Large Language Models and Otolaryngology: A Review

Area of Science:

• Artificial intelligence
• Natural language processing
• Medical informatics

Background:

• Large language models (LLMs) are advancing rapidly, with significant potential to revolutionize healthcare administration, clinical practice, and research.
• While AI adoption is accelerating across medicine, otolaryngology (ENT) has lagged behind other specialties in leveraging LLMs.
• Otolaryngology's reliance on complex, multimodal data (text, imaging, electrophysiology, video) presents unique opportunities for LLM application.

Purpose of the Study:

• To review the advantages and innovations of LLMs relevant to otolaryngology.
• To provide otolaryngologists with a foundation for understanding and advancing LLM technologies within their field.
• To encourage the broader adoption and application of LLMs in otolaryngology research and patient care.

Main Methods:

• Review of existing literature on LLM applications in medicine, with a focus on innovations in other specialties.
• Analysis of otolaryngology's specific data characteristics and clinical complexities in relation to LLM capabilities.
• Identification of current limitations and future directions for LLM implementation in otolaryngology.

Main Results:

• LLMs offer powerful language abilities by combining deep learning and natural language processing.
• Other medical specialties have successfully applied LLMs to structured data conversion, automated phenotyping, administrative task streamlining, decision support, and multimodal data integration.
• Otolaryngology research has primarily focused on limited question-answering tasks, often using closed-source models, hindering clinical utility.

Conclusions:

• LLMs hold significant promise for enhancing otolaryngology research and patient care due to the field's rich data.
• Moving beyond feasibility studies to clinical validation, open-source development, and domain-specific fine-tuning is crucial for progress.
• Responsible LLM implementation, including secure deployment and legal oversight, can improve efficiency, decision support, and patient outcomes in otolaryngology.