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 0.360.69; ischemic core volume on CTP ranked as the most important imaging biomarker with relative weights of 0.290.47. 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. ABBREVIATIONS: AIS ¼ acute ischemic stroke; CBS ¼ clot burden score; GBM ¼ Gradient Boosting Machine; IQR ¼ interquartile range; NECT ¼ non-contrast-enhanced CT; SPAN ¼ Stroke Prognostication Using Age and National Institutes of Health Stroke Scale; TIMI ¼ Thrombolysis in Myocardial Infarction; XGB ¼ Extreme Gradient Boosting; AUC ¼ area under the receiver operating curve
BACKGROUND AND PURPOSE: The contrast volume transfer coefficient (K trans), which reflects blood-brain barrier permeability, is influenced by circulation and measurement conditions. We hypothesized that focal low BBB permeability values can predict the spatial distribution of hemorrhagic transformation and global high BBB permeability values can predict the likelihood of hemorrhagic transformation. MATERIALS AND METHODS: We retrospectively enrolled 106 patients with hemispheric stroke who received intra-arterial thrombolytic treatment. K trans maps were obtained with first-pass perfusion CT data. The K trans values at the region level, obtained with the Alberta Stroke Program Early CT Score system, were compared to determine the differences between the hemorrhagic transformation and nonhemorrhagic transformation regions. The K trans values of the whole ischemic region based on baseline perfusion CT were obtained as a variable to hemorrhagic transformation possibility at the global level. RESULTS: Forty-eight (45.3%) patients had hemorrhagic transformation, and 21 (19.8%) had symptomatic intracranial hemorrhage. At the region level, there were 82 ROIs with hemorrhagic transformation and parenchymal hemorrhage with a mean K trans , 0.5 Ϯ 0.5/min, which was significantly lower than that in the nonhemorrhagic transformation regions (P Ͻ .01). The mean K trans value of 615 nonhemorrhagic transformation ROIs was 0.7 Ϯ 0.6/min. At the global level, there was a significant difference (P ϭ .01) between the mean K trans values of patients with symptomatic intracranial hemorrhage (1.3 Ϯ 0.9) and those without symptomatic intracranial hemorrhage (0.8 Ϯ 0.4). Only a high K trans value at the global level could predict the occurrence of symptomatic intracranial hemorrhage (P Ͻ .01; OR ϭ 5.04; 95% CI, 2.01-12.65). CONCLUSIONS: Global high K trans values can predict the likelihood of hemorrhagic transformation or symptomatic intracranial hemorrhage at the patient level, whereas focal low K trans values can predict the spatial distributions of hemorrhagic transformation at the region level. ABBREVIATIONS: AIS ϭ acute ischemic stroke; HI ϭ hemorrhagic infarction; HT ϭ hemorrhagic transformation; IAT ϭ intra-arterial thrombolysis; K trans ϭ contrast volume transfer coefficient; PCT ϭ perfusion CT; PH ϭ parenchymal hemorrhage; sICH ϭ symptomatic intracranial hemorrhage
BACKGROUND AND PURPOSE: CT is considered the standard reference both for quantification and characterization of carotid artery calcifications. Our aim was to investigate the relationship among different types of calcium configurations detected with CT within the plaque with a novel classification and to investigate the prevalence of cerebrovascular events. MATERIALS AND METHODS:Seven hundred ninety patients (men = 332; mean age, 69.7 [SD,13] years; 508 symptomatic for cerebrovascular symptoms and 282 asymptomatic) who underwent computed tomography of the carotid arteries were retrospectively included in this institutional review board-approved study. The plaque was classified into 6 types according to the different types of calcium configurations as the following: type 1, complete absence of calcification within the plaque; type 2, intimal or superficial calcifications; type 3, deep or bulky calcifications; type 4, adventitial calcifications with internal soft plaque of ,2 mm thickness; type 5, mixed patterns with intimal and bulky calcifications; and type 6, positive rim sign. RESULTS:The highest prevalence of cerebrovascular events was observed for type 6, for which 89 of the 99 cases were symptomatic. Type 6 plaque had the highest degree of correlation with TIA, stroke, symptoms, and ipsilateral infarct for both sides with a higher prevalence in younger patients. The frequency of symptoms observed by configuration type significantly differed between right and left plaques, with symptoms observed more frequently in type 6 calcification on the right side (50/53; 94%) than on the left side (39/46; 85%, P , .001). CONCLUSIONS:We propose a novel carotid artery plaque configuration classification that is associated with the prevalence of cerebrovascular events. If confirmed in longitudinal analysis, this classification could be used to stratify the risk of occurrence of ischemic events.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.