THRIVE Score Predicts Ischemic Stroke Outcomes and Thrombolytic Hemorrhage Risk in VISTA

Traditional and machine learning models for predicting haemorrhagic transformation in ischaemic stroke: a systematic review and meta-analysis

BACKGROUND

Haemorrhagic transformation (HT) is a severe complication after ischaemic stroke, but identifying patients at high risks remains challenging. Although numerous prediction models have been developed for HT following thrombolysis, thrombectomy, or spontaneous occurrence, a comprehensive summary is lacking. This study aimed to review and compare traditional and machine learning-based HT prediction models, focusing on their development, validation, and diagnostic accuracy.

METHODS

PubMed and Ovid-Embase were searched for observational studies or randomised controlled trials related to traditional or machine learning-based models. Data were extracted according to Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist and risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool (PROBAST). Performance data for prediction models that were externally validated at least twice and showed low risk of bias were meta-analysed.

RESULTS

A total of 100 studies were included, with 67 focusing on model development and 33 on model validation. Among 67 model development studies, 44 were traditional model studies involving 47 prediction models (with National Institutes of Health Stroke Scale score being the most frequently used predictor in 35 models), and 23 studies focused on machine learning prediction models (with support vector machines being the most common algorithm, used in 10 models). The 33 validation studies externally validated 34 traditional prediction models. Regarding study quality, 26 studies were assessed as having a low risk of bias, 11 as unclear, and 63 as high risk of bias. Meta-analysis of 15 studies validating eight models showed a pooled area under the receiver operating characteristic curve of approximately 0.70 for predicting HT.

CONCLUSION

While significant progress has been made in developing HT prediction models, both traditional and machine learning-based models still have limitations in methodological rigour, predictive accuracy, and clinical applicability. Future models should undergo more rigorous validation, adhere to standardised reporting frameworks, and prioritise predictors that are both statistically significant and clinically meaningful. Collaborative efforts across research groups are essential for validating these models in diverse populations and improving their broader applicability in clinical practice.

SYSTEMATIC REVIEW REGISTRATION

International Prospective Register of Systematic Reviews (CRD42022332816).

Stroke-SCORE: Personalizing Acute Ischemic Stroke Treatment to Improve Patient Outcomes

Background/Objectives: Acute ischemic stroke (AIS) is a leading cause of disability and mortality worldwide. Despite advances in interventions such as thrombolysis (TL) and mechanical thrombectomy (MT), current treatment protocols remain largely standardized, focusing on general eligibility rather than individual patient characteristics. To address this gap, we introduce the Stroke-SCORE (Simplified Clinical Outcome Risk Evaluation), a predictive tool designed to personalize AIS management by providing data-driven, individualized recommendations to optimize treatment strategies and improve patient outcomes. Methods: The Stroke-SCORE was derived using retrospective data from 793 AIS patients admitted to the University of Pécs (February 2023-September 2024). Logistic regression analysis identified age, National Institutes of Health Stroke Scale (NIHSS) score at admission, and pre-morbid modified Rankin Scale (pre-mRS) score as key predictors of unfavorable outcomes at 90 days (defined as modified Rankin Scale [mRS] score > 2). Based on these predictors, a simplified risk score was developed to stratify patients into low-, moderate-, and high-risk groups, guiding treatment decisions on TL, MT, combination therapy (TL + MT), or standard care (SC). Internal validation was performed to assess the model's predictive performance via receiver operating characteristic (ROC) analysis and isotonic regression calibration with bootstrapping. Results: The Stroke-SCORE was moderately positively correlated with a 90-day mRS score > 2 (odds ratio [OR] = 0.70, 95% confidence interval [CI]: 0.58-0.83, p < 0.001), with an area under the curve (AUC) of 0.86, a sensitivity and specificity of 79% and 81%, respectively, and an overall accuracy of 80%. Simulations indicated that personalized treatment guided by the Stroke-SCORE significantly reduced unfavorable outcomes. Conclusions: The Stroke-SCORE demonstrates strong predictive performance as a practical, data-driven approach for personalizing AIS treatment decisions. In the future, external, multicenter prospective validation is needed to confirm its applicability in real-world settings.

Neurologic prognostication in coma and disorders of consciousness

Coma and disorders of consciousness (DoC) are clinical syndromes primarily resulting from severe acute brain injury, with uncertain recovery trajectories that often necessitate prolonged supportive care. This imposes significant socioeconomic burdens on patients, caregivers, and society. Predicting recovery in comatose patients is a critical aspect of neurocritical care, and while current prognostication heavily relies on clinical assessments, such as pupillary responses and motor movements, which are far from precise, contemporary prognostication has integrated more advanced technologies like neuroimaging and electroencephalogram (EEG). Nonetheless, neurologic prognostication remains fraught with uncertainty and significant inaccuracies and is impacted by several forms of prognostication biases, including self-fulfilling prophecy bias, affective forecasting, and clinician treatment biases, among others. However, neurologic prognostication in patients with disorders of consciousness impacts life-altering decisions including continuation of treatment interventions vs withdrawal of life-sustaining therapies (WLST), which have a direct influence on survival and recovery after severe acute brain injury. In recent years, advancements in neuro-monitoring technologies, artificial intelligence (AI), and machine learning (ML) have transformed the field of prognostication. These technologies have the potential to process vast amounts of clinical data and identify reliable prognostic markers, enhancing prediction accuracy in conditions such as cardiac arrest, intracerebral hemorrhage, and traumatic brain injury (TBI). For example, AI/ML modeling has led to the identification of new states of consciousness such as covert consciousness and cognitive motor dissociation, which may have important prognostic significance after severe brain injury. This chapter reviews the evolving landscape of neurologic prognostication in coma and DoC, highlights current pitfalls and biases, and summarizes the integration of clinical examination, neuroimaging, biomarkers, and neurophysiologic tools for prognostication in specific disease states. We will further discuss the future of neurologic prognostication, focusing on the integration of AI and ML techniques to deliver more individualized and accurate prognostication, ultimately improving patient outcomes and decision-making process in neurocritical care.

NIH stroke scale and unfavourable outcomes in acute ischaemic stroke: retrospective study

OBJECTIVES

To evaluate the determining factors of severe functional impairment (SFI) outcome at discharge and in-hospital mortality in patients who had an acute ischaemic stroke and thus favouring early implementation of primary palliative care (PC).

METHODS

A retrospective descriptive study by the analysis of 515 patients who had an acute ischaemic stroke admitted at stroke unit, aged≥18 years, from January 2017 to December 2018. Previous clinical and functional status data, National Institute of Health Stroke Scale (NIHSS) on admission, and data related to the evolution during hospitalisation were evaluated, relating them to the SFI outcome at discharge and death. The significance level was set at 5%.

RESULTS

Of 515 patients included, 15% (77) died, 23.3%(120) had an SFI outcome and 9.1% (47) were evaluated by the PC team. It was observed that NIHSS Score≥16 is responsible for a 15.5-fold increase in the occurrence of death outcome. The presence of atrial fibrillation was responsible for a 3.5-fold increase in the risk of this outcome.

CONCLUSION

NIHSS Score is an independent predictor of in-hospital death and SFI outcomes at discharge. Knowledge about the prognosis and risk of developing unfavourable outcomes is important for planning the care of patients affected by a potentially fatal and limiting acute vascular insult.

Hemorrhagic transformation after thrombolytic therapy for acute ischemic stroke: Accuracy and improvement of existing predictive models in a rural population of the Midwest

BACKGROUND

Hemorrhagic transformation (HT) after rtPA in acute ischemic stroke is a known complication of thrombolytic therapy. Several grading scales have been introduced in clinical practice, aiming to quantify the risk of HT before rtPA is administered. The goals of this study are to evaluate the performance of existing grading scales in a rural population of the Midwest and improve the existing models.

METHODS

This is a retrospective study of stroke patients treated with thrombolytics at Southern Illinois Healthcare from July 2017 to August 2024. Demographics, clinical presentations, laboratory values, neuroimaging, and stroke metrics were collected. HT found on neuroimaging within 24 h after rtPA was reviewed. mRS at 30 days was noted. The cohort was divided in two groups: HT and no-HT. The two groups were compared by univariate analyses. SEDAN, HAT, MSS, and THRIVE scores were calculated, and multivariable logistic regression analysis was run for each model. Area under the receiver operating characteristic curve (AUC) with its 95 % confidence interval was calculated for each grading scale. P value was set at 0.05.

RESULTS

Out of 279 patients included in this study, HT occurred in 8.6 % of patients (n = 24), whereas 91.4 % (n = 255) had no-HT. The two groups were similar in baseline characteristics and stroke severity. HT group had significantly worse mRS 0-2 at 30 days (42 % vs. 69 %, p < 0.05). SEDAN score demonstrated the highest accuracy in predicting HT after rtPA (AUC = 0.65, 95 % CI:0.56-0.75). Adding 1 point for smoking to the score, SEDAN-S, improved the accuracy of the model (AUC = 0.67, 95 % CI:0.57-0.77).

CONCLUSIONS

Existing predictive scales of HT after rtPA underperform in our rural population. Among those, SEDAN score is the most accurate predictor. Adding smoking status to the score improves its accuracy. Further larger studies in similar rural populations should be performed to confirm our results.

Machine learning for early dynamic prediction of functional outcome after stroke

BACKGROUND

Prediction of outcome after stroke is critical for treatment planning and resource allocation but is complicated by fluctuations during the first days after onset. We propose a machine learning model that can provide hourly predictions based on the integration of continuous variables acquired within 72 h of hospital admission.

METHODS

We analyzed 2492 admissions for ischemic stroke in the Geneva University Hospital from 01.01.2018 to 31.12.2021, amounting to 2'131'752 unique data points. We developed a transformer model that continuously included clinical, physiological, imaging, and biological data recorded within 72 h of admission. This model was trained to generate hourly predictions of mortality and morbidity. Shapley additive explanations were used to identify the most relevant predictors to explain outcomes for each patient. The MIMIC-III database was used for external validation.

RESULTS

Our transformer model predicts mortality, with an area under the receiver operating characteristic curve of 0.830 (95% CI 0.763-0.885) on admission, reaching 0.893 (95% CI 0.839-0.933) 72 h later for a 3-month outcome. Validated in an independent cohort, it outperforms all static models. Based on their mean explanatory weights, the top predictors included continuous clinical evaluation, baseline patient characteristics, timing from admission to acute treatment, and markers of inflammation and organ dysfunction.

CONCLUSIONS

The performance of our transformer model demonstrates the potential of machine learning models integrating clinical, physiological, imaging, and biological variables over time after stroke. The clinical applicability of our model is further strengthened by access to hourly updated predictions along with accompanying explanations.

External validation of different predictive scores for symptomatic intracranial hemorrhage after intravenous thrombolysis in Asian stroke patients

OBJECTIVE

This study aimed to externally validate different predictive scores for symptomatic intracranial hemorrhage (SICH) after intravenous thrombolysis (IVT), with a particular focus on their predictive abilities in Asian stroke patients.

METHODS

We retrospectively enrolled stroke patients who received a standard dose of alteplase within 4.5 hours from symptom onset at the First Affiliated Hospital of Dalian Medical University from July 2010 to August 2023. SICH was defined as the hemorrhagic transformation detected on the head CT scan completed within 48 h post-IVT, accompanied by a clinical deterioration of at least a 4-point increase in NIHSS score. Predictive abilities of the HAT, MSS, SEDAN, SPAN-100, and GRASPS scores were tested. Discrimination and calibration were performed using the area under the receiver operating characteristic curve (ROC-AUC), DeLong test, and Hosmer-Lemeshow (H-L) goodness-of-fit test.

RESULTS

The study included 1007 stroke patients, of whom 31 (3.08 %) developed SICH. ROC-AUCs for predicting SICH were: 0.796 (95 %CI: 0.726-0.866) for the GRASPS score, 0.724 (95 %CI: 0.644-0.804) for the MSS score, 0.715 (95 %CI: 0.619-0.811) for the SEDAN score, 0.714 (95 %CI: 0.611-0.817) for the HAT score, and 0.605 (95 %CI: 0.491-0.720) for the SPAN-100 score (all P < 0.05). DeLong tests showed that the GRASPS score demonstrated significantly better discrimination than the MSS score (P = 0.010), the SEDAN score (P = 0.009), the HAT score (P = 0.049), and the SPAN-100 score (P = 0.000). H-L tests indicated good calibrations which were ranked HAT > SEDAN > MSS > SPAN-100 > GRASPS scores.

CONCLUSION

The GRASPS score showed reasonable predictive ability for SICH, indicating its potential utility for Asian stroke patients receiving IVT.

External validation of the iScore, ASTRAL score, DRAGON score, and THRIVE score and development of a nomogram to predict outcome in patients with large vessel occlusion-acute ischemic stroke

OBJECTIVE

This study aimed to validate the iScore, ASTRAL score, DRAGON score, and THRIVE score for assessing large vessel occlusion-acute ischemic stroke (AIS-LVO) and establish a predictive model for AIS-LVO patients that has better performance to guide clinical practice.

METHODS

We retrospectively included 439 patients with AIS-LVO and collected baseline data from all of them. External validation of the iScore, ASTRAL score, DRAGON score, and THRIVE score was performed. All variables were compared between groups via univariate analysis, and the results are expressed as ORs and 95 % CIs. Independent variables with P < 0.25 were included in the multivariate logistic analysis, and statistically significant differences (P < 0.05) were identified as risk factors for prognosis in AIS-LVO patients. Receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA) were used to evaluate the predictive value of our model.

RESULTS

Our external validation resulted in an iScore under the curve (AUC) of 0.8475, an ASTRAL AUC of 0.8324, a DRAGON AUC of 0.8196, and a THRIVE AUC of 0.8039. In our research, multivariate Cox regression revealed 8 independent predictors. We used a nomogram to visualize the results of the data analysis. The AUC for the training cohort was 0.8855 (95 % CI, 0.8487-0.9222), and that in the validation cohort was 0.8992 (95 % CI, 0.8496-0. 9488).

CONCLUSIONS

In this study, we verified that the above scores have excellent efficacy in predicting the prognosis of AIS-LVO patients. The nomogram we developed was able to predict the prognosis of AIS-LVO more accurately and may contribute to personalized clinical decision-making and treatment for future clinical work.

Multi-modality multi-task model for mRS prediction using diffusion-weighted resonance imaging

This study focuses on predicting the prognosis of acute ischemic stroke patients with focal neurologic symptoms using a combination of diffusion-weighted magnetic resonance imaging (DWI) and clinical information. The primary outcome is a poor functional outcome defined by a modified Rankin Scale (mRS) score of 3-6 after 3 months of stroke. Employing nnUnet for DWI lesion segmentation, the study utilizes both multi-task and multi-modality methodologies, integrating DWI and clinical data for prognosis prediction. Integrating the two modalities was shown to improve performance by 0.04 compared to using DWI only. The model achieves notable performance metrics, with a dice score of 0.7375 for lesion segmentation and an area under the curve of 0.8080 for mRS prediction. These results surpass existing scoring systems, showing a 0.16 improvement over the Totaled Health Risks in Vascular Events score. The study further employs grad-class activation maps to identify critical brain regions influencing mRS scores. Analysis of the feature map reveals the efficacy of the multi-tasking nnUnet in predicting poor outcomes, providing insights into the interplay between DWI and clinical data. In conclusion, the integrated approach demonstrates significant advancements in prognosis prediction for cerebral infarction patients, offering a superior alternative to current scoring systems.

Predicting ischemic stroke patients' prognosis changes using machine learning in a nationwide stroke registry

Accurately predicting the prognosis of ischemic stroke patients after discharge is crucial for physicians to plan for long-term health care. Although previous studies have demonstrated that machine learning (ML) shows reasonably accurate stroke outcome predictions with limited datasets, to identify specific clinical features associated with prognosis changes after stroke that could aid physicians and patients in devising improved recovery care plans have been challenging. This study aimed to overcome these gaps by utilizing a large national stroke registry database to assess various prediction models that estimate how patients' prognosis changes over time with associated clinical factors. To properly evaluate the best predictive approaches currently available and avoid prejudice, this study employed three different prognosis prediction models including a statistical logistic regression model, commonly used clinical-based scores, and a latest high-performance ML-based XGBoost model. The study revealed that the XGBoost model outperformed other two traditional models, achieving an AUROC of 0.929 in predicting the prognosis changes of stroke patients followed for 3 months. In addition, the XGBoost model maintained remarkably high precision even when using only selected 20 most relevant clinical features compared to full clinical datasets used in the study. These selected features closely correlated with significant changes in clinical outcomes for stroke patients and showed to be effective for predicting prognosis changes after discharge, allowing physicians to make optimal decisions regarding their patients' recovery.

Prognostic value of combining 24-hour ASPECTS and hemoglobin to red cell distribution width ratio to the THRIVE score in predicting in-hospital mortality among ischemic stroke patients treated with intravenous thrombolysis

BACKGROUND

Acute ischemic stroke (AIS) is a significant global health issue, directly impacting mortality and disability. The Totaled Health Risks in Vascular Events (THRIVE) score is appreciated for its simplicity and ease of use to predict stroke clinical outcomes; however, it lacks laboratory and neuroimaging data, which limits its ability to predict outcomes precisely. Our study evaluates the impact of integrating the 24-hour Alberta Stroke Program Early CT Score (ASPECTS) and hemoglobin-to-red cell distribution width (HB/RDW) ratio into the THRIVE score using the multivariable fractional polynomial (MFP) method (combined THRIVE-MFP model) compared to the THRIVE-c model. We aim to assess their added value in predicting in-hospital mortality (IHM) prognosis.

MATERIALS AND METHODS

A retrospective study from January 2015 to July 2022 examined consecutive AIS patients receiving intravenous thrombolysis. Data on THRIVE scores, 24-hour ASPECTS, and HB/RDW levels were collected upon admission. The model was constructed using logistic regression and the MFP method. The prognostic value was determined using the area under the receiver operating characteristic curve (AuROC). Ischemic cerebral lesions within the middle cerebral artery territory were evaluated with non-contrast computed tomography (NCCT) after completing 24 hours of intravenous thrombolysis (24-hour ASPECTS).

RESULTS

Among a cohort of 345 patients diagnosed with AIS who received intravenous thrombolysis, 65 individuals (18.8%) experienced IHM. The combined THRIVE-MFP model was significantly superior to the THRIVE-c model in predicting IHM (AuROC 0.980 vs. 0.876, p<0.001), 3-month mortality (AuROC 0.947 vs. 0.892, p<0.001), and 3-month poor functional outcome (AuROC 0.910 vs. 0.853, p<0.001).

CONCLUSION

The combined THRIVE-MFP model showed excellent predictive performance, enhancing physicians' ability to stratify patient selection for intensive neurological monitoring and guiding treatment decisions. Incorporating 24-hour ASPECTS on NCCT and HB/RDW proved valuable in mortality prediction, particularly for hospitals with limited access to advanced neuroimaging resources.

Radiologic and Blood Markers Predicting Long-Term Neurologic Outcome Following Decompressive Craniectomy for Malignant Ischemic Stroke: A Preliminary Single-Center Study

BACKGROUND

 Malignant ischemic stroke (MIS) is defined by progressive cerebral edema leading to increased intracranial pressure (ICP), compression of neural structures, and, eventually, death. Decompressive craniectomy (DC) has been advocated as a lifesaving procedure in the management of patients with MIS. This study aims to identify pre- and postoperative predictive variables of neurologic outcomes in patients undergoing DC for MIS.

METHODS

 We conducted a retrospective study of patients undergoing DC in a single center from April 2016 to April 2020. Preoperative workup included baseline clinical status, laboratory data, and brain computed tomography (CT). The primary outcome was the 6-month modified Rankin score (mRS). The secondary outcome was the 30-day mortality.

RESULTS

 During data capture, a total of 58 patients fulfilled the criteria for MIS, of which 22 underwent DC for medically refractory increased ICP and were included in the present analysis. The overall median age was 58.5 years. An immediate (24 hour) postoperative extended Glasgow Outcome Scale (GOSE) score ≥5 was associated with a good 6-month mRS (1-3; p = 0.004). Similarly, low postoperative neutrophils (p = 0.002), low lymphocytes (p = 0.004), decreased neutrophil-to-lymphocyte ratio (NLR; p = 0.02), and decreased platelet-to-lymphocytes ratio (PLR; p = 0.03) were associated with good neurologic outcomes. Preoperative variables independently associated with worsened 6-month mRS were the following: increased age (odds ratio [OR]: 1.10; 95% confidence interval [CI]: 1.01-1.20; p = 0.02), increased National Institutes of Health Stroke Scale (NIHSS) score (OR: 7.8; 95% CI: 2.5-12.5; p = 0.035), Glasgow Coma Scale (GCS) score less than 8 at the time of neurosurgical referral (OR: 21.63; 95% CI: 1.42-328; p = 0.02), and increased partial thromboplastin time (PTT) before surgery (OR: 2.11; 95% CI: 1.11-4; p = 0.02). Decreased postoperative lymphocytes confirmed a protective role against worsened functional outcomes (OR: 0.01; 95% CI: 0.01-0.4; p = 0.02). Decreased postoperative lymphocyte count was associated with a protective role against increased mRS (OR: 0.01; 95% CI: 0.01-0.4; p = 0.02). The occurrence of hydrocephalus at the postoperative CT scan was associated with 30-day mortality (p = 0.005), while the persistence of postoperative compression of the ambient and crural cistern showed a trend towards higher mortality (p = 0.07).

CONCLUSIONS

 This study reports that patients undergoing DC for MIS showing decreased postoperative blood inflammatory markers achieved better 6-month neurologic outcomes than patients with increased inflammatory markers. Similarly, poor NIHSS score, poor GCS score, increased age, and larger PTT values at the time of surgery were independent predictors of poor outcomes. Moreover, the persistence of postoperative compression of basal cisterns and the occurrence of hydrocephalus are associated with 30-day mortality.

Deep learning model integrating radiologic and clinical data to predict mortality after ischemic stroke

Objective

Most prognostic indexes for ischemic stroke mortality lack radiologic information. We aimed to create and validate a deep learning-based mortality prediction model using brain diffusion weighted imaging (DWI), apparent diffusion coefficient (ADC), and clinical factors.

Methods

Data from patients with ischemic stroke who admitted to tertiary hospital during acute periods from 2013 to 2019 were collected and split into training (n = 1109), validation (n = 437), and internal test (n = 654). Data from patients from secondary cardiovascular center was used for external test set (n = 507). The algorithm for predicting mortality, based on DWI and ADC (DLP_DWI), was initially trained. Subsequently, important clinical factors were integrated into this model to create the integrated model (DLP_INTG). The performance of DLP_DWI and DLP_INTG was evaluated by using time-dependent area under the receiver operating characteristic curves (TD AUCs) and Harrell concordance index (C-index) at one-year mortality.

Results

The TD AUC of DLP_DWI was 0.643 in internal test set, and 0.785 in the external dataset. DLP_INTG had a higher performance at predicting one-year mortality than premise score in internal dataset (TD- AUC: 0.859 vs. 0.746; p = 0.046), and in external dataset (TD- AUC: 0.876 vs. 0.808; p = 0.007). DLP_DWI and DLP_INTG exhibited strong discrimination for the high-risk group for one-year mortality.

Interpretation

A deep learning model using brain DWI, ADC and the clinical factors was capable of predicting mortality in patients with ischemic stroke.

The systemic inflammation score is a prognostic factor for patients with ischemic stroke who have not undergone intravenous thrombolysis or endovascular thrombectomy therapy

BACKGROUND

The systemic inflammation score (SIS) has been utilised as a representative biomarker for evaluating nutritional and inflammation status. However, the predictive value of SIS has not been reported in patients with acute ischemic stroke (AIS). We aimed to evaluate whether SIS is associated with prognosis in stroke.

METHODS

A total of 4801 patients with AIS were included in the study. The primary outcome was a modified Rankin Scale score>2 at the 3-month follow-up. A total of 4801 patients were randomly allocated into training (n=3361) and validation cohorts (n=1440) at a ratio of 7:3. Model performance was validated using the receiver operating characteristic (ROC) curve and calibration curve. Additionally, a comparison was made between the nomogram and the THRIVE score in regards to their respective predictive capabilities.

RESULTS

Overall, 1091(32.5%) patients in the training cohort and 446 (31.0%) patients in the validation cohort experienced an unfavorable outcome. The multivariate logistic regression analysis revealed that a high SIS, age, NIHSS, diabetes and prior stroke were associated with unfavorable outcome. Our nomogram was developed based on the variables mentioned above. The area under the curve (AUC) of the training set and the validation set are 0.702 and 0.708, respectively, indicating that the model has modest agreement and discrimination. The results of AUC, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) showed that nomogram had significantly higher predictive value than THRIVE scores (all P<0.001). However, unlike the THRIVE publication, all patients who had undergone intravenous thrombolysis or endovascular thrombectomy therapy were excluded in our study. In consequence, our derived THRIVE scores cannot be compared to those in the original THRIVE study.

CONCLUSION

The SIS exhibits potential as a simple prognostic biomarker, and the nomogram, which utilizes the SIS, may serve as a valuable tool for clinicians in the early identification of patients at heightened risk for unfavorable outcomes.

Supra-Blan2 t score as a multisystem-based risk score to predict poor 3-month outcome in acute ischemic stroke patients with intravenous thrombolysis

AIM

To develop and validate a novel weighted score integrating multisystem laboratory and clinical variables to predict poor 3-month outcome (mRS score of 3-6) in acute ischemic stroke (AIS) patients with intravenous thrombolysis (IVT) therapy.

METHODS

We retrospectively analyzed data from Trial of Revascularization Treatment for Acute Ischemic Stroke study. The Supra-Blan2 t score was derived using the data on age, the National Institutes of Health Stroke Scale score, history of atrial fibrillation, blood sugar level, neutrophil count, direct bilirubin level, platelet-lymphocyte ratio, and TnI level in the derivation cohort of 433 patients, and validated in a cohort of 525 patients. Furthermore, we compared the performance of the Supra-Blan2 t score with DRAGON, TURN, and SPAN-100 scores.

RESULTS

The discrimination capacity in the derivation and validation cohorts was good for poor 3-month outcome (the area under the curve was 0.821 and 0.843, respectively). The cumulative incidence of poor 3-month outcome significantly increased across risk categories in the derivation (low-risk, 9.2%; medium-risk, 17.4%; and high-risk, 58.8%) and validation cohorts (12.7%, 36.5%, and 73.6%, respectively). The performance of the Supra-Blan2 t score was similar to or superior to DRAGON, TURN, and SPAN-100 scores.

CONCLUSION

The Supra-Blan2 t score, based on easily available multisystem laboratory and clinical variables, reliably predicted poor 3-month functional outcome in AIS patients treated with IVT therapy featuring good calibration and discrimination.

The combination model of serum occludin and clinical risk factors improved the efficacy for predicting hemorrhagic transformation in stroke patients with recanalization

Background

and Purpose: Hemorrhagic transformation (HT) is one of the severe complications in acute ischemic stroke, especially for the patients who undergo recanalization treatment. It is crucial to screen patients who have high risk of HT before recanalization. However, current prediction models based on clinical factors are not ideal for clinical practice. Serum occludin, a biomarker for cerebral ischemia-induced blood-brain barrier disruption, has potential for predicting HT. This study was to investigate whether the combination of serum occludin and clinical risk factors improved the efficacy of predicting HT.

Methods

This was a single-center prospective observational study. Baseline clinical data and blood samples of recanalization patients were collected upon admission to our hospital. The level of serum occludin was measured using enzyme-linked immunosorbent assay. The diagnosis of HT was confirmed by CT scans within 36 h post recanalization.

Results

A total of 324 patients with recanalization were enrolled and 68 patients presented HT occurrence. HT patients had the higher level of baseline occludin than patients without HT (p < 0.001). Multivariate regression analysis showed that serum occludin level, Alberta Stroke Program Early CT Scores and endovascular therapy were independent risk factors (p < 0.05) for HT after adjusting potential confounders. The combination of serum occludin and clinical risk factors significantly improved the accuracy of predicting HT [area under the curve (AUC, 0.821 vs 0.701, p < 0.001), and net reclassification improvement (31.1 %), integrated discrimination improvement (21.5 %), p < 0.001] compared to a model employing only clinical risk factors. The modified AUC (0.806) of combined model based on 10-fold-cross-validation was still higher than clinical risk model (0.701).

Conclusion

The combination of serum occludin and clinical risk factors significantly improved the prediction efficacy for HT, providing a novel potential prediction model to screen for patients with high risk of HT before recanalization in acute ischemic stroke.

External Validation and Updating of Published Models for Predicting 7-day Risk of Symptomatic Intracranial Hemorrhage after Receiving Alteplase for Acute Ischemic Stroke: A Retrospective Cohort Study

Background

Prediction scores for symptomatic intracranial hemorrhage (sICH) in acute ischemic stroke patients receiving thrombolytic therapy have been widely developed, but the external validation of these scores, especially in the Thai population, is lacking. This study aims to externally validate existing models and update the selected model to enhance its performance in our specific context.

Methods

This cohort study retrospectively collected data from medical records between 2013 and 2022. Acute ischemic stroke patients who received thrombolysis were included. All predictors were gathered at admission. External validation was performed on eight published prediction models; in addition, the observed and expected probabilities of sICH were compared. The most effective model for discrimination was then chosen for further updating using multivariable logistic regression and was bootstrapped for internal validation. Finally, a points-based system for clinical practice was developed from the optimism-corrected model.

Results

Fifty patients (10% of the 502 included cohort members) experienced sICH after undergoing thrombolysis. The SICH score outperformed the other seven models in terms of discrimination (area under the receiver operating characteristic [AuROC] curve = 0.74 [95% confidence interval {CI} 0.67 to 0.81]), but it still overstated risk (expected-to-observed outcomes [E/O] ratio = 1.7). Once updated, the optimism-corrected revised SICH model showed somewhat better calibration (E/O = 1 and calibration-in-the-large = 0), slightly worse underprediction in the moderate-to-high risk group (calibration slope = 1.152), and marginally better discrimination (AuROC = 0.78). The points-based system also demonstrated substantial agreement (88.1%) with the risk groups predicted by the logistic regression model (kappa statistic = 0.78).

Conclusion

Since the SICH score outperformed seven models in terms of discrimination, it was then modified to the Revised-SICH score, which predicted that patients with at least 5.5 points were at high risk of having sICH.

Developing and Validating a New Model to Predict the Risk of Poor Neurological Status of Acute Ischemic Stroke After Intravenous Thrombolysis

OBJECTIVES

The objective of this study was to develop and validate a predictive model for the risk of poor neurological status in in-hospital patients with acute ischemic stroke (AIS) after intravenous thrombolysis.

METHODS

This 2-center retrospective study included patients with AIS treated at the Advanced Stroke Center of the Second Hospital of Hebei Medical University and Baoding No.1 Central Hospital between January 2018 and January 2020). The neurological function status at day 7 of AIS onset was used as the endpoint of the study, which was evaluated using the National Institute of Health Stroke Scale (NIHSS) score.

RESULTS

A total of 878 patients were included in the study and divided into training (n=652) and validation (n=226) sets. Seven variables were selected as predictors to establish the risk model: age, NIHSS before thrombolysis (NIHSS1), NIHSS 24 hours after thrombolysis (NIHSS3), high-density lipoprotein, antiplatelet, cerebral computed tomography after thrombolysis (CT2), and lower extremity venous color Doppler ultrasound. The risk prediction model achieved good discrimination (the areas under the Receiver Operating Characteristic curve in the training and validation sets were 0.9626 and 0.9413, respectively) and calibration (in the training set Emax=0.072, Eavg=0.01, P =0.528, and in the validation set Emax=0.123, Eavg=0.019, P =0.594, respectively). The decision curve analysis showed that the model could achieve a good net benefit.

CONCLUSIONS

The prediction model obtained in this study showed good discrimination, calibration, and clinical efficacy. This new nomogram can provide a reference for predicting the risk of poor neurological status in patients with acute ischemic stroke after intravenous thrombolysis.

Clinical Grading Scales and Neuroprognostication in Acute Brain Injury

Prediction of neurological clinical outcome after acute brain injury is critical because it helps guide discussions with patients and families and informs treatment plans and allocation of resources. Numerous clinical grading scales have been published that aim to support prognostication after acute brain injury. However, the development and validation of clinical scales lack a standardized approach. This in turn makes it difficult for clinicians to rely on prognostic grading scales and to integrate them into clinical practice. In this review, we discuss quality measures of score development and validation and summarize available scales to prognosticate outcomes after acute brain injury. These include scales developed for patients with coma, cardiac arrest, ischemic stroke, nontraumatic intracerebral hemorrhage, subarachnoid hemorrhage, and traumatic brain injury; for each scale, we discuss available validation studies.

Predictors of 7-day symptomatic hemorrhagic transformation in patients with acute ischemic stroke and proposal of a novel screening tool: A retrospective cohort study

OBJECTIVES

Hemorrhagic transformation (HT) is significantly related to poor neurological outcomes and mortality. Although variables and models that predict HT have been reported in the literature, the need for a model with high diagnostic performance continues. We aimed to propose a model that can accurately predict symptomatic HT within 7 days of acute ischemic stroke (AIS).

METHODS

Patients with AIS admitted to the emergency department of a tertiary training and research hospital between November 07, 2021, and August 26, 2022, were included in this single-center retrospective study. For the model, binary logistics with the forced-entry method was used and the model was validated with 3-fold cross-validation. After the final model was created, the optimal cutoff point was determined with Youden's index. Another cut-off point was determined at which the sensitivity was the highest.

RESULTS

The mean age of the 423 patients included in the study was 70 (60-81) and 53.7% (n = 227) of the patients were male. Symptomatic HT was present in 31 (7.3%) patients. Mechanical thrombectomy, atrial fibrillation, and diabetes mellitus were the independent predictors (P < 0.001, P = 0.003, P = 0.006, respectively). The mean area under the curve of the receiver operating characteristics of the model was 0.916 (95% confidence interval [CI] = 0.876-0.957). The sensitivity for the optimal cut-off point was 90.3% (95% CI = 74.3%-97.9%) and specificity was 80.6% (95% CI = 76.4%-84.4%). For the second cutoff point where the sensitivity was 100%, the specificity was 60.5% (95% CI = 55.4%-65.3%).

CONCLUSION

The diagnostic performance of our model was satisfactory and it seems to be promising for symptomatic HT. External validation studies are required to implement our results into clinical use.

Predictive Factors for Hemorrhagic Transformation in Acute Ischemic Stroke in the REAL-World Clinical Practice

BACKGROUND

Few data exists on predictive factors of hemorrhagic transformation (HT) in real-world acute ischemic stroke patients. The aims of this study were: (i) to identify predictive variables of HT (ii) to develop a score for predicting HT.

METHODS

We retrospectively analyzed the clinical, radiographic, and laboratory data of patients with acute ischemic stroke consecutively admitted to our Stroke Unit along two years. Patients with HT were compared with those without HT. A multivariate logistic regression analysis was performed to identify independent predictors of HT on CT scan at 24 hours to develop a practical score.

RESULTS

The study population consisted of 564 patients with mean age 77.5±11.8 years. Fifty-two patients (9.2%) showed HT on brain CT at 24 hours (4.9% symptomatic). NIHSS score ≥8 at Stroke Unit admission (3 points), cardioembolic etiology (2 points), acute revascularization by systemic thrombolysis and/or mechanical thrombectomy (1 point), history of previous TIA/stroke (1 point), and major vessel occlusion (1 point) were found independent risk factors of HT and were included in the score (Hemorrhagic Transformation Empoli score (HTE)). The predictive power of HTE score was good with an AUC of 0.785 (95% CI: 0.749-0.818). Compared with 5 HT predictive scores proposed in the literature (THRIVE, SPAN-100, MSS, GRASPS, SITS-SIC), the HTE score significantly better predicted HT.

CONCLUSIONS

NIHSS score ≥8 at Stroke Unit admission, cardioembolism, urgent revascularization, previous TIA/stroke, and major vessel occlusion were independent predictors of HT. The HTE score has a good predictive power for HT. Prospective studies are warranted.

Mechanical Thrombectomy Versus Best Medical Treatment in the Late Time Window in Non-DEFUSE-Non-DAWN Patients: A Multicenter Cohort Study

BACKGROUND

We assessed the efficacy and safety of mechanical thrombectomy (MT) in adult stroke patients with anterior circulation large vessel occlusion presenting in the late time window not fulfilling the DEFUSE-3 (Thrombectomy for Stroke at 6 to 16 Hours With Selection by Perfusion Imaging trial) and DAWN (Thrombectomy 6 to 24 Hours After Stroke With a Mismatch Between Deficit and Infarct trial) inclusion criteria.

METHODS

Cohort study of adults with anterior circulation large vessel occlusion admitted between 6 and 24 hours after last-seen-well at 5 participating Swiss stroke centers between 2014 and 2021. Mismatch was assessed by computer tomography or magnetic resonance imaging perfusion with automated software (RAPID or OLEA). We excluded patients meeting DEFUSE-3 and DAWN inclusion criteria and compared those who underwent MT with those receiving best medical treatment alone by inverse probability of treatment weighting using the propensity score. The primary efficacy end point was a favorable functional outcome at 90 days, defined as a modified Rankin Scale score shift toward lower categories. The primary safety end point was symptomatic intracranial hemorrhage within 7 days of stroke onset; the secondary was all-cause mortality within 90 days.

RESULTS

Among 278 patients with anterior circulation large vessel occlusion presenting in the late time window, 190 (68%) did not meet the DEFUSE-3 and DAWN inclusion criteria and thus were included in the analyses. Of those, 102 (54%) received MT. In the inverse probability of treatment weighting analysis, patients in the MT group had higher odds of favorable outcomes compared with the best medical treatment alone group (modified Rankin Scale shift: acOR, 1.46 [1.02-2.10]; P=0.04) and lower odds of all-cause mortality within 90 days (aOR, 0.59 [0.37-0.93]; P=0.02). There were no significant differences in symptomatic intracranial hemorrhage (MT versus best medical treatment alone: 5% versus 2%, P=0.63).

CONCLUSIONS

Two out of 3 patients with anterior circulation large vessel occlusion presenting in the late time window did not meet the DEFUSE-3 and DAWN inclusion criteria. In these patients, MT was associated with higher odds of favorable functional outcomes without increased rates of symptomatic intracranial hemorrhage. These findings support the enrollment of patients into ongoing randomized trials on MT in the late window with more permissive inclusion criteria.

Development and Validation of a Novel Score for Predicting Long-Term Mortality after an Acute Ischemic Stroke

BACKGROUND

Long-term mortality prediction can guide feasible discharge care plans and coordinate appropriate rehabilitation services. We aimed to develop and validate a prediction model to identify patients at risk of mortality after acute ischemic stroke (AIS).

METHODS

The primary outcome was all-cause mortality, and the secondary outcome was cardiovascular death. This study included 21,463 patients with AIS. Three risk prediction models were developed and evaluated: a penalized Cox model, a random survival forest model, and a DeepSurv model. A simplified risk scoring system, called the C-HAND (history of Cancer before admission, Heart rate, Age, eNIHSS, and Dyslipidemia) score, was created based on regression coefficients in the multivariate Cox model for both study outcomes.

RESULTS

All experimental models achieved a concordance index of 0.8, with no significant difference in predicting poststroke long-term mortality. The C-HAND score exhibited reasonable discriminative ability for both study outcomes, with concordance indices of 0.775 and 0.798.

CONCLUSIONS

Reliable prediction models for long-term poststroke mortality were developed using information routinely available to clinicians during hospitalization.

Pre-treatment spectral CT combined with CT perfusion can predict hemorrhagic transformation after thrombolysis in patients with acute ischemic stroke

OBJECTIVE

To investigate the value of pre-treatment spectral CT angiography (CTA) in predicting hemorrhagic transformation (HT) after intravenous thrombolysis (IVT) treatment in acute ischemic stroke (AIS) patients.

MATERIALS AND METHODS

AIS patients who underwent IVT with recombinant tissue plasminogen activator and pre-treatment head and neck spectral CTA and head CT perfusion (CTP) from January 2018 to June 2020 were reviewed retrospectively. Finally, 20 patients were included in the HT group and 22 age-matched patients were included in the non-HT group. Spectral and CTP parameters of the region of interest on pre-treatment CTA axial raw images and CTP images, including the infarct core (IC) and ischemic penumbral (IP) regions, were recorded. The differences in clinical variables, CTP, collateral scores and spectral parameters between the two groups were analyzed. Three multivariate logistic regression models were then developed, where model 1 included clinical and spectral parameters, model 2 included clinical and CTP parameters and a combined model included clinical, CTP, and spectral parameters. Receiver operating characteristic analysis was used to evaluate the performance of the multivariate model.

RESULTS

Patients with HT had higher Safe Implementation of Treatments in Stroke (SITS) score (p = 0.023), the volume of perfusion lesions (p = 0.005), the volume of IP (p = 0.003), the mean transit time (MIT) in the IC area (p = 0.012), as well as the TTP in IP area (p = 0.015) compared with patients without HT. The HT group showed significantly lower CBF in the IC area (p = 0.019), iodine concentration (p = 0.017) and the effective atomic number (p = 0.024) in the IP area than non-HT group. And the slope of the spectral curve of the HT group in the IP region was larger than that of the non-HT group (p = 0.023). Gender, age, SITS score, the volume of entire perfusion lesion, CBF and MIT in the IC area, TTP in the IP area, as well as iodine concentration in the IP area were included in the final multivariate model for predicting HT. And CBF in the IC area (OR = 0.779, 95 % CI:0.609-0.996, p = 0.046) as well as the iodine concentration of IP area (OR = 0.343, 95 % CI: 0.131-0.901, p = 0.030) were proved to be independent predictors for HT. The combined model including clinical, spectral, and CTP parameters, showed improved accuracy compared to the other two models, while the Delong test did not suggest a statistically significant difference (both p values > 0.05).

CONCLUSIONS

The iodine concentration of IP area derived from pre-treatment spectral CTA was an independent predictor of HT after IVT treatment for AIS patients. Moreover, multivariate models combined with clinical, spectral, and CTP parameters may be able to predict HT.

The SON2A2 score: A novel grading scale for predicting hemorrhage and outcomes after thrombolysis

Objectives

This study aimed to develop a score including novel putative predictors for predicting the risk of sICH and outcomes after thrombolytic therapy with intravenous (IV) recombinant tissue-type plasminogen activator (r-tPA) in acute ischemic stroke patients.

Methods

All patients with acute ischemic stroke treated with IV r-tPA at three university-based hospitals in Chongqing, China, from 2014 to 2019 were retrospectively studied. Potential risk factors associated with sICH (NINDS criteria) were determined with multivariate logistic regression, and we developed our score according to the magnitude of logistic regression coefficients. The score was validated in another independent cohort. Area under the receiver operating characteristic curve (AUC-ROC) was used to assess the performance of the score. Calibration was evaluated using the Hosmer-Lemeshow goodness-of-fit method.

Results

The SON2A2 score (0 to 8 points) consisted of history of smoking (no = 1, yes = 0, β = 0.81), onset-to-needle time (≥3.5 = 1,<3.5=0, β = 0.74), NIH Stroke Scale on admission (>10 = 2, ≤10 = 0, β = 1.22), neutrophil percentage (≥80.0% = 1, <80% = 0, β = 0.81), ASPECT score (≤11 = 2, >11 = 0, β = 1.30), and age (>65 years = 1, ≤65 years = 0, β = 0.89). The SON2A2 score was strongly associated with sICH (OR 1.98; 95%CI 1.675-2.34) and poor outcomes (OR 1.89; 95%CI 1.68-2.13). AUC-ROC in the derivation cohort was 0.82 (95%CI 0.77-0.86). Similar results were obtained in the validation cohort. The Hosmer-Lemeshow test revealed that predicted and observed event rates in derivation and validation cohorts were very close.

Conclusion

The SON2A2 score is a simple, efficient, quick, and easy-to-perform scale for predicting the risk of sICH and outcome after intravenous r-tPA thrombolysis within 4.5 h in patients with ischemic stroke, and risk assessment using this test has the potential for early and personalized management of this disease in high-risk patients.

A novel nomogram to predict hemorrhagic transformation in ischemic stroke patients after intravenous thrombolysis

Background

Hemorrhagic transformation (HT) is the most serious complication of ischemic stroke patients after intravenous thrombolysis and leads to a poor clinical prognosis. This study aimed to determine the independent predictors associated with HT in stroke patients with intravenous thrombolysis and to establish and validate a nomogram that combines with predictors to predict the probability of HT after intravenous thrombolysis in patients with ischemic stroke.

Method

This study enrolled ischemic stroke patients with intravenous thrombolysis from December 2016 to June 2022. All the patients were divided into training and validation cohorts. The nomogram was composed of the significant predictors for HT in the training cohort as obtained by the multivariate logistic regression analysis. The area under the receiver operating characteristic curve was used to assess the discriminative performance of the nomogram. The calibration performance of the nomogram was assessed by the Hosmer–Lemeshow goodness-of-fit test and calibration plots. Decision curve analysis was used to test the clinical validity of the nomogram.

Results

A total of 394 patients with intravenous thrombolysis were enrolled in the study. In the training cohort (n = 257), 45 patients had HT after intravenous thrombolysis. Multivariate logistic regression analysis demonstrated early infarct signs (OR, 7.954; 95% CI, 3.553-17.803; P < 0.001), NIHSS scores (OR, 1.110; 95% CI, 1.054-1.168; P < 0.001), uric acid (OR, 0.993; 95% CI, 0.989–0.997; P = 0.001), and albumin-to-globulin ratio (OR, 0.109; 95% CI, 0.023–0.508; P = 0.005) were independent predictors for HT and constructed the nomogram. In the training and validation cohorts, the AUC of the nomogram was 0.859 and 0.839, respectively. The Hosmer–Lemeshow goodness-of-fit test and calibration plot showed good concordance between predicted and observed probability in the training and validation cohorts. Decision curve analysis indicated that the nomogram was significantly useful for predicting HT in the training and further confirmed in the validation cohort.

Conclusion

This study proposes a novel and practical nomogram based on early infarct signs, NIHSS scores, uric acid, and albumin-to-globulin ratio that can well predict the probability of HT after intravenous thrombolysis in patients with ischemic stroke.

Intelligible Models for HealthCare: Predicting the Probability of 6-Month Unfavorable Outcome in Patients with Ischemic Stroke

Early prediction of unfavorable outcome after ischemic stroke is significant for clinical management. Machine learning as a novel computational modeling technique could help clinicians to address the challenge. We aim to investigate the applicability of machine learning models for individualized prediction in ischemic stroke patients and demonstrate the utility of various model-agnostic explanation techniques for machine learning predictions. A total of 499 consecutive patients with Unfavorable [modified Rankin Scale (mRS) score 3-6, n = 140] and favorable (mRS score 0-2, n = 359) outcome after 6-month from ischemic stroke were enrolled in this study. Four machine learning models, including Random Forest [RF], eXtreme Gradient Boosting [XGBoost], Adaptive Boosting [Adaboost] and Support Vector Machine [SVM] were performed with the area-under-the-curve (AUC): (90.20 ± 0.22)%, (86.91 ± 1.05)%, (86.49 ± 2.35)%, (81.89 ± 2.40)%, respectively. Three global interpretability techniques (Feature Importance shows the contribution of selected features, Partial Dependence Plot aims to visualize the average effect of a feature on the predicted probability of unfavorable outcome, Feature Interaction detects the change in the prediction that occurs by varying the features after considering the individual feature effects) and one local interpretability technique (Shapley Value indicates the probability of unfavorable outcome of different instances) have been applied to present the interpretability techniques via visualization. Thereby, the current study is important for better understanding intelligible healthcare analytics via explanations for the prediction of local and global levels, and potentially reduction of the mortality of patients with ischemic stroke by assisting clinicians in the decision-making process.

Deep Learning Applications for Acute Stroke Management

Brain imaging is essential to the clinical care of patients with stroke, a leading cause of disability and death worldwide. Whereas advanced neuroimaging techniques offer opportunities for aiding acute stroke management, several factors, including time delays, inter-clinician variability, and lack of systemic conglomeration of clinical information, hinder their maximal utility. Recent advances in deep machine learning (DL) offer new strategies for harnessing computational medical image analysis to inform decision making in acute stroke. We examine the current state of the field for DL models in stroke triage. First, we provide a brief, clinical practice-focused primer on DL. Next, we examine real-world examples of DL applications in pixel-wise labeling, volumetric lesion segmentation, stroke detection, and prediction of tissue fate postintervention. We evaluate recent deployments of deep neural networks and their ability to automatically select relevant clinical features for acute decision making, reduce inter-rater variability, and boost reliability in rapid neuroimaging assessments, and integrate neuroimaging with electronic medical record (EMR) data in order to support clinicians in routine and triage stroke management. Ultimately, we aim to provide a framework for critically evaluating existing automated approaches, thus equipping clinicians with the ability to understand and potentially apply DL approaches in order to address challenges in clinical practice. ANN NEUROL 2022.

Interleukin-6 and YKL-40 predicted recurrent stroke after ischemic stroke or TIA: analysis of 6 inflammation biomarkers in a prospective cohort study

OBJECTIVE

Contribution of individual and combined inflammatory markers in prognosis after stroke was still undefined. We aimed to investigate the association of systemic and local vascular inflammatory markers and recurrent stroke as well as impact on poor functional outcome.

METHODS

In this pre-specified substudy of the Third China National Stroke Registry (CNSR-III), 10,472 consecutive acute ischemic stroke or TIA patients with available centralized-measured levels of Interleukin-6 (IL-6), high sensitive C-reactive protein (hsCRP), IL-1 receptor antagonist (IL-1Ra), lipoprotein-associated phospholipase A₂ mass (Lp-PLA₂) and activity (Lp-PLA₂-A), and YKL-40 from 171 sites were enrolled. The primary outcomes consisted of stroke recurrence and poor functional outcome defined as modified Rankin Scale (mRS) score of 2-6 within 1 year.

RESULTS

There were 1026 (9.8%) and 2395 (23.4%) patients with recurrent stroke and poor functional outcome within 1 year. The highest quartiles of IL-6 (adjusted HR, 1.36; 95% CI 1.13-1.64; P = 0.001), hsCRP (adjusted HR, 1.41; 95% CI 1.17-1.69; P = 0.0003) and YKL-40 (adjusted HR, 1.28; 95% CI 1.06-1.56; P = 0.01) were associated with increased risk of recurrent stroke; and the highest quartiles of IL-6 (adjusted OR 1.93; 95% CI 1.64-2.27; P < 0.0001), IL-1Ra (adjusted OR 1.60; 95% CI 1.37-1.87; P < 0.0001), hsCRP (adjusted OR 1.60; 95% CI 1.37-1.86; P < 0.0001) and YKL-40 (adjusted OR 1.21; 95% CI 1.03-1.42; P = 0.02) were correlated with increased risk of poor functional outcome. In the multivariate stepwise regression analysis including all markers with backward selection, elevated levels of IL-6 or YKL-40 were associated with recurrent stroke (IL6: OR, 1.34; 95% CI 1.19-1.52; P < 0.0001; YKL-40: OR, 1.01; 95% CI 1.01-1.03; P = 0.004) and poor functional outcome (IL6: OR, 1.68; 95% CI 1.46-1.93; P < 0.0001; YKL-40: OR, 1.02; 95% CI 1.01-1.03; P = 0.0001). Adding IL-6 and YKL-40 significantly increased the area under the receiver operating characteristic curves for the prediction models of Essen Stroke Risk Score (0.03, P < 0.0001) and Totaled Health Risks in Vascular Events Score (0.07, P < 0.0001), and yielded continuous net reclassification improvement (19.0%, P < 0.0001; 33.0, P < 0.0001).

CONCLUSIONS

In the patients with ischemic stroke or TIA, IL-6 and YKL-40 were independently associated with recurrent stroke and poor functional outcome, and improved risk classification of clinical risk algorithms.

Precision medicine in stroke: towards personalized outcome predictions using artificial intelligence

Stroke ranks among the leading causes for morbidity and mortality worldwide. New and continuously improving treatment options such as thrombolysis and thrombectomy have revolutionized acute stroke treatment in recent years. Following modern rhythms, the next revolution might well be the strategic use of the steadily increasing amounts of patient-related data for generating models enabling individualized outcome predictions. Milestones have already been achieved in several health care domains, as big data and artificial intelligence have entered everyday life. The aim of this review is to synoptically illustrate and discuss how artificial intelligence approaches may help to compute single-patient predictions in stroke outcome research in the acute, subacute and chronic stage. We will present approaches considering demographic, clinical and electrophysiological data, as well as data originating from various imaging modalities and combinations thereof. We will outline their advantages, disadvantages, their potential pitfalls and the promises they hold with a special focus on a clinical audience. Throughout the review we will highlight methodological aspects of novel machine-learning approaches as they are particularly crucial to realize precision medicine. We will finally provide an outlook on how artificial intelligence approaches might contribute to enhancing favourable outcomes after stroke.

Triglyceride-Glucose Index and Intravenous Thrombolysis Outcomes for Acute Ischemic Stroke: A Multicenter Prospective–Cohort Study

Background

The triglyceride-glucose (TyG) index has recently been proposed as a reliable marker of insulin resistance. There is insufficient evidence to verify that the TyG index is correlated with functional outcomes and hemorrhagic transformation and in patients with stroke treated with intravenous thrombolysis (IVT).

Methods

We designed a multicenter cohort study, which enrolled patients with acute ischemic stroke treated with IVT between December 2004 and December 2016. The TyG index was divided into tertiles and calculated on a continuous scale. Unfavorable functional outcomes were defined by the modified Rankin Scale of 3–6 at 90 days and the incident rates of symptomatic intracranial hemorrhage (SICH) within 36 h of IVT onset were surveyed. Stroke severity was defined as mild (4–8), moderate (9–15), or high (≥16) based on the National Institutes of Health Stroke Scale (NIHSS) scores.

Results

Among 914 enrolled patients, the tertiles of the TyG index were 8.48 for T1, 8.48–9.04 for T2, and 9.04 for T3. T3 showed an increased risk of unfavorable functional outcomes at 90 days [odds ratio (OR): 1.76; P = 0.0132]. The TyG index was significantly associated with unfavorable functional outcomes at 90 days (OR: 1.32; P = 0.0431 per unit increase). No association was found between the TyG index and SICH. These findings were applicable for T3 with stroke of moderate (OR, 2.35; P = 0.0465) and high severity (OR: 2.57, P = 0.0440) patients with stroke.

Conclusion

This study supports the strong association between the increased TyG index and increased unfavorable functional outcomes at 90 days in patients with acute ischemic stroke treated with IVT. These findings were found to be robust in patients with moderate and high stroke severity.

Comparison of Different Dosages of Alteplase in Atrial Fibrillation-Related Acute Ischemic Stroke After Intravenous Thrombolysis: A Nationwide, Multicenter, Prospective Cohort Study in Taiwan

Background

Insufficient evidence is available for patients with acute ischemic stroke with atrial fibrillation (AF) to determine the efficacy and safety of different dosages of intravenous thrombolysis treatment. This study examined clinical outcomes in Chinese patients with stroke with and without AF after intravenous thrombolysis treatment with different intravenous thrombolysis doses.

Methods and Results

This multicenter, prospective cohort study recruited 2351 patients with acute ischemic stroke (1371 with AF and 980 without AF) treated with intravenous thrombolysis using alteplase. The Totaled Health Risks in Vascular Events score is a validated risk‐scoring tool used for assessing patients with acute ischemic stroke with and without AF. We evaluated favorable functional outcome at day 90 and symptomatic intracranial hemorrhage within 24 to 36 hours and outcomes of the patients receiving different doses of alteplase. Compared with the non‐AF group, the AF group exhibited a 2‐ to 3‐fold increased risk of symptomatic intracranial hemorrhage according to the National Institute of Neurological Disorders and Stroke standard (relative risk [RR], 2.10 [95% CI, 1.35–3.26]). Favorable functional outcome at 90 days and symptomatic intracranial hemorrhage rates according to the European Cooperative Acute Stroke Study II and the Safe Implementation of Thrombolysis in Stroke‐Monitoring Study standards did not significantly differ between the AF and non‐AF groups. In addition, the low‐dose alteplase subgroup exhibited an increased risk of symptomatic intracranial hemorrhage according to the National Institute of Neurological Disorders and Stroke standard (RR, 2.84 [95% CI, 1.63–4.96]). A validation study confirmed these findings after adjustment for scores determined using different stroke risk‐scoring tools.

Conclusions

Different alteplase dosages did not affect functional status at 90 days in the AF and non‐AF groups. Thus, the adoption of low‐dose alteplase simply because of AF is not recommended.

Predicting Functional Outcome After Decompressive Craniectomy for Malignant Hemispheric Infarction: Clinical and Novel Imaging Factors

OBJECTIVE

Decompressive craniectomy (DC) is an established optional treatment for malignant hemispheric infarction (MHI). We analyzed relevant clinical factors and computed tomography (CT) measurements in patients with DC for MHI to identify predictors of functional outcome 3-6 months after stroke.

METHODS

This study was performed at 2 comprehensive stroke centers. The inclusion criteria required DC for MHI, no additional intraoperative procedures (strokectomy or cerebral ventricular drain placement), and documented functional status 3-6 months after the stroke. We classified functional outcome as acceptable if the modified Rankin Scale score was <5, or as unacceptable if it was 5 or 6 (bedbound and totally dependent on others or death). Multiple logistic regression analyzed relevant clinical factors and multiple perioperative CT measurements to identify predictors of acceptable functional outcome.

RESULTS

Of 87 identified consecutive patients, 66 met the inclusion criteria. Acceptable functional outcome occurred in 35 of 66 (53%) patients. Likelihood of acceptable functional outcome decreased significantly with increasing age (OR 0.92, 95% CI 0.82-0.97, P = 0.004) and with increasing post-DC midline brain shift (OR 0.78, 95% CI 0.64-0.96, P = 0.016), and decreased non-significantly with left-sided stroke (OR 0.30, 95% CI 0.08-1.10, P = 0.069) and with increasing craniectomy barrier thickness (OR 0.92, 95% CI 0.85-1.01, P = 0.076).

CONCLUSIONS

Patient age and the post-DC midline shift may be useful in prognosticating functional outcome after DC for MHI. Stroke side and craniectomy barrier thickness merit further ideally prospective outcome prediction testing.

Development and Validation of a Predictive Model for Spontaneous Hemorrhagic Transformation After Ischemic Stroke

Background: Hemorrhagic transformation (HT) after reperfusion therapy for acute ischemic stroke (AIS) has been well studied; however, there is scarce research focusing on spontaneous HT (sHT). Spontaneous HT is no less important with a relatively high incidence and could be associated with neurological worsening. We aimed to develop and validate a simple and practical model to predict sHT after AIS (SHAIS) and compared the predictive value of the SHAIS score against the models of post-Reperfusion HT for sHT.

Methods: Patients with AIS admitted within 24 h of onset were prospectively screened to develop and validate the SHAIS score. The primary outcome was sHT during hospitalization (within 30 days after onset), and the secondary outcomes were symptomatic sHT and parenchymal hematoma (PH). Clinical information, laboratory, and neuroimaging data were screened to construct the SHAIS score. We selected six commonly used scales for predicting HT after reperfusion therapy and compared their predictive ability for sHT with the SHAIS score using Delong's test.

Results: The derivation cohort included 539 patients (mean age, 68.1 years; men, 61.4%), of whom 91 (16.9%) patients developed sHT with 25.3% (23/91) being symptomatic sHT and 62.6% (57/91) being PH. Five variables (atrial fibrillation, NIHSS score ≥ 10, hypodensity > 1/3 of middle cerebral artery territory, hyperdense artery sign, and anterior circulation infarction) composed the SHAIS score, which ranged from 0 to 11 points. The area under the receiver-operating characteristic curve (AUC) was 0.86 (95% CI 0.82–0.91, p < 0.001) for the overall sHT, 0.85 (95% CI 0.76–0.92, p < 0.001) for symptomatic sHT, and 0.89 (95% CI 0.85–0.94, p < 0.001) for PH. No evidence of miscalibration of the SHAIS score was found to predict the overall sHT (p = 0.19), symptomatic sHT (p = 0.44), and PH (p = 0.22). The internal (n = 245) and external validation cohorts (n = 200) depicted similar predictive performance compared to the derivation cohort. The SHAIS score had a higher AUC to predict sHT than any of the six pre-Existing models (p < 0.05).

Conclusions: The SHAIS score provides an easy-to-use model to predict sHT, which could help providers with decision-making about treatments with high bleeding risk, and to counsel patients and families on the baseline risk of HT, aligning expectations with probable outcomes.

Hemorrhagic Transformation After Ischemic Stroke: Mechanisms and Management

Symptomatic hemorrhagic transformation (HT) is one of the complications most likely to lead to death in patients with acute ischemic stroke. HT after acute ischemic stroke is diagnosed when certain areas of cerebral infarction appear as cerebral hemorrhage on radiological images. Its mechanisms are usually explained by disruption of the blood-brain barrier and reperfusion injury that causes leakage of peripheral blood cells. In ischemic infarction, HT may be a natural progression of acute ischemic stroke and can be facilitated or enhanced by reperfusion therapy. Therefore, to balance risks and benefits, HT occurrence in acute stroke settings is an important factor to be considered by physicians to determine whether recanalization therapy should be performed. This review aims to illustrate the pathophysiological mechanisms of HT, outline most HT-related factors after reperfusion therapy, and describe prevention strategies for the occurrence and enlargement of HT, such as blood pressure control. Finally, we propose a promising therapeutic approach based on biological research studies that would help clinicians treat such catastrophic complications.

Prediction of Clinical Outcome in Patients with Large-Vessel Acute Ischemic Stroke: Performance of Machine Learning versus SPAN-100

BACKGROUND AND PURPOSE

Traditional statistical models and pretreatment scoring systems have been used to predict the outcome for acute ischemic stroke patients (AIS). Our aim was to select the most relevant features in terms of outcome prediction on the basis of machine learning algorithms for patients with acute ischemic stroke and to compare the performance between multiple models and the Stroke Prognostication Using Age and National Institutes of Health Stroke Scale (SPAN-100) index model.

MATERIALS AND METHODS

A retrospective multicenter cohort of 1431 patients with acute ischemic stroke was subdivided into recanalized and nonrecanalized patients. Extreme Gradient Boosting machine learning models were built to predict the mRS score at 90 days using clinical, imaging, combined, and best-performing features. Feature selection was performed using the relative weight and frequency of occurrence in the models. The model with the best performance was compared with the SPAN-100 index model using area under the receiver operating curve analysis.

RESULTS

In 3 groups of patients, the baseline NIHSS was the most significant predictor of outcome among all the parameters, with relative weights of 0.36∼0.69; ischemic core volume on CTP ranked as the most important imaging biomarker with relative weights of 0.29∼0.47. The model with the best-performing features had a better performance than the other machine learning models. The area under the curve of the model with the best-performing features was higher than SPAN-100 model and reached statistical significance for the total (P < .05) and the nonrecanalized patients (P < .001).

CONCLUSIONS

Machine learning-based feature selection can identify parameters with higher performance in outcome prediction. Machine learning models with the best-performing features, especially advanced CTP data, had superior performance of the recovery outcome prediction for patients with stroke at admission in comparison with SPAN-100.

A Nomogram to Predict Symptomatic Intracranial Hemorrhage After Intravenous Thrombolysis in Chinese Patients

BACKGROUND AND AIMS

A reliable predictive score system to identify the risk of symptomatic intracranial hemorrhage (sICH) after intravenous thrombolysis (IVT) in acute ischemic stroke patients is of great essence. We aimed to develop a nomogram for predicting the risk of sICH after IVT in Chinese patients.

METHODS

We recruited acute ischemic stroke patients who were treated with IVT from five advanced stroke centers in China from April 2014 to November 2020. sICH was diagnosed according to the European Cooperative Acute Stroke Study II (ECASS-II) definition. Multivariable logistic regression was performed to construct the best-fit nomogram. The discrimination and calibration of the nomogram were evaluated by the area under the receiver operating characteristic curve (AUC-ROC) and calibration plot.

RESULTS

A total of 1200 patients were enrolled, of whom 66 (5.5%) developed sICH. In the multivariate logistic regression model, atrial fibrillation (odds ratio [OR] 3.25; 95% confidence interval [CI], 1.89-5.60; P < 0.001), baseline glucose level (OR, 1.13; 95% CI, 1.07-1.20; P < 0.001), neutrophil to lymphocyte ratio (OR, 1.05; 95% CI, 1.01-1.09; P = 0.024) and baseline National Institute of Health Stroke Scale (NIHSS) (OR, 1.07; 95% CI, 1.04-1.10; P < 0.001) were independent predictors for sICH and were used to generate the nomogram. The nomogram demonstrated good discrimination as the AUC-ROC value was 0.788 (95% CI, 0.737-0.840). The calibration plot revealed good calibration.

CONCLUSION

The nomogram consisted of atrial fibrillation, baseline glucose level, neutrophil to lymphocyte ratio, and NIHSS score may predict the risk of sICH in Chinese acute ischemic stroke patients treated with IVT.

A nomogram to predict unfavourable outcome in patients receiving oral anticoagulants for atrial fibrillation after stroke

Introduction

It is unknown whether the type of treatment (direct oral anticoagulant versus vitamin K antagonist) and the time of treatment introduction (early versus late) may affect the functional outcome in stroke patients with atrial fibrillation. We aimed to develop and validate a nomogram model including direct oral anticoagulant/vitamin K antagonist and early/late oral anticoagulant introduction for predicting the probability of unfavourable outcome after stroke in atrial fibrillation-patients.

Patients and Methods

We conducted an individual patient data analysis of four prospective studies. Unfavourable functional outcome was defined as three-month modified Rankin Scale score 3 -6. To generate the nomogram, five independent predictors including age (<65 years, reference; 65--79; or 80), National Institutes of Health Stroke Scale score (0--5 points, reference; 6--15; 16--25; or >25), acute revascularisation treatments (yes, reference, or no), direct oral anticoagulant (reference) or vitamin K antagonist, and early (7 days, reference) or late (8--30) anticoagulant introduction entered into a final logistic regression model. The discriminative performance of the model was assessed by using the area under the receiver operating characteristic curve.

Results

A total of 2102 patients with complete data for generating the nomogram was randomly dichotomised into training (n = 1553) and test (n = 549) sets. The area under the receiver operating characteristic curve was 0.822 (95% confidence interval, CI: 0.800--0.844) in the training set and 0.803 (95% CI: 0.764--0.842) in the test set. The model was adequately calibrated (9.852; p = 0.276 for the Hosmer--Lemeshow test).

Discussion and Conclusion

Our nomogram is the first model including type of oral anticoagulant and time of treatment introduction to predict the probability of three-month unfavourable outcome in a large multicentre cohort of stroke patients with atrial fibrillation.

Artificial neural network based prediction of postthrombolysis intracerebral hemorrhage and death

Despite the salient benefits of the intravenous tissue plasminogen activator (tPA), symptomatic intracerebral hemorrhage (sICH) remains a frequent complication and constitutes a major concern when treating acute ischemic stroke (AIS). This study explored the use of artificial neural network (ANN)-based models to predict sICH and 3-month mortality for patients with AIS receiving tPA. We developed ANN models based on evaluation of the predictive value of pre-treatment parameters associated with sICH and mortality in a cohort of 331 patients between 2009 and 2018. The ANN models were generated using eight clinical inputs and two outputs. The generalizability of the model was validated using fivefold cross-validation. The performance of each model was assessed according to the accuracy, precision, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). After adequate training, the ANN predictive model AUC for sICH was 0.941, with accuracy, sensitivity, and specificity of 91.0%, 85.7%, and 92.5%, respectively. The predictive model AUC for 3-month mortality was 0.976, with accuracy, sensitivity, and specificity of 95.2%, 94.4%, and 95.5%, respectively. The generated ANN-based models exhibited high predictive performance and reliability for predicting sICH and 3-month mortality after thrombolysis; thus, its clinical application to assist decision-making when administering tPA is envisaged.

A new nomogram for individualized prediction of the probability of hemorrhagic transformation after intravenous thrombolysis for ischemic stroke patients

Background

A reliable scoring tool to detect the risk of intracerebral hemorrhage (ICH) after intravenous thrombolysis for ischemic stroke is warranted. The present study was designed to develop and validate a new nomogram for individualized prediction of the probability of hemorrhagic transformation (HT) in patients treated with intravenous (IV) recombinant tissue plasminogen activator (rt-PA).

Methods

We enrolled patients who suffered from acute ischemic stroke (AIS) with IV rt-PA treatment in our emergency green channel between August 2016 and July 2018. The main outcome was defined as any type of intracerebral hemorrhage according to the European Cooperative Acute Stroke Study II (ECASS II). All patients were randomly divided into two cohorts: the primary cohort and the validation cohort. On the basis of multivariate logistic model, the predictive nomogram was generated. The performance of the nomogram was evaluated by Harrell’s concordance index (C-index) and calibration plot.

Results

A total of 194 patients with complete data were enrolled, of whom 131 comprised the primary cohort and 63 comprised the validation cohort, with HT rate 12.2, 9.5% respectively. The score of chronic disease scale (CDS), the global burden of cerebral small vascular disease (CSVD), National Institutes of Health Stroke Scale (NIHSS) score ≥ 13, and onset-to-treatment time (OTT) ≥ 180 were detected important determinants of ICH and included to construct the nomogram. The nomogram derived from the primary cohort for HT had C- Statistics of 0.9562 and the calibration plot revealed generally fit in predicting the risk of HT. Furthermore, we made a comparison between our new nomogram and several other risk-assessed scales for HT with receiver operating characteristic (ROC) curve analysis, and the results showed the nomogram model gave an area under curve of 0.9562 (95%CI, 0.9221–0.9904, P < 0.01) greater than HAT (Hemorrhage After Thrombolysis), SEDAN (blood Sugar, Early infarct and hyper Dense cerebral artery sign on non-contrast computed tomography, Age, and NIHSS) and SPAN-100 (Stroke Prognostication using Age and NIHSS) scores.

Conclusions

This proposed nomogram based on the score of CDS, the global burden of CSVD, NIHSS score ≥ 13, and OTT ≥ 180 gives rise to a more accurate and more comprehensive prediction for HT in patients with ischemic stroke receiving IV rt-PA treatment.

Supplementary Information

Supplementary information accompanies this paper at 10.1186/s12883-020-02002-w.

Predicting 6-Month Unfavorable Outcome of Acute Ischemic Stroke Using Machine Learning

Background and Purpose: Accurate prediction of functional outcome after stroke would provide evidence for reasonable post-stroke management. This study aimed to develop a machine learning-based prediction model for 6-month unfavorable functional outcome in Chinese acute ischemic stroke (AIS) patient. Methods: We collected AIS patients at National Advanced Stroke Center of Nanjing First Hospital (China) between September 2016 and March 2019. The unfavorable outcome was defined as modified Rankin Scale score (mRS) 3-6 at 6-month. We developed five machine-learning models (logistic regression, support vector machine, random forest classifier, extreme gradient boosting, and fully-connected deep neural network) and assessed the discriminative performance by the area under the receiver-operating characteristic curve. We also compared them to the Houston Intra-arterial Recanalization Therapy (HIAT) score, the Totaled Health Risks in Vascular Events (THRIVE) score, and the NADE nomogram. Results: A total of 1,735 patients were included into this study, and 541 (31.2%) of them had unfavorable outcomes. Incorporating age, National Institutes of Health Stroke Scale score at admission, premorbid mRS, fasting blood glucose, and creatinine, there were similar predictive performance between our machine-learning models, while they are significantly better than HIAT score, THRIVE score, and NADE nomogram. Conclusions: Compared with the HIAT score, the THRIVE score, and the NADE nomogram, the RFC model can improve the prediction of 6-month outcome in Chinese AIS patients.

Early prediction of the 3-month outcome for individual acute ischemic stroke patients who received intravenous thrombolysis using the N2H3 nomogram model

Background:

The aim of this study was to establish a nomogram model for individualized early prediction of the 3-month prognosis in patients with acute ischemic stroke (AIS) who were treated with intravenous recombinant tissue plasminogen activator (rt-PA) thrombolysis.

Methods:

A total of 691 patients were included in this study; 564 patients were included in the training cohort, while 127 patients were included in the test cohort. The main outcome measure was a 3-month unfavorable outcome (modified Rankin Scale 3–6). To construct the nomogram model, stepwise logistic regression analysis was applied to select the significant predictors of the outcome. The discriminative performance of the model was assessed by calculating the area under the receiver operating characteristic curve (AUC-ROC). A decision curve analysis was used to evaluate prognostic value of the model.

Results:

The initial National Institutes of Health Stroke Scale [NIHSS, odds ratio (OR), 1.35; 95% confidence interval (CI), 1.28–1.44; p < 0.001], delta NIHSS (changes in the NIHSS score from baseline to 24 h, OR, 0.75; 95% CI, 0.70–0.79; p < 0.001), hypertension (OR, 2.07; 95% CI, 1.32–3.31; p = 0.002), hyperhomocysteinemia (Hhcy, OR, 2.18; 95% CI, 1.20–4.11; p = 0.013), and the ratio of high-density lipoprotein cholesterol (HDL-C) to low-density lipoprotein cholesterol (LDL-C) (HDL-C/LDL-C, OR, 3.29; 95% CI, 1.00–10.89; p = 0.049) (N2H3) were found to be independent predictors of a 3-month unfavorable outcome from multivariate logistic regression analysis and were incorporated in the N2H3 nomogram model. The AUC-ROC of the training cohort was 0.872 (95% CI, 0.841–0.902), and the AUC-ROC of the test cohort was 0.900 (95% CI, 0.848–0.953).

Conclusion:

The study presented the N2H3 nomogram model, with initial NIHSS score, delta NIHSS, hypertension, Hhcy, and HDL-C/LDL-C as predictors. It therefore provides an individualized early prediction of the 3-month unfavorable outcome in AIS patients treated with intravenous rt-PA thrombolysis.

Predicting hemorrhagic transformation in patients not submitted to reperfusion therapies

BACKGROUND

Well studied in patients with ischemic stroke after reperfusion therapies (RT), hemorrhagic transformation (HT) is also common in patients not treated with RT and can lead to disability even in initially asymptomatic cases. The best predictors of HT in patients not treated with RT are not well established. Therefore, we aimed to identify predictors of HT in patients not submitted to RT and create a user-friendly predictive score (PROpHET).

MATERIAL AND METHODS

Patients admitted to a Comprehensive Stroke Center from 2015 to 2017 were prospectively evaluated and randomly selected to the derivation cohort. A multivariable logistic regression modeling was built to produce a predictive grading score for HT. The external validation was assessed using datasets from 7 Comprehensive Stroke Centers using the area under the receiver operating characteristic curve (AUROC).

RESULTS

In the derivation group, 448 patients were included in the final analysis. The validation group included 2,683 patients. The score derived from significant predictors of HT in the multivariate logistic regression analysis was male sex (1 point), ASPECTS ≤ 7 (2 points), presence of leukoaraiosis (1 point), hyperdense cerebral middle artery sign (1 point), glycemia at admission ≥180 mg/dL (1 point), cardioembolism (1 point) and lacunar syndrome (-3 points) as a protective factor. The grading score ranges from -3 to 7. A Score ≥3 had 78.2% sensitivity and 75% specificity, and AUROC of 0.82 for all cases of HT. In the validation cohort, our score had an AUROC of 0.83.

CONCLUSIONS

The PROpHET is a simple, quick, cost-free, and easy-to-perform tool that allows risk stratification of HT in patients not submitted to RT. A cost-free computerized version of our score is available online with a user-friendly interface.

External Validation of the WORSEN Score for Prediction the Deterioration of Acute Ischemic Stroke in a Chinese Population

Background: Early neurological deterioration (END) has been recognized as a serious neurological complication after acute ischemic stroke. However, to date, the WORSEN score was the only one scoring system specifically developed to detect END events in acute ischemic stroke patients. The purpose of this study was to investigate the WORSEN score's utility in China, and to determine the potential predictors of END in acute stroke patients. Methods: Consecutive patients with acute ischemic stroke admitted to the Department of Neurology, Aerospace Center Hospital between March 2015 to February 2017 were recruited into the study's cohort and divided into two groups: patients with and without END. END was defined as either an increase in two or more NIHSS points, an increment of at least one point in motor power or a description of fluctuating of clinical symptoms in medical reports during the first 7 days after admission. Severe END was defined as an increase of NIHSS ≥ 4 points from baseline during the first 7 days after admission. Results: Three hundred fifty four patients with acute ischemic stroke were enrolled in the present study and 67.5% were male. END occurred in 90 of these patients and severe END occurred in 55 of these patients. Logistic regression analysis showed that an initial NIHSS score ≥8, diameter of infarction, striatocapsular infarction, and TOAST type of large arterial atherosclerosis were independent predictors for END. The area under the ROC curve (AUC) of the WORSEN score for the prediction of END was 0.80 (95%CI 0.75-0.84), with a sensitivity of 62.22%, a specificity of 88.26%, positive predictive values of 64.37% and negative predictive values of 87.27%. Meanwhile, the AUC of the WORSEN score for the prediction of severe END was 0.82 (95%CI 0.78-0.86), with a sensitivity of 70.91%, specificity of 83.95%, positive predictive values of 44.83% and negative predictive values of 94.01%. Conclusion: END is a relatively common neurological complication in patients with acute ischemic stroke. Our findings showed that the WORSEN score had a good predictive value for identifying patients with END in a Chinese population. Moving forward, multi-center studies are required for further validations.

Development and Validation of a Prognostic Tool for Direct Enteral Tube Insertion After Acute Stroke

Background and Purpose- We aimed to create a novel prognostic risk score to estimate outcomes after direct enteral tube placement in acute stroke. Methods- We used the Ontario Stroke Registry and linked databases to obtain clinical information on all patients with direct enteral tube insertion after ischemic stroke or intracerebral hemorrhage from July 1, 2003 to June 30, 2010 (derivation cohort) and July 1, 2010 to March 31, 2013 (validation cohort). We used multivariable regression to assign scores to predictor variables for 3 outcomes after tube placement: favorable outcome (discharge modified Rankin Scale score 0-3 and alive at 90 days), poor outcome (discharge modified Rankin Scale score 5 or death at 90 days), and 30-day mortality. Results- Variables associated with a favorable outcome were younger age, preadmission independence, ischemic stroke rather than intracerebral hemorrhage, lower stroke severity, and a shorter time between stroke and tube placement. Variables associated with a poor outcome were older age, preadmission dependence, atrial fibrillation, greater stroke severity, and tracheostomy. Age, preadmission dependence, atrial fibrillation, cancer, chronic obstructive pulmonary disease, and shorter time to tube placement were associated with increased 30-day mortality. Using these variables, we created an online calculator to facilitate estimation of individual patient risk of favorable and poor outcomes. C-statistic in the validation cohort was 0.82 for favorable outcome, 0.65 for poor outcome, and 0.62 for 30-day mortality, and calibration was adequate. Conclusions- We developed risk scores to estimate outcomes after direct enteral tube insertion for acute dysphagic stroke. This information may be useful in discussions with patients and families when there is prognostic uncertainty surrounding outcomes with direct enteral tube placement after stroke.

Predicting Functional Outcome Based on Linked Data After Acute Ischemic Stroke: S-SMART Score

Prediction of outcome after stroke may help clinicians provide effective management and plan long-term care. We aimed to develop and validate a score for predicting good functional outcome available for hospitals after ischemic stroke using linked data. A total of 22,005 patients with acute ischemic stroke from the Clinical Research Center for Stroke Registry between July 2007 and December 2014 were included in the derivation group. We assessed functional outcomes using a modified Rankin scale (mRS) score at 3 months after ischemic stroke. We identified predictors related to good 3-month outcome (mRS score ≤ 2) and developed a score. External validations (geographic and temporal validations) of the developed model were performed. The prediction model performance was assessed using the area under the receiver operating characteristic curve (AUC) and the calibration test. Stroke severity, sex, stroke mechanism, age, pre-stroke mRS, and thrombolysis/thrombectomy treatment were identified as predictors for 3-month good functional outcomes in the S-SMART score (total 34 points). Patients with higher S-SMART scores had an increased likelihood of a good outcome. The AUC of the prediction score was 0.805 (0.798-0.811) in the derivation group and 0.812 (0.795-0.830) in the geographic validation group for good functional outcome. The AUC of the model was 0.812 (0.771-0.854) for the temporal validation group. Moreover, they had good calibration. The S-SMART score is a valid and useful tool to predict good functional outcome following ischemic stroke. This prediction model may assist in the estimation of outcomes to determine care plans after stroke.

Radiological predictors of hemorrhagic transformation after acute ischemic stroke: An evidence-based analysis

Hemorrhagic transformation (HT) is one of the most common adverse events related to acute ischemic stroke (AIS) that affects the treatment plan and clinical outcome. Identification of a sensitive radiological marker may influence the controversial thrombolytic decision in the setting of AIS and may at a minimum indicate more intensive monitoring or further prophylactic interventions. In this article we summarize possible radiological biomarkers and the role of different radiological modalities including computed tomography (CT), magnetic resonance imaging, angiography, and ultrasound in predicting HT. Different radiological indices of early ischemic changes, large ischemic lesion volume, severe blood flow restriction, blood-brain barrier disruption, poor collaterals and high blood flow velocities have been reported to be associated with higher risk of HT. The current levels of evidence of the available studies highlight the role of the different CT perfusion parameters in predicting HT. Further large standardized studies are recommended to compare the sensitivity and specificity of the different radiological markers combined and delineate the most reliable predictor.

Deep Learning in the Prediction of Ischaemic Stroke Thrombolysis Functional Outcomes: A Pilot Study

RATIONALE AND OBJECTIVES

Intravenous thrombolysis decision-making and obtaining of consent would be assisted by an individualized risk-benefit ratio. Deep learning (DL) models may be able to assist with this patient selection.

MATERIALS AND METHODS

Clinical data regarding consecutive patients who received intravenous thrombolysis across two tertiary hospitals over a 7-year period were extracted from existing databases. The noncontrast computed tomography brain scans for these patients were then retrieved with hospital picture archiving and communication systems. Using a combination of convolutional neural networks (CNN) and artificial neural networks (ANN) several models were developed to predict either improvement in the National Institutes of Health Stroke Scale of ≥4 points at 24 hours ("NIHSS24"), or modified Rankin Scale 0-1 at 90 days ("mRS90"). The developed CNN and ANN were then applied to a test set. The THRIVE, HIAT, and SPAN-100 scores were also calculated for the patients in the test set and used to predict NIHSS24 and mRS90.

RESULTS

Data from 204 individuals were included in the project. The best performing DL model for prediction of mRS90 was a combination CNN + ANN based on clinical data and computed tomography brain (accuracy = 0.74, F1 score = 0.69). The best performing model for NIHSS24 prediction was also the combination CNN + ANN (accuracy = 0.71, F1 score = 0.74).

CONCLUSION

DL models may aid in the prediction of functional thrombolysis outcomes. Further investigation with larger datasets and additional imaging sequences is indicated.

Contralateral Hemispheric Cerebral Blood Flow Measured With Arterial Spin Labeling Can Predict Outcome in Acute Stroke

Background and Purpose- Imaging is frequently used to select acute stroke patients for intra-arterial therapy. Quantitative cerebral blood flow can be measured noninvasively with arterial spin labeling magnetic resonance imaging. Cerebral blood flow levels in the contralateral (unaffected) hemisphere may affect capacity for collateral flow and patient outcome. The goal of this study was to determine whether higher contralateral cerebral blood flow (cCBF) in acute stroke identifies patients with better 90-day functional outcome. Methods- Patients were part of the prospective, multicenter iCAS study (Imaging Collaterals in Acute Stroke) between 2013 and 2017. Consecutive patients were enrolled after being diagnosed with anterior circulation acute ischemic stroke. Inclusion criteria were ischemic anterior circulation stroke, baseline National Institutes of Health Stroke Scale score ≥1, prestroke modified Rankin Scale score ≤2, onset-to-imaging time <24 hours, with imaging including diffusion-weighted imaging and arterial spin labeling. Patients were dichotomized into high and low cCBF groups based on median cCBF. Outcomes were assessed by day-1 and day-5 National Institutes of Health Stroke Scale; and day-30 and day-90 modified Rankin Scale. Multivariable logistic regression was used to test whether cCBF predicted good neurological outcome (modified Rankin Scale score, 0-2) at 90 days. Results- Seventy-seven patients (41 women) met the inclusion criteria with median (interquartile range) age of 66 (55-76) yrs, onset-to-imaging time of 4.8 (3.6-7.7) hours, and baseline National Institutes of Health Stroke Scale score of 13 (9-20). Median cCBF was 38.9 (31.2-44.5) mL per 100 g/min. Higher cCBF predicted good outcome at day 90 (odds ratio, 4.6 [95% CI, 1.4-14.7]; P=0.01), after controlling for baseline National Institutes of Health Stroke Scale, diffusion-weighted imaging lesion volume, and intra-arterial therapy. Conclusions- Higher quantitative cCBF at baseline is a significant predictor of good neurological outcome at day 90. cCBF levels may inform decisions regarding stroke triage, treatment of acute stroke, and general outcome prognosis. Clinical Trial Registration- URL: https://www.clinicaltrials.gov. Unique identifier: NCT02225730.