Skip to main content
AAO-HNSF 2025 Annual Meeting & OTO EXPO Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Scientific Oral Presentations

Comprehensive Otolaryngology

Session: Scientific Oral Presentations 1 - Comprehensive Otolaryngology

Clinical Applications of Multimodal Artificial Intelligence in Otolaryngology: A State of the Art Review

Saturday, October 11, 2025
10:18 AM - 10:24 AM EDT
Location: 242
Disclosure(s):
Ying Jie Li:

faculty for this accredited education activity has no relationship(s) with ineligible companies to disclose.

Introduction: The artificial intelligence (AI) healthcare market has surged by billions in the last 3 years, with emerging models capable of simultaneously processing visual, auditory, and textual inputs, providing users with ‘multimodal’ AI models. Given the potential of multimodal AI for clinical integration, it's essential for otolaryngologists to stay informed, as current use for AI is concentrated to areas such patient education and research assistance. This study aims to review current literature on multimodal AI in otolaryngology.

Methods: Following the preferred reporting items for systematic reviews and meta-analyses extension for scoping reviews checklist, the MEDLINE, EMBASE, SCOPUS, Cochrane Library, Web of Science, and CINAHL databases were searched from date of inception to Oct 29th, 2024. Studies on the application of multimodal AI in otolaryngology were included.

Results: 31 studies were included, with 61% (19/31) published in 2024. Image and text were the most commonly combined modalities (84%, 25/31), with emerging combinations including video with vector data (3%, 1/31) and omics with text and/or image (16%, 5/31). Head and neck cancer (HNC) was the subspecialty of focus (68%, 21/31), followed by general ENT (13%, 4/31). All studies applied the multimodal AI models for clinical education and decision support, assessing performance in areas such as disease classification or cancer prognostication (AUC 0.57-0.90, median 0.84), and board-style examination accuracy (47.3-86.0%).

Conclusions: To date, the rise in multimodal AI utilization in otolaryngology has concentrated in HNC, demonstrating potential applications for staging, prognostication, and treatment planning. However, such tools are not yet reliable for clinical use and require further validation in the clinical setting prior to deployment.

    Presenting Author(s)

  • YL

    Ying Jie Li

    Medical Student
    University of British Columbia, Faculty of Medicine
    Coquitlam, British Columbia, Canada

    Disclosure(s):

    faculty for this accredited education activity has no relationship(s) with ineligible companies to disclose.

    • Co-Author(s)

  • NB

    Norbert Banyi, n/a

    University of British Columbia, Faculty of Medicine, Canada

    Disclosure information not submitted.

  • FS

    Flora Su, n/a

    University of British Columbia, Faculty of Science, British Columbia, Canada

    Disclosure information not submitted.

  • PE

    Philip Edgcumbe, MD, PhD

    Department of Radiology, University of British Columbia, Canada

    Disclosure information not submitted.

  • AT

    Andrew Thamboo, MD, FRCSC, MHSC

    Division of Otolaryngology–Head and Neck Surgery, Department of Surgery, University of British Columbia, Canada

    Disclosure information not submitted.

    • Senior Author(s)

  • Ameen Amanian, MD FRCSC MSE

    Clinical Instructor
    University of Kansas Medical Center
    Kelowna, BC, Canada

    Disclosure information not submitted.