Prognostic and predictive value of metformin in the European Organisation for Research and Treatment of Cancer 1325/KEYNOTE-054 phase III trial of pembrolizumab versus placebo in resected high-risk stage III melanoma

Background: Metformin is a commonly prescribed and well-tolerated medication. In laboratory studies, metformin suppresses BRAF wild-type melanoma cells but accelerates the growth of BRAF-mutated cells. This study investigated the prognostic and predictive value of metformin, including with respect to BRAF mutation status, in the European Organisation for Research and Treatment of Cancer 1325/KEYNOTE-054 randomised controlled trial. Methods: Patients with resected high-risk stage IIIA, IIIB, or IIIC melanoma received 200 mg of pembrolizumab (n = 514) or placebo (n = 505) every 3 weeks for twelve months. Pembrolizumab prolonged recurrence-free survival (RFS) and distant metastasis-free survival (DMFS) at approximately 42 months median follow-up (Eggermont et al., TLO, 2021). Multivariable Cox regression was used to estimate associations of metformin with RFS and DMFS. Interaction terms were used to model effect modification by treatment and BRAF mutation. Results: Fifty-four patients (0.5%) used metformin at baseline. Metformin was not significantly associated with RFS (hazard ratio [HR] 0.87, 95% confidence interval [CI] 0.52–1.45) and DMFS (HR 0.82, 95% CI 0.47–1.44). The interaction between metformin and the treatment arm was not significant for either RFS (p = 0.92) or DMFS (p = 0.93). Among patients with mutated BRAF, the association of metformin with RFS (HR 0.70, 95% CI 0.37–1.33) was greater in magnitude though not significantly different to those without mutated BRAF (HR 0.98, 95% CI 0.56–1.69). Conclusions: There was no significant impact of metformin use on pembrolizumab efficacy in resected high-risk stage III melanoma. However, larger studies or pooled analyses are needed, particularly to explore a possible effect of metformin in BRAF-mutated melanoma.