{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T15:02:26Z","timestamp":1789570946354,"version":"build-2803163510"},"reference-count":31,"publisher":"MDPI AG","issue":"18","license":[{"start":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T00:00:00Z","timestamp":1789516800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Cancers"],"abstract":"<jats:p>Background\/Objectives: Cross-institutional transfer of endoscopic deep learning models may be limited by acquisition and preprocessing differences. We reassessed persistent domain shift after correcting Facility B input linkage and validation design. Methods: A preflight audit corrected two Facility B image\u2013mask linkage omissions without changing the 402 source lesions or 469 still images. Criterion B was a qualitatively selected, Facility-B-informed HSV patch-selection adjustment; the same Tenengrad focus criterion was used at both facilities. Class-wise MMD2 was calculated using all three Facility A fold-specific feature extractors. Each Facility A source model was evaluated on the same corrected criterion B Facility B cohort. Facility B models were initialized from ImageNet and fully fine-tuned for four prespecified epochs under lesion-grouped three-fold cross-validation, with pooled out-of-fold patch-level evaluation. Historical Facility A internal results were descriptive only. Results: Criterion B retained 142,878 Facility B patches. Residual MMD2 was greatest for SM1 (0.185) and SM2 (0.147). Cross-site accuracy, macro-F1, and balanced accuracy were 0.711, 0.397, and 0.422; site-specific values were 0.770, 0.472, and 0.473. Recall improved for BG, LG, and SM1 but decreased for HG and SM2. Conclusions: Site-specific development partially improved aggregate local performance, but persistent domain shift and class-specific limitations remained; universal cross-institutional generalizability was not established.<\/jats:p>","DOI":"10.3390\/cancers18183002","type":"journal-article","created":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T13:50:27Z","timestamp":1789566627000},"page":"3002","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Persistent Cross-Institutional Domain Shift in Narrow-Band Imaging Classification of Colorectal Lesions and the Utility of Site-Specific Fine-Tuning"],"prefix":"10.3390","volume":"18","author":[{"given":"Hiroki","family":"Nomiya","sequence":"first","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5163-9563","authenticated-orcid":false,"given":"Hidezumi","family":"Kikuchi","sequence":"additional","affiliation":[{"name":"Division of Endoscopy, Hirosaki University Hospital, 53 Honcho, Hirosaki 036-8563, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2908-6007","authenticated-orcid":false,"given":"Takeshi","family":"Shimizu","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Sendai City Medical Center, Sendai Open Hospital, 5-22-1 Tsurugaya, Miyagino-Ku, Sendai 983-0824, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Saki","family":"Nakano","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Taka","family":"Asari","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yohei","family":"Sawada","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shohei","family":"Igarashi","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6232-4727","authenticated-orcid":false,"given":"Naoki","family":"Higuchi","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"},{"name":"Department of Gastroenterology, Misawa City Hospital, 164-65 Horiguchi, Misawa 033-0022, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tetsuya","family":"Tatsuta","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1690-5923","authenticated-orcid":false,"given":"Daisuke","family":"Chinda","sequence":"additional","affiliation":[{"name":"Division of Endoscopy, Hirosaki University Hospital, 53 Honcho, Hirosaki 036-8563, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7856-5277","authenticated-orcid":false,"given":"Yoshihiro","family":"Sasaki","sequence":"additional","affiliation":[{"name":"Department of Community Medicine, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6075-1282","authenticated-orcid":false,"given":"Hirotake","family":"Sakuraba","sequence":"additional","affiliation":[{"name":"Department of Gastroenterology, Hematology, and Clinical Immunology, Hirosaki University Graduate School of Medicine, 5 Zaifucho, Hirosaki 036-8562, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,9,16]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"7510","DOI":"10.1038\/s41598-023-34750-3","article-title":"A trial deep learning-based model for four-class histologic classification of colonic tumor from narrow band imaging","volume":"13","author":"Shimizu","year":"2023","journal-title":"Sci. Rep."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"568","DOI":"10.1053\/j.gastro.2017.10.010","article-title":"Accurate classification of diminutive colorectal polyps using computer-aided analysis","volume":"154","author":"Chen","year":"2018","journal-title":"Gastroenterology"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"94","DOI":"10.1136\/gutjnl-2017-314547","article-title":"Real-time differentiation of adenomatous and hyperplastic diminutive colorectal polyps during analysis of unaltered videos of standard colonoscopy using a deep learning model","volume":"68","author":"Byrne","year":"2019","journal-title":"Gut"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"30","DOI":"10.1038\/s41598-019-56697-0","article-title":"Endoscopic diagnosis and treatment planning for colorectal polyps using a deep-learning model","volume":"10","author":"Song","year":"2020","journal-title":"Sci. Rep."},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Lu, Y., Wu, J., Zhuo, X., Hu, M., Chen, Y., Luo, Y., Feng, Y., Zhi, M., Li, C., and Sun, J. (2022). Real-time artificial intelligence-based histologic classifications of colorectal polyps using narrow-band imaging. Front. Oncol., 12.","DOI":"10.3389\/fonc.2022.879239"},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Young, E., Edwards, L., and Singh, R. (2023). The role of artificial intelligence in colorectal cancer screening: Lesion detection and lesion characterization. Cancers, 15.","DOI":"10.3390\/cancers15215126"},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Zech, J.R., Badgeley, M.A., Liu, M., Costa, A.B., Titano, J.J., and Oermann, E.K. (2018). Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: A cross-sectional study. PLoS Med., 15.","DOI":"10.1371\/journal.pmed.1002683"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"1173","DOI":"10.1109\/TBME.2021.3117407","article-title":"Domain adaptation for medical image analysis: A survey","volume":"69","author":"Guan","year":"2022","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"283","DOI":"10.1056\/NEJMc2104626","article-title":"The clinician and dataset shift in artificial intelligence","volume":"385","author":"Finlayson","year":"2021","journal-title":"N. Engl. J. Med."},{"key":"ref_10","first-page":"4396","article-title":"Domain generalization: A survey","volume":"45","author":"Zhou","year":"2023","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"109256","DOI":"10.1016\/j.compbiomed.2024.109256","article-title":"A systematic review of generalization research in medical image classification","volume":"183","author":"Matta","year":"2024","journal-title":"Comput. Biol. Med."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Kelly, C.J., Karthikesalingam, A., Suleyman, M., Corrado, G., and King, D. (2019). Key challenges for delivering clinical impact with artificial intelligence. BMC Med., 17.","DOI":"10.1186\/s12916-019-1426-2"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"453","DOI":"10.1111\/den.14489","article-title":"Updates in narrow-band imaging for colorectal polyps: Narrow-band imaging generations, detection, diagnosis, and artificial intelligence","volume":"35","author":"Teramoto","year":"2023","journal-title":"Dig. Endosc."},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Park, J., Hwang, Y., Kim, H.G., Lee, J.S., Kim, J.-O., Lee, T.H., Jeon, S.R., Hong, S.J., Ko, B.M., and Kim, S. (2022). Reduced detection rate of artificial intelligence in images obtained from untrained endoscope models and improvement using domain adaptation algorithm. Front. Med., 9.","DOI":"10.3389\/fmed.2022.1036974"},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Krotkov, E. (1989). Active Computer Vision by Cooperative Focus and Stereo, Springer.","DOI":"10.1007\/978-1-4613-9663-5"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"1415","DOI":"10.1016\/j.patcog.2012.11.011","article-title":"Analysis of focus measure operators for shape-from-focus","volume":"46","author":"Pertuz","year":"2013","journal-title":"Pattern Recognit."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27\u201330 June 2016, IEEE.","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref_18","unstructured":"Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., and Antiga, L. (2019, January 8\u201314). PyTorch: An imperative style, high-performance deep learning library. Proceedings of the Advances in Neural Information Processing Systems 32, Vancouver, BC, Canada."},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). ImageNet: A large-scale hierarchical image database. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Miami, FL, USA, 20\u201325 June 2009, IEEE.","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref_20","first-page":"723","article-title":"A kernel two-sample test","volume":"13","author":"Gretton","year":"2012","journal-title":"J. Mach. Learn. Res."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"102309","DOI":"10.1016\/j.media.2021.102309","article-title":"Embracing the disharmony in medical imaging: A simple and effective framework for domain adaptation","volume":"76","author":"Wang","year":"2022","journal-title":"Med. Image Anal."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Kim, M.J., Kim, S.H., Kim, S.M., Nam, J.H., Hwang, Y.B., and Lim, Y.J. (2023). The advent of domain adaptation into artificial intelligence for gastrointestinal endoscopy and medical imaging. Diagnostics, 13.","DOI":"10.3390\/diagnostics13193023"},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Zhu, J.-Y., Park, T., Isola, P., and Efros, A.A. (2017). Unpaired image-to-image translation using cycle-consistent adversarial networks. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy, 22\u201329 October 2017, IEEE.","DOI":"10.1109\/ICCV.2017.244"},{"key":"ref_24","first-page":"1","article-title":"Domain-adversarial training of neural networks","volume":"17","author":"Ganin","year":"2016","journal-title":"J. Mach. Learn. Res."},{"key":"ref_25","unstructured":"Sun, B., and Saenko, K. (15\u201316, January 8\u201310). Deep CORAL: Correlation alignment for deep domain adaptation. Proceedings of the European Conference on Computer Vision Workshops (ECCV Workshops), Amsterdam, The Netherlands."},{"key":"ref_26","unstructured":"Long, M., Cao, Y., Wang, J., and Jordan, M.I. (2015, January 6\u201311). Learning transferable features with deep adaptation networks. Proceedings of the 32nd International Conference on Machine Learning, Lille, France."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"70","DOI":"10.1097\/SLA.0000000000002693","article-title":"Artificial intelligence in surgery: Promises and perils","volume":"268","author":"Hashimoto","year":"2018","journal-title":"Ann. Surg."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"1054","DOI":"10.1038\/s41591-019-0462-y","article-title":"Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer","volume":"25","author":"Kather","year":"2019","journal-title":"Nat. Med."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"1299","DOI":"10.1109\/TMI.2016.2535302","article-title":"Convolutional neural networks for medical image analysis: Full training or fine tuning?","volume":"35","author":"Tajbakhsh","year":"2016","journal-title":"IEEE Trans. Med. Imaging"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"119","DOI":"10.1038\/s41746-020-00323-1","article-title":"The future of digital health with federated learning","volume":"3","author":"Rieke","year":"2020","journal-title":"npj Digit. Med."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"12598","DOI":"10.1038\/s41598-020-69250-1","article-title":"Federated learning in medicine: Facilitating multi-institutional collaborations without sharing patient data","volume":"10","author":"Sheller","year":"2020","journal-title":"Sci. Rep."}],"container-title":["Cancers"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2072-6694\/18\/18\/3002\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T14:04:48Z","timestamp":1789567488000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2072-6694\/18\/18\/3002"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,16]]},"references-count":31,"journal-issue":{"issue":"18","published-online":{"date-parts":[[2026,9]]}},"alternative-id":["cancers18183002"],"URL":"https:\/\/doi.org\/10.3390\/cancers18183002","relation":{},"ISSN":["2072-6694"],"issn-type":[{"value":"2072-6694","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,16]]}}}