Research Summary

My Ph.D research is in Biostatistics. My primary focus is clinical trials, specifically adaptive randomization in biomarker-driven trial design. The following is a more detailed overview of my research interest and objective.

With the rapid development of gene sequencing techniques, most cancers now have refined classifications recognizing tumor heterogeneity on a molecular level, and treatment strategies have shifted from eradicating tumors to interrupting their molecular drivers of growth and spread. Statistical designs for clinical trials, however, have struggled to keep pace with continually evolving diagnostic classifications and novel treatment strategies, despite their central role in testing therapeutic hypotheses and translating scientific findings into clinical practice. There is an urgent need for novel clinical trial designs that are capable using emerging information (data) regarding the prognostic and predictive effects of tumor molecular features to adapt trial endpoints, objectives, and patient accrual while the trial is still ongoing.

The ultimate purpose of my research is to develop novel statistical methodology to support “next generation” clinical trial designs for molecularly targeted cancer therapies. Specifically, I will develop a new statistical theoretical framework for biomarker-driven trial designs in which one or more naturally continuous markers with possibly non-linear or non-monotone effects are of interest to drive adaptive decision-making at interim and final analysis timepoints. These designs will allow for more accurate and precise real-time identification of patient subgroups who are benefiting from an experimental targeted therapy due to their avoidance of standard design manipulations and assumptions (e.g., effect linearity or monotonicity). Instead, the goal is to create a flexible Bayesian modeling “engine” will use accumulating patient, molecular, and outcome data to continuously update estimates of design-driving marker relationships, quantify posterior uncertainty around these effects, and translate the statistical information to personalize treatment allocation.

Research Projects

Publications

Presentations