One of the main causes of death and permanent disability is ischemic stroke, for which prompt and precise diagnosis is essential to successful treatment. This study introduces a novel dual-stream deep learning framework for ischemic stroke classification using Computed Tomography (CT) images, specifically addressing challenges in accuracy, computational efficiency, and clinical interpretability. Three significant innovations are included in the suggested architecture: (1) a hybrid Dual Attention Mechanism that combines Dynamic Routing and Cross-Attention for improved region-specific feature discrimination; (2) a Multi-Scale Feature Extraction Module with parallel convolutional pathways that captures both contextual and fine-grained features; and (3) an Adaptive Random Vector Functional Link layer that significantly reduces training time while maintaining high classification performance. When tested on a single-center dataset, the model achieves state-of-the-art classification accuracy of 98.83% across normal, acute and chronic stroke categories. We demonstrate the strong generalization capabilities of the proposed framework by achieving 92.42% accuracy on a diverse, multi-center dataset of 7,842 CT images. The integration of explainable Artificial Intelligence tools improves clinical trustworthiness by offering clear insight into the model’s decision-making process. These outcomes demonstrate the model’s potential for use in actual clinical settings for quick and accurate stroke diagnosis, along with its interpretability and computational efficiency.

Azman RR, et al. Korean J Radiol. 2019 Jan;20:e31. https://doi.org/10.3348/kjr.2018.0416

Background: Additional bifurcations at the left main coronary artery (LMCA) could modify the geometry of the left coronary system, disturbing haemodynamic flow patterns and potentially altering endothelial shear stress (ESS). A low ESS has been implicated in atherogenesis. The emergence of the ramus intermedius (RI) from the LMCA creates additional branching, but the specific role of the RI in plaque deposition at the left coronary system remains unclear. This study sought to elucidate the potential effects of the RI on plaque formation at the LMCA and its bifurcation. Methods: A retrospective cross-sectional single-centre study was conducted using data from 139 female patients who were identified to have low risk of cardiovascular disease. These patients underwent cardiac computed tomography angiography between January 2017 and December 2018. Contrasted multiplanar coronary images taken during the best diastolic phase were analysed for the presence (experimental group) or absence (control group) of the RI. Measurements of plaques were done at the LMCA and at a 10 mm distance from the ostia of daughter arteries. Plaque data at the left bifurcation region were analysed using descriptive statistics, chi-square, and binary logistic regression tests. A p-value of <0.05 was considered statistically significant. Results: Amongst these low-risk patients, 33.8% (n = 47) had an RI. In the presence of RI, there was an eight-fold increased risk of plaque deposition at the LMCA (adjusted odds ratio, aOR = 8.5) and a three-fold increased risk of plaque deposition at the proximal left anterior descending (pLAD), especially on its lateral wall (aOR = 3.5). However, the RI did not influence plaque deposition at the distance of 10 mm from the ostium of the proximal left circumflex artery. Conclusions: These findings suggest that the RI increases the risk for atherosclerosis plaque deposition by three to eight-fold at the pLAD artery and the LMCA.