{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T19:50:00Z","timestamp":1792785000342,"version":"4.5.0"},"reference-count":37,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2023,1,31]],"date-time":"2023-01-31T00:00:00Z","timestamp":1675123200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,1,31]],"date-time":"2023-01-31T00:00:00Z","timestamp":1675123200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100003725","name":"National Research Foundation of Korea","doi-asserted-by":"publisher","award":["2020R1C1C1010079"],"award-info":[{"award-number":["2020R1C1C1010079"]}],"id":[{"id":"10.13039\/501100003725","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Ki-Soo Kim Research Fund","award":["2022"],"award-info":[{"award-number":["2022"]}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Sci Rep"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>We sought to establish an unsupervised algorithm with a three\u2013dimensional (3D) variational autoencoder model (VAE) for the detection of abnormal extraocular muscles in small datasets of orbital computed tomography (CT) images. 334 CT images of normal orbits and 96 of abnormal orbits diagnosed as thyroid eye disease were used for training and validation; 24 normal and 11 abnormal orbits were used for the test. A 3D VAE was developed and trained. All images were preprocessed to emphasize extraocular muscles and to suppress background noise (e.g., high signal intensity from bones). The optimal cut-off value was identified through receiver operating characteristic (ROC) curve analysis. The ability of the model to detect muscles of abnormal size was assessed by visualization. The model achieved a sensitivity of 79.2%, specificity of 72.7%, accuracy of 77.1%, F1-score of 0.667, and AUROC of 0.801. Abnormal CT images correctly identified by the model showed differences in the reconstruction of extraocular muscles. The proposed model showed potential to detect abnormalities in extraocular muscles using a small dataset, similar to the diagnostic approach used by physicians. Unsupervised learning could serve as an alternative detection method for medical imaging studies in which annotation is difficult or impossible to perform.<\/jats:p>","DOI":"10.1038\/s41598-023-28082-5","type":"journal-article","created":{"date-parts":[[2023,1,31]],"date-time":"2023-01-31T08:04:29Z","timestamp":1675152269000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":10,"title":["Detection of abnormal extraocular muscles in small datasets of computed tomography images using a three-dimensional variational autoencoder"],"prefix":"10.1038","volume":"13","author":[{"given":"Yeon Woong","family":"Chung","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"In Young","family":"Choi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2023,1,31]]},"reference":[{"key":"28082_CR1","doi-asserted-by":"publisher","first-page":"102","DOI":"10.1016\/j.zemedi.2018.11.002","volume":"29","author":"AS Lundervold","year":"2019","unstructured":"Lundervold, A. S. & Lundervold, A. An overview of deep learning in medical imaging focusing on MRI. Z. Med. Phys. 29, 102\u2013127. https:\/\/doi.org\/10.1016\/j.zemedi.2018.11.002 (2019).","journal-title":"Z. Med. Phys."},{"key":"28082_CR2","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1186\/s40537-021-00444-8","volume":"8","author":"L Alzubaidi","year":"2021","unstructured":"Alzubaidi, L. et al. Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. J Big Data 8, 53. https:\/\/doi.org\/10.1186\/s40537-021-00444-8 (2021).","journal-title":"J Big Data"},{"key":"28082_CR3","doi-asserted-by":"publisher","first-page":"e715","DOI":"10.7717\/peerj-cs.715","volume":"7","author":"L Alzubaidi","year":"2021","unstructured":"Alzubaidi, L. et al. Deepening into the suitability of using pre-trained models of imagenet against a lightweight convolutional neural network in medical imaging: An experimental study. PeerJ Comput. Sci. 7, e715. https:\/\/doi.org\/10.7717\/peerj-cs.715 (2021).","journal-title":"PeerJ Comput. Sci."},{"key":"28082_CR4","doi-asserted-by":"publisher","first-page":"377","DOI":"10.1002\/jum.15413","volume":"40","author":"L Blaivas","year":"2021","unstructured":"Blaivas, L. & Blaivas, M. Are convolutional neural networks trained on imagenet images wearing rose-colored glasses?: A quantitative comparison of imagenet, computed tomographic, magnetic resonance, chest x-ray, and point-of-care ultrasound images for quality. J. Ultrasound Med. 40, 377\u2013383. https:\/\/doi.org\/10.1002\/jum.15413 (2021).","journal-title":"J. Ultrasound Med."},{"key":"28082_CR5","doi-asserted-by":"publisher","first-page":"788","DOI":"10.1076\/ceyr.21.4.788.5548","volume":"21","author":"SW Cheung","year":"2000","unstructured":"Cheung, S. W. & Cho, P. Endothelial cells analysis with the TOPCON specular microscope SP-2000P and IMAGEnet system. Curr. Eye Res. 21, 788\u2013798. https:\/\/doi.org\/10.1076\/ceyr.21.4.788.5548 (2000).","journal-title":"Curr. Eye Res."},{"key":"28082_CR6","doi-asserted-by":"publisher","first-page":"5915","DOI":"10.1038\/s41467-021-26216-9","volume":"12","author":"S Wang","year":"2021","unstructured":"Wang, S. et al. Annotation-efficient deep learning for automatic medical image segmentation. Nat. Commun. 12, 5915. https:\/\/doi.org\/10.1038\/s41467-021-26216-9 (2021).","journal-title":"Nat. Commun."},{"key":"28082_CR7","doi-asserted-by":"publisher","DOI":"10.3390\/s22041596","author":"CB Rasmussen","year":"2022","unstructured":"Rasmussen, C. B., Kirk, K. & Moeslund, T. B. The challenge of data annotation in deep learning-a case study on whole plant corn silage. Sensors (Basel) https:\/\/doi.org\/10.3390\/s22041596 (2022).","journal-title":"Sensors (Basel)"},{"key":"28082_CR8","doi-asserted-by":"publisher","first-page":"101952","DOI":"10.1016\/j.media.2020.101952","volume":"69","author":"C Baur","year":"2021","unstructured":"Baur, C., Denner, S., Wiestler, B., Navab, N. & Albarqouni, S. Autoencoders for unsupervised anomaly segmentation in brain MR images: A comparative study. Med. Image Anal. 69, 101952. https:\/\/doi.org\/10.1016\/j.media.2020.101952 (2021).","journal-title":"Med. Image Anal."},{"key":"28082_CR9","doi-asserted-by":"publisher","first-page":"149","DOI":"10.2214\/ajr.145.1.149","volume":"145","author":"G Forbes","year":"1985","unstructured":"Forbes, G., Gehring, D. G., Gorman, C. A., Brennan, M. D. & Jackson, I. T. Volume measurements of normal orbital structures by computed tomographic analysis. AJR Am. J. Roentgenol. 145, 149\u2013154. https:\/\/doi.org\/10.2214\/ajr.145.1.149 (1985).","journal-title":"AJR Am. J. Roentgenol."},{"key":"28082_CR10","doi-asserted-by":"publisher","first-page":"94","DOI":"10.4274\/tjo.80688","volume":"47","author":"E Sahli","year":"2017","unstructured":"Sahli, E. & Gunduz, K. Thyroid-associated ophthalmopathy. Turk. J. Ophthalmol. 47, 94\u2013105. https:\/\/doi.org\/10.4274\/tjo.80688 (2017).","journal-title":"Turk. J. Ophthalmol."},{"key":"28082_CR11","doi-asserted-by":"publisher","first-page":"553","DOI":"10.1016\/s0161-6420(81)34993-8","volume":"88","author":"SL Trokel","year":"1981","unstructured":"Trokel, S. L. & Jakobiec, F. A. Correlation of CT scanning and pathologic features of ophthalmic Graves\u2019 disease. Ophthalmology 88, 553\u2013564. https:\/\/doi.org\/10.1016\/s0161-6420(81)34993-8 (1981).","journal-title":"Ophthalmology"},{"key":"28082_CR12","doi-asserted-by":"publisher","first-page":"675","DOI":"10.1148\/radiology.177.3.2243967","volume":"177","author":"RA Nugent","year":"1990","unstructured":"Nugent, R. A. et al. Graves orbitopathy: Correlation of CT and clinical findings. Radiology 177, 675\u2013682. https:\/\/doi.org\/10.1148\/radiology.177.3.2243967 (1990).","journal-title":"Radiology"},{"key":"28082_CR13","first-page":"737","volume":"4","author":"G Forbes","year":"1983","unstructured":"Forbes, G., Gorman, C. A., Gehring, D. & Baker, H. L. Jr. Computer analysis of orbital fat and muscle volumes in graves ophthalmopathy. AJNR Am. J. Neuroradiol. 4, 737\u2013740 (1983).","journal-title":"AJNR Am. J. Neuroradiol."},{"key":"28082_CR14","doi-asserted-by":"publisher","first-page":"1327","DOI":"10.6061\/clinics\/2012(11)18","volume":"67","author":"AC Goncalves","year":"2012","unstructured":"Goncalves, A. C., Gebrim, E. M. & Monteiro, M. L. Imaging studies for diagnosing graves\u2019 orbitopathy and dysthyroid optic neuropathy. Clinics (Sao Paulo) 67, 1327\u20131334. https:\/\/doi.org\/10.6061\/clinics\/2012(11)18 (2012).","journal-title":"Clinics (Sao Paulo)"},{"key":"28082_CR15","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1186\/s12859-020-03936-1","volume":"22","author":"C Han","year":"2021","unstructured":"Han, C. et al. MADGAN: Unsupervised medical anomaly detection GAN using multiple adjacent brain MRI slice reconstruction. BMC Bioinform. 22, 31. https:\/\/doi.org\/10.1186\/s12859-020-03936-1 (2021).","journal-title":"BMC Bioinform."},{"key":"28082_CR16","doi-asserted-by":"publisher","first-page":"e27822","DOI":"10.2196\/27822","volume":"23","author":"Y Han","year":"2021","unstructured":"Han, Y. et al. Application of an anomaly detection model to screen for ocular diseases using color retinal fundus images: Design and evaluation study. J. Med. Internet Res. 23, e27822. https:\/\/doi.org\/10.2196\/27822 (2021).","journal-title":"J. Med. Internet Res."},{"key":"28082_CR17","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1038\/s41598-020-79336-5","volume":"11","author":"TT Ho","year":"2021","unstructured":"Ho, T. T. et al. A 3D-CNN model with CT-based parametric response mapping for classifying COPD subjects. Sci. Rep. 11, 34. https:\/\/doi.org\/10.1038\/s41598-020-79336-5 (2021).","journal-title":"Sci. Rep."},{"key":"28082_CR18","doi-asserted-by":"publisher","first-page":"1413","DOI":"10.1007\/s11548-021-02451-9","volume":"16","author":"M Bengs","year":"2021","unstructured":"Bengs, M., Behrendt, F., Kruger, J., Opfer, R. & Schlaefer, A. Three-dimensional deep learning with spatial erasing for unsupervised anomaly segmentation in brain MRI. Int. J. Comput. Assist. Radiol. Surg. 16, 1413\u20131423. https:\/\/doi.org\/10.1007\/s11548-021-02451-9 (2021).","journal-title":"Int. J. Comput. Assist. Radiol. Surg."},{"key":"28082_CR19","doi-asserted-by":"publisher","first-page":"4598","DOI":"10.1093\/bioinformatics\/btac529","volume":"38","author":"J Wang","year":"2022","unstructured":"Wang, J. et al. 3D GAN image synthesis and dataset quality assessment for bacterial biofilm. Bioinformatics 38, 4598\u20134604. https:\/\/doi.org\/10.1093\/bioinformatics\/btac529 (2022).","journal-title":"Bioinformatics"},{"key":"28082_CR20","doi-asserted-by":"publisher","DOI":"10.3390\/cells8091040","author":"L Zhang","year":"2019","unstructured":"Zhang, L., Chen, X. & Yin, J. Prediction of potential miRNA-disease associations through a novel unsupervised deep learning framework with variational autoencoder. Cells https:\/\/doi.org\/10.3390\/cells8091040 (2019).","journal-title":"Cells"},{"key":"28082_CR21","doi-asserted-by":"publisher","first-page":"105683","DOI":"10.1016\/j.compbiomed.2022.105683","volume":"147","author":"M Ogawa","year":"2022","unstructured":"Ogawa, M. et al. Utility of unsupervised deep learning using a 3D variational autoencoder in detecting inner ear abnormalities on CT images. Comput. Biol. Med. 147, 105683. https:\/\/doi.org\/10.1016\/j.compbiomed.2022.105683 (2022).","journal-title":"Comput. Biol. Med."},{"key":"28082_CR22","doi-asserted-by":"publisher","first-page":"3816","DOI":"10.1002\/mp.15603","volume":"49","author":"T Escobar","year":"2022","unstructured":"Escobar, T. et al. Voxel-wise supervised analysis of tumors with multimodal engineered features to highlight interpretable biological patterns. Med. Phys. 49, 3816\u20133829. https:\/\/doi.org\/10.1002\/mp.15603 (2022).","journal-title":"Med. Phys."},{"key":"28082_CR23","doi-asserted-by":"publisher","first-page":"635","DOI":"10.1111\/his.14571","volume":"80","author":"LM Hondelink","year":"2022","unstructured":"Hondelink, L. M. et al. Development and validation of a supervised deep learning algorithm for automated whole-slide programmed death-ligand 1 tumour proportion score assessment in non-small cell lung cancer. Histopathology 80, 635\u2013647. https:\/\/doi.org\/10.1111\/his.14571 (2022).","journal-title":"Histopathology"},{"key":"28082_CR24","doi-asserted-by":"publisher","first-page":"4494","DOI":"10.1002\/mp.15627","volume":"49","author":"G Lin","year":"2022","unstructured":"Lin, G. et al. Improving sensitivity and connectivity of retinal vessel segmentation via error discrimination network. Med. Phys. 49, 4494\u20134507. https:\/\/doi.org\/10.1002\/mp.15627 (2022).","journal-title":"Med. Phys."},{"key":"28082_CR25","doi-asserted-by":"publisher","first-page":"11093","DOI":"10.1109\/ACCESS.2019.2891970","volume":"7","author":"J Enguehard","year":"2019","unstructured":"Enguehard, J., O\u2019Halloran, P. & Gholipour, A. Semi supervised learning with deep embedded clustering for image classification and segmentation. IEEE Access 7, 11093\u201311104. https:\/\/doi.org\/10.1109\/ACCESS.2019.2891970 (2019).","journal-title":"IEEE Access"},{"key":"28082_CR26","doi-asserted-by":"publisher","first-page":"215","DOI":"10.1186\/s12911-020-01230-x","volume":"20","author":"H Zhu","year":"2020","unstructured":"Zhu, H., Fang, Q., Huang, Y. & Xu, K. Semi-supervised method for image texture classification of pituitary tumors via CycleGAN and optimized feature extraction. BMC Med. Inform. Decis. Mak. 20, 215. https:\/\/doi.org\/10.1186\/s12911-020-01230-x (2020).","journal-title":"BMC Med. Inform. Decis. Mak."},{"key":"28082_CR27","doi-asserted-by":"publisher","first-page":"4133","DOI":"10.1038\/s41598-022-07685-4","volume":"12","author":"T Monjo","year":"2022","unstructured":"Monjo, T., Koido, M., Nagasawa, S., Suzuki, Y. & Kamatani, Y. Efficient prediction of a spatial transcriptomics profile better characterizes breast cancer tissue sections without costly experimentation. Sci. Rep. 12, 4133. https:\/\/doi.org\/10.1038\/s41598-022-07685-4 (2022).","journal-title":"Sci. Rep."},{"key":"28082_CR28","doi-asserted-by":"publisher","first-page":"601","DOI":"10.1007\/s10637-017-0524-2","volume":"36","author":"RR Colen","year":"2018","unstructured":"Colen, R. R. et al. Radiomics to predict immunotherapy-induced pneumonitis: Proof of concept. Invest New Drugs 36, 601\u2013607. https:\/\/doi.org\/10.1007\/s10637-017-0524-2 (2018).","journal-title":"Invest New Drugs"},{"key":"28082_CR29","doi-asserted-by":"publisher","first-page":"1478","DOI":"10.1007\/s00259-020-05075-4","volume":"48","author":"H Wang","year":"2021","unstructured":"Wang, H. et al. Decoding COVID-19 pneumonia: Comparison of deep learning and radiomics CT image signatures. Eur. J. Nucl. Med. Mol. Imaging 48, 1478\u20131486. https:\/\/doi.org\/10.1007\/s00259-020-05075-4 (2021).","journal-title":"Eur. J. Nucl. Med. Mol. Imaging"},{"key":"28082_CR30","doi-asserted-by":"publisher","first-page":"232","DOI":"10.1016\/j.neucom.2015.08.104","volume":"184","author":"Y Wang","year":"2016","unstructured":"Wang, Y., Yao, H. & Zhao, S. Auto-encoder based dimensionality reduction. Neurocomputing 184, 232\u2013242 (2016).","journal-title":"Neurocomputing"},{"key":"28082_CR31","unstructured":"Kingma D.P., W. M. Auto-encoding variational bayes. In Proceedings of the 2nd International Conference on Learning Representations (ICLR)); Banff, AB, Canada (2014)."},{"key":"28082_CR32","first-page":"355","volume":"39","author":"W Xu","year":"2019","unstructured":"Xu, W., Keshmiri, S. & Wang, W. Adversarially approximated autoencoder for image generation and manipulation. IEEE Trans. Multimedia 39, 355\u2013368 (2019).","journal-title":"IEEE Trans. Multimedia"},{"key":"28082_CR33","doi-asserted-by":"publisher","unstructured":"Kingma, D.P., W. M. An introduction to variational autoencoders. arXiv doi: https:\/\/doi.org\/10.1561\/2200000056. Available online: https:\/\/arxiv.org\/abs\/1906.026911906.02691 (2019).","DOI":"10.1561\/2200000056"},{"key":"28082_CR34","doi-asserted-by":"publisher","DOI":"10.3390\/jimaging7050083","author":"M Elbattah","year":"2021","unstructured":"Elbattah, M. et al. Variational autoencoder for image-based augmentation of eye-tracking data. J. Imaging https:\/\/doi.org\/10.3390\/jimaging7050083 (2021).","journal-title":"J. Imaging"},{"key":"28082_CR35","doi-asserted-by":"publisher","first-page":"146","DOI":"10.1214\/aop\/1176996454","volume":"3","author":"I Csiszar","year":"1975","unstructured":"Csiszar, I. I-Divergence geometry of probability distributions and minimization problems. Ann. Probab. 3, 146\u2013158. https:\/\/doi.org\/10.1214\/aop\/1176996454 (1975).","journal-title":"Ann. Probab."},{"key":"28082_CR36","doi-asserted-by":"publisher","unstructured":"Odaibo, S. Tutorial: Deriving the Standard Variational Autoencoder (VAE) Loss Function. arXiv (2019). Doi: https:\/\/doi.org\/10.48550\/arXiv.1907.08956","DOI":"10.48550\/arXiv.1907.08956"},{"key":"28082_CR37","doi-asserted-by":"publisher","first-page":"23","DOI":"10.1016\/s0167-5877(00)00115-x","volume":"45","author":"M Greiner","year":"2000","unstructured":"Greiner, M., Pfeiffer, D. & Smith, R. Principles and practical application of the receiver-operating characteristic analysis for diagnostic tests. Rev. Prev. Vet. Med. 45, 23\u201341. https:\/\/doi.org\/10.1016\/s0167-5877(00)00115-x (2000).","journal-title":"Rev. Prev. Vet. Med."}],"container-title":["Scientific Reports"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41598-023-28082-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-023-28082-5","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-023-28082-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,1,31]],"date-time":"2023-01-31T08:15:13Z","timestamp":1675152913000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41598-023-28082-5"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,1,31]]},"references-count":37,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2023,12]]}},"alternative-id":["28082"],"URL":"https:\/\/doi.org\/10.1038\/s41598-023-28082-5","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-2031362\/v1","asserted-by":"object"}]},"ISSN":["2045-2322"],"issn-type":[{"value":"2045-2322","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,1,31]]},"assertion":[{"value":"4 September 2022","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"12 January 2023","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"31 January 2023","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors declare no competing interests.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"1765"}}