- ホーム
- > 洋書
- > 英文書
- > Science / Mathematics
Full Description
This book provides a comprehensive overview of statistical methodologies used in cancer clinical trials, presenting the essential statistical background and practical details needed to apply these methods in real-world studies. By bringing together key methodologies in one place, the book helps readers select appropriate statistical approaches based on the specific objectives and requirements of a given trial. The statistical methods are organized according to the phases of oncology drug development. For early-phase trials, the book presents a balanced and concise discussion of dose-escalation methods, including approaches that account for late-onset toxicity. It also introduces statistical frameworks for go/no-go decision-making in proof-of-concept studies and master protocols, covering methods such as Simon's two-stage design, LaLonde criteria, predictive power, and Bayesian hierarchical modeling. In addition, the book provides a comprehensive overview of statistical approaches for dose-optimization trials, which are becoming increasingly important in modern oncology drug development. For late-stage or submission trials, the book covers methods for sample size determination under both fixed (single-look) and group-sequential designs, including procedures for futility and efficacy monitoring. It describes the standard statistical techniques for analyzing time-to-event and binary endpoints, including the log-rank test, Cox proportional hazards model, and Clopper-Pearson confidence intervals. The book also discusses multiple testing procedures, including graphical testing strategies and adaptive design approaches such as sample size re-estimation using the promising zone approach, the Cui-Hung-Wang method, p-value combination methods, and the 2-in-1 design. Seamless trial designs and alternatives to the log-rank test in the presence of delayed treatment effects—along with their extensions to group-sequential settings—are also presented.
Beyond confirmatory trial methodology, the book covers important exploratory statistical methods for post-hoc analyses, including subgroup identification, evaluation of treatment effects by phase in multi-phase treatment regimens, and statistical adjustments for bias introduced by treatment switching or crossover. The book also discusses several contemporary topics in oncology trials, including the estimand framework, event projection, and quantitative decision-making strategies. Emerging areas such as the application of artificial intelligence and machine learning in clinical trials are introduced, along with modern trial designs such as basket and umbrella trials. Additional topics include group-sequential designs with multiple endpoints maturing at different time points, the use of real-world data and real-world evidence in cancer drug development, and the Q-TWiST method.
The primary audience for this book is statisticians supporting oncology clinical trials from first-in-human (Phase 1) through pivotal submission (Phase 3) studies. The book may also serve as a postgraduate-level textbook for students aspiring to pursue careers as clinical trial statisticians in the pharmaceutical, biotechnology, or regulatory sectors.
Contents
Section 1: Introduction.- Chapter 1: Overview of cancer trials.- Chapter 2: Application of Estimands framework.- Section 2: Statistical methods for Early phase trials.- Chapter 3: Dose-escalation designs.- Chapter 4: Proof-of-concept and Dose-optimization designs.- Section 3: Late phase.- Chapter 5: Designing pivotal cancer trial along with sample size.- Chapter 6: Complex designs.- Chapter 7: Non-proportionality and other challenges related to immunotherapy.- Chapter 8: Post-hoc analyses (specific to time-to-event endpoint).- Section 4: Additional topics.- Chapter 9: Quantitative decision making and Event projection.- Chapter 10: AI/ML methods and precision medicine.- Chapter 11: Basket and umbrella trials.- Appendices.
