Artificial intelligence in otology

Introduction This narrative review explores the integration of artificial intelligence (AI) into the field of otology. AI, characterized by its ability to perform tasks without explicit human programming, is gaining prominence in the medical field, including otology.

Material and methods Inclusion criteria involved original clinical studies related to AI in otology and reviews on the same topic. A PubMed search has been conducted by utilizing specific otology and AI-related keywords. Of the initial 1,076 results generated, 912 were excluded after reviewing titles and abstracts. This led to a selection of 164 articles for full-text screening, resulting in 24 studies and two reviews for inclusion.

Results In the External Ear, AI primarily serves the purpose of detecting auricular abnormalities. In the Middle Ear, AI's application focuses on otoscopy and radiology. AI systems, including convolutional neural networks (CNNs), are trained to identify various middle ear conditions by analyzing images. In the Inner Ear, the vestibular system, radiology, audiology, and cochlear implantation are the key areas where AI demonstrates its potential.

Conclusion / Discussion In the majority of the reviewed studies, a variety of AI models were used, with CNNs being the most prevalent. Evaluation metrics included accuracy, precision, sensitivity, specificity, F-score, and more. In summary, AI holds the potential to significantly impact otology by saving time, making early outcome predictions, supporting telemedicine, and objectifying findings. However, practical application in clinical settings faces challenges that necessitate further research and development.

Publikationsverlauf

Artikel online veröffentlicht:
19. April 2024