Stroke: Vascular and Interventional Neurology

Background Endovascular treatment of unruptured intracranial aneurysms (UIAs) often requires dual antiplatelet therapy with aspirin and a P2Y12 inhibitor. Clopidogrel remains widely used, but variable bioactivation and high on-treatment platelet reactivity have prompted interest in alternatives such as prasugrel. We compared thromboembolic and hemorrhagic outcomes after clopidogrel versus prasugrel use in endovascular UIA treatment. Methods We queried the TriNetX Global Collaborative Network for adults with UIA who underwent endovascular treatment between 2016 and 2025 and received aspirin plus clopidogrel or prasugrel. Cohorts were matched 1:1 using propensity score matching based on demographics and comorbidities. Outcomes were assessed at 6 months and 1 year and reported as odds ratios (ORs) with 95% confidence intervals (CIs). Results Among 15,496 adults, 711 matched pairs were included. At 6 months, no significant differences were observed for in-stent thrombosis (OR 0.77, 95% CI 0.38?1.57; p=0.47), cerebral infarction (OR 1.04, 95% CI 0.75?1.46; p=0.80), cerebral hemorrhage (OR 1.07, 95% CI 0.80?1.44; p=0.65), or gastrointestinal hemorrhage (OR 0.84, 95% CI 0.43?1.64; p=0.61). Findings remained nonsignificant at 1 year. Conclusions In this propensity-matched TriNetX analysis, clopidogrel and prasugrel were not associated with significant differences in thromboembolic or hemorrhagic outcomes after endovascular UIA treatment. Prasugrel may be a reasonable alternative P2Y12 inhibitor in selected patients, including those with suspected or documented clopidogrel non-response.

Purpose: The aims of this study were to develop an externally validated deep learning vascular segmentation model for digital enhancement of vertebral artery dissection detection on CTA and to assess clinical utility of the model through a paired crossover reader study. Materials and Methods: This retrospective, IRB-approved study conducted from September 2024 to July 2026 included an internal training cohort of 84 manually segmented CTAs plus 17 CTAs from the RSNA Intracranial Aneurysm Challenge (101 CTAs total) and an external testing cohort of 40 CTAs (22 positive, 18 negative for vertebral artery dissection). A nnU-Net (version 2) model was trained using five-fold cross-validation with the entire training cohort. Technical performance was evaluated using the Dice similarity coefficient (DSC). A two-part crossover reader study included 10 readers who assessed diagnostic accuracy, confidence, and interpretation time with and without segmentation-based augmentation. Statistical analyses included McNemar's exact test (accuracy) and Wilcoxon signed-rank test (confidence, time), with P < .05 considered significant. Results: The model achieved a DSC of 0.96 (SD 0.01). In the reader study, diagnostic accuracy was lower with augmentation than without (0.72 vs 0.84, P < 0.001), while confidence (3.98 vs 3.96, P = .77) and interpretation time (137.3 vs 166.7 seconds, P = .54) did not differ significantly. Subjectively, 7 of 10 readers reported they would use the tool routinely in acute or trauma settings. Conclusion: Although the deep learning model demonstrated accurate segmentation of vascular structures, further validation is needed to establish its utility for enabling accurate diagnosis of VAD in fast-paced clinical settings.

Background: Dexmedetomidine (DEX) and midazolam (MDZ) are commonly used agents for conscious sedation in patients with acute ischemic stroke (AIS) treated with endovascular thrombectomy (EVT). While DEX offers superior sedation quality and potential neuroprotective effects, it is also associated with risk of hemodynamic instability. High-quality evidence directly comparing their impact on functional outcomes after EVT is lacking. This trial aims to determine whether conscious sedation with DEX, compared to MDZ, improves functional outcomes in patients with AIS due to anterior circulation large vessel occlusion (LVO) undergoing EVT. Methods: PEACE (Dexmedetomidine versus Midazolam for Conscious Sedation during Endovascular Stroke Thrombectomy) is a prospective, randomized, open label, blinded-endpoint, multicenter trial. Up to 810 patients with AIS due to LVO in the anterior circulation within 24 hours of last known well will be randomly assigned in a 1:1 ratio to receive either DEX or MDZ for conscious sedation during EVT across 30 centers in China. Results: The primary end point is the proportion of patients achieving a modified Rankin Scale (mRS) score 0-2 at 90 days, analyzed in the intention-to-treat population using a modified Poisson regression model, with results expressed as a risk ratio. Key secondary outcomes include the ordinal distribution of mRS scores at 90 days, and successful recanalization. Safety outcomes include symptomatic intracranial hemorrhage within 48 hours, and mortality at 90 days. Conclusion: The PEACE trial will provide definitive, high-level evidence on the comparative efficacy and safety of DEX versus MDZ for conscious sedation during EVT in patients with AIS due to LVO in the anterior circulation, with the goal of guiding clinical practice.

Background Endovascular thrombectomy (EVT) is used selectively for vertebrobasilar artery occlusion (VBAO), but its net value in mild deficits remains uncertain. We performed a hierarchical win-ratio analysis of EVT versus best medical management (BMM) in patients with acute VBAO and admission National Institutes of Health Stroke Scale (NIHSS) scores of 10 or less. Methods We performed a secondary analysis of a multicenter retrospective registry of acute VBAO with admission NIHSS scores of 10 or less. EVT plus BMM was compared with BMM alone in the overall cohort and in admission NIHSS strata of 0-5 and 6-10. The primary hierarchy ranked death status through common follow-up, 90-day modified Rankin Scale score, and symptomatic intracranial hemorrhage (sICH). All EVT-BMM pairs were compared sequentially and classified as an EVT win, BMM win, or tie. The adjusted analysis used stabilized inverse probability of treatment weighting (IPTW), with weights truncated at the 1st and 99th percentiles in a sensitivity analysis. Results The analysis included 1,232 patients: 428 underwent EVT and 804 received BMM. In the unweighted overall analysis, EVT had 146,674 wins, BMM had 149,958 wins, and 47,480 ties, yielding a WR of 0.98 (95% CI, 0.84-1.15; P = 0.749). After stabilized IPTW, the WR was 0.96 (95% CI, 0.80-1.14; P = 0.618). Sensitivity analyses did not show a robust overall association favoring EVT. In exploratory NIHSS-stratified analyses, the WR was 0.80 (95% CI, 0.61-1.05; P = 0.101) for NIHSS scores of 0-5 and 1.36 (95% CI, 1.10-1.68; P = 0.004) for scores of 6-10. Conclusions EVT was not associated with an overall net advantage over BMM in this hierarchical analysis. The exploratory signal among patients with admission NIHSS scores of 6-10 suggests that EVT may be feasible in selected patients and warrants confirmation in randomized trials; for patients with NIHSS scores of 0-5, EVT use should remain cautious and individualized.

Introduction Mechanical thrombectomy has been well-established as the standard of care treatment for patients with moderate to severe acute ischaemic stroke secondary to large vessel occlusion. However, the clinical value of endovascular treatment (EVT) in patients presenting with minor stroke and low baseline National Institutes of Health Stroke Scale scores remains incompletely defined, with limited high-quality clinical evidence. This protocol outlines the methodology of a randomized trial designed to evaluate whether EVT combined with best medical therapy (BMT) achieves superior functional outcomes relative to BMT alone among patients with minor stroke. Methods and analysis Eligible patients with acute minor anterior circulation ischaemic stroke and confirmed large vessel occlusion (n=264) will be randomized via a computer generated block randomization sequence to receive either EVT plus BMT (intervention arm) or BMT alone (control arm). The primary efficacy outcome is the proportion of participants achieving excellent functional recovery, defined as a modified Rankin Scale score of 0 or 1 at 90 days. Primary safety outcomes comprise symptomatic intracranial haemorrhage within 48 hours after randomization (classified according to Heidelberg criteria) and all cause mortality at 90 days. Participants will be considered eligible if they satisfy all of the following inclusion criteria: (1) enrolment within 24 hours of stroke symptom onset or time last known well; (2) baseline NIHSS score of 2 - 5 accompanied by at least one focal neurological deficit (including altered consciousness, facial palsy, limb weakness, aphasia or hemispatial neglect); (3) angiographically confirmed acute occlusion of the intracranial internal carotid artery, the M1 segment of the middle cerebral artery, or a dominant M2 branch; (4) a hypoperfusion volume (Tmax >6s) [&ge;] 50mL on computed tomography or magnetic resonance perfusion imaging. Both intention to treat and per protocol analyses will be undertaken, and potential sources of heterogeneity in treatment effect magnitude will be explored. Ethics and Dissemination This study was approved by the ethnic committee from the First Affiliated Hospital of Wannan Medical University (2023-IRB90). The findings of this study will be disseminated through publication in peer-reviewed journals. Registration URL:https://www.clinicaltrial.gov; Unique identifier: NCT06143488.

Background Passive advancement--antegrade movement of the aspiration catheter upon inner delivery device withdrawal without active operator force--is a physical phenomenon observed during mechanical thrombectomy, with anecdotal reports of favorable recanalization outcomes. This study characterizes its prevalence and associations with recanalization. Methods Consecutive thrombectomy cases at two high-volume stroke centers (January 2019-December 2025) were retrospectively reviewed. Stored fluoroscopy and angiograms were assessed for passive advancement and final catheter tip position at arterial branch points. Per-pass and per-vessel mixed-effects multivariable regression identified independent associations with eTICI score, first-pass effect (FPE, eTICI [&ge;]2c-3), modified FPE (mFPE, eTICI [&ge;]2b-3), secondary occlusion, and complications. Results Of 1,824 passes in 995 consecutive patients (1,027 occlusions), 733 had imaging adequate for passive advancement assessment. Passive advancement occurred in 525/733 passes (71.6%), including to an arterial branch in 306/733 (41.7%). On per-pass analysis, passive advancement was associated with higher eTICI [&ge;]2b-3 (71.0% vs 19.7%) and eTICI [&ge;]2c-3 (57.5% vs 10.1%); advancement to a branch amplified these associations (90.5% vs 32.1% and 79.4% vs 18.7%; all p<0.001). Among 429 first passes, mFPE was 74.1% vs 10.8% and FPE 63.4% vs 7.5% with vs without passive advancement; rates reached 95.0% and 87.2%, respectively, at a branch point (all p<0.001). Passive advancement to a branch was inversely associated with secondary occlusion (13.3% vs 31.3%; p<0.001). On multivariable analysis, passive advancement independently predicted eTICI score, eTICI [&ge;]2b-3, and eTICI [&ge;]2c-3 (all p<0.001); final catheter position at a branch was the dominant independent predictor of FPE, mFPE, secondary occlusion, any complication, and intracranial complication on per-vessel analysis (all p[&le;]0.046). Conclusions Passive advancement is a prevalent physical phenomenon and an independent predictor of successful recanalization. When advancing to an arterial branch point, it is associated with mFPE >87%. Passive advancement should be a technical objective of aspiration thrombectomy.

Background Accurate prediction of 90-day outcomes following ischemic stroke is crucial for optimizing clinical strategies. This study aimed to systematically review and meta-analyze studies using deep learning (DL) and hand-crafted radiomics (HCR) models for predicting 90-day modified Rankin Scale (mRS) outcomes after ischemic stroke. Methods We followed the PRISMA guidelines for systematic reviews. Four databases (PubMed, Embase, Web of Science, and Cochrane Library) were searched comprehensively for studies published up to September 2025. Inclusion criteria encompassed original research focusing on ischemic stroke patients with mRS at 90 days as the primary outcome and predictive models utilizing imaging biomarkers from CT or MRI. Quality assessment of the included studies was conducted using the PROBAST and and METhodological RadiomICs Score (METRICS) tools. Results Seventeen studies were included in the quantitative meta-analysis. Models demonstrated good prognostic performance for 90-day outcomes, with pooled AUCs of 0.85 (95% CI: 0.80-0.86) for internal and 0.74 (95% CI: 0.67-0.89) for external validation sets, with no significant difference (p=0.42). Deep learning and hand-crafted radiomics models showed comparable diagnostic accuracy in the primary analysis (p=0.667); a difference in sensitivity favouring hand-crafted radiomics emerged only in a post-hoc analysis after exclusion of one outlier study. Models incorporating clinical variables significantly outperformed imaging-only models (p<0.01). CT- and MRI-based models showed similar performance. Treatment-specific analyses revealed higher sensitivity for IVT-based models and higher specificity for combined MT+IVT models. Conclusions DL and HCR models demonstrate robust potential for predicting 90-day outcomes in ischemic stroke. Future research should focus on standardizing model evaluation metrics and incorporating explainable AI techniques to improve clinical applicability.

BackgroundMinor ischemic stroke is heterogeneous, and low NIHSS does not reliably identify patients at risk of deterioration or disability. We investigated whether integrating CT perfusion and angiographic vessel status identifies prognostically distinct phenotypes in a consecutive stroke-code cohort. MethodsWe analyzed consecutive adults with ischemic stroke, baseline NIHSS 0-5, prestroke modified Rankin Scale (mRS) [&le;]2, presentation within 24 hours of symptom onset or last known well, and emergency multiparametric CT. Patients were classified as no perfusion deficit/no vessel occlusion (PF-/VO-), perfusion deficit without visible occlusion (PF+/VO-), or perfusion deficit with vessel occlusion (PF+/VO+). The primary outcome was mRS [&ge;]3 at 90 days; early neurological deterioration (END), defined as an NIHSS increase [&ge;]2 points by 24 hours, was secondary. Multivariable logistic regression adjusted for prespecified covariates and additional variables associated with each outcome in univariable analysis. ResultsOf 2211 patients evaluated through the stroke-code pathway, 534 met inclusion criteria; 391 (73.2%) were PF-/VO-, 61 (11.4%) PF+/VO-, and 82 (15.4%) PF+/VO+. END occurred in 8.1% and mRS [&ge;]3 at 90 days in 17.6%. END occurred in 6.6%, 4.9%, and 17.1% of PF-/VO-, PF+/VO-, and PF+/VO+ patients, respectively (P=0.004), whereas mRS [&ge;]3 occurred in 13.3%, 34.4%, and 25.6%, respectively (P<0.001). Compared with PF-/VO-, PF+/VO+ was independently associated with END (adjusted odds ratio [aOR], 2.50 [95% CI, 1.03-6.06]), whereas PF+/VO- was independently associated with mRS [&ge;]3 at 90 days (aOR, 3.36 [95% CI, 1.67-6.73]). Findings were consistent across definitions of disabling versus nondisabling stroke and after exclusion of lacunar stroke. ConclusionsMultiparametric CT identified distinct risk phenotypes in minor ischemic stroke. Perfusion deficit without visible vessel occlusion was associated with 90-day disability, whereas combined perfusion deficit and vessel occlusion identified patients at increased risk of END. Tissue perfusion and vessel status may provide complementary information for acute risk stratification and future trial enrichment.

Background: Stroke imposes substantial mortality and socioeconomic burden in India, particularly among younger adults. However, prospective evidence on the prognostic value of baseline clinical and radiomics features for mortality prediction remains limited. Objective: To assess mortality at 28, 90, 180, and 365 days after stroke; characterize survival patterns across stroke subtypes and vascular territories; and identify baseline demographic, clinical, and radiomics predictors of mortality. Methods: This prospective cohort study recruited consecutive patients with stroke admitted to a tertiary care hospital in Mangalore, India. Demographic, clinical, laboratory, and radiomics features were collected at admission, and mortality was assessed at 28, 90, 180, and 365 days. Kaplan Meier analysis was used to compare survival across stroke subtypes and vascular territories, while Cox proportional hazards regression was used to identify independent baseline predictors of mortality. Results: Among 63 confirmed stroke cases, 38 (60%) had ischemic stroke (IS) and 25 (40%) had haemorrhagic stroke (HS). Within the middle cerebral artery (MCA) territory, HS patients demonstrated better survival than IS patients due to a greater burden of adverse baseline characteristics such as previous stroke, atrial fibrillation, and higher lesion volume in the IS-MCA subgroup. The combined clinical-radiomics model demonstrated superior prognostic performance than individual models, with higher concordance indices and consistently lower Akaike Information Criterion values. In the combined model, stroke severity (aHR = 2.35, 3.0, 2.93, 2.41) and DWI texture heterogeneity of lesion (aHR = 2.35, 2.11, 2.24, 2.10) were the strongest predictors of mortality across all time points. The CT attenuation heterogeneity (aHR = 1.81, 28 days; 1.76, 90 days; 1.80, 180 days) of lesion and age (aHR = 1.48, 365 days) were also significant predictors. Conclusion: The radiomics features provide critical complementary information to clinical parameters for predicting 28-day mortality, unlike during mid- to long-term follow-up where their incremental contribution appears to be limited. Our results may have important implications in resource-constrained settings for supporting personalized post-stroke management and rehabilitation.

Introduction: Hematoma expansion is a main cause of early worsening after spontaneous intracerebral hemorrhage (ICH). Many hospitals in low income countries have no invasive pressure monitoring and no automated pupillometer. Staff use the Glasgow Coma Scale (GCS) and a penlight to check the pupils, and pupil changes seen with a penlight usually come late. The speed of pupil constriction may change earlier, and smartphones record video fast enough to measure it. Methods and analysis: This will be a single-centre prospective cohort study of 120 adults with ICH on non-contrast computed tomography (NCCT) who arrive within 24 hours of symptom onset. At admission and every 6 hours until 48 hours, we will record a 6 second smartphone video of each eye reacting to a flash, using an eyecup that blocks ward light. Open-source software will measure the pupil diameter in every frame. The main measure is constriction velocity (CV). The primary outcome is hematoma expansion (an increase of more than 6 mL or more than 33%) on NCCT at 24 hours. Secondary outcomes are a fall of 2 or more GCS points within 48 hours and in-hospital death. The primary analysis estimates the area under the receiver operating characteristic curve for the percentage fall in CV from each patient's first recording. Ethics and dissemination: Patients or their legal representatives will give written consent. Results will be submitted to a peer-reviewed journal, and the analysis code will be made public.

Abstract Background: Despite the high stroke burden in the elderly, their thrombolysis utilization is suboptimal partly due to heightened hemorrhagic risk. Observational data suggest that low-dose thrombolysis may offer comparable efficacy to standard-dose with a trend toward reduced symptomatic intracranial hemorrhage (sICH), yet high-level evidence for low-dose thrombolysis in this broad population remains absent. Objective: To evaluate the safety and efficacy of low-dose intravenous tenecteplase thrombolysis in elderly stroke patients within 4.5 hours of ischaemic stroke onset. Methods and design: The DATE-AGING trial is a multicenter, prospective, randomized, open-label, blinded-end point (PROBE) non-inferiority study. Eligible participants are elderly patients (aged [&ge;]70 years)with acute ischemic stroke who meet current guideline recommendations for intravenous thrombolysis within 4.5 hours of symptom onset will be randomly assigned to Low-dose Tenecteplase group (0.175 mg/kg, max 17.5 mg) or Standard-dose Tenecteplase group (0.25 mg/kg, max 25 mg). The primary outcome was excellent functional outcome, defined as a mRS 0-1 assessed at 90-day post-randomization. Safety outcomes included sICH within 36 hours and mortality at 90 days. Discussions: The DATE-AGING trial is intended to provide high-quality evidence for low-dose tenecteplase intravenous thrombolysis in elderly patients with acute ischemic stroke. Trial registry number: NCT07294209(https://clinicaltrials.gov/study/07294209). Keywords: elderly patients, acute ischemic stroke, intravenous thrombolysis, low-dose tenecteplase

Introduction and objectives Hydrocephalus is a common complication of aneurysmal subarachnoid haemorrhage. This is one of several standard indications for the insertion of an external ventricular drain, but there are no standard guidelines for timing of extraventricular drain discontinuation in this population. Therefore, the study aims were to examine the 1) factors associated with extraventricular drain duration in aneurysmal subarachnoid haemorrhage and 2) association between extraventricular drain duration and clinical and resource-based outcomes. Methods In this single centre retrospective cohort study, data for adult patients treated in the ICU between 2015-2023 with aneurysmal subarachnoid haemorrhage and requiring an extraventricular drain were analysed. Data were extracted from both a prospective patient registry and the electronic medical record. Linear regression modelling was used to identify the factors associated with extraventricular drain duration. Measures used to examine the association between extraventricular drain duration, and the outcomes of interest included hazard ratio, relative risk, and mean differences as appropriate for a specific outcome. Results Of the participants (n=205), 68% (n = 140) were female, mean age was 60.7 (SD 13.3) years and 71% (n = 145) had a modified Fisher score of 4. Mean extraventricular drain duration was 15.3 (SD 8.6) days. The number of days mechanically ventilated and extraventricular drain challenge by clamp trial were associated with an increased duration of extraventricular drain treatment. Increased extraventricular drain duration was associated with an increased incidence of ventriculoperitoneal shunt insertion, and an increase in ICU and hospital length of stay when extraventricular drain duration exceeded 20 days. Conclusion The factors associated with an increased duration of extraventricular drain treatment were a longer duration of mechanical ventilation. The outcomes associated with an increased duration of extraventricular drain treatment were an increased incidence of ventriculoperitoneal shunt insertion, an increased incidence of ventriculitis when extraventricular drain duration exceeded 20 days, and an increase in ICU length of stay. These study findings may inform the conduct of a larger multi-centre study to examine the factors associated with duration of treatment with extraventricular drain, and outcomes associated with extraventricular drain duration.

Background In Kenya, stroke is one of the top five causes of mortality, with a high rate of inpatient mortality. The American Heart Association (AHA) and American Stroke Association (ASA) translated scientific evidence into evidence-based clinical practice guidelines aimed at reducing inpatient mortality and improving the quality of care for stroke patients. Despite the proven benefits of the ASA stroke guidelines in improving stroke care, translation of the guidelines in various settings remains a key challenge. This study aimed to quantify the gap between evidence-based recommendations and current stroke care practices in a peri-urban referral facility in Kiambu, Kenya. The study was conducted at xxxxxxxxxx Hospital, a peri-urban referral facility in Kiambu County, Kenya. Methods: We reviewed records of stroke patients attending xxxxxxxxxxx from January 1, 2020, to December 31, 2023. The study evaluated acute stroke management practices and secondary prevention against the 2019 American Heart Association/American Stroke Association (AHA/ASA) Guidelines for the Early Management of Acute Ischemic Stroke. Results: The median distance travelled by stroke patients to the facility was 47.8 kilometres; further, the median onset-to-presentation time was 24 hours. The study evaluated eight clinical practices against the ASA guidelines. The facility lacked a dedicated stroke unit. Remarkably, patients arriving within the therapeutic window for intravenous alteplase and thrombectomy did not receive these treatments. Secondary prevention measures, including discharge on an antiplatelet regimen, were moderately adhered to; among the 86 eligible ischemic stroke patients, 67 (77.9%) were correctly discharged on aspirin. Similarly, 84/123 (68.3%) of the patients were discharged on statin therapy. Conclusion: Evidence-based stroke care recommendations were minimally adhered to at this peri-urban referral facility in Kiambu, Kenya. Contribution: This study quantifies the gap between evidence-based stroke care recommendations and current clinical practice in a peri-urban referral facility in Kenya, providing context-specific evidence to inform quality improvement initiatives and guideline implementation efforts in similar resource-constrained settings.

Background: Immunothrombosis is increasingly recognized in large-vessel-occlusion acute ischemic stroke (LVO-AIS), yet most biomarker studies infer thrombus biology from circulating neutrophil extracellular trap (NET) markers or from pre-defined thrombus subtypes. We prospectively studied 160 patients with anterior-circulation LVO-AIS treated with mechanical thrombectomy to determine whether intrathrombus NETosis markers identify thrombus phenotypes with distinct clinical trajectories, and whether plasma NETosis biomarkers reflect the molecular composition of thrombi. Methods: Thrombi underwent quantitative histopathology and immunohistochemistry for citrullinated histone H3 (CitH3), neutrophil elastase (NE) and CD61. Thrombi were analyzed without prior etiological or histological stratification, and intrathrombus markers were dichotomized at cohort medians (CitH3 6.62%, NE 3.95%, CD61 23.09%) to define internally standardized candidate phenotypes. Associations with thrombus composition, collateral circulation (Tan 2-3), 3-month functional independence (modified Rankin Scale 2 or less) and long-term mortality were assessed by correlation analyses and progressively adjusted multivariable models, with cross-compartment correlations corrected by the Benjamini-Hochberg false discovery rate (FDR). Results: Intrathrombus markers defined prognostically distinct phenotypes: CD61 tracked functional recovery, NE preserved collateral circulation and CitH3 long-term survival, each remaining significant after multivariable adjustment. Circulating NE, myeloperoxidase and leukocyte count instead tracked the macroscopic red-blood-cell-fibrin composition of retrieved thrombi. Cross-compartment concordance was limited: after FDR correction the only significant plasma-thrombus association was an inverse correlation between plasma cell-free DNA and intrathrombus NE, and no intrathrombus marker correlated positively with its circulating counterpart. Conclusions: Local immunothrombotic signatures therefore provide clinically relevant information not captured by plasma markers, and median-defined intrathrombus thresholds represent hypothesis-generating cutoffs for external validation and NET-targeted treatment stratification.

Background: Spontaneous intracerebral hemorrhage (ICH) has high disability and mortality. Accurate early prognostication remains challenging in routine practice. Conventional CT assessment relies mainly on hematoma volume and rough anatomical location, limiting individualized prognosis. Radiomics combined with machine learning may improve prognostic assessment by extracting quantitative texture and heterogeneity features.Methods: In this retrospective multicenter study, we included 2,680 consecutive patients with spontaneous ICH: training (n=1,876), internal validation (n=804), and independent external validation (n=196). Deep learning automatically segmented hematomas, and 1,690 radiomics features were extracted from admission non-contrast CT (NCCT). After reproducibility assessment, univariate screening, and LASSO selection, 42 features were integrated into a Rad-score. The Rad-score was combined with demographic, clinical, and conventional CT variables to construct two complementary machine learning models for functional outcome prediction and comprehensive prognostic stratification. Performance was assessed using R2, mean squared error (MSE), area under the receiver operating characteristic curve (AUC), calibration, decision curve analysis, and external validation. Both models were internally and externally validated. Ambient temperature variation at symptom onset was explored as a potential prognostic modifier. Results: Forty-two reproducible radiomics features constituted the Rad-score. Adding radiomics substantially improved performance over conventional clinical and imaging models. For functional outcome prediction, the integrated model increased R2 from 0.35 to 0.93 and reduced MSE from 387 to 61. For prognostic risk stratification, the clinical-only model achieved an AUC of 0.59, whereas the fully integrated model achieved 0.95 and maintained robust external performance (AUC=0.86). Conclusions: We developed and validated a machine learning-based clinical decision support system integrating conventional CT radiomics and clinical variables for prognostic prediction and risk stratification in ICH. By capturing quantitative imaging information beyond conventional CT markers, this framework improved prognostic discrimination and generalizability. These findings support integrating quantitative imaging biomarkers into clinical decision-making for spontaneous ICH. The association between ambient temperature variation and disease progression warrants prospective investigation. Further prospective validation is warranted.

Background. Prediction models after stroke increasingly add neuroimaging information to bedside data. Whether findings documented in radiology reports improve short term risk stratification beyond a patient's evolving physiological state is unknown. Methods. We studied 4,827 adults with acute ischaemic stroke, intracerebral haemorrhage or subarachnoid haemorrhage in MIMIC IV (2008 to 2019). At eight landmarks from 6 to 48 hours after ICU admission, models ranked patients by their risk of deterioration or death within 24 hours. The baseline model combined demographic and clinical characteristics, a previously validated dynamic ICU state representation, and the physiology, treatment and imaging status at each landmark. A locally run language model extracted imaging phenotypes from 6,992 head CT and MRI reports; three phenotypes and four location categories met a pre specified reliability criterion for automated extraction in independent validation and entered the model. The primary outcome was the difference in the proportion of events captured among the 10% of patients ranked highest at each landmark, in temporal validation (2017 to 2019; 8,923 patient landmarks, 666 events). Results. Among the highest ranked 10%, adding imaging phenotypes captured no additional event (348 of 666 with and without them): a difference of 0.00 percentage points (95% CI -2.20 to +2.01). Discrimination (AUROC 0.855 vs 0.849), calibration and net benefit were similar. No prespecified sensitivity analysis, stroke subtype or availability-adjusted analysis showed statistically supported incremental capture. In a post hoc redundancy analysis, imaging phenotypes improved capture when added to demographic and clinical characteristics alone (+12.91 points, 7.54 to 17.81), but this gain was no longer evident once the dynamic state representation was included. Conclusions. Neuroimaging phenotypes derived from routine radiology reports were associated with short-term deterioration, but we found no evidence that they added risk stratification value beyond the dynamic ICU state and contemporaneous physiology.

Objectives: To determine how strongly clinical referral text, rather than image content, drives the decisions of multimodal large language models (LLMs) differentiating glioma from brain abscess on MRI. Materials and methods: This retrospective single-center study included 60 contrast-enhanced T1-weighted brain MRI examinations (30 gliomas, 30 abscesses; reference-standard diagnosis). One axial slice per case was paired with five referral texts (explicitly inflammatory to explicitly tumor-suggestive) under three prompt instructions and submitted 20 times each to GPT-4o, Claude Sonnet 5, Gemini 3.5 Flash, and Qwen 3.6 (72,000 answers). The primary outcome was the text-driven span (difference in tumor-call frequency between explicit inflammatory and explicit tumor text). Two blinded radiologists classified the images without text; sensitivity, specificity, and accuracy were compared descriptively. Results: Patients (29 women) had a mean age of 55.8 +/- 16.4 years. With neutral text, models answered tumor in 45-67% of queries (specificity, 0.38-0.78), versus 52% and 60% for radiologists (specificity, 0.77 and 0.63). The text-driven span was 0.79-1.00 (p < 0.001 for all models), whereas confirmed gliomas and abscesses differed by only 0.01-0.10 in three models and 0.25 in Gemini 3.5 Flash. Subtle wording produced at least 83% of the span in three models. An instruction to ignore the text reduced the span to 0.48-0.70 in three models and to 0.04 in Gemini 3.5 Flash, partly through reversion to a tumor default. Model accuracy was 0.52-0.73 versus 0.78 and 0.73 for radiologists. Conclusion: Multimodal LLMs deferred to referral text over imaging evidence in differentiating glioma from brain abscess, and prompt instructions did not reliably correct this.

BACKGROUND: Existing models for discharge to long-term institutional care after stroke are developed within single health systems and pool to an area under the curve near 0.80, but none is accompanied by an estimate of how much of the outcome is a property of the health system rather than the patient. METHODS: Secondary analysis of the International Stroke Trial, a randomised trial in 19,435 patients with suspected acute ischaemic stroke recruited in 1991-1996 across 36 countries. Among 14,885 survivors to six months with residence recorded, we developed and internally validated a logistic prognostic model for residential or nursing-home residence using 17 admission-day terms, and compared it against a pre-specified benchmark of age and stroke syndrome alone. Between-country variation was estimated by a method-of-moments decomposition with a cluster bootstrap. RESULTS: Institutional residence was recorded for 1,984 of 14,885 survivors (13.3%); the model was fitted on 14,121 patients containing 1,935 events. In 10-fold cross-validation with the whole procedure refitted per fold, the out-of-fold c-statistic was 0.789, calibration slope 0.987, calibration-in-the-large 0.1371 predicted against 0.1370 observed, and Brier score 0.1013. A benchmark of age and stroke syndrome alone reached an optimism-corrected 0.774 against 0.790 for the full model, a corrected difference of 0.0160. Between-country variation gave an intraclass correlation of 0.0409 (0.0191 to 0.0673), and in leave-one-country-out validation the c-statistic ranged from 0.635 to 0.823 across the 14 countries with enough events to estimate it. CONCLUSIONS: Institutional residence after stroke can be stratified at admission with discrimination comparable to existing models and calibration that holds in unseen patients, but almost all of the signal is carried by age and stroke syndrome. Performance varies substantially between countries, pointing to placement systems as the limit on what a pooled model can do.

Condensed Abstract: Using the 2016?2022 National Inpatient Sample, this study evaluated temporal trends and demographic adoption patterns of percutaneous left atrial appendage occlusion (LAAO). Among 190,020 weighted procedures, annual volume increased from 4,875 to 60,445. In-hospital mortality was exceptionally low (0.1%). Age-adjusted procedural rates rose from 34.19 to 379.87 per 100,000 population. However, significant disparities persisted, characterized by heavy centralization in urban teaching hospitals (87.0%) and disproportionately lower absolute age-adjusted utilization among female, Black, Hispanic, and Asian/Pacific Islander patients compared to male and White cohorts. Targeted initiatives are needed to ensure equitable access to structural heart interventions.

Objective To estimate the mortality proportion among patients with severe traumatic brain injury (sTBI) and to assess variation in mortality according to geographical region, study characteristics, patient populations, and mortality assessment time points. Design Systematic review and meta-analysis. Data sources PubMed/MEDLINE, Embase, Scopus, Web of Science, Google Scholar, and Cochrane Library were searched from database inception to January 2026. Eligibility criteria Studies reporting mortality among patients with severe TBI were eligible. Severe TBI was defined as a Glasgow Coma Scale (GCS) score of [&le;]8, equivalent to GCS <9 or GCS 3-8. Studies involving mixed TBI severity were included only when data for patients with severe TBI could be separately extracted. Data extraction and synthesis Two reviewers independently screened studies, extracted data, and assessed methodological quality. Mortality proportions were pooled using a random-effects model. Statistical heterogeneity was assessed using Cochran's Q and the I2 statistic. Subgroup and sensitivity analyses were undertaken where sufficient data were available. Results Twenty-five studies comprising 12,829 patients with severe TBI were included in the systematic review. Across studies, 5,550 deaths were reported. Mortality proportions ranged from 17.5% to 71.1%. The pooled mortality proportion was 40.6% (95% CI 35.7% to 45.7%), with substantial heterogeneity (I2=96.8%, p<0.001). Mortality varied considerably according to geographical region and timing of outcome assessment, with higher estimates generally observed in African settings and at longer follow-up periods. Conclusion Mortality among patients with severe TBI remains substantial, with approximately two in five patients dying across the included studies. However, the very high heterogeneity indicates that the pooled estimate should not be interpreted as a universal mortality risk. Differences in healthcare resources, patient characteristics, injury severity, management practices, and mortality assessment time points may contribute to the observed variation.