{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T13:44:00Z","timestamp":1792158240967,"version":"4.4.0"},"reference-count":29,"publisher":"Wiley","issue":"10","license":[{"start":{"date-parts":[[2026,8,9]],"date-time":"2026-08-09T00:00:00Z","timestamp":1786233600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"},{"start":{"date-parts":[[2026,8,9]],"date-time":"2026-08-09T00:00:00Z","timestamp":1786233600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Transfusion"],"published-print":{"date-parts":[[2026,10]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Antibody identification testing is essential for preventing hemolytic transfusion reactions from unexpected red cell antibodies. Although automated interpretation is desirable to improve efficiency and consistency, conventional algorithmic approaches show limited accuracy, particularly in complex cases. Large language models (LLMs) offer a potential alternative; however, their performance is known to be sensitive to tabular data, which is a standard format for antigen profiles in antibody identification.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Study Design and Methods<\/jats:title>\n                    <jats:p>We generated 100 simulated antibody identification cases comprising single alloantibodies, dual alloantibodies, autoantibodies, and unidentifiable patterns. Seven LLMs (GPT\u20104o, GPT\u20104.1, o3, GPT\u20105, Gemini 2.0 Flash, Gemini 2.5 Flash, and Gemini 2.5 Pro) were evaluated. Antigen profile tables were provided using five text\u2010based formats (Markdown, CSV, JSON, list\u2010based, and sentence\u2010based serialization) and as uploaded spreadsheet files. Each model interpreted all cases 10 times per format, and accuracy was assessed overall and by antibody category.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Reasoning\u2010capable LLMs significantly outperformed non\u2010reasoning models, achieving a median accuracy of 84.5%. Among text\u2010based formats, sentence\u2010based serialization yielded the highest accuracy, whereas CSV performed poorest. The maximum accuracy difference between table formats reached 40.5 percentage points. GPT\u20105 and o3 performed comparably when antigen profiles were supplied as uploaded spreadsheet files, achieving accuracy similar to that of the sentence\u2010based format.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Reasoning LLMs demonstrated measurable baseline capability for interpreting antibody identification results when accompanied by appropriately structured antigen profile inputs. However, performance varied markedly with data representation and case complexity, and the use of LLMs in transfusion practice should remain limited to an assistive role under specialist supervision.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1111\/trf.70364","type":"journal-article","created":{"date-parts":[[2026,8,10]],"date-time":"2026-08-10T06:49:12Z","timestamp":1786344552000},"page":"1983-1992","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Interpreting antibody identification tests with large language models: Critical role of tabular data formatting"],"prefix":"10.1111","volume":"66","author":[{"given":"Jiwoo","family":"Lee","sequence":"first","affiliation":[{"name":"Department of Laboratory Medicine Seoul National University Hospital  Seoul Republic of Korea"},{"name":"Department of Laboratory Medicine Seoul National University College of Medicine  Seoul Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yousun","family":"Chung","sequence":"additional","affiliation":[{"name":"Department of Laboratory Medicine Kangdong Sacred Heart Hospital  Seoul Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dae\u2010Hyun","family":"Ko","sequence":"additional","affiliation":[{"name":"Department of Laboratory Medicine, Asan Medical Center University of Ulsan College of Medicine  Seoul Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eun Young","family":"Song","sequence":"additional","affiliation":[{"name":"Department of Laboratory Medicine Seoul National University Hospital  Seoul Republic of Korea"},{"name":"Department of Laboratory Medicine Seoul National University College of Medicine  Seoul Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hyungsuk","family":"Kim","sequence":"additional","affiliation":[{"name":"Department of Laboratory Medicine Seoul National University Hospital  Seoul Republic of Korea"},{"name":"Department of Laboratory Medicine Seoul National University College of Medicine  Seoul Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,8,9]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/B978-0-12-813726-0.00022-2"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.transci.2014.08.020"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.transci.2018.10.022"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10916-023-01925-4"},{"key":"e_1_2_8_6_1","unstructured":"OpenAI.Reasoning models.https:\/\/platform.openai.com\/docs\/guides\/reasoning?api-mode=responsesAccessed 14 July 2025."},{"key":"e_1_2_8_7_1","unstructured":"Google.Gemini thinking.https:\/\/ai.google.dev\/gemini-api\/docs\/thinkingAccessed 14 July 2025."},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1800"},{"key":"e_1_2_8_9_1","unstructured":"JahinA ZidanAH ZhangW BaoY LiuT.Evaluating mathematical reasoning across large language models: a fine\u2010grained approach. arXiv Preprint.2025arXiv:2503.10573."},{"key":"e_1_2_8_10_1","unstructured":"OpenAI.OpenAI o3 and o4\u2010mini System Card.2025."},{"key":"e_1_2_8_11_1","unstructured":"WangX WeiJ SchuurmansD LeQ ChiE NarangS et al.Self\u2010consistency improves chain of thought reasoning in language models. arXiv Preprint.2022arXiv:2203.11171."},{"key":"e_1_2_8_12_1","unstructured":"XuF HaoQ ZongZ WangJ ZhangY WangJ et al.Towards large reasoning models: a survey of reinforced reasoning with large language models. arXiv Preprint.2025arXiv:2501.09686."},{"key":"e_1_2_8_13_1","unstructured":"YaoZ LiuY ChenY ChenJ FangJ HouL et al.Are reasoning models more mrone to hallucination? arXiv Preprint.2025arXiv:2505.23646."},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-023-02448-8"},{"key":"e_1_2_8_15_1","doi-asserted-by":"publisher","DOI":"10.2196\/63430"},{"key":"e_1_2_8_16_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pdig.0000198"},{"key":"e_1_2_8_17_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-023-10213-1"},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMsr2214184"},{"key":"e_1_2_8_19_1","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-024-00944-1"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1111\/trf.17526"},{"key":"e_1_2_8_21_1","doi-asserted-by":"publisher","DOI":"10.1145\/3539618.3591708"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.1145\/3616855.3635752"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11704-024-40763-6"},{"key":"e_1_2_8_24_1","unstructured":"NguyenG BrugereI SharmaS KariyappaS NguyenAT LecueF.Interpretable LLM\u2010based table question answering. arXiv Preprint.2024arXiv:2412.12386."},{"key":"e_1_2_8_25_1","unstructured":"SarkarS LausenL.Testing the limits of unified sequence to sequence LLM pretraining on diverse table data tasks. arXiv Preprint.2023arXiv:2310.00789."},{"key":"e_1_2_8_26_1","unstructured":"WangZ ZhangH LiC\u2010L EisenschlosJM PerotV WangZ et al.Chain\u2010of\u2010table: evolving tables in the reasoning chain for table understanding. arXiv Preprint.2024arXiv:2401.04398."},{"issue":"3","key":"e_1_2_8_27_1","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3654979","article-title":"Table\u2010gpt: Table fine\u2010tuned gpt for diverse table tasks","volume":"2","author":"Li P","year":"2024","journal-title":"Proce ACM Manage Data"},{"key":"e_1_2_8_28_1","doi-asserted-by":"publisher","DOI":"10.1109\/FiCloud.2015.82"},{"key":"e_1_2_8_29_1","first-page":"5883","article-title":"Using JSON for data exchanging in web service applications","volume":"7","author":"Peng D","year":"2011","journal-title":"J Comput Inf Syst"},{"key":"e_1_2_8_30_1","unstructured":"Google.Gemini 2.5 Pro.https:\/\/cloud.google.com\/vertex-ai\/generative-ai\/docs\/models\/gemini\/2-5-pro?hl=ko. Accessed 14 July 2025"}],"container-title":["Transfusion"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/trf.70364","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1111\/trf.70364","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/trf.70364","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T12:44:54Z","timestamp":1792154694000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1111\/trf.70364"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,9]]},"references-count":29,"journal-issue":{"issue":"10","published-print":{"date-parts":[[2026,10]]}},"alternative-id":["10.1111\/trf.70364"],"URL":"https:\/\/doi.org\/10.1111\/trf.70364","archive":["Portico"],"relation":{},"ISSN":["0041-1132","1537-2995"],"issn-type":[{"value":"0041-1132","type":"print"},{"value":"1537-2995","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,9]]},"assertion":[{"value":"2026-01-29","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-07-07","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-09","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}