Transitioning from 3+3 to N-CRM at GlaxoSmithKline

• Activity Number: 30
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Sharon Murray*+ and Allison M. Florance
• Companies: GlaxoSmithKline and GlaxoSmithKline
• Keywords: N-CRM ; Oncology ; Early-phase designs

Abstract

The algorithm based 3+3 or a modified 3+3 design had been the standard for dose escalation trials in Oncology at GlaxoSmithKline (GSK). Recently, the New Continual Reassessment Method (N-CRM) (Neuenschwander et al, Stat Med, 2008) has been adopted as the primary design for monotherapy dose escalation trials. The N-CRM is a variation of the original CRM method that uses a two-parameter Bayesian logistic model for the dose-toxicity curve with overdose control via dose-specific posterior probabilities for excessive or unacceptable toxicity. This presentation will review our recent experience of implementation of the N-CRM design in four different drug development programs. Details will include the software used to evaluate the operating characteristics of the design, standardization of protocol text, awareness sessions provided for statisticians and clinicians, and standardization of results provided for discussion at dose escalation meetings. Collaboration with the pharmacokineticists and elicitation of the prior based on pre-clinical toxicity data will also be discussed.