Sign in Subscribe
Artificial intelligence

Flatiron Health Presents AI-Driven Cancer Progression Extraction Research at AACR Conference

Strategic Insights

— 1 min read

Flatiron Health's recent presentation at the American Association for Cancer Research (AACR) Special Conference in Cancer Research: Artificial Intelligence and Machine Learning 2025 showcased the potential of large language models (LLMs) to accurately extract real-world cancer progression events from unstructured electronic health record (EHR) data across 14 cancer types. The research demonstrated that LLMs achieved F1 scores similar to expert human abstractors and produced nearly identical real-world progression-free survival estimates, underscoring the potential of AI to scale high-quality clinical endpoint extraction for oncology research and care.

Key Takeaways:

  • Flatiron Health presented two new research studies on AI-driven cancer progression extraction at the AACR Special Conference in Cancer Research: Artificial Intelligence and Machine Learning 2025.
  • The research demonstrated the potential of LLMs to accurately and efficiently extract real-world cancer progression events from EHR data across 14 cancer types.
  • LLMs achieved F1 scores similar to expert human abstractors and produced nearly identical real-world progression-free survival estimates.
  • The study utilized the VALID Framework to assess the quality of LLM-extracted real-world data and compared it to expert human abstractors.
  • LLMs were provided by Anthropic, the leading AI safety and research company that builds the Claude family of models.
  • The research has the potential to accelerate clinical research, improve patient outcomes, and set a new standard for evidence generation in cancer care.

Statistics:

  • 14 cancer types were evaluated in the study.
  • LLMs achieved F1 scores similar to expert human abstractors.
  • Real-world progression-free survival estimates produced by LLMs were nearly identical to those produced by expert human abstractors.
  • The study utilized the VALID Framework to assess the quality of LLM-extracted real-world data.

Sources:

  • "Using Large Language Models for Scalable Extraction of Real-World Progression Events across Multiple Cancer Types" (presentation at AACR Special Conference in Cancer Research: Artificial Intelligence and Machine Learning 2025)
  • Flatiron Health Press Release (July 14)

Read more