Using Novel Precision Medicine Tools to Identify Variability in Hematopoietic Cell Transplantation

Abstract

In cancer patients, chemotherapy can be personalized based on the characteristics of the tumor and the patient. This thesis focuses on using patient characteristics to personalize an old chemotherapy drug, specifically busulfan, in allogeneic hematopoietic cell transplant (HCT) recipients. For the past 30 years, busulfan doses are often personalized to a target plasma area under the concentration-time curve (AUC) using an individual patient’s pharmacokinetic characteristics, a process termed pharmacokinetic-guided busulfan dosing (PKbusulfan). Because of this narrow therapeutic index necessitating PKbusulfan and the extended time after initial regulatory approval (~70 years), busulfan is a unique exemplar of adapting the ‘learn – confirm’ paradigm used in phase I – III drug development to the ‘predict, learn, confirm – implement cycle’ for model informed precision dosing (MIPD) of a phase IV drug. We used real-world data combined with population pharmacokinetic modeling (Part 1) and various -omics tools (Parts 2 and 3) to improve the effectiveness and toxicity of busulfan.

Part 1 focuses on improving PKbusulfan. Chapter 1 was a ‘call to action,’ as it summarized the lack of evidence-based decisions for several aspects of PKbusulfan. Chapter 2 sought to harmonize the worldwide HCT community to one busulfan plasma exposure unit, which can capture busulfan AUC in HCT registry databases. Chapter 3 created an international proficiency program for each step in PKbusulfan using the unique resources within the Netherlands of the KKGT. Finally, Chapter 4 summarizes a busulfan population pharmacokinetic model that can estimate the initial intravenous (IV) busulfan dose and be used to enable MIPDbusulfan. Specifically, MIPDbusulfan would be real-time Maximum A posteriori Probability (MAP) Bayesian estimation of the individual patient’s pharmacokinetic parameters, incorporating a blend of individualized pharmacokinetic data and population parameter priors.

Part 2 describes our efforts to use novel precision medicine tools to predict IV busulfan clearance. In Chapter 5, we used the candidate gene approach to evaluate whether IV busulfan clearance was associated with polymorphisms in the genes regulating the predominant metabolizing enzymes involved in busulfan conjugation. We then evaluated if IV busulfan clearance could be predicted with endogenous metabolomic compounds (EMCs). We adapted the concept of pharmacometabonomics, using pre-dose EMC profiling to predict drug response, to using pre-dose EMC profiling to predict IV busulfan clearance. Chapters 6 and 7 used a retrospective cohort to evaluate if IV busulfan clearance could be predicted with a targeted panel of 200 identified EMCs (Chapter 6) or a global panel of 1885 unknown ions. In Chapter 8, we sought to validate the findings of Chapters 6 and 7 and identify new pathways associated with IV busulfan clearance in a new cohort of prospectively collected EMC and busulfan pharmacokinetic data.

In Part 3, Chapter 9 addresses if additional biomarkers, specifically EMCs, are associated with relapse (the effectiveness of busulfan and the allograft) or graft versus host disease (GVHD, a significant contributor to nonrelapse mortality in allogeneic HCT patients) in a prospective cohort of 84 patients receiving PKbusulfan.