Adaptive Design of Early Phase Cancer Clinical Trials

Targeted therapy in cancer has identified a number of gene-specific biomarkers, enabling the development of targeted anticancer drugs. An umbrella Phase I study is a Phase I trial assessing more than one anticancer drug in a single clinical trial.

Specific biomarkers in patients are identified and those patients receive drugs that target their particular biomarkers. Strong preclinical data are required prior to the start of this kind of trial.

Researchers at the Center for Childhood Cancer Research, including Yimei Li, PhD, are working on Bayesian statistical models for the design of this type of trial, testing three drugs in neuroblastoma and evaluating the operating characteristics of such a design through computer simulations. Many variables have to be considered: dose range and dose toxicity, multiple drugs and prioritizing the drugs in the dosing finding.

Subjects will be randomly assigned to a potential candidate drug, unless there is a clear benefit for one drug, in which case that drug will be assigned first for a greater number of patients in that cohort. This potential adaptive randomization of drug dosing is another consideration in the design of the trial, in addition to finding the optimal dose for each of the drugs.

By testing multiple agents in a single patient, information can be generated on which drug works best for which biomarker. Results from computer simulations can inform the design of actual trials that assess the clinical efficacy of targeted therapeutics in cancer and other diseases.

Our Collaborators