BACKGROUND: Stroke is a leading cause of death and disability worldwide. Rapid and accurate diagnosis is crucial for minimizing brain damage and optimizing treatment plans.
OBJECTIVE: This review aims to summarize the methods of artificial intelligence (AI)-assisted stroke diagnosis over the past 25 years, providing an overview of performance metrics and algorithm development trends. It also delves into existing issues and future prospects, intending to offer a comprehensive reference for clinical practice.
METHODS: A total of 50 representative articles published between 1999 and 2024 on using AI technology for stroke prevention and diagnosis were systematically selected and analyzed in detail.
RESULTS: AI-assisted stroke diagnosis has made significant advances in stroke lesion segmentation and classification, stroke risk prediction, and stroke prognosis. Before 2012, research mainly focused on segmentation using traditional thresholding and heuristic techniques. From 2012 to 2016, the focus shifted to machine learning (ML)-based approaches. After 2016, the emphasis moved to deep learning (DL), which brought significant improvements in accuracy. In stroke lesion segmentation and classification as well as stroke risk prediction, DL has shown superiority over ML. In stroke prognosis, both DL and ML have shown good performance.
CONCLUSIONS: Over the past 25 years, AI technology has shown promising performance in stroke diagnosis.

UNASSIGNED: Disparities in time to hospital presentation and prehospital stroke care may be important drivers in inequities in acute stroke treatment rates, functional outcomes, and mortality. It is unknown how patient-level factors, such as race and ethnicity and county-level socioeconomic status, affect these aspects of prehospital stroke care.
UNASSIGNED: Cross-sectional study of patients with ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage in the Get With the Guidelines-Stroke registry, presenting from July 2015 to December 2019, with symptom onset <24 hours. Multivariable logistic regression and quantile regression were used to investigate the outcomes of interest: emergency medical services (EMS) transport (versus private vehicle), EMS prehospital notification (versus no prehospital notification), and stroke symptom onset to time of arrival at the emergency department. Prespecified covariates included patient-level, hospital-level, and county-level characteristics.
UNASSIGNED: The inclusion criteria was met by the 606 369 patients. Of the patients, 51.2% were men and 69.9% White, with a median National Institutes of Health Stroke Severity of 4 (IQR, 2-10), and median social deprivation index (SDI) of 51 (IQR, 27-75). Median symptom onset to arrival time was 176 minutes (IQR, 64-565). Black race was significantly associated with prolonged symptom onset to emergency department arrival time (+28.21 minutes [95% CI, 25.59-30.84]), and decreased odds of EMS prehospital notification (OR, 0.80 [95% CI, 0.78-0.82]). SDI was not associated with differences in EMS use but was associated with lower odds of EMS prehospital notification (upper SDI tercile versus lowest, OR, 0.79 [95% CI, 0.78-0.81]). SDI was also significantly associated with stroke symptom onset to emergency department arrival time (upper SDI tercile versus lowest +2.56 minutes [95% CI, 0.58-4.53]).
UNASSIGNED: In this national cross-sectional study, Black race was associated with prolonged onset to time of arrival intervals and significantly decreased odds of EMS prehospital notification, despite similar use of EMS transport. Greater county-level deprivation was also associated with reduced odds of EMS prehospital notification and slightly prolonged stroke symptom onset to emergency department arrival time. Efforts to reduce place-based disparities in stroke care must address significant inequities in prehospital care of acute stroke and continue to address health inequities associated with race and ethnicity.

BACKGROUND: In stroke patients with acute large vessel occlusion, endovascular therapy (EVT) may be performed with or without sedation. Our aim is to describe self-reported intraprocedural comfort in patients undergoing EVT depending on sedation type.
METHODS: We performed a prospective observational single-center study of patients undergoing EVT. Patients were systematically interviewed on the day following intervention using a structured questionnaire addressing five domains (nausea/vomiting, pain of any kind, physical discomfort, emotional discomfort, and medical team interaction). Each domain scored 0 to 2 points for a maximum total of 10 points (a higher score indicating greater discomfort). In addition, satisfaction with procedural comfort was rated on a visual analog scale (VAS), and patients reported whether they would have preferred more, less, or the same amount of sedation. Patients who underwent EVT without sedation (local anesthesia, LA) were compared to those who received procedural sedation (conscious sedation, CS).
RESULTS: Seventy-seven questionnaires were completed: 37 (48%) patients underwent EVT with CS while 40 (52%) were treated under LA. Median scores on the self-reported discomfort scale (1[0-2] vs 1[0-2], p = 0.70) and mean scores on VAS (76 ± 25 vs 81 ± 24, p = 0.37) were similar between the CS and the LA group. The proportion of patients who were satisfied with the adopted sedation strategy was similar between groups.
CONCLUSIONS: EVT without prior sedation seems to be well tolerated. Systematic self-evaluation of patient comfort appears feasible and may become integrated into routine clinical care. Patient-oriented outcomes should be included in future trials of sedation during thrombectomy.

BACKGROUND: Left atrial (LA) fibrosis is a marker of atrial cardiomyopathy and has been reported to be associated with both atrial fibrillation and ischemic stroke. Elucidating this relationship is clinically important as LA fibrosis could serve as a surrogate biomarker of LA cardiomyopathy. The objective of this study is to investigate the association of LA fibrosis and embolic stroke of undetermined source (ESUS) using cardiac magnetic resonance imaging.
RESULTS: Following an International Prospective Register of Systematic Reviews-registered protocol, 3 blinded reviewers performed a systematic review for studies that quantified the degree of LA fibrosis in patients with ESUS as compared with healthy patients from inception to February 2024. A meta-analysis was conducted in the mean difference. From 7 studies (705 patients), there was a significantly higher degree of LA fibrosis in patients with ESUS compared with healthy controls (MD, 5.71% [95% CI, 3.55%-7.87%], P<0.01). The degree of LA fibrosis was significantly higher in patients with atrial fibrillation than healthy controls (MD, 8.22% [95% CI, 5.62%-10.83%], P<0.01). A similar degree of LA fibrosis was observed in patients with ESUS compared with patients with atrial fibrillation (MD, -0.92% [95% CI, -2.29% to 0.44%], P=0.35).
CONCLUSIONS: A significantly higher degree of LA fibrosis was found in patients with ESUS as compared with healthy controls. This suggests that LA fibrosis may play a significant role in the pathogenesis of ESUS. Further research is warranted to investigate LA fibrosis as a surrogate biomarker of atrial cardiomyopathy and recurrent stroke risk in patients with ESUS.

BACKGROUND: Differences of treatment outcome between full or reduced dose of tissue plasminogen activator (tPA) for bridge mechanical thrombectomy (MT) in the extended time window have not been clearly established. We aimed to present real-world results of bridge MT with different tPA dosages in the standard and extended windows.
METHODS: Patients with anterior circulation stroke treated with MT between 2017 and 2021 at two stroke referral centers were retrospectively reviewed. Bridge MT with tPA were categorized as full (0.9 mg/kg) or reduced (<0.9 mg/kg) dose. Standard window (SW) cohort was defined as MT performed within 6 h of acute ischemic stroke onset, while those beyond 6 h as the extended window (EW) cohort. 90 days Modified Rankin Scale (mRS) score, technical treatment success, in-hospital mortality, and post-treatment hemorrhage were analyzed.
RESULTS: A total of 423 patients met the inclusion criteria, 218 of which treated in the SW, while 205 treated in the EW. Within the SW cohort, the full-dose tPA group demonstrated a higher proportion of good functional outcome (GFO) at 90 days (mRS0-3) versus reduced (49% vs 21%, p = 0.0358). The overall GFO of SW was higher than that of the EW cohort (33% vs 20%, p = 0.0480). Within the EW cohort, GFO was similar between full and reduced dose groups. Successful reperfusion rate was lower in SW versus EW cohorts (39% vs 58%, p = 0.0199).
CONCLUSIONS: In real-world practice, the GFO of bridge MT is better than MT alone. The tPA dosage is not a determining factor of GFO in EW MT.

Delirium is a common complication in acute stroke patients, occurring in 15-35% of all stroke unit admissions and is associated with prolonged hospital stay and a poor post-stroke prognosis. Managing delirium in acute stroke patients necessitates an intensive and multiprofessional therapeutic approach, placing a significant burden on healthcare staff. However, dedicated practical recommendations for delirium management developed for the population of acute stroke patients are lacking. For this purpose, the Austrian Stroke Society, in cooperation with the Austrian Society of Neurology, the Austrian Society of Neurorehabilitation, and the Austrian Society of Psychiatry, Psychotherapy, and Psychosomatics has formulated an evidence-based position paper addressing the management of delirium in acute stroke patients. The paper outlines practical recommendations on the three pillars of care in stroke patients with delirium: (a) Key aspects of delirium prevention including stroke-specific delirium risk factors and delirium prediction scores are described. Moreover, a non-pharmacological delirium prevention bundle is presented. (b) The paper provides recommendations on timing and frequency of delirium screening to ensure early diagnosis of delirium in acute stroke patients. Moreover, it reports on the use of different delirium screening tools in stroke populations. (c) An overview of non-pharmacological and pharmacological treatment strategies in patients with delirium and acute stroke is presented and summarized as key recommendation statements.

OBJECTIVE: Heart failure may result in reduced brain perfusion, limiting the blood flow needed to achieve clinical recovery. We investigated whether plasma levels of brain natriuretic peptide (BNP), a biological marker of heart failure, were related to clinical outcomes after mechanical thrombectomy (MT).
METHODS: Data were analyzed from stroke patients with internal carotid or middle cerebral artery occlusion enrolled in the SKIP trial for whom plasma level of BNP was evaluated on admission. Favorable outcome was defined as a modified Rankin scale score of 0-2 at 3 months.
RESULTS: Among 169 patients (median age, 74 years; 62% men, median National Institutes of Health Stroke Scale score, 18), 104 (62%) achieved favorable outcomes. Median plasma BNP level was lower in the favorable outcome group (124.1 pg/mL; interquartile range [IQR], 62.1-215.5 pg/mL) than in the unfavorable outcome group (198.0 pg/mL; IQR, 74.8-334.0 pg/mL; p=0.005). In multivariate regression analysis, the adjusted odds ratio for BNP for favorable outcomes was 0.971 (95% confidence interval, 0.993-0.999; p=0.048). At 3 months after onset, the favorable outcome rate was lower in the ≥186 pg/mL group (45%) than in the <186 pg/mL group (72%; p=0.001). This significant difference remained regardless of the presence of atrial fibrillation (AF), with rates of 47% and 76%, respectively, in AF patients (p=0.003) and 33% and 68%, respectively, in patients without AF (p=0.046).
CONCLUSIONS: High plasma BNP concentration appears associated with unfavorable outcomes after MT.

Functional electrical stimulation (FES) can support functional restoration of a paretic limb post-stroke. Hebbian plasticity depends on temporally coinciding pre- and post-synaptic activity. A tight temporal relationship between motor cortical (MC) activity associated with attempted movement and FES-generated visuo-proprioceptive feedback is hypothesized to enhance motor recovery. Using a brain-computer interface (BCI) to classify MC spectral power in electroencephalographic (EEG) signals to trigger FES-delivery with detection of movement attempts improved motor outcomes in chronic stroke patients. We hypothesized that heightened neural plasticity earlier post-stroke would further enhance corticomuscular functional connectivity and motor recovery. We compared subcortical non-dominant hemisphere stroke patients in BCI-FES and Random-FES (FES temporally independent of MC movement attempt detection) groups. The primary outcome measure was the Fugl-Meyer Assessment, Upper Extremity (FMA-UE). We recorded high-density EEG and transcranial magnetic stimulation-induced motor evoked potentials before and after treatment. The BCI group showed greater: FMA-UE improvement; motor evoked potential amplitude; beta oscillatory power and long-range temporal correlation reduction over contralateral MC; and corticomuscular coherence with contralateral MC. These changes are consistent with enhanced post-stroke motor improvement when movement is synchronized with MC activity reflecting attempted movement.

UNASSIGNED: To evaluate performance of synthetic and real FLAIR for identifying early stroke in a multicenter cohort.
UNASSIGNED: This retrospective study was conducted using DWI and FLAIR extracted from the Endovascular Treatment in Ischemic Stroke image registry (2017-2021). The database was partitioned into subsets according to MRI field strength and manufacturer, and randomly divided into training set (70%) used for model fine-tuning, validation set (15%), and test set (15%). In test set, five readers, blinded to FLAIR sequence type, assessed DWI-FLAIR mismatch using real and synthetic FLAIR. Interobserver agreement for DWI-FLAIR rating and concordance between synthetic and real FLAIR were evaluated with kappa statistics. Sensitivity and specificity for identification of ⩽4.5 h AIS were compared in patients with known onset-to-MRI delay using McNemar\'s test.
UNASSIGNED: 1454 complete MRI sets (1172 patients, median (IQR) age: 73 years (62-82); 762 women) acquired on 125 MRI units were analyzed. In test set (207 MRI), interobserver reproducibility for DWI-FLAIR mismatch labeling was substantial for real and synthetic FLAIR (Fleiss κ = 0.79 (95%CI: 0.73-0.84) and 0.77 (95%CI: 0.71-0.82), respectively). After consensus, concordance between real and synthetic FLAIR was excellent (κ = 0.85 (95%CI: 0.78-0.92)). In 141 MRI sets with known onset-to-MRI delay, diagnostic performances for ⩽4.5 h AIS identification did not differ between real and synthetic FLAIR (sensitivity: 60/71 (85%) vs 59/71 (83%), p = .56; specificity: 65/70 (93%) vs 65/70 (93%), p > 0.99).
UNASSIGNED: A deep-learning-based FLAIR fine-tuned on multicenter data can provide comparable performances to real FLAIR for early AIS identification. This approach may help reducing MR protocol duration and motion artifacts.

BACKGROUND: To better target stroke awareness efforts (pre and post first stroke) and thereby decrease the time window for help-seeking, this study aims to assess quantitatively whether stroke awareness is associated with appropriate help-seeking at symptom onset, and to investigate qualitatively why this may (not) be the case.
METHODS: This study conducted in a German regional stroke network comprises a convergent quantitative-dominant, hypothesis-driven mixed methods design including 462 quantitative patient questionnaires combined with qualitative interviews with 28 patients and seven relatives. Quantitative associations were identified using Pearson\'s correlation analysis. Open coding was performed on interview transcripts before the quantitative results were used to further focus qualitative analysis. Joint display analysis was conducted to mix data strands. Cooperation with the Patient Council of the Department of Neurology ensured patient involvement in the study.
RESULTS: Our hypothesis that stroke awareness would be associated with appropriate help-seeking behaviour at stroke symptom onset was partially supported by the quantitative data, i.e. showing associations between some dimensions of stroke awareness and appropriate help-seeking, but not others. For example, knowing stroke symptoms is correlated with recognising one\'s own symptoms as stroke (r = 0.101; p = 0.030*; N = 459) but not with no hesitation before calling help (r = 0.003; p = 0.941; N = 457). A previous stroke also makes it more likely to recognise one\'s own symptoms as stroke (r = 0.114; p = 0.015*; N = 459), but not to be transported by emergency ambulance (r = 0.08; p = 0.872; N = 462) or to arrive at the hospital on time (r = 0.02; p = 0.677; N = 459). Qualitative results showed concordance, discordance or provided potential explanations for quantitative findings. For example, qualitative data showed processes of denial on the part of patients and the important role of relatives in initiating appropriate help-seeking behaviour on patients\' behalf.
CONCLUSIONS: Our study provides insights into the complexities of the decision-making process at stroke symptom onset. As our findings suggest processes of denial and inabilities to translate abstract disease knowledge into correct actions, we recommend to address relatives as potential saviours of loved ones, increased use of specific situational examples (e.g. lying on the bathroom floor) and the involvement of patient representatives in the preparation of informational resources and campaigns. Future research should include mixed methods research from one sample and more attention to potential reporting inconsistencies.