Duration of survival for disseminated malignant melanoma: results of a meta-analysis

The aim of this study was to estimate the survival experience of patients with disseminated malignant (stage IV) melanoma regardless of the treatment received. Articles, including abstracts, published in the English language medical literature between 1985 and 1999 were identified through an electronic key word literature search using MEDLINE and PUB-MED together with the reference lists of review articles. Some literature published prior to 1985 was identified through these searches and was also included. Median survivals were calculated across all the data, for the post- and pre-1985 periods, and for the larger studies alone (n > or = 20) using a weighting scheme based on sample size and study characteristics. An attempt was also made to estimate the frequency of long-term survival to 2, 3 and 5 years in this patient population (also stratified according to the study publication time). A total of 83 studies were identified comprising a total of 6322 patients. The overall median survival from diagnosis of stage IV melanoma was estimated to be 8.1 months (approximate 95% confidence interval [CI] 7.3-8.9 months). For the 59 studies published since 1985 (covering 3715 patients), the median survival was 8.9 months (approximate 95% CI 7.9-9.9 months). Prior to 1985, the median survival was 5.8 months (approximate 95% CI 4.5-7.1 months). For the 67 studies with at least 20 patients (a total of 6024 patients), the median survival was 8.1 months (approximate 95% CI 7.2-9.0 months). Long-term survival for the post- and pre-1985 periods over 2, 3 and 5 years was estimated to be 13.6%, 9.7% and 2.3%, respectively, and was consistent across the two time periods. In conclusion, while the survival experience in stage IV melanoma patients has improved since 1985, the use of varied approaches to treat this disease has not produced a favourable long-term prognosis. This meta-analysis will allow the survival results from current and future trials to be compared with the expected survival based on cumulative findings to date.