{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T09:14:36Z","timestamp":1790241276126,"version":"4.1.0"},"reference-count":53,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T00:00:00Z","timestamp":1790208000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neurol."],"abstract":"\n Background<\/jats:title>\n Early risk assessment is a routine part of neurocritical care, yet mortality prediction remains difficult because neurological injury often coexists with respiratory, circulatory, renal, and inflammatory dysfunction. This study aimed to develop and externally validate an interpretable machine learning (ML) model for predicting in-hospital mortality in neurocritical-care patients using variables available within the first 24 h after intensive care unit (ICU) admission.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n We conducted a retrospective cohort study using two public critical-care databases. The Medical Information Mart for Intensive Care IV (MIMIC-IV) database was used for model development and internal testing, and the eICU Collaborative Research Database (eICU-CRD) was used for independent external validation. The report follows the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) statement. Adult ICU patients with neurocritical-care diagnoses were identified by predefined International Classification of Diseases codes. Candidate predictors were restricted to information recorded during the first 24 h after ICU admission and included demographics, comorbidities, vital signs, laboratory results, neurological status, organ-support therapies, and severity scores. Least absolute shrinkage and selection operator (LASSO) regression was used for feature selection. Logistic regression, random forest, XGBoost, LightGBM, CatBoost, and multilayer perceptron models were trained. Discrimination was assessed by area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC). Calibration, decision curve analysis (DCA), and SHapley Additive exPlanations (SHAP) were used to evaluate probability reliability, potential clinical utility, and interpretability. The operating threshold used for classification metrics was fixed in the MIMIC-IV development data before internal and external evaluation.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n The MIMIC-IV cohort included\n 6,842<\/jats:bold>\n patients, of whom\n 1,014<\/jats:bold>\n died during hospitalization; the held-out internal test set contained\n 1,368<\/jats:bold>\n patients and\n 203<\/jats:bold>\n deaths. The external validation cohort included\n 5,236<\/jats:bold>\n eICU-CRD patients, of whom\n 686<\/jats:bold>\n died. LightGBM showed the best overall performance. In the internal test cohort, it achieved an AUROC of\n 0.841<\/jats:bold>\n (95% CI:\n 0.821\u20130.861<\/jats:bold>\n ) and an AUPRC of\n 0.518<\/jats:bold>\n (95% CI:\n 0.486\u20130.552<\/jats:bold>\n ). In external validation, the AUROC was\n 0.801<\/jats:bold>\n (95% CI:\n 0.779\u20130.824<\/jats:bold>\n ) and the AUPRC was\n 0.413<\/jats:bold>\n (95% CI:\n 0.380\u20130.449<\/jats:bold>\n ), with sensitivity of\n 72.8%<\/jats:bold>\n , specificity of\n 73.9%<\/jats:bold>\n , positive predictive value of\n 30.3%<\/jats:bold>\n , negative predictive value of\n 94.7%<\/jats:bold>\n , and F1 score of\n 0.428<\/jats:bold>\n . Thus, approximately 70% of patients classified as high risk at this operating point survived to hospital discharge. Calibration was acceptable in the external cohort, with a Brier score of\n 0.101<\/jats:bold>\n , calibration intercept of\n 0.050<\/jats:bold>\n , and calibration slope of\n 0.84<\/jats:bold>\n . SHAP analysis identified GCS total score, invasive mechanical ventilation, age, lactate, SOFA score, vasopressor use, and creatinine as leading contributors to predicted mortality risk.\n <\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n A model based on routinely collected first-day ICU data provided externally validated and interpretable risk stratification for in-hospital mortality in neurocritical care. Its low positive predictive value at the reported operating point precludes use as a stand-alone basis for prognostic pessimism, treatment limitation, or withdrawal of life-sustaining treatment. Prospective validation in diverse countries and health systems is required before clinical implementation.<\/jats:p>\n <\/jats:sec>","DOI":"10.3389\/fneur.2026.1888888","type":"journal-article","created":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T08:55:16Z","timestamp":1790240116000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Development and external validation of an interpretable machine learning model for early prediction of in-hospital mortality in neurocritical care"],"prefix":"10.3389","volume":"17","author":[{"given":"Xiao","family":"Cheng","sequence":"first","affiliation":[{"name":"The Neurology Department of Shanxi Provincial People's Hospital, Shanxi Medical University, Shanxi Key Laboratory of Neurobehavior and Brain Disease Control","place":["Taiyuan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jianxin","family":"Ye","sequence":"additional","affiliation":[{"name":"Department of Neurology, The 900th Hospital of the Joint Logistics Support Force of the Chinese People's Liberation Army","place":["Fuzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongen","family":"Wei","sequence":"additional","affiliation":[{"name":"The Neurology Department of Shanxi Provincial People's Hospital, Shanxi Medical University, Shanxi Key Laboratory of Neurobehavior and Brain Disease Control","place":["Taiyuan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,24]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"459","DOI":"10.1016\/S1474-4422(18)30499-X","article-title":"Global, regional, and national burden of neurological disorders, 1990\u20132016: a systematic analysis for the Global Burden of Disease Study 2016","volume":"18","author":"GBD","year":"2019","journal-title":"Lancet Neurol."},{"key":"B2","doi-asserted-by":"publisher","first-page":"795","DOI":"10.1016\/S1474-4422(21)00252-0","article-title":"Global, regional, and national burden of stroke and its risk factors, 1990\u20132019: a systematic analysis for the Global Burden of Disease Study 2019","volume":"20","author":"GBD","year":"2021","journal-title":"Lancet Neurol"},{"key":"B3","doi-asserted-by":"publisher","first-page":"344","DOI":"10.1016\/S1474-4422(24)00038-3","article-title":"Global, regional, and national burden of disorders affecting the nervous system, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021","volume":"23","author":"GBD","year":"2024","journal-title":"Lancet Neurol"},{"key":"B4","doi-asserted-by":"publisher","first-page":"e021499","DOI":"10.1161\/JAHA.121.021499","article-title":"Age-specific and sex-specific trends in life-sustaining care after acute stroke","volume":"10","author":"Joundi","year":"2021","journal-title":"J Am Heart Assoc."},{"key":"B5","doi-asserted-by":"publisher","first-page":"2346","DOI":"10.1161\/01.STR.31.10.2346","article-title":"Cost and outcome of mechanical ventilation for life-threatening stroke","volume":"31","author":"Mayer","year":"2000","journal-title":"Stroke"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1085","DOI":"10.1097\/CCM.0000000000003127","article-title":"Clinical predictors of survival and functional outcome of stroke patients admitted to critical care","volume":"46","author":"van Valburg","year":"2018","journal-title":"Crit Care Med"},{"key":"B7","doi-asserted-by":"publisher","first-page":"235","DOI":"10.1186\/s12871-022-01777-4","article-title":"Outcomes of patients admitted to the ICU for acute stroke: a retrospective cohort","volume":"22","author":"Carval","year":"2022","journal-title":"BMC Anesthesiol."},{"key":"B8","doi-asserted-by":"publisher","first-page":"3","DOI":"10.5334\/aogh.2413","article-title":"Global critical care: moving forward in resource-limited settings","volume":"85","author":"Diaz","year":"2019","journal-title":"Ann Global Health"},{"key":"B9","doi-asserted-by":"publisher","first-page":"81","DOI":"10.1016\/S0140-6736(74)91639-0","article-title":"Assessment of coma and impaired consciousness: a practical scale","volume":"2","author":"Teasdale","year":"1974","journal-title":"Lancet"},{"key":"B10","doi-asserted-by":"publisher","first-page":"707","DOI":"10.1007\/BF01709751","article-title":"The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction\/failure","volume":"22","author":"Vincent","year":"1996","journal-title":"Intensive Care Med."},{"key":"B11","doi-asserted-by":"publisher","first-page":"818","DOI":"10.1097\/00003246-198510000-00009","article-title":"APACHE II a severity of disease classification system","volume":"13","author":"Knaus","year":"1985","journal-title":"Crit Care Med"},{"key":"B12","doi-asserted-by":"publisher","first-page":"878","DOI":"10.1097\/01.CCM.0000201881.58644.41","article-title":"Mortality predictions in the intensive care unit: comparing physicians with scoring systems","volume":"34","author":"Sinuff","year":"2006","journal-title":"Crit Care Med"},{"key":"B13","doi-asserted-by":"publisher","first-page":"447","DOI":"10.1186\/s13054-021-03868-z","article-title":"Demystifying machine learning for mortality prediction","volume":"25","author":"Smit","year":"2021","journal-title":"Crit Care"},{"key":"B14","doi-asserted-by":"publisher","first-page":"1263","DOI":"10.1161\/STROKEAHA.118.024293","article-title":"Machine learning-based model for prediction of outcomes in acute stroke","volume":"50","author":"Heo","year":"2019","journal-title":"Stroke"},{"key":"B15","doi-asserted-by":"publisher","first-page":"26","DOI":"10.1177\/17423953211067435","article-title":"Machine learning-based patient classification system for adults with stroke: a systematic review","volume":"19","author":"Ruksakulpiwat","year":"2023","journal-title":"Chronic Illn"},{"key":"B16","doi-asserted-by":"publisher","first-page":"1385013","DOI":"10.3389\/fneur.2024.1385013","article-title":"Explainable machine learning for predicting neurological outcome in hemorrhagic and ischemic stroke patients in critical care","volume":"15","author":"Wei","year":"2024","journal-title":"Front Neurol"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1007\/s12028-022-01606-z","article-title":"A machine learning approach for the prediction of severe acute kidney injury following traumatic brain injury","author":"Peng","year":"2022","journal-title":"Neurocrit Care"},{"key":"B18","doi-asserted-by":"publisher","first-page":"1734264","DOI":"10.3389\/fneur.2026.1734264","article-title":"An interpretable machine learning model for predicting sepsis risk in ICU patients with non-traumatic subarachnoid hemorrhage: development and validation using the MIMIC-IV database","volume":"17","author":"Guo","year":"2026","journal-title":"Front Neurol."},{"key":"B19","doi-asserted-by":"publisher","first-page":"100674","DOI":"10.1016\/j.imu.2021.100674","article-title":"Explainable machine learning prediction of ICU mortality","volume":"25","author":"Chia","year":"2021","journal-title":"Inform Med Unlocked"},{"key":"B20","doi-asserted-by":"publisher","first-page":"e0262895","DOI":"10.1371\/journal.pone.0262895","article-title":"E-CatBoost: an efficient machine learning framework for predicting ICU mortality using the eICU Collaborative Research Database","volume":"17","author":"Safaei","year":"2022","journal-title":"PLoS ONE"},{"key":"B21","doi-asserted-by":"publisher","first-page":"1068","DOI":"10.3390\/diagnostics12051068","article-title":"Establishment of ICU mortality risk prediction models with machine learning algorithm using MIMIC-IV database","volume":"12","author":"Pang","year":"2022","journal-title":"Diagnostics"},{"key":"B22","doi-asserted-by":"publisher","first-page":"e1091","DOI":"10.1097\/CCM.0000000000004550","article-title":"An explainable artificial intelligence predictor for early detection of sepsis","volume":"48","author":"Yang","year":"2020","journal-title":"Crit Care Med"},{"key":"B23","doi-asserted-by":"publisher","first-page":"905","DOI":"10.1016\/S2213-2600(18)30300-X","article-title":"Machine learning for real-time prediction of complications in critical care: a retrospective study","volume":"6","author":"Meyer","year":"2018","journal-title":"Lancet Respir Med"},{"key":"B24","doi-asserted-by":"publisher","first-page":"76","DOI":"10.1186\/s13054-024-04866-7","article-title":"Real-time machine learning model to predict short-term mortality in critically ill patients: development and international validation","volume":"28","author":"Lim","year":"2024","journal-title":"Crit Care"},{"key":"B25","doi-asserted-by":"publisher","first-page":"e179","DOI":"10.1016\/S2589-7500(20)30018-2","article-title":"Dynamic and explainable machine learning prediction of mortality in patients in the intensive care unit: a retrospective study of high-frequency data in electronic patient records","volume":"2","author":"Thorsen-Meyer","year":"2020","journal-title":"Lancet Digital Health"},{"key":"B26","doi-asserted-by":"publisher","first-page":"733","DOI":"10.1038\/s41746-025-02114-y","article-title":"Unlocking the potential of real-time ICU mortality prediction: redefining risk assessment with continuous data recovery","volume":"8","author":"Xie","year":"2025","journal-title":"NPJ Dig Med"},{"key":"B27","doi-asserted-by":"publisher","first-page":"750","DOI":"10.1007\/s00134-021-06446-7","article-title":"Moving from bytes to bedside: a systematic review on the use of artificial intelligence in the intensive care unit","volume":"47","author":"van de Sande","year":"2021","journal-title":"Intensive Care Med"},{"key":"B28","doi-asserted-by":"publisher","first-page":"1240","DOI":"10.1177\/08850666241277134","article-title":"Artificial intelligence-based models for prediction of mortality in ICU patients: a scoping review","volume":"40","author":"Olang","year":"2025","journal-title":"J Intensive Care Med"},{"key":"B29","doi-asserted-by":"publisher","first-page":"55","DOI":"10.7326\/M14-0697","article-title":"Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement","volume":"162","author":"Collins","year":"2015","journal-title":"Ann Intern Med"},{"key":"B30","doi-asserted-by":"publisher","first-page":"51","DOI":"10.7326\/M18-1376","article-title":"PROBAST: a tool to assess the risk of bias and applicability of prediction model studies","volume":"170","author":"Wolff","year":"2019","journal-title":"Ann Intern Med"},{"key":"B31","doi-asserted-by":"publisher","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":"B32","doi-asserted-by":"publisher","first-page":"e082505","DOI":"10.1136\/bmj-2024-082505","article-title":"PROBAST+AI: an updated quality, risk of bias, and applicability assessment tool for prediction model studies using regression or artificial intelligence methods","volume":"388","author":"Moons","year":"2025","journal-title":"BMJ"},{"key":"B33","doi-asserted-by":"publisher","first-page":"1453","DOI":"10.1016\/S0140-6736(07)61602-X","article-title":"The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies","volume":"370","author":"von Elm","year":"2007","journal-title":"Lancet"},{"key":"B34","doi-asserted-by":"publisher","first-page":"e1001885","DOI":"10.1371\/journal.pmed.1001885","article-title":"The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement","volume":"12","author":"Benchimol","year":"2015","journal-title":"PLoS Med"},{"key":"B35","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41597-022-01899-x","article-title":"MIMIC-IV, a freely accessible electronic health record dataset","volume":"10","author":"Johnson","year":"2023","journal-title":"Sci Data"},{"key":"B36","doi-asserted-by":"publisher","first-page":"180178","DOI":"10.1038\/sdata.2018.178","article-title":"The eICU Collaborative Research Database, a freely available multi-center database for critical care research","volume":"5","author":"Pollard","year":"2018","journal-title":"Sci Data"},{"key":"B37","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","year":"2011","journal-title":"J Stat Softw."},{"key":"B38","doi-asserted-by":"publisher","first-page":"267","DOI":"10.1111\/j.2517-6161.1996.tb02080.x","article-title":"Regression shrinkage and selection via the lasso","volume":"58","author":"Tibshirani","year":"1996","journal-title":"J R Stat Soc."},{"key":"B39","first-page":"2825","article-title":"Scikit-learn: machine learning in Python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J Mach Learn Res."},{"key":"B40","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Mach Learn"},{"key":"B41","doi-asserted-by":"publisher","first-page":"1189","DOI":"10.1214\/aos\/1013203451","article-title":"Greedy function approximation: a gradient boosting machine","volume":"29","author":"Friedman","year":"2001","journal-title":"Ann Stat"},{"key":"B42","doi-asserted-by":"publisher","first-page":"785","DOI":"10.1145\/2939672.2939785","article-title":"XGBoost: a scalable tree boosting system","author":"Chen","year":"2016","journal-title":"Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining"},{"key":"B43","first-page":"3146","article-title":"LightGBM: a highly efficient gradient boosting decision tree","author":"Ke","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"B44","first-page":"6638","article-title":"CatBoost: unbiased boosting with categorical features","author":"Prokhorenkova","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"B45","doi-asserted-by":"publisher","first-page":"233","DOI":"10.1145\/1143844.1143874","article-title":"The relationship between precision-recall and ROC curves","author":"Davis","year":"2006","journal-title":"Proceedings of the 23rd international conference on machine learning"},{"key":"B46","doi-asserted-by":"publisher","first-page":"e0118432","DOI":"10.1371\/journal.pone.0118432","article-title":"The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets","volume":"10","author":"Saito","year":"2015","journal-title":"PLoS ONE"},{"key":"B47","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-16399-0","author":"Steyerberg","year":"2019","journal-title":"Clinical Prediction Models: A Practical Approach to Development, Validation, and Updating"},{"key":"B48","doi-asserted-by":"publisher","first-page":"m441","DOI":"10.1136\/bmj.m441","article-title":"Calculating the sample size required for developing a clinical prediction model","volume":"368","author":"Riley","year":"2020","journal-title":"BMJ"},{"key":"B49","doi-asserted-by":"publisher","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 Dec Mak"},{"key":"B50","first-page":"4765","article-title":"A unified approach to interpreting model predictions","author":"Lundberg","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"B51","doi-asserted-by":"publisher","first-page":"140","DOI":"10.21203\/rs.3.rs-3779429\/v1","article-title":"Using machine learning to predict neurologic injury in venovenous extracorporeal membrane oxygenation recipients: an ELSO Registry analysis","volume":"21","author":"Kalra","year":"2024","journal-title":"JTCVS Open"},{"key":"B52","doi-asserted-by":"publisher","first-page":"403","DOI":"10.1007\/s12028-025-02233-0","article-title":"A novel, interpretable machine learning model to predict neurological outcomes following venoarterial extracorporeal membrane oxygenation","volume":"43","author":"Shou","year":"2025","journal-title":"Neurocr Care"},{"key":"B53","doi-asserted-by":"publisher","first-page":"e208081","DOI":"10.1212\/WNL.0000000000208081","article-title":"Long-term neuropsychiatric, neurocognitive, and functional outcomes of patients receiving ECMO: a systematic review and meta-analysis","volume":"102","author":"Kalra","year":"2024","journal-title":"Neurology"}],"container-title":["Frontiers in Neurology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fneur.2026.1888888\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T08:55:17Z","timestamp":1790240117000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fneur.2026.1888888\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,24]]},"references-count":53,"alternative-id":["10.3389\/fneur.2026.1888888"],"URL":"https:\/\/doi.org\/10.3389\/fneur.2026.1888888","relation":{},"ISSN":["1664-2295"],"issn-type":[{"value":"1664-2295","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,24]]},"article-number":"1888888"}}