Creating a translational statistical environment for biomarker discovery for I/O trialSITUATION
- Client is a global pharma initiating 30+ phase 1 trials in the Immune-oncology area; protocols including generating 10+ biomarker datasets at various time points to understand the tumor microenvironment and the peripheral blood biomarkers
- The problem statement: immense amounts of data was produced on biomarkers across multiple labs; there was no common environment integrated for biomarker data, so any biomarker data analysis involved manual data integration and programming
- Datasets included: EDC(Clinical), RNASeq, Cancer DNA Panels, flow cytometry, multicolor IFN, mplex ELISA for cytokines, IHC/Pathology/H&E, Radiology
- Can we predict which patients are likely to show early recurrence after first line therapy by combing multiple data-sets and with advanced analytics
- Which data-platform will enable us to seamlessly integrated NGS data, circulating tumor DNA data and PET-CT data?
- What advanced algorithms need to be developed?
- Conducted detailed interviews with translational and biomarker leads, biostats leads as well as medical leads to understand the data requirements
- Create product / solution requirements
- Evaluated vendors and selected one of them for this project
- Led the process of development of a pilot with our technology partner to test solution’
- Led two workshops with participation from client and their biopharma customers to refine value proposition and opportunity
- After a successful pilot that was limited to one study, the biomarker statistical solution being rolled out across a broader set of trials
