Targeted cancer clinical trials

TitleTargeted cancer clinical trials
Publication TypeBook Chapter
Year of Publication2012
AuthorsGeorge, Stephen L., and Xiaofei Wang
Book TitleDesigns for Clinical Trials
CityNew York
ISBN Number978-1-4614-0139-1

Randomized clinical trials (RCTs) are the most reliable scientific tool for the identification of safe and effective therapies for cancer. Indeed, it can be argued that such trials are indispensible for this purpose. The design and analysis of these trials are generally aimed at differences in overall population parameters. The trials compare outcome measures such as response rates, progression-free survival (PFS), disease-free survival, and overall survival (OS) for patients assumed to be representative of the population under study, determined by the specified eligibility criteria on the trial. Although no assumption of homogeneity of the patients studied or of the relative effects of treatment is required, major conclusions and, for new agents, subsequent regulatory approvals are generally based on the overall population. For example, a conclusion might be that some new therapy is better than a standard therapy in some population with respect to PFS, if the observed hazard ratio is significantly less than unity (in favor of the new therapy), even if the “average” benefit on other scales (e.g., difference in the PFS medians) is quite small. This is a legitimate and reasonable approach, one that has over time yielded major advances in cancer treatments. However, if the eligibility criteria are sufficiently broad, a generally desirable feature of traditional RCTs, unrecognized heterogeneity may have a substantial effect on the power of the trial (Betensky et al. 2002; Zhang et al. 2006). Indeed, there may be reasons to suspect that the treatment under study might benefit only a subset of patients in the population and that any observed difference is due solely or largely to the results in this subset.If this is true and if such patients can be reliably identified beforehand, major benefits could accrue by targeting the subset population rather than the broader one. Benefits include more efficient clinical trials, avoidance of unnecessary treatment of patients who will not benefit from treatment, faster drug development times, and so on. Indeed, the concept of “personalized” medicine is predicated on the existence of such exploitable patient differences. Of course, the reliable identification of the subset population, assuming such exists, is not easy, but is of critical importance (George 2008).

Original PublicationTargeted cancer clinical trials.