This research demonstrates how large language models can automate the generation of clinical summaries from CT simulation orders in radiation oncology, improving workflow efficiency.

• Successfully deployed a locally-hosted Llama 3.1 405B model to extract keywords and generate summaries from CT simulation orders
• Evaluated performance across 607 CT simulation orders from a real clinical database
• Demonstrated potential for reducing documentation burden for healthcare professionals while maintaining clinical accuracy

This innovation has significant implications for medical practice by streamlining administrative workflows, reducing manual documentation time, and potentially improving treatment planning efficiency in oncology settings.

Evaluating The Performance of Using Large Language Models to Automate Summarization of CT Simulation Orders in Radiation Oncology