Predicting hemorrhagic transformation after large vessel occlusion stroke in the era of mechanical thrombectomy

Prediction of cerebral hemorrhagic transformation after thrombectomy using a deep learning of dual-energy CT

OBJECTIVES

To develop and validate a deep learning model for predicting hemorrhagic transformation after endovascular thrombectomy using dual-energy computed tomography (CT).

MATERIALS AND METHODS

This was a retrospective study from a prospective registry of acute ischemic stroke. Patients admitted between May 2019 and February 2023 who underwent endovascular thrombectomy for acute anterior circulation occlusions were enrolled. Hemorrhagic transformation was defined using follow-up magnetic resonance imaging or CT. The deep learning model was developed using post-thrombectomy dual-energy CT to predict hemorrhagic transformation within 72 h. Temporal validation was performed with patients who were admitted after July 2022. The deep learning model's performance was compared with a logistic regression model developed from clinical variables using the area under the receiver operating characteristic curve (AUC).

RESULTS

Total of 202 patients (mean age 71.4 years ± 14.5 [standard deviation], 92 men) were included, with 109 (54.0%) patients having hemorrhagic transformation. The deep learning model performed consistently well, showing an average AUC of 0.867 (95% confidence interval [CI], 0.815-0.902) upon five-fold cross validation and AUC of 0.911 (95% CI, 0.774-1.000) with the test dataset. The clinical variable model showed an AUC of 0.775 (95% CI, 0.709-0.842) on the training dataset (p < 0.01) and AUC of 0.634 (95% CI, 0.385-0.883) on the test dataset (p = 0.06).

CONCLUSION

A deep learning model was developed and validated for prediction of hemorrhagic transformation after endovascular thrombectomy in patients with acute stroke using dual-energy computed tomography.

CLINICAL RELEVANCE STATEMENT

This study demonstrates that a convolutional neural network (CNN) can be utilized on dual-energy computed tomography (DECT) for the accurate prediction of hemorrhagic transformation after thrombectomy. The CNN achieves high performance without the need for region of interest drawing.

KEY POINTS

• Iodine leakage on dual-energy CT after thrombectomy may be from blood-brain barrier disruption. • A convolutional neural network on post-thrombectomy dual-energy CT enables individualized prediction of hemorrhagic transformation. • Iodine leakage is an important predictor of hemorrhagic transformation following thrombectomy for ischemic stroke.

MMP-9 upregulation may predict hemorrhagic transformation after endovascular thrombectomy

Background

Hemorrhagic transformation (HT) is a serious complication after endovascular thrombectomy (EVT) for patients with acute ischemic stroke (AIS). We analyzed the plasma levels of MMP-9 before and after EVT and assessed the temporal changes of MMP-9 that may be associated with, and therefore predict, HT after EVT.

Methods

We enrolled 30 AIS patients who received EVT, and 16 (53.3%) developed HT. The levels of MMP-9 in plasma collected from the arteries of AIS patients before and immediately after EVT were measured using ELISA. The percent change in MMP-9 after EVT (after/before) was calculated and compared between patients with and without HT.

Results

The median age of the AIS patients was 70 years, and 13 patients (43.3%) were men. The median National Institutes of Health Stroke Scale (NIHSS) scores of patients with HT were 18 on admission and 18 after EVT. The median NIHSS scores of patients without HT were 17 on admission and 11 after EVT. Patients with HT demonstrated significantly greater percentage increases in arterial MMP-9 levels after EVT.

Conclusion

Patients with AIS who developed HT had significantly increased arterial MMP-9 levels after EVT, suggesting that the upregulation of MMP-9 following EVT could serve as a predictive biomarker for HT.

Acute ischemic STROKE - from laboratory to the Patient's BED (STROKELABED): A translational approach to reperfusion injury. Study Protocol

Cerebral edema (CE) and hemorrhagic transformation (HT) are frequent and unpredictable events in patients with acute ischemic stroke (AIS), even when an effective vessel recanalization has been achieved. These complications, related to blood-brain barrier (BBB) disruption, remain difficult to prevent or treat and may offset the beneficial effect of recanalization, and lead to poor outcomes. The aim of this translational study is to evaluate the association of circulating and imaging biomarkers with subsequent CE and HT in stroke patients with the dual purpose of investigating possible predictors as well as molecular dynamics underpinning those events and functional outcomes. Concurrently, the preclinical study will develop a new mouse model of middle cerebral artery (MCA) occlusion and recanalization to explore BBB alterations and their potentially harmful effects on tissue. The clinical section of the study is based on a single-center observational design enrolling consecutive patients with AIS in the anterior circulation territory, treated with recanalization therapies from October 1, 2015 to May 31, 2020. The study will employ an innovative evaluation of routine CT scans: in fact, we will assess and quantify the presence of CE and HT after stroke in CT scans at 24 h, through the quantification of anatomical distortion (AD), a measure of CE and HT. We will investigate the relationship of AD and several blood biomarkers of inflammation and extracellular matrix, with functional outcomes at 3 months. In parallel, we will employ a newly developed mouse model of stroke and recanalization, to investigate the emergence of BBB changes 24 h after the stroke onset. The close interaction between clinical and preclinical research can enhance our understanding of findings from each branch of research, enabling a deeper interpretation of the underlying mechanisms of reperfusion injury following recanalization treatment for AIS.

Hemorrhagic Conversion of Acute Ischemic Stroke

Stroke is a leading cause of morbidity and mortality worldwide; a serious complication of ischemic stroke is hemorrhagic transformation. Current treatment of acute ischemic stroke includes endovascular thrombectomy and thrombolytic therapy. Both of these treatment options are linked with increased risks of hemorrhagic conversion. The diagnosis and timely management of patients with hemorrhagic conversion is critically important to patient outcomes. This review aims to discuss hemorrhagic conversion of acute ischemic stroke including discussion of the pathophysiology, review of risk factors, imaging considerations, and treatment of patients with hemorrhagic conversion.

Eosinophil: A New Circulating Biomarker for Risk of Poor Outcome in Stroke Patients Undergoing Mechanical Thrombectomy

Objective

Acute ischemic stroke (AIS), caused by occlusion of large vessel, is a serious life-threatening disease. This study aimed to comprehensively investigate the association of 14 common and readily available circulating biomarkers with the 90-day modified Rankin Scale (mRS) score in patients undergoing mechanical thrombectomy (MT).

Methods

This study included patients with anterior circulation large vessel occlusive stroke treated with MT from 05/2017 to 12/2021. Baseline comparisons of poor outcome were performed among enrolled patients. Factors that may be associated with the mRS score were assessed using correlation analysis. Univariate and multivariate logistic regression analyses were used to evaluate the predictive value of circulating biomarkers and poor outcome.

Results

The mRS score has a strong correlation with neutrophil to lymphocyte ratio (NLR) and eosinophil levels (all r_s>0.4 in absolute value and all P<0.001) in addition to a high correlation with National Institute of Health Stroke Scale (NIHSS) score (r_s=0.40, P<0.001). There was also a high correlation between NLR and eosinophil (r_s=-0.58, P<0.001). In the multivariate regression analysis, only neutrophil (adjusted OR=1.301, 95% CI: 1.155-1.465, P<0.001), eosinophil (adjusted OR<0.001, 95% CI: <0.001-0.016, P<0.001), and NLR (adjusted OR=1.158, 95% CI: 1.082-1.241, P<0.001) were independently associated with poor outcome.

Conclusion

This study evaluated a series of circulating biomarkers and found that neutrophil, eosinophil, and NLR independently predicted poor outcome after MT in AIS patients. There was a significant negative correlation between eosinophil and NLR levels.

Matrix Metalloproteinases in Cardioembolic Stroke: From Background to Complications

Matrix metalloproteinases (MMPs) are endopeptidases participating in physiological processes of the brain, maintaining the blood-brain barrier integrity and playing a critical role in cerebral ischemia. In the acute phase of stroke activity, the expression of MMPs increase and is associated with adverse effects, but in the post-stroke phase, MMPs contribute to the process of healing by remodeling tissue lesions. The imbalance between MMPs and their inhibitors results in excessive fibrosis associated with the enhanced risk of atrial fibrillation (AF), which is the main cause of cardioembolic strokes. MMPs activity disturbances were observed in the development of hypertension, diabetes, heart failure and vascular disease enclosed in CHA₂DS₂VASc score, the scale commonly used to evaluate the risk of thromboembolic complications risk in AF patients. MMPs involved in hemorrhagic complications of stroke and activated by reperfusion therapy may also worsen the stroke outcome. In the present review, we briefly summarize the role of MMPs in the ischemic stroke with particular consideration of the cardioembolic stroke and its complications. Moreover, we discuss the genetic background, regulation pathways, clinical risk factors and impact of MMPs on the clinical outcome.

Biomarkers predict hemorrhagic transformation and stroke severity after acute ischemic stroke

OBJECTIVES

Hemorrhagic transformation (HT) is a complication occurring in patients with acute ischemic stroke (AIS) either spontaneously or post-thrombolysis leading to significant morbidity and mortality. We assessed circulating matrix metalloproteinase-9 (MMP-9), Claudin-5, and soluble serum stimulation-2 (sST2) in HT and stroke severity in AIS based on their temporal distribution.

MATERIALS AND METHODS

We prospectively enrolled 111 AIS patients within 12 h from onset. Patient demographic, clinical, and imaging details were documented. Follow-up imaging was conducted 24-48 h after admission. Blood samples were taken at three time-points from stroke onset. HT was classified according to the European Co-operative Acute Stroke Study-III(ECASS-III). Stroke severity was assessed using the National Institutes of Health Stroke Scale (NIHSS). Multiple logistic regression and receiver operating characteristic curve were conducted to determine the discriminative capacity.

RESULTS

Mean age was 62.3 ± 11.7 years and median baseline NIHSS was 12[IQR 8.0-18.0]. HT was detected in 30(27%) patients. Biomarker levels at 12 h were elevated with median MMP-9 concentration of 153.9 ng/mL[IQR 110.6-309 ng/mL] indicating a trend toward significant positive correlation with HT(P = 0.05). Claudin-5 levels at 12 h was elevated but was not statistically significant (43.1 pg/mL[IQR:26.7-72.6 pg/mL] vs 59.4 pg/mL[IQR:24.5-100.8 pg/mL];P = 0.4). Multiple logistic regression indicated Claudin-5 levels at 12 h (OR 9.46;95% CI:1.97-64.6;P = 0.010) and baseline low ASPECTS score(OR 20.3;95% CI:3.46-193; P = 0.003) independently predicted HT. MMP-9 at 12 h was significantly elevated in patients with moderate to severe strokes (P = 0.04).

CONCLUSIONS

Claudin-5 and low ASPECTS independently predicted HT. MMP-9 was positively correlated with baseline stroke severity.