Summary authored by Yiduo Hu, MD, PhD

Issues in Design of Translational Research Trials

Yiduo Hu, MD, PhD

Dr. Hilsenbeck led an important discussion on the key considerations when designing a translational trial, especially one that focuses on testing biomarkers and targeted therapy. She started with an emphasis on the importance of a complete circle in translational research that includes bringing lab findings to test in trials, using trial results to benefit the patient, and using the findings in the trial to take back to the lab and test the new hypothesis.

She then continued to discuss the key elements of designing translational trials on targeted therapy: (1) integration of biomarker(s); (2) endpoints vs outcomes; (3) basic considerations of study design, statistical power, sample size, and whether to incorporate adaptive vs non-adaptive design; as well as (4) pitfalls and concerns. She then discussed the characteristics of biomarkers and how biomarkers should be used. She reviewed the steps of a trial when biomarkers are integrated, including analytic validation to confirm the assay performance, clinical validation to assess the association between test results and pathophysiology state, and clinical utility of the biomarkers. It is important to understand the different types of biomarkers and that their utility is context specific. 

She next discussed the roles of biomarkers in a study. (1) Integral biomarkers are required for the conduction of a trial. A CLIA lab test might be required, as well as an IND or IDE. (2) Integrated biomarkers are used to validate secondary outcomes. (3) Ancillary or exploratory biomarkers which are not directly related to the study but important for exploratory studies derived from the main research.

Dr. Hilsenbeck then moved on to discuss the importance of appropriate objectives and the difference between objectives and outcomes. She emphasized the importance of having early involvement and collaboration with biostatisticians during the design of a study. Finally, she discussed the advantages and pitfalls of interim analysis. We briefly reviewed study design models in which interim analysis can be integrated appropriately, including Simon two-stage design and Bayesian methods.