{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T21:37:42Z","timestamp":1790285862976,"version":"4.1.0"},"reference-count":30,"publisher":"MDPI AG","issue":"9","license":[{"start":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T00:00:00Z","timestamp":1788307200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Medicina"],"abstract":"<jats:p>Background and Objectives: Delayed cerebral ischemia (DCI) is a major determinant of poor outcome after aneurysmal subarachnoid hemorrhage (aSAH), yet early risk prediction remains difficult, particularly in sedated or ventilated intensive care patients. We evaluated whether a large language model (LLM) could predict DCI from phase-based dynamic clinical data, compared with intensive care specialists. Materials and Methods: In this single-center, retrospective study, 216 consecutive patients with aSAH were assessed at three predefined phases of accumulating clinical data (day 1; days 1 + 3; days 1 + 3 + 5). For each patient\u2013phase, an LLM (ChatGPT, GPT-5.5 Thinking) and two blinded intensive care specialists predicted DCI risk using only the data available up to that time point. DCI was adjudicated by a blinded three-member panel. Discrimination was assessed by the area under the receiver operating characteristic curve (AUC), with non-inferiority defined a priori as \u0394 = 0.10. Calibration, decision-curve analysis, and reproducibility were also assessed. Results: Of 216 patients, 60 (27.8%) were DCI-positive and 15 were indeterminate; the primary sample comprised 201 patients. In the prespecified primary analysis, the LLM met the non-inferiority criterion relative to both specialists across all three phases (LLM AUC 0.703\u20130.747; specialists 0.712\u20130.764); however, in an equal-granularity sensitivity analysis, non-inferiority remained supported only in Phases 2 and 3 and was not demonstrated in Phase 1. Discrimination increased numerically as data accumulated (Phase 1 vs. 3, p = 0.072), an increase that was attenuated in a landmark-restricted analysis accounting for DCI-onset timing. The LLM showed the lowest false-reassurance rate (3.3\u201311.7%), reflecting a more cautious threshold rather than better discrimination. Confidence did not reliably indicate accuracy (~25% of high-confidence predictions were wrong); outputs were highly reproducible (Fleiss \u03ba 0.887; intraclass correlation coefficient (ICC) 0.968). Conclusions: An LLM achieved DCI discrimination that was non-inferior to\u2014but not better than\u2014that of experienced specialists; its unreliable confidence scores support clinician-supervised rather than autonomous use.<\/jats:p>","DOI":"10.3390\/medicina62091680","type":"journal-article","created":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T07:53:11Z","timestamp":1788335591000},"page":"1680","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Phase-Based Dynamic Prediction of Delayed Cerebral Ischemia After Aneurysmal Subarachnoid Hemorrhage: Comparison of a Large Language Model with Intensive Care Specialists"],"prefix":"10.3390","volume":"62","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6210-1083","authenticated-orcid":false,"given":"Mustafa","family":"Ay","sequence":"first","affiliation":[{"name":"Department of Intensive Care, Antalya Training and Research Hospital, 07100 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0242-6307","authenticated-orcid":false,"given":"Dilara","family":"T\u00fcfek \u00d6ztan","sequence":"additional","affiliation":[{"name":"Department of Intensive Care, Antalya Training and Research Hospital, 07100 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"investigation","vocabulary":"credit"},{"role":"data-curation","vocabulary":"credit"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"given":"\u015eule","family":"Asri","sequence":"additional","affiliation":[{"name":"Department of Intensive Care, Antalya Training and Research Hospital, 07100 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"data-curation","vocabulary":"credit"},{"role":"writing-review-editing","vocabulary":"credit"},{"role":"investigation","vocabulary":"credit"}]},{"given":"Arzu","family":"Karaveli","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Antalya Training and Research Hospital, 07100 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"validation","vocabulary":"credit"},{"role":"investigation","vocabulary":"credit"},{"role":"supervision","vocabulary":"credit"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"given":"Nazife","family":"\u00d6zt\u00fcrk","sequence":"additional","affiliation":[{"name":"Research and Development Department, Antalya Training and Research Hospital, 07100 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ahmet \u015e\u00fckr\u00fc","family":"Alparslan","sequence":"additional","affiliation":[{"name":"Department of Radiology, Antalya Training and Research Hospital, 07100 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"validation","vocabulary":"credit"},{"role":"investigation","vocabulary":"credit"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3871-9884","authenticated-orcid":false,"given":"Serap","family":"Avc\u0131 Ay","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Kemer State Hospital, 07980 Antalya, T\u00fcrkiye"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,9,2]]},"reference":[{"key":"ref_1","first-page":"e314","article-title":"2023 guideline for the management of patients with aneurysmal subarachnoid hemorrhage: A guideline from the American Heart Association\/American Stroke Association","volume":"54","author":"Hoh","year":"2023","journal-title":"Stroke"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"1711","DOI":"10.1161\/STR.0b013e3182587839","article-title":"Guidelines for the management of aneurysmal subarachnoid hemorrhage: A guideline for healthcare professionals from the American Heart Association\/American Stroke Association","volume":"43","author":"Connolly","year":"2012","journal-title":"Stroke"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"1365","DOI":"10.1136\/jnnp.2007.117655","article-title":"Incidence of subarachnoid haemorrhage: A systematic review with emphasis on region, age, gender and time trends","volume":"78","author":"Linn","year":"2007","journal-title":"J. Neurol. Neurosurg. Psychiatry"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"2391","DOI":"10.1161\/STROKEAHA.110.589275","article-title":"Definition of delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage as an outcome event in clinical trials and observational studies: Proposal of a multidisciplinary research group","volume":"41","author":"Vergouwen","year":"2010","journal-title":"Stroke"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"e81","DOI":"10.1161\/STROKEAHA.122.040072","article-title":"Early brain injury after subarachnoid hemorrhage: Pathophysiology and emerging therapies","volume":"54","author":"Lauzier","year":"2023","journal-title":"Stroke"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"1457","DOI":"10.1136\/jnnp.51.11.1457","article-title":"A universal subarachnoid hemorrhage scale: Report of a committee of the World Federation of Neurosurgical Societies","volume":"51","author":"Teasdale","year":"1988","journal-title":"J. Neurol. Neurosurg. Psychiatry"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"2012","DOI":"10.1161\/hs0901.095677","article-title":"Effect of cisternal and ventricular blood on risk of delayed cerebral ischemia after subarachnoid hemorrhage","volume":"32","author":"Claassen","year":"2001","journal-title":"Stroke"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"137","DOI":"10.1093\/neuros\/nyx364","article-title":"The Subarachnoid Hemorrhage Early Brain Edema Score predicts delayed cerebral ischemia and clinical outcomes","volume":"83","author":"Ahn","year":"2018","journal-title":"Neurosurgery"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"24","DOI":"10.1038\/s41591-018-0316-z","article-title":"A guide to deep learning in healthcare","volume":"25","author":"Esteva","year":"2019","journal-title":"Nat. Med."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"31","DOI":"10.1038\/s41591-021-01614-0","article-title":"AI in health and medicine","volume":"28","author":"Rajpurkar","year":"2022","journal-title":"Nat. Med."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"e54121","DOI":"10.2196\/54121","article-title":"Machine learning for the early prediction of delayed cerebral ischemia in patients with subarachnoid hemorrhage: Systematic review and meta-analysis","volume":"27","author":"Zhang","year":"2025","journal-title":"J. Med. Internet Res."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Kung, T.H., Cheatham, M., Medenilla, A., Sillos, C., De Leon, L., Elepa\u00f1o, C., Madriaga, M., Aggabao, R., Diaz-Candido, G., and Maningo, J. (2023). Performance of ChatGPT on USMLE: Potential for AI-assisted medical education using large language models. PLoS Digit Health, 2.","DOI":"10.1371\/journal.pdig.0000198"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Sallam, M. (2023). ChatGPT utility in healthcare education, research, and practice: Systematic review on the promising perspectives and valid concerns. Healthcare, 11.","DOI":"10.3390\/healthcare11060887"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"44","DOI":"10.1038\/s41591-018-0300-7","article-title":"High-performance medicine: The convergence of human and artificial intelligence","volume":"25","author":"Topol","year":"2019","journal-title":"Nat. Med."},{"key":"ref_15","doi-asserted-by":"crossref","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":"ref_16","doi-asserted-by":"crossref","first-page":"e745","DOI":"10.1016\/S2589-7500(21)00208-9","article-title":"The false hope of current approaches to explainable artificial intelligence in health care","volume":"3","author":"Ghassemi","year":"2021","journal-title":"Lancet Digit Health"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"60","DOI":"10.1038\/s41591-024-03425-5","article-title":"The TRIPOD-LLM reporting guideline for studies using large language models","volume":"31","author":"Gallifant","year":"2025","journal-title":"Nat. Med."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"e078378","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","year":"2024","journal-title":"BMJ"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"328","DOI":"10.1148\/radiol.12120725","article-title":"How to demonstrate similarity by using noninferiority and equivalence statistical testing in radiology research","volume":"267","author":"Ahn","year":"2013","journal-title":"Radiology"},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"1076","DOI":"10.1007\/s10278-024-01258-9","article-title":"Automated ASPECTS Segmentation and Scoring Tool: A Method Tailored for a Colombian Telestroke Network","volume":"38","author":"Ortiz","year":"2025","journal-title":"J. Imaging Inf. Med."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"2445","DOI":"10.1056\/NEJMp1303712","article-title":"From mindless to mindful practice\u2014Cognitive bias and clinical decision making","volume":"368","author":"Croskerry","year":"2013","journal-title":"N. Engl. J. Med."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Saposnik, G., Redelmeier, D., Ruff, C.C., and Tobler, P.N. (2016). Cognitive biases associated with medical decisions: A systematic review. BMC Med. Inf. Decis. Mak., 16.","DOI":"10.1186\/s12911-016-0377-1"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach","volume":"44","author":"DeLong","year":"1988","journal-title":"Biometrics"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Van Calster, B., McLernon, D.J., van Smeden, M., Wynants, L., and Steyerberg, E.W. (2019). Topic Group \u201cEvaluating diagnostic tests and prediction models\u201d of the STRATOS initiative. Calibration: The Achilles heel of predictive analytics. BMC Med., 17.","DOI":"10.1186\/s12916-019-1466-7"},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"565","DOI":"10.1177\/0272989X06295361","article-title":"Decision curve analysis: A novel method for evaluating prediction models","volume":"26","author":"Vickers","year":"2006","journal-title":"Med. Decis. Mak."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"928","DOI":"10.1001\/jamasurg.2024.1621","article-title":"Large language model capabilities in perioperative risk prediction and prognostication","volume":"159","author":"Chung","year":"2024","journal-title":"JAMA Surg."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"38890","DOI":"10.1038\/s41598-025-22634-7","article-title":"A large language model for delirium prediction in the intensive care unit using structured electronic health records","volume":"15","author":"Contreras","year":"2025","journal-title":"Sci. Rep."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"175","DOI":"10.1038\/s41746-025-01543-z","article-title":"A systematic review and meta-analysis of diagnostic performance comparison between generative AI and physicians","volume":"8","author":"Takita","year":"2025","journal-title":"npj Digit Med."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"e301654","DOI":"10.1136\/bmjment-2025-301654","article-title":"Reasoning language models for more transparent prediction of suicide risk","volume":"28","author":"McCoy","year":"2025","journal-title":"BMJ Ment. Health"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"e2522866","DOI":"10.1001\/jamanetworkopen.2025.22866","article-title":"Operationalization of artificial intelligence applications in the intensive care unit: A systematic review","volume":"8","author":"Berkhout","year":"2025","journal-title":"JAMA Netw. Open"}],"container-title":["Medicina"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1648-9144\/62\/9\/1680\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T08:21:49Z","timestamp":1788337309000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1648-9144\/62\/9\/1680"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,2]]},"references-count":30,"journal-issue":{"issue":"9","published-online":{"date-parts":[[2026,9]]}},"alternative-id":["medicina62091680"],"URL":"https:\/\/doi.org\/10.3390\/medicina62091680","relation":{},"ISSN":["1648-9144"],"issn-type":[{"value":"1648-9144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,2]]}}}