{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T17:29:24Z","timestamp":1787074164574,"version":"build-2736575974"},"reference-count":45,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"Abstract<\/jats:title>\n \n Background<\/jats:title>\n Postoperative neurological complications (PNCs) after acute type A aortic dissection (ATAAD) surgery are clinically emergent and require multidimensional perioperative risk assessment. Large language models (LLMs) have shown potential in clinical prediction, but the incremental value of structured multiagent collaboration remains unclear.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aimed to develop and validate a multidisciplinary team (MDT)\u2013based LLM framework for predicting PNC after ATAAD surgery and to compare its performance with that of traditional machine learning (ML) models and single-agent LLM settings.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n \n A retrospective cohort from January 2020 to June 2024 (N=763) was randomly divided into a training set (n=533) and an internal validation set (n=230). A prospective cohort from July 2024 to June 2025 (n=120) was used for prospective validation. The outcome was PNC, defined as stroke, cerebral hemorrhage, paraplegia, or coma. Population-level in-context learning used outcome-stratified summary statistics from the training cohort, including predictor distributions in patients with and without PNC and between-group\n P<\/jats:italic>\n values. Two LLMs (DeepSeek-V3 and ChatGPT [GPT-5]) were evaluated under 4 settings: no MDT without in-context learning, MDT without in-context learning, no MDT with in-context learning, and MDT with in-context learning. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, accuracy,\n F<\/jats:italic>\n 1<\/jats:sub>\n -score, and Brier score.\n <\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n PNC occurred in 13.0% (99\/763) of patients in the retrospective cohort and 15.8% (19\/120) in the prospective cohort. Among the ML models, the random forest achieved the highest AUC in the internal validation set (AUC 0.7857, 95% CI 0.6945\u20100.8768). Among the LLM configurations, GPT-5 with MDT and in-context learning achieved the highest AUC (AUC 0.8419, 95% CI 0.7398\u20100.9440), with a sensitivity of 80.00%, specificity of 87.00%, and Brier score of 0.0876. Although this configuration showed a significantly higher AUC than the fully unaided GPT-5 baseline (\n P<\/jats:italic>\n =.006), this improvement reflected the combined effect of MDT and in-context learning, as adding the MDT framework within matched settings did not significantly improve AUC over corresponding single-agent settings in either cohort. The AUC of GPT-5 with MDT and in-context learning was also not significantly higher than that of the random forest model (\n P<\/jats:italic>\n =.18). Word frequency analysis showed that the generated rationales following in-context learning more frequently mentioned variables that were statistically significant in the provided context.\n <\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n The GPT-5 configuration combining MDT-style collaboration with population-level in-context learning showed promising discrimination for PNC prediction after ATAAD surgery. However, the MDT layer did not produce a statistically significant incremental improvement in AUC over the corresponding single-agent settings, and its predictive contribution remains to be confirmed. The framework generated structured, role-specific rationales, supporting its further evaluation as a proof-of-concept approach for postoperative risk stratification. Larger multicenter studies are required before routine clinical implementation.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/91257","type":"journal-article","created":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T17:10:45Z","timestamp":1787073045000},"page":"e91257-e91257","source":"Crossref","is-referenced-by-count":0,"title":["A Multidisciplinary Team\u2013Based Large Language Model Framework for Predicting Postoperative Neurological Complications in Acute Type A Aortic Dissection: Model Development and Validation Study"],"prefix":"10.2196","volume":"28","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7051-0503","authenticated-orcid":false,"given":"Jili","family":"Li","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-4542-9941","authenticated-orcid":false,"given":"Julin","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-3792-2359","authenticated-orcid":false,"given":"Xingrui","family":"Tao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-3711-310X","authenticated-orcid":false,"given":"Yaoye","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-8198-8175","authenticated-orcid":false,"given":"Siqi","family":"Xun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-3772-2006","authenticated-orcid":false,"given":"Guangshuo","family":"Qin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6497-1020","authenticated-orcid":false,"given":"Kaiyue","family":"Diao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7340-0628","authenticated-orcid":false,"given":"Chaoyi","family":"Qin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,8,18]]},"reference":[{"issue":"1","key":"R1","doi-asserted-by":"publisher","first-page":"9835","DOI":"10.1038\/s41467-025-64802-3","article-title":"Age modulates the link between stress-related neural activity and mortality","volume":"16","author":"Mikail","journal-title":"Nat Commun"},{"key":"R2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijmedinf.2025.106177","article-title":"Medication-based mortality prediction in COPD using machine learning and conventional statistical methods","volume":"206","author":"Pena-Gralle","journal-title":"Int J Med Inform"},{"issue":"8","key":"R3","doi-asserted-by":"publisher","DOI":"10.2196\/38082","article-title":"Predicting mortality in intensive care unit patients with heart failure using an interpretable machine learning model: retrospective cohort study","volume":"24","author":"Li","journal-title":"J Med Internet Res"},{"issue":"3","key":"R4","doi-asserted-by":"publisher","first-page":"869","DOI":"10.1038\/s41591-024-03398-5","article-title":"Prediction of checkpoint inhibitor immunotherapy efficacy for cancer using routine blood tests and clinical data","volume":"31","author":"Yoo","journal-title":"Nat Med"},{"issue":"11","key":"R5","doi-asserted-by":"publisher","first-page":"e848","DOI":"10.1093\/cid\/ciaa1576","article-title":"Predicting antibiotic resistance in hospitalized patients by applying machine learning to electronic medical records","volume":"72","author":"Lewin-Epstein","journal-title":"Clin Infect Dis"},{"issue":"1","key":"R6","doi-asserted-by":"publisher","DOI":"10.1038\/s43856-023-00275-z","article-title":"Prediction of ciprofloxacin resistance in hospitalized patients using machine learning","volume":"3","author":"Mintz","journal-title":"Commun Med (Lond)"},{"issue":"12","key":"R7","doi-asserted-by":"publisher","first-page":"2670","DOI":"10.1093\/jamia\/ocab211","article-title":"Predicting next-day discharge via electronic health record access logs","volume":"28","author":"Zhang","journal-title":"J Am Med Inform Assoc"},{"key":"R8","volume":"2024","author":"Zhang","journal-title":"AMIA Annu Symp Proc"},{"issue":"2","key":"R9","doi-asserted-by":"publisher","first-page":"332","DOI":"10.1097\/TA.0000000000004805","article-title":"Augmenting decision making in acute care surgery: a systematic review of machine learning-driven risk prediction models","volume":"100","author":"Lee","journal-title":"J Trauma Acute Care Surg"},{"issue":"11","key":"R10","doi-asserted-by":"publisher","DOI":"10.1002\/mco2.70460","article-title":"Artificial intelligence for the diagnosis and management of cancers: potentials and challenges","volume":"6","author":"Wang","journal-title":"MedComm (2020)"},{"key":"R11","doi-asserted-by":"publisher","DOI":"10.2196\/85057","article-title":"Large language model\u2013based analysis of statin therapy discussions and sentiment on social media: cross-sectional observational study","volume":"28","author":"Liu","journal-title":"J Med Internet Res"},{"issue":"9","key":"R12","doi-asserted-by":"publisher","first-page":"2114","DOI":"10.1093\/jamia\/ocae074","article-title":"Large language models for biomedicine: foundations, opportunities, challenges, and best practices","volume":"31","author":"Sahoo","journal-title":"J Am Med Inform Assoc"},{"issue":"1","key":"R13","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-024-01366-4","article-title":"Probabilistic medical predictions of large language models","volume":"7","author":"Gu","journal-title":"NPJ Digit Med"},{"issue":"9","key":"R14","doi-asserted-by":"publisher","first-page":"2977","DOI":"10.1038\/s41591-025-03856-8","article-title":"Large language model-based biological age prediction in large-scale populations","volume":"31","author":"Li","journal-title":"Nat Med"},{"key":"R15","doi-asserted-by":"publisher","DOI":"10.2196\/48659","article-title":"Assessing the utility of ChatGPT throughout the entire clinical workflow: development and usability study","volume":"25","author":"Rao","journal-title":"J Med Internet Res"},{"key":"R16","doi-asserted-by":"publisher","DOI":"10.2196\/48808","article-title":"ChatGPT-generated differential diagnosis lists for complex case-derived clinical vignettes: diagnostic accuracy evaluation","volume":"11","author":"Hirosawa","journal-title":"JMIR Med Inform"},{"issue":"12","key":"R17","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pdig.0000680","article-title":"CPLLM: clinical prediction with large language models","volume":"3","author":"Ben Shoham","journal-title":"PLOS Digit Health"},{"issue":"1","key":"R18","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-025-01684-1","article-title":"Evaluating large language model workflows in clinical decision support for triage and referral and diagnosis","volume":"8","author":"Gaber","journal-title":"NPJ Digit Med"},{"issue":"5","key":"R19","doi-asserted-by":"publisher","first-page":"811","DOI":"10.1093\/jamia\/ocaf038","article-title":"Large language models are less effective at clinical prediction tasks than locally trained machine learning models","volume":"32","author":"Brown","journal-title":"J Am Med Inform Assoc"},{"issue":"1","key":"R20","doi-asserted-by":"publisher","first-page":"78","DOI":"10.1001\/jama.2023.8288","article-title":"Accuracy of a generative artificial intelligence model in a complex diagnostic challenge","volume":"330","author":"Kanjee","journal-title":"JAMA"},{"issue":"10","key":"R21","doi-asserted-by":"publisher","first-page":"813","DOI":"10.1093\/eurheartj\/ehac757","article-title":"Acute aortic dissection: evidence, uncertainties, and future therapies","volume":"44","author":"Rylski","journal-title":"Eur Heart J"},{"key":"R22","doi-asserted-by":"publisher","DOI":"10.1016\/j.xjon.2026.101945","article-title":"Hemiarch repair versus total-arch repair with frozen elephant trunk for acute type A aortic dissection: A propensity score\u2013matched study","author":"Zhang","journal-title":"JTCVS Open"},{"issue":"17","key":"R23","doi-asserted-by":"publisher","first-page":"1846","DOI":"10.1161\/CIRCULATIONAHA.117.031264","article-title":"Insights from the International Registry of Acute Aortic Dissection: a 20-year experience of collaborative clinical research","volume":"137","author":"Evangelista","journal-title":"Circulation"},{"issue":"1","key":"R24","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1016\/j.athoracsur.2020.01.026","article-title":"Immediate surgery in acute type A dissection and neurologic dysfunction: fighting the inevitable?","volume":"110","author":"Dumfarth","journal-title":"Ann Thorac Surg"},{"issue":"6","key":"R25","doi-asserted-by":"publisher","first-page":"2082","DOI":"10.1016\/s0003-4975(00)02160-3","article-title":"Outcome with high blood lactate levels during cardiopulmonary bypass in adult cardiac operation","volume":"70","author":"Demers","journal-title":"Ann Thorac Surg"},{"issue":"6","key":"R26","doi-asserted-by":"publisher","first-page":"1128","DOI":"10.1016\/j.athoracsur.2024.02.026","article-title":"Axillary vs femoral arterial cannulation in acute type A dissection: international multicenter data","volume":"117","author":"Elbatarny","journal-title":"Ann Thorac Surg"},{"issue":"18","key":"R27","doi-asserted-by":"publisher","DOI":"10.3390\/jcm14186392","article-title":"Unilateral vs. bilateral selective cerebral perfusion for acute type A aortic dissection with frozen elephant trunk: systematic review and meta-analysis","volume":"14","author":"Baudo","journal-title":"J Clin Med"},{"issue":"1","key":"R28","doi-asserted-by":"publisher","DOI":"10.1136\/bmjoq-2022-002050","article-title":"Quality and process improvement of the multidisciplinary Heart Team meeting using Lean Six Sigma","volume":"12","author":"Hoefsmit","journal-title":"BMJ Open Qual"},{"issue":"1","key":"R29","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-025-01550-0","article-title":"Enhancing diagnostic capability with multi-agents conversational large language models","volume":"8","author":"Chen","journal-title":"NPJ Digit Med"},{"issue":"8","key":"R30","doi-asserted-by":"publisher","DOI":"10.1002\/aisy.202400840","article-title":"A proactive agent collaborative framework for zero-shot multimodal medical reasoning","volume":"7","author":"Gu","journal-title":"Adv Intell Syst"},{"key":"R31","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2020.00085","article-title":"The multidisciplinary team (MDT) approach and quality of care","volume":"10","author":"Taberna","journal-title":"Front Oncol"},{"key":"R32","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijmedinf.2025.106136","article-title":"MedARC: adaptive multi-agent refinement and collaboration for enhanced medical reasoning in large language models","volume":"206","author":"Miao","journal-title":"Int J Med Inform"},{"issue":"10","key":"R33","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pdig.0000787","article-title":"Collaborative intelligence in AI: evaluating the performance of a council of AIs on the USMLE","volume":"4","author":"Shaikh","journal-title":"PLOS Digit Health"},{"key":"R34","doi-asserted-by":"publisher","DOI":"10.1136\/bmj-2023-078378","article-title":"TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods","volume":"385","author":"Collins","journal-title":"BMJ"},{"key":"R35","unstructured":"Lundberg SM Lee SI . A unified approach to interpreting model predictions. In: Guyon I Von Luxburg U Bengio S Wallach H Fergus R Vishwanathan S Garnett R , editors. Advances in Neural Information Processing System. Neural Information Processing Systems Foundation; 2017:4765-4774. URL: https:\/\/proceedings.neurips.cc\/paper_files\/paper\/2017\/file\/8a20a8621978632d76c43dfd28b67767-Paper.pdf [Accessed 31-07-2026]"},{"key":"R36","unstructured":"Wu Q Bansal G Zhang J Wu Y Li B Zhu E et al. AutoGen: enabling next-gen LLM applications via multi-agent conversations. Presented at: Presented at: First Conference on Language Modeling (COLM 2024); Oct 7-9, 2024; Philadelphia, PA. URL: https:\/\/openreview.net\/pdf?id=BAakY1hNKS [Accessed 19-07-2026]"},{"issue":"3","key":"R37","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v045.i03","article-title":"MICE: multivariate imputation by chained equations in R","volume":"45","author":"Van Buuren","journal-title":"J Stat Softw"},{"issue":"10","key":"R38","doi-asserted-by":"publisher","DOI":"10.7759\/cureus.94300","article-title":"Performance of ChatGPT and large language models on medical licensing exams worldwide: a systematic review and network meta-analysis with meta-regression","volume":"17","author":"Kasagga","journal-title":"Cureus"},{"issue":"9","key":"R39","doi-asserted-by":"publisher","first-page":"2613","DOI":"10.1038\/s41591-024-03097-1","article-title":"Evaluation and mitigation of the limitations of large language models in clinical decision-making","volume":"30","author":"Hager","journal-title":"Nat Med"},{"key":"R40","doi-asserted-by":"crossref","unstructured":"Tang X Zou A Zhang Z et al. MedAgents: large language models as collaborators for zero-shot medical reasoning. In: Ku LW Martins A Srikumar V , editors. Findings of the Association for Computational Linguistics. Association for Computational Linguistics; 2024:599-621. [doi: 10.18653\/v1\/2024.findings-acl.33]","DOI":"10.18653\/v1\/2024.findings-acl.33"},{"issue":"1","key":"R41","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-026-02443-6","article-title":"Benchmarking large language model-based agent systems for clinical decision tasks","volume":"9","author":"Liu","journal-title":"NPJ Digit Med"},{"key":"R42","doi-asserted-by":"publisher","DOI":"10.2196\/69742","article-title":"Advantages and inconveniences of a multi-agent large language model system to mitigate cognitive biases in diagnostic challenges","volume":"27","author":"Bousquet","journal-title":"J Med Internet Res"},{"issue":"3","key":"R43","doi-asserted-by":"publisher","first-page":"943","DOI":"10.1038\/s41591-024-03423-7","article-title":"Toward expert-level medical question answering with large language models","volume":"31","author":"Singhal","journal-title":"Nat Med"},{"key":"R44","volume":"2024","author":"Cui","journal-title":"AMIA Annu Symp Proc"},{"issue":"1","key":"R45","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-025-01519-z","article-title":"Retrieval augmented generation for 10 large language models and its generalizability in assessing medical fitness","volume":"8","author":"Ke","journal-title":"NPJ Digit Med"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","deposited":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T17:10:47Z","timestamp":1787073047000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2026\/1\/e91257"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,18]]},"references-count":45,"URL":"https:\/\/doi.org\/10.2196\/91257","relation":{},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,18]]},"article-number":"v28i1e91257"}}