Citations

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.

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.

Coronary artery calcification (CAC) accompanies the development of advanced atherosclerosis. Its role in atherosclerosis holds great interest because the presence and burden of coronary calcification provide direct evidence of the presence and extent of coronary artery disease; furthermore, CAC predicts future events independently of concomitant conventional cardiovascular risk factors and to a greater extent than any other noninvasive biomarker of this disease. Nevertheless, the relationship between CAC and the susceptibility of a plaque to provoke a thrombotic event remains incompletely understood. This review summarizes the current understanding and literature on CAC. It outlines the pathophysiology of CAC and reviews laboratory, histopathological, and genetic studies, as well as imaging findings, to characterize different types of calcification and to elucidate their implications. Some patterns of calcification such as microcalcification portend increased risk of rupture and cardiovascular events and may improve prognosis assessment noninvasively. However, contemporary computed tomography cannot assess early microcalcification. Limited spatial resolution and blooming artifacts may hinder estimation of degree of coronary artery stenosis. Technical advances such as photon counting detectors and combination with nuclear approaches (eg, NaF imaging) promise to improve the performance of cardiac computed tomography. These innovations may speed achieving the ultimate goal of providing noninvasively specific and clinically actionable information.

Background Distal medium vessel occlusions (DMVOs) and minor strokes represent emerging frontiers in mechanical thrombectomy (MT). Although several randomized clinical trials (RCTs) are underway, the design characteristics of these trials and the specific questions they aim to address have not been extensively explored. This current study sought to investigate the design and data elements reported in active prospective DMVO and minor stroke studies. Methods The ClinicalTrials.gov database was searched for ongoing prospective studies assessing the role of MT in patients with DMVOs or minor strokes. The Nested Knowledge AutoLit platform was utilized to categorize reported outcomes and inclusion/exclusion criteria. Frequencies of reported data elements were extracted from study protocols. Results A total of 10 (8 DMVO and 2 minor stroke) studies enrolling 3520 patients were included. All DMVO studies employ different criteria regarding target occlusion locations. Five DMVO studies use stent retrievers as the first-line thrombectomy technique (62.5%, 5/8), while three studies allow any MT techniques, generally at the operator’s discretion. Four DMVO studies permit intravenous thrombolysis (IVT) utilization in both intervention and control arms (50%, 4/8). The DISTALS trial excludes patients receiving IVT, while the DUSK trial and Tigertriever registry only enroll patients who are ineligible for IVT or for whom IVT failed to achieve reperfusion. DMVO studies exhibit notable heterogeneity in symptom onset duration thresholds for inclusion (<6 hours: 2 studies; <12 hours: 2 studies; <24 hours: 3 studies). Minor stroke trials employ similar inclusion criteria and outcome measures except for symptom duration thresholds for inclusion (8 hours for ENDOLOW and 23 hours for MOSTE). Conclusions There is considerable heterogeneity among active DMVO trials regarding potential target DMVO locations and time thresholds for inclusion based on the last known well time. Furthermore, our review indicates that the utility of aspiration thrombectomy in DMVOs and the advantages of MT without IVT over IVT alone will remain largely unexplored even after completion of active DMVO trials.

Carotid artery stenosis is a major cause of morbidity and mortality. The journey to understanding carotid disease has developed over time and radiology has a pivotal role in diagnosis, risk stratification and therapeutic management. This paper reviews the history of diagnostic imaging in carotid disease, its evolution towards its current applications in the clinical and research fields, and the potential of new technologies to aid clinicians in identifying the disease and tailoring medical and surgical treatment.

Introduction: The carotid artery rim sign on CTA is characterized by soft-plaque components surrounded by a rind of thin, adventitial calcification and its role in plaque instability is not established yet. Intraplaque hemorrhage (IPH) and the inflammatory microenvironment in carotid atheroma are considered features of atherosclerotic plaque vulnerability leading to cerebrovascular events. Hypothesis: The association between rim sign on CTA and carotid IPH and inflammation remains scantly investigated. In this study, we aimed to determine whether rim sign is associated with IPH on MRA and with the extent of IPH and inflammation on histology. We hypothesized that rim sign may be a radiological predictor of carotid IPH and inflammation. Methods: We retrospectively enrolled 100 consecutive patients who underwent carotid endarterectomy (CEA) and neck CTA before CEA. On CTA, the ipsilateral rim sign was defined as the presence of adventitial calcifications with internal soft-tissue plaque of ≥2 mm in maximum thickness. Of the included patients, 25 had MRA with 3D magnetization-prepared rapid acquisition with gradient echo (MPRAGE) sequence. Carotid plaque specimens were collected and stained with H&E to quantify IPH (%). Histologically, we defined IPH as an area of carotid plaque with blood elements, and inflammatory microenvironment as the presence of lymphocytes and plasma cells (chronic inflammation) and foam cells. Results: The rim sign on CTA was present in 56 (56%) of patients. IPH on MRA was present in 11 (44%) patients. The rim sign was significantly associated with the presence of IPH on MRA [9 (81.8%) vs 2 (14.3%); p=0.003]. On histology, the rim sign was associated with a higher extent of IPH [5.0 % (IQR 2.0-25.0) vs 5.0% (IQR 0.0-10.0); p=0.017], with chronic inflammation [44 (78.6%) vs 12 (21.4%); p=0.003], and foam cells [51 (91.1%) vs 5 (8.9%); p<0.001]. Conclusions: These findings suggest that rim sign on CTA may be proposed as a potential imaging marker of IPH as well as a histologic predictor of IPH extent and carotid plaque inflammatory microenvironment.

Abstract Background and Motivation: Due to the intricate relationship between the small non-coding ribonucleic acid (miRNA) sequences, the classification of miRNA species, namely Human, Gorilla, Rat, and Mouse is challenging. Previous methods are not robust and accurate. In this study, we present GeneAI 3.0 (AtheroPoint™, Roseville, CA, USA), a powerful, novel, and generalized method for extracting features from the fixed patterns of purines and pyrimidines in each miRNA sequence in ensemble paradigms in machine learning (EML) and convolutional neural network (CNN)-based deep learning (EDL) frameworks. Method: GeneAI 3.0 utilized five conventional (Entropy, Dissimilarity, Energy, Homogeneity, and Contrast), and three contemporary (Shannon entropy, Hurst exponent, Fractal dimension) features, to generate a compositefeature set from given miRNA sequences which were then passed into our ML and DL classification framework. A set of 11 new classifiers was designed consisting of five EML and six EDL for binary/multiclass classification. It was benchmarked against 9 solo ML (SML), 6 solo DL (SDL), 12 hybrid DL (HDL) models, resulting in a total of 11+27=38 models were designed. Four hypotheses were formulated and validated using explainable AI (XAI) as well as reliability/statistical tests. Results: The order of the mean performance using accuracy (ACC)/area-under-the-curve (AUC) of the 24 DL classifiers was: EDL>HDL>SDL. The mean performance of EDL models with CNN layers was superior to that without CNN layers by 0.73%/0.92%. Mean performance of EML models was superior to SML models with improvements of ACC/AUC by 6.24%/6.46%. EDL models performed significantly better than EML models, with a mean increase in ACC/AUC of 7.09%/6.96%. The GeneAI 3.0 tool produced expected XAI feature plots, and the statistical tests showed significant p-values. Conclusions: Ensemble models with composite features are highly effective and generalized models for effectively classifying miRNA sequences.

Background and motivation: Lung computed tomography (CT) techniques are high-resolution and are well adopted in the intensive care unit (ICU) for COVID-19 disease control classification. Most artificial intelligence (AI) systems do not undergo generalization and are typically overfitted. Such trained AI systems are not practical for clinical settings and therefore do not give accurate results when executed on unseen data sets. We hypothesize that ensemble deep learning (EDL) is superior to deep transfer learning (TL) in both non-augmented and augmented frameworks. Methodology: The system consists of a cascade of quality control, ResNet–UNet-based hybrid deep learning for lung segmentation, and seven models using TL-based classification followed by five types of EDL’s. To prove our hypothesis, five different kinds of data combinations (DC) were designed using a combination of two multicenter cohorts—Croatia (80 COVID) and Italy (72 COVID and 30 controls)—leading to 12,000 CT slices. As part of generalization, the system was tested on unseen data and statistically tested for reliability/stability. Results: Using the K5 (80:20) cross-validation protocol on the balanced and augmented dataset, the five DC datasets improved TL mean accuracy by 3.32%, 6.56%, 12.96%, 47.1%, and 2.78%, respectively. The five EDL systems showed improvements in accuracy of 2.12%, 5.78%, 6.72%, 32.05%, and 2.40%, thus validating our hypothesis. All statistical tests proved positive for reliability and stability. Conclusion: EDL showed superior performance to TL systems for both (a) unbalanced and unaugmented and (b) balanced and augmented datasets for both (i) seen and (ii) unseen paradigms, validating both our hypotheses.