As an unexpected end of the year treat in December 2012, FDA released a new draft guidance document (Guidance for Industry: Enrichment Strategies for Clinical Trials to Support Approval of Human Drugs and Biological Products). This document broadly describes a variety of strategies that sponsors might use to optimize the selection of patients for the study of a new drug. Of particular interest to the IVD world are important sections of this document providing current FDA thinking on study designs for use of companion diagnostics. This is the first technical analysis by FDA of this issue since publication of a draft concept paper (Drug-Diagnostic Co-Development Concept Paper) in 2005.
Enrichment is defined as “the prospective use of any patient characteristic to select a study population in which detection of a drug effect … is more likely than it would be in an unselected population.” Three strategies are identified:
- Strategies to decrease heterogeneity: these include selecting patients with baseline measurements in a narrow range and excluding patients whose disease or symptoms improve spontaneously or whose measurements are highly variable.
- Prognostic enrichment strategies: selecting patients with a greater likelihood of having disease-related endpoint events or a substantial worsening in condition.
- Predictive enrichment strategies: selecting patients more likely to respond to a drug treatment than other patients with the condition being treated.
The guidance provides in detail options for using tests to identify subjects for demonstrating the efficacy and safety of a new drug. Of particular interest it addresses the popular albeit controversial use of drug studies looking at efficacy in biomarker positive patients only. At least four high visibility FDA approvals of companion diagnostics (Her2, EGFR, BRAF, and ALK) have been based on this technique. However, as the guidance carefully notes “because a study that uses a marker-positive only population will provide no direct information about the marker-negative populations, its use should generally be limited to situations in which information about the marker-negative population is not needed or is not feasible.”
FDA is sufficiently concerned about this type of study design that it specifically calls for use of “an established, FDA approved, laboratory test explicitly labeled for this purpose as a companion diagnostic.” The guidance, however, fails to address the core challenge in the use of the biomarker positive study design, which is the failure of this type of study to provide information on the sensitivity, specificity, or negative predictive value of a test. These limitations make it a bit of a leap of faith to characterize the new biomarker assay itself as safe and effective. A more narrow interpretation is that in the pre-selected populations the test and drug working together appear to have benefits that outweigh the risks. Ongoing studies of Her 2 and ALK promise to offer insights into how significant the limitations of this study approach may or may not be.
Building off the work of Simon (NCI) and Wang (FDA) the document also for the first time addresses how adaptive designs may be applied to companion diagnostics. This is a useful public recognition of techniques being proposed by cutting edge statisticians and trialists for the past several years. It is to be hoped that the descriptions of these approaches in this guidance may encourage future exploration of these ideas in actual real-world submissions.
Comments on this draft guidance are welcome until February 15, 2013. It is not clear whether the final draft will involve minor tweaks or more substantive refinements. However it is clear that FDA is providing intellectual leadership in the area of companion diagnostics and working to create a flexible but scientifically sound approach toward how it regulates these important products. The guidance, although perhaps only a starting point toward mastering companion diagnostics, is nonetheless most welcome.