Inside Precision Medicine March 3, 2025
Flatiron Health researchers used large language models (LLMs) to extract PD-L1 biomarker testing details from cancer patient data for clinical decisions
In the ever-evolving intersection of artificial intelligence (AI) and healthcare, a groundbreaking AI in Precision Oncology study highlights how large language models (LLMs) could transform cancer care by unlocking critical data hidden in unstructured electronic health records (EHRs).
Researchers at Flatiron Health, led by Aaron B. Cohen, MD, head of research oncology, clinical data, looked into how open-source LLMs could get detailed information about PD-L1 biomarker testing—a key part of customizing cancer treatments—from EHRs. The results show a big step forward: When these AI systems were fine-tuned with high-quality, expert-labeled data, they did better than traditional deep learning models,...
Today's Sponsors
Today's Sponsor