{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T19:15:35Z","timestamp":1792782935980,"version":"4.5.0"},"reference-count":27,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T00:00:00Z","timestamp":1789516800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T00:00:00Z","timestamp":1789516800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Monash University","id":[{"id":"https:\/\/ror.org\/02bfwt286","id-type":"ROR","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Neurosurg Rev"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Patients increasingly consult large language models (LLMs) before specialist review, yet whether frontier models agree with one another in nuanced clinical domains such as unruptured intracranial aneurysm (UIA) management remains uncharacterised. We therefore quantified inter-model variability across ChatGPT, Gemini and Claude, anchored against neurovascular multidisciplinary team (MDT) consensus and the Unruptured Intracranial Aneurysm Treatment Score (UIATS).\u00a0Sixty-seven UIA cases referred to our neurovascular service (January\u2013December 2025) were retrospectively analysed. De-identified clinical vignettes were submitted to Claude Opus 4.6, ChatGPT-5.4 and Gemini 3 Pro Thinking, each run five times. Within-model reproducibility was assessed using Fleiss\u2019 \u03ba; inter-model agreement via Cohen\u2019s \u03ba and McNemar\u2019s test; and each LLM\u2019s majority-vote anchored against MDT and UIATS using the same methods.\u00a0Within-model reproducibility was almost perfect (Fleiss\u2019 \u03ba 0.837\u20130.860). Pairwise inter-model agreement was asymmetric: ChatGPT\u2013Gemini behaved near-identically (Cohen\u2019s \u03ba\u2009=\u20090.850, 95% CI 0.71\u20130.97), whereas Claude diverged from both (\u03ba\u2009=\u20090.688 and 0.667). Recommendations were non-unanimous in 13\/67 cases (19.4%); Claude was the sole outlier in 8\/13 (conservative in 7). Gemini was significantly more pro-treatment than Claude (McNemar\n                    <jats:italic>p<\/jats:italic>\n                    \u2009=\u20090.0117). Against MDT, Gemini and ChatGPT showed significant pro-treatment propensity (\n                    <jats:italic>p<\/jats:italic>\n                    \u2009=\u20090.0022,\n                    <jats:italic>p<\/jats:italic>\n                    \u2009=\u20090.0153). Claude was significantly more conservative than UIATS (\n                    <jats:italic>p<\/jats:italic>\n                    \u2009=\u20090.0162).\u00a0Frontier LLMs are highly reproducible internally but diverge from one another asymmetrically. ChatGPT and Gemini behave near-identically while Claude diverges conservatively. Clinicians should anticipate AI-driven treatment expectations and future work should explore prompting, patient sentiment, and multicentre moderators.\n                  <\/jats:p>","DOI":"10.1007\/s10143-026-04498-1","type":"journal-article","created":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T06:39:15Z","timestamp":1789540755000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Inter-Model Variability of ChatGPT, Gemini and Claude in Treatment Recommendations for Unruptured Intracranial Aneurysms"],"prefix":"10.1007","volume":"49","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-5326-6949","authenticated-orcid":false,"given":"Anish","family":"Narayan","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Frederick","family":"Mariajoseph","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Malik","family":"Farooq","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Adrian","family":"Praeger","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ronil","family":"Chandra","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Idrees","family":"Sher","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lee-Anne","family":"Slater","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Calvin","family":"Gan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrew","family":"Gauden","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hamed","family":"Asadi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Justin","family":"Moore","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,9,16]]},"reference":[{"key":"4498_CR1","doi-asserted-by":"publisher","unstructured":"Abed A, Abu Assab M, Merdas ZJH et al (2026) Confronting antimicrobial resistance in Jordan: regulatory, economic, and behavioral determinants of non-prescription antibiotic dispensing in community pharmacies\u2014a mixed-methods study. Front Med (Lausanne) 12. https:\/\/doi.org\/10.3389\/fmed.2025.1742205","DOI":"10.3389\/fmed.2025.1742205"},{"issue":"3","key":"4498_CR2","doi-asserted-by":"publisher","first-page":"282","DOI":"10.1001\/jamaneurol.2018.4165","volume":"76","author":"AM Algra","year":"2019","unstructured":"Algra AM, Lindgren A, Vergouwen MDI et al (2019) Procedural clinical complications, case-fatality risks, and risk factors in endovascular and neurosurgical treatment of unruptured intracranial aneurysms. JAMA Neurol 76(3):282. https:\/\/doi.org\/10.1001\/jamaneurol.2018.4165","journal-title":"JAMA Neurol"},{"key":"4498_CR3","unstructured":"Anthropic (2023) Claude\u2019s Constitution. Accessed 9 May 2023. https:\/\/www.anthropic.com\/news\/claudes-constitution"},{"issue":"17","key":"4498_CR4","doi-asserted-by":"publisher","first-page":"1600","DOI":"10.1212\/wnl.0000000000003865","volume":"88","author":"D Backes","year":"2017","unstructured":"Backes D, Rinkel GJE, Greving JP et al (2017) ELAPSS score for prediction of risk of growth of unruptured intracranial aneurysms. Neurology 88(17):1600\u20131606. https:\/\/doi.org\/10.1212\/wnl.0000000000003865","journal-title":"Neurology"},{"key":"4498_CR5","unstructured":"Bai Y, Jones A, Ndousse K et al (2022) Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback. arXiv:2204.05862. https:\/\/arxiv.org\/abs\/2204.05862"},{"key":"4498_CR6","doi-asserted-by":"publisher","unstructured":"Basu S (2026) The Value Sensitivity Gap: How Clinical Large Language Models Respond to Patient Preference Statements in Shared Decision-Making. arXiv:2603.00076. https:\/\/doi.org\/10.48550\/arXiv.2603.00076","DOI":"10.48550\/arXiv.2603.00076"},{"key":"4498_CR7","doi-asserted-by":"publisher","first-page":"e68146","DOI":"10.2196\/68146","volume":"17","author":"H Campos","year":"2025","unstructured":"Campos H, Wolfe D, Luan H, Sim I (2025) Generative AI as third agent: large language models and the transformation of the clinician-patient relationship. J Particip Med 17:e68146. https:\/\/doi.org\/10.2196\/68146","journal-title":"J Particip Med"},{"key":"4498_CR8","doi-asserted-by":"publisher","DOI":"10.3389\/frhs.2025.1694139","volume":"5","author":"Q Chen","year":"2025","unstructured":"Chen Q (2025) Artificial intelligence in healthcare: rethinking doctor-patient relationship in megacities. Front Health Serv 5:1694139. https:\/\/doi.org\/10.3389\/frhs.2025.1694139","journal-title":"Front Health Serv"},{"key":"4498_CR9","doi-asserted-by":"publisher","DOI":"10.1126\/science.aec8352","author":"M Cheng","year":"2026","unstructured":"Cheng M, Lee C, Khadpe P, Yu S, Han D, Jurafsky D (2026) Sycophantic AI decreases prosocial intentions and promotes dependence. Science. https:\/\/doi.org\/10.1126\/science.aec8352","journal-title":"Science"},{"issue":"10","key":"4498_CR10","doi-asserted-by":"publisher","first-page":"881","DOI":"10.1212\/wnl.0000000000001891","volume":"85","author":"N Etminan","year":"2015","unstructured":"Etminan N, Brown RD, Beseoglu K et al (2015) The unruptured intracranial aneurysm treatment score. Neurology 85(10):881\u2013889. https:\/\/doi.org\/10.1212\/wnl.0000000000001891","journal-title":"Neurology"},{"issue":"3","key":"4498_CR11","doi-asserted-by":"publisher","first-page":"1625","DOI":"10.1007\/s10143-020-01356-6","volume":"44","author":"J Feghali","year":"2021","unstructured":"Feghali J, Gami A, Caplan JM, Tamargo RJ, McDougall CG, Huang J (2021) Management of unruptured intracranial aneurysms: correlation of UIATS, ELAPSS, and PHASES with referral center practice. Neurosurg Rev 44(3):1625\u20131633. https:\/\/doi.org\/10.1007\/s10143-020-01356-6","journal-title":"Neurosurg Rev"},{"issue":"1","key":"4498_CR12","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1016\/S1474-4422(13)70263-1","volume":"13","author":"JP Greving","year":"2014","unstructured":"Greving JP, Wermer MJH, Brown RD et al (2014) Development of the PHASES score for prediction of risk of rupture of intracranial aneurysms: a pooled analysis of six prospective cohort studies. Lancet Neurol 13(1):59\u201366. https:\/\/doi.org\/10.1016\/S1474-4422(13)70263-1","journal-title":"Lancet Neurol"},{"key":"4498_CR13","doi-asserted-by":"publisher","DOI":"10.1007\/s10143-025-03341-3","author":"A Hadjiathanasiou","year":"2025","unstructured":"Hadjiathanasiou A, Goelz L, Muhn F et al (2025) Artificial intelligence in neurovascular decision-making: a comparative analysis of ChatGPT-4 and multidisciplinary expert recommendations for unruptured intracranial aneurysms. Neurosurg Rev. https:\/\/doi.org\/10.1007\/s10143-025-03341-3","journal-title":"Neurosurg Rev"},{"issue":"24","key":"4498_CR14","doi-asserted-by":"publisher","first-page":"1725","DOI":"10.1056\/nejm199812103392401","volume":"339","author":"International Study of Unruptured Intracranial Aneurysms Investigators","year":"1998","unstructured":"International Study of Unruptured Intracranial Aneurysms Investigators (1998) Unruptured intracranial aneurysms \u2014 risk of rupture and risks of surgical intervention. N Engl J Med 339(24):1725\u20131733. https:\/\/doi.org\/10.1056\/nejm199812103392401","journal-title":"N Engl J Med"},{"issue":"8","key":"4498_CR15","doi-asserted-by":"publisher","first-page":"1600","DOI":"10.3174\/ajnr.A3470","volume":"34","author":"T Ishibashi","year":"2013","unstructured":"Ishibashi T, Murayama Y, Saguchi T, Ebara M, Arakawa H, Irie K et al (2013) Justification of unruptured intracranial aneurysm repair: a single-center experience. AJNR Am J Neuroradiol 34(8):1600\u20131605. https:\/\/doi.org\/10.3174\/ajnr.A3470","journal-title":"AJNR Am J Neuroradiol"},{"key":"4498_CR16","doi-asserted-by":"crossref","unstructured":"Kirk HR, Whitefield A, R\u00f6ttger P et al (2024) The PRISM alignment dataset: what participatory, representative and individualised human feedback reveals about the subjective and multicultural alignment of large language models. In: Advances in Neural Information Processing Systems, vol 37. Curran Associates, pp 105236\u2013105344","DOI":"10.52202\/079017-3342"},{"issue":"2","key":"4498_CR17","doi-asserted-by":"publisher","first-page":"177","DOI":"10.1136\/neurintsurg-2020-016238","volume":"13","author":"VM Lu","year":"2021","unstructured":"Lu VM, Chen SH, Young CC, Starke RM (2021) Nature, content and shifts over time of the most impactful unruptured intracranial aneurysms articles: a bibliometric analysis. J Neurointerv Surg 13(2):177\u2013181. https:\/\/doi.org\/10.1136\/neurintsurg-2020-016238","journal-title":"J Neurointerv Surg"},{"key":"4498_CR18","unstructured":"Maes P (2025) People overtrust AI-generated medical advice despite low accuracy. MIT Media Lab. https:\/\/www.media.mit.edu\/publications\/NEJM-AI-people-overtrust-ai-generated-medical-advice-despite-low-accuracy\/. Accessed 2025"},{"key":"4498_CR19","unstructured":"OpenAI (2026) Introducing ChatGPT Health. Accessed 14 April 2026. https:\/\/openai.com\/index\/introducing-chatgpt-health\/"},{"issue":"10","key":"4498_CR20","doi-asserted-by":"publisher","DOI":"10.3390\/brainsci15101061","volume":"15","author":"M Overstijns","year":"2025","unstructured":"Overstijns M, Nazeeruddin S, Scheffler P, Roelz R, Beck J, El Rahal A (2025) Comparative analysis of AI models in predicting treatment strategies for unruptured intracranial aneurysms. Brain Sci 15(10):1061. https:\/\/doi.org\/10.3390\/brainsci15101061","journal-title":"Brain Sci"},{"issue":"7","key":"4498_CR21","doi-asserted-by":"publisher","first-page":"764","DOI":"10.3390\/brainsci15070764","volume":"15","author":"P Penchev","year":"2025","unstructured":"Penchev P, Ivanov K, Milanova-Ilieva D et al (2025) Mental health and quality of life in patients with untreated unruptured intracranial aneurysms: a systematic review and meta-analysis of 417,152 patients with trial sequential analysis. Brain Sci 15(7):764. https:\/\/doi.org\/10.3390\/brainsci15070764","journal-title":"Brain Sci"},{"key":"4498_CR22","doi-asserted-by":"publisher","DOI":"10.1002\/14651858.cd013312","author":"FG Pontes","year":"2019","unstructured":"Pontes FG, de Vasconcelos B, Baptista-Silva V, da Silva JC EM (2019) Treatments for unruptured intracranial aneurysms. Cochrane Database Syst Rev. https:\/\/doi.org\/10.1002\/14651858.cd013312","journal-title":"Cochrane Database Syst Rev"},{"key":"4498_CR23","doi-asserted-by":"publisher","DOI":"10.2196\/69790","author":"MB Schoening","year":"2024","unstructured":"Schoening MB, Cotliar D (2024) Patients and caregivers leveraging AI to improve their healthcare journey: a case study and lessons learned (Preprint). J Particip Med. https:\/\/doi.org\/10.2196\/69790","journal-title":"J Particip Med"},{"key":"4498_CR24","doi-asserted-by":"publisher","DOI":"10.2196\/60391","volume":"4","author":"AH Shmilovitch","year":"2025","unstructured":"Shmilovitch AH, Katson M, Cohen-Shelly M, Peretz S, Aran D, Shelly S (2025) GPT-4 as a clinical decision support tool in ischemic stroke management: evaluation study. JMIR AI 4:e60391. https:\/\/doi.org\/10.2196\/60391","journal-title":"JMIR AI"},{"issue":"6","key":"4498_CR25","doi-asserted-by":"publisher","first-page":"E18","DOI":"10.3171\/2025.9.focus25723","volume":"59","author":"VE Staartjes","year":"2025","unstructured":"Staartjes VE, Adamides AA, Anania P et al (2025) Contemporary adverse event profile of microsurgery for intracranial unruptured aneurysms in high-volume microsurgical centers: the international PRAEMIUM study. Neurosurg Focus 59(6):E18. https:\/\/doi.org\/10.3171\/2025.9.focus25723","journal-title":"Neurosurg Focus"},{"issue":"26","key":"4498_CR26","doi-asserted-by":"publisher","first-page":"2474","DOI":"10.1056\/nejmoa1113260","volume":"366","author":"UCAS Japan Investigators","year":"2012","unstructured":"UCAS Japan Investigators (2012) The natural course of unruptured cerebral aneurysms in a Japanese cohort. N Engl J Med 366(26):2474\u20132482. https:\/\/doi.org\/10.1056\/nejmoa1113260","journal-title":"N Engl J Med"},{"key":"4498_CR27","doi-asserted-by":"publisher","first-page":"104","DOI":"10.1016\/j.jocn.2022.01.038","volume":"98","author":"T Wende","year":"2022","unstructured":"Wende T, Kasper J, Wilhelmy F, Prasse G, Qu\u00e4schling U, Haase A et al (2022) Comparison of the unruptured intracranial aneurysm treatment score recommendations with clinical treatment results \u2014 a series of 322 aneurysms. J Clin Neurosci 98:104\u2013108. https:\/\/doi.org\/10.1016\/j.jocn.2022.01.038","journal-title":"J Clin Neurosci"}],"container-title":["Neurosurgical Review"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10143-026-04498-1.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10143-026-04498-1","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10143-026-04498-1.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T06:39:22Z","timestamp":1789540762000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10143-026-04498-1"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,16]]},"references-count":27,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["4498"],"URL":"https:\/\/doi.org\/10.1007\/s10143-026-04498-1","relation":{},"ISSN":["1437-2320"],"issn-type":[{"value":"1437-2320","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,16]]},"assertion":[{"value":"2 June 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"5 June 2026","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"12 September 2026","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"16 September 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"This study approved under the institution\u2019s ethics review board. Approval was granted by the institutional review board (reference: RES-25-0000-885Q).","order":1,"name":"Ethics","label":"Ethics approval","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Individual patient consent was waived by the institutional review board owing to the retrospective, fully de-identified nature of the data. No identifiable patient information was submitted to any LLM at any point.","order":2,"name":"Ethics","label":"Consent to participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable \u2014 this manuscript contains no identifiable individual patient data.","order":3,"name":"Ethics","label":"Consent to publish","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","label":"Competing interests","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":5,"name":"Ethics","label":"Clinical trial number","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":6,"name":"Ethics","label":"Human ethics and consent to participate declarations","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"This study was prospectively registered on the Open Science Framework (https:\/\/doi.org\/10.17605\/OSF.IO\/358MQ).","order":7,"name":"Ethics","label":"Registration","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"569"}}