Carotid atherosclerosis is a major contributor to ischemic stroke. While luminal stenosis has historically guided treatment decisions, growing evidence indicates that plaque composition, vascular inflammation and perivascular adipose tissue (PVAT) may be more closely linked to clinical outcomes and plaque vulnerability. This study aimed to characterize carotid PVAT using photon-counting computed tomography (PCCT) and to evaluate its spatial behavior and variability in a cohort of asymptomatic patients. We retrospectively analyzed PCCT angiography data from 20 asymptomatic patients. A custom-developed Python algorithm was used to segment concentric perivascular layers from 1 mm to 5 mm around the carotid artery. For each layer, we quantified attenuation values in Hounsfield Units (HU) and voxel counts. Statistical comparisons were performed across layers and between sides. Mean PVAT attenuation decreased progressively with increasing distance from the carotid wall. Significant differences were observed between inner and outer layers, particularly between the 1 mm and 3–5 mm annuli. Circle-by-circle analysis revealed substantial inter-individual variability in HU trends. Voxel count increased with annular thickness, but variability (SD and CV) also rose in outer layers. No significant differences were found between left and right carotid arteries in either attenuation or voxel distribution. Photon-counting CT enables detailed, layer-specific assessment of carotid PVAT. The observed attenuation patterns and inter-individual variability suggest that PVAT profiling may provide valuable insights into local vascular inflammation and plaque vulnerability. These findings support the potential of PCCT as a noninvasive tool for vascular risk stratification beyond luminal stenosis. Question Can photon-counting CT enable a reliable, layer-by-layer quantitative characterization of carotid perivascular adipose tissue in asymptomatic patients beyond luminal stenosis assessment? Findings Photon-counting CT demonstrated a progressive decrease in PVAT attenuation with increasing distance from the carotid wall and marked inter-individual variability across concentric layers. Clinical relevance Layer-specific PVAT profiling with photon-counting CT may provide a noninvasive imaging marker of local vascular inflammation, supporting improved carotid risk stratification beyond stenosis severity, even in asymptomatic individuals.

To provide a useful and practical Machine Learning framework to facilitate the diagnosis of Neuropsychiatric Systemic Lupus Erythematosus (NPSLE) and Systemic Lupus Erythematosus (SLE) from Magnetic Resonance Imaging (MRI) derived features. Twenty-seven SLE patients (14 NPSLE, 13 SLE; 24 females and 3 males; average age: 43 years, age range: 21 to 62) and 20 healthy controls (17 females and 3 males; average age 41, age range: 21 to 56), were included in this cross-sectional study. VolBrain online platform was used to quantitatively assess brain structural features (regional cortical thickness) which were used as input for the Machine Learning models. Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF) and XGBoost were trained and tested, using a fivefold cross validation in the process. The Random Forest model demonstrated superior performance with an accuracy of 90

BackgroundNative T1 mapping can quantitatively depict small variations in tissue T1 values.PurposeTo evaluate T1 value distribution and differences within affected ileum, in active Crohn's disease (CD) patients, using native T1-maps derived from axial Modified Look-Locker inversion recovery (MOLLI) sequences.Material and MethodsMOLLI sequences were added to the magnetic resonance enterography (MRE) standard protocol. The mid-point of the ileum with the most significant wall thickening was selected. The T1 values were first measured on a single region of interest (ROI), then divided into two, and further into six equal segments (mesenteric: posterior [P], medial posterior [MP], lateral posterior [LP]; anti-mesenteric: anterior [A], medial anterior [MA], lateral anterior [LA]). The Wilcoxon signed-rank test was used to compare T1 values.ResultsIn total, 40 patients with active CD (sMARIA ≥2) were retrospectively included (22 men, 18 women; mean age = 51.1 ± 14.3 years). The cohort was stratified into two groups: mild-to-moderate (wall thickness = 4-8 mm, n = 20) and moderate-to-severe (>8 mm, n = 20). In the mild-to-moderate group, significant differences were found in anti-mesenteric side segments, between MA vs. A (z = 2.606; p = 0.008) and MA vs. LA (z = 2.014; p = 0.038); and comparing opposite segments, between P vs. A (z = 2.014; p = 0.038); P vs. LA (z = 2.132; p = 0.028) and MP vs. A segments (z = 2.014; p = 0.038). In the moderate-to-severe group, significant differences were found between A vs. LP (z = 2.014; p = 0.038) and LA vs. LP (z = 2.606; p = 0.008).ConclusionNative T1 maps can display heterogeneous T1 value distribution within active severe CD ileal segments, particularly in mild-to-moderate wall thickness patients. The T1 value could be a promising imaging biomarker for patient phenotyping.

BackgroundCarotid 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.MethodsStandardized 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.ResultsAmong 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.ConclusionCarotid 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 transition from qualitative to quantitative analysis marks a pivotal shift in medical sciences, moving from subjective assessments to objective, data-driven approaches. Historically, fields like laboratory medicine and genetics relied on qualitative methods, which often resulted in variability and inconsistency. The introduction of quantitative techniques, such as spectrophotometry in laboratory medicine and PCR in genetics, revolutionized diagnostics and research, leading to standardized criteria and personalized medicine. Radiology has embraced, and pathology is also embracing this transition, with quantitative imaging biomarkers and digital pathology enhancing diagnostic accuracy and treatment planning. Artificial Intelligence (AI) is accelerating this shift, transforming medical practice by analyzing vast amounts of data with unprecedented speed and precision. In radiology, AI algorithms automate image interpretation, reducing variability and improving diagnostic accuracy. In pathology, AI enhances the quantification of histological features, providing reproducible and objective data. AI-driven predictive models forecast disease progression and patient outcomes, guiding personalized treatment strategies. However, the integration of AI introduces challenges, particularly the “black-box” phenomenon, where complex AI algorithms make accurate predictions without transparent decision-making processes. This lack of transparency can hinder validation, accountability, and trust. Additionally, the potential for bias in AI models necessitates rigorous training on large cohorts of patients, internal and external validation, and testing. A balanced regulatory framework is crucial to ensure the safe and ethical use of AI in healthcare. This includes ensuring transparency, rigorous validation, ethical standards, and continuous monitoring. By achieving this balance, the continued evolution of quantitative analysis and AI integration can enhance patient outcomes, ushering in a new era of precision medicine where data-driven insights inform every aspect of healthcare, from prevention to diagnosis and treatment. In conclusion, the shift to quantitative analysis and the integration of AI, or better, its “embedding” in clinical workflows, represents a fundamental breakthrough in medical sciences, driving innovation and improving the precision and reliability of medical practice.

Atherosclerotic abdominal aortic aneurysms (AAAs) are a common vascular pathology with significant morbidity and mortality risks. Timely diagnosis, accurate characterization, and standardized reporting are critical for effective management and monitoring of atherosclerotic AAAs. Imaging modalities, particularly computed tomography angiography (CTA), play a pivotal role in the detection, treatment planning, and identification of both primary and secondary complications, as well as distinguishing AAAs from other etiologies. This narrative review provides a comprehensive exploration of the spectrum of imaging findings in atherosclerotic AAAs on CTA, underscoring the importance of structured reporting. Additionally, it examines therapeutic approaches and complications, and it differentiates AAAs from inflammatory, mycotic, and traumatic variants, serving as a primer for radiologists in AAA evaluation.

Coronary plaque vulnerability, more than luminal stenosis, drives acute coronary syndromes. Optical coherence tomography (OCT), intravascular ultrasound (IVUS), and coronary computed tomography angiography (CCTA) visualize plaque morphology in vivo, but manual interpretation is time-consuming and operator-dependent. We performed a narrative literature survey of artificial intelligence (AI) applications—focusing on machine learning (ML) architectures—for automated coronary plaque segmentation and risk characterization across OCT, IVUS, and CCTA. Recent ML models achieve expert-level lumen and plaque segmentation, reliably detecting features linked to vulnerability such as a lipid-rich necrotic core, calcification, positive remodelling, and a napkin-ring sign. Integrative radiomic and multimodal frameworks further improve prognostic stratification for major adverse cardiac events. Nonetheless, progress is constrained by small, single-centre datasets, heterogeneous validation metrics, and limited model interpretability. AI-enhanced plaque assessment offers rapid, reproducible, and comprehensive coronary imaging analysis. Future work should prioritize large multicentre repositories, explainable architectures, and prospective outcome-oriented validation to enable routine clinical adoption.

This study examines the interaction between age and sex on hippocampal volume across the lifespan, focusing on structural changes and asymmetry patterns. Extensive research has demonstrated sex differences in hippocampal structure and function, but the effects of aging on these patterns remain underexplored. Our study aims to elucidate how age-related changes in hippocampal volume differ between men and women, and their implications for cognitive and emotional processes. We performed correlational analyses of hippocampal subfield volumes, total hippocampal volume, and asymmetry indices in males and females. Age-related trends were assessed using linear mixed models (LMM), focusing on hippocampal subfields such as CA1, CA2-3, and CA4-DG. Sex-specific aging patterns were also examined using principal component analysis (PCA) to account for variance in hippocampal volume. Males showed stronger asymmetry patterns, particularly between CA4-DG and CA2-3 subfields (p < 0.001), whereas females showed a distinct asymmetry pattern between CA4-DG and total hippocampal asymmetry (p < 0.001). Significant age-related hippocampal volume loss was observed in males (p = 0.003) but not in females (slope = 0.0026, p = 0.673). Sex differences were most pronounced in the 40s and 50s groups (p < 0.005). Additionally, males showed significant increase in CA1

Atherosclerotic plaque buildup in the coronary and carotid arteries is pivotal in the onset of acute myocardial infarctions or cerebrovascular events, leading to heightened levels of illness and death. Atherosclerosis is a complex and multistep disease, beginning with the deposition of low-density lipoproteins in the arterial intima and culminating in plaque rupture. Modern technology favors non-invasive imaging techniques to assess atherosclerotic plaque and offer insights beyond mere artery stenosis. Among these, computed tomography stands out for its widespread clinical adoption and is prized for its speed and accessibility. Nonetheless, some limitations persist. The introduction of photon-counting computed tomography (PCCT), with its multi-energy capabilities, enhanced spatial resolution, and superior soft tissue contrast with minimal electronic noise, brings significant advantages to carotid and coronary artery imaging, enabling a more comprehensive examination of atherosclerotic plaque composition. This narrative review aims to provide a comprehensive overview of the main concepts related to PCCT. Additionally, we aim to explore the existing literature on the clinical application of PCCT in assessing atherosclerotic plaque. Finally, we will examine the advantages and limitations of this recently introduced technology.

BACKGROUND: This study aimed to develop and validate a computed tomography angiography based machine learning model that uses plaque composition data and degree of carotid stenosis to detect symptomatic carotid plaques in patients with carotid atherosclerosis. METHODS: The machine learning based model was trained using degree of stenosis and the volumes of 13 computed tomography angiography derived intracarotid plaque subcomponents (eg, lipid, intraplaque hemorrhage, calcium) to identify plaques associated with cerebrovascular events. The model was internally validated through repeated 10-fold cross-validation and tested on a dedicated testing cohort according to discrimination and calibration. RESULTS: This retrospective, single-center study evaluated computed tomography angiography scans of 268 patients with both symptomatic and asymptomatic carotid atherosclerosis (163 for the derivation set and 106 for the testing set) performed between March 2013 and October 2019. The area-under-receiver-operating characteristics curve by machine learning on the testing cohort (0.89) was significantly higher than the areas under the curve of traditional logit analysis based on the degree of stenosis (0.51, P<0.001), presence of intraplaque hemorrhage (0.69, P<0.001), and plaque composition (0.78, P<0.001), respectively. Comparable performance was obtained on internal validation. The identified plaque components and associated cutoff values that were significantly associated with a higher likelihood of symptomatic status after adjustment were the ratio of intraplaque hemorrhage to lipid volume (>= 50%, 38.5 [10.1-205.1]; odds ratio, 95% CI) and percentage of intraplaque hemorrhage volume (>= 10%, 18.5 [5.7-69.4]; odds ratio, 95% CI). CONCLUSIONS: This study presented an interpretable machine learning model that accurately identifies symptomatic carotid plaques using computed tomography angiography derived plaque composition features, aiding clinical decision-making.

Vascular inflammation is widely recognized as an important factor in the atherosclerotic process, particularly in terms of plaque development and progression. Conventional tests, such as measuring circulating inflammatory biomarkers, lack the precision to identify specific areas of vascular inflammation. In this context, noninvasive imaging modalities can detect perivascular fat changes, serving as a marker of vascular inflammation. This review aims to provide a comprehensive overview of the key concepts related to perivascular carotid fat and its pathophysiology. Additionally, we examine the existing literature on the association of pericarotid fat with features of plaque vulnerability and cerebrovascular events. Finally, we scrutinize the advantages and limitations of the noninvasive assessment of pericarotid fat.

BACKGROUND AND PURPOSE: Systemic lupus erythematosus is a complex autoimmune disease known for its diverse clinical manifestations, including neuropsychiatric systemic lupus erythematosus, which impacts a patient's quality of life. Our aim was to explore the relationships among brain MR imaging morphometric findings, neuropsychiatric events, and laboratory values in patients with systemic lupus erythematosus, shedding light on potential volumetric biomarkers and diagnostic indicators for neuropsychiatric systemic lupus erythematosus. MATERIALS AND METHODS: Twenty-seven patients with systemic lupus erythematosus (14 with neuropsychiatric systemic lupus erythematosus, 13 with systemic lupus erythematosus), 24 women and 3 men (average age, 43?years, ranging from 21 to 62 years) were included in this cross-sectional study, along with 10 neuropsychiatric patients as controls. An MR imaging morphometric analysis, with the VolBrain online platform, to quantitatively assess brain structural features and their differences between patients with neuropsychiatric systemic lupus erythematosus and systemic lupus erythematosus, was performed. Correlations and differences between MR imaging morphometric findings and laboratory values, including disease activity scores, such as the Systemic Lupus Erythematosus Disease Activity Index and the Systemic Lupus International Collaborating Clinics Damage Index, were explored. An ordinary least squares regression analysis further explored the Systemic Lupus Erythematosus Disease Activity Index and Systemic Lupus International Collaborating Clinics Damage Index relationship with MR imaging features. RESULTS: For neuropsychiatric systemic lupus erythematosus and non-neuropsychiatric systemic lupus erythematosus, the brain regions with the largest difference in volumetric measurements were the insular central operculum volume (P value = .003) and the occipital cortex thickness (P = .003), which were lower in neuropsychiatric systemic lupus erythematosus. The partial correlation analysis showed that the most correlated morphometric features with neuropsychiatric systemic lupus erythematosus were subcallosal area thickness asymmetry (P < .001) and temporal pole thickness asymmetry (P = .011). The ordinary least squares regression analysis yielded an R-2 of 0.725 for the Systemic Lupus Erythematosus Disease Activity Index score, with calcarine cortex volume as a significant predictor, and an R-2 of 0.715 for the Systemic Lupus International Collaborating Clinics Damage Index score, with medial postcentral gyrus volume as a significant predictor. CONCLUSIONS: The MR imaging volumetric analysis, along with the correlation study and the ordinary least squares regression analysis, revealed significant differences in brain regions and their characteristics between patients with neuropsychiatric systemic lupus erythematosus and those with systemic lupus erythematosus, as well as between patients with different Systemic Lupus Erythematosus Disease Activity Index and Systemic Lupus International Collaborating Clinics Damage Index scores.

Introduction In the current paper, the “carotid artery calcium score” method is presented with the target to offer a metric method to quantify the amount of calcification in the carotid artery. Model and Definition The Volume of Interest (VOI) should be extracted and those voxels, with a Hounsfield Unit (HU) value ≥130, should be considered. The total weight value is determined by calculating the sum of the HU attenuation values of all voxels with values ≥130 HU. This value should be multiplied by the conversion factor (“or voxel size”) and divided by a weighting factor, the attenuation threshold to consider a voxel as calcified (and therefore 130 HU): this equation determines the Carotid Artery Calcium Score (CACS). Results In order to provide the demonstration of the potential feasibility of the model, the CACS was calculated in 131 subjects (94 males; mean age 72.7 years) for 235 carotid arteries (in 27 subjects, unilateral plaque was present) considered. The CACS value ranged from 0.67 to 11716. A statistically significant correlation was found (rho value = 0.663, p value = .0001) between the CACS in the right and left carotid plaques. Moreover, a statistically significant correlation between the age and the total CACS was present (rho value = 0.244, p value = .005), whereas no statistically significant difference was found in the distribution of CACS by gender ( p = .148). The CACS was also tested at baseline and after contrast and no statistically significant difference was found. Conclusion In conclusion, this method is of easy application, and it weights at the same time the volume and the degree of calcification in a unique parameter. This method needs to be tested to verify its potential utility, similar to the coronary artery calcium score, for the risk stratification of the occurrence of cerebrovascular events of the anterior circulation. Further studies using this new diagnostic tool to determine the prognostic value of carotid calcium quantification are needed.

Purpose: The standard bibliometric indexes (“ m-quotient “ H-,” “ H2-,” “ g-,” “ a-,” “ m-,” and “ r-” index) do not considered the research’ position in the author list of the paper. We proposed a new methodology, System of Authorship Best Assessment (SABA), to characterize the scientific output based on authors’ position. Material and Methods: Four classes S1A, S1B, S2A, and S2B include only papers where the researcher is in first, first/last, first/second/last, and first/second/second-last/last position respectively were used for the calculation of H-index and number of citations The system was tested with Noble prize winners controlled with researchers matched for H-index. The different in percentage between standard bibliometric index and S2B was calculated and compared. Results: The percentage differences in Noble prize winners between S2B- H-index versus Global H-index and number of citations is very lower comparing with control group (median 4.15% [adjusted 95% CI, 2.54–5.30] vs 9.00 [adjusted 95% CI, 7.16–11.84], p < 0.001; average difference 8.7% vs 20.3%). All different in percentage between standard bibliometric index and S2B except two ( H2- and m-index) were significantly lower among Noble prize compared with control group. Conclusion: The SABA methodology better weight the research impact by showing that for excellent profiles the S2B is similar to global values whereas for other researchers there is a significant difference.

The global mortality rate is known to be the highest due to cardiovascular disease (CVD). Thus, preventive, and early CVD risk identification in a non-invasive manner is vital as healthcare cost is increasing day by day. Conventional methods for risk prediction of CVD lack robustness due to the non-linear relationship between risk factors and cardiovascular events in multi-ethnic cohorts. Few recently proposed machine learning-based risk stratification reviews without deep learning (DL) integration. The proposed study focuses on CVD risk stratification by the use of techniques mainly solo deep learning (SDL) and hybrid deep learning (HDL). Using a PRISMA model, 286 DL-based CVD studies were selected and analyzed. The databases included were Science Direct, IEEE Xplore, PubMed, and Google Scholar. This review is focused on different SDL and HDL architectures, their characteristics, applications, scientific and clinical validation, along with plaque tissue characterization for CVD/stroke risk stratification. Since signal processing methods are also crucial, the study further briefly presented Electrocardiogram (ECG)-based solutions. Finally, the study presented the risk due to bias in AI systems. The risk of bias tools used were (I) ranking method (RBS), (II) region-based map (RBM), (III) radial bias area (RBA), (IV) prediction model risk of bias assessment tool (PROBAST), and (V) risk of bias in non-randomized studies-of interventions (ROBINS-I). The surrogate carotid ultrasound image was mostly used in the UNet-based DL framework for arterial wall segmentation. Ground truth (GT) selection is vital for reducing the risk of bias (RoB) for CVD risk stratification. It was observed that the convolutional neural network (CNN) algorithms were widely used since the feature extraction process was automated. The ensemble-based DL techniques for risk stratification in CVD are likely to supersede the SDL and HDL paradigms. Due to the reliability, high accuracy, and faster execution on dedicated hardware, these DL methods for CVD risk assessment are powerful and promising. The risk of bias in DL methods can be best reduced by considering multicentre data collection and clinical evaluation.

Purpose: The aims of our study were to investigate with cardiovascular magnetic resonance (CMR) the role of Epicardial Fat Volume (EFV) and distribution in patients with Takotsubo cardiomyopathy (TTC). Moreover, we explored EFV in patients with TTC and related this to comorbidities, cardiac biomarkers, and cardiac function.Methods: This retrospective study performed CMR scans in 30 consecutive TTC patients and 20 healthy controls. The absolute amount of EFV was quantified in consecutive short-axis cine stacks through the modified Simpson's rule. In addition, the left atrio-ventricular groove (LV) and right ventricle (RV) Epicardial Fat Thickness (EFT) were measured as well. Besides epicardial fat, LV myocardial strain parameters and T2 mapping measurements were obtained.Results: TTC patients and controls were of comparable age, sex, and body mass index. Compared to healthy controls, patients with TTC demonstrated a significantly increased EFV, epicardial fat mass, and EFV indexed for body 7surface area (p = 0.005; p = 0.003; p = 0.008; respectively). In a multiple regression model including age, sex, BMI, atrial fibrillation, and dyslipidemia, TTC remained an independent association with EFV (p = 0.008). Global T2 mapping and Global longitudinal strain in patients with TTC were correlated with EFV (r = 0.63, p = 0.001, and r = 0.44, p = 0.02, respectively).Conclusion: Patients with TTC have increased EFV compared to healthy controls, despite a similar body mass index. The amount of epicardial fat was associated with CMR markers of myocardial inflammation and sub -clinical contractile dysfunction.

BACKGROUND:No quantitative computed tomography (CT) biomarker is actually sufficiently accurate to assess Crohn's disease (CD) lesion activity, with adequate precision to guide clinical decisions. PURPOSE:To assess the available literature on the use of iodine concentration (IC), from multi-spectral CT acquisition, as a quantitative parameter able to distinguish healthy from affected bowel and assess CD bowel activity and heterogeneity of activity along the involved segments. MATERIAL AND METHODS:A literature search was conducted to identify original research studies published up to February 2022. The inclusion criteria were original research papers (>10 human participants), English language publications, focus on dual-energy CT (DECT) of CD with iodine quantification (IQ) as an outcome measure. The exclusion criteria were animal-only studies, languages other than English, review articles, case reports, correspondence, and study populations <10 patients. RESULTS:Nine studies were included in this review; all of which showed a strong correlation between IC measurements and CD activity markers, such as CD activity index (CDAI), endoscopy findings and simple endoscopic score for Crohn's disease (SES-CD), and routine CT enterography (CTE) signs and histopathologic score. Statistically significant differences in IC were reported between affected bowel segments and healthy ones (higher P value was P < 0.001), normal segments and those with active inflammation (P < 0.0001) as well as between patients with active disease and those in remission (P < 0.001). CONCLUSION:The mean normalized IC at DECTE could be a reliable tool in assisting radiologists in the diagnosis, classification and grading of CD activity.

The role of calcium in atherosclerosis is controversial and the relationship between vascular calcification and plaque vulnerability is not fully understood. Although calcifications are present in ≈50% to 60% of carotid plaques, their association with cerebrovascular ischemic events remains unclear. In this review, we summarize current understanding of carotid plaque calcification. We outline the role of calcium in atherosclerotic carotid disease by analyzing laboratory studies and histopathologic studies, as well as imaging findings to understand clinical implications of carotid artery calcifications. Differences in mechanism of calcium deposition express themselves into a wide range of calcification phenotypes in carotid plaques. Some patterns, such as rim calcification, are suggestive of plaques with inflammatory activity with leakage of the vasa vasourm and intraplaque hemorrhage. Other patterns such as dense, nodular calcifications may confer greater mechanical stability to the plaque and reduce the risk of embolization for a given degree of plaque size and luminal stenosis. Various distributions and patterns of carotid plaque calcification, often influenced by the underlying systemic pathological condition, have a different role in affecting plaque stability. Modern imaging techniques afford multiple approaches to assess geometry, pattern of distribution, size, and composition of carotid artery calcifications. Future investigations with these novel technologies will further improve our understanding of carotid artery calcification and will play an important role in understanding and minimizing stroke risk in patients with carotid plaques.

Background: The previous COVID-19 lung diagnosis system lacks both scientific validation and the role of explainable artificial intelligence (AI) for understanding lesion localization. This study presents a cloud-based explainable AI, the “COVLIAS 2.0-cXAI” system using four kinds of class activation maps (CAM) models. Methodology: Our cohort consisted of ~6000 CT slices from two sources (Croatia, 80 COVID-19 patients and Italy, 15 control patients). COVLIAS 2.0-cXAI design consisted of three stages: (i) automated lung segmentation using hybrid deep learning ResNet-UNet model by automatic adjustment of Hounsfield units, hyperparameter optimization, and parallel and distributed training, (ii) classification using three kinds of DenseNet (DN) models (DN-121, DN-169, DN-201), and (iii) validation using four kinds of CAM visualization techniques: gradient-weighted class activation mapping (Grad-CAM), Grad-CAM++, score-weighted CAM (Score-CAM), and FasterScore-CAM. The COVLIAS 2.0-cXAI was validated by three trained senior radiologists for its stability and reliability. The Friedman test was also performed on the scores of the three radiologists. Results: The ResNet-UNet segmentation model resulted in dice similarity of 0.96, Jaccard index of 0.93, a correlation coefficient of 0.99, with a figure-of-merit of 95.99%, while the classifier accuracies for the three DN nets (DN-121, DN-169, and DN-201) were 98%, 98%, and 99% with a loss of ~0.003, ~0.0025, and ~0.002 using 50 epochs, respectively. The mean AUC for all three DN models was 0.99 (p < 0.0001). The COVLIAS 2.0-cXAI showed 80% scans for mean alignment index (MAI) between heatmaps and gold standard, a score of four out of five, establishing the system for clinical settings. Conclusions: The COVLIAS 2.0-cXAI successfully showed a cloud-based explainable AI system for lesion localization in lung CT scans.