Other paradigms: randomized discontinuation trial design

Phase II oncologic clinical trials for agents that are mainly growth inhibitory and benefit only a selected patient population are challenging. The randomized discontinuation trial design is one approach by which this can be accomplished. A broad patient population is enrolled and all patients receive the investigational agent. Those with tumor shrinkage sufficient to be deemed of likely clinical significance after a prespecified period continue treatment, those with growth sufficient to be deemed clinically adverse and those with toxicity discontinue, and the remaining patients with "stable disease" are randomized to continue or discontinue therapy in a double-blind manner. The primary end point is the fraction of patients who remain progression free after an additional postrandomization period or the time to progression after randomization. By enriching for a possible sensitive population and then testing whether this was due to the agent or selection of an indolent disease group, the randomized discontinuation trial design efficiently assesses the putative growth inhibitory properties of an investigational agent and furthermore minimizes the number of patients exposed to placebo.