{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T17:57:38Z","timestamp":1791827858683,"version":"4.3.4"},"reference-count":33,"publisher":"SAGE Publications","issue":"4","license":[{"start":{"date-parts":[[2026,3,3]],"date-time":"2026-03-03T00:00:00Z","timestamp":1772496000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"},{"start":{"date-parts":[[2026,3,3]],"date-time":"2026-03-03T00:00:00Z","timestamp":1772496000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/journals.sagepub.com\/page\/policies\/text-and-data-mining-license"}],"content-domain":{"domain":["journals.sagepub.com"],"crossmark-restriction":true},"short-container-title":["Journal of VitreoRetinal Diseases"],"published-print":{"date-parts":[[2026,7]]},"abstract":"<jats:p>\n                    <jats:bold>Purpose:<\/jats:bold>\n                    To evaluate the diagnostic accuracy of GPT-4.o, Claude 3.7 Sonnet, and Gemini 2.5 Pro for real-world retinal cases.\n                    <jats:bold>Methods:<\/jats:bold>\n                    Forty retina cases from the University of Iowa\u2019s EyeRounds repository were assessed under 2 conditions: (1) full clinical context (textual history, examination, and image descriptions); and (2) image-only (raw clinical images). Outputs were evaluated for diagnostic accuracy, symptom\/sign recognition, differential diagnosis, and treatment recommendations. Cochran Q, McNemar test, analysis of variance, and Cohen kappa were used to make statistical comparisons.\n                    <jats:bold>Results:<\/jats:bold>\n                    In full-context cases, GPT-4.o achieved the highest diagnostic accuracy (78.4%), followed by Claude (73.0%) and Gemini (29.7%) (\n                    <jats:italic toggle=\"yes\">P<\/jats:italic>\n                    &lt; .001). In image-only cases, Claude was superior (73.7%), outperforming GPT-4.o (63.2%) and Gemini (31.6%) (\n                    <jats:italic toggle=\"yes\">P<\/jats:italic>\n                    &lt; .001). GPT-4.o was best at identifying signs\/symptoms in full-context cases (64.0%), while Claude excelled in image-only cases (63.3%). GPT-4.o (42.9%) and Claude (37.0%) outperformed Gemini (25.9%) on differential diagnosis (\n                    <jats:italic toggle=\"yes\">P<\/jats:italic>\n                    \u2264 .011). Claude showed superior treatment accuracy in image-only cases (61.1%), while GPT-4.o was superior in full-context cases (57.3%), and both outperformed Gemini (35.8%). GPT-4.o and Claude showed substantial agreement in image-based diagnoses (\u03ba=0.658), while Gemini showed minimal agreement (\u03ba\u22640.196).\n                    <jats:bold>Conclusions:<\/jats:bold>\n                    GPT-4.o and Claude demonstrated strong diagnostic and clinical reasoning in retinal cases, with Claude excelling in image-based analysis and GPT-4.o in text-rich contexts. Gemini\u2019s lower performance underscores the importance of careful model selection in clinical applications.\n                  <\/jats:p>","DOI":"10.1177\/24741264261423759","type":"journal-article","created":{"date-parts":[[2026,3,11]],"date-time":"2026-03-11T18:38:13Z","timestamp":1773254293000},"page":"732-739","update-policy":"https:\/\/doi.org\/10.1177\/sage-journals-update-policy","source":"Crossref","is-referenced-by-count":3,"title":["Multimodal Diagnostic Accuracy of GPT-4.o, Claude 3.7, and Gemini 2.5 on Real-World Retina Cases"],"prefix":"10.1177","volume":"10","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-4140-696X","authenticated-orcid":false,"given":"Iden","family":"Amiri","sequence":"first","affiliation":[{"name":"Department of Ophthalmology, University of North Carolina-Chapel Hill, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Cheng","family":"Jiao","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, University of North Carolina-Chapel Hill, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5194-5256","authenticated-orcid":false,"given":"Alice Yang","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, University of North Carolina-Chapel Hill, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"179","published-online":{"date-parts":[[2026,3,3]]},"reference":[{"key":"e_1_3_3_2_2","doi-asserted-by":"publisher","DOI":"10.1155\/2018\/5278196"},{"key":"e_1_3_3_3_2","doi-asserted-by":"publisher","DOI":"10.1136\/bjophthalmol-2018-313173"},{"key":"e_1_3_3_4_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-020-0867-7"},{"key":"e_1_3_3_5_2","doi-asserted-by":"publisher","DOI":"10.1016\/S2214-109X(13)70113-X"},{"key":"e_1_3_3_6_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ophtha.2014.05.013"},{"key":"e_1_3_3_7_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-018-0107-6"},{"key":"e_1_3_3_8_2","doi-asserted-by":"publisher","DOI":"10.1136\/bjophthalmol-2021-319030"},{"key":"e_1_3_3_9_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06291-2"},{"key":"e_1_3_3_10_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pdig.0000198"},{"key":"e_1_3_3_11_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.xops.2024.100600"},{"key":"e_1_3_3_12_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.xops.2024.100667"},{"key":"e_1_3_3_13_2","doi-asserted-by":"publisher","DOI":"10.1007\/s40123-023-00805-x"},{"key":"e_1_3_3_14_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.patter.2024.100943"},{"key":"e_1_3_3_15_2","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2024.37711"},{"key":"e_1_3_3_16_2","doi-asserted-by":"publisher","DOI":"10.1001\/jamaophthalmol.2024.0017"},{"key":"e_1_3_3_17_2","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2016.17216"},{"key":"e_1_3_3_18_2","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2017.18152"},{"key":"e_1_3_3_19_2","doi-asserted-by":"publisher","DOI":"10.12688\/f1000research.8996.2"},{"key":"e_1_3_3_20_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.preteyeres.2018.07.004"},{"key":"e_1_3_3_21_2","unstructured":"EyeRounds. Case reports tutorials videos images from Univ of Iowa Dept of Ophthalmology. Accessed June 12 2025. https:\/\/www.eyerounds.org\/#gsc.tab=0"},{"key":"e_1_3_3_22_2","doi-asserted-by":"publisher","DOI":"10.1001\/jamaophthalmol.2023.6917"},{"key":"e_1_3_3_23_2","unstructured":"Holmes J Ye S Li Y et al. Evaluating large language models in ophthalmology. arXiv [preprint]. Published online November 7 2023. arXiv:2311.04933. https:\/\/arxiv.org\/abs\/2311.04933"},{"key":"e_1_3_3_24_2","doi-asserted-by":"publisher","DOI":"10.7759\/cureus.78069"},{"key":"e_1_3_3_25_2","doi-asserted-by":"publisher","DOI":"10.1016\/S2589-7500(21)00132-1"},{"key":"e_1_3_3_26_2","doi-asserted-by":"crossref","unstructured":"Pal A Sankarasubbu M. Gemini goes to med school: exploring the capabilities of multimodal large language models on medical challenge problems & hallucinations. In: Naumann T Ben Abacha A Bethard S eds. Proceedings of the 6th Clinical Natural Language Processing Workshop. Association for Computational Linguistics; 2024:21-46. https:\/\/aclanthology.org\/2024.clinicalnlp-1.3\/","DOI":"10.18653\/v1\/2024.clinicalnlp-1.3"},{"key":"e_1_3_3_27_2","doi-asserted-by":"crossref","unstructured":"Yan Q He X Yue X Wang XE. Worse than random? an embarrassingly simple probing evaluation of large multimodal models in medical VQA. In: Che W Nabende J Shutova E eds. Findings of the Association for Computational Linguistics: ACL 2025. Association for Computational Linguistics; 2025:19188-19205. https:\/\/aclanthology.org\/2025.findings-acl.981\/","DOI":"10.18653\/v1\/2025.findings-acl.981"},{"key":"e_1_3_3_28_2","doi-asserted-by":"publisher","DOI":"10.1097\/IAE.0000000000004351"},{"key":"e_1_3_3_29_2","unstructured":"Anthropic. Introducing the next generation of Claude. March 4 2024. Accessed June 8 2025. https:\/\/www.anthropic.com\/news\/claude-3-family"},{"key":"e_1_3_3_30_2","unstructured":"Oladele S. GPT-4o vs. Gemini 1.5 Pro vs. Claude 3 Opus: multimodal AI model comparison. May 15 2024. Accessed June 8 2025. https:\/\/encord.com\/blog\/gpt-4o-vs-gemini-vs-claude-3-opus\/"},{"key":"e_1_3_3_31_2","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics15101221"},{"key":"e_1_3_3_32_2","doi-asserted-by":"publisher","DOI":"10.1186\/s13054-023-04473-y"},{"key":"e_1_3_3_33_2","doi-asserted-by":"publisher","DOI":"10.1136\/bjophthalmol-2022-321141"},{"key":"e_1_3_3_34_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijmedinf.2025.105886"}],"container-title":["Journal of VitreoRetinal Diseases"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/journals.sagepub.com\/doi\/pdf\/10.1177\/24741264261423759","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/journals.sagepub.com\/doi\/full-xml\/10.1177\/24741264261423759","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/journals.sagepub.com\/doi\/pdf\/10.1177\/24741264261423759","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T15:26:40Z","timestamp":1790695600000},"score":1,"resource":{"primary":{"URL":"https:\/\/journals.sagepub.com\/doi\/10.1177\/24741264261423759"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,3]]},"references-count":33,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,7]]}},"alternative-id":["10.1177\/24741264261423759"],"URL":"https:\/\/doi.org\/10.1177\/24741264261423759","relation":{},"ISSN":["2474-1264","2474-1272"],"issn-type":[{"value":"2474-1264","type":"print"},{"value":"2474-1272","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,3]]}}}