P208: An Immunotherapy Registry to Study Safety Signals

Poster Presenter

Adil Alaoui

Director, Health IT and Operations
Georgetown University
United States

Objectives

A platform to identify the potential immunotherapy toxicity and immune-related adverse events that may occur during cancer treatment, as well as the methods used to develop a comprehensive I/O registry.

Method

A multi-disciplinary team of experts from the Georgetown-MedStar Cancer Network and the Hackensack Meridian Cancer Center with extensive knowledge of cancer treatment, bioinformatics, registry data compilation and research, software engineering and clinical genomics.

Results

Our approach was to use EHR data of MedStar Health and Hackensack Meridian Health as the foundation of the registry. A cohort of individuals receiving ICI treatment within the organizations has been identified and created in the database. We evaluated patients outcomes as overall survival (OS) and time to treatment failure (TTF) according to immune Response Evaluation Criteria in Solid Tumours (iRECIST) and irAEs by the CTCAE v4.03. Data of approximately 1,800 cancer patients including demographic details such as age, gender and ethnicity, as well as tumor pathology, molecular data, immunotherapy treatment, response to therapy, lab tests, and medication were collected by our team. Clinicians extracted the irAEs from the EHRs and clinical notes by conducting an evaluation of records from patients having advanced cancer treated with ICI agents. In this poster, we will also present the AI/ML-driven data modeling approaches used as a promising solution to the challenge and to enable effective risk stratification and identification of potential biomarkers that could assist in predicting response to treatment and a chance to develop irAEs, through improved risk stratification and screening algorithms.

Conclusion

The value of this unique registry and the developed data science tools and methods promise great insights into a better understanding of key irAE causes. Current collaborations and publications at medical and bioinformatics conferences, as well as manuscripts in medical journals have already influenced treatment decisions that are currently being implemented, and is already generating interest in commercial entities interested in learning how to improve the immunotherapy treatment given to oncology patients.