Predictor Of Stroke Outcome Research Articles

Benchmarking Stroke Outcome Prediction through Comprehensive Data Analysis - NeuralCup 2023.

Stroke is a significant cause of mortality and long-term disability worldwide, with variable recovery trajectories posing substantial challenges in anticipating post-event care and rehabilitation planning. The NeuralCup 2023 consortium was established to address these challenges by comparing the predictability of stroke outcome models through a collaborative, data-driven approach. This study presents the consortium's findings, which involved 15 participating teams worldwide. Using a comprehensive dataset, which included clinical and imaging data, we conducted an open competition to identify and compare predictors of motor, cognitive, and neuropsychological (emotional) outcomes one-year post-stroke. Analyses incorporated both traditional and novel methods, including machine learning algorithms. These efforts culminated in the search for 'optimal recipes' for predicting each domain through an exhaustive exploration of the features of all the approaches. Key predictors included lesion characteristics, T1-weighted MRI sequences, and demographic factors. Notably, integrating FLAIR imaging and white matter tract analysis emerged as crucial to improving the accuracy of cognitive and motor outcome predictions, respectively. These findings advocate for a tailored, multifaceted approach to stroke outcome prediction, underscoring the potential of collaborative data science in addressing complex neurological prognostication challenges. This study also sets a new benchmark methodology in stroke research, offering a foundational step toward personalized care strategies that could significantly impact recovery planning and quality of life for stroke survivors.

The relationship of hematologic parameters and lipid profile with acute ischemic stroke

Objective: This study aims to investigate the effect of routine laboratory parameters on clinical outcomes in patients with acute ischemic stroke (AIS). Materials and Methods: Our study was designed as retrospective and cross-sectional. 94 patients with AIS who received inpatient treatment in our clinic with a diagnosis of AIS were included in the study. Laboratory data, demographic, and clinical characteristics were recorded at the time of admission. Neurological disabilities of the patients 3 months after treatment were evaluated with the modified Rankin Scale (mRS). Patients were divided into two groups (good prognosis (mRS 0-2) and poor prognosis (mRS 3-6)). Results:Of the 94 patients included in the study, 47 were men and 47 were women. The mean age was 71.18±11.92 years. Laboratory parameters hemoglobin, eosinophil, triglyceride, total-cholesterol, and ldl-cholesterol values were significantly lower in the group with mRS≥3. As a result of ROC analysis, the area under the curve of hemoglobin, eosinophil, total-cholesterol, ldl-cholesterol, and triglyceride was significantly higher[(AUC: 0.710 95%CI: 0.60-0.81); Sensitivity=67.6%, Specificity=67.9%, p=0.001]. In the multivariable logistic regression model, only the NIHSS score (National Institutes Of Health Stroke Scale Scores) was independently associated with the degree of disability (p=0.001). NIHSS is an independent factor in predicting stroke outcomes. Conclusion: It was determined that there was a significant inverse relationship between the routine laboratory parameters hemoglobin level, eosinophil level, LDL and total cholesterol levels, and infarct volume with the NIHSS scores of hemoglobin and LDL-cholesterol.

Improving Stroke Outcome Prediction Using Molecular and Machine Learning Approaches in Large Vessel Occlusion.

Introduction: Predicting stroke outcomes in acute ischemic stroke (AIS) can be challenging, especially for patients with large vessel occlusion (LVO). Available tools such as infarct volume and the National Institute of Health Stroke Scale (NIHSS) have shown limited accuracy in predicting outcomes for this specific patient population. The present study aimed to confirm whether sudden metabolic changes due to blood-brain barrier (BBB) disruption during LVO reflect differences in circulating metabolites and RNA between small and large core strokes. The second objective was to evaluate whether integrating molecular markers with existing neurological and imaging tools can enhance outcome predictions in LVO strokes. Methods: The infarction volume in patients was measured using magnetic resonance diffusion-weighted images, and the 90-day stroke outcome was defined by a modified Rankin Scale (mRS). Differential expression patterns of miRNAs were identified by RNA sequencing of serum-driven exosomes. Nuclear magnetic resonance (NMR) spectroscopy was used to identify metabolites associated with AIS with small and large infarctions. Results: We identified 41 miRNAs and 11 metabolites to be significantly associated with infarct volume in a multivariate regression analysis after adjusting for the confounders. Eight miRNAs and ketone bodies correlated significantly with infarct volume, NIHSS (severity), and mRS (outcome). Through integrative analysis of clinical, radiological, and omics data using machine learning, our study identified 11 top features for predicting stroke outcomes with an accuracy of 0.81 and AUC of 0.91. Conclusions: Our study provides a future framework for advancing stroke therapeutics by incorporating molecular markers into the existing neurological and imaging tools to improve predictive efficacy and enhance patient outcomes.

A Review on Stroke’s Pathophysiology, Associated Risk Factors, and Management

Background: Stroke is a medical disorder defined by focal neurological deficits caused by vascular impairment (hemorrhage or infarction) in the CNS (Central Nervous System). Currently the second biggest cause of morbidity worldwide, strokes continue to be a deadly disease. Objective: This review aimed to elucidate the stroke’s pathophysiology, both changeable and unchangeable risk factors associated with stroke, including prevention of stroke, its diagnosis, as well as treatment. Methods: The articles were collected involving the research on stroke with the help of different sources, like PubMed, Science Direct, and Google. Many manuscripts were collected and analyzed, and the relevant ones were considered for the present manuscript. Results: A brain CT scan has been found to be the first step in the management of Cerebrovascular Accident (CVA) in order to distinguish between ischemic and hemorrhagic stroke types. Beyond stroke classification, imaging has been observed to serve as a powerful tool for ruling out alternative diagnoses, predicting stroke outcomes, and facilitating appropriate treatment planning. Conclusion: Stroke is the second foremost cause of death as well as disability worldwide, leading to substantial economic losses. Therefore, better treatment and improved management after stroke are a global health priority. Modifying risk factors can decrease the chances of occurrence of stroke. Also, correct diagnosis should be done to diagnose the presence of stroke and its type. Proper treatment should be provided, like perfusion, glucose management, antiplatelet therapy, antihypertensive therapy, stem cell therapy, neural response and rehabilitation, increased awareness, lifestyle changes, stress management, etc. Exercise can also help positively to reduce the chances of stroke.

From data to decisions: AI and functional connectivity for diagnosis, prognosis, and recovery prediction in stroke.

Stroke is a severe medical condition which may lead to permanent disability conditions. The initial 8weeks following a stroke are crucial for rehabilitation, as most recovery occurs during this period. Personalized approaches and predictive biomarkers are needed for tailored rehabilitation. In this context, EEG brain connectivity and Artificial Intelligence (AI) can play a crucial role in diagnosing and predicting stroke outcomes efficiently. In the present study, 127 patients with subacute ischemic lesions and 90 age- and gender-matched healthy controls were enrolled. EEG recordings were obtained from each participant within 15days of stroke onset. Clinical evaluations were performed at baseline and at 40-days follow-up using the National Institutes of Health Stroke Scale (NIHSS). Functional connectivity analysis was conducted using Total Coherence (TotCoh) and Small Word (SW). Quadratic support vector machines (SVM) algorithms were implemented to classify healthy subjects compared to stroke patients (Healthy vs Stroke), determine the affected hemisphere (Left vs Right Hemisphere), and predict functional recovery (Functional Recovery Prediction). In the classification for Functional Recovery Prediction, an accuracy of 94.75%, sensitivity of 96.27% specificity of 92.33%, and AUC of 0.95 were achieved; for Healthy vs Stroke, an accuracy of 99.09%, sensitivity of 100%, specificity of 98.46%, and AUC of 0.99 were achieved. For Left vs Right Hemisphere classification, accuracy was 86.77%, sensitivity was 91.44%, specificity was 80.33%, and AUC was 0.87. These findings highlight the potential of utilizing functional connectivity measures based on EEG in combination with AI algorithms to improve patient outcomes by targeted rehabilitation interventions.

Oral condition at admission predicts functional outcomes and hospital-acquired pneumonia development among acute ischemic stroke patients

IntroductionOral care is crucial for the prevention of cardiovascular events and pneumonia. However, few studies have evaluated the associations between multidimensional assessments of oral status or functional outcomes and hospital-acquired pneumonia (HAP).MethodsConsecutive patients with acute ischemic stroke (AIS) were retrospectively analyzed. We evaluated the modified oral assessment grade (mOAG) and investigated its association with a modified Rankin scale (mRS) score of 0‒2 (good stroke outcome) and HAP. The mOAG was developed to evaluate 8 categories (lip, tongue, coated tongue, saliva, mucosa, gingiva, preservation, and gargling) on a 4-point scale ranging from 0 to 3. We analyzed the effectiveness of the mOAG score for predicting stroke outcome or HAP using receiver operating characteristic (ROC) curve analysis.ResultsIn total, 247 patients with AIS were analyzed. The area under the ROC curve of the mOAG for predicting poor outcomes was 0.821 (cutoff value: 7), and that for HAP incidence was 0.783 (cutoff value: 8). mOAG (a one-point increase) was associated with poor stroke outcome (odds ratio [OR] 1.31, 95% confidence interval [CI] 1.17‒1.48, P < 0.001) and HAP (OR 1.21, 95% CI 1.07‒1.38, P = 0.003) after adjusting for baseline clinical characteristics, including age and stroke severity.ConclusionsLower mOAG scores at admission were independently associated with good outcomes and a decreased incidence of HAP. Comprehensive oral assessments are essential for acute stroke patients in clinical settings.

Machine Learning Approaches for Stroke Risk Prediction: Findings from the Suita Study.

Stroke constitutes a significant public health concern due to its impact on mortality and morbidity. This study investigates the utility of machine learning algorithms in predicting stroke and identifying key risk factors using data from the Suita study, comprising 7389 participants and 53 variables. Initially, unsupervised k-prototype clustering categorized participants into risk clusters, while five supervised models including Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosted Machine (LightGBM) were employed to predict stroke outcomes. Stroke incidence disparities among identified risk clusters using the unsupervised k-prototype clustering method are substantial, according to the findings. Supervised learning, particularly RF, was a preferable option because of the higher levels of performance metrics. The Shapley Additive Explanations (SHAP) method identified age, systolic blood pressure, hypertension, estimated glomerular filtration rate, metabolic syndrome, and blood glucose level as key predictors of stroke, aligning with findings from the unsupervised clustering approach in high-risk groups. Additionally, previously unidentified risk factors such as elbow joint thickness, fructosamine, hemoglobin, and calcium level demonstrate potential for stroke prediction. In conclusion, machine learning facilitated accurate stroke risk predictions and highlighted potential biomarkers, offering a data-driven framework for risk assessment and biomarker discovery.

Neuroimaging biomarkers for predicting stroke outcomes: A systematic review.

Stroke is a prominent cause of long-term adult impairment globally and a significant global health issue. Only 14% of stroke survivors achieve full recovery, while 25% to 50% require varying degrees of support, and over half become dependent. The aftermath of a stroke brings profound changes to an individual's life, with early choices significantly impacting their quality of life. This review aims to establish the efficacy of neuroimaging data in predicting long-term outcomes and recovery rates following a stroke. A scientific literature search was conducted using the Centre of Reviews and Dissemination (CRD) criteria and PRISMA guidelines for a combined meta-narrative and systematic quantitative review. The methodology involved a structured search in databases like PubMed and The Cochrane Library, following inclusion and exclusion criteria to identify relevant studies on neuroimaging biomarkers for stroke outcome prediction. Data collection utilized the Microsoft Edge Zotero plugin, with quality appraisal conducted via the CASP checklist. Studies published from 2010 to 2024, including observational, randomized control trials, case reports, and clinical trials. Non-English and incomplete studies were excluded, resulting in the identification of 11 pertinent articles. Data extraction emphasized study methodologies, stroke conditions, clinical parameters, and biomarkers, aiming to provide a thorough literature overview and evaluate the significance of neuroimaging biomarkers in predicting stroke recovery outcomes. The results of this systematic review indicate that integrating advanced neuroimaging methods with highly successful reperfusion therapies following a stroke facilitates the diagnosis of the condition and assists in improving neurological impairments resulting from stroke. These measures reduce the possibility of death and improve the treatment provided to stroke patients. These findings highlight the crucial role of neuroimaging in advancing our understanding of post-stroke outcomes and improving patient care.

Ischemic stroke outcome prediction with diversity features from whole brain tissue using deep learning network.

This study proposed an outcome prediction method to improve the accuracy and efficacy of ischemic stroke outcome prediction based on the diversity of whole brain features, without using basic information about patients and image features in lesions. In this study, we directly extracted dynamic radiomics features (DRFs) from dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI) and further extracted static radiomics features (SRFs) and static encoding features (SEFs) from the minimum intensity projection (MinIP) map, which was generated from the time dimension of DSC-PWI images. After selecting whole brain features Ffuse from the combinations of DRFs, SRFs, and SEFs by the Lasso algorithm, various machine and deep learning models were used to evaluate the role of Ffuse in predicting stroke outcomes. The experimental results show that the feature Ffuse generated from DRFs, SRFs, and SEFs (Resnet 18) outperformed other single and combination features and achieved the best mean score of 0.971 both on machine learning models and deep learning models and the 95% CI were (0.703, 0.877) and (0.92, 0.983), respectively. Besides, the deep learning models generally performed better than the machine learning models. The method used in our study can achieve an accurate assessment of stroke outcomes without segmentation of ischemic lesions, which is of great significance for rapid, efficient, and accurate clinical stroke treatment.

A Manta-Ray Hill Climbing Vision Transformer Model for Predicting Ischemic Stroke Outcome

An ischemic stroke attack can cause permanent damage to healthy brain tissue, leading to a permanent loss of motor or sensory function. It can also result in disability or death if not diagnosed and treated promptly. Early prediction of the outcome of the first stroke, such as disability or death, can help many patients by administering appropriate medications to save their lives. Additionally, early prediction of a recurrent stroke within 14 days of the initial stroke can contribute to prevent its recurrence. This paper first proposes a modified Manta-Ray Foraging Optimizer (MMRFO) to enhance the characteristics of the MRFO technique. This approach is based on incorporating the Hill Climbing methodology into the original MRFO in order to improve the exploitation phase, which is responsible for locating the promising zone in the search area. The proposed approach is then utilized to determine the appropriate hyperparameters of the Vision Transformer(ViT) model to predict stroke outcomes prior to its occurrence. To transform categorical data to numerical values, an ASCII encoder module is included. In the feature selection step, the Harris Hawk Optimization approach (HHO) is used to identify the most important elements that may define the stroke. A comparative study has been performed to confirm the effectiveness of the proposed methodology. The results demonstrate that the proposed technique with a Vision Transformer achieves superior results compared to state-of-the-art algorithms. The accuracy of the proposed technique was improved to 87% for the first dataset and 83% for the second, which is clearly superior to that of the other models and earlier research.

Point of View on Outcome Prediction Models in Post-Stroke Motor Recovery

Stroke is a leading cause of disability worldwide which can cause significant and persistent upper limb (UL) impairment. It is difficult to predict UL motor recovery after stroke and to forecast the expected outcomes of rehabilitation interventions during the acute and subacute phases when using clinical data alone. Accurate prediction of response to treatment could allow for more timely and targeted interventions, thereby improving recovery, resource allocation, and reducing the economic impact of post-stroke disability. Initial motor impairment is currently the strongest predictor of post-stroke motor recovery. Despite significant progress, current prediction models could be refined with additional predictors, and an emphasis on the time dependency of patient-specific predictions of UL recovery profiles. In the current paper a panel of experts provide their opinion on additional predictors and aspects of the literature that can help advance stroke outcome prediction models. Potential strategies include close attention to post-stroke data collection timeframes and adoption of individual-computerized modeling methods connected to a patient’s health record. These models should account for the non-linear and the variable recovery pattern of spontaneous neurological recovery. Additionally, input data should be extended to include cognitive, genomic, sensory, neural injury, and function measures as additional predictors of recovery. The accuracy of prediction models may be further improved by including standardized measures of outcome. Finally, we consider the potential impact of refined prediction models on healthcare costs.

ASSOCIATION OF THE TYPE OF INTRACEREBRAL HEMORRHAGE WITH SERIOUS COMPLICATIONS AND PREDICTIVE FACTORS FOR HEMORRHAGIC TRANSFORMATION AFTER THROMBOLYTIC TREATMENT IN PATIENTS WITH ACUTE ISCHEMIC STROKE

Background: Accurate classification of postthrombolytic intracerebral hemorrhage (ICH) subtypes is vital for predicting stroke outcomes and managing ICH. Currently, the recommended classification criteria are the European Cooperative Acute Stroke Study III criteria, including two primary categories: hemorrhagic infarction (HI) and parenchymal hematoma (PH). Objectives: The primary objective of this study was to assess the contribution of various ICH subtypes to serious complications, with the secondary aim to identify associated predictors.Methods: The study examined medical records of patients with acute ischemic stroke receiving thrombolysis at Saraburi Hospital from 2014 to 2022. The logit model with the margins command assessed the association of ICH subtypes with serious complications, and multinomial logistic regression identified potential predictors for HI and PH. Results: Among 345 patients, HI-1, HI-2, PH-1 and PH-2 had prevalence rates of 3.2, 7.8, 4.9 and 7.5%, respectively, while 76.5% did not have ICH. PH-2 demonstrated the strongest correlation with inhospital mortality (adjusted risk ratio [RR] 2.83, 95% CI 1.56-5.13), invasive mechanical ventilator requirement (adjusted RR 3.93, 95% CI 2.09-7.39) and hematoma evacuation (adjusted RR 4.58, 95% CI 1.17-17.95) compared with patients of non-ICH. HI demonstrated a significant prolongation of hospitalization. (adjusted RR 3.30, 95% CI 1.53-7.12). Multinomial logistic regression analysis revealed that prior use of antiplatelet drugs, antihypertensive treatment before rt-PA, white blood cell count ≥11,750 cells/mm3 and baseline Alberta stroke program early CT scores ≤7 were independent predictors for PH. The adjusted odds ratios were 3.06 (95% CI, 1.23-7.57), 6.95 (95% CI, 2.62-18.45), 6.01 (95% CI, 2.17-16.65) and 5.01 (95% CI, 2.00-12.60), respectively. Conclusion: The PH-2 subtype was associated with the highest mortality, while our study demonstrated that the HI subtype, previously considered relatively benign with successful early recanalization, showed a significant prolongation of hospitalization compared with that of patients of non-ICH. High-risk patients of ICH require intensive monitoring to reduce complications.

The analysis of association between single features of small vessel disease and stroke outcome shows the independent impact of the number of microbleeds and presence of lacunes

The impact of small vessel disease (SVD) on stroke outcome was investigated either separately for its single features in isolation or for SVD sum score measuring a qualitative (binary) assessment of SVD-lesions. We aimed to investigate which SVD feature independently impacts the most on stroke outcome and to compare the continuous versus binary SVD assessment that reflects pronouncement and presence correspondingly. Patients with a first-ever anterior circulation ischemic stroke were retrospectively investigated. We performed an ordered logistic regression analysis to predict stroke outcome (mRS 3 months, 0–6) using age, stroke severity, and pre-stroke disability as baseline input variables and adding SVD-features (lacunes, microbleeds, enlarged perivascular spaces, white matter hyperintensities) assessed either continuously (model 1) or binary (model 2). The data of 873 patients (age 67.9 ± 15.4, NIHSS 24 h 4.1 ± 4.8) was analyzed. In model 1 with continuous SVD-features, the number of microbleeds was the only independent predictor of stroke outcome in addition to clinical parameters (OR 1.21; 95% CI 1.07–1.37). In model 2 with the binary SVD assessment, only the presence of lacunes independently improved the prediction of stroke outcome (OR 1.48, 1.1–1.99). In a post hoc analysis, both the continuous number of microbleeds and the presence of lacunes were independent significant predictors. Thus, the number of microbleeds evaluated continuously and the presence of lacunes are associated with stroke outcome independent from age, stroke severity, pre-stroke disability and other SVD-features. Whereas the presence of lacunes is adequately represented in SVD sum score, the microbleeds assessment might require another cutoff and/or gradual scoring, when prediction of stroke outcome is needed.

Abstract WMP69: Brain Atrophy and White Matter Disease Agreement on NCCT and MRI in Ischemic Stroke

Background: NCCT is the most widely used parenchymal imaging for acute stroke. Evaluations of brain frailty measures like atrophy and white matter disease (WMD) are becoming increasingly relevant in stroke outcome prediction but are conventionally thought to be best seen on MRI. We assessed the agreement between baseline NCCT and follow-up MRI ratings of brain atrophy and WMD and compared their predictive validity in relation to 90-day functional outcomes in acute ischemic stroke. Methods: In this post-hoc analysis of baseline NCCT and follow-up MRI data from the Alteplase compared to Tenecteplase (AcT) randomised-controlled trial, expert readers (stroke neurologists and radiologists) assessed atrophy using the global cortical atrophy (GCA), Koedam, and medial temporal lobe atrophy scales. WMD was measured using the Fazekas scale. Binary agreement (none-mild vs. moderate-severe) and agreement across the full range of scores between atrophy and WMD measures on NCCT and MRI were calculated using Gwet’s agreement coefficient (AC1). Logistic regressions and Delong’s test were used to compare the area under the curve (AUC) for the prediction of 90-day modified Rankin score (mRS) 0-1 when using NCCT vs MRI-based ratings for each atrophy and WMD variable. Results: Of 1577 patients included in the AcT trial, 491 had interpretable NCCT+MRI. Binary agreement was substantial (AC1:0.68-0.80) for deep and total WMD scores, Koedam scale and frontal GCA. Almost perfect binary agreement (AC1:0.81-0.97) was found for all other measures. Deep white matter showed a significant difference between NCCT and MRI ratings for predicting 90-day mRS 0-1 (MRI AUC = 0.61 [0.56-0.66], CT AUC = 0.56 [0.50-0.60], p = 0.01, Delong p = 0.02) There was no significant difference in predicting 90-day mRS for the other variables (all p &gt; 0.12, GCA: MRI AUC = 0.54 [0.49-0.60] CT AUC = 0.53 [0.47-0.58], p = 0.58, Delong p = 0.86). Conclusion: NCCT ratings of brain atrophy and WMD by experts have substantial to almost-perfect agreement with MRI atrophy and WMD ratings and achieve similar prediction of 90-day mRS 0-1 functional outcomes in acute stroke patients. This implies that it is reasonable to use NCCT scans to evaluate these brain frailty measures in clinical practice and in stroke trials.

Abstract WP266: Prediction of Stroke Incidence Using Machine Learning: The Suita Study

Background: This population-based study investigated the potential of machine learning algorithms to predict stroke incidence and identify important risk factors. This study aimed to evaluate the accuracy of these algorithms in constructing a stroke prediction model. Methods: Participants from the Suita study were included, and baseline measurements were used to predict stroke outcomes over a 15-year follow-up period. In total, 7,389 participants and 51 variables were investigated, including demographics, medical history, medical imaging, laboratory data, and lifestyle habits. Initially, unsupervised K-prototype clustering was used to group participants based on their stroke risk. Subsequently, five supervised models (logistic regression, random forest, support vector machine, extreme gradient boosting, and light gradient boosted machine) were applied to predict the stroke outcomes. The Shapley Additive Explanations (SHAP) method determined the most critical variables. Results: Unsupervised clustering revealed significant differences in stroke incidence among the three identified risk clusters (9.1%, 6.6%, and 3.2%). These clusters were categorized into high-, medium-, and low-risk groups. Among the supervised models, the random forest algorithm demonstrated the best performance. The top ten most important variables for predicting stroke incidence were identified using the SHAP, with age being the most influential variable. Other significant risk markers included systolic blood pressure, hypertension, estimated glomerular filtration rate, metabolic syndrome, and blood sugar level. Additionally, elbow joint thickness and fructosamine, hemoglobin, and calcium levels were found to be potential predictors of stroke risk. Notably, the variables identified by the SHAP were consistent with those obtained from the unsupervised clustering approach in the high-risk group. Conclusion: Machine learning algorithms provide accurate predictions of stroke incidence and offer valuable insights into subclinical markers without the need for prior assumptions of causality. This study presents a data-driven machine-learning framework for stroke risk prediction and biomarker identification.

Automated evaluation of collateral circulation for outcome prediction in acute ischemic stroke

IntroductionThe assessment of collateral circulation in acute ischemic stroke management is essential. Modern tools, such as Brainomix's e-CTA, powered by artificial intelligence, provide detailed insights into collateral assessment. This retrospective study aims to identify factors contributing to favorable collateral status and compare outcomes between patients with good collaterals (grade 3) and fair collaterals (grade 0–2). MethodThis retrospective study included 97 patients admitted to the Stroke Unit at the Hospital de Clínicas of the Federal University of Paraná, Brazil, from September 2021 to January 2023. Comparative analyses involved demographic factors, cardiovascular risk factors, and the combined outcome of mortality and moderate to severe disability at discharge, 30-day, and 90-day follow-ups. ResultsAmong the 97 cases, 58.8 % showed 'good collaterals' with a grade 3 status. Variables affecting collateral status included age (p = 0.042), neutrophil-lymphocyte ratio (p = 0.005), and initial NIHSS scores (p<0.001). The presence of good collaterals according to e-CTA reduced the odds of death and moderate-severe disability at discharge (p = 0.003; OR 0.27) and at 30 days (p = 0.015; OR 0.33), although this effect diminished at the 90-day mark after multivariate analysis. DiscussionAutomated collateral assessment through e-CTA is a valuable tool in acute ischemic stroke evaluation. Good e-CTA collateral score serve as a promising imaging biomarker, guiding informed clinical decisions during Stroke Unit hospitalizations. This study highlights the relationship between collaterals and stroke outcomes and underscores the potential for AI-driven tools to enhance stroke care management.

Adaptation of the Concept of Brain Reserve for the Prediction of Stroke Outcome: Proxies, Neural Mechanisms, and Significance for Research.

The prediction of stroke outcome is challenging due to the high inter-individual variability in stroke patients. We recently suggested the adaptation of the concept of brain reserve (BR) to improve the prediction of stroke outcome. This concept was initially developed alongside the one for the cognitive reserve for neurodegeneration and forms a valuable theoretical framework to capture high inter-individual variability in stroke patients. In the present work, we suggest and discuss (i) BR-proxies-quantitative brain characteristics at the time stroke occurs (e.g., brain volume, hippocampus volume), and (ii) proxies of brain pathology reducing BR (e.g., brain atrophy, severity of white matter hyperintensities), parameters easily available from a routine MRI examination that might improve the prediction of stroke outcome. Though the influence of these parameters on stroke outcome has been partly reported individually, their independent and combined impact is yet to be determined. Conceptually, BR is a continuous measure determining the amount of brain structure available to mitigate and compensate for stroke damage, thus reflecting individual differences in neural resources and a capacity to maintain performance and recover after stroke. We suggest that stroke outcome might be defined as an interaction between BR at the time stroke occurs and lesion load. BR in stroke can potentially be influenced, e.g., by modifying cardiovascular risk factors. In addition to the potential power of the BR concept in a mechanistic understanding of inter-individual variability in stroke outcome and establishing individualized therapeutic approaches, it might help to strengthen the synergy of preventive measures in stroke, neurodegeneration, and healthy aging.

A Deep Learning Approach to Predicting Stroke Outcomes from Brain Imaging Data

A Deep Learning Approach to Predicting Stroke Outcomes from Brain Imaging Data

Outcome Prediction by Combining Corticospinal Tract Lesion Load with Diffusion-tensor Fractional Anisotropy in Patients after Hemorrhagic Stroke.

The objective of this study was to evaluate the predictive precision of combining the corticospinal tract lesion load (CST-LL) with the diffusion-tensor fractional anisotropy of the corticospinal tract (CST-FA) in the lesioned hemispheres regarding motor outcomes. Patients with putaminal and/or thalamic hemorrhage who had undergone computed tomography (CT) soon after onset in our hospital were retrospectively enrolled. The CST-LL was calculated after registration of the CT images to a standard brain. Diffusion-tensor imaging was performed during the second week after onset. Standardized automated tractography was employed to calculate the CST-FA. Outcomes were assessed at discharge from our affiliated rehabilitation facility using total scores of the motor component of the Stroke Impairment Assessment Set (SIAS-motor total; null to full, 0 to 25). Multivariate regression analysis was performed with CST-LL and CST-FA as explanatory variables and SIAS-motor total as a target value. Twenty-five patients participated in this study. SIAS-motor total ranged from 0 to 25 (median, 17). CST-LL ranged from 0.298 to 7.595 (median, 2.522) mL, and the lesion-side CST-FA ranged from 0.211 to 0.530 (median, 0.409). Analysis revealed that both explanatory variables were detected as statistically significant contributory factors. The estimated t values indicated that the contributions of these two variables were almost equal. The obtained regression model accounted for 63.9% of the variability of the target value. Incorporation of the CST-LL with the lesion-side CST-FA enhances the precision of the stroke outcome prediction model.

Machine learning application in ischemic stroke diagnosis, management, and outcome prediction: a narrative review

Stroke is a leading cause of disability and death. The condition requires prompt diagnosis and treatment. The quality of care provided to patients with stroke can vary depending on the availability of medical resources, which in turn, can affect prognosis. Recently, there has been growing interest in using machine learning (ML) to support stroke diagnosis and treatment decisions based on large medical data sets. Current ML applications in stroke care can be divided into two categories: analysis of neuroimaging data and clinical informationbased predictive models. Using ML to analyze neuroimaging data can increase the efficiency and accuracy of diagnoses. Commercial software that uses ML algorithms is already being used in the medical field. Additionally, the accuracy of predictive ML models is improving with the integration of radiomics and clinical data. is expected to be important for improving the quality of care for patients with stroke.