Development and Validation of Multiple Machine Learning Models Integrating Neutrophil‐Lymphocyte Ratio for Prediction of Hemorrhagic Transformation After Intravenous Thrombolysis in Acute Ischemic Stroke

To Treat or Not to Treat: Low‐Dose Alteplase for Patients With Acute Ischemic Stroke Taking Direct Oral Anticoagulants Within 24 Hours

Commentary

Background: The presence of effective collateral blood flow patterns may influence response to intravenously administered tissue plasminogen activator (IV-tPA) in acute ischemic stroke (AIS). We compared various existing methods of scoring collaterals on the pre-treatment computed tomographic angiogram (CTA) of the brain for a reliable prediction of functional outcome in AIS patients. Methods: Consecutive AIS patients treated with IV-tPA within 4.5 hours of symptom-onset during 2007-2011 were included. Data were collected for demographics, vascular risk factors, National Institute of Health Stroke Scale (NIHSS) scores and stroke subtypes. Intracranial collaterals were evaluated by 2 independent blinded neuroradiologists via 4 predefined criteria- Miteff’s system that grades middle cerebral artery (MCA) collateral branches with respect to the sylvian fissure; Maas system that compares collaterals on the affected hemisphere against the unaffected side; modified Tan’s scale where collaterals in 50% or more of the MCA territory are classified as good; and a 20-point collateral grading scale in regions corresponding to Alberta Stroke Program Early CT score (ASPECTS) methodology. Good functional outcomes at 3-months were determined by modified Rankin scale (mRS) scores of 0-1. Results: CTA was performed in 115 patients with anterior circulation AIS before IV-tPA bolus. Median age 66yrs (range 35-92), 42% males, median NIHSS 19 points (range 3-30) and median onset-to-treatment time 155 minutes. Overall, 52 (45.2%) patients achieved good functional outcome at 3-months. Univariable analysis revealed younger age, absence of diabetes, lower pre-tPA NIHSS scores and good collaterals according to ASPECTS methodology as significantly associated with good functional outcomes. On multivariable logistic regression, only lower NIHSS (OR 1.111 per NIHSS point; 95% CI 1.023-1.206, p=0.013) and good collaterals by ASPECTS methodology (OR 1.117 per point; 95%CI 1.006-1.241, p=0.039) were found as independent predictors of good outcomes. Conclusion: Of the existing intracranial collaterals scoring systems, only the ASPECTS methodology serves as a reliable predictor of favorable outcomes at 3-months in patients with anterior circulation AIS.

IntroductionAcute ischemic stroke (AIS) patients often experience poor functional outcomes post-intravenous thrombolysis (IVT). Novel computational methods leveraging machine learning (ML) architectures increasingly support medical decision-making. We aimed to develop and validate a machine learning model to predict 3-month unfavorable functional outcome after IVT in AIS patients.MethodsThis retrospective study developed ML prognostic models for 3-month functional outcome (modified Rankin scale scores of 3–6) in IVT-treated AIS patients. A derivation cohort (n = 938) was split 7:3 for training/testing, with an independent external validation cohort (n = 324). The least absolute shrinkage and selection operator (LASSO) regression selected predictors from clinical/neuroimaging/laboratory variables. Eight ML algorithms (including Logistic Regression, Random Forest, Extreme Gradient Boosting, Multilayer Perceptron, Support Vector Machine, Light Gradient Boosting Machine, Decision Tree, and K-Nearest Neighbors) were trained using 10-fold cross-validation and evaluated on test/external sets via the area under the curve (AUC), accuracy, precision, recall and F1-score. Additionally, the SHapley Additive exPlanations (SHAP) interpreted the optimal model.Results938 patients constituted the derivation cohort (training: n = 656, test: n = 282) and 324 patients the external validation cohort. Unfavorable 3-month outcomes (mRS 3–6) occurred in 25.7% and 22.8%, respectively. LASSO regression selected five predictors: the neutrophil-to-lymphocyte ratio (NLR), admission National Institutes of Health Stroke Scale (NIHSS) score, the Alberta Stroke Program Early CT Score (ASPECTS), atrial fibrillation, and blood glucose. While tree-based methods like XGBoost and LightGBM showed elevated training performance (e.g., XGBoost training AUC = 0.878) but significant drops in validation (AUC = 0.791), LR demonstrated optimal performance: robust training AUC (0.792), minimal validation degradation (AUC = 0.787). LR model was subsequently employed as classification method demonstrating optimal performance with (AUC = 0.777) in the test dataset. External validation confirmed LR’s stability (AUC = 0.797). SHAP analysis ranked NLR as the strongest predictor (followed by NIHSS/ASPECTS), with higher values increasing risk. Learning curves indicated no overfitting. A nomogram enabled individualized risk quantification.ConclusionA parsimonious 5-variable LR model robustly predicts 3-month post-IVT outcomes, combining clinical utility, interpretability, and generalizability. NLR-driven inflammation is critical to prognosis. This tool facilitates early high-risk patient identification for personalized intervention.

Background: ASPECTS was developed for the semi-quantitative assessment of early ischemic changes (EIC) on non-contrast computed tomography (NCCT) in acute ischemic stroke (AIS). Artificial intelligence (AI)-based automated tools for the ASPECT scoring system were developed to automate the diagnosis and improve the agreement with radiologists of AIS. The performance of the automated software compared to physicians should be tested before the software is further used in clinical practice as a tool for clinicians. Objective: To evaluate the agreement with radiologists of an AI-based automated post-processing software for detecting EIC and calculating ASPECTS on NCCT images in AIS patients using a radiologist's assessment as a reference. Materials and Methods: NCCT of AIS patients were retrospectively reviewed (Stroke Fast Track Service July 2022 - December 2023). The complete set of clinical data and imaging data from both baseline and follow-up were analyzed by a radiologist as a reference. Two additional observers provided individual ASPECTS from the baseline NCCT only (observer 1 was a radiologist who independently reviewed only the baseline NCCT with stroke window setting. Observer 2 was a radiologist on service which was from the pool of 20 radiologists onsite and online). Recon&GO Inline ASPECTS software (Somaris X, VA40A, Siemens Healthineers AG, Erlangen, Germany) was applied. Both ASPECT score analysis and ASPECTS region analysis were evaluated. Positive percent agreement (PPA) and negative percent agreement (NPA) were calculated. Interobserver agreement was assessed using the Cohen's kappa coefficient and the intraclass correlation coefficient (ICC). Results: 111 patients with a mean age of 67.8 years (±11.9), 56 (50.5%) females, a mean National Institute of Health Stroke Scale (NIHSS) score of 14.2 (±8.8), and a mean time to baseline NCCT of 123.9 minutes (±58.7) were included. For dichotomized ASPECTS, the automated software showed lower PPA (14.6% vs. 27.1%) but higher NPA (100.0% vs. 93.7%) than observer 2. For the region-based analysis, both the automated software and observer 2 differed in terms of regional contribution. The automated software showed low PPA but rather high NPA with perfect (100%) NPA in lentiform nucleus and M2. The automated software showed higher agreement with the reference and two observers in deep/central regions than cortical regions. For total ASPECTS, the automated software showed a moderate agreement of total ASPECTS with the reference and observer 1 (ICC = 0.545 and 0.545). Observer 2 showed a poor agreement of total ASPECTS with the reference, observer 1, and the automated software (ICC = 0.349, 0.422, and 0.301, respectively). Conclusion: For total ASPECT score, the agreement of the tested AI software is lower compared to observer 1 obtained by a radiologist using the stroke window on NCCT, but better compared to a pool of radiologists on service with a time limit of 30 minutes to interpret the ASPECT score. When analyzing the ASPECTS regions, there are different advantages for the assessment of the deep regions and the cortical regions. The tested AI software shows higher agreement in deep/central regions than cortical regions. From the result, the tested AI software retains its potential for use in emergency situations, particularly for radiologists with limited experience and limited time to report.

The Alberta Stroke Program Early CT Score (ASPECTS) scale was developed for monitoring early ischemic changes on CT, being associated with clinical outcomes. The ASPECTS can also associate with peripheral biomarkers that reflect the pathophysiological response of the brain to the ischemic stroke. To investigate the association between peripheral biomarkers with the Alberta Stroke Program Early CT Score (ASPECTS) in individuals after ischemic stroke. Patients over 18 years old with acute ischemic stroke were enrolled in this study. No patient was eligible for thrombolysis. The patients were submitted to non-contrast CT in the first 24 hours of admission, being the Alberta Stroke Program Early CT Score and clinical and molecular evaluations applied on the same day. The National Institutes of Health Stroke Scale (NIHSS), modified Rankin scale and the Mini-Mental State Examination for clinical evaluation were also applied to all subjects. Plasma levels of BDNF, VCAM-1, VEGF, IL-1β, sTNFRs and adiponectin were determined by ELISA. Worse neurological impairment (NIHSS), cognitive (MEEM) and functional (Rankin) performance was observed in the group with changes in the NCTT. Patients with NCTT changes also exhibited higher levels of IL-1β and adiponectin. In the linear multivariate regression, an adjusted R coefficient of 0.515 was found, indicating adiponectin and NIHSS as independent predictors of ASPECTS. Plasma levels of adiponectin are associated with the ASPECTS scores.

Background: Various collateral channels are recruited to provide alternative pathways in acute ischemic stroke (AIS), however the relationship with patient outcomes remain unclear. We compared various existing methods of scoring collaterals on the pre-treatment and day-2 computed tomographic angiogram (CTA) of the brain in thrombolyzed AIS patients. Methods: We included 115 consecutive patients in whom CTA was performed both pre-tPA and on day-2. Intracranial collaterals were evaluated by 2 independent neuroradiologists using 4 existing and one modified method- Miteff’s system (grades middle cerebral artery (MCA) collateral branches with respect to sylvian fissure); Maas system (compares collaterals in affected hemisphere against the contralatral side); Modified Tan’s scale (collaterals in 50% or more of MCA territory classified as good); and 20-point collateral grading scale by Alberta Stroke Program Early CT score (ASPECTS) methodology. For the modified scoring system we adapted ASPECTS methodology into a 14 point score for cortical and internal cerebral veins (ICV) and removing basal ganglia area from scoring. Symptomatic intracranial hemorrhage (SICH) was defined by new bleeding on the CT scan and an increase in NIH stroke scale (NIHSS) by 4 points or more. Results: On univariate analysis collateral recruitment via the Tan scoring system, ASPECTS method (improvement of ≥6 points), modified scoring system (improvement ≥7 points), hypertension and higher NIHSS score were associated with SICH. On multivariate analysis only collateral recruitment on the Tan scoring system (OR 3.286 95% CI 1.014-11.025, p =0.049), Collateral recruitment on ASPECTS ≥6 points (OR 2.839 95% CI 1.064- 7.576, p = 0.037) and collateral recruitment on the modified scoring system ≥ 7 (OR 4.174 95% CI 1.212-14.372, p = 0.023) were independent predictors of SICH. Interestingly, collateral failure on the day-2 CTA did not show any association with SICH. Conclusion: Large recruitment of the collateral channels on the day-2 CTA is strongly associated with SICH after thrombolysis . Perhaps, an unregulated cerebral hyperperfusion contributed to SICH and close monitoring along with aggressive blood pressure control might prevent complications.

Extensive ischemic change on stroke presentation is uncommon: A population-based study

BackgroundEarly neurological deterioration (END), defined as neurological decline after endovascular thrombectomy (EVT) for acute ischemic stroke (AIS), is a critical complication associated with poor long-term prognosis. Although END significantly impacts clinical outcomes, its underlying mechanisms and modifiable risk factors remain poorly understood. This study aimed to (1) characterize the clinical profile of END and (2) evaluate its association with 90-day functional outcomes in AIS patients who underwent EVT.MethodsThis retrospective study included consecutive patients with acute ischemic stroke (AIS) secondary to large vessel occlusion (LVO) who were treated with endovascular thrombectomy (EVT) between January 2021 and December 2024. END was defined as either a ≥4-point increase in the National Institutes of Health Stroke Scale (NIHSS) score or a ≥1-point increase in the NIHSS consciousness subscore (Item Ia) within 72 h after EVT. Patients meeting these criteria were classified into the END group; others comprised the non-END group. Variables with a p-value of <0.05 in univariate analysis were included in a multivariable logistic regression model to identify independent predictors of END. The predictive performance of the Alberta Stroke Program Early CT Score (ASPECTS) was assessed through receiver operating characteristic (ROC) analysis, with the optimal cutoff determined by maximizing the Youden index.ResultsThe study included 177 consecutive AIS patients with LVO who were treated with EVT. END occurred in 52 patients (29.4%). The END group had significantly lower baseline ASPECTS values than the non-END group (median [IQR]: 7 [6–8] vs. 9 [8–9], p < 0.001). In the adjusted analysis, higher ASPECTS was independently associated with a reduced risk of END (OR = 0.59 per 1-point increase, 95% CI: 0.44–0.80, p = 0.001). ROC analysis identified an ASPECTS ≤7 as the optimal cutoff for predicting END (AUC = 0.761; sensitivity 75.0%, specificity 76.8%). At 90 days, functional independence [modified Rankin Scale (mRS) 0–2] was achieved in 52.0% of the non-END group versus 7.6% of the END group (p < 0.001).ConclusionEND after EVT independently predicts poor 90-day functional independence (mRS 0–2) in LVO-related AIS. Our findings support incorporating a baseline ASPECTS ≤7 into risk stratification protocols to identify high-risk patients requiring intensive neuromonitoring (hourly NIHSS assessments for 72 h post-EVT). Early detection of END signs, such as a ≥ 4-point increase in the NIHSS score 4, may enable prompt interventions (e.g., blood pressure control or edema management) to improve long-term outcomes.

Interventional neuroradiology.

Background &amp; Objective: Computed tomography angiography (CTA) in acute stroke has been widely used to demonstrate arterial occlusion. Alberta Stroke Program Early CT Score (ASPECTS) is used to detect early ischemic signs in non-contrast computed tomography (NCCT) in the middle cerebral artery region. We hypothesized that computed tomography angiography source image (CTA-SI) is superior to NCCT in predicting final infarct volume, 24 hour National Institutes of Health Stroke Scale (NIHSS) score and 90-day clinical outcome. Methods: Patients who had an acute ischemic stroke due to middle cerebral artery (MCA) occlusion and treated with tissue plasminogen activator (tPA) were retrospectively evaluated. ASPECTS was evaluated by two experienced stroke neurologists in acute NCCT, CTA-SI, and follow up imaging. The final ASPECTS was compared with the mean baseline ASPECTS of NCCT and CTA-SI. The relation of both scores with 24-hour NIHSS and clinical outcome was compared. The Modified Rankin Scale (mRS) was utilized to evaluate the 90-day outcomes. mRS score of 0-2 was considered a “good outcome”. Results: Fifty-three patients were evaluated. We observed a significant relation among CTA-SI ASPECTS and after treatment 24hr ASPECTS (y= -3.9 + 1.4 x; 95% CI, -7.6 to -0.2) (y= -26.04 + 3.5 x; CI, -41 to -10). The median baseline 24-hr NHISS was 6 (0 - 22). We found a better correlation between CTA-SI ASPECTS and 24-hr NHISS (y= 363.06 + -37.03 x; CI, -148 to 864) than between NCCT ASPECTS and 24h NHISS (y=529.80 + -62.55 x; CI, 180 - 829). Median 90 days mRS score was 2 (0 - 6). According to Deming regression analysis, the CTA-SI ASPECTS (y= 76.10 + -7.69 x; 95% CI, -36 to 188) was more consistent with the 90 day mRS compared to NCCT ASPECTS (y=149.86 + -17.67 x; 95% CI, 23 - 267) CTA-SI was superior in predicting 24hr NIHSS and day 90 mRS compared to NCCT ASPECTS. Conclusion: Prediction of CTA-SI ASPECTs is better than NCCT ASPECTs at 24hr NIHSS, 3-month mRS and final infarct size in acute ischemic stroke patients treated with tPA.

The management of acute ischemic stroke has undergone a sea of change with the introduction of intravenous thrombolysis (IVT). Current guidelines state that the window period for IVT using rTPA is 4.5 hours. The MERCI, Multi Merci, and Penumbra trials in which patients with acute ischemic stroke were treated using endovascular treatment demonstrated better recanalisation in patients having a large vessel occlusion. However, recently published data from the three large trials IMS 3, Synthesis Expansion, and MR rescue, which compared endovascular treatment with intravenous therapy, failed to demonstrate superiority of endovascular treatment over IVT. In these trials, stent retrievers were used in very few patients. We present our results from a tertiary care center in India where patients are treated using intravenous as well as endovascular modalities. Among the 53 patients with acute ischemic stroke treated between 2010 and 2012, 23 were treated with IVT and 30 with endovascular methods. Stent retriever was used in majority of the endovascular cases.Aims:To compare the outcomes of acute ischemic stroke patients treated with IVT versus those who were managed using endovascular therapy. To evaluate outcomes of patients with acute ischemic stroke with a large vessel occlusion in whom endovascular modalities were used and to compare them with those of patients who were treated with IVT in presence of a large vessel occlusion.Settings and Design:Data of patients who underwent thrombolysis at our centre was collected over a 3-year period, that is, from 2010 to 2012. Endovascular treatment was done by an interventional neurologist.Materials and Methods:Data of patients with acute ischemic stroke who underwent IVT or endovascular treatment at our centre between 2010 and 2012 was analyzed. Parameters included age, National Institutes of Health Stroke Scale (NIHSS) on admission, door to needle time, stroke subtype, modality of treatment, outcome based on modified Rankin Scale (mRS) Score at 90 days follow up and mortality rates at 90 days.Statistical Analysis:Tabulated results were analysed using INSTAT Graphpad analyser. Data were analysed using paired and unpaired t-test, Chi-square test, and Fishers test as applicable. P value was considered significant when it was <0.05.Results:Upon comparison of the outcomes of patients with acute ischemic stroke and large vessel disease treated with endovascular therapy with those treated with IVT, it was found that the former group had better outcomes. We also found that in spite of there being a significant difference in the NIHSS on admission and a significant difference in the door to needle time, the outcomes of patients treated using intravenous or endovascular therapy were similar. There was no statistically significant difference in mortality rates between intravenous and endovascular groups.Conclusions:IVT is currently the standard of care in the management of acute ischemic stroke. Endovascular treatment during the window period is reserved for those patients with contraindication to IVT. In this study, we found that patients with documented large vessel disease with no evidence of cross flow through Willisian collaterals benefit from endovascular treatment. We recommend that all patients of acute ischemic stroke, be subjected to a baseline angiogram either computed tomography (CT) or magnetic resonance imaging (MRI) to document vessel status. This will help in identifying patients who may benefit from early endovascular treatment, if they fail to improve with IVT. Further, large trials using stent retrievers are needed, to prove that endovascular treatment is superior to IVT, in presence of documented large vessel disease.

PurposeMachine learning (ML) models were constructed according to non-contrast computed tomography (NCCT) images as well as clinical and laboratory information to assess risk stratification for the occurrence of hemorrhagic transformation (HT) in acute ischemic stroke (AIS) patients.MethodsA retrospective cohort was constructed with 180 AIS patients who were diagnosed at two centers between January 2019 and October 2023 and were followed for HT outcomes. Patients were analyzed for clinical risk factors for developing HT, infarct texture features were extracted from NCCT images, and the radiomics score (Rad-score) was calculated. Then, five ML models were established and evaluated, and the optimal ML algorithm was used to construct the clinical, radiomics, and clinical-radiomics models. Receiver operating characteristic (ROC) curves were used to compare the performance of the three models in predicting HT.ResultsBased on the outcomes of the AIS patients, 104 developed HT, and the remaining 76 had no HT. The HT group consisted of 27 hemorrhagic infarction (HI) and 77 parenchymal-hemorrhage (PH). Patients with HT had a greater neutrophil-to-lymphocyte ratio (NLR), baseline National Institutes of Health Stroke Scale (NIHSS) score, infarct volume, and Rad-score and lower Alberta stroke program early CT score (ASPECTS) (all p < 0.01) than patients without HT. The best ML algorithm for building the model was logistic regression. In the training and validation cohorts, the AUC values for the clinical, radiomics, and clinical-radiomics models for predicting HT were 0.829 and 0.876, 0.813 and 0.898, and 0.876 and 0.957, respectively. In subgroup analyses with different treatment modalities, different infarct sizes, and different stroke time windows, the assessment accuracy of the clinical-radiomics model was not statistically meaningful (all p > 0.05), with an overall accuracy of 79.5%. Moreover, this model performed reliably in predicting the PH and HI subcategories, with accuracies of 82.9 and 92.9%, respectively.ConclusionML models based on clinical and NCCT radiomics characteristics can be used for early risk evaluation of HT development in AIS patients and show great potential for clinical precision in treatment and prognostic assessment.

We aimed to verify the application value of the Alberta Stroke Program Early CT Score (ASPECTS) based on multiple post-labeling delay (multi-PLD) arterial spin labeling (ASL) for outcome assessment in acute ischemic stroke (AIS) patients. The endpoint was modified Rankin scale score at 90 days (90-day mRS). Patients were divided into the good outcome (0-2) and poor outcome (3-6) groups. The independent samples t-test, Mann-Whitney U-test, and χ2-test were used to compare clinical and imaging parameters between groups. We used partial correlation analysis to evaluate the relationships between ASPECTS and outcomes. Multivariate logistic regression analysis was used to examine potential independent prognostic indicators. The receiver operating characteristic (ROC) curve analysis was used to evaluate the performance of the independent prognostic indicators in predicting outcomes. Fifty-five AIS patients were included. The good outcome group had a lower baseline National Institutes of Health Stroke Scale (NIHSS; Z = -3.413, P < 0.001) and infarct core volume (ICV; Z = -3.114, P = 0.002) as well as higher cerebral blood flow (CBF)-ASPECTS (Z = -3.835, P < 0.001) and cerebral blood volume (CBV)-ASPECTS (Z = -4.099, P < 0.001). Higher CBF-ASPECTS (r = -0.459, P = 0.001), and CBV-ASPECTS (r = -0.502, P < 0.001) were associated with a lower 90-day mRS. The baseline NIHSS, CBF-ASPECTS, and CBV-ASPECTS were identified as independent prognostic indicators. The AUCs of the baseline NIHSS, CBF-ASPECTS, and CBV-ASPECTS were 83.3, 87.4, and 89.9%, respectively. Combining NIHSS with CBF-ASPECTS and CBV-ASPECTS, the AUC significantly improved to 96.3%. The combined three factors showed a significant difference compared to the baseline NIHSS (Z = 2.039, P = 0.041) and CBF-ASPECTS (Z = 2.099, P = 0.036), but no difference with CBV-ASPECTS (Z = 1.176, P = 0.239). The ASPECTS based on multi-PLD ASL is a valuable tool for identifying independent prognostic indicators and assessing clinical outcomes in AIS patients. The baseline NIHSS, combined with CBF-ASPECTS and CBV-ASPECTS, enhances the predictive efficacy of clinical outcomes in AIS patients. The CBV-ASPECTS alone can offer comparable predictive efficacy to the combination.

Exclusion of Isolated Cortical Swelling Can Increase Efficacy of Baseline Alberta Stroke Program Early CT Score in the Prediction of Prognosis in Acute Ischemic Stroke Patients Treated with Thrombolysis