Glottis Recognition Software Development Using Artificial Intelligence

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Anesthesiology

Background:

• Video laryngoscopes improve vocal cord visualization for tracheal intubation.
• Artificial intelligence (AI) image recognition can potentially enhance anatomical comprehension during procedures.

Purpose of the Study:

• To investigate the accuracy of an AI model in identifying the glottis and surrounding structures in video laryngoscope images.
• To assess AI's ability to aid in visual identification of anatomical landmarks during simulated tracheal intubation.

Main Methods:

• An AI model was developed using image data from manikin bronchoscopy with recording capabilities.
• Image preprocessing and annotation of vocal cords, epiglottis, and glottis were performed, including human annotation.
• AI-determined anatomical structures were color-coded for enhanced visual identification.

Main Results:

• The AI model achieved over 95% accuracy (0.9516) in recognizing the epiglottis and vocal cords.
• The AI successfully marked the glottis, epiglottis, and vocal cords during simulated tracheal intubation.
• AI-generated markings significantly improved the visual identification accuracy of these structures.

Conclusions:

• AI image recognition is effective in identifying critical anatomical structures during video laryngoscopy.
• The developed AI model demonstrates high accuracy and potential to assist in tracheal intubation procedures.
• AI holds promise for enhancing anatomical understanding and procedural guidance in airway management.