CLINICAL IMPACT OF ARTIFICIAL INTELLIGENCE IN ENT RADIOLOGY: A SYSTEMATIC REVIEW AND META-ANALYSIS khalilsaado@gmail.com

 

CLINICAL IMPACT OF ARTIFICIAL INTELLIGENCE IN ENT RADIOLOGY: A SYSTEMATIC REVIEW AND META-ANALYSIS

7

 

 

 

 

Australia

Oral or Poster

Artificial intelligence (AI) is rapidly reshaping paradigms in otology, particularly through its application in radiological interpretation of middle and inner ear pathology. Convolutional neural networks (CNNs) are a form of AI that can be trained using large volumes of accurately labelled training data (supervised learning) to automatically classify unlabeled test data with increasing accuracy. AI holds promise for enhancing diagnostic precision, supporting clinical decision-making, streamlining preoperative planning and expanding access to care, especially in resource-limited settings. Despite rapid technological advances, its comparative clinical efficacy remains poorly defined.

Systematic review with PRISMA guidelines using seven online databases of articles examining artificial intelligence algorithms in otology was performed. English language studies with primary or secondary endpoints pertaining to diagnostic efficiency were included. Editorials, non-clinical studies and literature reviews were excluded.

The database search identified 2124 articles. Titles and abstracts were reviewed for eligibility by two independent investigators. After full-text retrieval, 11 studies were included. Most applications of AI for interpretation of middle and inner ear radiological imaging included segmentation of CT and MRI scans, point-of-care diagnosis assistance, and improving management of middle and inner ear disorders. Meta-analysis of 11 studies assessing patient outcomes indicated high levels of accuracy when compared to otolaryngologists and radiologists, with overall accuracy of 92% vs. 85%, respectively.

AI-assisted interpretation of radiological imaging shows strong potential to enhance otologic care by improving diagnostic accuracy, streamlining management, and expanding access to healthcare, especially in underserved areas. Further research should focus on optimizing the user interface across diverse clinical settings.