Artificial intelligence (AI) is increasingly integrated into acute stroke imaging, offering new opportunities for automated image interpretation, diagnostic support, and workflow optimization. However, understanding its clinical value requires distinguishing the underlying computational methods from their diagnostic applications and demonstrated effects on patient care. This narrative educational review introduces the fundamental principles of machine learning, deep learning, and computer vision, with particular emphasis on convolutional neural networks and their applications in cerebrovascular imaging. Current evidence is examined across intracranial hemorrhage detection and volumetry, recognition of early ischemic changes, automated Alberta Stroke Program Early CT Score (ASPECTS) assessment, arterial occlusion detection, cerebral perfusion analysis, and ischemic lesion segmentation. Emerging applications in stroke etiology assessment are also discussed. Particular attention is given to persistent diagnostic challenges involving acute lacunar infarcts and cerebral small vessel disease, where conventional imaging limitations and insufficient external validation constrain current AI approaches. The review also examines generalizability, workflow integration, and regulatory considerations. AI can assist selected diagnostic tasks, but its clinical usefulness depends on external validation, appropriate implementation, and evidence of patient benefit.
Background: Deep vein thrombosis (DVT), chronic venous insufficiency (CVI), and atherosclerosis have traditionally been regarded as distinct conditions, although they may share mechanisms involving inflammation and endothelial dysfunction. Objective: To explore the relationship between DVT, CVI, inflammatory biomarkers, and carotid plaque burden using machine learning. Methods: We included 10 patients with DVT (>1 year after diagnosis), 10 with CVI, and 10 age-, sex-, and cardiovascular risk factor- frequency-matched controls. Circulating endothelial cells (CEC), ICAM-1, CXCL7, and high-sensitivity C-reactive protein (hs-CRP) were measured. Participants underwent carotid magnetic resonance imaging, with plaques classified according to the American Heart Association (AHA) score. Kernel principal component analysis (kPCA) was followed by k-nearest neighbors (kNN) (k = 3). Results: Only ICAM-1 differed significantly between clinical groups (median [IQR]: controls, 0.41 [0.37–0.43]; DVT, 0.70 [0.61–0.79]; CVI, 0.64 [0.61–0.79] pg/mL; p < 0.001). kPCA followed by kNN identified three classes: Class 1 included 9 controls; Class 2 included 8 CVI participants and 1 control; and Class 3 included 10 DVT and 2 CVI. The prevalence of AHA plaque scores ≥3 was descriptively higher in Classes 2 and 3 than in Class 1 (77.8%, 83.3%, and 33.3%, respectively), although the overall difference did not reach statistical significance (p = 0.071). Conclusions: ICAM-1 was the only biomarker differing significantly between clinical groups. Exploratory kPCA and kNN analyses identified multidimensional patterns associated with DVT and CVI. Differences in the prevalence of AHA plaque scores ≥3 across kNN classes did not reach statistical significance. These findings are hypothesis-generating and require validation in larger cohorts.
Aims: Long-term exposure to fine particulate matter (PM2.5) is a near-universal and modifiable exposure. We synthesised the systematic review-level evidence linking PM2.5 to type 2 diabetes (T2D) and examined how far agreement between reviews reflects independent replication rather than repeated analysis of shared primary studies. Methods and Results: We searched PubMed, Embase, Scopus, Web of Science and Google Scholar to 10 May 2026 (PROSPERO CRD42025642487), appraised reviews with the JBI checklist, and graded certainty using GRADE separately for each outcome domain. Thirteen systematic reviews and meta-analyses were included, spanning more than 200 million cumulative participant observations. All seven reviews reporting incident T2D found a positive association; six clustered at hazard or relative risks of 1.10–1.16 per 10 µg/m3, with a dose–response gradient and comparable estimates in Asian and Western populations. Associations were also reported for prevalent T2D (OR 1.12–1.25), diabetes-related mortality (RR 1.08, 1.03–1.13) and fasting plasma glucose (+1.18%, 0.86–1.50); HOMA-IR was null. Positive associations persisted in populations with annual means near the WHO guideline value of 5 µg/m3. Certainty was low for incident T2D—the ceiling attainable for an non-randomisable observational exposure—and very low for secondary outcomes; heterogeneity exceeded 90%, and because the reviews draw on shared primary studies their agreement is not independent replication. Conclusions: Across the whole review-level literature, long-term PM2.5 exposure is consistently associated with a 10–16% higher risk of T2D per 10 µg/m3. Because exposure is near-universal, an association of this size at the individual level implies a substantial attributable burden, greatest in South Asia. PM2.5 belongs within cardiometabolic prevention and within the cardiovascular–kidney–metabolic continuum, although causality and reversibility remain untested.
Background: Premature coronary artery disease (PCAD) represents a growing clinical concern. Although insulin resistance is central to atherogenesis and the triglyceride–glucose (TyG) index serves as a practical surrogate for insulin resistance, its specific prognostic capacity within younger patient cohorts remains poorly understood. This study sought to evaluate the prognostic value of the TyG index regarding major adverse cardiovascular events (MACE) and its correlation with angiographic disease severity in PCAD patients. Methods: Enrolling 382 consecutive patients presenting with acute coronary syndrome (ACS) undergoing primary or early invasive coronary angiography between July 2020 and March 2023, this retrospective study evaluated subjects stratified into premature CAD (males ≤ 45, females ≤ 55 years; n = 138) and late-onset CAD (n = 244) cohorts. Lesion complexity was evaluated via the Gensini score. Primary prognosis was evaluated via cumulative MACE during a mean follow-up of 40.7 months. Results: Baseline values for the TyG index did not exhibit significant variance between early-onset and late-onset cohorts (9.09 ± 0.71 vs. 9.06 ± 0.65, p = 0.716). Stratification by median TyG index in early-onset CAD showed expected metabolic clustering but no difference in Gensini scores (41.68 ± 25.90 vs. 44.12 ± 22.75, p = 0.321). In univariable logistic regression, baseline TyG index was not associated with MACE (OR = 1.10, 95% CI: 0.65–1.85, p = 0.732), whereas older age (OR = 1.06, 95% CI: 1.01–1.11, p = 0.017) and elevated baseline CRP levels (OR = 1.04, 95% CI: 1.01–1.08, p = 0.007) showed significant univariable associations with adverse clinical outcomes. Sensitivity analysis in non-diabetic patients (n = 275) confirmed no significant association between TyG index and MACE (OR = 0.721, 95% CI: 0.327–1.591, p = 0.418). Conclusion: In this study, the admission TyG index did not show a statistically significant relationship with coronary anatomical severity or long-term clinical outcomes in premature CAD. Risk assessment in acute coronary patients should continue to prioritize established clinical, acute inflammatory, and angiographic parameters.
Cardiovascular magnetic resonance (CMR) increasingly relies on advanced contrast strategies to characterize vascular anatomy, hemodynamics, and biological disease activity. Although gadolinium-based contrast agents remain the clinical standard, iron oxide nanoparticles offer distinctive magnetic and pharmacokinetic properties that enable applications extending beyond conventional contrast enhancement. Among these agents, ultrasmall superparamagnetic iron oxide particles (USPIOs), particularly ferumoxytol, combine prolonged intravascular persistence with delayed uptake by cells of the mononuclear phagocyte system. This review distinguishes three levels of evidence. First, ferumoxytol-enhanced blood-pool magnetic resonance angiography is a clinically feasible, although off-label, application, with particular value in congenital heart disease and in patients with impaired renal function. Second, T2*/R2* imaging of macrophage-associated inflammation in atherosclerotic plaques, abdominal aortic aneurysms, and, to a lesser extent, the myocardium is an exploratory biomarker approach supported mainly by small studies; in the largest cohort (MA3RS), USPIO enhancement did not improve prediction of clinical events beyond conventional risk factors, and ΔR2* reflects nanoparticle accumulation rather than a direct measure of macrophage burden. Third, surface-functionalized iron oxide nanoparticles for active molecular targeting of endothelial activation, thrombosis, and vascular remodeling, and for targeted drug delivery, remain at a preclinical stage; they represent a future direction rather than an additional temporal phase of conventional ferumoxytol imaging. This review summarizes the physical and pharmacokinetic principles of iron-based contrast agents, their cardiovascular applications, and their safety considerations and discusses the requirements (standardization, multicenter reproducibility, and outcome-based validation) for their integration into multiparametric cardiovascular imaging.
