Abstract Introduction Electroencephalography (EEG) has shown potential as a prehospital triage instrument for anterior circulation large vessel occlusion (LVO-a) stroke. In this study, we determined the diagnostic accuracy of EEG-based machine learning algorithms for LVO-a stroke detection. Patients and methods Anterior circulation large vessel occlusion stroke detection machine learning models were developed using EEG data from the ELECTRA-STROKE (NCT03699397) and AI-STROKE (NCT05437237) studies. Single 2- to 3-min EEG recordings were obtained using 8 dry electrodes in adults with suspected stroke in prehospital and in-hospital settings. Patients with a clean EEG segment of ≥ 5 s (visually selected) were included and EEG symmetry features were extracted. We assessed the performance of multiple machine learning models using nested cross-validation (CV), optimising models with 3-fold CV (inner loop) and evaluating them with leave-one-out CV (outer loop). Results Of 1293 patients with suspected stroke, 336 (26%) were included (74% prehospital, 26% in-hospital). Median age was 72 (IQR 61–80) years, and 231 (69%) were male. Sixteen (5%) patients had LVO-a stroke, 140 (42%) non-LVO-a ischaemic stroke, 52 (15%) TIA, 13 (4%) haemorrhagic stroke and 115 (34%) a stroke mimic. Naïve Bayes achieved the highest observed area under the receiver operating characteristic curve (0.85 [95% CI, 0.77–0.93]), with a sensitivity of 69% (95% CI, 44%–86%) and specificity of 86% (95% CI, 82%–89%). Discussion and conclusion Machine learning models for LVO-a stroke detection based solely on EEG showed moderately high diagnostic accuracy. These results support the potential of EEG-based LVO-a stroke detection, but improved EEG data acquisition is needed before clinical implementation can be considered.
Abstract Introduction Endovascular thrombectomy (EVT) is established for large infarcts, but remains cautiously applied in extensive infarcts with an Alberta Stroke Program Early CT Score (ASPECTS) of 0–2. Real-life outcome data remain limited. Patients and methods Retrospective analysis of the German Stroke Registry-Endovascular Treatment comparing extensive (ASPECTS 0–2) and large infarcts (ASPECTS 3–5) in patients with anterior-circulation occlusion and EVT. The primary outcome was the ordinal 90-day mRS distribution; secondary outcomes were mRS 0–3, survival, quality of life (3-level European Quality of Life-5 Dimensions) and living condition. Results Of 18,069 screened patients, 889 were analysed (ASPECTS 1–2, n = 89; ASPECTS 3–5, n = 800); none had ASPECTS 0. With ASPECTS 1–2, 20.2% achieved mRS 0–3 and 55.1% died within 90 days. In the model including successful reperfusion, no statistically significant association was detected between infarct group and the 90-day mRS distribution (common odds ratio [cOR] 0.80; 95% CI, 0.53–1.23) or mRS 0–3 (adjusted odds ratio [aOR] 0.89; 95% CI, 0.64–1.25), whereas successful reperfusion was associated with both (cOR 3.68; 95% CI, 2.31–5.85 and aOR 4.10; 95% CI, 2.45–6.86). Among survivors, median European Quality of Life-5 Dimensions utility index was 0.445 vs 0.541 (n = 27/324), and 51.5% vs 50.9% lived at home (n = 33/411). Discussion An extensive infarct carried a challenging prognosis, yet acceptable outcome and non-institutional living remained achievable in a relevant minority. An independent association of infarct group with outcome was not detected. Conclusion An acceptable outcome remained attainable in selected patients with extensive infarction, particularly after successful reperfusion. Lacking a medically managed comparator, these data describe prognosis rather than treatment effect.
Abstract Introduction Brain-derived tau (BD-tau) selectively quantifies tau of central nervous system origin in blood and has demonstrated strong correlations with infarct volume and functional outcome in general ischaemic stroke, but data in lacunar stroke and cerebral small vessel disease (CSVD) are lacking. Patients and methods In a prospectively collected cohort of 101 lacunar stroke patients with an MRI-confirmed recent small subcortical infarct (RSSI) and 47 controls, we measured serum BD-tau using the HD-X SIMOA platform at baseline and 3 and 15 months after stroke. We examined associations of BD-tau with infarct size, chronic CSVD neuroimaging features (white matter hyperintensities [WMHs], lacunes, microbleeds, total SVD score), CSVD progression on follow-up MRI and functional outcome. Results Baseline serum BD-tau was higher in RSSI patients than controls (median 9.54 [7.32–12.20] vs 7.75 [6.18–9.44] pg/mL; P = .003) and correlated with infarct diameter (rs = 0.222, P = 0.039). Brain-derived tau did not correlate with time from stroke symptom onset to blood sampling (rs = 0.150, P = .179). Brain-derived tau was not associated with chronic CSVD neuroimaging features (WMHs, lacunes, microbleeds, SVD score; all P > .05) at any timepoint. Longitudinally, BD-tau returned to control levels by 3 months (P = .374) and remained stable at 15 months (P = .985). Brain-derived tau was neither associated with new vascular MRI lesions on follow-up nor functional neurological outcome. Conclusion Serum BD-tau is elevated acutely in CSVD-related RSSI, correlates with infarct size and normalises by 3 months. Brain-derived tau was not associated with chronic CSVD neuroimaging features or disease progression, suggesting sensitivity to acute neuronal injury rather than chronic or progressive CSVD.
Background Reduced water exchange rate (k(w)) 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 k(w) is associated with indirect MRI-based markers of brain fluid dynamics and whether these markers statistically explain the relationship between k(w) and disease severity. Patients and Methods Seventy genetically confirmed CADASIL patients underwent clinical, cognitive and multimodal MRI assessments. Whole-brain k(w) 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. Results Lower whole-brain k(w) was associated with higher WM FW fraction (beta = -0.419, P-corrected = .012), lower DTI-ALPS index (beta = 0.431, P-corrected = .028) and larger CP volume (beta = -0.510, corrected P < .001). Lower k(w) 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 k(w) and brain volume through WM FW fraction, DTI-ALPS index and CP volume. Conclusions These cross-sectional findings indicate that lower k(w) 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.
Abstract Introduction Dementia is a common and disabling complication after stroke. Exercise-based cardiac rehabilitation (EXCR) may influence vascular and inflammatory mechanisms relevant to cognitive decline. We investigated the association between EXCR initiation after stroke and incident dementia. Patients and methods Using TriNetX, we identified adults with ischaemic stroke or TIA. Patients with recorded ICD-10-CM codes indicating an NIHSS score ≥ 16 were excluded from the primary analysis. Patients initiating EXCR within 6 months were compared with those without documented post-index EXCR. Propensity score matching balanced demographics, comorbidities, medication use and available stroke-severity measures. Outcomes were assessed at 3 and 5 years. Results The unmatched cohorts included 4226 patients initiating EXCR (mean age, 64.7 years; 60.8% male) and 1,889,796 without EXCR (mean age, 66.3 years; 48.2% male). After matching, 4138 patients remained in each cohort. At 5 years, EXCR was associated with lower hazards of incident dementia (HR, 0.70; 95% CI, 0.59–0.82), all-cause mortality (HR, 0.50; 95% CI, 0.45–0.55) and recurrent stroke (HR, 0.75; 95% CI, 0.61–0.91), but a higher hazard of rehospitalisation (HR, 1.21; 95% CI, 1.06–1.38); no significant association was observed for acute myocardial infarction. The association with incident dementia remained consistent across sensitivity analyses. Discussion The observed association with lower dementia risk suggests that the potential benefits of post-stroke EXCR may extend beyond cardiovascular and functional outcomes. Conclusions EXCR initiation after ischaemic stroke or TIA was associated with a lower hazard of incident dementia. This association remained consistent across the 3-year, 30-day landmark, 1-year lagged and all-stroke-severity sensitivity analyses.
