The Neuroradiology Journal

The aortic arch (AA) influences catheter navigation during endovascular treatment (EVT) for acute ischemic stroke. Whether routine inclusion of the AA in stroke CT angiography (CTA) protocols is necessary remains debated, particularly in settings aiming to streamline imaging workflows.

To evaluate whether machine learning algorithms (MLAs) can use routinely available clinical and laboratory data to estimate AA configuration in patients with stroke risk factors.

In this retrospective study, 108 patients who underwent CTA, including the AA, were analyzed. Eight MLAs were trained to classify AA anatomy as favorable or unfavorable using clinical and laboratory variables available at presentation. Class imbalance was addressed using the Synthetic Minority Oversampling Technique. Model performance was evaluated using five-fold cross-validation and summarized with seven performance metrics.

After class balancing, mean accuracy across classifiers ranged from 0.52 to 0.77. Gaussian copula and Naive Bayes classifiers showed the highest accuracy, while k-nearest neighbors and support vector machines demonstrated higher sensitivity. Specificity was highest for regularized discriminant analysis and the Gaussian copula. AUROC values ranged from 0.49 to 0.80 across models.

Routinely available clinical and laboratory data contain a limited yet detectable signal, enabling several machine learning algorithms to estimate AA configuration with moderate accuracy. These findings are exploratory and do not support modification of CTA protocols. However, they suggest that pre-imaging clinical data may provide supportive information for future workflow-oriented decision-support tools. Further validation in larger, externally tested cohorts is required before clinical integration can be considered.

Ischemic stroke, often caused by intracranial atherosclerotic stenosis (ICAS), is a leading cause of disability and death globally. 7T MRI offers improved signal-to-noise ratio (SNR) and contrast, enhancing MRA image quality. This study assessed the accuracy of 7T high-resolution (HR) MRI and time-of-flight MRA (TOF-MRA) in measuring middle cerebral artery (MCA) stenosis compared to digital subtraction angiography (DSA).

In this prospective observational study, patients, diagnosed with MCA atherosclerotic stenosis, underwent 7T HR-MRI, TOF-MRA and DSA. Spearman correlation analysis was used to assess the linear correlation between HR-MRI, TOF-MRA, and DSA measurements of MCA stenosis. Agreement for stenosis classification data was evaluated using weighted kappa values. Receiver Operating Characteristic (ROC) curves were utilized to predict severe stenosis and occlusion of MCA in HR-MRI and TOF-MRA.

This study included 86 lesions from 81 patients with MCA atherosclerotic stenosis. The stenosis degree measurements between 7T HR-MRI and DSA indicated a very strong correlation (r = 0.8000; p < .0001). The correlation between 7T TOF-MRA and DSA assessments was high (r = 0.7188; p < .001). A robust correlation between 7T HR-MRI and TOF-MRA evaluations was shown (r = 0.839; p < .0001). 7T TOF-MRA (K = 0.673) demonstrated substantial agreement with DSA measurements, showing slight superiority over HR-MRI (K = 0.589). ROC analysis indicated both 7T HR-MRI and TOF-MRA effectively predicted severe MCA stenosis and occlusion (HR-MRI, AUC: 0.888; TOF-MRA, AUC: 0.878).

7T HR-MRI and TOF-MRA may emerge as comparable modalities for assessing MCA atherosclerotic stenosis, and HR-MRI allows for precise evaluation through direct visualization of the vessel lumen and plaque.

Alzheimer’s disease requires early detection for effective intervention with disease-modifying treatments, yet significant implementation barriers persist in clinical practice, including limited computational infrastructure and the gap between research model performance and practical deployment in resource-constrained settings.

To develop and evaluate a computer-aided diagnosis system for classifying cognitive states (AD, MCI, and CN) from structural MRI, with automated preprocessing and cost-effective cloud deployment suitable for resource-constrained healthcare facilities.

Computer-aided diagnosis system integrating neural architecture search with serverless cloud infrastructure.

A multi-view MRI analysis model was optimized through neural architecture search, incorporating Universal Inverted Bottleneck blocks and Kolmogorov–Arnold Networks. Automated MRI preprocessing using FSL was deployed through cloud-based serverless functions for scalable image processing. Evaluation used the ADNI dataset (1687 individuals: 368 AD, 625 MCI, and 694 CN). A web application was developed providing patient management, MRI visualization, and automated diagnostic prediction.

The model achieved 86.7% accuracy and 0.900 AUC in three-class classification with 1.7 million parameters. High specificity was observed across all classes (CN: 91.0%, MCI: 91.8%, AD: 97.3%), with 100% CN-AD specificity ensuring no AD cases were misclassified as cognitively normal. Operational costs were approximately 0.028 USD per diagnosis for typical hospital workloads.

This system provides a cost-effective approach for early Alzheimer’s diagnosis accessible to resource-constrained environments. Despite challenges in MCI classification, the combination of neural architecture search with serverless deployment demonstrates progress toward clinically deployable automated AD detection. Future work should focus on prospective clinical validation and integration of interpretability features.

To identify the peculiar structural imaging features of atypical parkinsonian syndromes and recognize the role of the current state-of-the-art neuroimaging modalities (particularly MRI and nuclear medicine techniques) in discriminating Parkinson’s disease from its mimics.

Sickle cell disease (SCD) confers a three-to fivefold increase in the prevalence of intracranial aneurysms (IA). Flow diversion (FD) requires mandatory dual antiplatelet therapy (DAPT), yet perioperative transfusion to reduce hemoglobin S (HbS) risks delayed hemolytic transfusion reaction (DHTR). No guidelines address this conflict between DHTR-associated coagulopathy and mandatory DAPT. We report the largest FD series in SCD.

Retrospective single-center case series of SCD patients undergoing FD for IA (2017 to 2026). Of 12 SCD patients treated for aneurysms, seven received FD and comprised the analytic cohort.

Seven patients (5 HbSS, 2 HbS/β⁰-thalassemia; median age 38; 71% female) harbored 29 aneurysms; 12 were treated with FD across 8 procedures using 15 devices. HbS <30% was achieved in 3 of 7 (43%); one patient underwent FD at HbS 78.5% due to alloimmunization. Clopidogrel hyporesponsiveness was identified in 2 of 4 TEG-tested patients (50%). Procedure-related mortality was 14% (1/7): fatal DHTR with DIC on postoperative day 8, the first reported DHTR-DAPT collision. Procedure-related morbidity included one intraoperative thrombus and one vasospasm episode, both resolved without sequelae (2/8 procedures, 25%). Additional events included asymptomatic in-stent stenosis and delayed stroke from SCD vasculopathy. At last follow-up, 5 of 6 survivors maintained mRS 0 to 1.

FD is technically feasible in SCD, but the 14% procedure-related mortality from a fatal DHTR exposes the unresolved conflict between transfusion-associated hyperhemolysis and mandatory DAPT. CYP2C19-mediated clopidogrel resistance and alloimmunization pose additional challenges that require predefined protocols and prospective multicenter registries.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has been associated with neurological complications beyond respiratory involvement. Although children usually experience mild systemic disease, immune-mediated central nervous system (CNS) disorders have emerged as an important cause of morbidity. Neuroimaging plays a central role in detecting these complications and supporting diagnostic decisions. This study aims to describe neuroimaging patterns in children who developed immune-mediated CNS disorders following SARS-CoV-2 infection.

We retrospectively reviewed pediatric patients aged 0–18 years with laboratory-confirmed SARS-CoV-2 infection followed at a tertiary pediatric neurology center. Among post-COVID admissions, four patients with severe neurological involvement and characteristic neuroimaging abnormalities were identified. Clinical features, laboratory findings, cerebrospinal fluid analyses, electroencephalography results, and multimodal neuroimaging studies were evaluated.

Four previously healthy male children developed immune-mediated neurological disorders associated with SARS-CoV-2 infection, including febrile infection–related epilepsy syndrome, acute disseminated encephalomyelitis, central nervous system vasculitis, and anti–N-methyl-D-aspartate receptor encephalitis. Brain MRI demonstrated disorder-specific patterns such as cortical diffusion restriction, multifocal demyelinating lesions, inflammatory vasculitic changes, and progressive parenchymal injury. In one patient, serial imaging showed progression from acute inflammation to chronic structural damage. All patients received immunomodulatory therapy, with variable neurological outcomes.

SARS-CoV-2 infection may trigger immune-mediated CNS disorders in children with distinct neuroimaging features. Recognition of these patterns is essential for early diagnosis, treatment guidance, and prognostic assessment. Neuroimaging also provides insight into the evolution of post–COVID-19 neuroinflammatory injury.

To characterize demographic variation in quantitative MR neurography parameters of the peroneal nerve and to assess the methodological impact of epineurium inclusion on signal intensity–based measurements.

This retrospective single-center study included 114 screened, presumably healthy subjects undergoing routine 3-T knee MRI. Cross-sectional area (CSA), signal intensity (SI), and nerve-to-muscle SI ratios of the peroneal nerve were measured at three predefined levels, with and without epineurium inclusion. Associations with age, body mass index (BMI), and sex were evaluated using univariable linear regression. Agreement between SI measurements with and without epineurium was assessed using cubic spline regression. Exploratory fuzzy c-means clustering was applied to visualize demographic trends.

SI ratios measured with and without epineurium showed excellent agreement without evidence of a clinically relevant systematic bias. Increasing age was significantly associated with larger CSA (p < .05). In contrast, SI ratios showed no significant age-related association and only weak trends. BMI demonstrated no significant association with either CSA or SI, and no relevant sex-related differences were observed. Fuzzy clustering corroborated regression findings, demonstrating gradual demographic trends with substantial overlap between clusters.

Quantitative MR neurography parameters of the peroneal nerve exhibit small, gradual variations with age and body mass index, primarily affecting cross-sectional area, while signal-intensity ratios remain largely stable. Measurements with and without epineurium inclusion show strong agreement, indicating that demographic-related variability is modest and should be interpreted as physiological variation rather than pathological change.

Chronic subdural hematoma (cSDH) remains a prevalent neurosurgical condition in elderly patients, traditionally managed via surgical evacuation when symptomatic. Middle meningeal artery embolization (MMAE) has recently emerged as a minimally invasive alternative. However, the comparative efficacy and safety of standalone MMAE versus standalone surgical evacuation remain inadequately defined.

A comprehensive search of PubMed, Scopus, Embase, and Web of Science, was performed on March 20, 2025. Studies comparing standalone MMAE and standalone surgical evacuation in patients with cSDH were included. The primary outcome was recurrence; secondary outcomes included complication and mortality rates. A random-effects pairwise meta-analysis model using restricted maximum likelihood (REML) was applied to account for between-study variability.

Ten studies encompassing 50,138 patients were included. Recurrence was significantly lower with MMAE (RR = 0.51, 95% CI: 0.33–0.80, p = .00), with no statistically detected heterogeneity (I² = 0%), although clinical variability across studies cannot be excluded. Mortality (RR = 0.70, 95% CI: 0.44–1.12, p = .14) and complications (RR = 1.04, 95% CI: 0.67–1.64, p = .85) did not differ significantly between both groups. The GRADE assessment rated the certainty of evidence for recurrence as moderate, while the certainty for both mortality and complications were rated as very low, underscoring the need for further high-quality studies.

Standalone MMAE significantly reduces recurrence risk compared to surgical evacuation in cSDH patients, while showing no significant difference in mortality or complication rates. These findings highlight MMAE as a promising first-line treatment option, particularly for patients at high surgical risk. Prospective randomized trials are warranted to further validate these outcomes.

Intracranial arterial stenosis is a major cause of ischemic stroke in Asian populations. High-resolution magnetic resonance imaging (HR-MRI) of the vessel wall is an emerging tool that offers valuable insights.

We conducted a retrospective study of adult patients (≥18 years) with acute ischemic stroke and M1 MCA stenosis who underwent HR-MRI at the University Medical Center Ho Chi Minh City between January 2019 and May 2025.

Sixty-two patients were included (mean age, 56 ± 13 years; male-to-female ratio, 2:1). The average stenosis degree was 66.3 ± 17.5, with severe stenosis in 38.7% of patients. The predominant pattern was eccentric stenosis (80.6%), more frequently observed in older patients. The mean remodeling index was 0.87 ± 0.21, and most cases showed negative remodeling. The majority of lesions at the stenotic site exhibited post-contrast enhancement, most commonly grade 2 enhancement (43.6%). The most frequent cause of stenosis was atherosclerosis (43 cases), followed by dissection (2 cases), Moyamoya disease (1 case), and indeterminate etiology (16 cases). Among infarct-associated atherosclerotic lesions, plaques with intraplaque hemorrhage (58.1%), irregular plaque surface (86.1%), and strong (grade 3) enhancement (46.5%) were observed more frequently than in the small non–infarct-associated incidental group; however, this exploratory comparison was limited by the very small size of the comparator group.

HR-MRI of the MCA vessel wall provides detailed information on stenotic lesion characteristics, aiding in the identification of the underlying etiology. These findings highlight the potential role of HR-MRI in stroke diagnosis and risk stratification in Vietnamese patients.

Arterial calcification is commonly found in patients with cerebrovascular disease and may negatively affect outcomes following endovascular thrombectomy (EVT) for acute ischemic stroke (AIS). This systematic review and meta-analysis aimed to evaluate the impact of intracranial or aortic arch calcification on procedural success and functional outcomes in AIS patients undergoing anterior circulation EVT.

We systematically searched the PubMed, EMBASE, and Cochrane databases to identify studies comparing EVT outcomes between AIS patients with and without arterial calcification. Outcomes assessed included first-pass effect (FPE), successful reperfusion at the end of the procedure, modified Rankin scale (mRS), mortality, and symptomatic intracerebral hemorrhage (sICH). Risk ratios (RRs) with 95% confidence intervals (CIs) were pooled using random effects models, with heterogeneity evaluated with I² statistics.

A total of six observational studies were included, with a total of 3834 patients (2303 with calcification and 1531 controls). Arterial calcification was associated with a lower chance of achieving FPE (RR: 0.50, 95% CI: 0.36–0.69, p < 0.001) and successful reperfusion (RR: 0.90, 95% CI: 0.86–0.94, p < 0.001). Patients with calcification had a greater risk of poor outcome (mRS 3–6) (RR: 1.29, 95% CI: 1.16–1.43, p < 0.001), mortality (RR: 1.51, 95% CI: 1.14–2.00, p < 0.001), and sICH (RR: 1.29, 95% CI: 1.16–1.43, p < 0.001). Heterogeneity was low, with I² ≤ 34%.

Arterial calcification may be indicative of advanced vascular disease, potentially leading to diminished technical success and worse functional outcomes following EVT for AIS. These findings underscore that vascular calcification acts as an associative marker of procedural complexity and advanced vascular disease in EVT candidates.

To evaluate virtual non-contrast (VNC) accuracy at multiple radiation doses and iodine concentrations in a head CT phantom using dual-energy integrating-detector CT (EID-CT; TwinSpiral DECT) and photon-counting detector CT (PCD-CT).

A cylindrical head-sized acrylic CT phantom with brain tissue inserts and iodine (0, 43.75, 175, and 350 mg/mL; 87.5 mg/mL was excluded due to beam-hardening artifacts) was scanned three times on each system at 100%, 80%, 75%, 65%, 50%, and 25% of the standard radiation dose to generate mixed and VNC images. Attenuations (HU) were measured in standard and corresponding VNC images; signed and absolute VNC errors were analyzed with two-way ANOVA (dose, iodine) with Tukey post hoc tests (p < .05), using nonparametric alternatives when assumptions were violated. Phantom HU for fat, bone, and white matter at 100% dose on EID-CT were compared with 16 patient scans to confirm realism.

Radiation dose significantly affected VNC errors in both scanners, but only 25% versus 100% dose in PCD-CT was significant after Tukey (p = .0383); no pairwise dose differences were significant in EID-CT (p ≥ .07). Iodine concentration had a stronger effect. In EID-CT, 350 mg/mL differed from 0, 43.75, and 175 mg/mL (p < .001), with errors up to +13.7 ± 1.1 HU. In PCD-CT, 175 and 350 mg/mL differed from 0 mg/mL (p < .05) and from 43.75 mg/mL (p < .001), with a negative bias up to −10.6 ± 1.5 HU.

Dose reduction had limited impact on VNC accuracy, whereas high iodine (175–350 mg/mL) caused significant errors with opposite signed biases (EID-CT overestimation, PCD-CT underestimation). Both scanners still achieved substantial visual contrast removal.

Only limited studies have directly compared apparent diffusion coefficient (ADC) values at term-equivalent age (TEA) between normal term infants and very preterm infants. We hypothesized that regional ADC values at TEA would differ between preterm infants and term infants, even in the absence of MRI-detectable brain injury. The purpose of this study was to compare ADC values at TEA between term infants and preterm infants born at less than 32 weeks of gestation.

This retrospective study enrolled consecutive infants who underwent neonatal brain MRI over a 7-year period from 2016 to 2022. From this cohort, very preterm infants were selected after confirmation of normal MRI findings and normal neurodevelopmental outcomes. ADC values were measured at six predefined brain regions. ADC values were compared between term infants and preterm infants, and further subgroup analyses were performed among term, very preterm, and extremely preterm infants.

ADC values at the basal ganglia were significantly higher in preterm infants compared with term infants. No significant differences were observed at the other five brain regions. Subgroup analysis among term, very preterm, and extremely preterm infants demonstrated significant differences in basal ganglia ADC values, with both very preterm and extremely preterm infants showing higher ADC values than term infants.

ADC values at the basal ganglia at TEA were significantly higher in preterm infants than in term infants, despite normal conventional MRI findings and favorable neurodevelopmental outcomes. These findings suggest region-specific differences in brain microstructural maturation associated with preterm birth.

Carotid artery calcification represents a common feature of atherosclerotic plaques. However, the geometric relationships of calcific clusters have not been systematically investigated. This study aimed to develop a geometry-based atlas, independent of symptom status, characterizing the three-dimensional properties of calcium clusters within carotid plaques and quantifying population-level distributions and data-driven couplings between size, shape, and density.

Standardized three-dimensional calcium masks of clinical CT scans were analyzed using connected-component labeling. Extracted features included cluster volume, aspect ratio, eccentricity, compactness, and CT attenuation values (mean μ, standard deviation σ). Associations between cluster features were assessed using Spearman correlations with Benjamini–Hochberg false discovery rate correction. Independent associations were determined using partial Spearman correlation and restricted cubic spline regression.

Among 107 plaques, 149 distinct clusters were identified, most frequently presenting as a single elongated cluster (aspect ratio = 2.23). Independent associations (all p < 0.001) were identified for: eccentricity with μ (ρ = −0.39), volume with μ (ρ = 0.47), compactness with μ (ρ = 0.39), and volume with eccentricity (ρ = 0.82). Multi-cluster plaques exhibited smaller mean cluster volume (76.1 vs 359.5 mm³; p < 0.001) and lower eccentricity (0.5 vs 1.6; p < 0.001) compared with single-cluster plaques.

Carotid calcification most frequently manifests as a single elongated cluster, with robust couplings between cluster size, shape, and density. These geometric archetypes provide a quantifiable framework for future biomechanical and biological studies of plaque vulnerability.

The deep cerebral venous system plays a critical role in surgical approaches to the pineal and posterior third ventricular regions—areas that are particularly challenging in pediatric neurosurgery. While adult venous anatomy is well described, data on normal pediatric venous development remain scarce. Surgical planning in children often relies on extrapolating adult anatomy, which may not reflect age-related anatomical variations that occur during rapid brain growth.

We retrospectively analyzed high-resolution 3.0 T contrast-enhanced MRIs from 85 pediatric patients (aged 15 days–17 years) performed at a single tertiary center. Key venous angles (ICV–GV, GV–SS, SS–SSS, BV–GV) and distances (BV–BV, ICV–ICV, ICV–BV) were measured using multiplanar reconstruction. Correlations with age were assessed using Pearson correlation and linear regression. Patients were divided into three age groups (0–3 years, 4–12 years, 13–18 years), and group differences were evaluated with ANOVA and Bonferroni correction.

Significant age-related variations were identified in the conformation of the deep venous system. The ICV–GV, GV–SS, and SS–SSS angles widened with age (p < .001), while BV–GV angles decreased. BV–BV distance increased progressively from infancy to adolescence (p < .001). The ICV–BV relationship remained stable across age groups. Group comparisons confirmed significant differences between the youngest (0–3 years) and oldest (13–18 years) cohorts for most parameters.

The venous anatomy of the pineal region evolves significantly with age. Understanding these changes can optimize preoperative planning and surgical strategies in pediatric patients. These data provide age-specific anatomical references that may support neurosurgeons worldwide, including those working in low-resource settings without routine access to neuronavigation.