Chairs:
Zhaoyang Teng, PhD (FDA)
Yeh-Fong Chen, PhD (FDA)
Abstract: To be announced
Speaker: Zoe Hua, PhD (Servier Bio-Innovation)
Title: Methodological Aspects and Practical Application of Drug Quantitative Benefit-Risk Assessment, a Case Study
Abstract: The quantitative benefit-risk assessment (BRA) is an essential tool in evaluating new therapies, ensuring a balanced understanding of clinical benefits and potential risks for informed decision-making. We present the methodological aspects and practical challenges of quantitatively assessing the benefit-risk profile of a novel oncology drug, utilizing insights from a panel of external clinical experts.
A Multi-Criteria Decision Analysis (MCDA) model was applied through a step-by-step approach. A value tree was created in collaboration with clinical experts, identifying key criteria, including survival outcomes, safety measures, and quality-of-life measures. Scales and weights for these criteria were determined through expert elicitation using the swing-weighting method. The primary analysis was based on the Scale Loss Score (SLoS) model, with sensitivity analyses performed using alternative models such as the linear and product models to confirm the robustness of the BRA results.
Our findings showcase the practical application of these quantitative methods in oncology, demonstrating how they provide objective and transparent support for regulatory and clinical decision-making. They contribute to the growing body of evidence advocating the use of quantitative BRA methodologies to optimize therapeutic decision-making in complex clinical environments.
Speaker: Bo Fu, PhD (Astellas Pharma)
Title: A Regulator‑Aligned Joint‑Model, Preference‑Robust Framework for Quantitative Benefit–Risk Assessment in Early‑Phase Trials
Abstract: Early-phase trials often have many endpoints but few patients. Teams must balance short-term benefit signals with several safety concerns and make clear, defensible decisions. We present JM-qBRA, a practical framework that combines a simple Bayesian joint model (to borrow strength across mixed endpoints) with a preference-robust benefit–risk layer using Dirichlet-weighted SMAA. Each endpoint is translated into a 0–1 “value,” anchored by clinical targets. We then sample plausible weights to reflect uncertainty in preferences and report an acceptability index for each regimen plus tipping-point diagnostics that show how conclusions might change.
A class-informed safety gate excludes over-toxic options, and simulation-calibrated thresholds control the chances of advancing an inferior or too-toxic regimen. Outputs—an effects table, acceptability bars, and a safety panel—fit directly into CTD 2.5.6 and the FDA Benefit–Risk Framework. JM-qBRA complements model-assisted dose-escalation methods by serving as the selection and communication layer once safety-admissible regimens are identified. It is lightweight to implement and well suited for PoC and early expansion decisions across therapeutic areas.
Speaker: Wenquan Wang, PhD (Pfizer, Inc.)
Title: Application of Win Statistics in Benefit-Risk Assessment for Hemophilia
Abstract: Chronic diseases such as hemophilia are complex to study in clinical trials. Patients often face multiple challenges at once: symptoms that need relief, the risk of disease progression or serious flare-ups, tolerability issues of the treatment, and complexity of treatment administration. Balancing these different factors is not easy in assessing a new therapy.
A benefit-risk assessment of a novel therapy usually starts with the separate assessment of each critical benefit and risk endpoint on a population-level, followed by the value judgement of combination of the marginal assessments. Therefore, some drawbacks follow naturally. One might miss the association between the benefits and risks, e.g. if patients benefiting from the therapy are more or less likely to suffer the risks. Also, determining the importance of different benefits and risks relative to each other could introduce subjectivity into the assessment.
One family of non-parametric techniques, win statistics, have good small-sample properties. They are suitable for the analysis of prioritized endpoints including benefit-risk assessment. Ranking outcomes is a more straightforward process than weighting, and they can handle the challenge related to the association between outcomes relatively easily. One can also perform sensitivity analyses with varying priorities and thresholds to check the robustness of the data. Patient-centric endpoints, such as patient-reported outcomes, can also be incorporated in prioritized win statistics, which is of importance as more and more focus is given to this area.
In this presentation, win statistics are applied to benefit-risk assessment of a novel therapy in hemophilia through re-analyses of clinical trial data. We will discuss whether these statistics can help trial investigators, regulators, and payers better understand the benefit/risk profile, and make decisions that truly reflect patient priorities. We will also discuss the pitfalls to watch out for when prioritizing outcomes.
Speaker: Lola Fashoyin-Aje, PhD (Parexel)
Title: To be announced
Abstract: To be announced
Discussant: Yeh-Fong Chen, PhD (FDA)
