Randomized Phase II Cancer Clinical Trials
Purchasing Options
Features
- Evaluates the design and analysis methods of optimal single-arm phase II trials on response rate and progression-free survival
- Covers various statistical methods for randomized phase II trials, including Fisher’s exact test and two-sample log-rank test
- Describes flexible phase II clinical trial designs for comparing survival distributions between maintenance therapies
- Includes many examples throughout and references at the end of each chapter
Summary
In cancer research, a traditional phase II trial is designed as a single-arm trial that compares the experimental therapy to a historical control. This simple trial design has led to several adverse issues, including increased false positivity of phase II trial results and negative phase III trials. To rectify these problems, oncologists and biostatisticians have begun to use a randomized phase II trial that compares an experimental therapy with a prospective control therapy.
Randomized Phase II Cancer Clinical Trials explains how to properly select and accurately use diverse statistical methods for designing and analyzing phase II trials. The author first reviews the statistical methods for single-arm phase II trials since some methodologies for randomized phase II trials stem from single-arm phase II trials and many phase II cancer clinical trials still use single-arm designs. The book then presents methods for randomized phase II trials and describes statistical methods for both single-arm and randomized phase II trials. Although the text focuses on phase II cancer clinical trials, the statistical methods covered can also be used (with minor modifications) in phase II trials for other diseases and in phase III cancer clinical trials.
Suitable for cancer clinicians and biostatisticians, this book shows how randomized phase II trials with a prospective control resolve the shortcomings of traditional single-arm phase II trials. It provides readers with numerous statistical design and analysis methods for randomized phase II trials in oncology.
Table of Contents
Introduction
Single-Arm Phase II Trial Designs
Single-Stage Designs
Two-Stage Designs
Two-Stage Designs with Both Upper and Lower Stopping Values
Inference on the Binomial Probability in Single-Arm Multistage Clinical Trials
Point Estimation
Confidence Intervals
P-Values
When Realized Sample Size Is Different from That Specified in Design
Single-Arm Phase II Clinical Trials with Time-to-Event Endpoints
A Test Based on Median Survival Time
Maximum Likelihood Method for Exponential Distribution
One-Sample Log-Rank Test
Two-Stage Trials Using One-Sample Log-Rank Test
Binomial Testing on t-Year Survival Probability
Single-Arm Phase II Trials with Heterogeneous Patient Populations: Binary and Survival Outcomes
Binary Outcome Case
Survival Outcome Case: Stratified One-Sample Log-Rank Test
Randomized Phase II Trials for Selection: No Prospective Control Arms
With a Historical Control
When No Historical Control Exists
Extension to More Than Two Arms
Randomized Phase II Cancer Clinical Trials with a Prospective Control on Binary Endpoints (I): Two-Sample Binomial Test
Two-Sample Binomial Test
Two-Stage Designs with Both Upper and Lower Stopping Values
Discussions
Randomized Phase II Cancer Clinical Trials with a Prospective Control on Binary Endpoints (II): Fisher’s Exact Test
Single-Stage Design
Two-Stage Design
Extensions
Discussions
Randomized Phase II Trials with Heterogeneous Patient Populations: Stratified Fisher’s Exact Test
Single-Stage Stratified Fisher’s Exact Test
Two-Stage Designs with an Interim Futility Test
Randomized Phase II Clinical Trials Based on Survival Endpoints: Two-Sample Log-Rank Test
Two-Sample Log-Rank Test
Two-Stage Log-Rank Test
Stratified Two-Sample Log-Rank Test for Single-Stage Designs
Some Flexible Phase II Clinical Trial Designs
Comparing Survival Distributions under General Hypothesis Testing
Randomized Phase II Trials for Comparing Maintenance Therapies
Index
References appear at the end of each chapter.
Author Bio(s)
Sin-Ho Jung is a professor of biostatistics and bioinformatics at Duke University School of Medicine. He earned a PhD from the University of Wisconsin-Madison. His research interests include clinical trial design, survival analysis, longitudinal data analysis, clustered data analysis, ROC curve analysis, and microarray studies.
