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.

Best of Scientific Oral Presentations

Pediatric Otolaryngology

Session: Best of Scientific Orals 1

Computer Vision Model for Automated Detection of Abnormal Pediatric Vocal Cord Movement on Video Laryngoscopy

Sunday, October 12, 2025
1:54 PM - 2:00 PM EDT
Location: 120
Introduction: Current methods for evaluating vocal cord movement in children are qualitative and rely on subjective interpretation. This study aims to develop a computer vision model for real-time tracking of pediatric vocal cord movements from video laryngoscopy and to objectively detect movement abnormalities.

Methods: One hundred pediatric laryngoscopy videos were annotated to extract maximum and minimum anterior glottic angles (AGA). A ResNet-based computer vision model, fine-tuned for feature extraction, was combined with linear regression to predict AGA values. Model performance, evaluated using mean absolute error and Pearson correlation coefficients, was compared to manual expert annotations. The system was developed to characterize vocal mobility as normal or abnormal (hypomobile or immobile) based on AGA predictions.

Results: The model’s AGA estimations correlated strongly with expert annotations (r = 0.87, p < 0.001) and had a mean absolute difference of 2.91° (3.02°). Children with abnormal VC mobility had a significantly smaller maximum AGA [mean (SD): 34.92° (12.53°) vs. 55.48° (11.73°); p < 0.001] and reduced AGA range [mean (SD): 29.79° (12.24°) vs. 51.82° (12.76°); p < 0.001]. The model showed a high discriminatory ability for detecting vocal cord movement abnormality [AUC 0.877, 95% CI: 0.812-0.939; p < 0.001].

Conclusions: This computer vision model successfully quantified AGA from pediatric video laryngoscopy, accurately differentiating between normal and abnormal vocal cord movements. Results may potentially be used as a diagnostic aid for monitoring outcomes for pediatric vocal cord movement abnormalities.

    Presenting Author(s)

  • Yasmine Madan, BSc

    Medical Student
    Hospital for Sick Children
    Toronto, ON, Canada

    Disclosure(s):

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

    • Co-Author(s)

  • NC

    Nida Copty, n/a

    The Hospital for Sick Children, United States

    Disclosure information not submitted.

  • Karan S. Chahal

    Undergraduate Student
    Hospital for Sick Children
    Brampton, ON, Canada

    Disclosure information not submitted.

  • Frank Chen

    Medical Student
    Queen's University
    Kingston, ON, Canada

    Disclosure information not submitted.

  • PE

    Paul Esemu-Ezewu, n/a

    The Hospital for Sick Children, United States

    Disclosure information not submitted.

  • Nikolaus Ernst E. Wolter, MD, MSc, FRCSC, FACS

    Otolaryngologist
    Hospital for Sick Children
    Toronto, ON, Canada

    Disclosure information not submitted.

  • Evan J. Propst, MD, MSc

    Professor
    Hospital for Sick Children
    Toronto, ON, Canada

    Disclosure information not submitted.

  • CM

    Clyde Matava, MD

    The Hospital for Sick Children, United States

    Disclosure information not submitted.

    • Senior Author(s)

  • JS

    Jennifer M. Siu, MD, MPH

    Otolaryngologist
    University of Toronto
    Toronto, ON, Canada

    Disclosure information not submitted.