The 3rd Stat4Onc Annual Symposium
Emerging new approaches in early-phase oncology drug
Speaker Biography and Abstract
Richard Patt, M.D.
Patt is co-founder and principal in RadMD, and is a board-certified radiologist
with > 25 year’s experience customizing imaging endpoints for oncology
trials and training and managing performance of an independent reader group
that has performed reviews for over 500 oncology trials. He has served as head
of MRI, Georgetown University Medical Center, and Director, Imaging Clinical
Development at Berlex Labs. He has also co-founded The
Blinded Reader and Investigator Training Institute (BRITI) which focuses on
web-based training of research sites on imaging efficacy criteria for hundreds
of research sites, site readers, and trial personnel globally. He has special
interest in advanced reader training and performance management methods, and
utilizing imaging to better define drug mechanism of action in early phase
Imaging AI, Radiomics, and Live Central Image Reviews are Transforming Early
blinded independent central review was generally utilized in later phase trials
as surrogate endpoints of efficacy. It was the results of site image
interpretations, however, that was used for early signal detection and go no-go
decision making. The added complexity of imaging immune-oncology agents,
combined with new methods of evaluating images for early efficacy signals (radiomics
and artificial intelligence (AI) resulted in greater demands on image
reviewers. This presentation will provide an overview of how centralized image
review is changing early phase oncology development.
Chang-Heok Soh, Ph.D.
has more than 16 years of experience in medical research and
biotech/pharmaceutical industry, spanning early- to late-stage drug
development. She is currently Head of
Early Oncology Statistics at AbbVie, providing leadership to the early-stage
oncology statistics groups at AbbVie’s sites in California and Illinois.
to joining AbbVie, she was Director of Biostatistics at Alnylam Pharmaceuticals
in Cambridge, Massachusetts, helping to advance RNA interference therapies for
multiple rare diseases. Before Alnylam,
Chang-Heok worked at Genentech/Roche for 10 years, providing statistical
leadership on key oncology and Alzheimer’s disease programs, including
successful regulatory filings. Before
expanding her career in biotech/pharma industry, Chang-Heok was involved in
pediatrics AIDS research in research institute setting and diverse disease
areas in hospital setting. Her career
has spanned geographic locations in USA, Europe and Asia, including 2.5 years
in Switzerland representing Genentech/Roche Biostatistics group in internal and
cross-industry collaborations in Alzheimer’s disease.
received her master’s and Ph.D. degrees in Biostatistics from Harvard
Interim Monitoring for Faster Decision-Making in Early Oncology Trials
approaches for performing interim analysis of early phase oncology trials
typically involve examining the data at specified sample size(s). There is often no formal mechanism if the actual
timing of the interim analysis deviates from plan. For example, such situations may arise in
practice when there is a need to perform an interim analysis before the
specified sample size due to slow accrual on the clinical trial.
Bayesian interim monitoring approach allows more flexibility in the timing of
interim analysis and could accelerate decision-making in early-stage oncology
trials. Simulations show high level of
concordance between the decision made at interim analysis and the decision that
would have been made should the trial continue to its planned end.
Ananthakrishnan works as a Biostatistician at Celgene on designing, analyzing
and interpreting immuno-oncology trials.
has a broad interdisciplinary background of math, statistics, physics and
biology and is interested in various aspects of Oncology clinical trials. She has worked on several early phase
Oncology trials as well as trials for regulatory submission for solid tumors as
well as blood cancers.
of the TEQR and mTPI designs including non-monotone efficacy in addition to
toxicity in dose selection
the emergence of immunotherapy and other novel therapies, the traditional
assumption that the efficacy of the study drug increases monotonically with
dose levels is not always true. Therefore, dose-finding methods evaluating only
toxicity data may not be adequate. Hence, this talk will cover three new early
phase designs that consider efficacy in addition to safety in dose selection.
The first two designs are the extended TEQR and mTPI designs – in these
designs, the optimal dose for safety and efficacy is determined by applying
isotonic regression to the observed toxicity and efficacy rates, once the early
phase trial is completed. The third design is the 2D TEQR design, the
frequentist counterpart of an existing Bayesian design called the TEPI
(Toxicity Efficacy Probability Interval) design. We conduct simulation studies
to investigate the operating characteristics of the proposed designs for
various underlying DLT and response rates and compare them to existing designs.
We found that the extended mTPI design selects the optimal dose for safety and
efficacy more accurately than the other designs considered for most of the
scenarios considered. Although for the same sample size and cohort size, the
frequentist 2D TEQR design is less accurate than the Bayesian TEPI design in
selecting the optimal dose, the accuracy of optimal dose selection of the 2D
TEQR design can be increased, in many cases, with a moderate increase in cohort