Presentation on theme: "Role of the Statistician and the Bioinformatics Scientist in Cancer Clinical Trials Larry Rubinstein, PhD Biometric Research Branch, NCI International."— Presentation transcript:
Role of the Statistician and the Bioinformatics Scientist in Cancer Clinical Trials Larry Rubinstein, PhD Biometric Research Branch, NCI International Clinical Trial Workshop ASCO, FLASCA, NCI, ONS Cordoba, Argentina September 11, 2014
Financial Disclosure I have nothing to disclose.
Oncology Statisticians in the U.S. There are 100 statisticians working at the NCI. There are 5 oncology cooperative research groups in the U.S. (4 adult, 1 pediatric). Each has a statistical center with approximately 20 statisticians, most of them PhD’s. The NCI Biometric Research Branch (NCI/BRB) has 11 PhD statisticians involved in clinical trials, including translational research, imaging and genomics, and 7 bioinformatics scientists.
Design of Cancer Clinical Trials The cooperative group statisticians write the statistical sections, collaborating with the PI’s. The disease oriented steering committees have monthly teleconferences to review concepts for randomized phase 2 and phase 3 trials, for which there are formal statistical reviews by NCI and extramural statisticians. The NCI Cancer Therapy Evaluation Program (NCI/CTEP) has weekly meetings to review phase 1, phase 2, and phase 3 protocols in concept and final form.
Design and Conduct of Clinical Trials The NCI/CTEP weekly meetings include formal reviews of all protocols by the NCI statisticians, sometimes including separate translational research and imaging reviews. Once initiated, Data Monitoring Committees review the progress of randomized phase 2 and phase 3 trials, stopping early for futility or dramatic efficacy, assessing adequacy of accrual, and making necessary amendments, all with close collaboration between NCI and extramural clinicians and statisticians.
Design and Conduct of Clinical Trials Each cooperative group has semi-annual meetings in which group clinicians and statisticians collaborate in the conduct of ongoing trials and the planning of new trials, with some NCI advisory collaboration. Group statisticians do independent analyses of completed trials and collaborate with clinicians on publications, with no NCI involvement.
Clinical Trials Statistical Methodology NCI and group statisticians have developed and published most of the major statistical methodology for the design, monitoring, and analysis of cancer clinical trials. NCI and group statisticians have developed and published statistical methodology for incorporating biomarkers into clinical trials, including those predictive of treatment effect and those prognostic of outcome, independent of treatment, both of which are increasingly relevant to trials of molecularly targeted agents.
Bioinformatics - Statistical Genomics Statisticians analyze the significance of gene mutations and expression in relation to patient outcome and response to particular treatments. This involves new and special high- dimensional statistical methodology to avoid a high false discovery rate when a large number of variables are simultaneously explored. NCI/BRB has developed and maintained, over the last 15 years, a toolkit (BRB-ArrayTools) for genomic analyses, with over 15,000 users world-wide.