European Stroke Journal最新文献

{"title":"Effects of oral hygiene care on swallowing recovery in hospitalised patients with post-stroke dysphagia: a systematic review and meta-analysis.","authors":"Sungkrit Pojmonpiti, Mariska Meersschaut, Katja Batens, Margot Baudelet, Veerle De Herdt, Barbara Janssens","doi":"10.1093/esj/aakag126","DOIUrl":"10.1093/esj/aakag126","url":null,"abstract":"<p><strong>Objective: </strong>To evaluate the effects of oral hygiene care on swallowing-related outcomes in hospitalised patients with post-stroke dysphagia (PSD).</p><p><strong>Methods: </strong>A systematic literature search was conducted in MEDLINE, Embase, Web of Science, Scopus, CINAHL, CENTRAL and ClinicalTrials.gov up to May 2026. Randomised controlled trials (RCTs) and non-RCTs with a comparison group that reporting swallowing-related outcomes were included. Risk of bias was assessed using RoB 2 for RCTs and ROBINS-I for non-RCTs. Where appropriate, meta-analyses were conducted using random-effects models to estimate mean differences (MDs), odds ratios (ORs) and risk differences (RDs) with 95% confidence intervals (CIs). The certainty of the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework.</p><p><strong>Results: </strong>Nine studies, comprising 7 RCTs and two non-RCTs, were included, with risk-of-bias ratings ranging from low risk to some concerns for RCTs and moderate risk for non-RCTs. Oral hygiene care increased the odds of achieving independent oral intake compared with usual care (OR = 2.54, 95% CI, 1.80; 3.59; I2 = 0%), corresponding to a 13% absolute increase (RD = 0.13, 95% CI, 0.03; 0.23; I2 = 0%). Functional Oral Intake Scale (FOIS; scores 1-7), with higher scores indicating more independent and unrestricted oral intake, favoured oral hygiene care (MD = 0.40, 95% CI, 0.07; 0.72; I2 = 0%). Similarly, Gugging Swallowing Screen (GUSS; scores 0-20), assessing swallowing safety and dysphagia severity, favoured oral hygiene care (MD = 2.93, 95% CI, 1.01; 4.85; I2 = 0%). The certainty of the evidence for achieving independent oral intake was rated as high.</p><p><strong>Conclusion: </strong>Oral hygiene care supports swallowing recovery in patients with PSD, particularly independent oral intake and should be considered an active component of dysphagia rehabilitation.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 10","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13689882/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149246793","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Initial magnetic resonance imaging/magnetic resonance angiography may miss intracranial artery dissection in young adults with stroke: the role of subsequent imaging.","authors":"Naoki Makita, Tomoyuki Ohara, Masahiro Makino, Eijirou Tanaka, Daiki Fukunaga, Naoki Tokuda, Akihiro Fujii, Hidesato Takezawa, Keisuke Imai, Takehiro Yamada, Yoshinari Nagakane, Shiori Ogura, Keiko Maezono-Kandori, Toshiki Mizuno","doi":"10.1093/esj/aakag115","DOIUrl":"10.1093/esj/aakag115","url":null,"abstract":"<p><strong>Introduction: </strong>Intracranial artery dissection (IAD) is an important cause of ischaemic stroke in young adults; however, its diagnosis at initial presentation remains challenging. We aimed to determine how often initial magnetic resonance imaging (MRI)/magnetic resonance angiography (MRA) and subsequent imaging identified IAD, and to clarify the timing of IAD diagnosis.</p><p><strong>Methods: </strong>We performed a secondary analysis of a prospective, multicentre registry of 275 adults aged 18-50 years admitted with acute ischaemic stroke within 14 days of onset. All patients underwent MRI/MRA at admission, followed by scheduled follow-up MRI/MRA 7 days later. Vessel wall MRI (VW-MRI), digital subtraction angiography (DSA) and computed tomography angiography (CTA) were performed when clinically indicated. The proportion of patients in whom each examination identified IAD and the time to diagnosis were assessed.</p><p><strong>Results: </strong>IAD was diagnosed in 53 patients. Initial MRI/MRA identified IAD in only 9 (17%; 95% CI, 8-30). The median time to diagnosis was 5 days after admission; 77% had been diagnosed by Day 7, and the latest diagnosis was made on Day 35. Among patients who underwent each examination, the observed diagnostic yield was 38 of 45 for VW-MRI (84%; 95% CI, 71-94), 41 of 53 for follow-up MRI/MRA (77%; 95% CI, 64-88), 15 of 30 for DSA (50%; 95% CI, 31-69) and 7 of 34 for CTA (21%; 95% CI, 9-38).</p><p><strong>Conclusion: </strong>Among young adults with ischaemic stroke who ultimately met the imaging criteria for IAD, initial MRI/MRA alone was insufficient to establish the diagnosis in most cases. Repeated and multimodal imaging contributed to the diagnosis of IAD.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13588095/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149002682","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Tenecteplase versus standard care for minor ischaemic stroke according to the baseline occlusion status in TEMPO-2 trial.","authors":"Silja Räty, Daniel Strbian, Nishita Singh, Shabnam Vatanpour, Liisa Tomppo, Aravind Ganesh, Bruce C V Campbell, Mark W Parsons, Michael D Hill, Shelagh B Coutts","doi":"10.1093/esj/aakag111","DOIUrl":"10.1093/esj/aakag111","url":null,"abstract":"<p><strong>Introduction: </strong>Patients with minor ischaemic stroke (NIHSS ≤ 5) and intracranial occlusion or focal perfusion abnormality ≤ 12 h of onset did not benefit from intravenous tenecteplase compared to non-thrombolytic standard care (control) in the Tenecteplase vs Standard of Care for Minor Ischaemic Stroke with Proven Occlusion (TEMPO-2) trial. We tested whether tenecteplase was associated with better outcomes in subgroups defined by baseline vessel occlusion status.</p><p><strong>Patients and methods: </strong>In this secondary analysis of the randomised, controlled, blinded TEMPO-2 trial, patients were divided into 3 subgroups: complete occlusion (revised arterial occlusive lesion score 0), near occlusion (score 1) and no visible occlusion (focal perfusion abnormality present). The primary outcome was return to pre-stroke neurological functioning or better, assessed with 90-day mRS. Outcomes were analysed using a generalised linear model adjusted for age, sex, onset-to-randomisation time and baseline stroke severity.</p><p><strong>Results: </strong>Of 881 patients, complete occlusion was present in 218 in the control and 216 in the tenecteplase arm, near occlusion in 102 and 92 and no visible occlusion in 130 and 123, respectively. At baseline, there were longer treatment delays in the control arm with near occlusion, and older age, more hypertension and higher creatinine levels in the tenecteplase arm with complete occlusion. We found no difference in the primary outcome between tenecteplase and control in any subgroup (no visible occlusion: adjusted RR 1.03 [95% CI, 0.90-1.18]; near occlusion: adjusted RR 0.85 [95% CI, 0.71-1.01]; complete occlusion: adjusted RR 1.00 [95% CI, 0.89-1.12]).</p><p><strong>Discussion and conclusion: </strong>There was no difference in outcome between tenecteplase and standard care for patients with mild stroke based on their baseline vessel occlusion status.</p><p><strong>Trial registration: </strong>Clinicaltrials.gov: NCT02398656.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13573578/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148971776","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Response to \"Perhaps Fewer, but Faster: Should we Rethink the 90% Stroke Unit Admission Standard?\"","authors":"Aleš Tomek, Melinda Berg Roaldsen, Bjorn Logi Thórarinsson, Arlene Wilkie, Francesca Romana Pezzella, Hanne Christensen","doi":"10.1093/esj/aakag059","DOIUrl":"10.1093/esj/aakag059","url":null,"abstract":"","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13546758/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148901244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Tracking the changes in longitudinal MRI-detected perivascular spaces following ischaemic stroke.","authors":"William Pham, Mohamed S Khlif, Zhibin Chen, Alexander Jarema, Luke A Henderson, Vaughan G Macefield, Amy Brodtmann","doi":"10.1093/esj/aakag108","DOIUrl":"10.1093/esj/aakag108","url":null,"abstract":"<p><strong>Introduction: </strong>Stroke is a leading cause of mortality and morbidity worldwide. Magnetic resonance imaging-visible perivascular spaces (PVS) are an emerging marker of cerebral small vessel disease and may have prognostic value in stroke. We aimed to characterise longitudinal changes in PVS volume and cluster count following ischaemic stroke.</p><p><strong>Patients and methods: </strong>Perivascular space volumes and cluster counts were compared between stroke survivors (n = 124; 39 women; median [Q1, Q3] age = 70 [62, 76] years) and healthy controls (n = 39; 15 women; median age = 69 [66, 72.5] years). Magnetic resonance imaging scans were acquired at 3 months, 12 months and 36 months post-stroke. Perivascular spaces were automatically segmented from T1-weighted MRI using a validated deep learning algorithm (nnU-Net). Generalised linear mixed-effects models were used to assess group differences and longitudinal changes in PVS, adjusting for baseline age, sex, total intracranial volume and body mass index.</p><p><strong>Results: </strong>At the 3-month timepoint, whole-brain PVS metrics were significantly greater in the stroke group than in controls. By 36 months, stroke patients demonstrated significant PVS reductions relative to controls in the frontal lobe (PVS volume: exp(β) = 0.93; 95% CI, 0.82, 0.99, P = .032; PVS cluster count: exp(β) = 0.91; 95% CI, 0.83, 1; P = .037).</p><p><strong>Discussion: </strong>Contrary to age-related PVS enlargement observed in controls, stroke survivors showed significant longitudinal reductions in PVS volume and cluster counts over 36 months. These findings suggest that MRI-visible PVS trajectories after stroke are dynamic and region-specific, challenging the interpretation of PVS enlargement as a simple marker of cerebrovascular disease burden.</p><p><strong>Conclusion: </strong>Ischaemic stroke is associated with dynamic and regional changes in PVS volume and count.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13547861/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148901289","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MRI markers of brain fluid dynamics link reduced blood-brain barrier water exchange rate to disease severity in CADASIL.","authors":"Yunqing Ying, Rong Zhou, Jianqi Li, Qi Yang, Weijun Tang, Xingfeng Shao, Xin Cheng","doi":"10.1093/esj/aakag114","DOIUrl":"10.1093/esj/aakag114","url":null,"abstract":"<p><strong>Background: </strong>Reduced water exchange rate (kw) across the blood-brain barrier has been linked to disease severity in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), but the mechanisms underlying this association remain unclear. We investigated whether kw is associated with indirect MRI-based markers of brain fluid dynamics and whether these markers statistically explain the relationship between kw and disease severity.</p><p><strong>Patients and methods: </strong>Seventy genetically confirmed CADASIL patients underwent clinical, cognitive and multimodal MRI assessments. Whole-brain kw was measured using diffusion-prepared pseudo-continuous arterial spin labelling. Brain fluid dynamics were assessed using indirect MRI-derived markers, including white matter free water (WM FW) fraction, diffusion tensor imaging along the perivascular space (DTI-ALPS) index, choroid plexus (CP) volume and enlarged perivascular space burden.</p><p><strong>Results: </strong>Lower whole-brain kw was associated with higher WM FW fraction (β = -0.419, P-corrected = .012), lower DTI-ALPS index (β = 0.431, P-corrected = .028) and larger CP volume (β = -0.510, corrected P < .001). Lower kw was also associated with smaller brain volume and slower processing speed. Magnetic resonance imaging markers of brain fluid dynamics were related to lacune count, cerebral microbleed count, WM hyperintensity volume and brain volume. Exploratory mediation analyses identified significant indirect associations between kw and brain volume through WM FW fraction, DTI-ALPS index and CP volume.</p><p><strong>Conclusions: </strong>These cross-sectional findings indicate that lower kw is associated with MRI measures consistent with altered brain fluid dynamics and with greater disease burden in CADASIL. Longitudinal and mechanistic studies are needed to determine the temporal and biological relationships among these measures.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13598894/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149026223","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prehospital triage of patients with suspected acute stroke (PRESTO-2): protocol for a controlled before-and-after intervention trial.","authors":"Ruben M van de Wijdeven, Naomi S de Ruijter, Femke C C Kremers, Esmee Venema, Patrick J van der Geest, Johannes C M IJzermans, Merel Willeboer, Kirsten R I S Dorresteijn, Annemarie D Wijnhoud, Bas van der Veen, Jonathan M Coutinho, Diederik W J Dippel, Henk Kerkhoff, Bob Roozenbeek","doi":"10.1093/esj/aakag106","DOIUrl":"10.1093/esj/aakag106","url":null,"abstract":"<p><strong>Background: </strong>Prehospital triage of suspected stroke patients aims to reduce onset-to-groin time by eliminating the need for interhospital transfers for patients with ischaemic stroke caused by an LVO eligible for EVT, thereby improving long-term functional outcome. We aim to evaluate the implementation of a novel decision support tool, the Stroke Triage App (STA), which provides personalised triage advice by integrating stroke severity, onset time, GPS-based driving times, hospital-specific workflow metrics and time-dependent treatment effects.</p><p><strong>Study design: </strong>PRESTO-2 is a multicentre controlled before-and-after intervention trial. The study aims to evaluate the outcomes before and after implementation of prehospital triage with the STA in 1 acute-care region in the south-west of the Netherlands, compared with a control region. The STA is a smartphone-based application that uses an integrated decision model to estimate the probability of good functional outcome after 3 months (mRS score 0-2) under 2 strategies: transportation to the nearest primary stroke centre or transportation to the more distant intervention centre. Paramedics will use the STA for suspected stroke patients with a positive FAST test of 1 point or more.</p><p><strong>Study endpoints: </strong>The primary outcome is onset-to-groin time. Effects will be estimated using difference-in-differences linear regression analysis. Secondary outcomes include the onset-to-needle time, 3-month functional outcome, proportion of patients treated with IVT and EVT within guideline-recommended time windows, overtriage and undertriage, ambulance utilisation, compliance and adherence to triage advice.</p><p><strong>Summary: </strong>The PRESTO-2 study uses a pragmatic and efficient design for implementation studies, and is expected to provide insights into the optimisation of prehospital stroke care pathways and the challenges associated with integration of novel digital health technologies in existing clinical practice. These findings may inform scientists and clinicians to more efficiently design studies and organise stroke care pathways.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13547856/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148901230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction to: Endovascular Acute Stroke Intervention-Tandem Occlusion (EASI-TOC) study: protocol and rationale.","authors":"","doi":"10.1093/esj/aakag116","DOIUrl":"10.1093/esj/aakag116","url":null,"abstract":"","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13573592/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148971766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Efficacy and safety of minocycline in the treatment of intracerebral haemorrhage: study design and protocol for a multicentre, randomised, open-label, blinded endpoint clinical trial.","authors":"Yang Qu, Zhen-Ni Guo, Shuang Qi, Reziya Abuduxukuer, Peng Zhang, Huimin Li, Xin-Ni Lv, Zeng-Xin Xu, Yi Yang","doi":"10.1093/esj/aakag084","DOIUrl":"10.1093/esj/aakag084","url":null,"abstract":"<p><strong>Background: </strong>Although minocycline is a multi-target neuroprotective drug, its potential benefit in patients with ICH remains unclear. This study primarily aims to assess the efficacy and safety of minocycline for the treatment of ICH.</p><p><strong>Study design: </strong>Efficacy and Safety of Minocycline in the Treatment of ICH (ENHANCE) is an ongoing prospective, multicentre, randomised, open-label, blinded-endpoint clinical trial. A total of 1248 patients with ICH within 24 h of onset will be randomly assigned to the minocycline or control group at a ratio of 1:1. The minocycline group will receive an initial dose of 200 mg minocycline hydrochloride capsules in addition to the guideline-based treatment, followed by 100 mg orally every 12 h for 7 days, resulting in a total of 14 administrations. The control group will receive only the guideline-based treatment.</p><p><strong>Study endpoints: </strong>The primary efficacy outcome measure will be the proportion of patients with good functional outcomes (mRS score, 0-2) at 180 days. Safety outcome measures will include all adverse events and severe adverse events occurring during the study period.</p><p><strong>Summary: </strong>The ENHANCE trial is designed to evaluate the effectiveness and safety of minocycline in the treatment of ICH.</p><p><strong>Trial registration: </strong>Chinese Clinical Trials Registry identifier: ChiCTR2500114607.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13580077/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148983885","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning versus conventional methods for prehospital detection of stroke due to large vessel occlusion or intracranial haemorrhage.","authors":"Bjarty L Garcia, Luuk Dekker, Rene Bekker, Marieke J H Wermer, Ellis S van Etten, Erik W van Zwet, Jan Bosch, Adriaan C G M van Es, Ido R van den Wijngaard, Loet M H Kloos, Leo A M Aerden, Els L L M de Schryver, Henk Kerkhoff, Diederik W J Dippel, Jasper D Daems, Kirsten R I S Dorresteijn, Niels H Chavannes, Bob Roozenbeek, Nyika D Kruyt, For The Lpss And Presto Investigators","doi":"10.1093/esj/aakag112","DOIUrl":"10.1093/esj/aakag112","url":null,"abstract":"<p><strong>Introduction: </strong>Adequate prehospital triage of anterior-circulation LVO (aLVO) or ICH enables direct allocation to appropriate stroke centres. Traditional triage based on clinical scales or logistic regression may miss complex predictor interactions, whereas machine learning approaches may improve diagnostic accuracy. We compared 3 prehospital triage methods to detect a composite outcome of aLVO or ICH using internal-external cross-validation.</p><p><strong>Patients and methods: </strong>We analysed emergency medical services-activated stroke codes (2018-2019) from the Leiden Prehospital Stroke Study and the Prehospital Triage of Patients With Suspected Stroke Study. We compared XGBoost (supervised machine learning [sML]), conventional logistic regression (cLR) and the Rapid Arterial oCclusion Evaluation (RACE) scale using 2-fold internal-external cross-validation. Two variable sets were evaluated: (1) routine prehospital variables (demographics, symptom duration, vital signs, glucose, neurological deficits) and (2) an extended set including medical history, anticoagulant use and point-of-care international normalised ratio. Discrimination and calibration were assessed using balanced accuracy, sensitivity, specificity, Brier scores and calibration plots.</p><p><strong>Results: </strong>Among 3320 patients, 538 (16%) had aLVO or ICH. Using routine variables, mean balanced accuracy was 82.1% for sML, 79.0% for cLR and 74.5% for RACE (P < .01). With extended variables, performance improved to 84.4% for sML and 80.7% for cLR (P < .01). The sML model achieved the highest sensitivity (76.7%) and specificity (92.1%). Calibration was good for both models, although Brier scores were lower for cLR.</p><p><strong>Conclusion: </strong>XGBoost showed modestly higher discriminative performance than cLR or RACE, while cLR showed better calibration. These findings suggest that machine learning may improve prehospital classification in selected settings, but the absolute gains were limited and require further external validation and prospective implementation studies.</p>","PeriodicalId":46821,"journal":{"name":"European Stroke Journal","volume":"11 9","pages":""},"PeriodicalIF":5.2,"publicationDate":"2026-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13639102/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149002661","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}