|
Activity Number:
|
128
- SPEED: Biometrics and Biostatistics Part 1
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 29, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #301805
|
|
|
Title:
|
Adaptive Design with Biomarker Population Deselection and Enrichment for Oncology Trials
|
|
Author(s):
|
Pingye Zhang* and Yue Shentu and Qi Liu
|
|
Companies:
|
and Merck & Co., Inc. and Merck & Co., Inc.
|
|
Keywords:
|
Subgroup;
Adaptive design;
De-selection
|
|
Abstract:
|
Subgroup analyses are becoming increasingly important in Oncology trials as more and more therapies showed evidence to be more beneficial to a subgroup defined by a certain biomarker. Subgroup analyses could serve as an alternative assessment of evidence from confirmatory trials in which conclusions for the overall population may not hold. Subgroup analyses can have different purposes. Here we focused on the framework that a study treatment in a clinical trial is thought to be particularly beneficial to a certain subgroup of patients and may not be performing for another subgroup of patients. We would like to test the treatment effect for the overall population at a reduced alpha level so that treatment effect can be tested in subgroups of interest with fractions of total alpha level saved from the overall population analysis. We propose a new testing procedure that allows researchers to have a futility check on the potentially non-performing subgroup and de-select the non-performing subgroup early in the trial. We used simulations to confirm the Type I error rate for the new procedure and compare the power of it to some existing procedures.
|
Authors who are presenting talks have a * after their name.
