{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,5]],"date-time":"2026-11-05T07:12:38Z","timestamp":1793862758557,"version":"4.6.1"},"reference-count":34,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,3,16]],"date-time":"2026-03-16T00:00:00Z","timestamp":1773619200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Integrated Traditional Chinese and Western Medicine Chronic Disease Management Research Project of China","award":["CXZH2024014-10"],"award-info":[{"award-number":["CXZH2024014-10"]}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neurol."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Timely identification of patients at high risk for early neurological deterioration (END) is critical for effective intervention after acute ischemic stroke; however, existing prediction models largely rely on structured clinical data and underutilize semantic information from imaging findings.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Patients and methods<\/jats:title>\n                    <jats:p>In this retrospective single-center study at Xiangyang No.1 People\u2019s Hospital (January 2018\u2013December 2023), 426 patients with acute ischemic stroke and imaging-confirmed middle cerebral artery occlusion who received non-endovascular treatment were included. Patients with other arterial occlusions, endovascular therapy, or incomplete data were excluded. END was defined as a \u22652-point increase in total National Institutes of Health Stroke Scale (NIHSS) score or a \u22651-point increase in the motor subscore within 7\u202fdays after admission. Structured clinical variables and radiology report text collected at admission were integrated into a multimodal deep learning model. Model performance was evaluated using AUC, accuracy, recall, precision, F1-score, calibration, and decision curve analysis, with interpretability assessed using SHAP and Integrated Gradients.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Among all patients, 38.0% exhibited END (30.3% early; 7.7% late). The multimodal Concat-Fusion model achieved AUCs of 0.877 (training) and 0.771 (test), surpassing single-modality models, with strong predictive capabilities for early (AUC\u202f=\u202f0.842) and late END (AUC\u202f=\u202f0.855). Subgroup analyses confirmed consistent performance across NIHSS scores, D-dimer levels, hypertension, and atrial fibrillation, with significantly higher AUC in diabetic patients (\n                      <jats:italic>p<\/jats:italic>\n                      \u202f=\u202f0.002). SHAP analysis identified D-dimer, diastolic blood pressure, and heart rate as key contributors. Integrated Gradients revealed symptom descriptors including \u201cunsteadiness,\u201d \u201cwalking,\u201d and \u201cvision\u201d as significant textual predictors. Risk stratification effectively distinguished high- and low-risk groups with significantly different cumulative END incidences within 48\u202fh and 7\u202fdays (log-rank\n                      <jats:italic>p<\/jats:italic>\n                      \u202f&amp;lt;\u202f0.0001). High-risk patients exhibited poorer NIHSS trajectories during 14-day follow-up.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This multimodal prediction model may improve early identification of END and support individualized clinical management. Larger prospective studies are required to validate its clinical utility.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fneur.2026.1787921","type":"journal-article","created":{"date-parts":[[2026,3,16]],"date-time":"2026-03-16T05:23:34Z","timestamp":1773638614000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["A multimodal deep learning model for predicting early neurological deterioration in patients with acute ischemic stroke"],"prefix":"10.3389","volume":"17","author":[{"given":"Chenglin","family":"Sun","sequence":"first","affiliation":[{"name":"Department of Neurology, Renmin Hospital of Wuhan University","place":["Wuhan, Hubei, China"]},{"name":"Xiangyang No.1 People\u2019s Hospital, Hubei University of Medicine","place":["Xiangyang, Hubei, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yaqiong","family":"Zhang","sequence":"additional","affiliation":[{"name":"Xiangyang No.1 People\u2019s Hospital, Hubei University of Medicine","place":["Xiangyang, Hubei, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peiyang","family":"Zhou","sequence":"additional","affiliation":[{"name":"Xiangyang No.1 People\u2019s Hospital, Hubei University of Medicine","place":["Xiangyang, Hubei, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zuneng","family":"Lu","sequence":"additional","affiliation":[{"name":"Department of Neurology, Renmin Hospital of Wuhan University","place":["Wuhan, Hubei, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,3,16]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"17562864251357274","DOI":"10.1177\/17562864251357274","article-title":"Early intensive therapy for preventing neurological deterioration in branch atheromatous disease","volume":"18","author":"Huang","year":"2025","journal-title":"Ther Adv Neurol Disord"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"52","DOI":"10.31083\/j.jin2202052","article-title":"Predictors of early neurological deterioration occurring within 24 H in acute ischemic stroke following reperfusion therapy: a systematic review and Meta-analysis","volume":"22","author":"Shi","year":"2023","journal-title":"J Integr Neurosci"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"216","DOI":"10.1007\/s00415-017-8490-3","article-title":"Revisiting 'progressive stroke': incidence, predictors, pathophysiology, and Management of Unexplained Early Neurological Deterioration Following Acute Ischemic Stroke","volume":"265","author":"Seners","year":"2018","journal-title":"J Neurol"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"87","DOI":"10.1136\/jnnp-2014-308327","article-title":"Incidence, causes and predictors of neurological deterioration occurring within 24 H following acute Ischaemic stroke: a systematic review with pathophysiological implications","volume":"86","author":"Seners","year":"2015","journal-title":"J Neurol Neurosurg Psychiatry"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"e549","DOI":"10.1016\/j.jstrokecerebrovasdis.2013.06.012","article-title":"A proposal for the classification of etiologies of neurologic deterioration after acute ischemic stroke","volume":"22","author":"Siegler","year":"2013","journal-title":"J Stroke Cerebrovasc Dis"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"383","DOI":"10.1016\/j.jstrokecerebrovasdis.2015.10.010","article-title":"The etiologies of early neurological deterioration after thrombolysis and risk factors of ischemia progression","volume":"25","author":"Kim","year":"2016","journal-title":"J Stroke Cerebrovasc Dis"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1161\/strokeaha.125.053320","article-title":"Management strategies for early neurological deterioration in noncardioembolic ischemic stroke","author":"Kim","year":"2025","journal-title":"Stroke"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"179","DOI":"10.5853\/jos.2022.02936","article-title":"Collateral circulation in ischemic stroke: an updated review","volume":"25","author":"Maguida","year":"2023","journal-title":"J Stroke"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"2615","DOI":"10.1161\/STROKEAHA.120.030982","article-title":"Early neurological deterioration after intravenous thrombolysis","volume":"51","author":"Mitsias","year":"2020","journal-title":"Stroke"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"1518685","DOI":"10.3389\/fstro.2025.1518685","article-title":"Early neurological deterioration following acute stroke: association with reperfusion therapies and National Institutes of Health stroke scale score","volume":"4","author":"Yu","year":"2025","journal-title":"Front Stroke"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"e24746","DOI":"10.1016\/j.heliyon.2024.e24746","article-title":"Hemodynamic predictors of early neurological deterioration and clinical outcome after endovascular treatment in large artery occlusion","volume":"10","author":"Xu","year":"2024","journal-title":"Heliyon"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"870","DOI":"10.1161\/STR.0b013e318284056a","article-title":"Guidelines for the early management of patients with acute ischemic stroke","volume":"44","author":"Jauch","year":"2013","journal-title":"Stroke"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"23052","DOI":"10.1038\/s41598-024-73570-x","article-title":"An intelligent learning system based on electronic health Records for Unbiased Stroke Prediction","volume":"14","author":"Saleem","year":"2024","journal-title":"Sci Rep"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"1534","DOI":"10.1093\/jamia\/ocy102","article-title":"The stroke navigator: meaningful use of the electronic health record to efficiently report inpatient stroke care quality","volume":"25","author":"Nathan","year":"2018","journal-title":"J Am Med Inform Assoc"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"2966","DOI":"10.1161\/STROKEAHA.125.051993","article-title":"Accuracy of large language models to identify stroke subtypes within unstructured electronic health record data","volume":"56","author":"Owens","year":"2025","journal-title":"Stroke"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"286","DOI":"10.3390\/jpm10040286","article-title":"Prediction of stroke outcome using natural language processing-based machine learning of radiology report of brain Mri","volume":"10","author":"Heo","year":"2020","journal-title":"J Pers Med"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"e0234908","DOI":"10.1371\/journal.pone.0234908","article-title":"Machine learning and natural language processing methods to identify ischemic stroke, acuity and location from radiology reports","volume":"15","author":"Ong","year":"2020","journal-title":"PLoS One"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"e023486","DOI":"10.1161\/jaha.121.023486","article-title":"Natural language processing enhances prediction of functional outcome after acute ischemic stroke","volume":"10","author":"Sung","year":"2021","journal-title":"J Am Heart Assoc"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"1722965","DOI":"10.3389\/fneur.2025.1722965","article-title":"Ai-driven integration of imaging and radiology language improves stroke mortality prediction","volume":"16","author":"Alotaibi","year":"2025","journal-title":"Front Neurol"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"291","DOI":"10.1007\/s12028-022-01513-3","article-title":"Natural language processing of radiology reports to detect complications of ischemic stroke","volume":"37","author":"Miller","year":"2022","journal-title":"Neurocrit Care"},{"key":"ref21","doi-asserted-by":"publisher","first-page":"197","DOI":"10.1007\/s40263-024-01145-5","article-title":"Adjunctive Pcsk9 inhibitor evolocumab in the prevention of early neurological deterioration in non-cardiogenic acute ischemic stroke: a multicenter, prospective, randomized, open-label, clinical trial","volume":"39","author":"Tian","year":"2025","journal-title":"CNS Drugs"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1613\/jair.953","article-title":"Smote: synthetic minority over-sampling technique","volume":"16","author":"Chawla","year":"2002","journal-title":"J Artif Intell Res"},{"key":"ref23","author":"Feihong Yang","year":"2021"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"61","DOI":"10.1016\/j.jphys.2013.12.012","article-title":"National Institutes of Health stroke scale (Nihss)","volume":"60","author":"Kwah","year":"2014","journal-title":"J Physiother"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"902","DOI":"10.1016\/s0140-6736(20)31961-9","article-title":"Diagnosis of non-consensus transient Ischaemic attacks with focal, negative, and non-progressive symptoms: population-based validation by investigation and prognosis","volume":"397","author":"Tuna","year":"2021","journal-title":"Lancet"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"1416","DOI":"10.5551\/jat.65653","article-title":"Hemostatic activation markers and early neurological deterioration in branch atheromatous disease-related stroke","volume":"32","author":"Hoshino","year":"2025","journal-title":"J Atheroscler Thromb"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"198","DOI":"10.3414\/me14-04-0003","article-title":"Youden's index and the weight of evidence","volume":"54","author":"Hughes","year":"2015","journal-title":"Methods Inf Med"},{"key":"ref28","doi-asserted-by":"publisher","first-page":"1433010","DOI":"10.3389\/fneur.2024.1433010","article-title":"Predictors of early neurological deterioration in patients with acute ischemic stroke","volume":"15","author":"Zhou","year":"2024","journal-title":"Front Neurol"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"e70080","DOI":"10.1111\/ene.70080","article-title":"Associations of inflammatory markers with neurological dysfunction and prognosis in patients with progressive stroke","volume":"32","author":"Wang","year":"2025","journal-title":"Eur J Neurol"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"880","DOI":"10.1186\/s40001-025-03171-5","article-title":"Nomogram for predicting the occurrence of progressive ischemic stroke: a single-center retrospective study","volume":"30","author":"Liu","year":"2025","journal-title":"Eur J Med Res"},{"key":"ref31","doi-asserted-by":"publisher","first-page":"1105496","DOI":"10.3389\/fpsyt.2022.1105496","article-title":"Prognosis of ischemic stroke predicted by machine learning based on multi-modal Mri Radiomics","volume":"13","author":"Yu","year":"2022","journal-title":"Front Psych"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"2136","DOI":"10.1166\/jmihi.2020.3158","article-title":"Prediction of risk factors for progressive and non-progressive stroke by transcranial Doppler ultrasound information technology","volume":"10","author":"Bai","year":"2020","journal-title":"J Med Imaging Health Inform"},{"key":"ref33","doi-asserted-by":"publisher","first-page":"3800","DOI":"10.1007\/s00415-022-11010-4","article-title":"Imaging-based outcome prediction in posterior circulation stroke","volume":"269","author":"Kniep","year":"2022","journal-title":"J Neurol"},{"key":"ref34","doi-asserted-by":"publisher","first-page":"e0212778","DOI":"10.1371\/journal.pone.0212778","article-title":"Natural language processing and machine learning algorithm to identify brain Mri reports with acute ischemic stroke","volume":"14","author":"Kim","year":"2019","journal-title":"PLoS One"}],"container-title":["Frontiers in Neurology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fneur.2026.1787921\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,16]],"date-time":"2026-03-16T05:23:35Z","timestamp":1773638615000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fneur.2026.1787921\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,16]]},"references-count":34,"alternative-id":["10.3389\/fneur.2026.1787921"],"URL":"https:\/\/doi.org\/10.3389\/fneur.2026.1787921","relation":{},"ISSN":["1664-2295"],"issn-type":[{"value":"1664-2295","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,16]]},"article-number":"1787921"}}