OBJECTIVES: In the last 3 years, many new advanced melanoma treatments have come to market. Ipilimumab was the first compound to demonstrate a statistically significant improvement in overall survival in two Phase III RCTs and observational studies, with prolonged long-term survival up to 10 years. METHODS: Data from the advanced melanoma trials: CA184-024, MDX010-20 and BRIM-3 were used to develop an economic model in the UK and to predict the conditional survival of patients treated with ipilimumab, dacarbazine and vemurafenib who survived 2 years. Dabrafenib, pembrolizumab and nivolumab were not valid UK comparators. The model uses patient level Kaplan–Meier data for the first 2 years for ipilimumab and dacarbazine, gompertz parametric curves fitted to patient level data from 2 to 5 years, and American Joint Committee on Cancer registry data from 5 years to 40 years. The curves were a good fit to the trial data and consistent with published longer-term data for ipilimumab (up to 10 years). Vemurafenib data were taken from the Roche submission to NICE, with covariate adjustment to account for differences in patient characteristics between trials: M-stage, brain metastases, LDH, gender and ECOG.  RESULTS: Once a previously untreated patient has survived 2 years, the modelled probability of being alive at 5 years is 70.9% for ipilimumab, 44.3% for dacarbazine and 50.2% for vemurafenib. At 10 years conditional survival is 45.4%, 28.4% and 31.6%, respectively; at 20 years 19.2%, 12.0% and 12.9%. Difference in mean survival was 1.2 years vs dacarbazine; 0.4 vs vemurafenib. CONCLUSIONS: A substantial proportion of patients treated with ipilimumab surviving to 2 years are likely to have a sustained survival benefit: 45.4% of such ipilimumab patients are alive at 10 years and 19.2% at 20 years. The level of sustained survival is lower with dacarbazine and vemurafenib.