I-SPY2: Identifying Biomarker Signatures for Therapeutic Agents in Neoadjuvant Breast Cancer

• Activity Number: 332
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Donald Arthur Berry+ and Kyle Wathen* and Nebiyou Bekele and Laura Esserman
• Companies: MD Anderson Cancer Center and MD Anderson Cancer Center and MD Anderson Cancer Center and University of California, San Francisco
• Keywords: Adaptive design ; Bayesian predictions ; Adaptive randomization ; Biomarker signatures ; Phase 2 screening trials

Abstract

I-SPY2 in a phase 2 drug screening process designed to identify which therapies are effective in breast cancer and which patients benefit from them. Patients are characterized by the biomarker signatures of their tumors. We consider multiple drugs, adaptively randomizing against control therapy using Bayesian methods. Drugs can graduate to phase 3 at any time based on Bayesian predictive probabilities of success in phase 3, with indications of which subsets of patients will benefit. Drugs are dropped from consideration if they fail to show a benefit for any biomarker signature. The primary endpoint is response at surgery, 6 months after randomization. We use a longitudinal model to predict this endpoint based on interim imaging measurements of tumor size.