Authored by Namrata Chandhok, MD Yale University School of Medicine, New Haven, Connecticut, USA

Dr. Hilsenbeck’s talk was focused on critical issues in the design of translational research development, with an emphasis on biomarkers. She started her presentation by defining translational research as a field of research that “aims to ‘translate’ findings in fundamental research into medical practice and meaningful health outcomes” and noting that while this encompasses a broad spectrum of research opportunities, our discussion is primarily focused on early phase trials. The role of biomarkers evolves throughout the drug development process (possibly in a cyclic manner), and while biomarkers should be developed and assessed concurrently with drug trials, the same study should not assess both the drug and the biomarker because, if the study fails, it is unclear whether the failure is due to the wrong biomarker or the wrong drug. Dr Hilsenbeck also made a point to explain that biomarkers are context-specific and serve a particular function; it may be diagnostic, prognostic, predictive or response, indicative or progression, etc. She then went on to discuss the difference in prognostic versus predictive biomarkers (a critical distinction) before moving onto the role of biomarkers in clinical trials.

We discussed the various types of biomarkers that may be incorporated in a trial, namely integral, integrated, and exploratory biomarkers. An integral biomarker must be assessed for trial participation and may inform eligibility, treatment assignment, or group stratification within the trial. Due to their important role in trials, these biomarkers must be analyzed in a CLIA-certified lab to ensure uniformity in the results. Integrated biomarkers, which are meant to validate an assay or a biomarker based on a prospectively established hypothesis, are performed in a less stringent manner, and exploratory biomarkers are used to develop an initial understanding of a drug or disease biology.

The next major theme discussed by Dr. Hilsenbeck was choosing an evaluable endpoint for a study, which is dependent on several factors such as the treatment setting (neoadjuvant, adjuvant, or metastatic setting) and on the natural history of the disease being evaluated. She explained how interval outcomes may help reduce sample size, which is always a consideration in effectively limiting patient exposure to an agent where efficacy is unknown.

We concluded the session by discussing areas of trial development in which it would be prudent to exercise caution. For example, we must consciously avoid over-interpretation of subgroup analyses. We also discussed the importance of analyzing p-value based on the pre-specified threshold as opposed to “optimal” cut- points, as results may appear meaningful even in the absence of any clinical significance entirely due to chance.

The lecture made clear points in trial design and analysis.

Authored by by Yoshinori Imamura, MD, PhD, Kobe University Hospital, Kobe, Japan

In her talk, Dr. Susan Hilsenbeck focused on the development of biomarkers. First, she showed that there are three components to establishment of new biomarkers; analytic validation, clinical validation, and clinical utility. Second, clinical trials play an important role in validation of new biomarkers, exploration of biomarkers in phase I trials, clarification of the relationship between biomarkers and clinical outcome in phase II trials, and in locking down the biomarkers’ assays in phase III trials. Third, biomarkers can be used to assess risk, diagnosis, prognosis, prediction of treatment benefit, pharmacodynamics or dose adjustment, and/or as a surrogate endpoint for disease progression (prognostic value) or treatment response (predictive value). Finally, she related that there are three aspects in biomarkers:

• Integral biomarkers are involved in the eligibility, treatment assignment, or group stratification.
• Integrated biomarkers have pre-planned analysis intended to validate assays.
• Exploratory biomarkers are used to develop biomarkers or assays and understand the biology of a new agent.

In summary, Dr. Hilsenbeck provided valuable tips for design of translational research trials. It is desirable to set optimal endpoints, to pre-specify hypotheses to be tested with appropriate sample size, and to prepare for elimination of multiplicity.