screening is difficult to perform in the general population. Instead, screening patients diagnosed with uveal melanoma for metastases can be performed by several methods:
Liver function tests
Ultrasonographic examination of internal organs
Magnetic resonance imaging (MRI) of internal organs, especially liver
Computed tomographic scanning of the chest, abdomen and pelvis, from different angles
Positron emission tomography (PET) scan to detect malignant cells in the body.
Uveal melanomas can be classified into anterior uveal melanomas, when the tumor appears in the iris, and posterior uveal melanomas, when the tumor arises in the choroid or ciliary body.
2. a) Clinical aspects. Uveal melanomas remain asymptomatic for long periods. In general, in case of an anterior uveal melanoma, the ocular symptoms appear much later. According to Garcia-Valenzuela et al.(5), uveal melanoma may have the following symptoms:
Blurred visual acuity
Progressive and painless visual field loss
Severe ocular pain
Weight loss, marked fatigue, cough or changes in bowel or bladder habits (which suggests a primary non-ocular malignancy with choroidal metastases).
According to the same authors, ophthalmological examination may reveal the following(5):
Small choroidal melanomas typically have a nodular or well-circumscribed dome-shaped appearance located below the retinal pigment epithelium.
As choroidal melanomas grow, they may adopt more irregular shapes, such as bilobular, multilobular or fungal shape.
Diffuse choroidal melanoma, characterized by lateral growth throughout the choroid with minimal elevation of the retinal plane, is more difficult to diagnose and often produces exudative retinal detachment.
Choroidal melanomas can have a variable coloration, ranging from amelanotic to intensely pigmented; some are partially pigmented.
If the tumor is light-colored, its abnormal vascularization can usually be seen on ophthalmological examination.
Usually, there can be seen retinal pigment epithelial changes (like ocular drusen), areas of atrophy and orange discoloration overlying the choroidal melanoma; orange modifications in fundus eye can appear in both malignant and benign tumors.
Choroidal melanoma may remain undiagnosed underneath a large serous retinal detachment, retinal hemorrhage or a vitreous hemorrhage.
Rare presentation of advanced choroidal melanoma may be in the form of a painful red eye with cataracts and proptosis due to transscleral extension of the tumor in the orbit.
A malignant ciliary body melanoma may have sentinel vessels (dilated episcleral blood vessels visible through the conjunctiva) that feed the metabolically active tumor.
Transscleral growth of a choroidal malignant melanoma, especially an anterior uveal melanoma (mainly through emissary channels) may occur on ophthalmological examination as a small subconjunctival area of abnormal hyperpigmentation.
2. b) Paraclinical investigations
A-scan ultrasonography is useful for tumors thicker than 2-3 mm(5).
B-scan ultrasonography of the eye is a routine test used to assess any mass of the posterior segment, being especially useful in patients with opaque ocular media(5).
For choroidal melanomas, B-scan ultrasonography is useful in establishing the diagnosis, in evaluating the possible extraocular extension, in estimating the tumor size for periodic follow-up and for planning the therapeutic management(5).
Ultrasound biomicroscopy (UBM) has the following advantages(5):
It may differentiate anterior choroidal melanomas from those of the ciliary body.
It can help define the anterior edges of the tumor.
It provides excellent resolution of the anterior ocular structures.
Fluorescein and indocyanine green angiography does not show pathognomonic signs for uveal malignant melanoma but may help establish the diagnosis.
Small choroidal melanomas may present angiographic changes similar to those of choroidal nevi with changes that may range from normal angiography to hypofluorescence secondary to blockage of background fluorescence.
Large choroidal melanomas may show a patchy pattern of early hypofluorescence and hyperfluorescence followed by late intense staining.
Simultaneous fluorescence of the retinal and choroidal circulation in the tumor is fairly suggestive for choroidal melanomas.
Chest X-ray to rule out possible lung metastases.
Computed tomographic scanning of the globe and orbit is useful for visualizing extraocular extension and can differentiate a choroidal or retinal detachment from a solid tumor mass.
Magnetic resonance imaging of the globe and orbit can be used to determine the extrascleral extension of melanoma and to differentiate the fluid around the tumor.
3. Peculiarities of evolution
According to Grossniklauss et al.(6), in the 1970s it became apparent that a substantial number of people with uveal melanoma died due to metastases, mainly in the liver, after enucleation of the eye with melanoma. Data from the Institute of Pathology of the Armed Forces showed that the mortality rate peak at about 2 to 3 years after enucleation(6). In 1978, Zimmerman et al.(7), from the Institute of Pathology of the Armed Forces, published these findings and hypothesized that the act of enucleation may have contributed to the release of tumor cells into the systemic circulation.
This observation of a maximum mortality rate 2-3 years after enucleation became known as the Zimmerman Effect, while the iatrogenic contribution to the systemic spread of melanoma was known as the Zimmerman Hypothesis(6). The Zimmerman Hypothesis led to the establishment of the National Eye Institute, sponsored by the Collaborative Ocular Melanoma Study (COMS)(8). COMS consisted of two randomized prospective clinical trials and a third observational study on small melanomas to determine their natural history(6). In the largest study for medium-sized uveal melanomas, patients were separated into those with enucleation and those with brachytherapy with radioactive plaque(6).
The third study for large uveal melanomas was to determine if external beam irradiation before enucleation provided a survival benefit compared to enucleation alone(6).
For medium-sized melanomas, the survival statistics were virtually identical for patients who were treated with plaque brachytherapy compared to patients treated with enucleation(9). For large melanomas, external beam radiation did not provide any survival benefit and in some cases the prognosis was worse(10).
In addition, COMS showed that there was a maximum incidence of mortality at approximately 3 years in plaque brachytherapy as well as in the case of enucleation(11). Although there was a maximum incidence of mortality at about 3 years after the treatment of uveal melanoma, there was no evidence that enucleation surgery promoted the spread of tumor cells(6). Thus, the Zimmerman Effect was supported and the Zimmerman Hypothesis was disproven(12). It should be noted that this effect has been observed in other malignancies(6).
Grossniklauss et al.(6) consider that there are two main components of the metastatic process that may explain the Zimmerman Effect: (1) the properties of primary uveal melanoma and (2) the host mechanisms involved in the control of preexisting metastatic disease. These authors assumed that the microenvironment of the liver, the final organ of most metastases, participates in limiting the local tumor growth. This microenvironment niche includes sinusoidal spaces and periportal areas. This led these authors to believe that metastases are ongoing and random, but there is a change in the hepatic microenvironment after enucleation/irradiation of the primary tumor that allows micrometastases to survive and grow(13).
Terai and et al.(14) consider that the mechanisms of development of liver metastases are still very speculative. They assume that multiple factors contribute to the development of metastases and the growth of uveal melanoma cells in the liver. The mechanisms proposed by these authors are summarized as follows:
Slow hepatic blood circulation. The hepatic sinusoids are located at a confluence of arterial blood (hepatic artery) and venous blood (portal vein), where oxygen-rich arterial blood in the hepatic artery mixes with nutrient-rich venous blood in the portal vein. The hepatic sinusoids are a type of blood capillary with discontinuous, fenestrated endothelium. Slow blood flow at this level maximizes the contact between liver cells and pathogenic molecules to allow their filtration before reaching the systemic circulation. Slow and tortuous sinusoidal blood flow can hold uveal melanoma cells in the liver(15,16).
The interaction between chemoattractants and their receptors. Chemokines produced in the liver could attract uveal melanoma cells to the liver and interact with chemokine receptors on the surface of liver cells(14). A typical example is the interaction between CXCR4 and its CXCL12 ligand which is present in large amounts in the liver(17). The primary cell lines of uveal melanoma express CXCR4. Blockade of CXCR4 in uveal melanoma significantly reduces the migration of its cells to a human liver extract(18). An alternative explanation for chemokine-related liver tropism is the loss of chemokine receptors in the liver. It has been reported that extracts from the human liver downregulated the expression of CXCR4 and CCR7 on primary uveal melanoma cell lines(18). The retention of uveal melanoma cells in the liver cannot only be linked to a gradient of chemokines toward the liver, but could also be related to the loss of chemokine receptors once uveal melanoma cells reach the liver.
Another example is c-Met, a receptor for hepatic growth factor (HGF). Uveal melanoma cells expressing c-Met interact with the growth factor produced in the liver(16). Cells of primary uveal melanomas that metastasized had high levels of c-Met receptor expression compared to primary uveal tumors that did not metastasize. The expression of the c-Met receptor in primary tumors of uveal melanoma represents an increased risk for subsequent liver metastases(16).
Growth factors rich in the liver. Insulin-like growth factor 1 (IGF-1) plays a major role in tumor transformation, maintaining the malignant phenotype, promoting tumor growth and preventing apoptosis. It is produced mainly in the liver(14). Elevated insulin-like factor 1 (IGF-1R) receptor expressions were detected in uveal melanoma liver metastasis specimens(19). The association between IGR-1R expression in the tumor and the progression of uveal melanoma has also been reported(20). In addition, human growth factor (HGF) could facilitate hepatic growth of uveal melanoma cells expressing the c-Met receptor(14).
Chromosomal and genetic abnormalities. Uveal melanoma has unique genetic profile abnormalities compared to cutaneous melanoma(14). The BRCA-1 mutation associated with protein 1 (BAP-1) located at the level of chromosome 3p21 has been frequently identified in metastatic uveal melanoma(21-23). Particularly, it has been reported that the BAP-1 mutation present in uveal melanoma cells can cause liver tropism(24). However, this explanation is too simplistic for liver tropism, as a certain fraction of metastatic uveal melanoma that has retained BAP-1 expression and monosomy 3 is not always present in hepatic metastases(14). Polysomy 8q is rather a common feature of metastatic uveal melanoma and the role of this chromosomal abnormality in liver metastases should be investigated in the future(25,26).
Expression of adhesion molecules in hepatic sinusoids. Vascular cell adhesion molecule 1 (VCAM-1) is expressed on hepatic endothelial sinusoidal cells and can retain tumor cells at this level due to slow blood flow(15,27). VCAM-1 is expressed in endothelial cells under inflammatory conditions and mediates the flow and adhesion of various subset of leukocytes, as well as tumor cells for their recruitment and retention of these cells from the blood stream(14). In animal models, partial hepatectomy induced the expression of inflammatory cytokines such as tumor necrosis factor a (TNF-a), interleukin (IL)-1 b and IL-6, but also the expression of VCAM-1 and facilitated liver metastasis(28). The expression of VCAM-1 in endothelial cells showed the adhesion of human malignant melanoma cells that expressed the activation antigen 4 (VLA-4) very late on their surface(29).
Factors of angiogenesis rich in the liver. Interleukin 8 (IL-8) and endothelial vascular growth factor (VEGF) are present in increased amounts in the liver and may promote tumor angiogenesis in the hepatic microenvironment(30). Hepatic stellate cells (HSCs) in the tumor stroma predominantly produce interleukin 8, and neutralizing interleukin 8 with antibodies dramatically reduces the effects of angiogenesis(30). IL-8 also induces the expression of VEGFR2 and VEGF on endothelial cells by activating NF-kB (nuclear factor kappa light chain enhancer of activated B cells), a complex protein that controls the DNA transcription, cytokine production and cell survival and mediates the autocrine and paracrine stimulation of vascular endothelium(14).
Hepatic immunomodulatory microenvironment. The liver is considered to be an immunomodulatory organ and this complex immunological microenvironment could promote tumor metastasis and growth in the liver(31,32).
4. Modern therapies
Although uveal melanoma and cutaneous melanoma both occur in melanocytes and have similar risk factors(33), these subtypes of melanomas differ substantially biologically and clinically(34). Cutaneous melanoma has an extremely high mutational burden with up to 100 mutations per megabase(35,36), whereas in the case of uveal melanoma the number of mutations is extremely low and the typical cutaneous melanoma driver mutations are usually not present in uveal melanoma, and vice versa(37).
Several therapeutic approaches for primary uveal melanoma are appropriate to gain sufficient control over local tumor growth and in most cases even preserve the vision of the affected eye(38,39). These therapeutic options for primary uveal melanoma include:
External beam radiotherapy
Photon-based radiation or
Surgical approaches, such as local resection or enucleation of the affected eye(39-41).
However, up to 50% of patients develop distant metastases(42) and, unlike cutaneous melanoma, uveal melanoma primarily metastasizes to the liver(43). The reason for this selective behavior in uveal melanoma has not yet been fully elucidated(33). Metastatic uveal melanoma has a poor prognosis. No significant benefits have been reported with systemic therapy and no standard treatment has been established.
Immunotherapy is an option for metastatic uveal melanoma, especially due to its low toxicity. In the metastatic stage, the treatment of uveal melanoma was largely adopted from the treatment of cutaneous melanoma(33).
Immune checkpoints inhibitors (ICIs) are some of the most successful therapeutic agents used in the immunotherapy of cutaneous melanoma(44). They are monoclonal antibodies that inhibit the interactions between immune checkpoints proteins, such as programmed death-1 (PD-1)/PD-ligand 1 and cytotoxic T-lymphocyte antigen 4 (CTLA-4)/B7-1/B7-2, allowing T cells to recognize and attack cancer cells(45).
The response rates among patients with metastasized cutaneous melanoma are in the range of 40-50% for anti-programmed cell death protein 1 (PD-1) monotherapy (nivolumab or pembrolizumab)(44) and up to 60% for combined immunotherapy with anti-cytotoxic T-lymphocyte-associated protein 4 (CLTA-4) (ipilimumab) and anti-PD-1 (nivolumab)(46,47), thus making metastasized cutaneous melanoma a potentially curative disease(44).
Zimmer et al. conducted a phase II clinical study in 53 patients with metastatic uveal melanoma who received ipilimumab therapy(44). Although most patients (85%) received previous systemic treatment, only 38% had elevated lactate dehydrogenase (LDH)(44). The best result was a stabilization of the disease (SD) with a disease control rate (DCR) of 47%, but progression-free survival (PFS) was only 2.8 months and the overall survival rate (OS) was 6.8 months(48).
Adverse effects have been reported in 66% of patients, the most common being gastrointestinal disorders (diarrhea and colitis), skin-related toxicities (pruritus and rash) and hepatitis (increased transaminases)(44). Grade 3 or higher adverse effects – mainly diarrhea and colitis – were observed in approximately 36% of cases(44). When the patients were treated according to therapeutic protocols, these side effects were usually reversible(44).
In a retrospective study of 96 patients, nivolumab monotherapy, pembrolizumab and combined immunotherapy (ipilimumab + nivolumab) were administered to 32, 54 and 15 patients, respectively(44). Most patients participating in the study had not received previous systemic treatment and were treated with PD-1 inhibitor monotherapy as first-line therapy. About 50% of patients had an elevated LDH(44). Patients who received combined immunotherapy were much younger than those who received PD-1 monotherapy(44).
In the case of anti-PD-1 monotherapy (nivolumab or pembrolizumab), an overall response rate (ORR) of approximately 5% and a progression-free survival (PFS) of approximately 3 months were observed(49). For combination immunotherapy (ipilimumab and nivolumab), the overall response rate (ORR) was approximately 17%, with a similar progression-free survival of 3 months(49).
The disease control rate (DCR) was approximately 20% for anti-PD-1 monotherapy and 33% for combined immunotherapy(44). Two cases of partial response (PR) occurred under combined immunotherapy and under nivolumab therapy, and one case of partial response (PR) under pembrolizumab therapy(44). Three patients (13%) in the nivolumab group reported adverse effects of grade 3 or higher (colitis, cardiac toxicity, arthralgia and fatigue)(44). One of these three patients died(44). In the combined immunotherapy group, two patients (13%) presented adverse effects of grade 3 or higher (hypophysitis, colitis and thyroiditis)(44).
Another retrospective study of 56 patients treated with PD-1 or PD-L1 antibodies presented similarly disappointing results(44). In this study, elevated lactate dehydrogenase values (LDH) were present in approximately 70% of patients and were an indicator of advanced disease(44). Forty-eight patients (85%) had received previous systemic treatment according to the therapeutic protocols for stage IV disease. The overall response rate (ORR) was 3.6% with two partial responses (PR), and the progression-free survival (PFS) was 2.6 months(50). Moreover, seven patients (12.5%) reported grade 3 adverse effects (nausea, vomiting, arthralgia, colitis, hyperbilirubinemia, fatigue and lymphopenia) and one patient discontinued the treatment as a result of arthralgia(44).
Some authors believe that a subset of patients with metastatic uveal melanoma achieved promising results after combined therapy with entinostat and pembrolizumab(51). The combination of entinostat (SNDX-275, MS-275) and pembrolizumab (Keytruda®) yielded long-term responses in a subgroup of patients with metastatic uveal melanoma, according to findings from the PEMDAC phase 2 study(52) (NCT02697630). The authors of the PEMDAC study(52) consider that they demonstrated for the first time that immunotherapy combined with epigenetic therapy could produce tumor regression in a subgroup of patients with metastatic uveal melanoma. They considered that the low tumor burden (ctDNA and LDH), the presence of tumor suppressor gene BAP1 and an outlier tumor mutational burden in an iris melanoma are closely related to intrinsic tumor genetics and the tumor microenvironment(52).
Patients who received the combined treatment entinostat and pembrolizumab had an overall response rate (ORR) of 14%, including four confirmed partial responses. Investigators reported a clinical benefit rate of 28% at 18 weeks(51).
A total of 66% of patients developed adverse effects of grade 3 or higher which commonly included elevated blood alkaline phosphatase levels, neutropenia, elevated aspartate/alanine aminotransferases and rash(51). Immune-related adverse effects occurred in 86% of patients and 28% developed grade 3 or higher immune-related adverse effects(51). Grade 3 immune-related adverse effects included hepatitis, colitis and stomatitis, skin toxicity, but also a case of grade 4 hypophisis(51).
Immunotherapy represents one of the most effective methods for the treatment of many types of cancer. Although cutaneous melanoma originates in similar melanocytes and the therapeutic results with immunotherapy offer remarkable results, the therapeutic algorithm in cutaneous melanoma cannot be applied in the case of uveal melanoma, the results of immunotherapy in this case being considered a failure. Some authors had attributed this low response rate in uveal melanoma immunotherapy to the low tumor burden.
Gold et al.(53) believe that genetic analysis can help to personalize treatment and could lead to better therapeutic outcomes in uveal melanoma; they are followers of personalized medicine or precision medicine, a relatively new term used to describe therapies targeted to the patient’s needs based on genetic phenotypic or psychosocial characteristics that differentiate one patient from another one with similar clinical presentations(54). Genetic testing can help today to determine the aggressiveness of the tumor and the likelihood of its metastasis prior to a local treatment(55). The discovery of two classes of molecular genetic signatures for uveal melanoma was first reported in 2004 by Onken et al., using gene expression profiling of messenger RNA(56).
The results of gene expression profile (GEP) analysis can be used to classify uveal melanoma as having a low risk of metastasis (class 1a or 1b) or an increased risk of metastasis (class 2)(56). Gene expression profile (GEP) has been shown to exceed clinical and histological prognostic methods in its ability to predict uveal melanoma metastasis and mortality(57). Although large tumors, epitheloid cell histopathological profile and older patient age are significantly associated with class 2 tumors, no definitive clinical marker has been found to differentiate between class 1 and class 2 tumors(58).
Although Gold and colleagues(53) presented preliminary results, they suggest that the analysis of gene expression profile can play an important role in the treatment of uveal melanoma. In patients with uveal melanoma, gene expression profile analysis can be used as a new form of personalized medicine that appears to neutralize low-risk tumors, to quickly identify high-risk tumors and subsequently to limit the number of patients who would likely experience radiation-associated side effects in the future, thus limiting long-term intraocular radiation management(53).
Conflict of interests: The authors declare no conflict of interests.