Home Page >
Research areas >
Methodology >
Applied stat methodology >
Design and analysis of trials involving biomarkers
Design and analysis of trials involving biomarkers
Evaluating stratified approaches to
treatment, typically according to the value of one or more
biomarkers, is an important area of medical research. This is
particularly in cancer where personalised medicine is becoming a
real possibility.
At the CTU, examples of such biomarker
trials are FOCUS3 and
FOCUS4 (currently in design) in colorectal cancer. We are planning
to run further trials, but there are competing approaches to
evaluating strategies based on biomarkers.
One consists of randomising patients to one
of two or more treatment strategies according to their values of a
biomarker. An alternative design is to randomise patients to
each of two or more treatments, independently of the biomarker, and
then analyse the treatment effects according to the value of the
biomarker (taken at baseline).
In this area, we are reviewing the different
approaches, with the aim of making recommendations to researchers
to fill any identified gaps in the methodology. If required, we
will carry out simulation studies to identify risks and benefits
(power and false positive rates) of the different approaches under
feasible scenarios.
The speed of science means that potential
new biomarkers are being discovered each year, and in the lifetime
of a long-term trial, the biomarker that led to the trial design
may not ultimately be the most appropriate one for the therapy
under investigation. Therefore, one of the key practical issues is
to try to develop designs of biomarker trials that are
‘future-proof’.
Key
projects
- Review of competing trial designs, and
approaches to sample size calculation
- Evaluation and possible further development
of statistical methods for the analysis of biomarker trials,
particularly in terms of new biomarkers emerging during the course
of the trial
