Authors: Shashi Kant Gupta, Aditya Basu, Mauro Nievas, Jerrin Thomas, Nathan Wolfrath, Adhitya Ramamurthi, Bradley Taylor, Anai N. Kothari, Therica M. Miller, Sorena Nadaf-Rahrov, Yanshan Wang, Hrituraj Singh

Published on: April 23, 2024

Impact Score: 8.0

Arxiv code: Arxiv:2404.15549

Summary

• What is new: The first end-to-end large-scale empirical evaluation of clinical trial matching using real-world electronic health records.
• Why this is important: Clinical trial matching is labor-intensive, requiring detailed examination of patient records against trial criteria, leading to many patients missing potential treatments.
• What the research proposes: Utilization of Large Language Models (LLMs) for automating the matching of patients to clinical trials, showcasing a proprietary model, OncoLLM.
• Results: OncoLLM not only outperforms GPT-3.5 but also matches the performance of qualified medical doctors in clinical trial matching.

Technical Details

Technological frameworks used: nan

Models used: GPT-4, GPT-3.5, custom fine-tuned model OncoLLM

Data used: Real-world EHRs from a single cancer center in the United States

Potential Impact

Clinical trial matching services, healthcare IT solutions, cancer centers, and pharmaceutical companies conducting clinical trials

Want to implement this idea in a business?

We have generated a startup concept here: OncoMatch.