CardioOnco-AI: AI-Empowered Cardiotoxicity Risk Prediction Among Breast Cancer Survivors Using Multi-Site Real-World Data

External Institution: University of Minnesota (subaward to University of Texas Health Science Center)

University of Minnesota Collaborators: Rui Zhang, PhD, MS (co-PI), Bhavadharini Ramu, MBBS; Ju Sun, PhD, MS; Anne Blaes, MD, MS; Genevieve Melton-Meaux, MD, PhD, MA; Paul Drawz, MD, MS

University of Texas Health Science Center Collaborators: Hongfang Liu, PhD, MS, (co-PI); Liwei Wang, MD, PhD, MS; Bijal Balasubramanian, MBBS, PhD, MPH

FDA Collaborators: Laleh Amiri-Kordestani, MD; Mori Krantz, MD; Abhilasha Nair, MD; Vaibhav Kumar, MD, MS; Melanie Royce, MD, PhD; Jianjin Xu, PhD

Date Started: 9/22/25

Regulatory Science Challenge

Advances in cancer diagnosis and treatment have significantly improved the survival of breast cancer patients. However, patients sometimes experience heart damage as a result of their treatment. Current guidelines for preventing and managing heart damage were not designed for cancer patients and often miss important details. Thus, there is a need for better artificial intelligence (AI) tools that can accurately predict personalized risk of heart damage in breast cancer patients.

Project Goals and Objectives

The objective of this project is to create and validate CardioOnco-AI, an AI tool that uses real-world health data to predict the risk of heart damage in breast cancer patients. The goals are to: 1) collect real-world datasets (RWD) for heart risk prediction using AI to pull important patient information, 2) develop and test CardioOnco-AI in two electronic health record (EHR) systems, and 3) assess whether CardioOnco-AI can be generalizable in two larger EHR consortia. Overall, this project will provide a new AI tool to better predict heart risks in breast cancer patients, leading to more personalized care.

For more detailed information on this project, please see this link.