Phase II clinical trials of experimental treatments play an essential role in drug development. Historically, trials were conducted sequentially, starting from a phase I trial in which dose and safety are evaluated advancing to phase II, looking for some sign of efficacy on a short-term end-point such as tumour shrinkage to screen out inefficacious experimental agents, and to large-scale randomised phase III trials evaluating properly the efficacy of a new treatment on objective clinical end-points. To expedite this process, phase II trials historically used single-arm designs to treat patients with experimental therapies only [1]. However, the overall success rate of drug development programs in oncology from the start of phase I to registration has been estimated recently to be only 3.4%, much lower compared with that of other diseases [2]. Part of this low success rate has been argued to be due to suboptimal use of randomised trial designs in the phase II space.