Skip to content

Uveal melanoma

TL;DR — Uveal melanoma shares a name with cutaneous melanoma and almost nothing else: it is driven by GNAQ/GNA11 rather than BRAF, its prognosis is set by monosomy 3 and BAP1 loss, it metastasises preferentially to liver, and checkpoint blockade does not reproduce the cutaneous result. US age-adjusted incidence is 5.6 per million (95% CI 5.5–5.7) over 1975–2020 with 5-year relative survival stable at 82.8% despite a complete shift from surgery to radiation as primary treatment (Weinberger 2025, PMID 40225965). The COMS randomised trial of iodine-125 brachytherapy against enucleation, accruing 1986–1998 with follow-up to 2003, found no difference in survival and little difference in quality of life (Hawkins 2011, PMID 22013172). In metastatic disease tebentafusp is the first agent to improve overall survival: 1-year OS 73% vs 59% (HR 0.51, 95% CI 0.37–0.71, P < .001) in 378 HLA-A02:01-positive patients, median OS 21.6 vs 16.9 months at 3 years (HR 0.68, 0.54–0.87), and 5-year OS 16% vs 8% (stratified HR 0.67, 0.54–0.85) (Nathan 2021, PMID 34551229; Hassel 2023, PMID 37870955; Piperno-Neumann 2026, PMID 42162665). Tebentafusp requires HLA-A02:01, which excludes roughly half the population.

Why it is a different disease

Feature Cutaneous melanoma Uveal melanoma
Dominant drivers BRAF, RAS, NF1 (Cancer Genome Atlas Network 2015, PMID 26091043) GNAQ/GNA11 — 83% of uveal melanomas carry a somatic mutation in one of them (Van Raamsdonk 2010, PMID 21083380)
Prognostic genomics No outcome correlation for the four-way genomic class (PMID 26091043) Monosomy 3 and BAP1 loss define the poor-prognosis class (Robertson 2017, PMID 28810145)
Mutational aetiology UV signature at every evolutionary stage (Shain 2015, PMID 26559571) Not UV-signature-driven
Metastatic tropism Widely distributed, CNS common (M1d) Liver-predominant
5-year relative survival 89% (all stages, SEER 1988–2010) (Bishop 2014, PMID 24272143) 78% (same series); 82.8% and unchanged 1975–2016 (PMID 40225965)
Survival trend over the immunotherapy era Improved No change — neither ocular nor mucosal melanoma improved across 1988–2010 (PMID 24272143), and 5-year relative survival was flat 1975–2016 despite the treatment shift (PMID 40225965)
First-line metastatic therapy Checkpoint blockade Tebentafusp, a bispecific gp100×CD3 ImmTAC requiring HLA-A*02:01 (PMID 34551229)

Molecular architecture

  • GNAQ/GNA11. Sequencing 713 melanocytic neoplasms (186 uveal melanomas, 139 blue nevi, 106 other nevi, 282 other melanomas) found mutually exclusive somatic mutations at Q209 (exon 5) and R183 (exon 4). GNA11 Q209 mutations occurred in 7% of blue nevi, 32% of primary uveal melanomas and 57% of uveal melanoma metastases; GNAQ Q209 in 55% of blue nevi, 45% of uveal melanomas and 22% of metastases. R183 mutations were less common (2% of blue nevi, 6% of uveal melanomas). GNA11 mutation induced spontaneously metastasising tumours in mice and activated the MAPK pathway. Overall 83% of uveal melanomas carried a GNAQ or GNA11 mutation (Van Raamsdonk 2010, PMID 21083380).
  • Four molecular subsets. Multiplatform analysis of 80 uveal melanomas identified four clinically relevant subtypes: two with poor-prognosis monosomy 3 and two with better-prognosis disomy 3. BAP1 loss follows monosomy 3 and correlates with a distinct global DNA methylation state; the monosomy-3 group divides into subsets with divergent genomic aberrations, transcription and outcomes; change-of-function SRSF2 mutations were identified; within disomy 3, EIF1AX- and SRSF2/SF3B1-mutant tumours have distinct copy-number and methylation profiles corresponding to low- versus intermediate-risk clinical categories (Robertson 2017, PMID 28810145).
  • Germline BAP1. BAP1 tumour predisposition syndrome spans uveal melanoma, mesothelioma, cutaneous melanoma, renal cell carcinoma and BAP1-inactivated melanocytic tumours across 181 families carrying 140 unique variants (Walpole 2018, PMID 30517737) — see germline predisposition. Multigene panel testing in 70 uveal melanoma patients meeting NCCN criteria found pathogenic or likely pathogenic variants in 13% of unrelated individuals across BAP1, BRCA1, BRCA2, MBD4, MUTYH, POT1 and XRCC2, with 8 of 10 carriers missed by BAP1-only testing (Byrne 2026, PMID 42415513).
  • Prognostic assays. A 15-gene expression profile predicts metastatic risk (Harbour 2014, PMID 24258991); immunohistochemistry for BAP1 has been evaluated for detecting BAP1 mutation (Koopmans 2014, PMID 24633195), and prognostic parameters have been related to BAP1 expression (van Essen 2014, PMID 25147369). GWAS meta-analysis has identified novel uveal melanoma risk loci with population effect-size heterogeneity (PMID 40495383).

Primary tumour management

The COMS randomised trial of iodine-125 brachytherapy versus enucleation accrued from 1986 to 1998, followed patients 5–15 years, and reported no difference in survival outcomes and little difference in quality of life; five-year survival was substantially better than the trial's design assumptions predicted (Hawkins 2011, PMID 22013172; overview Margo 2004, PMID 15377989). Quality-of-life results were reported separately at five years (PMID 16476893), and threshold analysis of COMS mortality outcomes has been revisited (PMID 40323297).

SEER data show the practice consequence: the proportion treated with surgery alone fell from 93% (1975–1977) to 21% (2017–2020) while primary radiation rose from 1% to 58% — with no change in 5-year relative survival (82.8%) across the same period (Weinberger 2025, PMID 40225965). Eye preservation was achieved without a survival cost, and without a survival gain.

Staging uses the AJCC ophthalmic classification, internationally validated in 3,809 patients from ten ocular oncology centres on four continents diagnosed 2001–2011 (AJCC Ophthalmic Oncology Task Force 2015, PMID 25555246). A competing-risk reanalysis of 6,528 consecutively registered patients from three centres over 1981–2022 proposes volume-based size categories and seven stages instead (Stålhammar 2024, PMID 38071620) — see staging.

Radiotherapy modality

Proton beam therapy is the main alternative to plaque brachytherapy for globe-conserving treatment. A review of clinical outcomes found most treated tumours choroidal and medium- or large-sized, receiving 50–70 Cobalt Gray equivalent with lower doses in more recent series: 5-year local control exceeded 90% and persisted at 10 and 15 years; 5-year overall survival 70–85%; 5-year metastasis-free and disease-specific survival 75–90%; 5-year enucleation rates consistently 7–10% after removing smaller studies; and 60–70% of patients showed a post-treatment visual-acuity decrease while retaining purposeful vision (>20/200), with better figures in more recent higher-volume series and complication rates improved on historical plaque-brachytherapy data (Verma 2016, PMID 26915706). Salvage proton therapy for local recurrence has been reported (PMID 25038327), and treatment and prognostication reviewed together (PMID 28399342).

The pattern matches COMS's conclusion: modality choice affects eye and vision outcomes, and the survival evidence does not distinguish between the globe-conserving options.

Metastatic surveillance

Because metastasis is liver-predominant and prognostic class is knowable from the primary tumour, uveal melanoma has a surveillance literature that cutaneous melanoma lacks.

Study Finding
Choudhary 2016, PMID 26633182 Retrospective cohort of primary uveal melanoma treated 2003–2012 under a standardised protocol: contrast-enhanced CT of chest, abdomen and pelvis at staging, then hepatic ultrasonography and liver function tests every 6 months for 5 years and annually thereafter, with abnormal findings categorised as cyst/haemangioma, indeterminate, suspicious, or consistent with metastasis
Rantala 2020, PMID 32278769 Reliability analysis in 215 patients diagnosed with hepatic metastases by ultrasound within 60 days of staging CT/MRI, 1999–2016; 67% had biopsy-confirmed metastases, screening was regular in 98% and 66% were asymptomatic. Ultrasound was fully consistent with CT/MRI on the presence and number of metastases in only 113 of 215 (53%)
Zabor 2026, PMID 42616744 349 patients (median age 62, 54% male, 65% gene-expression-profile Class 1, 75% choroidal) triaged by commercial GEP: Class 1 offered 6-monthly hepatic ultrasound, Class 2 recommended enhanced surveillance by higher frequency (3-monthly) or enhanced modality (hepatic CT/MRI)

This is risk-stratified surveillance in practice — the strategy that cutaneous melanoma has modelled but not implemented (see screening and overdiagnosis) — and its weak point is the modality: ultrasound agreed fully with cross-sectional imaging in only just over half of cases.

Metastatic disease: tebentafusp

Analysis Result
IMCgp100-202 primary (Nathan 2021, PMID 34551229) 378 previously untreated HLA-A02:01-positive patients randomised 2:1 to tebentafusp or investigator's choice (pembrolizumab, ipilimumab or dacarbazine), stratified by LDH. 1-year OS 73% vs 59%, HR 0.51 (0.37–0.71), P < .001. PFS at 6 months 31% vs 19%, HR 0.73 (0.58–0.94), P* = .01. Commonest treatment-related events were cytokine-mediated (T-cell activation) and skin-related (gp100-positive melanocytes)
3-year (Hassel 2023, PMID 37870955) Median OS 21.6 vs 16.9 months, HR 0.68 (0.54–0.87); 3-year survival 27% vs 18%. Any-grade rash 83%, pyrexia 76%, pruritus 70%, hypotension 38%; most events early, none new with long-term administration
5-year (Piperno-Neumann 2026, PMID 42162665) Median OS unchanged at 21.6 vs 16.9 months, stratified HR 0.67 (0.54–0.85); 5-year OS 16% vs 8%. Benefit persisted in poor-prognosis groups including baseline tumours ≥10 cm and patients whose best RECIST response was progressive disease with target growth >20%. Post hoc, treatment beyond radiographic progression was associated with longer survival; longer OS associated with undetectable ctDNA
Versus combination checkpoint blockade (Piulats 2024, PMID 38048850) Propensity-score inverse-probability-weighted comparison of 240 tebentafusp patients from IMCgp100-202 against 45 nivolumab-plus-ipilimumab patients from GEM-1402, balancing age, sex, LDH, alkaline phosphatase, disease location, ECOG status and time to metastasis: OS HR 0.52 (0.35–0.78); 1-year OS 73% vs 50%. All sensitivity-analysis HRs ≤0.61. Pembrolizumab versus nivolumab-plus-ipilimumab showed no significant difference (HR 0.72, 0.50–1.06)
Previously treated disease (Sacco 2024, PMID 38844408) Long-term survival follow-up for tebentafusp after prior therapy

The three most consequential features of the tebentafusp result are unusual. First, survival benefit without a conventional response benefit — benefit extended even to patients whose best response was progression with >20% tumour growth (PMID 42162665). Second, HLA restriction: the drug requires HLA-A*02:01, whose frequency varies by population, with direct consequences for access (a Brazilian analysis quantified this, PMID 38785402). Third, the comparison against checkpoint blockade is a propensity-weighted cross-trial analysis, not a randomised comparison (PMID 38048850).

Liver-directed therapy

Liver-predominant metastasis is the defining clinical problem.

Approach Evidence
Melphalan hepatic delivery system (percutaneous hepatic perfusion) Open-label randomised study, 85 patients randomised 1:1 to melphalan/HDS (3 mg/kg ideal body weight q6–8 weeks, max 6 cycles) or best alternative care; the design was amended to single-arm because of slow enrolment and patient reluctance to receive best alternative care, so all efficacy analyses are exploratory. Median OS 18.5 vs 14.5 months; median PFS 9.1 vs 3.3 months; ORR 27.5% vs 9.4%; disease control 80.0% vs 46.9%. Serious adverse events 51.2% vs 21.9%, commonest thrombocytopenia (19.5%), neutropenia (9.8%), leukopenia (9.8%), febrile neutropenia (7.3%) (Zager 2025, PMID 40192993)
Perfusion plus checkpoint blockade CHOPIN, single-centre open-label randomised phase 2 (NCT04283890): 76 eligible of 80 screened, unresectable liver-only or liver-dominant disease, no prior systemic therapy, randomised 1:1 to percutaneous hepatic perfusion alone or with ipilimumab 1 mg/kg + nivolumab 3 mg/kg at weeks 0, 3, 6 and 9 without maintenance; two melphalan perfusions at weeks 1 and 7. Primary endpoint 1-year progression-free survival (van den Hoek 2026, PMID 41785896)
Isolated and percutaneous hepatic perfusion generally Meta-analysis of both techniques for uveal melanoma liver metastases (PMID 34572953); PFS after perfusion as first- or second-line therapy (PMID 39174837); comparison against selective internal radiotherapy (PMID 37894309); troponin elevation as a procedural harm (PMID 38605434); immunoembolisation series (PMID 42109660)

Why checkpoint blockade underperforms

The candidate explanations are mutational burden (uveal melanoma lacks the UV-driven burden that generates neoantigens in cutaneous disease), the GNAQ/GNA11 versus BRAF driver divide, and the immunological privilege of the eye. Real-world data on checkpoint inhibitors in metastatic uveal melanoma exist (PMID 31623302), and the propensity-weighted analysis above puts 1-year survival on nivolumab plus ipilimumab at 50% against 73% with tebentafusp (PMID 38048850). The explanation is described rather than tested as a predictive framework: no study has used it to predict in advance which other tumours will and will not respond to checkpoint blockade. Single-cell work on uveal melanoma heterogeneity and its microenvironment has been reviewed (PMID 38966635), and phenotypic plasticity demonstrated in an immune-competent genetically engineered mouse model (PMID 41661679).

Interpretation rules for this page

  • Do not transfer cutaneous melanoma evidence to uveal melanoma. Different drivers, different tropism, different first-line agent, and no survival improvement over the era in which cutaneous survival improved (PMID 24272143; PMID 40225965).
  • Report HLA status. Tebentafusp's evidence base is entirely HLA-A*02:01-positive (PMID 34551229).
  • Tebentafusp separates survival from response. Judging it on RECIST response understates its effect (PMID 42162665).
  • The tebentafusp-versus-checkpoint comparison is propensity-weighted across trials, not randomised (PMID 38048850).
  • The melphalan/HDS randomised comparison was abandoned mid-trial; its efficacy analyses are exploratory (PMID 40192993).
  • Eye preservation is not a survival trade-off — COMS showed equivalence, and SEER shows a complete practice shift with unchanged survival (PMID 22013172; PMID 40225965).
  • Uveal melanoma is an adult intraocular malignancy; the paediatric intraocular malignancy is retinoblastoma, a separate disease — see retinoblastoma.

Open questions

  • Why does checkpoint blockade transform cutaneous but not uveal melanoma, and does that explanation predict anything in advance (PMID 39282897; PMID 38048850)?
  • What is the treatment for HLA-A*02:01-negative patients, who are excluded from the only agent with a survival benefit (PMID 34551229; PMID 38785402)?
  • Why does tebentafusp prolong survival in patients with progressive disease by RECIST (PMID 42162665)?
  • Does adding checkpoint blockade to hepatic perfusion improve outcomes? CHOPIN is the first randomised test (PMID 41785896).
  • Can adjuvant therapy after primary treatment reduce metastatic risk in monosomy-3/BAP1-loss tumours, where risk is known but no intervention is established (PMID 28810145)?
  • Should volume-based staging replace the current AJCC size categories (PMID 38071620)?
  • Why has 5-year survival been flat for five decades despite a complete change in primary treatment (PMID 40225965; PMID 22013172)?

References

  1. Weinberger Y, et al. Uveal Melanoma: 5-Year Update on Incidence, Treatment, and Survival (SEER 1975-2020). Ocular oncology and pathology. 2025;11:30-36. PMID 40225965
  2. Hawkins BS. Collaborative ocular melanoma study randomized trial of I-125 brachytherapy. Clinical trials (London, England). 2011;8:661-73. PMID 22013172
  3. Nathan P, et al. Overall Survival Benefit with Tebentafusp in Metastatic Uveal Melanoma. The New England journal of medicine. 2021;385:1196-1206. PMID 34551229
  4. Hassel JC, et al. Three-Year Overall Survival with Tebentafusp in Metastatic Uveal Melanoma. The New England journal of medicine. 2023;389:2256-2266. PMID 37870955
  5. Piperno-Neumann S, et al. Five-year survival with tebentafusp in metastatic uveal melanoma. Annals of oncology : official journal of the European Society for Medical Oncology. 2026;37:1266-1277. PMID 42162665
  6. Cancer Genome Atlas Network. Genomic Classification of Cutaneous Melanoma. Cell. 2015;161:1681-96. PMID 26091043
  7. Van Raamsdonk CD, et al. Mutations in GNA11 in uveal melanoma. The New England journal of medicine. 2010;363:2191-9. PMID 21083380
  8. Robertson AG, et al. Integrative Analysis Identifies Four Molecular and Clinical Subsets in Uveal Melanoma. Cancer cell. 2017;32:204-220.e15. PMID 28810145
  9. Shain AH, et al. The Genetic Evolution of Melanoma from Precursor Lesions. The New England journal of medicine. 2015;373:1926-36. PMID 26559571
  10. Bishop KD, et al. Epidemiology and survival outcomes of ocular and mucosal melanomas: a population-based analysis. International journal of cancer. 2014;134:2961-71. PMID 24272143
  11. Walpole S, et al. Comprehensive Study of the Clinical Phenotype of Germline BAP1 Variant-Carrying Families Worldwide. Journal of the National Cancer Institute. 2018;110:1328-1341. PMID 30517737
  12. Byrne L, et al. Impact of Multigene Panel Testing in High-Risk Uveal Melanoma Patients. Pigment cell & melanoma research. 2026;39:e70108. PMID 42415513
  13. Harbour JW. A prognostic test to predict the risk of metastasis in uveal melanoma based on a 15-gene expression profile. Methods in molecular biology (Clifton, N.J.). 2014;1102:427-40. PMID 24258991
  14. Koopmans AE, et al. Clinical significance of immunohistochemistry for detection of BAP1 mutations in uveal melanoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2014;27:1321-30. PMID 24633195
  15. van Essen TH, et al. Prognostic parameters in uveal melanoma and their association with BAP1 expression. The British journal of ophthalmology. 2014;98:1738-43. PMID 25147369
  16. Mies G, et al. Meta-analysis of uveal melanoma genome-wide association studies identifies novel risk loci and population effect size heterogeneity. HGG advances. 2025;6:100465. PMID 40495383
  17. Margo CE. The Collaborative Ocular Melanoma Study: an overview. Cancer control : journal of the Moffitt Cancer Center. 2004;11:304-9. PMID 15377989
  18. Melia M, et al. Quality of life after iodine 125 brachytherapy vs enucleation for choroidal melanoma: 5-year results from the Collaborative Ocular Melanoma Study: COMS QOLS Report No. 3. Archives of ophthalmology (Chicago, Ill. : 1960). 2006;124:226-38. PMID 16476893
  19. Gill VT, et al. Threshold analysis of mortality outcomes in the collaborative ocular melanoma study (COMS). Melanoma management. 2025;12:2494977. PMID 40323297
  20. AJCC Ophthalmic Oncology Task Force. International Validation of the American Joint Committee on Cancer's 7th Edition Classification of Uveal Melanoma. JAMA ophthalmology. 2015;133:376-83. PMID 25555246
  21. Stålhammar G, et al. Improved Staging of Ciliary Body and Choroidal Melanomas Based on Estimation of Tumor Volume and Competing Risk Analyses. Ophthalmology. 2024;131:478-491. PMID 38071620
  22. Verma V, et al. Clinical Outcomes of Proton Radiotherapy for Uveal Melanoma. Clinical oncology (Royal College of Radiologists (Great Britain)). 2016;28:e17-27. PMID 26915706
  23. Riechardt AI, et al. Salvage proton beam therapy in local recurrent uveal melanoma. American journal of ophthalmology. 2014;158:948-56. PMID 25038327
  24. Dogrusöz M, et al. Uveal Melanoma Treatment and Prognostication. Asia-Pacific journal of ophthalmology (Philadelphia, Pa.). 2017;6:186-196. PMID 28399342
  25. Choudhary MM, et al. Hepatic Ultrasonography for Surveillance in Patients With Uveal Melanoma. JAMA ophthalmology. 2016;134:174-80. PMID 26633182
  26. Rantala ES, et al. Hepatic Ultrasonography Compared With Computed Tomography and Magnetic Resonance Imaging at Diagnosis of Metastatic Uveal Melanoma. American journal of ophthalmology. 2020;216:156-164. PMID 32278769
  27. Zabor EC, et al. Systemic surveillance for uveal melanoma: Approach to optimization. PloS one. 2026;21:e0355405. PMID 42616744
  28. Piulats JM, et al. Overall survival from tebentafusp versus nivolumab plus ipilimumab in first-line metastatic uveal melanoma: a propensity score-weighted analysis. Annals of oncology : official journal of the European Society for Medical Oncology. 2024;35:317-326. PMID 38048850
  29. Sacco JJ, et al. Long-term survival follow-up for tebentafusp in previously treated metastatic uveal melanoma. Journal for immunotherapy of cancer. 2024;12. PMID 38844408
  30. de Melo AC, et al. Frequency of HLA-A*02:01 in the Brazilian population and its impact on uveal melanoma systemic treatment. The oncologist. 2024;29:e1098-e1099. PMID 38785402
  31. Zager JS, et al. An Open-label, Randomized Study of Melphalan/Hepatic Delivery System Versus Best Alternative Care in Patients with Unresectable Metastatic Uveal Melanoma. Annals of surgical oncology. 2025;32:4976-4988. PMID 40192993
  32. van den Hoek L, et al. Percutaneous hepatic perfusion combined with ipilimumab and nivolumab for metastatic uveal melanoma (CHOPIN): a single-centre, open-label, randomised, phase 2 trial. The Lancet. Oncology. 2026;27:372-382. PMID 41785896
  33. Bethlehem MS, et al. Meta-Analysis of Isolated Hepatic Perfusion and Percutaneous Hepatic Perfusion as a Treatment for Uveal Melanoma Liver Metastases. Cancers. 2021;13. PMID 34572953
  34. Ghali H, et al. Hepatic and Overall Progression-Free Survival After Percutaneous Hepatic Perfusion (PHP) as First-Line or Second-Line Therapy for Metastatic Uveal Melanoma. Annals of surgical oncology. 2024;31:9150-9158. PMID 39174837
  35. Kolb M, et al. Selective Internal Radiotherapy (SIRT) and Chemosaturation Percutaneous Hepatic Perfusion (CS-PHP) for Metastasized Uveal Melanoma: A Retrospective Comparative Study. Cancers. 2023;15. PMID 37894309
  36. Chandrasekhar S, et al. Troponin Elevation in Patients Undergoing Percutaneous Hepatic Perfusion for Metastatic Uveal Melanoma. Cancer control : journal of the Moffitt Cancer Center. 2024;31:10732748241246898. PMID 38605434
  37. Morecroft R, et al. Immunoembolization in liver-predominant metastatic uveal melanoma: a single-center retrospective analysis. Frontiers in oncology. 2026;16:1752725. PMID 42109660
  38. Bol KF, et al. Real-World Impact of Immune Checkpoint Inhibitors in Metastatic Uveal Melanoma. Cancers. 2019;11. PMID 31623302
  39. Tang S, et al. Single cell RNA-sequencing in uveal melanoma: advances in heterogeneity, tumor microenvironment and immunotherapy. Frontiers in immunology. 2024;15:1427348. PMID 38966635
  40. Xu X, et al. A Multistep Immune-Competent Genetically Engineered Mouse Model Reveals Phenotypic Plasticity in Uveal Melanoma. Cancer research. 2026;86:3666-3686. PMID 41661679
  41. Wolchok JD, et al. Final, 10-Year Outcomes with Nivolumab plus Ipilimumab in Advanced Melanoma. The New England journal of medicine. 2025;392:11-22. PMID 39282897