{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,30]],"date-time":"2026-10-30T09:35:56Z","timestamp":1793352956400,"version":"4.5.0"},"reference-count":18,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,22]],"date-time":"2026-09-22T00:00:00Z","timestamp":1790035200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100018624","name":"Science and Technology Bureau of Quanzhou","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100018624","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"\n Purpose<\/jats:title>\n Large language models (LLMs) have demonstrated impressive performance across various medical examinations; however, their competency in subspecialty clinical reasoning, particularly in prenatal diagnosis, remains poorly characterized.<\/jats:p>\n <\/jats:sec>\n \n Materials and methods<\/jats:title>\n \n A cross-sectional comparative study was conducted by using 125 questions from China's nationally standardized Prenatal Ultrasound Screening Qualification Examination. Three LLMs and two human physicians were evaluated. Questions were classified along three independent taxonomic axes: modality, disease system, and question format. Cochran's\n Q<\/jats:italic>\n test was used for omnibus comparisons, with\n post-hoc<\/jats:italic>\n pairwise McNemar tests where were significant.\n <\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n Overall accuracy did not differ significantly among the five test-takers (\n p<\/jats:italic>\n \u2009=\u20090.126): the chief physician scored highest (89.60%, 112\/125), followed by Claude-Sonnet-4.6 (85.60%, 107\/125), Grok-4 (84.80%, 106\/125), ChatGPT-5 (84.00%, 105\/125), and the attending physician (80.00%, 100\/125). No significant differences were observed when stratified by modality (text-only,\n p<\/jats:italic>\n \u2009=\u20090.146; imaging diagnostic questions,\n p<\/jats:italic>\n \u2009=\u20090.603) or by any individual disease system (all\n p<\/jats:italic>\n \u2009&gt;\u20090.05). However, performance divergence emerged sharply across question types. In clinical vignette-based questions (\n n<\/jats:italic>\n \u2009=\u200920), the chief physician (95.00%) significantly outperformed all other test-takers (\n p<\/jats:italic>\n \u2009=\u20090.016), with a 25% gap over the best-performing LLM (Grok-4, 70.00%). In sequential item sets, Grok-4 (95.74%) and the chief physician (93.62%) both significantly outperformed the attending physician (76.60%; both\n p<\/jats:italic>\n \u2009&lt;\u20090.05). Conversely, LLMs excelled in knowledge-retrieval tasks, with ChatGPT-5 and Claude-Sonnet-4.6 both achieving 100% accuracy in matching questions. No significant differences were detected among test-takers for factual recall, matching, or multiple-choice multiple-answer questions (all\n p<\/jats:italic>\n \u2009&gt;\u20090.05).\n <\/jats:p>\n <\/jats:sec>\n \n Conclusion<\/jats:title>\n LLMs are well-positioned to serve as powerful knowledge augmentation tools, but are not yet ready to function as autonomous clinical reasoners in this complex, ethically consequential subspecialty.<\/jats:p>\n <\/jats:sec>","DOI":"10.3389\/fdgth.2026.1917573","type":"journal-article","created":{"date-parts":[[2026,9,22]],"date-time":"2026-09-22T13:15:24Z","timestamp":1790082924000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Discrepancy between medical knowledge and clinical reasoning: evaluating large language models in prenatal diagnosis"],"prefix":"10.3389","volume":"8","author":[{"given":"Qiuling","family":"Guo","sequence":"first","affiliation":[{"name":"Department of Ultrasound, Quanzhou Maternal and Child Health Hospital (Quanzhou Children\u2019s Hospital)","place":["Quanzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yanyun","family":"Du","sequence":"additional","affiliation":[{"name":"Department of Ultrasound, Quanzhou Maternal and Child Health Hospital (Quanzhou Children\u2019s Hospital)","place":["Quanzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jingcheng","family":"Dai","sequence":"additional","affiliation":[{"name":"Department of Gynecology and Obstetrics, Quanzhou Maternal and Child Health Hospital (Quanzhou Children\u2019s Hospital)","place":["Quanzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Weiru","family":"Lin","sequence":"additional","affiliation":[{"name":"Department of Ultrasound, Quanzhou Maternal and Child Health Hospital (Quanzhou Children\u2019s Hospital)","place":["Quanzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,22]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"CD014715","DOI":"10.1002\/14651858.CD014715.pub2","article-title":"Diagnostic accuracy of ultrasound screening for fetal structural abnormalities during the first and second trimester of pregnancy in low-risk and unselected populations","volume":"5","author":"Buijtendijk","year":"2024","journal-title":"Cochrane Database Syst Rev"},{"key":"B2","doi-asserted-by":"publisher","first-page":"11","DOI":"10.1002\/uog.23740","article-title":"First-trimester ultrasound detection of fetal heart anomalies: systematic review and meta-analysis","volume":"59","author":"Karim","year":"2022","journal-title":"Ultrasound Obstet Gynecol"},{"key":"B3","doi-asserted-by":"publisher","first-page":"446","DOI":"10.1002\/uog.27576","article-title":"Diagnostic accuracy of prenatal ultrasound in coarctation of aorta: systematic review and individual participant data meta-analysis","volume":"63","author":"Villala\u00edn","year":"2024","journal-title":"Ultrasound Obstet Gynecol"},{"key":"B4","doi-asserted-by":"publisher","first-page":"e0000198","DOI":"10.1371\/journal.pdig.0000198","article-title":"Performance of ChatGPT on USMLE: potential for AI-assisted medical education using large language models","volume":"2","author":"Kung","year":"2023","journal-title":"PLoS Digit Health"},{"key":"B5","doi-asserted-by":"publisher","first-page":"104717","DOI":"10.1016\/j.ijnurstu.2024.104717","article-title":"Assessing question characteristic influences on ChatGPT\u2019s performance and response-explanation consistency: insights from Taiwan\u2019s nursing licensing exam","volume":"153","author":"Su","year":"2024","journal-title":"Int J Nurs Stud"},{"key":"B6","doi-asserted-by":"publisher","first-page":"e80874","DOI":"10.7759\/cureus.80874","article-title":"Comparative analysis of ChatGPT-4o and Gemini advanced performance on diagnostic radiology in-training exams","volume":"17","author":"Huang","year":"2025","journal-title":"Cureus"},{"key":"B7","doi-asserted-by":"publisher","first-page":"478","DOI":"10.1002\/uog.29186","article-title":"Evaluation of first-trimester ultrasound screening strategy for fetal congenital heart disease","volume":"65","author":"Yang","year":"2025","journal-title":"Ultrasound Obstet Gynecol"},{"key":"B8","doi-asserted-by":"publisher","first-page":"575","DOI":"10.1002\/jcu.23673","article-title":"Prenatal ultrasound findings and prenatal diagnosis of fetal skeletal dysplasia","volume":"52","author":"Li","year":"2024","journal-title":"J Clin Ultrasound"},{"key":"B9","doi-asserted-by":"publisher","first-page":"843","DOI":"10.1007\/s00404-023-07267-3","article-title":"The value of combined detailed first-trimester ultrasound-biochemical analysis for screening fetal aneuploidy in the era of non-invasive prenatal testing","volume":"310","author":"Ye","year":"2024","journal-title":"Arch Gynecol Obstet"},{"key":"B10","doi-asserted-by":"publisher","first-page":"81","DOI":"10.1002\/uog.20856","article-title":"Ratio of fetal choroid plexus to head size: simple sonographic marker of open spina bifida at 11\u201313 weeks\u2019 gestation","volume":"55","author":"Chaoui","year":"2020","journal-title":"Ultrasound Obstet Gynecol"},{"key":"B11","doi-asserted-by":"publisher","first-page":"52","DOI":"10.1016\/j.placenta.2024.08.006","article-title":"Magnetic resonance imaging diagnosis of abnormal placental cord insertions","volume":"155","author":"Tian","year":"2024","journal-title":"Placenta"},{"key":"B12","doi-asserted-by":"publisher","first-page":"1296615","DOI":"10.3389\/fmed.2023.1296615","article-title":"ChatGPT\u2019s performance in German OB\/GYN exams\u2014paving the way for AI-enhanced medical education and clinical practice","volume":"10","author":"Riedel","year":"2023","journal-title":"Front Med"},{"key":"B13","doi-asserted-by":"publisher","first-page":"114960","DOI":"10.1016\/j.ejogrb.2026.114960","article-title":"Assessment of AI-generated responses to common pregnancy questions: a blinded expert evaluation","volume":"318","author":"Atay","year":"2026","journal-title":"Eur J Obstet Gynecol Reprod Biol"},{"key":"B14","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1038\/s41586-023-06291-2","article-title":"Large language models encode clinical knowledge","volume":"620","author":"Singhal","year":"2023","journal-title":"Nature"},{"key":"B15","doi-asserted-by":"publisher","first-page":"461","DOI":"10.1080\/0142159X.2018.1435855","article-title":"Training medical students for the twenty-first century: rationale and development of the Utrecht curriculum \u201cCRU+\u201d","volume":"40","author":"Ten Cate","year":"2018","journal-title":"Med Teach"},{"key":"B16","doi-asserted-by":"publisher","first-page":"e0000341","DOI":"10.1371\/journal.pdig.0000341","article-title":"Large language models approach expert-level clinical knowledge and reasoning in ophthalmology: a head-to-head cross-sectional study","volume":"3","author":"Thirunavukarasu","year":"2024","journal-title":"PLoS Digit Health"},{"key":"B17","doi-asserted-by":"publisher","first-page":"137","DOI":"10.1016\/j.clinimag.2023.06.008","article-title":"Decoding radiology reports: potential application of OpenAI ChatGPT to enhance patient understanding of diagnostic reports","volume":"101","author":"Li","year":"2023","journal-title":"Clin Imaging"},{"key":"B18","doi-asserted-by":"publisher","first-page":"548","DOI":"10.1016\/j.molmed.2024.11.009","article-title":"From pixels to patients: the evolution and future of deep learning in cancer diagnostics","volume":"31","author":"Yang","year":"2025","journal-title":"Trends Mol Med"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1917573\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,22]],"date-time":"2026-09-22T13:15:24Z","timestamp":1790082924000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1917573\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,22]]},"references-count":18,"alternative-id":["10.3389\/fdgth.2026.1917573"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2026.1917573","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,22]]},"article-number":"1917573"}}