Amgen Symposium: Statistical Innovations in Clinical Trials

Cytel, Amgen and PSI brought together UK-based Industry and academia trial design leaders on June 29th, 2015.

Download the slidesets:

Case Study of an SSR Adaptive Trial: The VALOR Trial , Cyrus Mehta, Cytel

Statistical Issues in the Design Of Multi-Arm Multi-Stage Trials, James Wason, MRC Biostattistics Unit

Tackling Real Problems in Multi-Regional Trials, Kevin Carroll, KJCStatistics

Defining, Understanding and Communicating Decision Criteria in Early Development, Peter Colman, UCB

Case Study of an Adaptive Trial for AML: The VALOR Trial

The recently completed VALOR trial comparing vosaroxin to cytarabine in acute myeloid leukemia (ASH 2014) accrued 711 patients, and comprises the largest body of evidence for AML from a randomized Phase 3 trial. This adaptive event driven trial was designed with an interim analysis that allowed for early efficacy stopping, early futility stopping, unblinded sample size re-estimation, or continue as planned. The sample size re-estimation option was implemented. We'll present trial results from the trial and discuss operational, regulatory and statistical challenges faced. To our knowledge this is the first Phase 3 confirmatory oncology trial incorporating unblinded sample size re-estimation.

Cyrus Mehta

Cytel Inc.

Cyrus Mehta, Ph.D., is President and Co-Founder of Cytel and Adjunct Professor at the Harvard School of Public Health. A widely respected biostatistician, Cyrus has made lasting contributions to the fields of exact and Monte Carlo methods, adaptive and group sequential clinical trial designs, and simulations based methods for enrollment and feasibility. Cyrus is the driving force behind East®, the leading software for modern trial design. He is a fellow of the American Statistical Association and recipient of its Snedecor Award. He's been elected to the International Statistical Institute. Cyrus has published over 80 articles in journals such as Nature, JASA, Biometrics, Biometrika, and Statistics in Medicine.

Statistical Issues in the Design Of Multi-Arm Multi-Stage Trials

Multi-arm trials are increasingly being recommended for use in diseases where multiple experimental treatments are awaiting testing. This is because they allow a shared control group, which considerably reduces the sample size required compared to separate randomised trials. Further gains in efficiency can be obtained by introducing interim analyses (multi-arm multi-stage, MAMS trials). At the interim analyses, a variety of modifications are possible, including changing the allocation to different treatments, dropping of ineffective treatments or stopping the trial early if sufficient evidence of a treatment being superior to control is found. In this talk I describe each of these designs and provide some comparison in how they perform. I will also discuss some future research directions, including the use of predictive biomarkers in multi-arm trials.

James Wason

MRC Biostatistics Unit, Cambridge

Dr James Wason is a senior investigator statistician at the MRC Biostatistics unit (BSU) in Cambridge. He has been there since doing his Ph.D. in 2006. Since 2009 he has worked in the BSU’s Hub for Trials Methodology Research, directed by Adrian Mander. His main research interests are adaptive designs for clinical trials, efficient analysis of composite endpoints, and the use of biomarkers in clinical trials. He is also the co-lead of the MRC Hub for Trials Methodology’s stratified medicine working group.

Tackling Real Problems in Multi-Regional Trials

MRCTs are an increasingly necessary feature of modern drug development. In areas like diabetes and cardiovascular disorders, or early adjuvant oncologic disease settings, or where there is a regulatory requirement to rule out small, but important safety issues such as in the ongoing cluster of trials investigating long acting beta agonists in the treatment of asthma, demand for very large trials drives the need for MRCTs. They are are needed to provide the power to address the underlying hypothesis of interest, but can only do so under the assumption of no true regional heterogeneity. I review the implications of true regional differences in MRCTs and illustrate the real statistical and regulatory challenges faced by reference to recent case studies, including the 18,000 patient PLATO trial in ACS.

Kevin Joseph Carroll

KJCStatistics Ltd.

Kevin Joseph Carroll, PhD, CStat, CSci, Honorary Senior Lecturer Medical Statistics is an Independent Statistical Consultant and owner of KJCStatistics Ltd. Kevin has 26 years drug development of experience across all trial phases and multiple therapeutic areas including Oncology, CV, Metabolism, Respiratory, CNS and GI. Most recently Kevin held the positions of VP Statistics and Chief Statistician at AstraZeneca Pharmaceuticals and Expert Statistician with Boehringer-Ingelheim. Kevin currently helps both pharma and biotechs tackle statistical issues, including the application of innovative approaches to design and analysis and regulatory product license applications.

Defining, Understanding and Communicating Decision Criteria in Early Clinical Development

It is not uncommon to investigate multiple indicators of potential clinical efficacy in an early study in patients. A broad swathe of biomarkers may be nominated to ensure that evidence of activity on the projected biochemical pathway is acquired and additional biomarkers may indicate the precursors of clinical benefit. Registration endpoints, surrogates thereof or less-qualified biomarkers of efficacy may also be captured. The number of endpoints may easily exceed the number of subjects in the study and so it is vital that the properties of any decisions are well understood and consistently communicated. We discuss one approach which elicits views on the combinations of results that would be considered positive or encouraging and seeks to assess the likely false-positive rates associated with the consequent decision criteria. We illustrate the ideas with some real examples.

Peter Colman

UCB

Peter spent the first 29 years of his statistical career with Pfizer at their UK research site in Sandwich. During this time he worked in Animal Health, Clinical Pharmacology and Early Clinical Development. He also experienced a phase III project for 6 months. He led a group of statisticians focussing on PK-PD Modelling & Simulation, Genetics, Clinical Technology and Outcomes Research. He returned to mainstream, hands-on project work and also contributed to a number of European collaborations (e.g. IMI SAFE-T). In 2011, upon closure of the Sandwich site, Peter moved to AZ at Alderley Park. He joined the UCB early development statistics team at UCB in 2013 where he contributes to the design of studies in the immunology and neuroscience areas.