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. 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. 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. 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. International Prospective Register of Systematic Reviews (CRD42022332816).

Introduction: Patients with hemorrhagic transformation (HT) were reported to have hemorrhage expansion. However, identification these patients with high risk of hemorrhage expansion has not been well studied. Objectives: We aimed to develop a radiomic score to predict hemorrhage expansion after HT among patients treated with thrombolysis/thrombectomy during acute phase of ischemic stroke. Methods: A total of 104 patients with HT after reperfusion treatment from the West China hospital, Sichuan University, were retrospectively included in this study between 1 January 2012 and 31 December 2020. The preprocessed initial non-contrast-enhanced computed tomography (NECT) imaging brain images were used for radiomic feature extraction. A synthetic minority oversampling technique (SMOTE) was applied to the original data set. The after-SMOTE data set was randomly split into training and testing cohorts with an 8:2 ratio by a stratified random sampling method. The least absolute shrinkage and selection operator (LASSO) regression were applied to identify candidate radiomic features and construct the radiomic score. The performance of the score was evaluated by receiver operating characteristic (ROC) analysis and a calibration curve. Decision curve analysis (DCA) was performed to evaluate the clinical value of the model. Results: Among the 104 patients, 17 patients were identified with hemorrhage expansion after HT detection. A total of 154 candidate predictors were extracted from NECT images and five optimal features were ultimately included in the development of the radiomic score by using logistic regression machine-learning approach. The radiomic score showed good performance with high area under the curves in both the training data set (0.91, sensitivity: 0.83; specificity: 0.89), test data set (0.87, sensitivity: 0.60; specificity: 0.85), and original data set (0.82, sensitivity: 0.77; specificity: 0.78). The calibration curve and DCA also indicated that there was a high accuracy and clinical usefulness of the radiomic score for hemorrhage expansion prediction after HT. Conclusions: The currently established NECT-based radiomic score is valuable in predicting hemorrhage expansion after HT among patients treated with reperfusion treatment after ischemic stroke, which may aid clinicians in determining patients with HT who are most likely to benefit from anti-expansion treatment.

Background: The retina and brain share similar neuronal and microvascular features. We aimed to investigate the retinal thickness and microvasculature in patients with thalamic infarcts compared with control participants. Material and methods: Swept-source optical coherence tomography (SS-OCT) was used to image the macular thickness (retinal nerve fiber layer, RNFL; ganglion cell-inner plexiform layer, GCIP), while OCT angiography was used to image the microvasculature (superficial vascular plexus, SVP; intermediate capillary plexus, ICP; deep capillary plexus, DCP). Inbuilt software was used to measure the macular thickness (µm) and microvascular density (%). Lesion volumes were quantitively assessed based on structural magnetic resonance images. Results: A total of 35 patients with unilateral thalamic infarction and 31 age–sex-matched controls were enrolled. Compared with control participants, thalamic infarction patients showed a significantly thinner thickness of RNFL (p < 0.01) and GCIP (p = 0.02), and a lower density of SVP (p = 0.001) and ICP (p = 0.022). In the group of patients, ipsilateral eyes showed significant reductions in SVP (p = 0.033), RNFL (p = 0.01) and GCIP (p = 0.043). When divided into three groups based on disease duration (6 months), no significant differences were found among these groups. After adjusting for confounders, SVP, ICP, DCP, RNFL, and GCIP were significantly correlated with lesion volume in patients. Conclusions: Thalamic infarction patients showed significant macular structure and microvasculature changes. Lesion size was significantly correlated with these alterations. These findings may be useful for further research into the clinical utility of retinal imaging in stroke patients, especially those with damage to the visual pathway.

Aims: We aimed to evaluate the retinal microvascular and structural changes in intracranial hypertension (IH) patients compared with an age‐ and sex‐matched control group. We also investigated the association between clinical parameters and retinal changes in IH patients.Methods: Intracranial hypertension patients were divided into eyes with papilledema (IH‐P) and eyes without papilledema (IH‐WP). IH patients underwent lumbar puncture to measure intracranial pressure (ICP); visual acuity was performed using the Snellen chart. Optical coherence tomography (OCT) was used to image and measure the retinal nerve fiber layer (RNFL) and ganglion cell‐inner plexiform layer (GCIPL) while OCT angiography was used to image and measure the superficial vascular complex (SVC) and deep vascular complex (DVC).Results: Intracranial hypertension patients showed reduced microvascular densities and thinner retinal thicknesses compared with the control group (all p < 0.001). Compared with the control group, IH‐P showed reduced microvascular densities and thinner retinal thicknesses (all p < 0.001). IH‐P showed reduced SVC density and thinner retinal thicknesses when compared with IH‐WP (p = 0.008 for SVC, p = 0.025 for RNFL, and p = 0.018 for GCIPL). ICP correlated with the microvascular densities and GCIPL thickness in IH patients (p = 0.025 for GCIPL, p = 0.004 for SVC, and p = 0.002 for DVC). A significant association of ICP with SVC (p = 0.010) and DVC (p = 0.005) densities were also found in IH‐P.Conclusions: Given the observed differences in these noninvasive retinal imaging markers, further research into their clinical utility in IH is needed.

Purpose: We explored the retinal microvascular changes in carotid artery stenosis (CAS) and their relationship with carotid plaque morphology. Methods: All participants were diagnosed with carotid artery stenosis by a neurologist. Participants underwent digital subtraction angiography (DSA) and optical coherence tomography angiography (OCTA) imaging. The degree and length of carotid plaque were obtained from the DSA tool. OCTA tool measured the densities in the superficial vascular complex (SVC) and deep vascular complex (DVC). Results: One hundred seventeen patients with CAS patients were included in our data analysis. Eyes with ipsilateral stenosis had reduced retinal microvascular densities when compared to contralateral eyes in patients with CAS (P = 0.016 for SVC, and P = 0.004 for DVC). Microvascular densities correlated with the length of carotid plaque (P = 0.015 for SVC, and P = 0.022 for DVC) in our CAS cohort, although they did not correlate with the degree of carotid plaque (P = 0.264 for SVC, and P = 0.298 for DVC). However, when stratified into moderate and severe subgroups, the degree of carotid plaque correlated with microvascular densities in patients with severe stenosis (P = 0.045 for SVC, and P = 0.038 for DVC). Conclusions: Our study suggests that OCTA can noninvasively detect retinal microvascular changes in patients with CAS and that these changes correlated with the length of the stenosis, but future studies are required to confirm these findings. Translational Relevance: Noninvasive and rapid acquisition of the OCTA image might have the potential to be used as a screening tool to detect microvascular changes in carotid artery stenosis.

Background: We aimed to assess the retinal structural and choroidal changes in carotid artery stenosis (CAS) patients and their association with cerebral hemodynamic changes. Asymptomatic and symptomatic patients with unilateral CAS were enrolled in our study. Material and methods: Swept-source optical coherence tomography (SS-OCT) was used to image the retinal nerve fiber layer (RNFL), ganglion cell-inner plexiform layer (GCIPL), while SS-OCT angiography (SS-OCTA) was used to image and measure the choroidal vascular volume (CVV) and choroidal vascular index (CVI). Computed Tomography Perfusion (CTP) was used to assess the cerebral perfusion parameters; relative perfusion (r) was calculated as the ratio of the value on the contralateral side to that on the ipsilateral side. Results: Compared with contralateral eyes, ipsilateral eyes showed significantly thinner RNFL (p < 0.001), GCIPL (p = 0.013) and CVV (p = 0.001). Relative cerebral blood volume (rCBV) showed a significant correlation with RNFL (p < 0.001), GCIPL (p < 0.001) and CVI (p = 0.027), while the relative permeability surface (rPS) correlated with RNFL (p < 0.001) and GCIPL (p < 0.001). Conclusions: Our report suggests that retinal and choroidal changes have the potential to detect hemodynamic changes in CAS patients and could predict the risk of stroke.