{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,22]],"date-time":"2026-10-22T17:37:29Z","timestamp":1792690649107,"version":"4.5.0"},"reference-count":23,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,5,28]],"date-time":"2026-05-28T00:00:00Z","timestamp":1779926400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,6,12]],"date-time":"2026-06-12T00:00:00Z","timestamp":1781222400000},"content-version":"vor","delay-in-days":15,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Universit\u00e4tsklinikum D\u00fcsseldorf. Anstalt \u00f6ffentlichen Rechts"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Oral Health"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Multimodal large language models (LLMs) are increasingly being evaluated for clinical image interpretation, but whether patient demographic cues influence their recommendations remains unclear in dental radiology\u2014a leading domain for AI-assisted diagnostics, and one marked by pronounced disparities in oral health.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We conducted a within-image, between-condition audit study using 705 panoramic dental radiographs from the publicly available DENTEX dataset. Each image was submitted to three commercial multimodal LLMs (Gemini 2.5 Flash, GPT-5.4, Claude Sonnet 4.6) under 12 experimental conditions: five race\/ethnicity groups \u00d7 two sex categories plus two controls, yielding 25,380 independent zero-temperature API calls. Each call returned a structured JSON object containing ten prespecified clinical recommendation variables. Delta variables were computed against the primary control, and linear mixed-effects models with image identity as a random intercept tested for race\/ethnicity (H1), sex (H2), and Race \u00d7 Sex interaction (H3) effects. The five primary ordinal outcomes were evaluated against a Bonferroni-corrected significance threshold of \u03b1\u2009=\u20090.01.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      All three models altered recommendations when demographic cues were introduced. Claude showed the broadest sensitivity, with significant race effects on treatment invasiveness, prognosis, and confidence (all\n                      <jats:italic>p<\/jats:italic>\n                      &lt; .001), and significant sex effects on prognosis, confidence, and invasiveness. GPT-5.4 effects were concentrated in confidence and in invasiveness for Hispanic patients (\u03b2 = +0.052,\n                      <jats:italic>p<\/jats:italic>\n                      &lt; .001). Gemini produced higher invasiveness for Black patients (\u03b2 = +0.026,\n                      <jats:italic>p<\/jats:italic>\n                      &lt; .001). No Race \u00d7 Sex interactions reached significance. All ordinal shifts were below 6% of one category width. Inter-model diagnostic agreement under control prompts was near chance (Cohen\u2019s \u03ba\u2009=\u20090.009\u20130.159). Findings were robust across four prespecified sensitivity analyses.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Demographic prompt cues produced small but systematic shifts in multimodal LLM clinical recommendations for dental radiographs, with model-specific sensitivity profiles. Because clinical accuracy was not benchmarked, these findings should be read as demographic-cue sensitivity rather than clinical bias. Inter-model agreement was near chance, indicating that model choice currently introduces more variability than any demographic label. Demographic-cue audits and model-specific reliability evaluations should both accompany clinical deployment, particularly in oral health, where racial and socioeconomic disparities are already documented.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1186\/s12903-026-08644-5","type":"journal-article","created":{"date-parts":[[2026,5,28]],"date-time":"2026-05-28T06:55:21Z","timestamp":1779951321000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Demographic prompt cues shift clinical recommendations in multimodal large language models: a multi-model audit of 25,380 dental radiograph assessments"],"prefix":"10.1186","volume":"26","author":[{"given":"Babak","family":"Saravi","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ingrid","family":"Winkler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Daman Deep","family":"Singh","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Julian","family":"Lommen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Max","family":"Wilkat","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lara","family":"Schorn","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andreas","family":"Vollmer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Michael","family":"Vollmer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Norbert","family":"K\u00fcbler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christoph","family":"Sproll","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Felix","family":"Schrader","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,5,28]]},"reference":[{"key":"8644_CR1","doi-asserted-by":"publisher","first-page":"259","DOI":"10.1038\/s41586-023-05881-4","volume":"616","author":"M Moor","year":"2023","unstructured":"Moor M, Banerjee O, Abad ZSH, Krumholz HM, Leskovec J, Topol EJ, et al. Foundation models for generalist medical artificial intelligence. Nature. 2023;616:259\u201365. https:\/\/doi.org\/10.1038\/s41586-023-05881-4.","journal-title":"Nature"},{"key":"8644_CR2","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1038\/s41586-023-06291-2","volume":"620","author":"K Singhal","year":"2023","unstructured":"Singhal K, Azizi S, Tu T, Mahdavi SS, Wei J, Chung HW, et al. Large language models encode clinical knowledge. Nature. 2023;620:172\u201380. https:\/\/doi.org\/10.1038\/s41586-023-06291-2.","journal-title":"Nature"},{"key":"8644_CR3","doi-asserted-by":"publisher","first-page":"6859","DOI":"10.3390\/jcm13226859","volume":"13","author":"N Turosz","year":"2024","unstructured":"Turosz N, Ch\u0119ci\u0144ska K, Ch\u0119ci\u0144ski M, Sielski M, Sikora M. Evaluation of Dental Panoramic Radiographs by Artificial Intelligence Compared to Human Reference: A Diagnostic Accuracy Study. J Clin Med. 2024;13:6859. https:\/\/doi.org\/10.3390\/jcm13226859.","journal-title":"J Clin Med"},{"key":"8644_CR4","doi-asserted-by":"publisher","first-page":"20210296","DOI":"10.1259\/dmfr.20210296","volume":"51","author":"M Estai","year":"2022","unstructured":"Estai M, Tennant M, Gebauer D, Brostek A, Vignarajan J, Mehdizadeh M, et al. Deep learning for automated detection and numbering of permanent teeth on panoramic images. Dentomaxillofac Radiol. 2022;51:20210296. https:\/\/doi.org\/10.1259\/dmfr.20210296.","journal-title":"Dentomaxillofac Radiol"},{"key":"8644_CR5","doi-asserted-by":"publisher","first-page":"1403","DOI":"10.1186\/s12909-025-07875-4","volume":"25","author":"BN Gokkurt Yilmaz","year":"2025","unstructured":"Gokkurt Yilmaz BN, Ozbey F, Yilmaz BE. Effect of artificial intelligence-assisted personalized feedback on radiographic diagnostic performance of dental students: a controlled study. BMC Med Educ. 2025;25:1403. https:\/\/doi.org\/10.1186\/s12909-025-07875-4.","journal-title":"BMC Med Educ"},{"key":"8644_CR6","doi-asserted-by":"publisher","unstructured":"Hao J, Fan Y, Sun Y, Guo K, Lin L, Yang J et al. Towards Better Dental AI: A Multimodal Benchmark and Instruction Dataset for Panoramic X-ray Analysis. arXiv; 2025. https:\/\/doi.org\/10.48550\/ARXIV.2509.09254.\u00a0 Cited 2026 Mar 8.","DOI":"10.48550\/ARXIV.2509.09254"},{"key":"8644_CR7","doi-asserted-by":"publisher","first-page":"447","DOI":"10.1126\/science.aax2342","volume-title":"Dissecting racial bias in an algorithm used to manage the health of populations","author":"Z Obermeyer","year":"2019","unstructured":"Obermeyer Z, Powers B, Vogeli C, Mullainathan S. Dissecting racial bias in an algorithm used to manage the health of populations. Volume 366. New York, N.Y.: Science; 2019. pp. 447\u201353. https:\/\/doi.org\/10.1126\/science.aax2342."},{"key":"8644_CR8","doi-asserted-by":"publisher","first-page":"e12","DOI":"10.1016\/S2589-7500(23)00225-X","volume":"6","author":"T Zack","year":"2024","unstructured":"Zack T, Lehman E, Suzgun M, Rodriguez JA, Celi LA, Gichoya J, et al. Assessing the potential of GPT-4 to perpetuate racial and gender biases in health care: a model evaluation study. Lancet Digit Health. 2024;6:e12\u201322. https:\/\/doi.org\/10.1016\/S2589-7500(23)00225-X.","journal-title":"Lancet Digit Health"},{"key":"8644_CR9","doi-asserted-by":"publisher","first-page":"1873","DOI":"10.1038\/s41591-025-03626-6","volume":"31","author":"M Omar","year":"2025","unstructured":"Omar M, Soffer S, Agbareia R, Bragazzi NL, Apakama DU, Horowitz CR, et al. Sociodemographic biases in medical decision making by large language models. Nat Med. 2025;31:1873\u201381. https:\/\/doi.org\/10.1038\/s41591-025-03626-6.","journal-title":"Nat Med"},{"key":"8644_CR10","doi-asserted-by":"publisher","first-page":"57","DOI":"10.1186\/s12939-025-02419-0","volume":"24","author":"M Omar","year":"2025","unstructured":"Omar M, Sorin V, Agbareia R, Apakama DU, Soroush A, Sakhuja A, et al. Evaluating and addressing demographic disparities in medical large language models: a systematic review. Int J Equity Health. 2025;24:57. https:\/\/doi.org\/10.1186\/s12939-025-02419-0.","journal-title":"Int J Equity Health"},{"key":"8644_CR11","doi-asserted-by":"publisher","first-page":"332","DOI":"10.1038\/s41746-025-01746-4","volume":"8","author":"A Bouguettaya","year":"2025","unstructured":"Bouguettaya A, Stuart EM, Aboujaoude E. Racial bias in AI-mediated psychiatric diagnosis and treatment: a qualitative comparison of four large language models. npj Digit Med. 2025;8:332. https:\/\/doi.org\/10.1038\/s41746-025-01746-4.","journal-title":"npj Digit Med"},{"key":"8644_CR12","doi-asserted-by":"publisher","first-page":"249","DOI":"10.1016\/S0140-6736(19)31146-8","volume":"394","author":"MA Peres","year":"2019","unstructured":"Peres MA, Macpherson LMD, Weyant RJ, Daly B, Venturelli R, Mathur MR, et al. Oral diseases: a global public health challenge. Lancet Lond Engl. 2019;394:249\u201360. https:\/\/doi.org\/10.1016\/S0140-6736(19)31146-8.","journal-title":"Lancet Lond Engl"},{"key":"8644_CR13","doi-asserted-by":"publisher","unstructured":"Hamamci IE, Er S, Durugol OF, Cakmak GR, de la Rosa E, Simsar E et al. DENTEX: Dental Enumeration and Tooth Pathosis Detection Benchmark for Panoramic X-ray. arXiv; 2023. https:\/\/doi.org\/10.48550\/ARXIV.2305.19112.\u00a0 Cited 2026 Mar 8.","DOI":"10.48550\/ARXIV.2305.19112"},{"key":"8644_CR14","doi-asserted-by":"publisher","unstructured":"Hamamci IE, Er S, Simsar E, Sekuboyina A, Gundogar M, Stadlinger B et al. Diffusion-Based Hierarchical Multi-Label Object Detection to Analyze Panoramic Dental X-rays. arXiv; 2023. https:\/\/doi.org\/10.48550\/ARXIV.2303.06500.\u00a0Cited 2026 Mar 8.","DOI":"10.48550\/ARXIV.2303.06500"},{"key":"8644_CR15","doi-asserted-by":"publisher","first-page":"159","DOI":"10.2307\/2529310","volume":"33","author":"JR Landis","year":"1977","unstructured":"Landis JR, Koch GG. The Measurement of Observer Agreement for Categorical Data. Biometrics. 1977;33:159. https:\/\/doi.org\/10.2307\/2529310.","journal-title":"Biometrics"},{"key":"8644_CR16","unstructured":"SCIORA, Power-Analyse. \/ Sample Size Calculator. SCIORA; 2026. https:\/\/sciora.me\/tools\/fallzahlrechner."},{"key":"8644_CR17","doi-asserted-by":"publisher","first-page":"719","DOI":"10.1038\/s41551-023-01056-8","volume":"7","author":"RJ Chen","year":"2023","unstructured":"Chen RJ, Wang JJ, Williamson DFK, Chen TY, Lipkova J, Lu MY, et al. Algorithmic fairness in artificial intelligence for medicine and healthcare. Nat Biomed Eng. 2023;7:719\u201342. https:\/\/doi.org\/10.1038\/s41551-023-01056-8.","journal-title":"Nat Biomed Eng"},{"key":"8644_CR18","doi-asserted-by":"publisher","first-page":"99","DOI":"10.1038\/s41746-020-0304-9","volume":"3","author":"JK Paulus","year":"2020","unstructured":"Paulus JK, Kent DM. Predictably unequal: understanding and addressing concerns that algorithmic clinical prediction may increase health disparities. npj Digit Med. 2020;3:99. https:\/\/doi.org\/10.1038\/s41746-020-0304-9.","journal-title":"npj Digit Med"},{"key":"8644_CR19","doi-asserted-by":"publisher","first-page":"220","DOI":"10.14366\/usg.25012","volume":"44","author":"T Han","year":"2025","unstructured":"Han T, Jeong WK, Shin J. Diagnostic performance of multimodal large language models in radiological quiz cases: the effects of prompt engineering and input conditions. Ultrasonography Seoul Korea. 2025;44:220\u201331. https:\/\/doi.org\/10.14366\/usg.25012.","journal-title":"Ultrasonography Seoul Korea"},{"key":"8644_CR20","doi-asserted-by":"publisher","first-page":"573","DOI":"10.1186\/s12903-025-05926-2","volume":"25","author":"BE Yilmaz","year":"2025","unstructured":"Yilmaz BE, Gokkurt Yilmaz BN, Ozbey F. Artificial intelligence performance in answering multiple-choice oral pathology questions: a comparative analysis. BMC Oral Health. 2025;25:573. https:\/\/doi.org\/10.1186\/s12903-025-05926-2.","journal-title":"BMC Oral Health"},{"key":"8644_CR21","doi-asserted-by":"publisher","first-page":"513","DOI":"10.1146\/annurev-publhealth-040119-094318","volume":"41","author":"ME Northridge","year":"2020","unstructured":"Northridge ME, Kumar A, Kaur R. Disparities in Access to Oral Health Care. Annu Rev Public Health. 2020;41:513\u201335. https:\/\/doi.org\/10.1146\/annurev-publhealth-040119-094318.","journal-title":"Annu Rev Public Health"},{"key":"8644_CR22","doi-asserted-by":"publisher","first-page":"176","DOI":"10.1038\/s43856-024-00601-z","volume":"4","author":"Y Yang","year":"2024","unstructured":"Yang Y, Liu X, Jin Q, Huang F, Lu Z. Unmasking and quantifying racial bias of large language models in medical report generation. Commun Med (Lond). 2024;4:176. https:\/\/doi.org\/10.1038\/s43856-024-00601-z.","journal-title":"Commun Med (Lond)"},{"key":"8644_CR23","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1038\/s41591-024-03425-5","volume":"31","author":"J Gallifant","year":"2025","unstructured":"Gallifant J, Afshar M, Ameen S, Aphinyanaphongs Y, Chen S, Cacciamani G, et al. The TRIPOD-LLM reporting guideline for studies using large language models. Nat Med. 2025;31:60\u20139. https:\/\/doi.org\/10.1038\/s41591-024-03425-5.","journal-title":"Nat Med"}],"container-title":["BMC Oral Health"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12903-026-08644-5","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12903-026-08644-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12903-026-08644-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,12]],"date-time":"2026-06-12T05:24:53Z","timestamp":1781241893000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1186\/s12903-026-08644-5"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,28]]},"references-count":23,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["8644"],"URL":"https:\/\/doi.org\/10.1186\/s12903-026-08644-5","relation":{},"ISSN":["1472-6831"],"issn-type":[{"value":"1472-6831","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,28]]},"assertion":[{"value":"4 April 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"13 May 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"28 May 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"Not applicable.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"1017"}}