Proceedings of the University of Pennsylvania Annual Conference on Statistical Issues in Clinical Trials: Early, translational and proof of concept studies: The ‘Go/No Go’ Decisions

Abstract

Conference Funded by the National Cancer Institute, with Co-Sponsorship of the American Statistical Association and the Society for Clinical Trials There is a continuing need for the development of scientifically sound, efficient, ethical, and safe methods for the assessment of human interventions. The National Cancer Institute of the National Institutes of Health awarded a five-year Conference Grant (R13 CA132565-01) to the Division of Biostatistics in the Center for Clinical Epidemiology and Biostatistics at the University of Pennsylvania School of Medicine to establish a conference series that would facilitate advances in approaches to the statistical design, implementation, and biostatistical analysis of clinical trials, through a continuing series of targeted symposia. The conference series is co-sponsored by the American Statistical Association and the Society for Clinical Trials. Each symposium will focus on a specific topic area within the commonly accepted definition of stages in the development of an intervention. The topic for each symposium will be chosen annually, based on the most pressing topics in current research, suitability of an area to benefit from a concentrated symposium, and its potential impact on the practice of clinical trials. We expect that the symposia will address problems and produce solutions that translate readily across disease areas, which historically has been the case for biostatistical methodology, regardless of the disease area that inspired its development. Every symposium will examine methodological approaches in the context of the clinical, ethical, and practical environments attendant to the problems being considered. The full title of the conference grant is From Bench to Bedside to Community: Statistical Issues in the Early to Late Assessment of Interventions. This title represents the influence of the Clinical and Translational Science Awards (CTSA) program under the NIH Roadmap. It is our intent to incorporate into the proposed symposia a key principle of the CTSA Program, that is, to break down traditional research barriers to achieve efficient and ethical science. Our specific objectives are to develop a series of clinical trials symposia, each focused on an area within one of the five stages of assessment of an intervention; bring together leading scientists for presentation and discussion of the state of the art and expected future development in these areas; to have pre-conference review by organizers in order to establish controversy and equipoise in the targeted area and incorporate such into program; to seek an audience with both expertise and commitment to participate in the symposia; and to publish the proceedings of the symposia, including discussion, as quickly as possible in a peer-reviewed journal. Our expectation is that meeting these objectives for the conference series will provide more than just a transfer of information, allowing for an expansion of the available state-of-the-art information through the contrast of ideas and the elements of discussion. We strongly believe that the nature of clinical trials methodology is not disease specific, although there will be a heavier use of certain methodologies in some areas of medicine (e.g. both cancer and cardiovascular clinical trials might be more oriented towards time-to-event outcomes and their attendant design and analytic approaches). Over a five-year grant period, the intent is to develop five annual single day symposia. Each symposium will cover a single topic roughly within each stage, covering all five stages within the five-year period. The scientific need for this series of symposia is strong. The scientific community faces increasing demands for efficient (with respect to both financial and participant resources), timely, and valid approaches to the assessment of new interventions. The streamlining of drug development is receiving increased emphasis with the advent of the CTSAs and FDA's Critical Path Initiative, and an increased focus on the complete and coordinated spectrum of such investigations from laboratory to population safety assessment. These symposia will provide for biostatisticians involved in biomedical research an organized synthesis of what is needed, what is available, and what needs to be developed. The program for the inaugural symposium, from which the subsequent papers were presented, was entitled 'Early translational and proof of concept studies—the ‘Go/NoGo’ decisions' and covered an overlap of the topics in the stages of early human and doses finding studies, and early efficacy and transition studies. Peter Gilbert from the University of Washington and The Fred Hutchinson Cancer Research Center points out that Clinical trials programs are increasingly following successful Phase 1/2 trials with focused intermediate-sized Phase 2b efficacy trials, instead of a traditional full-scale Phase 3 trial. These Phase 2b trials may be designed to directly contribute to the evidence-base for licensing a product, or, to test a scientific concept for moving the field forward, acknowledging that the particular product may never be licensable. Using the HIV vaccine field as a case study, he compares and contrasts these Phase 2b designs to each other and to a Phase 3 design, and describes a decision-theoretic framework for guiding the optimal choice of trial type and sample size for an initial efficacy trial. This framework may be helpful for designing initial efficacy trials. Lawrence Rubinstein, Seth Steinberg, Shivanni Kummar, Robert Kinders, Ralph Parchment, Anthony Murgo, Joseph Tomaszewski, and James Doroshow from the National Cancer Institute argue that a phase 0 trial can be used as a first-in-man proof of principle study to assess the effectiveness of a new agent against a molecular target, by means of a pharmacodynamic (PD) assay, in a very small number of patients. The dosage is meant to be pharmacologically active, but not toxic. Such a trial may be used for very early weeding out of ineffective agents, as a test of pre-clinical findings, and as an opportunity for clinical development of a PD assay. They present designs for such PD-driven studies which are statistically efficient and rigorous and conclude that the phase 0 trial promises to become an increasingly important tool for facilitating and speeding the development of new therapeutic agents, particularly in oncology. Sumithra Mandrekar, Rui Qin, and Daniel J. Sargent from the Mayo Clinic discuss novel therapies with specific biologic targets, non-monotone dose–efficacy curves, and mild toxicity profiles that mandate innovative clinical trial designs to identify biologically optimal doses in early phase trials. Two novel designs based on modeling the relationship between dose and efficacy and dose and toxicity have been recently proposed: proportional odds model and continuation ratio model. These models are based on the three-state nature of clinically observed dose outcomes for most novel agents: neither toxicity nor efficacy, acceptable toxicity with efficacy, and unacceptable toxicity. The theoretical framework, computational difficulties, and design issues are reviewed. Simulation studies demonstrate considerable promise with favorable operating characteristics for the two model-based adaptive designs. Suggestions based on their experience are offered to overcome the pragmatic challenges that hinder the application of such model-based designs in practice.