Secular Trends in Outcomes and Impact of Novel Oral Anticoagulants in Atrial Fibrillation-Related Acute Ischemic Stroke.

BACKGROUND: Novel oral anticoagulants (NOACs) are currently recommended for the secondary prevention of stroke in patients with acute ischemic stroke (AIS) accompanied by atrial fibrillation (AF). However, the impact of NOACs on clinical outcomes in real-world practice remains ambiguous. This study analyzes the trend of clinical events in patients with AF-related AIS and determines how much the introduction of NOACs has mediated this trend. METHODS: We identified patients with AIS and AF between January 2011 and December 2019 using a multicenter stroke registry. Annual rates of NOAC prescriptions and clinical events within 1 year were evaluated. The primary outcome was a composite of recurrent stroke, myocardial infarction, and all-cause mortality. To assess the mediation effect of NOACs on the relationship between the calendar year and these outcomes, we used natural effect models and conducted exposure-mediator, exposure-outcome, and mediator-outcome analyses using multivariable regression models or accelerated failure time models, adjusting for potential confounders. RESULTS: Among the 12 977 patients with AF-related AIS, 12 500 (average age: 74.4 years; 51.3% male) were analyzed after excluding cases of valvular AF. Between 2011 and 2019, there was a significant decrease in the 1-year incidence of the primary composite outcome from 28.3% to 21.7%, while the NOAC prescription rate increased from 0% to 75.6%. A 1-year increase in the calendar year was independently associated with delayed occurrence of the primary outcome (adjusted time ratio, 1.10 [95% CI, 1.07-1.14]) and increased NOAC prescription (adjusted odds ratio, 2.20 [95% CI, 2.14-2.27]). Increased NOAC prescription was associated with delayed occurrence of the primary outcome (adjusted time ratio, 3.82 [95% CI, 3.17 to 4.61]). Upon controlling for NOAC prescription (mediator), the calendar year no longer influenced the primary outcome (adjusted time ratio, 0.97 [95% CI, 0.94-1.00]). This suggests that NOAC prescription mediates the association between the calendar year and the primary outcome. CONCLUSIONS: Our study highlights a temporal reduction in major clinical events or death in Korean patients with AF-related AIS, mediated by increased NOAC prescription, emphasizing NOAC use in this population.

Network analysis of stroke systems of care in Korea.

Background: The landscape of stroke care has shifted from stand-alone hospitals to cooperative networks among hospitals. Despite the importance of these networks, limited information exists on their characteristics and functional attributes. Methods: We extracted patient-level data on acute stroke care and hospital connectivity by integrating national stroke audit data with reimbursement claims data. We then used this information to transform interhospital transfers into a network framework, where hospitals were designated as nodes and transfers as edges. Using the Louvain algorithm, we grouped densely connected hospitals into distinct stroke care communities. The quality and characteristics in given stroke communities were analysed, and their distinct types were derived using network parameters. The clinical implications of this network model were also explored. Results: Over 6 months, 19 113 patients with acute ischaemic stroke initially presented to 1009 hospitals, with 3114 (16.3%) transferred to 246 stroke care hospitals. These connected hospitals formed 93 communities, with a median of 9 hospitals treating a median of 201 patients. Derived communities demonstrated a modularity of 0.904, indicating a strong community structure, highly centralised around one or two hubs. Three distinct types of structures were identified: single-hub (n=60), double-hub (n=22) and hubless systems (n=11). The endovascular treatment rate was highest in double-hub systems, followed by single-hub systems, and was almost zero in hubless systems. The hubless communities were characterised by lower patient volumes, fewer hospitals, no hub hospital and no stroke unit. Conclusions: This network analysis could quantify the national stroke care system and point out areas where the organisation and functionality of acute stroke care could be improved.

Automated segmentation of MRI white matter hyperintensities in 8,421 patients with acute ischemic stroke.

BACKGROUND AND PURPOSE: To date, only a few small studies have attempted deep learning-based automatic segmentation of white matter hyperintensity (WMH) lesions in patients with cerebral infarction, which is complicated because stroke-related lesions can obscure WMH borders. We developed and validated deep learning algorithms to segment WMH lesions accurately in patients with cerebral infarction, using multisite datasets involving 8,421 patients with acute ischemic stroke. MATERIALS AND METHODS: We included 8,421 stroke patients from 9 centers in Korea. 2D UNet and SE-Unet models were trained using 2,408 FLAIR MRI from 3 hospitals and validated using 6,013 FLAIR MRIs from 6 hospitals. WMH segmentation performance was assessed by calculating DSC, correlation coefficient, and concordance correlation coefficient compared to a human-segmented gold standard. In addition, we obtained an uncertainty index that represents overall ambiguity in the voxel classification for WMH segmentation in each patient based on the Kullback-Leibler divergence. RESULTS: In the training dataset, the mean age was 67.4±13.0 years and 60.4% were men. The mean (95% CI) DSCs for Unet in internal testing and external validation were respectively 0.659 (0.649-0.669) and 0.710 (0.707-0.714), which were slightly lower than the reliability between humans (DSC=0.744; 95% CI=0.738-0.751; P=.031). Compared with the Unet, the SE-Unet demonstrated better performance, achieving a mean DSC of 0.675 (0.666-0.685; P<.001) in the internal testing and 0.722 (0.719-0.726; P<.001) in the external validation; moreover, it achieved high DSC values (ranging from 0.672 to 0.744) across multiple validation datasets. We observed a significant correlation between WMH volumes that were segmented automatically and manually for the Unet (r=0.917, P<.0001) and even stronger for the SE-Unet (r=0.933, P<.0001). The SE-Unet also attained a high concordance correlation coefficient (ranging from 0.841 to 0.956) in external test datasets. In addition, the uncertainty indices in the majority of patients (86%) in the external datasets were below 0.35, with an average DSC of 0.744 in these patients. CONCLUSIONS: We developed and validated deep learning algorithms to segment WMH in patients with acute cerebral infarction using the largest-ever MRI datasets. In addition, we showed that the uncertainty index can be used to identify cases where automatic WMH segmentation is less accurate and requires human review. ABBREVIATIONS: WMH = white matter hyperintensity; CNN = convolutional neural networks; SE = squeeze-and-excitation; KL = Kullback-Leibler; ReLU = rectified linear unit; LKW = last known well; mRS = modified Rankin Scale; NIHSS = National Institute of Health Stroke Scale; LAA = large artery atherosclerosis; SVO = small vessel occlusion; CE = cardioembolism.