{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T04:54:33Z","timestamp":1777870473467,"version":"3.51.4"},"posted":{"date-parts":[[2026,4,28]]},"group-title":"Health Informatics","reference-count":20,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T00:00:00Z","timestamp":1777334400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,4,28]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Retinal fluids, detectable through optical coherence tomography (OCT), are key biomarkers for retinal diseases such as diabetic macular edema and age-related macular degeneration, guiding treatment decisions and monitoring response to therapy. Automated segmentation of retinal fluids could support large-scale clinical research and the development of clinical decision support tools. Recent ophthalmic foundation models trained on massive retinal imaging datasets show promise across many downstream tasks, including disease risk prediction and biomarker segmentation, but their performance relative to task-specific models for specialized clinical tasks remains unclear. We compared a task-specific segmentation model (RetiFluidNet) and an ophthalmic foundation model (VisionFM) using a standard benchmarking dataset containing 4,248 OCT images from 48 patients with three retinal diseases. Models were evaluated using three-fold cross-validation and assessed for pixel-level segmentation accuracy and patient-level fluid burden estimation. The task-specific model achieved higher segmentation performance and more consistent fluid quantification across devices. These findings suggest that, for retinal fluid segmentation, specialized task-specific models currently remain more reliable than general-purpose foundation models, highlighting the need for targeted adaptation before clinical deployment.<\/jats:p>","DOI":"10.64898\/2026.04.26.26351792","type":"posted-content","created":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T22:10:42Z","timestamp":1777414242000},"source":"Crossref","is-referenced-by-count":0,"title":["One Size Fits All? Comparing Foundation and Task-specific Models for Retinal Fluid Segmentation"],"prefix":"10.64898","author":[{"given":"Xiaoyu","family":"Sun","sequence":"first","affiliation":[{"name":"Institute for Informatics, Data Science and Biostatistics, Washington University in St. Louis, St. Louis, MO, USA"},{"name":"Pitzer College, Claremont, CA, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Saiyu","family":"You","sequence":"additional","affiliation":[{"name":"Institute for Informatics, Data Science and Biostatistics, Washington University in St. Louis, St. Louis, MO, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Siqi","family":"Sun","sequence":"additional","affiliation":[{"name":"Institute for Informatics, Data Science and Biostatistics, Washington University in St. Louis, St. Louis, MO, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Cindy X.","family":"Cai","sequence":"additional","affiliation":[{"name":"Wilmer Eye Institute, Johns Hopkins School of Medicine, Baltimore, MD, USA"},{"name":"Department of Biomedical Informatics and Data Science, Johns Hopkins School of Medicine, Baltimore, MD, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Joanna","family":"Abraham","sequence":"additional","affiliation":[{"name":"Institute for Informatics, Data Science and Biostatistics, Washington University in St. Louis, St. Louis, MO, USA"},{"name":"Department of Anesthesiology, Washington University in St. Louis, St. Louis, MO, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Po-Yin","family":"Yen","sequence":"additional","affiliation":[{"name":"Institute for Informatics, Data Science and Biostatistics, Washington University in St. Louis, St. Louis, MO, USA"},{"name":"Department of Medicine, Washington University in St. Louis, St. Louis, MO, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4356-4645","authenticated-orcid":false,"given":"Linying","family":"Zhang","sequence":"additional","affiliation":[{"name":"Institute for Informatics, Data Science and Biostatistics, Washington University in St. Louis, St. Louis, MO, USA"},{"name":"Department of Medicine, Washington University in St. Louis, St. Louis, MO, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"key":"2026043009050563000_2026.04.26.26351792v1.1","article-title":"On the opportunities and risks of foundation models","year":"2021","journal-title":"arXiv [Preprint]"},{"key":"2026043009050563000_2026.04.26.26351792v1.2","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2023.102996"},{"key":"2026043009050563000_2026.04.26.26351792v1.3","doi-asserted-by":"publisher","DOI":"10.1186\/s40662-015-0026-2"},{"key":"2026043009050563000_2026.04.26.26351792v1.4","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2022.3228285"},{"key":"2026043009050563000_2026.04.26.26351792v1.5","doi-asserted-by":"publisher","DOI":"10.1056\/AIoa2300221"},{"key":"2026043009050563000_2026.04.26.26351792v1.6","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2019.2901398"},{"key":"2026043009050563000_2026.04.26.26351792v1.7","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.106391"},{"key":"2026043009050563000_2026.04.26.26351792v1.8","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929"},{"key":"2026043009050563000_2026.04.26.26351792v1.9","unstructured":"Zhou J , Wei C , Wang H , Shen W , Xie C , Yuille A , Kong T. iBOT: Image BERT pre-training with online tokenizer. In: International Conference on Learning Representations (ICLR). 2022."},{"key":"2026043009050563000_2026.04.26.26351792v1.10","doi-asserted-by":"crossref","unstructured":"Hatamizadeh A , Tang Y , Nath V , et al. UNETR: Transformers for 3D medical image segmentation. In: Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision (WACV). 2022. p. 574\u2013584.","DOI":"10.1109\/WACV51458.2022.00181"},{"key":"2026043009050563000_2026.04.26.26351792v1.11","article-title":"MONAI: an open-source framework for deep learning in healthcare","year":"2022","journal-title":"arXiv [Preprint]"},{"key":"2026043009050563000_2026.04.26.26351792v1.12","doi-asserted-by":"publisher","DOI":"10.21037\/qims-21-86"},{"key":"2026043009050563000_2026.04.26.26351792v1.13","doi-asserted-by":"publisher","DOI":"10.1016\/j.jcm.2016.02.012"},{"key":"2026043009050563000_2026.04.26.26351792v1.14","doi-asserted-by":"publisher","DOI":"10.21105\/joss.01026"},{"key":"2026043009050563000_2026.04.26.26351792v1.15","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2308.09331"},{"key":"2026043009050563000_2026.04.26.26351792v1.16","doi-asserted-by":"crossref","unstructured":"Kirillov A , Mintun E , Ravi N , et al. Segment anything. In: Proceedings of the IEEE\/CVF International Conference on Computer Vision (ICCV); 2023. p. 4015\u20134026.","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"2026043009050563000_2026.04.26.26351792v1.17","doi-asserted-by":"publisher","DOI":"10.3390\/bioengineering11090940"},{"key":"2026043009050563000_2026.04.26.26351792v1.18","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-025-01852-3"},{"key":"2026043009050563000_2026.04.26.26351792v1.19","doi-asserted-by":"publisher","DOI":"10.1038\/s41433-023-02615-8"},{"key":"2026043009050563000_2026.04.26.26351792v1.20","doi-asserted-by":"publisher","DOI":"10.1016\/j.ophtha.2017.10.031"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.04.26.26351792","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T16:05:14Z","timestamp":1777565114000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.64898\/2026.04.26.26351792"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,28]]},"references-count":20,"URL":"https:\/\/doi.org\/10.64898\/2026.04.26.26351792","relation":{},"subject":[],"published":{"date-parts":[[2026,4,28]]},"subtype":"preprint"}}