Skip to content

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

One-paragraph summary

IMCgp100-202 randomised 378 previously untreated, HLA-A02:01-positive patients with metastatic uveal melanoma 2:1 to tebentafusp — a bispecific molecule linking an affinity-enhanced gp100-directed T-cell receptor to an anti-CD3 effector — or to the investigator's choice of single-agent pembrolizumab, ipilimumab or dacarbazine, stratified by lactate dehydrogenase level. The primary endpoint was overall survival. Overall survival at 1 year was 73% with tebentafusp and 59% with control (hazard ratio for death 0.51, 95% CI 0.37–0.71, P < .001) in the intention-to-treat population. Progression-free survival was also higher (31% vs 19% at 6 months; HR 0.73, 0.58–0.94, P* = .01). The commonest treatment-related events were cytokine-mediated, from T-cell activation, and skin-related, from targeting gp100-positive melanocytes.

Key findings

  • 1-year overall survival 73% vs 59%; HR for death 0.51 (0.37–0.71), P < .001.
  • 6-month progression-free survival 31% vs 19%; HR 0.73 (0.58–0.94), P = .01.
  • At 3 years: median overall survival 21.6 vs 16.9 months, HR 0.68 (0.54–0.87), with 27% vs 18% alive; any-grade rash 83%, pyrexia 76%, pruritus 70%, hypotension 38%, with most events early and none new on long-term administration (Hassel 2023, PMID 37870955).
  • At 5 years: median overall survival unchanged at 21.6 vs 16.9 months, stratified HR 0.67 (0.54–0.85), 5-year overall survival 16% vs 8%; benefit persisted in poor-prognosis groups including patients whose best RECIST response was progressive disease with target growth exceeding 20% (Piperno-Neumann 2026, PMID 42162665).
  • Propensity-score-weighted comparison against first-line nivolumab plus ipilimumab gave OS HR 0.52 (0.35–0.78) with 1-year survival 73% vs 50% (Piulats 2024, PMID 38048850).

Limitations

  • Restricted to HLA-A*02:01-positive patients, which excludes a large share of the population and varies by ancestry, with direct access consequences (PMID 38785402).
  • Open-label design with investigator's-choice control, so the comparator is heterogeneous.
  • The survival benefit is not accompanied by a proportionate response benefit; judging the agent on RECIST criteria understates its effect, which complicates its use as a template for other agents.
  • The comparison against combination checkpoint blockade is propensity-weighted across two separate trials, not randomised (PMID 38048850).
  • Median overall survival did not change between the 3-year and 5-year analyses, so the benefit is concentrated in a tail rather than in a shift of the median.

Why it matters

This is the first therapy of any kind to improve overall survival in metastatic uveal melanoma — a disease whose 5-year relative survival was unchanged at 82.8% from 1975 to 2016 despite a complete shift in primary treatment from enucleation to radiation (Weinberger 2025, PMID 40225965), and in which checkpoint blockade never reproduced its cutaneous success. It also validates the ImmTAC platform, which is now being tested in cutaneous melanoma after checkpoint failure (NCT05549297) and first line against nivolumab with a PRAME-directed construct (NCT06112314). Together with darovasertib, adjuvant tebentafusp, adjuvant melatonin and oncolytic-virus trials, it has taken uveal melanoma from essentially no phase 3 pipeline to six active phase 3 or phase 2/3 efficacy trials (including NCT05502900 and NCT06581406) — see clinical trials landscape.

Cited by wiki pages

  • uveal melanoma
  • overview
  • clinical trials landscape