Predicting hemorrhagic transformation in patients not submitted to reperfusion therapies

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).

A predictive score for atrial fibrillation in poststroke patients

BACKGROUND

 Atrial fibrillation (AF) is a risk factor for cerebral ischemia. Identifying the presence of AF, especially in paroxysmal cases, may take time and lacks clear support in the literature regarding the optimal investigative approach; in resource-limited settings, identifying a higher-risk group for AF can assist in planning further investigation.

OBJECTIVE

 To develop a scoring tool to predict the risk of incident AF in the poststroke follow-up.

METHODS

 A retrospective longitudinal study with data collected from electronic medical records of patients hospitalized and followed up for cerebral ischemia from 2014 to 2021 at a tertiary stroke center. Demographic, clinical, laboratory, electrocardiogram, and echocardiogram data, as well as neuroimaging data, were collected. Stepwise logistic regression was employed to identify associated variables. A score with integer numbers was created based on beta coefficients. Calibration and validation were performed to evaluate accuracy.

RESULTS

 We included 872 patients in the final analysis. The score was created with left atrial diameter ≥ 42 mm (2 points), age ≥ 70 years (1 point), presence of septal aneurysm (2 points), and score ≥ 6 points at admission on the National Institutes of Health Stroke Scale (NIHSS; 1 point). The score ranges from 0 to 6. Patients with a score ≥ 2 points had a fivefold increased risk of having AF detected in the follow-up. The area under the curve (AUC) was of 0.77 (0.72-0.85).

CONCLUSION

 We were able structure an accurate risk score tool for incident AF, which could be validated in multicenter samples in future studies.

Hemorrhagic Transformation in Noncardioembolic Acute Ischemic Stroke: MRI Analysis From PACIFIC-STROKE

BACKGROUND

In the phase 2 PACIFIC-STROKE trial (Proper Dosing and Safety of the Oral FXIa Inhibitor BAY 2433334 in Patients Following Acute Noncardioembolic Stroke), asundexian, an oral factor XIa inhibitor, did not increase the risk of hemorrhagic transformation (HT). In this secondary analysis, we aimed to investigate the frequency, types, and risk factors of HT on brain magnetic resonance imaging (MRI).

METHODS

This was a secondary analysis of the PACIFIC-STROKE trial. Patients with mild-to-moderate acute noncardioembolic ischemic stroke were randomly assigned to asundexian or placebo plus guideline-based antiplatelet therapy. Brain MRIs were required at baseline (≤120 hours after stroke onset) and at 26 weeks or end-of-study. HT was defined using the Heidelberg classification and classified as early HT (identified on baseline MRI) or late HT (new HT by 26 weeks) based on iron-sensitive sequences. Multivariable logistic regression models were used to test factors that are associated with early HT and late HT, respectively.

RESULTS

Of 1745 patients with adequate baseline brain MRI (mean age, 67 years; mean National Institutes of Health Stroke Scale score, 2.8), early HT at baseline was detected in 497 (28.4%). Most were hemorrhagic infarctions (hemorrhagic infarction type 1: 15.2%; HI2: 12.7%) while a few were parenchymal hematomas (parenchymal hematoma type 1: 0.4%; parenchymal hematoma type 2: 0.2%). Early HT was more frequent with longer symptom onset-to-MRI interval. Male sex, diabetes, higher National Institutes of Health Stroke Scale large (>15 mm) infarct size, cortical involvement by infarct, higher number of acute infarcts, presence of chronic brain infarct, cerebral microbleed, and chronic cortical superficial siderosis were independently associated with early HT in the multivariable logistic regression model. Of 1507 with follow-up MRI, HT was seen in 642 (42.6%) overall, including 361 patients (23.9%) with late HT (new HT: 306; increased grade of baseline HT: 55). Higher National Institutes of Health Stroke Scale, large infarct size, cortical involvement of infarct, and higher number of acute infarcts predicted late HT.

CONCLUSIONS

About 28% of patients with noncardioembolic stroke had early HT, and 24% had late HT detectable by MRI. Given the high frequency of HT on MRI, more research is needed on how it influences treatment decisions and outcomes.

In-Hospital Aspirin Dose as a Risk Factor for Hemorrhagic Transformation in Patients Not Treated With Thrombolysis

BACKGROUND

Aspirin is widely used as secondary prophylaxis for acute ischemic stroke. However, its influence on the risk of spontaneous hemorrhagic transformation (HT) is still unclear. Predictive scores of HT have been proposed. We hypothesized that an increased aspirin dose might be harmful in patients at a high risk of HT. This study aimed to analyze the relationship between in-hospital daily aspirin dose (IAD) and HT in patients with acute ischemic stroke.

METHODS

We conducted a retrospective cohort study of patients admitted to our comprehensive stroke center between 2015 and 2017. The attending team defined IAD. All included patients underwent either computed tomography or magnetic resonance imaging within 7 days of admission. The risk of HT was assessed using the predictive score of HT in patients not undergoing reperfusion therapies. Regression models were used to evaluate the correlations between HT and IAD.

RESULTS

A total of 986 patients were included in the final analysis. The prevalence of HT was 19.2%, and parenchymatous hematomas type-2 (PH-2) represented 10% (n=19) of these cases. IAD was not associated with HT ( P =0.09) or PH-2 ( P =0.06) among all patients. However, in patients at a higher risk for HT (patients not undergoing reperfusion therapies ≥3), IAD was associated with PH-2 (odds ratio 1.01,95% CI 1.001-1.023, P =0.03) in an adjusted analysis. Taking 200 versus 300 mg aspirin was protective against PH-2 (odds ratio 0.102, 95% CI 0.018-0.563, P =0.009).

CONCLUSION

An increased in-hospital aspirin dose is associated with intracerebral hematoma in patients at a high risk of HT. Stratifying the risk of HT may lead to individualized daily aspirin dose choices. However, clinical trials on this topic are required.

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.

Roles of Bilirubin in Hemorrhagic Transformation of Different Types and Severity

BACKGROUND

Hemorrhagic transformation (HT) is a severe complication in patients with acute ischemic stroke (AIS). This study was performed to explore and validate the relation between bilirubin levels and spontaneous HT (sHT) and HT after mechanical thrombectomy (tHT).

METHODS

The study population consisted of 408 consecutive AIS patients with HT and age- and sex-matched patients without HT. All patients were divided into quartiles according to total bilirubin (TBIL) level. HT was classified as hemorrhagic infarction (HI) and parenchymal hematoma (PH) based on radiographic data.

RESULTS

In this study, the baseline TBIL levels were significantly higher in the HT than non-HT patients in both cohorts (p < 0.001). Furthermore, the severity of HT increased with increasing TBIL levels (p < 0.001) in sHT and tHT cohorts. The highest quartile of TBIL was associated with HT in sHT and tHT cohorts (sHT cohort: OR = 3.924 (2.051-7.505), p < 0.001; tHT cohort: OR = 3.557 (1.662-7.611), p = 0.006).

CONCLUSIONS

Our results suggest that an increased TBIL is associated with a high risk of patients with sHT and tHT, and that TBIL is more suitable as a predictor for sHT than tHT. These findings may help to identify patients susceptible to different types and severity of HT.

Predicting hemorrhagic transformation in posterior circulation stroke patients not treated with reperfusion therapies

INTRODUCTION

Posterior Circulation (PC) stroke represents one-fifth of all ischemic strokes, with peculiar physiological characteristics. Hemorrhagic Transformation (HT) is a dreaded complication among stroke patients. Many predictive scores of this complication have been proposed, but none is designed specifically for PC stroke patients - therefore, patients who are not eligible for reperfusion therapies (RT) represent about 80% of hospitalized cases. We propose a scoring system to assess the HT risk in PC stroke patients not submitted to RT.

METHODS

We retrospectively evaluated data of patients diagnosed with PC stroke not treated with RT from 5 Comprehensive Stroke Centers (four in Brazil, 1 in the US) from 2015 to 2018. All patients underwent CT scan or MRI at admission and a follow-up neuroimaging within seven days. Independent variables identified in a logistic regression analysis were used to produce a predictive grading score.

RESULTS

We included 952 patients in the final analysis. The overall incidence of HT was 8.7%. Male gender (1 point), NIH Stroke Scale at admission ≥ 5 points (1), blood glucose at admission ≥ 160 mg/dL (1), and cardioembolism (2) were independently associated with HT. The AUC of the grading score (0 to 5 points) was 0.713 (95% CI 0.65-0.78). Subjects with a score ≥ 3 points had an OR of 4.8 (95% CI 2.9-7.9, p < 0.001) for HT.

CONCLUSIONS

Our score has good accuracy in identifying patients at higher risk of HT. This score may be useful for evaluating secondary prevention and stratifying patients in the context of even clinical trials.

Differential Expression and Correlation Analysis of Global Transcriptome for Hemorrhagic Transformation After Acute Ischemic Stroke

In order to explore the epigenetic characteristics of hemorrhagic transformation (HT) after acute ischemic stroke, we used transcriptome sequencing technology to analyze the global transcriptome expression profile of patients with and without HT after acute ischemic stroke and to study the differential expression of messenger RNA (mRNA), long noncoding RNA (lncRNA), circular RNA (circRNA) and mircoRNA (miRNA) between the two groups. To further explore the role of differentially expressed genes in HT, we annotated the function of differentially expressed genes by using gene ontology (GO) and pathway analysis on the results and showed that there were 1,051 differential expressions of lncRNAs, 2,575 differential expressions of mRNAs, 447 differential expressions of circRNAs and 47 miRNAs in patients with HT compared with non-HT patients. Pathway analysis showed that ubiquitin-mediated proteolysis, MAPK signal pathway, axon guidance, HIF-1 signal pathway, NOD-like receptor signal pathway, beta-alanine metabolism, Wnt signal pathway, sphingolipid signal pathway, neuroactive ligand-receptor interaction, and intestinal immune network used in IgA production play an important role in HT. Terms such as iron homeostasis, defense response, immune system process, DNA conformational change, production of transforming growth factor beta-2, and oxidoreductase activity were enriched in the gene list, suggesting a potential correlation with HT. A total of 261 lncRNA-miRNA relationship pairs and 21 circRNA-miRNA relationship pairs were obtained; additionally, 5 circRNAs and 13 lncRNAs were screened, which can be used as competing endogenous RNA (ceRNA) to compete with miRNA in the co-expression network. Co-expression network analysis shows that these differentially expressed circRNA and lncRNA may play a vital role in HT and provide valuable information for new biomarkers or therapeutic targets.

Biomarkers in the Prediction of Hemorrhagic Transformation in Acute Stroke: A Systematic Review and Meta-Analysis

BACKGROUND

Hemorrhagic transformation (HT) is a complication that occurs spontaneously or after thrombolysis in acute ischemic stroke (AIS) and can increase morbidity and mortality. The association of biomarkers with the risk of HT has been variably reported. We conducted a systematic review of the literature and meta-analysis and sought to compare blood biomarkers associated with HT and its subtypes by evaluating its predictability and correlation with outcome in AIS.

METHODS

The study protocol was registered in the PROSPERO database (CRD42020201334) and adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Among 2,230 articles identified from Cochrane Library, PubMed, and Web of Science databases, 30 quality-appraised articles were found eligible. Meta-analysis was conducted for matrix metalloproteinase-9 (MMP-9), cellular fibronectin (c-Fn), ferritin, S100 calcium-binding protein B (S100B), and neutrophil-lymphocyte ratio (NLR). We also reviewed biomarkers for correlation with the functional outcome at 90 days from stroke onset (poor outcome modified Rankin scale >2).

RESULTS

The pooled diagnostic odds ratio (DORpooled) was the highest for baseline c-Fn levels (299.253 [95% CI, 20.508-4,366.709]), followed by MMP-9 (DORpooled, 29.571 [95% CI 17.750-49.267]) and ferritin (DORpooled, 24.032 [95% CI 2.557-225.871]). However, wide confidence intervals for ferritin and c-Fn suggested lesser reliability of the markers. Patients with MMP-9 levels ≥140 ng/mL were 29.5 times at higher risk of developing symptomatic HT after AIS (area under the curve = 0.881). S100B (DORpooled, 6.286 [95% CI, 1.861-21.230]) and NLR (DORpooled, 5.036 [95% CI, 2.898-8.749]) had lower diagnostic accuracies. Among the markers not included for meta-analysis, caveolin-1, thrombin-activated fibrinolysis inhibitor, plasminogen activator inhibitor-1, and soluble ST2 were highly sensitive. Elevated levels of MMP-9, ferritin, and NLR were found to be associated with poor functional outcomes and mortality.

CONCLUSION

Of the 5 biomarkers, there was enough evidence that MMP-9 has higher diagnostic accuracy for predicting the risk of HT before thrombolysis. MMP-9, ferritin, and NLR also predicted poor short-term outcomes.