Cancer

September 30, 2024

VUMC researchers awarded $1.9 million for AI-driven cancer data extraction

The project seeks to reduce the time for clinical chart abstraction by 90% while maintaining human-level accuracy.

(iStock image) (iStock image)

Vanderbilt University Medical Center researchers have secured a two-year funding award of up to $1.9 million from the Advanced Research Projects Agency for Health (ARPA-H) to develop an innovative AI-powered platform for extracting cancer patient data from medical records.

Led by principal investigator Daniel Fabbri, PhD, associate professor of Biomedical Informatics, and co-principal investigator Christine Micheel, PhD, research assistant professor of Medicine in the Division of Hematology and Oncology, the project aims to revolutionize the time-consuming process of patient chart curation in oncology.

“Our goal is to reduce the time for clinical chart abstraction by 90% while maintaining human-level accuracy,” said Fabbri. “This could dramatically accelerate cancer research and improve patient care.”

The team’s Democratized, AI-Guided Chart Abstraction Platform will use large language models to assist in extracting and organizing critical information from unstructured clinical notes and reports.

Micheel emphasized the project’s potential impact: “By streamlining data extraction, we can enable researchers to evaluate therapy efficacy more quickly and help oncologists quickly review patient histories.”

The platform is designed to be accessible to health care professionals across various resource settings.

“The tool is designed to benefit both large research institutions and smaller community hospitals and does not require programming expertise to operate,” Fabbri said.

The project aligns with ARPA-H’s mission to drive transformative biomedical breakthroughs. Established in 2022, ARPA-H is an agency within the U.S. Department of Health and Human Services that supports transformative high-risk, high-reward research to drive biomedical and health breakthroughs to benefit everyone.

The VUMC team plans to demonstrate the platform’s capabilities using various oncology datasets, including those from the American Association for Cancer Research Project GENIE and other collaborative projects.