6.1. Liver-Directed Approach
Approximately 30%–50% of UM patients successfully treated for primary tumour develop metastases. The liver is the most common site of metastases, whereas approximately 50% of patients have isolated liver metastases [1
]. An analysis of patients enrolled in the COMS study found that 93% of patients had liver metastases at time of death [106
]. Of those who had only one site of metastasis, the liver was involved in 95% of cases and the median overall survival (OS) for patients with liver metastases was six to 12 months [1
Several different systemic and regional treatment strategies have been explored, but survival rates have not improved [107
]. Nowadays, a surgical approach is considered the gold standard for achieving a curative resection (R0 resection) for UM metastases, whereas chemotherapy is used for unresectable tumours [101
]. Unfortunately, the number of patients with resectable liver metastasis is limited (<10%). Most liver metastases are multiple with both hepatic lobes involved [108
]. Systemic chemotherapy, both single substance (Dacarbazine or Temozolomide [109
]) or combination treatments (BOLD regimen [110
], gemcitabine with treosulfan [111
]) have shown response rates of less than 10%. In order to reduce systemic toxicity and side effects, alternative liver approaches have been developed. Both transarterial chemoembolization (TACE) with cisplatin and carboplatin and selective internal radiation therapy (SIRT) have shown a partial response (57% of patients with a median survival of nine months versus 62% of patients with a median survival of seven months, respectively) [112
]. In a retrospective analysis on 3873 patients with uveal melanoma, Mariani et al. [114
] reported on a median overall postoperative survival of 14 months for 255 patients who had undergone surgical resection. Notably, the median overall survival rate increased to 27 months in the selective cohort of 76 patients where a microscopically complete resection (R0) was obtained [114
]. Isolated hepatic perfusion (IHP), whether open or percutaneous, has emerged as a viable treatment option [115
]. These organ-specific approaches allow the liver to be perfused with high doses of cytotoxic chemotherapy, minimizing systemic exposure. Compared to systemic chemotherapy, hepatic intra-arterial chemotherapy (HIA) has shown a slightly higher overall response rate, but no increase in overall survival [116
]. In patients with refractory uveal melanoma liver metastases, IHP, performed with alkylating agent melphalan or tumour necrosis factor α (TNFα), has been reported to achieve hepatic response rates of greater than 50%. The progression-free survival was 12 months, thus exceeding outcomes obtained with systemic chemotherapy [117
]. The EORTC 18021 phase III trial (NCT00110123) [116
] stated that despite improved anti-tumour efficacy enhancing rate response (RR) and improving progression free survival (PFS), IHP did not increase the OS of UM patients with liver metastases only. Leyvraz and co-workers compared IHP with fotemustine to intravenous systemic chemotherapy in treating liver metastasis from UM, demonstrating that IHP is more effective than systemic chemotherapy in controlling liver metastases. A median PFS of 4.5 months and response rate of 10.5% with IHP versus 3.7 months and 2.4% was obtained, respectively, with intravenous treatment [116
A phase II follow-up trial confirmed a significant potential survival benefit of 14 months for patients treated with IHP compared to the longest surviving patients in Sweden during the same time period [118
A randomised controlled multi-centre study (SCANDIUM) is currently recruiting patients with isolated liver metastases from uveal melanoma to evaluate if IHP may increase overall survival compared with best alternative care (BAC) (NCT01785316).
Percutaneous hepatic perfusion (PHP) is an alternative endovascular procedure with vascular isolation and perfusion of the liver. When compared to best alternative care, PHP has shown an improvement in overall rate response (ORR) but not in OS [115
]. During PHP, chemotherapy is infused via a percutaneous catheter in the hepatic artery. The liver venous outflow is filtered through an extracorporeal filtration system using a double-balloon catheter positioned percutaneously in the retrohepatic inferior vena cava. Then, a catheter in the internal jugular vein re-transfuses the blood [117
]. In the FOCUS study (NCT02678572) an ongoing multi-centre clinical trial is randomising patients with hepatic dominant metastatic ocular melanoma to either PHP with melphalan or one of the following alternative care options: transarterial chemoembolization, dacarbazine, ipilimumab or pembrolizumab. The estimated primary completion date is in 2020.
Recent investigations have focused on embolisation techniques, especially in association with systemic immune checkpoint inhibitors. It has been postulated that these procedures may lead to an increased release of tumour antigens to the immune system, and the concurrent use of immune checkpoint inhibitors may be helpful in controlling the disease [101
]. Embolisation approaches include hepatic arterial chemoembolisation with different of agents (fotemustine, BCNU, cisplatin), bland embolisation, and radioembolisation with yttrium-90 (90Y)-labeled biocompatible microspheres injected into the hepatic artery. In 2011, Gonsalves et al. [120
] show results of 32 patients UM and liver metastasis treated with radioembolization therapy, proposed as salvage therapy after failure of TACE or immunembolisation. Based on pre-treatment tumour burden, authors divided patients into three categories: <25% (n
= 25), 25% to 50% (n
= 5), and >50% (n
= 2). At the end of the clinical follow-up, the overall survival for the entire cohort of patients ranged from 1.0 to 29.0 months (with a median of 10.0 months). A significantly longer median overall survival was found in the group with a pre-treatment tumour burden <25% (10.5 months; range 3.1–29.0 months) than in the other two groups (3.9 months; range 1.0–12.1 months). In term of treatment response, a complete response was observed in one patient. One patient had a partial response and 56.3% of patients (n
= 18) had stable disease. In comparison to patients with progressive disease, those with stable disease or better, had a significantly longer overall survival (4.9 versus 14.7 months, respectively).
In a similar study with a smaller cohort of 13 patients, Klingenstein et al. showed benefit for patients with liver metastases treated with radioembolisation. In 77% of patients, there was partial remission or a stabilising of the disease was observed [112
]. A clinical trial studying a combination of Y-labelled microspheres with sorafenib (NCT01893099) was concluded, but results have not yet been published. Immunoembolisation using granulocyte-macrophage colony-stimulating factor (GM-CSF), represents another approach, as does a randomised phase II study comparing immunembolisation (IE) and bland embolisation (BE). Statistically significant survival advantage was reported in the immunembolisation cohort of patients with at least 20% liver involvement. In contrast to the overall survival of 17.2 months obtained with BE, the IE group showed 21.5 months of OS [121
6.3. Targeted Therapy
Improved understanding of cutaneous and uveal melanoma mutations, particularly those involving the mitogen-activated protein kinase pathway have led to the development of targeted therapies. Targeted blockades of regulatory growth signalling pathways such as agents against components of the mitogen-activated protein kinase (MAPK) pathway, BRAF and NRAS have recently been introduced in the armamentarium for advanced cutaneous melanoma management. These agents significantly extended the survival of patients with cutaneous melanoma. Unfortunately, this success has not extended to UM [127
]. Since primary uveal melanoma tumours and liver metastases, characterised by mutations in GNAQ or GNA11, have constitutive activation MAPK signalling via alternative, similar therapies targeting downstream effectors of the MEK pathway, Akt and protein kinase C (PKC) were recently investigated. Preclinical models have shown that UM cell lines are susceptible to the inhibition of the MAPK pathway by MEK inhibitors [131
Selumetinib (AZD6244) is an oral, potent, highly selective non-ATP competitive inhibitor of MEK1/2 [132
]. A randomised, phase II study in 101 patients with advanced uveal melanoma compared selumetinib versus chemotherapy (temozolomide or dacarbazine). Selumetinib resulted in a modest improved progression-free survival PFS (15.9 versus 7 weeks, p
< 0.001) and response rate; however, no significant improvement in overall survival was reported (11.8 versus 9 months, p
= 0.09) [132
]. Subsequently, the SUMIT study (NCT01974752), a double-blind, placebo-controlled phase III trial was conducted, in which 129 patients with metastatic uveal melanoma with no prior systemic therapy were randomly assigned to selumetinib plus dacarbazine versus dacarbazine alone. Unfortunately, the primary endpoint, the median PFS, was not significantly improved in the selumetinib arm compared to dacarbazine alone (2.8 versus 1.8 months, p
= 0.32). The objective response rate was 3% with the selumetinib cohort and 0% with dacarbazine cohort with no significant difference (p
= 0.36) [133
To further investigate the role and optimise the efficacy of the MEK inhibitor in UM, while minimizing toxicity and the effect of drug-related feedback reactivation, other studies are underway [134
]. Selumetinib, with an intermittent dosing schedule is being tested in treatment-naïve metastatic uveal melanoma, as monotherapy in a phase Ib trial (NCT02768766). The randomised phase II SelPac trial investigated the activity of selumetinib in three treatment arms: continuous selumetinib, continuous and intermittent selumetinib in combination paclitaxel, based on evidence that the combination increases induction of apoptosis in preclinical models [135
Mutations in GNAQ and GNA11 activate both MEK and Akt, thus an alternative approach might be offered by a simultaneous inhibition of these two pathways. Trametinib, another MEK inhibitor, was evaluated in a phase II trial, both as monotherapy and in combination with an Akt/PKB (protein kinase B) inhibitor, GSK2141785, in 40 patients with metastatic uveal melanoma. Unfortunately, no benefit to PFS was detected with the addition of Akt inhibition [137
Constitutive protein kinase C (PKC) and phospholipase C β (PLCβ) activation has also been implicated in stimulations of the MAPK pathway in uveal melanoma. Investigation into agents targeting these components are also underway. A phase I study (NCT02601378) is recruiting patients to test the PKC inhibitor oral agent LXS196. Preliminary data shown during last Society of Melanoma Research (SRM) Congress [138
] suggest encouraging clinical activity of LXS196 as monotherapy in patients with metastatic UM. In 50 evaluated patients, two had partial responses (PR), and 33 patients showned stable disease. In another phase I study on PKC inhibitor AEB071 (Sotrastaurin), partial responses were observed in only four of 153 patients (3%). In total, 76 patients (50%) had reached stable disease conditions, while the median PFS was 3.5 months [139
]. Preclinical studies have documented the synergetic effect of PKC inhibitor AEB071 with PI3K inhibitor BYL719 in GNAQ- and GNA11-mutant cell lines, offering an alternative approach to target the signalling pathways [139
]. The combination of AEB071 and BYL719 is being evaluating in a phase Ib trial in patients with metastatic uveal melanoma (NCT02273219).
Furthermore, as a consequence of GNAQ/GNA11 mutation, an upregulation of MET has been implicated in uveal melanoma. MET and VEGF signalling are implicated in angiogenesis, invasion and metastasis.
Cabozantinib is an orally bioavailable inhibitor of tyrosine kinases including, c-MET, VEGF and Axl that has been recently approved by the Food and Drug Administration (FDA) for treatment of patients with advanced renal cell carcinoma or medullary thyroid cancer. In preclinical studies, it has shown anti-metastatic activity in a xenograft mouse model of the metastatic uveal melanoma [139
]. A phase II randomised discontinuation trial assessed cabozantinib (XL184), in a cohort of 23 patients with metastatic melanoma and reported that the median OS was 12.6 months and the median PFS was 4.8 months, thereby suggesting potential for clinical activity [139
]. Based on these promising results, a randomised phase II study compared the MET inhibitor Cabozantinib to Temozolomide or Dacarbazine in patients with metastatic uveal melanoma (NCT01835145), but results are yet to be published. Recently, it has been demonstrated the role of cMET in the primary resistance to MEK inhibitors in metastatic UM with GNAQ/11 mutation. Even if the microenvironment mechanism regulating these responses is poorly understood, the combined inhibition of cMET and MEK may improve the responses to targeted therapies of metastatic UM. Although this preliminary research is promising, further studies are needed [140
Sorafenib is another oral multikinase inhibitor with a potent activity against serine/threonine kinase isoforms, including RAF kinase, vascular endothelial growth factor receptor (VEGFR) and the platelet-derived growth factor receptor (PDGFR). A single arm phase II trial evaluated sorafenib as a single agent in patients with MUM and revealed stable disease at 24 weeks in 31.2% of patients [141
]. Previously, a phase II study had evaluated the efficacy of sorafenib in combination with carboplatin and paclitaxel (CP) in metastatic uveal melanoma. Although minor tumour responses and stable disease were observed (tumour regression <30% and the four-month PFS) no significant efficacy occurred (ORR = 0% [95% CI: 0–14%]) [142
Similar results indicating some level of anti-tumour activity were demonstrated in a randomised discontinuation, blinded, placebo-controlled phase II study (STREAM) (NCT01377025) [142
]. A cohort of 149 patients with chemotherapy-naive metastatic UM were enrolled. Seventy-eight patients, that showed stable disease conditions, were then randomised to continuation or discontinuation of sorafenib. Significant improvement in PFS compared to placebo (5.5 versus 1.9 months, p
= 0.0079) was associated with the continuation regimen. Notably, two patients (1.3%) achieved a partial response, suggesting some degree of clinical activity despite the overall low ORR. Although sorafenib seems to promote the stabilisation of the disease, it failed to improve OS of metastatic UM patients [142
Interesting results were recently reported by Valsecchi et al. [121
]. In this retrospective study, the use of low-dose sunitinib, a non-selective c-Kit inhibitor, in the adjuvant setting was associated with better overall survival. Investigators are currently validating the results of this study in a randomised, non-comparative (sunitinib and valproic acid) phase II clinical trial (NCT02068586) in patients with high-risk of uveal melanoma. Previously, the randomised multicentre phase II (SUAVE) trial failed to show any survival benefit with sunitinib compared to dacarbazine in 74 patients with untreated metastatic uveal melanoma [92
6.5. Anti-PD-1 Therapy
Programmed cell death-1 (PD-1), a member of the CD28 superfamily, delivers negative signals upon interaction with its two ligands, PD-L1 or PD-L2. PD-1 exerts a wider range of immune-regulatory roles in T cells activation and tolerance, inhibiting their activity and proliferation in peripheral tissues. Agents targeting the PD-1/PD-L1 signalling have shown promising response in the treatments of many types of solid tumours both in clinical trials and settings. Nowadays, two antibodies against PD-1, nivolumab and pembrolizumab have been recently approved by FDA and/or European Medicines Agency (EMA) for treatment of advanced cutaneous melanoma. These molecules have demonstrated greater outcomes and an improved toxicity profile compared to ipilimumab [157
]. Although the efficacy of these molecules has only been evaluated in retrospective case series on patients with advanced uveal melanoma, both activity and favourable toxicity profiles have been reported.
Kottschade et al. [160
] investigated a cohort of 10 patients with metastatic UM treated with pembrolizumab, given at a dose of 2 mg/kg every three weeks for up to two years, through the expanded access program (EAP). One complete response, one stable disease condition (10%), and two partial responses (20%) were reported with an ORR of 30%. No OS data were reported and PFS was only 18 weeks (range 3.14–49.3).
Encouraging results were documented by Karydis and co-workers [161
] in another retrospective case series. Of the 25 metastatic uveal melanoma patients enrolled and treated with pembrolizumab, two patients showed a partial response and six patients stable disease conditions. Similar results were reported in a large series of 58 patients with advance uveal melanoma who received PD-1 antibodies (pembrolizumab or nivolumab) or PD-L1 antibodies (atezolizumab) [162
]. In this cohort, objective tumour responses were observed in two patients for an overall response rate of 3.6% (95% CI 1.8–22.5%), and stable disease conditions in five (9%). Median PFS and OS were 2.6 months (95% CI 2.4–2.8 months) and 7.6 months (95% CI 0.7–14.6 months), respectively.
The first efficacy and updated safety data from CheckMate 172 trials were encouraging. This phase II trial (NCT02156804) is the biggest one that has evaluated efficacy and safety for nivolumab monotherapy (3 mg/kg) in 734 patients with different melanoma subtypes, including UM, after anti-CTLA-4 therapy. At the end of a follow-up of ≥one year, out of 75 patients in the UM subgroup, 15 patients had disease stabilisation (44%) and two showed PR. The OS rate at one year was 47% and the median OS was 11 months.
The results of these retrospective series underline that durable, objective tumour responses and sustained disease control occurred, but were rare. Thus, nowadays PD-1 and PD-L1 antibodies are beneficial only in a small subset of metastatic uveal melanoma. As Javed et al. postulated [163
], the lower PD-L1 expression in metastatic UM, compared to cutaneous melanoma, may suggest a rationale for failure of PD-1 and PD-L1 inhibitor agents, and an alternative mechanisms of immune evasion may occur. Further prospective studies are needed to confirm the low response rate.
6.7. Innovative Immune-Based Approaches
Despite the limited clinical activity obtained with immune checkpoint inhibition, other novel immune-based approaches showing promising results in treating this disease have recently been reported. Depending on these results, future scenarios could dramatically change. Thus far, different clinical trials have tested infusion autologous-modified dendritic cells. In a prospective study, 14 patients with metastatic uveal melanoma were vaccinated with autologous dendritic cells loaded with tumour peptides of gp100 and tyrosinase. Four of the vaccinated patients showed long overall survival, and the median OS was 19.2 months [166
]. Currently, a phase III trial (NCT01983748) is recruiting patients suffering from uveal melanoma typed positive for monosomy 3 and without evidence for metastases who will be vaccinated over a two-year period with dendritic cells loaded with autologous tumour RNA. The open-label phase III study plans to enrol 200 participants, with an estimated primary completion date of July 2022.
IMCgp100 is an investigational drug that can refocus a T cell against the gp100 protein in uveal melanoma cells. IMCgp100 is a bispecific biologic with two functional ends, one targeting the soluble affinity enhanced T cell receptor (TCR) and the other an anti-CD3 single-chain variable fragment (scFv). The glycoprotein 100 (gp100) is presented in the context of HLA-A02, thus eligible patients must be HLA-A*02 positive. Two phase I trials have evaluated safety, pharmacokinetics, pharmacodynamics and efficacy for IMCgp100, which was administered weekly in HLA-A2 patients. The first-in-human (FIH) study enrolled 84 patients total with advanced melanoma including a cohort with UM (n
= 16) (NCT01211262) [168
]. The second study enrolled 19 patients with advanced UM using an intra-patient dose escalation (IE) regimen (four dose levels ranging from 54 to 73 mcg), designed to reduce T cell-mediated toxicities that were previously reported (NCT02570308). In both of the tow study durable and objective responses were noted. Rash, pruritus and peripheral edema were the most frequent related AEs observed. Hypotension noted with introductory doses was consistent with chemokine release and movement of lymphocytes into tissues. Notably, immunofluorescence studies reveal an infiltration of PD-1+/CD8+ T cells in tumour specimens with PD-L1 expression, confirming immune response activation within 24 h of the initial treatment dose. Preliminary results of the phase II study using the intra-patient escalation dosing regimen were recently presented at the 2018 Association for Research in Vision and Ophthalmology (ARVO) Annual Meeting [168
]. The 68-mcg dose was identified to be the maximum tolerated dose. The most common all-grade treatment-emergent adverse events (TEAEs) were pruritus (90%), pyrexia (84%), fatigue (84%), hypotension (74%), chills (63%), nausea (68%), dry skin (63%) and peripheral edema (63%). The most frequent grade 3/4 adverse events were fatigue (16%), hypotension (16%), erythema (16%) and macular rash (11%). The phase II portion of this study is now recruiting with a plan to achieve a cohort of 150 patients. The estimated completion date for this expansion trial is April 2019 (NCT02570308). In addition, another pivotal phase II trial is currently enrolling treatment-naïve patients with uveal melanoma to compare IMCgp100 against investigator’s choice of systemic therapy, including dacarbazine or the immunotherapies pembrolizumab or ipilimumab. The estimated primary completion date is July 2020 with the hope of enrolling more than 300 participants (NCT03070392).
Recently, adoptive T-cell therapy has shown favourable responses in multiple refractory solid tumours. Subset of UM tumour infiltrating lymphocytes (TILs) exhibited a robust level of reactivity. Based on these findings, a phase II study (NCT01814046) was conducted to determine if TILs could induce regression of metastatic uveal melanoma [171
]. A total of 24 patients were enrolled and divided into two cohorts. One cohort of 22 patients received lymphodepleting conditioning chemotherapy of cyclophosphamide and fludarabine followed by a single intravenous infusion of autologous TILs and high-dose aldesleukin. The other cohort of two patients received the same regimen with no aldesleukin. Of the 22 evaluated patients, about 36% had objective tumour regression, such as seven partial responses (31.81%) and one complete response (4.54%). Common grade 3/4 adverse effects were related to the lymphodepleting preparative regimen. All patients experienced lymphopenia, neutropenia and thrombocytopenia, and in six patients (27.27%) an infection was documented [171
Glembatumumab vedotin, a new antibody-drug conjugate, is also being investigated in a phase II study (NCT02363283) [173
]. The accrual of 37 patients with metastatic uveal melanoma has recently been completed. Glembatumumab vedotin is a monoclonal antibody-drug conjugate directed at the glycoprotein NBM and linked to the cytotoxic microtubule inhibitor monomethyl auristatin E (MMAE), which combines an antibody directed at the transmembrane glycoprotein NMB (gpNMB) with the microtubule inhibitor monomethyl auristatin (MMAE). A preclinical study has demonstrated a high frequency of gpNMB expression in uveal melanoma and modest clinical activity has been demonstrated in patients with cutaneous melanoma. Thus, the rationale to explore antitumour activity in patients with uveal melanoma [174
6.8. Epigenetic Approaches
In the past decade, an important role has been attributed to epigenetic mechanisms in the development of many pathologies, and new treatments interfering with epigenetic regulation have begun to appear. As mentioned above, alterations in tumour suppressor genes or oncogenes can be generated by mutations or by transcriptional regulation by epigenetic mechanisms. These include DNA methylation or demethylation and/or histone acetylation or deacetylation. Notably, recent studies have found that changes in microRNAs and long ncRNAs play a role in the development and metastasis of UM [176
]. A relatively new class of anti-cancer agents has been developed: histone deacetylase (HDAC) inhibitors. These agents have been demonstrated to induce histone and protein acetylation, and to alter gene expression, inducing death, apoptosis and cell cycle arrest in cancer cells [178
]. Vorinostat is an orally bioavailable inhibitor of class I and II HDACs. Vorinostat is FDA approved for the management of cutaneous T cell lymphoma. It is being investigated both as mono-therapy and combination therapy for other types of cancers, including metastatic UM. A phase 2 study of Vorinostat (NCT01587352) in metastatic UM has terminated its enrolment, but results are yet to be published. Vorinostat is currently being evaluated in adjuvant and metastatic settings, respectively (NCT02068586, NCT01587352).
The cancer-testis antigen, a preferentially expressed antigen in melanoma (PRAME) is an independent prognostic biomarker in UM, which identifies increased metastatic risk in patients with Class 1 or disomy 3 tumour [75
]. It has been considered another therapeutic target based on its lack of expression in normal cells [76
]. Furthermore, in an experimental study using a retrospective cohort of 64 patients, Gezgin et al. [182
] stated that 69% of metastatic uveal melanoma tumours analysed expressed PRAME. Additionally, the authors observed a concomitant expression of PRAME and HLA class I in metastases, suggesting a potential role for treatment with PRAME-directed immunotherapy. A Phase I/II study is evaluating the safety and activity of rimiducid, a genetically-modified autologous T cell product incorporating an HLA-A2-restricted PRAME-directed TCR, in patients with relapsed acute myeloid leukaemia, previously treated for myelodysplastic syndrome, and metastatic uveal melanoma (NCT02743611).