ICIs and targeted therapy have improved survival for patients with newly diagnosed metastatic melanoma. Recently, a new ICI combination, nivolumab (nivo) and relatlimab (rela), was approved for patients with metastatic melanoma after demonstrating efficacy in untreated patients. The numerous treatment options available for patients with metastatic disease have complicated the decisions faced by physicians and patients. In addition, more and more patients are receiving therapy earlier in the course of their disease, in the neoadjuvant or adjuvant setting, so that they have potentially resistant disease when metastasis is found.

Early studies on the use of ICIs in the treatment of metastatic melanoma showed significant improvements in progression-free survival (PFS) and OS over previously available therapies. PD-1 inhibitors, such as pembrolizumab (pembro) and nivo,^1,2 and cytotoxic T-lymphocyte antigen 4 (CTLA-4) inhibitors, such as ipilimumab (ipi), function by preventing inactivation of T-lymphocyte activity by cancer cells, leading to improved immune system activation and targeting of neoplastic cells. Ipi was the first agent to be associated with an improvement in survival in metastatic melanoma with a response rate of 11% (95% CI, 6.3 to 17.4). Nivo and pembro were demonstrated to have even greater benefit with response rates in the 30%-40% range, and the greatest initial response rate of 58% (95% CI, 45.9 to 68.5) is observed when PD-1 inhibition is combined with CTLA-4 inhibition.^3,4 These results were found to be consistent across different patient cohorts, but some factors including tumor PD-L1 status and BRAF mutation status have trends that are sometimes used to help make initial treatment selection.^3,4

Recent development of the LAG-3 inhibitor, rela, expanded options for the treatment of metastatic melanoma. LAG-3 is a cell surface molecule that functions as a negative regulator of T-cell proliferation and function, and rela is a first-in-class lymphocyte activation gene 3 protein (LAG-3) blocking antibody that prevents T-cell inactivation by tumor cells. The 2022 study RELATIVITY-047 showed that combination of rela/nivo had twice the median PFS and a 25% reduction in risk of disease progression (PD) or death compared with nivo alone (hazard ratio [HR], 0.75; P = .006). The benefits observed with nivo/rela were similar to the benefits observed in combination PD-1/CTLA-4 therapy although the response rates on both arms were lower at 43% and 32%, respectively.⁵

The selection of frontline therapy for patients with metastatic disease is often based on patient characteristics and preferences. As a result, the autoimmune toxicities associated with different regimens play a huge role. While the rate of grade III/IV treatment-related AEs associated with single-agent PD-1 inhibition is roughly 10%-14%, the rate of grade III/IV treatment-related AEs associated with combination of ipi/nivo is roughly 55% and must be considered when selecting patients to receive this regimen. The rate of grade III-IV AEs observed with nivo/rela is somewhere between single-agent PD-1 inhibition and combination therapy at 18.9%.

Modifications in the dosing of combination of ipi/nivo regimens have shown some effectiveness in reducing the incidence of high-grade AEs associated with ICI treatment. The CheckMate 511 study found that using flipped-dose ipi/nivo (nivo 3 mg/kg plus ipi 1 mg/kg every 3 weeks for four doses followed by nivo maintenance) lowered the incidence of grade III-V AEs by 14% (P = .006) with overlapping OS and PFS curves.⁶ However, the data with this regimen are much less mature than those with regular dose ipi/nivo, leading some clinicians to be concerned that responses might not be as durable.

Approximately 40%-50% of melanomas are found to have a mutation in the BRAF gene at the V600 location, and there have been numerous studies focused on the use of targeted BRAF and MEK inhibition in treatment of melanomas.^7-10 BRAF and MEK are both genes functioning in the mitogen-activated protein kinase (MAPK) signal transduction pathway, leading to increased cell division and proliferation, and, when mutated, become overactive with resultant tumor growth. The BRAF inhibitors vemurafenib (vem) and dabrafenib (dab) have efficacy as monotherapies in metastatic melanoma with BRAF V600E or V600K mutations. However, acquired resistance to BRAF inhibitors often develops due to aberrant reactivation of the MAPK pathway. Studies have shown that combining dab with the MEK inhibitor trametinib (tram) leads to improvements in treatment response in this population of patients, when compared with dab alone, because of delayed emergence of resistance.⁷ These benefits have been observed even in long-term treatment, as evidenced in a 5-year landmark analysis showing long-term survival, particularly in patients with favorable baseline prognostic features.⁸ Other combinations of BRAF/MEK inhibitors have also demonstrated benefit in the treatment of melanoma, which include vemurafenib with cobimetinib and encorafenib with binimetinib.^9,10 However, the combination of BRAF/MEK regimens is not without side effects and treatment with them carries a rate of grade III-IV AEs of 55% although side effects often respond quickly to treatment holds or reduction.

The question of which therapy, ICIs or targeted therapy, to start in patients with advanced BRAF V600 E/K–mutant melanoma was recently answered by two trials: the DREAMSeq and SECOMBIT trials. The DREAMSeq trial investigated optimal sequencing of treatment for patients with BRAF V600E/K–mutant melanoma by comparing ipi/nivo followed by dab/tram at progression against dab/tram followed by ipi/nivo at progression.¹¹ The study found that starting with ipi/nivo was associated with higher 2-year OS and durability of disease response versus the inverse sequence of dab/tram followed by ipi/nivo in all clinical subgroups examined. Recent tumor biology studies have suggested that resistance to BRAF/MEK inhibitor therapy results in a locally immunosuppressive tumor environment, preventing effective antigen presentation and immune system activation, and that initial treatment with immunotherapy may actually enhance responsiveness to targeted therapies among BRAF-mutated cancers. This study concluded that among patients with metastatic melanoma and BRAF V600E/K–mutated tumors, ipi/nivo should be considered the preferred first-line therapy for the majority of patients and has cemented ICI as the first-line choice for all patients with melanoma regardless of mutation status.¹¹ The SECOMBIT study was a phase II trial that randomly assigned patients to a combination of ipi/nivo, enco/bini, or enco/bini for 8 weeks, followed by ipi/nivo. It found a 3-year OS of 54% (95% CI, 41 to 67) for enco/bini as initial therapy, 62% (95% CI, 48 to 76) for ipi/nivo as initial therapy, and 60% (95% CI, 58 to 72) for a short course of targeted therapy followed by immunotherapy.¹² This trial confirms the benefit of starting patients with BRAF V600 mutations on immunotherapy. It also provides an alternate approach to be used in patients started on targeted therapy who did well with a switch at 8 weeks to immunotherapy.

Trials have been performed looking at the combination of ICI with BRAF/MEK inhibition. One such study, evaluating the efficacy of triplet therapy with dabrafenib/trametinib/pembro, found that this regimen conferred longer PFS and durability of response, although with greater toxicity than dabrafenib/trametinib alone (with a grade III-V AE rate of 58%).¹³ Another trial examined vemurafenib, cobimetinib, and atezolizumab in combination and demonstrated a PFS benefit over targeted therapy alone.¹⁴ However, the clinical application of triplet therapies has been limited by high toxicity rates and lack of comparison with ipi/nivo, nivo/rela, or single-agent ICI treatment. Further studies are ongoing on how best to combine immunotherapy with BRAF/MEK-targeted therapy.

The question that clinicians are faced with when discussing immune checkpoint inhibition with patients who have newly diagnosed metastatic melanoma is which ICI regimen to start. The options include single-agent PD-1 inhibition, combination of ipi/nivo at normal dosing or flipped dosing, and combination of nivo/rela. Each of these regimens have different levels of efficacy and different rates of severe autoimmune toxicities. Factors that are used to make the decision include BRAF mutation, location of metastasis, symptoms associated with melanoma, and the underlying fitness and comorbid medical issues of patients. In addition, many patients have received ICI in the neoadjuvant or adjuvant setting before being diagnosed with metastatic disease, and this influences treatment choice (Table 1).

TABLE 1. Treatments for Metastatic Melanoma

View larger version (198K)

As noted earlier, many patients with metastatic melanoma are diagnosed after treatment with neoadjuvant or adjuvant PD-1 inhibition and thus have disease resistant to PD-1 inhibitor therapy. In patients previously treated with PD-1 inhibition alone, studies have demonstrated that the combination of immunotherapy with ipilimumab and nivolumab is more effective than ipilimumab alone with a rough response rate of 30% (95% CI, 18.4 to 40.6).^15,16 Data for nivolumab and relatlimab in patients who progressed on previous PD-1 inhibition are poor with a response rate of 12% (95% CI, 8.8 to 15.8).¹⁷ Thus, in patients who progressed on adjuvant therapy, the combination of ipi/nivo would be predicted to have the highest response rate.

Trials testing novel agents in combination with PD-1 inhibition in initial treatment of patients with metastatic melanoma have had mixed success. Although Lag-3 inhibition was successful, three recent trials aimed at increasing tumor immune infiltrate all failed to improve treatment over PD-1 inhibition alone. These include an IDO inhibitor (epacadostat), a modified injectable herpes virus (talimogene laherparepvec), and a modified interleukin-2 (IL-2) agonist (bempegaldesleukin). Current frontline clinical trials are testing PD-1 inhibition in combination with HDAC inhibition, novel immunotherapies, vaccines, and many other agents. There are also efforts to test the administration of tumor-infiltrating lymphocytes (TILs) in treatment-naïve patients. TIL therapy involves extraction of infiltrating lymphocytes within tumors, ex vivo outgrowth and expansion of these cells, and adoptive transfer of the cells back to the patient with preparative lymphodepleting chemotherapy and IL-2. Objective responses have been observed among approximately one third of patients, and even in patients with heavily pretreated disease, and it was shown to be superior to single-agent ipilimumab in PD-1–refractory patients.^18-20

Given the severe immune toxicities associated with ICIs, there are several important frontline trials looking at adding drugs to the combination of ipilimumab and nivolumab to reduce side effects. The addition of granulocyte-macrophage colony-stimulating factor to ipilimumab was previously shown to decrease toxicity, and a current study is testing this with the combination of ipi/nivo.²¹ Ongoing studies are showing some promise in the use of tocilizumab, an IL-6 inhibitor, to reduce immune-related adverse effects, after finding increased expression of IL-6 in tumor tissue from patients treated with immune checkpoint inhibitors (ClinicalTrials.gov identifier: NCT03999749).

Anti–PD-1 ICIs, both as a monotherapy and as a combination with either anti-CTLA4 or anti-LAG3 agents, provide robust and durable response rates in a large percentage of patients with advanced melanoma and are now a widely accepted backbone of frontline treatment, regardless of BRAFV600 mutational status.^5,22-24 However, a large percentage of patients exhibit either primary resistance or acquired resistance after anti–PD-1 therapy.²⁵ The management of these patients presents an unmet need.^25,26

Although continuous enrollment on prospective clinical trials remains the encouraged approach, we already have important insights available. The phase II SWOG S1616 trial focused on patients who exhibited a complete lack of response to anti–PD-1 (or PD-L1) monotherapy by treating them with ipi alone or in combination with continuous nivo. A significant PFS benefit was observed with the combination of ipi/nivo compared with ipi monotherapy (HR, 0.63; 90% CI, 0.41 to 0.97; P = .037) with the 6-month PFS of 34% and 13%, respectively.²⁷ This study also noted intriguing enhancement of intratumoral CD8+ cells in the tumor microenvironment (TME) with the combination of ICIs, particularly among those with a response, suggesting that anti-CTLA4 ICI may reverse primary resistance to the PD-1/L1 blockade. An additional prospective effort combining low-dose ipi (1 mg/kg every 3 weeks for four doses) with pembro in patients who progressed on anti–PD-1/L1 monotherapy noted a median PFS and OS of 5 and 24.7 months, respectively, and a duration of response of 16.6 months (95% CI, 7.9 to not reached).¹⁶ These observations support consideration of combination of ipi plus anti–PD-1 in patients exhibiting resistance to frontline anti–PD-1 monotherapy. A reasonable alternative currently used for patients exhibiting primary or secondary resistance to anti–PD-1 ICI (without CNS involvement) includes the combination of nivo/rela. This is based on a promising signal of enhanced objective response rates (ORRs) and disease control rates of 12% and 45%, respectively, in a phase I/II patient cohort with advanced melanoma that progressed on or after previous anti–PD-1.^5,28 However, these data remain relatively immature awaiting long-term follow-up.

Targeted BRAF plus MEK inhibition is an approved therapy for patients with anti–PD-1-refractory BRAFV6000-mutant melanoma. BRAF/MEK inhibitors are commonly used in patients (1) who are considered ineligible for further ICI agents, (2) who require rapid disease control, and/or (3) who progress on combination of ICIs. The latter scenario is founded on early signals, suggesting CTLA-4 resistance in those previously receiving anti-PD1 plus anti-LAG3 regimens, with further validation needed.²⁹ Additional support for subsequent BRAF/MEK inhibition includes a post hoc analysis of the phase III (KEYNOTE-006) trial, where subsequent BRAF/MEK inhibition in patients with advanced melanoma who progressed on pembro achieved an ORR of 30.5% (95% CI, 19.2 to 43.9).³⁰ Given the suboptimal outcomes of targeted agents when compared with ICI regimens in the frontline setting, alternative options are needed in the anti-PD1 refractory setting.^31,32

The underlying biology responsible for ICI resistance in melanoma is complex and remains a field of active investigation.^25,33,34 The TME appears to play a crucial role in the immunologic suppression by (1) lack of T-cell priming, (2) immune tolerance, (3) stromal adaptations, and (4) enhanced exhaustion of antitumoral CD8+ T cells with concurrent upregulation of regulatory T cells.³³ Given the high mutational burden associated with melanoma, the utilization of advanced single-cell spatial profiling of non–T-cell-inflamed and low interferon-γ expressing (cold) tumors is anticipated to provide neoepitope targets to be recognized by T lymphocytes and provides the basis of peptide vaccine therapies.³⁵ This approach has recently led to US Food and Drug Administration (FDA) fast-track approval of a personalized neoepitope vaccine (EVX-01) combined with anti–PD-1 (pembrolizumab) in the phase II study (KEYNOTE-D36).³⁶ Finally, although not FDA-approved at the time of this publication, an exciting non-ICI treatment option for patients with anti–PD-1 refractory melanoma includes adoptive cell therapy with TILs. A phase II trial with an autologous, centrally manufactured TIL (lifileucel) has observed an ORR of 36% (95% CI, 25 to 49), with 41.7% of these lasting for more than 24 moths and a median OS of 17.4 months, in a heavily pretreated patient cohort with PD-1–refractory melanoma.¹⁹ A randomized phase III (M14TIL) trial comparing TIL therapy with ipilimumab monotherapy noted an ORR of 49% (95% CI, 38 to 60) vs 21% (95% CI, 13 to 32) along with a median PFS of 7.2 vs 3.1 months, respectively, in a cohort of patients with mostly PD-1 monotherapy–refractory melanoma.²⁰ Finally, a plethora of alternative approaches for patients with PD-1–refractory melanoma are ongoing.^23,37

Metastatic disease involving CNS portends a poor prognosis. Melanoma brain metastases (MBM) occur at high rate (40%-60%) and account for approximately 54% of melanoma deaths.³⁸ This high-risk subset remains under-represented in trials given exclusion of MBM in major clinical trials to date. However, retrospective and small prospective studies dedicated to this group revealed important observations leading to substantial improvements in response and survival for certain subsets of patients.^38-40

The phase II COMBI-MB trial, with BRAF/MEK combination (dabrafenib plus trametinib) in patients with BRAFV600 mutation, achieved intracranial responses in 58% (95% CI, 46 to 69) of patients with asymptomatic and treatment-naïve disease, a median duration of response of 6.5 (95% CI, 4.9 to 10.3) months, a mPFS of 5.6 (95% CI, 5.3 to 7.4) months, and an OS of 10.8 (95% CI, 8.7 to 19.6) months.⁴¹ Additional studies involving BRAF and MEK inhibitors at higher doses (ClinicalTrials.gov identifier: NCT03911869) or in combination with ICIs (ClinicalTrials.gov identifiers: NCT03625141 and NCT04511013) and novel agents (ClinicalTrials.gov identifier: NCT02159066) for patients with BRAFV600-mutant MBM are ongoing.³⁸

In patients with asymptomatic MBM, a phase II trial with anti-CTLA4 (ipilimumab) achieved intracranial disease control at 12 weeks in 24% with a median OS of 7 months (95% CI, 4.1 to 10.8).⁴² A phase II trial with anti-PD1 (pembrolizumab) noted an ORR of 22% and a median OS of 17 months (95% CI, 10 to not reached) in a similar population.⁴³ Similar response rates (20%; 95% CI, 7 to 41) were observed in the phase II (ABC) trial using anti–PD-1 (nivolumab) for patients with asymptomatic MBM.⁴⁴ Five-year intracranial PFS and OS rates of 46% and 51% for the ipi/nivo ABC trial arm, compared with 15% and 34% for nivolumab monotherapy in the asymptomatic patient cohort, were observed. The combination of ipilimumab plus nivolumab phase II (CheckMate 204) trial observed an ORR of 54% (95% CI, 43.3 to 63.5) and a 36-month intracranial PFS of 54% (OS 72%) for patients with asymptomatic MBM.^45,46 Given the durable intra- and extracranial response and survival rates observed in these two trials, up-front combination of ICIs (ipilimumab plus nivolumab) is currently considered an optimal approach for patients with asymptomatic MBM. This combination is preferred over frontline targeted therapy in patients with asymptomatic MBM and BRAFV600 mutation, supported by expert consensus and prospective and real-world analyses.^31,32,47,48 Prospective trials are ongoing to elucidate the additional benefit of ICI combined with stereotactic radiosurgery (SRS) or BRAF/MEK (ClinicalTrials.gov identifier: NCT04511013).^40,49-53

Patients with symptomatic MBM, requiring >10 mg daily prednisone-equivalent steroids, exhibited substantially worse response and survival rates in all trials where such cohorts were included.^40,46 These likely reflect a larger disease burden, need for larger doses of systemic steroids, and unique TME.³⁸ Responses to ICI are reduced in those with exclusively intracranial disease, suggesting that ICI may augment CD8+ T-cell trafficking via peripheral immune T-cell expansion to achieve a more robust intracranial response.^40,54 Development of relevant biomarkers is an area of ongoing research.^26,55

Given the systemic autoinflammatory pathophysiology associated with irAEs, patients with an underlying autoimmune disease have historically been excluded from major clinical trials and therefore are largely under-represented within the currently available prospective data, supporting ICIs in a variety of cancer subtypes.⁵⁶ A systematic review of patients with cancer (mostly melanoma) and concurrent autoimmune disease noted 41% experienced a flare of their underlying condition, 25% developed de novo irAEs, and 9% experienced both preexisting flares and de novo events upon initiation of ICI, with a signal of anti-CTLA4 ICI more commonly associated with flares of the underlying disease and anti-PD1 ICI triggering new irAEs.⁵⁶ Importantly, there was no observed difference in irAE rates in those with active compared with inactive pre-existing autoimmunity at the time of ICI initiation, and although irAEs were controlled without the need to discontinue ICI in more than half of the patients, the rates of grade 5 irAEs were higher than the general population at 2.4%.⁵⁶ Additional cohort studies observed that 71% of patients with underlying autoimmune diseases developed either worsening of their pre-existing condition or a new irAE on initiation of ICI.⁵⁷ This study also noted a trend in reduced PFS in those with autoimmune flares and irAEs compared with those without such events, mostly in patients requiring systemic immunosuppressants and/or ICI discontinuation.⁵⁷ The high doses of systemic glucocorticoids, especially in the early phases of ICI initiation, may blunt the response to ICI that is historically associated with patients who develop irAE compared with those who do not.^58-60 These observations suggest that the use of systemic glucocorticoids, although oftentimes required in the management of severe irAE, should be considered judiciously in those experiencing low-grade irAE events and/or indolent autoimmune conditions especially in the early phases of ICI initiation.^58,61-64

UM is the most common primary intraocular malignancy in adults, albeit a rare melanoma subtype, characterized by the absence of BRAF, NRAS, or KIT mutation but recurrent, mutually exclusive, and early oncogenic mutations in the Gαq pathway (mostly GNAQ/11)⁶⁵; the inactivation of these Gα proteins results in activation of downstream signaling cascades, including the MAPK, PI3K-AKT, and YAP pathways.⁶⁶

Despite a highly effective management of the ocular tumor, up to 50% of patients develop metastases by hematogenous dissemination, with occult micrometastases before the detection of the primary eye tumor. The liver is the first site of metastasis in more than 90% of patients.⁶⁷ Because of limited efficacy of available regional and systemic therapies, OS is poor and the historical median OS is 1 year after the first metastatic event.^68,69 Until recently, no systemic treatment had demonstrated any survival benefit. In January 2022, tebentafusp became the first systemic therapy to receive regulatory approval in this treatment-resistant cancer.

Tebentafusp is a T-cell–redirecting bispecific fusion protein HLA A*02:01–restricted, using a high-affinity T-cell receptor targeting glycoprotein 100 (gp100), a melanocyte lineage–specific antigen highly expressed on UM cells, and fused to an anti-CD3 single-chain variable fragment. Once bound to the gp100-HLA complex, tebentafusp recruits and activates polyclonal T cells through CD3 ligation to release cytokines and cytolytic mediators.⁷⁰

In the open-label, phase III, IMCgp100-202 trial, 378 HLA A*02:01–positive patients with first-line systemic metastatic UM were randomly assigned (2:1) to receive tebentafusp or the investigator's choice of pembrolizumab (82%), ipilimumab, or dacarbazine. After a median follow-up of 14 months, the improvement in OS was highly significant, with an estimated median OS duration of 21.7 months in the tebentafusp group versus 16 months in the control group and a 1-year OS rate of 73% versus 59% (HR, 0.51; 95% CI, 0.37 to 0.71; P < .001).⁷¹ This unprecedented result is weighted by a modest although significant benefit in PFS (median PFS 3.3 v 2.9 months; 31% v 19% at 6 months; HR, 0.73, 95% CI, 0.58 to 0.94; P = .01) and a low ORR of 9% versus 5%.

Moreover, an OS benefit was observed in patients with best objective response of RECIST 1.1 PD on tebentafusp compared with the control group (HR, 0.43) and who continued tebentafusp beyond initial radiographic progression. According to the protocol, 109 patients received tebentafusp beyond progression and 39 had a clinical benefit (ORR or SD >3 months); the OS benefit was sustainable after adjusting for all covariates (HR, 0.7).⁷² Investigations are underway to understand the decoupling of RECIST-based radiographic assessment and OS benefit: There is a need for new measures of clinical activity with this new class of immuno-oncology therapies and for predictive biomarkers of response or resistance.

Indeed, the underlying practical question is closely related to the optimal way to monitor tebentafusp treatment: (1) in patients with stable disease for uninterrupted months or years of weekly infusions and (2) in patients with RECIST or irRECIST PD in the absence of any disease symptoms or treatment-related AEs. What is the optimal duration of treatment? Which discontinuation rules can be outlined?

Because of the lack of correlation between survival benefit and OR rate and a weak PFS gain in tebentafusp versus the investigator's choice, an update of the 202 trial with longer follow-up is very much awaited.

UM has a very low mutational burden, making analysis of circulating tumor DNA (ctDNA) challenging.⁷³ In a single-arm, phase II trial of tebentafusp in 127 previously treated patients with metastatic UM, 84% with evaluable serum samples had detectable ctDNA at baseline and on treatment, measured with multiplex PCR followed by next-generation sequencing for major UM-specific genes (GNAQ, GNA11, SF3B1, PLCB4, CYSLTR2, EIF1AX). The magnitude of ctDNA reduction after 9 weeks of tebentafusp correlated with improvement in OS using a logarithmic scale (R² = 0.88, P < .0001): 14% of patients achieved complete ctDNA clearance, including patients with a best response of PD.⁷⁴ TILs and tumor necrosis may mimic a radiographic progression with changes in tumor size because of immune infiltration and activation rather than tumor growth; further studies are needed to confirm that ctDNA is a surrogate marker more accurate than imaging response criteria to assess tebentafusp benefit in patients with metastatic UM. ctDNA results from the prospective phase III trial 202 are required before ctDNA analyses can be added to routine clinical practice to early identify patients benefiting most from tebentafusp.

Assessment of OR by RECIST or irRECIST criteria underestimates the clinical benefit from tebentafusp treatment. Considering growth kinetics and providing a quantitative evaluation of tumor volume changes over time on the basis of the sum of the largest target lesion diameters, the tumor growth rate may provide additional information compared with standard RECIST criteria on the basis of unidimensional assessment of target lesions.⁷⁵

Moreover, the mechanism of action and the safety profile of tebentafusp may encourage trials combining or sequencing tebentafusp with other systemic treatments. Preclinical data suggest that tebentafusp induces and potentiates antigen cross-presentation by dendritic cells.⁷⁶ The changes induced by tebentafusp in the tumor TME may increase the efficacy of the checkpoint blockade; this is attested by post-tebentafusp biopsies showing increased expression of PD-L1 and PD-1. To investigate the outcomes from patients treated with checkpoint inhibition (CPI), pre- or post-tebentafusp may help to build the next prospective clinical trials in UM.

Regional therapies could also be tested with tebentafusp in selected patients with UM with a limited metastatic disease, combining clinical outcomes and immunologic end points.

Finally, adjuvant and/or neoadjuvant studies should be launched in the near future for patients at high risk of metastasis through international collaborations and networks including the Collaborative Ocular Oncology Group, the International Rare Cancer Initiative, EURACAN, and Patient Advocacy Groups as Cure OM and Melanoma Patient Network Europe.⁷⁷

IMC-F106C, a new bispecific molecule targeting an HLA A*02:01–restricted epitope of PRAME, the most broadly expressed cancer testis antigen in many tumors, such as lung, ovarian, endometrial, or melanoma, is currently tested as a single agent and in combination with CPI (ClinicalTrials.gov identifier: NCT04262466). Preliminary results showed durable RECIST partial responses and ctDNA response in PRAME-positive patients across multiple tumor types, including UM.⁷⁸

A major limitation of immune-mobilizing monoclonal T-cell receptor against cancer (ImmTAC) molecules is HLA restriction. Main ImmTACs target HLA A*02, which is represented at its highest frequency (around 45%) in the White population.⁷⁹ For patients with metastatic UM who are not HLA A*02:01–positive, current options are locoregional strategies focused on the liver, alone or combined with systemic therapies, depending on the extent of metastases; CPI in monotherapy or combination with no evidence of survival benefit; or a clinical trial.⁶⁶

One alternative for half of patients with HLA A*02:01–negative UM could be the development of ImmTACs targeting peptides on other HLA alleles. Other challenges for T-cell–engaging bispecific molecules include selecting the most relevant targets, achieving optimal dosing, reaching higher efficacy, and overcoming tumor resistance.⁸⁰

Giving the immunosuppressive hepatic microenvironment in UM metastases, exploration of new immunotherapy strategies to enhance antitumor immune response is of high interest. Cellular therapies with autologous TILs harvested from UM metastases demonstrated an ORR of 35% in 20 evaluable patients in a single-center phase II study, but survival data are lacking.⁸¹

LAG-3 is a negative regulator of T cells, leading to immune escape of cancers through T-cell dysfunction and immune exhaustion.⁸² Single-cell RNA sequencing showed that most CD8 cytotoxic T cells in UM expressed LAG3 at a high level and correlated with high risk of metastasis.^83,84 A single-arm, Simon, two-stage, phase II trial of relatlimab and nivolumab in patients with metastatic UM is recruiting (ClinicalTrials.gov identifier: NCT04552223).

The combination of MDM2 inhibitor APG-115 with pembrolizumab has demonstrated synergy in vitro via depletion of M2 macrophages from the TME as a result of p53 activation and is tested in a phase I/II study in solid tumors (ClinicalTrials.gov identifier: NCT03611868), showing encouraging results in UM.⁸⁵ Two phase II studies combining lenvatinib and pembrolizumab in patients with CPI-naïve metastatic UM are also recruiting (ClinicalTrials.gov identifiers: NCT05308901 and NCT05282901).

Multiple other emerging strategies are being investigated, including agents targeting protein kinase C, a downstream element of the Gαq signaling pathway. After disappointing results with first-generation sotrastaurin⁸⁶ and limited activity of second-generation LXS196/IDE196 in monotherapy,⁸⁷ the combination of IDE196/darovasertib with crizotinib showed a promising ORR of 31% in 35 evaluable patients (ClinicalTrials.gov identifier: NCT03947385).⁶⁶

Focal adhesion kinase inhibitors target YAP oncogenic activation related to GNAQ/11 mutations and are currently evaluated in monotherapy or in combination with MEK inhibitors (ClinicalTrials.gov identifiers: NCT04109456 and NCT04720417).

DYP-688 is a first-in-class PMEL17 targeting antibody-drug conjugate (ADC), a melanocyte lineage protein highly expressed in cutaneous and UM. On ADC binding to PMEL17 on the target cells, the linker is cleaved to release SDZ475, which inhibits GNAQ/11 oncogenic signaling, resulting in dose-dependent apoptosis. A multicenter phase I/II trial is recruiting patients with metastatic UM and other GNAQ/11-mutant melanoma (ClinicalTrials.gov identifier: NCT05415072). Epigenetic approaches and combination of regional and systemic therapies are also developed.⁶⁶

Finally, patient preferences need to be integrated into clinical trials, with the aim of improving patient satisfaction regarding information and supportive care, through tailored patient-reported outcomes⁸⁸ and dedicated studies (ClinicalTrials.gov identifier: NCT04728113).