Prediction of Phase III Survival Outcomes Using Phase II Surrogate Marker Results Through Modeling and Simulation

• Activity Number: 652
• Type: Topic Contributed
• Date/Time: Thursday, August 4, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yi Liu*+ and Mingxiu Hu and Shan Yang and Guohui Liu and Hongliang Shi
• Companies: Millennium: The Takeda Oncology Company and Millennium: The Takeda Oncology Company and Iowa State University and Millennium: The Takeda Oncology Company and Millennium: The Takeda Oncology Company
• Keywords: survival ; modelling ; simulation ; marker

Abstract

For oncology drug development, regulatory approval is typically based on overall survival (OS) from Phase III trials. However, due to long duration and large study size for OS trials, Phase II proof of concept studies often use shorter-term measures such as tumor response, progression free survival, and other biomarkers. Without information on the eventual endpoint of OS, decision making for launching a Phase III trial has been challenging and difficult, which is one of the reasons that led to high failure rates of phase III oncology trials. Thus a statistical tool to connect phase II results to phase III outcomes is highly desirable.

This paper presents how to predict survival time through parametric modeling and simulations based on Phase II data. Several parametric models and selection criteria are evaluated. The pros and cons for each model and model selection criterion are discussed. A recommendation is then made for the prediction of median survival time and hazard ratio, which helps phase III go/no go decision making to increase the probability of success. A couple of examples will also be given to illustrate the applications of the technique.