Researchers at Seoul National University Bundang Hospital (SNUBH) has successfully completed the "Drug-Induced Sleep Endoscopy AI Learning Data Construction National Project" with the highest grade rating.

Drug-induced sleep endoscopy (DISE) is a widely used procedure for diagnosing and treating conditions such as sleep apnea and other upper airway obstructions. It allows for a three-dimensional view of the airway, facilitating accurate assessment and convenience in use.

However, the procedure's reliance on direct visualization means it can be subject to the examiner's subjective judgment, potentially leading to inconsistencies and inaccuracies in diagnosis and treatment.

To address this challenge, the team, led by Professor Lee Jae-seo of the Department of Otolaryngology, embarked on a national project to collect data for medical AI software, aiming to increase the consistency of upper airway endoscopy image interpretations.

The team gathered approximately 9,000 video data and 45,000 image data, making it accessible on the AI Hub for medical AI researchers worldwide.

The project received the highest grade of "Excellent" across various evaluation criteria, including diversity, grammatical accuracy, semantic accuracy, and validity.

"This project has allowed us to construct a vast database of endoscopy videos and images," Professor Lee said. "We believe these resources will be instrumental in standardizing drug-induced sleep endoscopy diagnoses through medical AI data utilization."