{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,24]],"date-time":"2026-10-24T10:00:18Z","timestamp":1792836018338,"version":"4.5.0"},"reference-count":17,"publisher":"Springer Science and Business Media LLC","issue":"4","license":[{"start":{"date-parts":[[2026,3,24]],"date-time":"2026-03-24T00:00:00Z","timestamp":1774310400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,3,24]],"date-time":"2026-03-24T00:00:00Z","timestamp":1774310400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100016981","name":"Necmettin Erbakan University","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100016981","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Eur J Pediatr"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    We evaluated three multimodal LLMs, ChatGPT (GPT-5.2), Gemini 3, and Microsoft Copilot, in pediatric ECG interpretation, focusing on clinically significant abnormalities and emergency arrhythmias with likelihood ratios as primary outcome measures. This prospective comparative diagnostic accuracy study (STARD\/STARD-AI) included 264 pediatric patients with 12-lead ECGs (November 2024\u2013November 2025). De-identified images were submitted via standardized zero-shot prompt. Three blinded pediatric cardiologists established the reference diagnosis by majority-vote consensus. Cases were classified as Tier 1 (normal), Tier 2 (abnormal, non-urgent), or Tier 3 (urgent). Two binary endpoints were assessed: clinically significant abnormality (Tier 2\u2009+\u20093 vs Tier 1) and emergency abnormality (Tier 3 vs Tier 1\u2009+\u20092). Clinically significant abnormalities were present in 54.5% of patients. AUC values ranged from 0.550 to 0.623, reflecting modest discrimination. For the clinically significant endpoint,\u2009+\u2009LR values were 2.05 (ChatGPT), 1.26 (Gemini), and 1.21 (Copilot);\u2009\u2212\u2009LR values were 0.68, 0.55, and 0.81, indicating limited rule-in and insufficient rule-out utility. For the emergency endpoint, Gemini achieved 100% sensitivity (95% CI\u2009=\u200985.1\u2013100.0) with\u2009\u2212\u2009LR 0.07 (95% CI\u2009=\u20090.00\u20131.12) in a small subgroup (\n                    <jats:italic>n<\/jats:italic>\n                    \u2009=\u200922); however, specificity of 30.2% and\u2009+\u2009LR of 1.40 indicate overcalling rather than diagnostic precision. No model achieved clinically meaningful rule-in utility for either endpoint.\n                  <\/jats:p>\n                  <jats:p>\n                    <jats:italic>Conclusions<\/jats:italic>\n                    : Current multimodal LLMs showed limited diagnostic utility in pediatric ECG interpretation, with\u2009+\u2009LR values near 1.0 across both endpoints. Standalone deployment is not supported; these tools may at most serve as adjunctive screening aids under clinician oversight.\n                    <jats:table-wrap>\n                      <jats:table>\n                        <jats:tbody>\n                          <jats:tr>\n                            <jats:td align=\"left\" colspan=\"2\">\n                              <jats:bold>What is Known:<\/jats:bold>\n                              \u2022\n                              <jats:italic>Deep learning algorithms trained on large ECG datasets perform well in adult populations, but evidence in pediatric ECG interpretation is limited.<\/jats:italic>\n                              \u2022\u00a0\n                              <jats:italic>General-purpose LLMs show variable accuracy in medical examinations; reliability in subspecialty domains such as pediatric cardiology remains unproven.<\/jats:italic>\n                            <\/jats:td>\n                          <\/jats:tr>\n                          <jats:tr>\n                            <jats:td align=\"left\" colspan=\"2\">\n                              <jats:bold>What is New:<\/jats:bold>\n                              \u2022\u00a0\n                              <jats:italic>This is the[FCA1] first head-to-head comparative diagnostic accuracy study of multimodal LLMs in pediatric ECG evaluation, using likelihood ratios as primary outcome measures.<\/jats:italic>\n                              \u2022\n                              <jats:italic>All three LLMs showed limited rule-in utility (+LR near 1.0); Gemini achieved potentially meaningful rule-out performance for emergency arrhythmias (\u2212LR = 0.07), but with wide confidence intervals reflecting the small emergency subgroup (n = 22).<\/jats:italic>\n                              \u2022\u00a0\n                              <jats:italic>Gemini\u2019s 100% sensitivity in the emergency subgroup reflects overcalling (specificity 30.2%) consistent with a triage\/screening behavior rather than diagnostic precision.<\/jats:italic>\n                            <\/jats:td>\n                          <\/jats:tr>\n                        <\/jats:tbody>\n                      <\/jats:table>\n                    <\/jats:table-wrap>\n                  <\/jats:p>","DOI":"10.1007\/s00431-026-06874-x","type":"journal-article","created":{"date-parts":[[2026,3,24]],"date-time":"2026-03-24T06:39:00Z","timestamp":1774334340000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Comparative evaluation of multimodal large language models for diagnostic accuracy in pediatric electrocardiography: a prospective comparative diagnostic accuracy study"],"prefix":"10.1007","volume":"185","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-0652-5834","authenticated-orcid":false,"given":"U\u011fur","family":"Sara\u00e7","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8002-1087","authenticated-orcid":false,"given":"Ay\u015fe B\u00fc\u015fra","family":"Payda\u015f","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9455-2735","authenticated-orcid":false,"given":"Mustafa","family":"Gen\u00e7eli","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-2385-0953","authenticated-orcid":false,"given":"Talha","family":"\u00dcst\u00fcnta\u015f","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6091-3205","authenticated-orcid":false,"given":"Mehtap","family":"Y\u00fccel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8454-5555","authenticated-orcid":false,"given":"Abd\u00fclkerim","family":"\u00c7okbi\u00e7er","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7870-9704","authenticated-orcid":false,"given":"Fatih","family":"\u015eap","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6145-1831","authenticated-orcid":false,"given":"Tamer","family":"Baysal","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1515-4654","authenticated-orcid":false,"given":"Mehmet Burhan","family":"Oflaz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,3,24]]},"reference":[{"key":"6874_CR1","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1016\/j.tcm.2020.11.007","volume":"32","author":"D Itchhaporia","year":"2022","unstructured":"Itchhaporia D (2022) Artificial intelligence in cardiology. Trends Cardiovasc Med 32:34\u201341","journal-title":"Trends Cardiovasc Med"},{"key":"6874_CR2","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1016\/j.metabol.2017.01.011","volume":"69","author":"P Hamet","year":"2017","unstructured":"Hamet P, Tremblay J (2017) Artificial intelligence in medicine. Metabolism 69:36\u201340","journal-title":"Metabolism"},{"key":"6874_CR3","first-page":"279","volume":"4","author":"I Skalidis","year":"2023","unstructured":"Skalidis I, Cagnina A, Luangphiphat W et al (2023) ChatGPT takes on the European exam in core cardiology: an artificial intelligence success story? Eur Heart J 4:279\u2013281","journal-title":"Eur Heart J"},{"key":"6874_CR4","doi-asserted-by":"publisher","first-page":"51","DOI":"10.1016\/j.ajem.2024.03.017","volume":"80","author":"S G\u00fcnay","year":"2024","unstructured":"G\u00fcnay S, \u00d6zt\u00fcrk A, \u00d6zerol H et al (2024) Comparison of emergency medicine specialist, cardiologist, and ChatGPT in electrocardiography assessment. Am J Emerg Med 80:51\u201360","journal-title":"Am J Emerg Med"},{"key":"6874_CR5","doi-asserted-by":"publisher","first-page":"815","DOI":"10.1016\/j.jacc.2024.05.062","volume":"84","author":"J Mayourian","year":"2024","unstructured":"Mayourian J, Gearhart A, La Cava WG et al (2024) Deep learning-based electrocardiogram analysis predicts biventricular dysfunction and dilation in congenital heart disease. J Am Coll Cardiol 84:815\u2013828","journal-title":"J Am Coll Cardiol"},{"key":"6874_CR6","doi-asserted-by":"crossref","unstructured":"Mayourian J, Geggel R, La Cava WG, Ghelani SJ, Triedman JK (2024) Pediatric electrocardiogram-based deep learning to predict secundum atrial septal defects. Pediatr Cardiol. Online ahead of print","DOI":"10.1007\/s00246-024-03540-7"},{"key":"6874_CR7","doi-asserted-by":"publisher","first-page":"917","DOI":"10.1161\/CIRCULATIONAHA.123.067750","volume":"149","author":"J Mayourian","year":"2024","unstructured":"Mayourian J, La Cava WG, Vaid A et al (2024) Pediatric ECG-based deep learning to predict left ventricular dysfunction and remodeling. Circulation 149:917\u2013931","journal-title":"Circulation"},{"key":"6874_CR8","doi-asserted-by":"publisher","DOI":"10.1016\/j.eclinm.2023.102141","volume":"63","author":"K Miura","year":"2023","unstructured":"Miura K, Yagi R, Miyama H et al (2023) Deep learning-based model detects atrial septal defects from electrocardiography: a cross-sectional multicenter hospital-based study. EClinicalMedicine 63:102141","journal-title":"EClinicalMedicine"},{"key":"6874_CR9","doi-asserted-by":"publisher","first-page":"1206","DOI":"10.1016\/S0140-6736(22)01637-3","volume":"400","author":"PA Noseworthy","year":"2022","unstructured":"Noseworthy PA, Attia ZI, Behnken EM et al (2022) Artificial intelligence-guided screening for atrial fibrillation using electrocardiogram during sinus rhythm: a prospective non-randomised interventional trial. Lancet 400:1206\u20131212","journal-title":"Lancet"},{"key":"6874_CR10","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCEP.119.007284","volume":"12","author":"ZI Attia","year":"2019","unstructured":"Attia ZI, Friedman PA, Noseworthy PA et al (2019) Age and sex estimation using artificial intelligence from standard 12-lead ECGs. Circ Arrhythm Electrophysiol 12:e007284","journal-title":"Circ Arrhythm Electrophysiol"},{"key":"6874_CR11","doi-asserted-by":"publisher","first-page":"1626","DOI":"10.1136\/hrt.2004.057307","volume":"91","author":"DF Dickinson","year":"2005","unstructured":"Dickinson DF (2005) The normal ECG in childhood and adolescence. Heart 91:1626\u20131630","journal-title":"Heart"},{"key":"6874_CR12","doi-asserted-by":"publisher","first-page":"25","DOI":"10.3390\/children12010025","volume":"12","author":"DM Leone","year":"2024","unstructured":"Leone DM, O\u2019Sullivan D, Bravo-Jaimes K (2024) Artificial intelligence in pediatric electrocardiography: a comprehensive review. Children Basel 12:25","journal-title":"Children Basel"},{"key":"6874_CR13","doi-asserted-by":"publisher","DOI":"10.1177\/20552076251325279","volume":"11","author":"V \u00c7amk\u0131ran","year":"2025","unstructured":"\u00c7amk\u0131ran V, Tun\u00e7 H, Achmar B et al (2025) Artificial intelligence (ChatGPT) ready to evaluate ECG in real life? Not yet! Digit Health 11:20552076251325279","journal-title":"Digit Health"},{"key":"6874_CR14","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1038\/s41591-018-0268-3","volume":"25","author":"AY Hannun","year":"2019","unstructured":"Hannun AY, Rajpurkar P, Haghpanahi M et al (2019) Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network. Nat Med 25:65\u201369","journal-title":"Nat Med"},{"key":"6874_CR15","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-020-15432-4","volume":"11","author":"AH Ribeiro","year":"2020","unstructured":"Ribeiro AH, Ribeiro MH, Paix\u00e3o GMM et al (2020) Automatic diagnosis of the 12-lead ECG using a deep neural network. Nat Commun 11:1760","journal-title":"Nat Commun"},{"key":"6874_CR16","doi-asserted-by":"publisher","first-page":"2349","DOI":"10.1017\/S1047951124036035","volume":"34","author":"MN Gritti","year":"2024","unstructured":"Gritti MN, Prajapati R, Yissar D, Morgan CT (2024) Precision of artificial intelligence in paediatric cardiology multimodal image interpretation. Cardiol Young 34:2349\u20132354","journal-title":"Cardiol Young"},{"key":"6874_CR17","doi-asserted-by":"publisher","first-page":"330","DOI":"10.1111\/j.1542-474X.2005.00642.x","volume":"10","author":"RM Hamilton","year":"2005","unstructured":"Hamilton RM, McLeod K, Houston AB, Macfarlane PW (2005) Inter- and intraobserver variability in LVH and RVH reporting in pediatric ECGs. Ann Noninvasive Electrocardiol 10:330\u2013333. https:\/\/doi.org\/10.1111\/j.1542-474X.2005.00642.x","journal-title":"Ann Noninvasive Electrocardiol"}],"container-title":["European Journal of Pediatrics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00431-026-06874-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s00431-026-06874-x","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00431-026-06874-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,23]],"date-time":"2026-04-23T07:53:40Z","timestamp":1776930820000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s00431-026-06874-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,24]]},"references-count":17,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2026,4]]}},"alternative-id":["6874"],"URL":"https:\/\/doi.org\/10.1007\/s00431-026-06874-x","relation":{},"ISSN":["1432-1076"],"issn-type":[{"value":"1432-1076","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,24]]},"assertion":[{"value":"13 January 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"10 March 2026","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"17 March 2026","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 March 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The study protocol was approved by the Necmettin Erbakan University Faculty of Medicine Ethics Committee (2025\/6232). All procedures were performed in accordance with the principles of the Declaration of Helsinki.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval"}},{"value":"Informed consent was obtained from parents or legal guardians of all participants.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent to participate"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"206"}}