Role of meta-analyses and of large randomized trials in the study of cancer treatments.

The paper contributes to the discussion on the usefulness of meta-analysis as compared to large scale trial. After reviewing the issues to consider when comparing the results of a meta-analysis and of a large trial, several examples, in particular in thoracic oncology, of complementary uses of meta-analysis and large trials are discussed. Three types of examples are presented: (a) trial designed to confirm the overall results of a meta-analysis; (b) trial designed to confirm an interaction between a covariate and the treatment effect detected in a meta-analysis; (c) trial designed to confirm an interaction between a treatment parameter and the treatment effect detected in a meta-analysis. In conclusion, well-conducted meta-analyses and large trials are complementary tools to evaluate new treatments.