Deep learning-based automatic scoring of drug-induced sleep endoscopy in obstructive sleep apnea
- Document Type
- Publications · Journal Article
- Registered DB
- SCIEScopus
- Journal Ranking
- JCR Q1Scopus Q1
- Journal
- NPJ DIGITAL MEDICINE9(1) · #504 · 2026
- Citation
- Web of Science 1Scopus 2
- FWCI (Scopus)
- 0
- JCR IF
- JCR 2024 15.1
- ISSN
- 2398-6352
Authors
- Kim, Jin YoupFirst authorDongguk Univ, Ilsan Hosp, Dept Otorhinolaryngol Head & Neck Surg, Goyang, Gyeonggi Do, South Korea
- Mun, Sue JeanPusan Natl Univ, Yangsan Hosp, Dept Otorhinolaryngol Head & Neck Surg, Yangsan, Gyeongsangnam, South Korea
- Baik, Young SeoGachon Univ, Dept Biomed Engn, Seongnam, Gyeonggi Do, South Korea
- Lee, Young SeopGil Med Ctr, Med Devices R&D Ctr, Incheon, South Korea
- Kim, Young JaeGachon Univ, Gil Med Ctr, Dept Gachon Biomed & Convergence Inst, Incheon, South Korea
- Oh, JayoungSeoul Natl Univ, Bundang Hosp, Dept Otorhinolaryngol Head & Neck Surg, Seongnam, Gyeonggi Do, South Korea
- Ryu, GwanghuiSungkyunkwan Univ, Sch Med, Samsung Med Ctr, Dept Otorhinolaryngol Head & Neck Surg, Seoul, South Korea
- Sung, Chung-ManChonnam Natl Univ, Med Sch, Dept Otolaryngol Head & Neck Surg, Gwangju, South Korea
- Heo, Sung JaeKyungpook Natl Univ, Sch Med, Dept Otorhinolaryngol Head & Neck Surg, Daegu, South Korea
- Yang, Hyung ChaeChonnam Natl Univ, Med Sch, Dept Otolaryngol Head & Neck Surg, Gwangju, South Korea
- Kim, Hyun JikSeoul Natl Univ, Coll Med, Dept Otorhinolaryngol Head & Neck Surg, Seoul, South Korea
- Kim, Hyo YeolSungkyunkwan Univ, Sch Med, Samsung Med Ctr, Dept Otorhinolaryngol Head & Neck Surg, Seoul, South Korea
- Cho, KyusupPusan Natl Univ, Pusan Natl Univ Hosp, Dept Otorhinolaryngol Head & Neck Surg, Sch Med, Busan, South Korea
- Kim, Kwang GiCorresponding authorGachon Univ, Dept Biomed Engn, Seongnam, Gyeonggi Do, South Korea
- Rhee, Chae-SeoCorresponding authorSeoul Natl Univ, Coll Med, Dept Otorhinolaryngol Head & Neck Surg, Seoul, South Korea
Abstract
Polysomnography is the standard tool for assessing obstructive sleep apnea (OSA) severity; however, it does not provide information regarding the anatomical site or extent of upper airway obstruction. Drug-induced sleep endoscopy (DISE) serves as a dynamic method to evaluate airway collapse under sleep-like conditions, thereby helping to bridge this gap. However, its clinical utility is limited by inter-observer variability and subjectivity in interpretation. We developed internally and externally validated deep learning models utilizing convolutional neural networks based on EfficientNet-B2 and Attention Multiple Instance Learning to predict the degree of airway obstruction (DISE-V-obs, DISE-OTE-obs) and the primary cause of obstruction (DISE-OTE-cause) using DISE videos from 1904 patients across five Korean hospitals. The F1 scores for DISE-V-obs, DISE-OTE-obs, and DISE-OTE-cause were 84.7%, 74.7%, and 88.2%, respectively. These objective predictions of obstruction degree and primary cause may enhance clinical decision-making and treatment planning for patients with OSA.
Keywords
Identifiers
- Scopus EID
- 2-s2.0-105043736257
- WoS UID
- 001810675400001



Login
Back to Publication list