Current Status and Future Directions of Research on Artificial Intelligence in Nasopharyngolaryngoscopy

BACKGROUND: The nasopharyngolaryngoscopy (NPL) has emerged as a valuable tool for detecting early cases of head and neck cancers. However, misdiagnoses and missed diagnoses are still common phenomena. The expertise of examining physicians often serves as the primary limiting factor, leading to issues such as incomplete visualization, imprecise identification, and unclear vision. Over recent years, the application of artificial intelligence (AI) in medical imaging, particularly in the realm of gastrointestinal endoscopy, has instigated revolutionary changes in site quality control, lesion identification, and report generation. However, there remains a lack of standardized guidelines for the proper application of NPL across various countries. SUMMARY: In this paper, we set our sights on reviewing the current clinical applications and summarizing the primary shortcomings of NPL. In addition, we encapsulate the progress of AI application within gastrointestinal endoscopy and NPL. Drawing from real-world clinical practice, we propose future directions and prospects for AI research in NPL. We firmly believe that the pace of clinical application of AI in NPL will accelerate significantly in the near future. KEY MESSAGES: Incomplete examination coverage, failure to detect and diagnose lesions, and poor image quality happens in the current use of NPL. Currently, NPL examinations lack third-party supervision and quality control. AI application has achieved great advancements in gastrointestinal endoscopy concerning endoscopic quality control, lesion identification, and standardized reporting. While AI-related research in NPL is still in its nascent stages, it shows substantial potential for clinical application and endoscopic training. The interaction of AI into NPL examinations is potential and inevitable in the era of big data.