Chair: Shibing Deng, PhD (Pfizer)
Abstract: Biomarkers have been increasingly used in various stages of drug development. Recent advances in digital technology as well as modern analytics, such as advanced imaging, digital health device and AI, have offered new opportunities to utilize biomarkers in drug development. At the same time, these advances have also presented considerable regulatory and statistical challenges and opportunities in the design and analysis of clinical trials. This session will focus on methodological development pertaining to biomarker use to inform and accelerate drug development, with special emphasis on biomarker-guided clinical trial design.
Speaker: Dai Feng, PhD (Abbvie)
Title: Statistical Considerations for Biomarker Real-World Studies Supporting the Launch of New Oncology Treatments
Abstract: The selection of biomarker-specific patient populations is essential in targeted cancer therapies to enhance precision and efficacy. To ensure a successful launch, it is vital to promote awareness and adoption of biomarker testing at diagnosis, tailor implementation strategies to accommodate local variations, and ensure testing is accessible and reimbursed. An integrated evidence generation plan should address critical questions, including the prevalence of biomarker expression and agreement between local and central labs, across different platforms and pathologists. Furthermore, understanding prognostic effects and associations with other biomarkers is of considerable interest.
Real-world studies play a key role in generating the evidence needed to answer these questions. However, they pose challenges, such as mitigating confounding factors, addressing biases like immortal time bias, and handling missing data. Additional complexities involve assessing agreement across correlated factors and dealing with low expression prevalence. In this presentation, we will explore key statistical considerations for overcoming these challenges, focusing on practical strategies for robust study design and analysis to ensure reliable results.
Speaker: Kaiyuan Hua, PhD (FDA, CDRH)
Title: Biomarker-Guided Adaptive Design with Threshold Detection and Patient Enrichment of the Restricted Mean Survival Time
Abstract: Biomarker-guided designs are increasingly used to evaluate personalized treatments based on patients’ biomarker status in Phase II and III clinical trials. With adaptive enrichment, these designs can improve the efficiency of evaluating the treatment effect in biomarker-positive patients by increasing their proportion in the randomized trial. While time-to-event outcomes are often used as the primary endpoint to measure treatment effects for a new therapy in severe diseases like cancer and cardiovascular diseases, there is limited research on biomarker-guided adaptive enrichment trials in this context. Such trials almost always adopt hazard ratio methods for statistical measurement of treatment effects. In contrast, restricted mean survival time (RMST) has gained popularity for analyzing time-to-event outcomes because it offers more straightforward interpretations of treatment effects and does not require the proportional hazard assumption. This paper proposes a two-stage biomarker-guided adaptive RMST design with threshold detection and patient enrichment. We develop sophisticated methods for identifying the optimal biomarker threshold and biomarker-positive subgroup, and approaches for type I error rate, power analysis, and sample size calculation. We present a numerical example of re-designing an oncology trial. An extensive simulation study is conducted to evaluate the performance of the proposed design.
Speaker: Mark Chang, PhD (Boston University)
Title: The Intransitivity of One-Sided Rank Tests, Its Impacts for Biomarker-Enabled Clinical Trial Designs, and Resolutions
Abstract: Intransitive dice are a classic example from game theory, illustrating a paradoxical circular winning relationship. Consider three dice: A ({2, 2, 4, 4, 9, 9}), B ({1, 1, 6, 6, 8, 8}), and C ({3, 3, 5, 5, 7, 7}). Rolling these dice, A wins over B, B wins over C, and C wins over A, each with a probability of 56%, creating an intransitive system. Now, imagine these dice represent three medical interventions, with face values corresponding to patient responses. In a randomized trial comparing A and B, standard rank-based methods might suggest that A outperforms B, even if the treatments are equally effective in a broader sense.
Rank-based statistical tests, such as Wilcoxon tests, win-ratio tests, and log-rank tests, often rely on “winning” concepts. This raises critical concerns about their transitivity properties when applied in clinical trial designs. We investigate the prevalence of intransitivity under these methods, identifying sufficient conditions for achieving transitivity—conditions that are rarely met in practice.
To address these challenges, we propose a framework that introduces the concept of “Probabilistically Preferable” outcomes, ensuring transitivity through a sufficient condition: Pr(A>B)>0.618, the “golden point.” Furthermore, we explore AI-driven precision medicine strategies based on the similarity principle to refine decision-making in biomarker-enabled trials, demonstrating its potential through a practical example. This work underscores the importance of examining and addressing intransitivity issues in rank-based analyses to enhance the reliability of clinical trial results.
