Using the University of California Clinical Data Warehouse to Evaluate the Safety and Efficacy of Biologics, including Blood Transfusions, Vaccines, CAR-T Cell Therapies, and Other CBER Regulated Products

CERSI Collaborator: University of California San Francisco: Atul Butte, MD, PhD; Michelle Wang, PharmD (UCSF)

FDA Collaborators: Center for Biologics Evaluation and Research: Barbee Whitaker, PhD; Artur Belov, PhD; Steven Anderson, PhD, MPP; Hui-Lee Wong, PhD; Angela Moy

Project Start Date: 12/2019
Project End Date: 09/2021

Regulatory Science Challenge

This project meets the third regulatory science objective of developing methods and tools to improve and streamline clinical and post-marketing evaluation of FDA-regulated products. Novel approaches are needed to support the assessment and generation of real-world evidence for the safety and effectiveness of biologic products, including the novel chimeric antigen receptor T cell (CAR T cell) therapies, blood products, as well as vaccines. The methods and real-world evidence (RWE) developed in this project may support future safety monitoring and regulatory decision making. This project is broken down into two subprojects focusing on 1) developing novel computational methods to generate RWE for the novel CAR T cell therapies and 2) developing novel computational approaches to identify rare and severe adverse reactions associated with blood product transfusions.

Project Description & Goals

The CAR T study aimed to characterize the real-world safety and effectiveness of CAR T cell therapies using patients from across the University of California (UC). The blood transfusion study aimed to develop novel approaches to identify adverse reactions and study potential risks associated with these severe adverse reactions.

Research Outcomes/Results

The safety and efficacy of CAR T cell therapy have been assessed in real-world patients using UC health clinical data warehouse. Risk factors and biomarkers associated with effectiveness and severe adverse reactions have been explored. A machine learning model has been developed to show potential to predict patient-level outcomes of CAR T cell therapy based on patients’ characteristics prior to receiving the therapy.

An algorithm has been developed to automate extraction of blood transfusion adverse reactions from clinical notes. We assessed rates of blood transfusion adverse reactions at UCSF, characterized clinical presentations of transfusion-associated circulatory overload, a severe adverse reaction associated with the highest mortality. Risk factors potentially associated with transfusion-associated circulatory overload were explored to potentially help inform clinical care of high-risk patients receiving blood transfusions.

Research Impacts

The CAR T cell therapy project resulted in findings that may provide critical insights into real-world utilization, patient-level characterizations, and outcomes of novel FDA-approved therapies such as CAR T cell therapies. This research work achieved information dissemination through conference presentations and a manuscript submitted for publication.

The blood transfusion reaction project provided an updated assessment of incidence rates, risk factors, and clinical implications to prevent transfusion associated circulatory overload. This research work achieved information dissemination through conference presentations and a manuscript submitted for publication.