Effectiveness of data-augmentation on deep learning in evaluating rapid on-site cytopathology at endoscopic ultrasound-guided fine needle aspiration

Fujii, Yuki; Uchida, Daisuke; Sato, Ryosuke; Obata, Taisuke; Akihiro, Matsumi; Miyamoto, Kazuya; Morimoto, Kosaku; Terasawa, Hiroyuki; Yamazaki, Tatsuhiro; Matsumoto, Kazuyuki; Horiguchi, Shigeru; Tsutsumi, Koichiro; Kato, Hironari; Inoue, Hirofumi; Cho, Ten; Tanimoto, Takayoshi; Ohto, Akimitsu; Kawahara, Yoshiro; Otsuka, Motoyuki

doi:10.1038/s41598-024-72312-3

Permalink : https://ousar.lib.okayama-u.ac.jp/67646

ID	67646
フルテキストURL	fulltext.pdf 2.33 MB
著者	Fujii, Yuki Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Uchida, Daisuke Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Sato, Ryosuke Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Obata, Taisuke Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Akihiro, Matsumi Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Miyamoto, Kazuya Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Morimoto, Kosaku Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Terasawa, Hiroyuki Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Yamazaki, Tatsuhiro Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Matsumoto, Kazuyuki Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science ORCID Kaken ID publons Horiguchi, Shigeru Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Tsutsumi, Koichiro Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science ORCID Kaken ID researchmap Kato, Hironari Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science ORCID Kaken ID researchmap Inoue, Hirofumi Department of Pathology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Cho, Ten Business Strategy Division, Ryobi Systems Co., Ltd. Tanimoto, Takayoshi Business Strategy Division, Ryobi Systems Co., Ltd. Ohto, Akimitsu Business Strategy Division, Ryobi Systems Co., Ltd. Kawahara, Yoshiro Department of Practical Gastrointestinal Endoscopy, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science Kaken ID researchmap Otsuka, Motoyuki Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Science
抄録	Rapid on-site cytopathology evaluation (ROSE) has been considered an effective method to increase the diagnostic ability of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA); however, ROSE is unavailable in most institutes worldwide due to the shortage of cytopathologists. To overcome this situation, we created an artificial intelligence (AI)-based system (the ROSE-AI system), which was trained with the augmented data to evaluate the slide images acquired by EUS-FNA. This study aimed to clarify the effects of such data-augmentation on establishing an effective ROSE-AI system by comparing the efficacy of various data-augmentation techniques. The ROSE-AI system was trained with increased data obtained by the various data-augmentation techniques, including geometric transformation, color space transformation, and kernel filtering. By performing five-fold cross-validation, we compared the efficacy of each data-augmentation technique on the increasing diagnostic abilities of the ROSE-AI system. We collected 4059 divided EUS-FNA slide images from 36 patients with pancreatic cancer and nine patients with non-pancreatic cancer. The diagnostic ability of the ROSE-AI system without data augmentation had a sensitivity, specificity, and accuracy of 87.5%, 79.7%, and 83.7%, respectively. While, some data-augmentation techniques decreased diagnostic ability, the ROSE-AI system trained only with the augmented data using the geometric transformation technique had the highest diagnostic accuracy (88.2%). We successfully developed a prototype ROSE-AI system with high diagnostic ability. Each data-augmentation technique may have various compatibilities with AI-mediated diagnostics, and the geometric transformation was the most effective for the ROSE-AI system.
備考	The version of record of this article, first published in Scientific Reports, is available online at Publisher’s website: http://dx.doi.org/10.1038/s41598-024-72312-3
発行日	2024-09-28
出版物タイトル	Scientific Reports
巻	14巻
号	1号
出版者	Nature Portfolio
開始ページ	22441
ISSN	2045-2322
資料タイプ	学術雑誌論文
言語	英語
OAI-PMH Set	岡山大学
著作権者	© The Author(s) 2024
論文のバージョン	publisher
PubMed ID	39341885
DOI	10.1038/s41598-024-72312-3
Web of Science KeyUT	001325241700059
関連URL	isVersionOf https://doi.org/10.1038/s41598-024-72312-3
ライセンス	http://creativecommons.org/licenses/by-nc-nd/4.0/
Citation	Fujii, Y., Uchida, D., Sato, R. et al. Effectiveness of data-augmentation on deep learning in evaluating rapid on-site cytopathology at endoscopic ultrasound-guided fine needle aspiration. Sci Rep 14, 22441 (2024). https://doi.org/10.1038/s41598-024-72312-3
助成機関名	Japan Society for the Promotion of Science Okayama prefecture
助成番号	23K11932 24K18948