SurgeryLLM: a retrieval-augmented generation large language model framework for surgical decision support and...

Area of Science:

• Artificial Intelligence in Medicine
• Surgical Informatics
• Clinical Decision Support Systems

Background:

• Large language models (LLMs) offer potential for processing complex medical information.
• Integrating evidence-based guidelines into clinical workflows remains a challenge.
• Surgical decision-making requires efficient access to up-to-date, guideline-based knowledge.

Purpose of the Study:

• To evaluate the capability of a novel LLM framework, SurgeryLLM, in incorporating domain-specific knowledge from surgical guidelines.
• To assess the performance of SurgeryLLM when presented with patient-specific data.
• To determine the potential impact of guideline-informed LLMs on surgical practice.

Main Methods:

• Development of SurgeryLLM, a Retrieval Augmented Generation (RAG) framework.
• Testing SurgeryLLM with simulated patient-specific data and current evidence-based surgical guidelines.
• Qualitative and quantitative analysis of the model's ability to retrieve and apply guideline information.

Main Results:

• SurgeryLLM demonstrably incorporated domain-specific knowledge from surgical guidelines.
• The framework successfully utilized patient-specific data to inform guideline application.
• Evidence of accurate retrieval and relevant application of guideline-based information was observed.

Conclusions:

• SurgeryLLM represents a significant advancement in applying AI to surgical knowledge management.
• Successful integration of guideline information enhances potential for improved surgeon efficiency.
• This technology shows promise for improving patient safety and optimizing surgical outcomes through AI-driven clinical support.