[Expert consensus on artificial intelligence-assisted data acquisition and clinical application in electronic...

Area of Science:

• Otorhinolaryngology
• Medical Imaging
• Artificial Intelligence

Background:

• Current nasopharyngolaryngoscopy faces challenges like inter-observer variability and missed early diagnoses.
• Inefficiencies exist in dynamic functional assessments during examinations.

Purpose of the Study:

• Establish a comprehensive technical framework for AI-enabled data acquisition and application in nasopharyngolaryngoscopy.
• Standardize imaging protocols and annotation for AI use.
• Define AI application pathways and ethical data standards.

Main Methods:

• Introduced tiered data acquisition standards (foundational and advanced).
• Standardized a dual-mode imaging protocol (white light and electronic chromoendoscopy).
• Proposed a three-tier annotation framework (regional, characteristic, diagnostic) with quality control and validation.

Main Results:

• Delineated AI application pathways for auxiliary detection, diagnosis, quality control, report generation, and training.
• Defined ethically compliant data security standards.
• Established a framework supporting AI in Otorhinolaryngology.

Conclusions:

• The proposed technical framework supports standardization in nasopharyngeal and laryngeal disease management.
• Facilitates multi-center research collaboration and promotes standardized AI application in Otorhinolaryngology.