Transcript Introduction - Challenges in development in Oncology

European Statistical meeting
on Oncology
Thursday 24th, June 2010
Introduction - Challenges in development
in Oncology
H.U. Burger, Hoffmann-La Roche
Some challenges in Oncology
• Early developments:
– Dose determination for safety and/or efficacy
– Proof of concept study designs
• Late development:
– PFS versus OS as endpoint
– Go-No Go decisions for phase III
– Personalized health care strategies
Early Developments in Oncology
• Since about 10 years the paradigm of developing
oncology drugs has changed
– More and more biological treatments are developed for oncology
indications
– Targeted therapies play a larger role (pathway and mechanism of
actions)
– Balance between safety and efficacy more important with more
effective treatments available
• Therefore, aspects which have not been so important in
the past become more dominant. This concerns
– Proof of concept for biologics or combination regimen
– Dose finding using alternative approaches (e.g. model-based
approaches)
The Classical development paradigm
in Oncology
• Phase I: Dose escalation to define the dose for
development as the highest tolerable dose (MTD, 3+3
design classically based on ~ 20 to 40 patients)
• Phase II: Proof of concept: “One” responder is sufficient
to prove anti tumor activity and to move into phase III
(based on ~ 40 patients)
• Phase III: Large randomized clinical trial to confirm
efficacy versus standard or in combination with standard
versus standard
Challenge: Proof of concept
• Classically, proof of concept for a cytotoxic therapy was
based on a single arm monotherapy study in phase II
where some tumor responses were sufficient to warrant
the compound to go into phase III
• What triggers today proof of concept ?
– For biologics, responses not necessarily expected
– For combination therapies, responses can originate from the
combination partner
=> Proof of concept more and more based on randomized
studies including time to event endpoints; increasing role
of the biomarker data
Challenge: Dose finding
• Dose finding originally based on MTD trials for safety
• CRM methods introduced with the potential to improve
the precision of such studies to determine a dose with a
certain toxicity threshold
• More flexible two-parameter Bayesian logistic models
developed to better characterize the dose-toxicity
relationship
• For most of biologic therapies in oncology, maximal
tolerated doses become irrelevant as therapeutic effects
are already achieved at lower doses
 Optimal biologic dose or dose range leading to phase
II dose finding studies
Late Developments in Oncology
• Go-No Go decisions become more complex
– More sophisticated methods used in early development. Efficacy
assessment not based on single responses observed anymore
– More competitive environment requiring new risk-benefit
assessments
• Discussion around phase III endpoints never ending
story
– Overall survival (OS) clinically most relevant but sometimes
difficult to observe and frequently leads to long studies
– What are suitable surrogate endpoints for OS
• Response rates ?
• MRD (minimal residual disease) ?
• Progression-free survival ?
Late Developments in Oncology
• Targeted therapies: Personalized health care major
development challenge in the future
• Three types of development scenarios
– New treatment for all comers
– New treatment only for a subset of patient (defined by biomarker)
– Right population unknown upfront
• For last case challenges in development potentially huge
Today's Session
08:45-09:15
09:15-09:30
09:30-10:10
10:10-10:50
10:50-11:15
11:15-11:55
11:55-12:35
Registration
Chair: Hans-Ulrich Burger (Hoffmann-La Roche)
Introduction – Challenges in early phase development of
Oncology, Hans-Ulrich Burger (Hoffmann-La Roche)
Understanding Progression-free Survival,
Bertil Jonsson, MD (Medical Products Agency)
IRESSA: A Journey of Experience from Broad to Biomarker
Populations, Claire Watkins (AstraZeneca)
Coffee
Chair: Pierre Verweij (Merck and EFSPI)
Adaptive Bayesian Designs for Phase I Oncology Trials,
Stuart Bailey (Novartis)
Oncology Dose Finding – A Case Study,
Jonas Wiedemann (Hoffmann-La Roche)
Today's Session
12:35-13:35
13:35-14:15
14:15-14:55
14:55-15:20
15:20-16:00
16:00-16:40
Lunch
Chair: Nigel Howitt (PRA International and EFSPI)
Bayesian Hierarchical Modelling of Clinical Response in
NSCLC Subpopulations, Simon Wandel (Novartis)
The Time to Progression Ratio for Phase II Trials of
Personalized Medicine,
Marc Buyse (IDDI and University of Hasselt)
Coffee
Chair: Hans-Ulrich Burger (Hoffmann-La Roche)
Optimal Cost-Effective Go-No Go Decisions in Late-Stage
Oncology Drug Development, Cong Chen (Merck)
Panel Discussion