Outcome Predictors in First-Ever Ischemic Stroke Patients: A Population-Based Study

A Multimodal Ensemble Deep Learning Model for Functional Outcome Prognosis of Stroke Patients

BACKGROUND AND PURPOSE

The accurate prediction of functional outcomes in patients with acute ischemic stroke (AIS) is crucial for informed clinical decision-making and optimal resource utilization. As such, this study aimed to construct an ensemble deep learning model that integrates multimodal imaging and clinical data to predict the 90-day functional outcomes after AIS.

METHODS

We used data from the Korean Stroke Neuroimaging Initiative database, a prospective multicenter stroke registry to construct an ensemble model integrated individual 3D convolutional neural networks for diffusion-weighted imaging and fluid-attenuated inversion recovery (FLAIR), along with a deep neural network for clinical data, to predict 90-day functional independence after AIS using a modified Rankin Scale (mRS) of 3-6. To evaluate the performance of the ensemble model, we compared the area under the curve (AUC) of the proposed method with that of individual models trained on each modality to identify patients with AIS with an mRS score of 3-6.

RESULTS

Of the 2,606 patients with AIS, 993 (38.1%) achieved an mRS score of 3-6 at 90 days post-stroke. Our model achieved AUC values of 0.830 (standard cross-validation [CV]) and 0.779 (time-based CV), which significantly outperformed the other models relying on single modalities: b-value of 1,000 s/mm2 (P<0.001), apparent diffusion coefficient map (P<0.001), FLAIR (P<0.001), and clinical data (P=0.004).

CONCLUSION

The integration of multimodal imaging and clinical data resulted in superior prediction of the 90-day functional outcomes in AIS patients compared to the use of a single data modality.

Machine learning approach for hemorrhagic transformation prediction: Capturing predictors' interaction

Background and purpose

Patients with ischemic stroke frequently develop hemorrhagic transformation (HT), which could potentially worsen the prognosis. The objectives of the current study were to determine the incidence and predictors of HT, to evaluate predictor interaction, and to identify the optimal predicting models.

Methods

A prospective study included 360 patients with ischemic stroke, of whom 354 successfully continued the study. Patients were subjected to thorough general and neurological examination and T2 diffusion-weighted MRI, at admission and 1 week later to determine the incidence of HT. HT predictors were selected by a filter-based minimum redundancy maximum relevance (mRMR) algorithm independent of model performance. Several machine learning algorithms including multivariable logistic regression classifier (LRC), support vector classifier (SVC), random forest classifier (RFC), gradient boosting classifier (GBC), and multilayer perceptron classifier (MLPC) were optimized for HT prediction in a randomly selected half of the sample (training set) and tested in the other half of the sample (testing set). The model predictive performance was evaluated using receiver operator characteristic (ROC) and visualized by observing case distribution relative to the models' predicted three-dimensional (3D) hypothesis spaces within the testing dataset true feature space. The interaction between predictors was investigated using generalized additive modeling (GAM).

Results

The incidence of HT in patients with ischemic stroke was 19.8%. Infarction size, cerebral microbleeds (CMB), and the National Institute of Health stroke scale (NIHSS) were identified as the best HT predictors. RFC (AUC: 0.91, 95% CI: 0.85-0.95) and GBC (AUC: 0.91, 95% CI: 0.86-0.95) demonstrated significantly superior performance compared to LRC (AUC: 0.85, 95% CI: 0.79-0.91) and MLPC (AUC: 0.85, 95% CI: 0.78-0.92). SVC (AUC: 0.90, 95% CI: 0.85-0.94) outperformed LRC and MLPC but did not reach statistical significance. LRC and MLPC did not show significant differences. The best models' 3D hypothesis spaces demonstrated non-linear decision boundaries suggesting an interaction between predictor variables. GAM analysis demonstrated a linear and non-linear significant interaction between NIHSS and CMB and between NIHSS and infarction size, respectively.

Conclusion

Cerebral microbleeds, NIHSS, and infarction size were identified as HT predictors. The best predicting models were RFC and GBC capable of capturing nonlinear interaction between predictors. Predictor interaction suggests a dynamic, rather than, fixed cutoff risk value for any of these predictors.

Quantitative EEG Correlates with NIHSS and MoCA for Assessing the Initial Stroke Severity in Acute Ischemic Stroke Patients

BACKGROUND: National Institutes of Health Stroke Scale (NIHSS) and Montreal Cognitive Assessment (MoCA) measure stroke severity by assessing the functional and cognitive outcome, respectively. However, they cannot be used to measure subtle evolution in clinical symptoms during the early phase. Quantitative EEG (qEEG) can detect any subtle changes in CBF and brain metabolism thus may also benefit for assessing the severity. AIM: This study aims to identify the correlation between qEEG with NIHSS and MoCA for assessing the initial stroke severity in acute ischemic stroke patients. METHODS: This was a cross-sectional study. We recruited 30 patients with first-ever acute ischemic stroke hospitalized in Dr. Sardjito General Hospital, Yogyakarta, Indonesia. We measured the NIHSS, MoCA score, and qEEG parameter during the acute phase of stroke. Correlation and regression analysis was completed to investigate the relationship between qEEG parameter with NIHSS and MoCA. RESULTS: Four acute qEEG parameter demonstrated moderate-to-high correlations with NIHSS and MoCA. DTABR had positive correlation with NIHSS (r = 0.379, p = 0.04). Meanwhile, delta-absolute power, DTABR, and DAR were negatively correlated with MoCA score (r = −0.654, p = 0.01; r = −0.397, p = 0.03; and r = −0.371, p = 0.04, respectively). After adjusted with the confounding variables, delta-absolute power was independently associated with MoCA score, but not with NIHSS (B = −2.887, 95% CI (−4.304–−1.470), p < 0.001). CONCLUSIONS: Several qEEG parameters had significant correlations with NIHSS and MoCA in acute ischemic stroke patients. The use of qEEG in acute clinical setting may provide a reliable and efficient prediction of initial stroke severity. Further cohort study with larger sample size and wide range of stroke severity is still needed.

Machine-learning-based outcome prediction in stroke patients with middle cerebral artery-M1 occlusions and early thrombectomy

Background and purpose

Clinical outcomes vary substantially among individuals with large vessel occlusion (LVO) stroke. A small infarct core and large imaging mismatch were found to be associated with good recovery. The aim of this study was to investigate whether those imaging variables would improve individual prediction of functional outcome after early (<6 h) endovascular treatment (EVT) in LVO stroke.

Methods

We included 222 patients with acute ischemic stroke due to middle cerebral artery (MCA)‐M1 occlusion who received EVT. As predictors, we used clinical variables and region of interest (ROI)‐based magnetic resonance imaging features. We developed different machine‐learning models and quantified their prediction performance according to the area under the receiver‐operating characteristic curves and the Brier score.

Results

The rate of successful recanalization was 78%, with 54% patients having a favorable outcome (modified Rankin scale score 0–2). Small infarct core was associated with favorable functional outcome. Outcome prediction improved only slightly when imaging was added to patient variables. Age was the driving factor, with a sharp decrease in likelihood of favorable functional outcome above the age of 78 years.

Conclusions

In patients with MCA‐M1 occlusion strokes referred to EVT within 6 h of symptom onset, infarct core volume was associated with outcome. However, ROI‐based imaging variables led to no significant improvement in outcome prediction at an individual patient level when added to a set of clinical predictors. Our study is in concordance with current practice, where imaging mismatch or collateral readouts are not recommended as factors for excluding patients with MCA‐M1 occlusion for early EVT.

Evaluation of inflammatory markers and mean platelet volume as short-term outcome indicators in young adults with ischemic stroke

Background

Studying outcome predictors in patients with onset of cerebral infarction in early adult life may enhance our knowledge of disease pathophysiology and prognosis.

Aim

The aim is to identify independent predictors of short-term outcome of first-ever ischemic stroke in young adults with special emphasis on inflammatory and thrombogenic markers.

Methods

We enrolled 33 patients aged 19–44 years with first-ever ischemic stroke admitted to Kasr Alainy Stroke Unit and 33 matched controls. Clinical, radiological, and laboratory (adhesion molecules, C-reactive protein, prolactin, and mean platelet volume) evaluations were carried out. Functional outcome at 7 days after stroke onset was assessed using the modified Rankin scale, and independent predictors were identified.

Results

The most frequently identified risk factor was cardiac abnormality. Patients exhibited significantly higher levels of baseline inflammatory and thrombogenic markers compared with controls. These markers were significantly correlated with the stroke severity. Logistic regression model showed that high National Institutes of Health Stroke Scale (NIHSS) score (odds ratios [OR] = 0.13; 95% confidence interval [CI], 0.04–0.24; P = 0.01) and large infarction size (OR = 0.11; 95% CI, 0.09–0.17; P = 0.04) but not the laboratory markers were independent predictors of unfavorable outcome.

Conclusion

Our data suggested that higher NIHSS scores and large infarction size served as independent predictors of short-term unfavorable outcome, while inflammatory and thrombogenic markers did not.

Prediction of the Lethal Outcome of Acute Recurrent Cerebral Ischemic Hemispheric Stroke

Background and objectives. Stroke-induced mortality is the third most common cause of death in developed countries. Intense interest has focused on the recurrent ischemic stroke, which rate makes up 30% during first 5 years after first-ever stroke. This work aims to develop criteria for the prediction of acute recurrent cerebral ischemic hemispheric stroke (RCIHS) outcome on the basis of comprehensive baseline clinical, laboratory, and neuroimaging examinations. Materials and Methods. One hundred thirty-six patients (71 males and 65 females, median age 74 (65; 78)) with acute RCIHS were enrolled in the study. All patients underwent a detailed clinical and neurological examination using National Institutes of Health Stroke Scale (NIHSS), computed tomography of the brain, hematological, and biochemical investigations. In order to detect the dependent and independent risk factors of the lethal outcome of the acute period of RCIHS, univariable and multivariable regression analysis were conducted. A receiver operating characteristic (ROC) analysis with the calculation of sensitivity and specificity was performed to determine the prediction variables. Results. Twenty-five patients died. The independent predictors of the lethal outcome of acute RCIHS were: Baseline NIHSS score (OR 95% CІ 1.33 (1.08-1.64), p = 0.0003), septum pellucidum displacement (OR 95% CI 1.53 (1.17-2.00), p = 0.0021), glucose serum level (OR 95% CI 1.28 (1.09-1.50), p = 0.0022), neutrophil-to-lymphocyte ratio (OR 95% CI 1.11 (1.00-1.21), p = 0.0303). The mathematical model, which included these variables was developed and it could determine the prognosis of lethal outcome of the acute RCIHS with an accuracy of 86.8% (AUC = 0.88 ± 0.04 (0.88-0.93), p < 0.0001).

Development, Reliability, and Validity of the Multiple Errands Test Home Version (MET–Home) in Adults With Stroke

OBJECTIVE. Our objective was to perform initial psychometric analysis of the Multiple Errands Test Home Version (MET–Home), which was designed to assess the influence of poststroke executive dysfunction on in-home task performance.

METHOD. We examined the reliability and validity of the MET–Home in adults with stroke (n = 23) and individually matched control participants (n = 23). All participants completed a series of assessments during a single in-home visit.

RESULTS. Notable differences in MET-Home subscores were discovered between participants with stroke and control participants. Participants with stroke omitted more tasks, broke more rules, passed by tasks more often, and were less efficient than matched control participants. The MET–Home demonstrated evidence of adequate internal consistency, excellent interrater reliability, and significant moderate associations with several tests.

CONCLUSION. This preliminary study suggests that the MET–Home differentiates between adults with stroke and matched control participants. The MET–Home provides evidence of initial reliability and validity among adults with stroke.

Associations between self-generated strategy use and MET-Home performance in adults with stroke

Self-generated strategy use has substantial potential for improving community living outcomes in adults with impaired executive function after stroke. However, little is known about how self-generated strategies support task performance in people with post-stroke executive function impairments living in the community. We explored strategy use among home-dwelling persons with stroke and neurologically-healthy control participants during the Multiple Errands Test-Home Version (MET-Home), a context-specific assessment with evidence of ecological validity designed to examine how post-stroke executive dysfunction manifests during task performance in the home environment. For persons with stroke, significant associations were identified between planning and tasks accurately completed on the MET-Home. Significant associations were also identified among the control participants for self-monitoring, multitasking, and "using the environment" strategies. These associations are related to enhanced MET-Home performance on sub-scores for levels of accuracy, passes, and total time. Rehabilitation interventions that focus on reinforcing self-generated strategy use may support community living outcomes in persons with post-stroke executive function impairments, but this area needs additional investigation.

Ferreira P. Assessing upper limb function: transcultural adaptation and validation of the Portuguese version of the Stroke Upper Limb Capacity Scale

BACKGROUND

Brachial hemiparesis is one of the most frequent sequelae of stroke, leading to important functional disability given the role of the upper limb in executing activities of daily living (ADL). The Stroke Upper Limb Capacity Scale (SULCS) is a stroke-specific assessment instrument that evaluates functional capacity of the upper limb based on the execution of 10 tasks. The objective of this study is the transcultural adaptation and psychometric validation of the Portuguese version of the SULCS.

METHODS

A Portuguese version of the SULCS was developed, using the process of forward-backward translation, after authorisation from the author of the original scale. Then, a multicentre study was conducted in Portuguese stroke patients (n = 122) to validate the psychometric properties of the instrument. The relationship between sociodemographic and clinical characteristics was used to test construct validity. The relationship between SULCS scores and other instruments was used to test criterion validity.

RESULTS

Semantic and linguistic adaptation of the SULCS was executed without substantial issues and allowed the development of a Portuguese version. The application of this instrument suggested the existence of celling effect (19.7% of participants with maximum score). Reliability was demonstrated through the intraclass correlation coefficient of 0.98. As for construct validity, SULCS was sensible to muscle tonus and aphasia. SULCS classification impacted the scores of the Motor Evaluation Scale for Upper Extremity in Stroke (MESUPES) and the Stroke Impact Scale (SIS).

CONCLUSIONS

The present version of SULCS shows valid and reliable cultural adaptation, with good reliability and stability.