Applications of Large Language Models in Pathology

Area of Science:

• Artificial Intelligence
• Computational Pathology
• Machine Learning

Background:

• Large language models (LLMs) are advanced neural networks capable of generating human-like text.
• These models are increasingly being explored for applications in various scientific fields, including medicine.

Purpose of the Study:

• To provide a historical overview of large language models.
• To highlight current and potential use cases of LLMs in the practice and education of pathology.
• To discuss the benefits and limitations of integrating LLMs into pathology workflows.

Main Methods:

• Review of existing literature on large language models.
• Analysis of transformer-based neural network architectures.
• Exploration of current and emerging applications in the field of pathology.

Main Results:

• LLMs can generate educational content, summarize text, extract data, and write code.
• Combined with vision models, LLMs can aid in histopathology image interpretation.
• Potential applications include assisting in case sign-out and report generation.

Conclusions:

• Large language models hold significant promise for revolutionizing pathology practice and education.
• Users must exercise caution, as LLMs can produce inaccuracies and hallucinations.
• Verification of AI-generated content and mindful integration are essential to mitigate risks like de-skilling and automation bias.