DEEP LEARNING FOR THE AUTOMATED DIAGNOSIS OF TYMPANIC MEMBRANE LESIONS USING DIGITAL OTOSCOPY: DIAGNOSTIC ACCURACY, CLINICAL APPLICATIONS, AND IMPLEMENTATION CHALLENGES
Keywords:
deep learning; tympanic membrane; otoscopy; artificial intelligence; otitis media; tympanic membrane perforation; cholesteatoma; medical imagingAbstract
DOI: https://doi.org/10.46296/yc.v10i19.0972
Abstract
Introduction: Otoscopy is central to tympanic membrane assessment, but interpretation depends on examiner expertise, image quality, and adequate visualization of anatomic landmarks. Deep learning has enabled automated classification, segmentation, and quantitative analysis of tympanic membrane abnormalities, creating potential applications in otolaryngology, primary care, and telemedicine. Objective: To critically review recent evidence on deep learning for automated diagnosis of tympanic membrane lesions using digital otoscopy, focusing on diagnostic performance, clinical utility, limitations, and priorities for safe implementation. Methods: A structured narrative review and evidence update was conducted, focusing mainly on studies published from January 2020 through August 2026. Original studies and indexed reviews evaluating deep neural networks on otoscopic or otoendoscopic images for classification, detection, segmentation, or quantification of tympanic membrane pathology were prioritized. Architecture, dataset size, diagnostic classes, validation strategy, and performance metrics were extracted. Results: Convolutional neural networks frequently achieve internal-validation accuracies above 90% for normal-versus-abnormal classification and for otitis media, tympanic membrane perforation, and cholesteatoma. Studies using 2,272–6,066 images reported accuracies of approximately 93–97%, and smartphone-based systems have demonstrated high performance under controlled conditions. Nonetheless, external validation shows a relevant performance drop; a multicohort study reported a decrease in mean AUROC from 0.95 internally to 0.76 on external datasets. Automated segmentation has achieved Dice coefficients around 0.93 and can quantify perforation morphology and derive functionally relevant image features. Conclusions: Deep learning applied to digital otoscopy is promising for diagnostic support, triage, and tele-otoscopy, but current evidence does not support autonomous clinical use. Cross-device and cross-population generalizability, robust reference standards, calibration, explainability, and prospective multicenter validation remain essential before routine deployment.
Keywords: deep learning; tympanic membrane; otoscopy; artificial intelligence; otitis media; tympanic membrane perforation; cholesteatoma; medical imaging.
Downloads
References
World Health Organization. World report on hearing. Geneva: WHO; 2021.
Schilder AGM, Chonmaitree T, Cripps AW, Rosenfeld RM, Casselbrant ML, Haggard MP, et al. Otitis media. Nat Rev Dis Primers. 2016;2:16063. doi:10.1038/nrdp.2016.63.
Lee JY, Choi SH, Chung JW. Automated classification of the tympanic membrane using a convolutional neural network. Appl Sci. 2019;9(9):1827. doi:10.3390/app9091827.
Cha D, Pae C, Seong SB, Choi JY, Park HJ. Automated diagnosis of ear disease using ensemble deep learning with a big otoendoscopy image database. EBioMedicine. 2019;45:606-614.
Khan MA, Kwon S, Choo J, Hong SM, Kang SH, Park IH, et al. Automatic detection of tympanic membrane and middle ear infection from oto-endoscopic images via convolutional neural networks. Neural Netw. 2020;126:384-394. doi:10.1016/j.neunet.2020.03.023.
Habib AR, Wong E, Sacks R, Singh N. Artificial intelligence to detect tympanic membrane perforations. J Laryngol Otol. 2020;134:311-315.
Livingstone D, Chau J. Otoscopic diagnosis using computer vision: an automated machine learning approach. Laryngoscope. 2020;130:1408-1413.
Cai Y, Yu JG, Chen Y, Liu C, Xiao L, Grais EM, et al. Investigating the use of a two-stage attention-aware convolutional neural network for the automated diagnosis of otitis media from tympanic membrane images: a prediction model development and validation study. BMJ Open. 2021;11:e041139. doi:10.1136/bmjopen-2020-041139.
Pham VT, Tran TT, Wang PC, Chen PY, Lo MT. EAR-UNet: a deep learning-based approach for segmentation of tympanic membranes from otoscopic images. Artif Intell Med. 2021;115:102065. doi:10.1016/j.artmed.2021.102065.
Byun H, Yu S, Oh J, Bae J, Yoon MS, Lee SH, et al. An assistive role of a machine learning network in diagnosis of middle ear diseases. J Clin Med. 2021;10(15):3198. doi:10.3390/jcm10153198.
Tsutsumi K, Goshtasbi K, Risbud A, Khosravi P, Pang JC, Lin HW, et al. A web-based deep learning model for automated diagnosis of otoscopic images. Otol Neurotol. 2021;42(9):e1382-e1388. doi:10.1097/MAO.0000000000003210.
Crowson MG, Hartnick CJ, Diercks GR, et al. Machine learning for accurate intraoperative pediatric middle ear effusion diagnosis. Pediatrics. 2021;147:e2020034546.
Chen YC, Chu YC, Huang CY, Lee YT, Lee WY, Hsu CY, et al. Smartphone-based artificial intelligence using a transfer learning algorithm for the detection and diagnosis of middle ear diseases: a retrospective deep learning study. eClinicalMedicine. 2022;51:101543. doi:10.1016/j.eclinm.2022.101543.
Habib AR, Kajbafzadeh M, Hasan Z, Wong E, Gunasekera H, Perry C, et al. Artificial intelligence to classify ear disease from otoscopy: a systematic review and meta-analysis. Clin Otolaryngol. 2022;47:401-413. doi:10.1111/coa.13925.
Cao Z, Chen F, Grais EM, Yue F, Cai Y, Swanepoel DW, Zhao F. Machine learning in diagnosing middle ear disorders using tympanic membrane images: a meta-analysis. Laryngoscope. 2023;133(4):732-741. doi:10.1002/lary.30291.
Habib AR, et al. Evaluating the generalizability of deep learning image classification algorithms to detect middle ear disease using otoscopy. Sci Rep. 2023;13. doi:10.1038/s41598-023-31921-0.
Tseng CC, Lim V, Jyung RW. Use of artificial intelligence for the diagnosis of cholesteatoma. Laryngoscope Investig Otolaryngol. 2023;8(1):201-211. doi:10.1002/lio2.1008.
Song D, Kim T, Lee Y, Kim J. Image-based artificial intelligence technology for diagnosing middle ear diseases: a systematic review. J Clin Med. 2023;12(18):5831. doi:10.3390/jcm12185831.
Nwosu O, Suresh K, Knoll R, Lee DJ, Crowson MG. A proof-of-concept computer vision approach for measurement of tympanic membrane perforations. Laryngoscope. 2024;134(6):2906-2911. doi:10.1002/lary.31270.
Dubois C, Eigen D, Simon F, Couloigner V, Gormish M, Chalumeau M, et al. Development and validation of a smartphone-based deep-learning-enabled system to detect middle-ear conditions in otoscopic images. NPJ Digit Med. 2024;7:162. doi:10.1038/s41746-024-01159-9.
Guo Q, Xie L, Zhou L. Deep learning for otitis media classification using otoscopic image. Medicine (Baltimore). 2025;104(48):e46218. doi:10.1097/MD.0000000000046218.
Liu TY, Chang HC, Wang PC, Hsu SY, Fang TY, Pham VT, et al. Deep learning-assisted prediction of air-bone gap using tympanic membrane perforation image features. Otolaryngol Head Neck Surg. 2026;174(2):403-412. doi:10.1002/ohn.70074.
Liu TY, Chang HC, Wang PC, Hsu SY, Fang TY, Pham VT, et al. Deep learning-assisted prediction of hearing outcomes after anatomically successful type I tympanoplasty. J Otolaryngol Head Neck Surg. 2026;55:19160216261450491. doi:10.1177/19160216261450491.
Mongan J, Moy L, Kahn CE Jr. Checklist for Artificial Intelligence in Medical Imaging (CLAIM): a guide for authors and reviewers. Radiol Artif Intell. 2020;2(2):e200029. doi:10.1148/ryai.2020200029.
Published
How to Cite
License
Copyright (c) 2026 REVISTA CIENTÍFICA MULTIDISCIPLINARIA ARBITRADA YACHASUN - ISSN: 2697-3456
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
