Treatments for metastatic melanoma: Synthesis of evidence from randomized trials

Summary Background Advanced melanomas (non-resectable Stage-III/IV) are fatal, with few effective treatments. It remains unclear if other drugs offer improvements over the standard, dacarbazine. Purpose We quantified objective response rates (Complete + Partial response) of dacarbazine versus comparators for advanced cutaneous melanoma. Methods We retrieved all head-to-head randomized controlled trials involving dacarbazine and other drugs/combinations. Two reviewers searched MEDLINE (1966–Jan 2006), EMBASE (1980–2006), CINAHL (1982–2006) and Cochrane library, then compared results. Differences were resolved through consensus. Rates were combined using random effects meta-analysis. χ 2 tested heterogeneity; points from Jadad’s method were assessed to examine study quality. Results We found 48 studies having 111 active treatment arms [24 with dacarbazine monotherapy ( n  = 1390), 75 with dacarbazine combinations ( n  = 4962), and 12 with non-dacarbazine treatments ( n  = 783)] treating 7135 patients. Overall, study quality was poor. Response to dacarbazine monotherapy ranged between 5.3% and 28.0% (average 15.3%), OR = 1.31, CI 95% : 1.06–1.61; N  = 3356. Partial responses comprised 73% of successes. Only adding interferons improved response rates (OR = 1.69, CI 95% : 1.07–2.68, N  = 778) but survival duration was not significantly longer ( P  = 0.32), and trials with larger sample sizes found lower success rates. All other treatments alone or in combination were ineffective P  > 0.05. Conclusions Dacarbazine generally produces poor outcomes. Adding other therapies offers minimal clinical advantages (possibly with interferons). In general, study quality was poor and sample sizes were small. This meta-analysis highlights the unmet need for effective treatment options for advanced melanoma.