Invited Session 7: Platform Design: Lessons Learned and the Future

Chair: Alice Chen, National Cancer Institute

MATCH Trials: The Present and the Future

Speaker: Alice Chen, National Cancer Institute

Dr. Chen directs the Developmental Therapeutics Clinic at the NCI where she facilitates the discovery and development of new anticancer drugs and drug targets, and piloting novel clinical trial designs for development of molecularly targeted agents on a national and international basis. She is the director the Advance Developmental Therapeutic Training Program which trains medical oncologists in early phase drug development. She is the co-primary investigator for the largest precision medicine trial, NCI-MATCH, primary investigator for the Molecular Profiling-Based Assignment of Cancer Therapy and PI for one of the arms of ComboMATCH. She is the Chief Specialty Editor for Frontiers Medicine: Precision Medicine. Her other experiences include nine years as Senior Investigator at the Cancer Therapy Evaluation Program, DCTD, NCI handling a portfolio in DNA repair and antiangiogenic agents. She has interest in improving early phase clinical trials and led 2 revisions of the Common Toxicities Criteria for Adverse Events and is a member of the Response Evaluation Criteria in Solid Tumors committee. She has expertise in rare tumor and sarcoma, and is the primary investigator for several multicenter sarcoma studies testing immunotherapy and new novel agents. She has authored over 170 manuscripts and presented at major oncology meetings internationally on early phase clinical trials, precision medicine, DNA repair and CTCAE.

Abstract

Platform design is increasingly used in precision medicine to efficiently test multiple agents in one histology, multiple histologies with one agent or multiple histologies with multiple agents. The National Cancer Institute’s Precision Medicine Initiative supports clinical trials such as LungMAP, MPACT, NCI MATCH and alchemist to develop new treatment strategies, targeted therapies, and personalized medicine approaches. The next generation of platform trials aim to provide a more comprehensive and effective approach to investigating cancer treatment options. The NIH MATCH trial, which aimed to provide personalized treatment options for cancer patients based on their genetic profile, led to the development of a followup trial called ComboMATCH focused on investigating the effectiveness of combining multiple target therapies for solid tumors. Past Trials inform the design and effectiveness of new trials and the launch of ComboMATCH represents a promising step forward in precision medicine using platform trial design.

Lessons Learned and the Future of Lung-MAP

Speaker: Mary W. Redman, Fred Hutchinson Cancer Center

Mary W. Redman, Ph.D. is a Professor in Clinical Biostatistics in the Clinical Research Division at Fred Hutchinson Cancer Center. She has extensive experience in clinical trials, in particular, in phase II and III trials incorporating biomarkers particularly in lung cancer. Dr. Redman is the Statistical Chair for the Lung Cancer Committee in the SWOG Cancer Research Network, the head of the Biostatistics Core for the Fred Hutch Lung SPORE, and the Statistical Chair for the Lung-MAP Master Protocol, the first of the master protocols launched within the National Clinical Trials Network, and a trial that has served as a model for master protocols and the FDA guidance on master protocols.

Abstract

Lung-MAP is master protocol with an established infrastructure to screen patients with stage IV or recurrent non-small cell lung cancer for multiple integral biomarkers and conduct multiple independent studies of investigational treatments within biomarker-defined populations and a “non-match” population for patients not eligible for any of the biomarker-driven studies. The study, which is supported by a public-private partnership, was launched in 2014 and was the first master protocol within the National Clinical Trials Network of the National Cancer Institute. Lung-MAP has had to continually adapt to a quickly evolving treatment landscape for the disease its focused on treating. The current LUNG-MAP program consists of biomarker-driven studies for patients previously treated with a line of standard of care treatment for stage IV or recurrent disease with either TKI-naïve or TKI-resistant disease. These studies are focused on either a first line of targeted therapy combination for TKI naïve disease or targeting/overcoming resistance mechanisms for TKI-treated and resistant disease. Additionally, the program includes non-match sub-studies for patients previously treated with both platinum-based chemotherapy and immunotherapy focused on overcoming resistance to immunotherapy. In its almost decade long existence many lessons have been learned and there are also exciting future plans as this study continues to evolve.

A Sequential Multi-Source Adaptive Platform Design with Information Sharing and Future Directions for Oncology Platform Trials

Speaker: Alex Kaizer, Colorado School of Public Health

Dr. Kaizer is an Assistant Professor in the Department of Biostatistics & Informatics and a faculty member in the Center for Innovative Design & Analysis (CIDA) at the University of Colorado-Anschutz Medical Campus. He is passionate about translational research and the development of novel adaptive clinical trial designs that more efficiently and effectively utilize available resources, past trials, and past studies. Dr. Kaizer strives to translate complex statistical topics into understandable material that is more than "just math" and something we can appreciate and utilize in our daily lives and research.

Abstract

The potential strengths of platform trials have been elucidated for oncology, infectious diseases, and other settings. In this talk we first present the lessons learned from a sequential platform trial used in the West Africa Ebola virus disease outbreak and how these may translate to the context of oncology research. For example, one shortcoming of the original design was that supplemental information from controls in previous trial segments was not utilized. We address this limitation by proposing an adaptive design methodology that facilitates information sharing across possibly non-exchangeable segments using multi-source exchangeability models (MEMs). The design uses multisource adaptive randomization to target information balance within a trial segment in relation to posterior effective sample size. Compared to the standard design, we demonstrate that MEMs with adaptive randomization can improve power with limited type-I error inflation. We conclude by discussing some of the future directions and potential applications of newer methodologies and applications for platform trial designs in oncology that both include and extend the lessons learned previously.