Existing multi-modality segmentation networks tend to aggregate heterogeneous features without discrimination, thereby overlooking the intrinsic discrepancy between ultrasound (US) and contrast-enhanced ultrasound (CEUS) in anatomical boundary representation and functional perfusion response. Even recent asymmetric architectures still mainly rely on spatial-domain fusion, which is vulnerable to ultrasound speckle noise and low-contrast boundaries, leading to insufficient cross-modal alignment and inaccurate plaque delineation. To alleviate these limitations, this paper proposes WAAF-Net, a wavelet-guided asymmetric attention fusion network tailored for dual-modal carotid plaque segmentation. Specifically, WAAF-Net adopts an asymmetric two-stream backbone to accommodate the modality-specific characteristics of US and CEUS, where the CNN branch focuses on local anatomical boundaries and textural patterns from US images, while the Transformer branch models CEUS-related cross-regional perfusion semantics and global contextual dependencies. Then, the Hybrid Wavelet Enhancement Block (HWEB) and Wavelet Fusion Enhancement Module (WFEM) extend cross-modal interaction from the spatial domain to the frequency domain. By explicitly decomposing features into low-frequency structural components and high-frequency boundary details, they enhance modality-specific representations while suppressing noise interference. Subsequently, the Shared-Projection Affinity Fusion Block (SPAFB) maps heterogeneous features into a common latent space, where affinity relationship modeling is performed to align and enhance modality-shared responses while compensating for modality-discrepant regions. Finally, the Prediction-Guided Feature Enhancement Block (PGFEB) introduces intermediate prediction maps as regional guidance to progressively aggregate multi-scale heterogeneous features, thereby recovering complex plaque morphology and weak boundary details. These three core modules follow a progressive design principle from modality-specific enhancement, to shared feature alignment, and further to multi-stage feature aggregation, forming a tightly coupled dual-modal segmentation framework. Experiments on a clinical US/CEUS carotid plaque dataset and the public CHAOS dataset demonstrate that WAAF-Net achieves superior segmentation performance and more robust generalization compared with state-of-the-art multi-modality baselines.

To develop and validate a multicenter ultrasound-based predictive model for fluorescence in situ hybridization (FISH) results in HER2 (2+) breast cancer patients, aiming to provide a convenient and cost-effective tool to support clinical decision-making. In this retrospective multicenter study, 5,888 breast cancer patients from six institutions were included. Radiomics features were extracted from ultrasound images using PyRadiomics, and deep learning features were obtained using a Vision Transformer (ViT). Feature selection was performed using least absolute shrinkage and selection operator (LASSO) regression. Multiple machine learning models were developed, and their performance was evaluated with the area under the curve (AUC). The DeLong test was used for model comparison. The proportion of FISH-positive cases ranged from 9.7

OBJECTIVE:To investigate whether the composite ultrasound feature-characterized by intraplaque punctate calcification with contrast enhancement-serves as an independent predictor for new ischemic stroke in these presumed "low-risk" plaques, and to develop and validate a nomogram prediction model that integrates clinical and imaging features. METHODS:In this retrospective study, 3317 patients initially underwent carotid ultrasound. After stringent exclusions, 404 with CP-RADS grade 2 plaques were enrolled and randomly split 7:3 into training (n = 282) and validation (n = 121) sets. Three senior doctors, blinded to clinical data, independently assessed ultrasound features with excellent intra‑ and inter‑observer agreement (ICC > 0.80). Univariate and multivariate logistic regression identified independent risk factors, and a nomogram was developed. Model performance was evaluated using ROC analysis, calibration curves, and decision curve analysis (DCA). RESULTS:The composite imaging feature (punctate calcification with contrast enhancement) was an independent predictor of new‑onset cerebral infarction (OR = 3.60, 95% CI: 1.33-9.77, p = 0.012). The nomogram showed excellent discrimination in the validation set (AUC = 0.84; sensitivity 79%, specificity 86%), with good calibration between predicted and actual risk. CONCLUSION:The composite "active plaque" feature-intraplaque punctate calcification with contrast enhancement-serves as an independent predictor of new‑onset cerebral infarction in CP‑RADS grade 2 carotid plaques. Our nomogram demonstrates good predictive performance and clinical applicability, providing reliable imaging support for early risk stratification and individualized intervention in this presumed "low‑risk" population.

Analyzing the correlation between carotid stiffness and atherosclerotic cardiovascular disease (ASCVD) risk, and determine elastic modulus (EM) threshold. In this single-center, cross-sectional retrospective observational study, 229 patients without carotid plaque were consecutively enrolled between July 2021 and August 2022. ASCVD risk was calculated by simpler Framingham model and categorized into low (N = 108), moderate (N = 58) and high risk groups (N = 63). Shear-wave elastography was used to quantify carotid stiffness, and the resulting values were analyzed for their association with ASCVD risk. Among 229 participants without carotid plaque, 108 (47.2

Aims to access the diagnostic performance of different regions ultrasound viscoelasticity parameters especially margin of breast cancer and to determine whether the use of margin viscoelasticity can improve its accuracy in the Breast Imaging Reporting and Data System (BI-RADS). 234 benign and 90 malignant lesions were subjected to standard breast ultrasound and viscoelasticity examinations. The doctors selected region of interest (ROI) to measure viscoelasticity. ROI-1, ROI-2, and ROI-3 represent the tumor, peritumoral, and peripheral areas, respectively. The viscosity modulus (VMean, VMin, VMax, VSD) and elasticity modulus (EMean, EMin, EMax, ESD) of 3 ROIs were analyzed. The diagnostic performance of viscoelasticity of three regions was assessed by receiver operating characteristic curves (ROC). Comparison of the effectiveness of B-Mode ultrasound and viscoelastic parameters in BI-RADS diagnosis of breast cancer based on true positive (TP) and false negative (FN). The optimal viscoelasticity-related parameters for differentiating breast lesions were determined to be 2-EMax and 2-VMax, with area under the curve (AUC) values of 0.84 (0.79–0.90) and 0.85 (0.80–0.90), respectively. Using ≥ 30.4 kPa and ≥ 3.3 Pa·s as cutoff values, the BI-RADS classification was then modified. The joint model improves diagnoses of benign lesions in category 4 (69/73). In the same way, 2-EMin + 2-VMin can improve the diagnosis rate of benign lesions (227/234). Viscoelastic parameters have better diagnostic performance than viscosity and elasticity alone. Ultrasound quantitative viscoelasticity parameters of breast mass especially lesion margin can show more comprehensive information. Viscoelasticity improves diagnostic accuracy of BI-RADS categories and reduces unnecessary biopsies. Peritumoral ultrasound viscoelasticity should also be taken seriously, providing a new idea for early detection of breast cancer. It potentially improves diagnostic accuracy of BI-RADS assessment for patients with breast lesions.

Objectives Conventional ultrasound (US) elastography lacks specificity in distinguishing benign from malignant breast lesions. This study employed US to assess breast tissue viscosity and elasticity. The primary objective was to evaluate the diagnostic performance of US-derived viscoelasticity parameters. Secondary objectives included investigating the consistency of parameters in the mechanical properties of breast tissue. Materials and methods Two doctors independently measured the viscosity and elasticity of specific positions in the breasts of 20 health females for consistency assessment. Then the doctors selected region of interest (ROI) to measure viscoelasticity. ROI-1, ROI-2, and ROI-3 represent the tumor, peritumoral, and peripheral areas, respectively. The viscosity modulus and elasticity modulus of 3 ROIs were analyzed. The viscosity and elasticity parameters with the highest area under the curve (AUC) were selected as the optimal ones. Finally, elasticity and viscosity parameters were combined to assess their diagnostic performance in differentiating breast lesions. Results US viscoelasticity parameters can be measured with high consistency. Among conventional US elasticity parameters, 1-Emax demonstrated the highest AUC (0.746) for differentiating benign and malignant breast lesions. In US viscoelasticity parameters, 2-Emax achieved the highest AUC of 0.801, while 2-Vmax showed the highest AUC of 0.835. The highest specificity (0.903) was observed when both 2-Emax and 2-Vmax exceeded their respective cutoff values (P u0026lt; 0.05 for all). Conclusion Quantitative ultrasound viscoelasticity parameters play a crucial role in breast cancer diagnosis, with tumor boundary parameters being particularly significant for cancer screening and prevention strategies.

Ultrasonography (US) and contrast-enhanced ultrasound (CEUS) are effective imaging tools for analyzing the spatial and temporal characteristics of lesions and diagnosing or predicting diseases. At the same time, US is characterized by blurred boundaries and strong noise interference. Therefore, evaluating plaques and depicting lesions frame-by-frame is a time-consuming task, which poses a challenge in analyzing US videos using deep learning techniques. However, despite the existing methods for US and CEUS image segmentation, there are still limited approaches capable of integrating the feature information from these two distinct image types. Furthermore, these methods require additional optimization to enhance their capacity for extracting comprehensive global contextual information. To address the problem, we propose a U-shaped structured network model based on Transformer in this paper. The network is composed of two parts, that is, the dual-modal information interaction fusion module and the enhanced feature extraction module. The first module is used to extract comprehensive US and CEUS features and fuse them at multiple scales. The second module is used to enhance feature extraction capabilities. This network enables precise localization of the lesion and clear depiction of the region of interest in US. Our model achieved a Dice of 91.62% and an IoU of 88.04% on the carotid plaque segmentation dataset. The experimental results show that the performance of our designed network on the carotid artery dataset is better than that of the SOTA models.

ObjectiveTo evaluate the diagnostic value of 3.0T high-resolution magnetic resonance imaging (3.0T HR-MRI), ultrasound imaging, and GATA3 protein expression in breast cancer (BC) and their prognostic implications.MethodsA retrospective analysis of 143 BC patients was conducted. All patients underwent preoperative 3.0T HR-MRI and ultrasound examinations. Diffusion tensor imaging (DTI) parameters, including mean diffusivity (MD), fractional anisotropy (FA), radial diffusivity (Dr), and axial diffusivity (Da), were assessed. Ultrasound features such as tumor morphology, vascularity, and posterior acoustic characteristics were analyzed. Immunohistochemistry was performed to detect GATA3 expression, and correlations with imaging parameters and prognosis were evaluated.ResultsGATA3 expression was significantly associated with BC pathological subtypes ( x2= 26.59, P < 0.0001). Compared with the GATA3-negative group, GATA3-positive tumors exhibited lower FA but higher MD, Dr, and Da values (P < 0.05), suggesting more preserved tissue structure. Ultrasound analysis showed that GATA3-positive tumors had less vascularity, posterior attenuation, and irregular margins (P < 0.05). Poor prognosis was associated with higher FA and lower MD, Dr, and Da values (P < 0.0001), as well as more aggressive ultrasound features. ROC analysis demonstrated superior prognostic performance when combining GATA3 expression, MRI, and ultrasound parameters (AUC = 0.9695, sensitivity = 83.54%, specificity = 96.88%).Conclusion3.0T HR-MRI and ultrasound provide complementary insights into BC characteristics. GATA3 expression is associated with better prognosis, and their combined analysis enhances diagnostic accuracy and prognostic evaluation.

Freehand 3D ultrasound enables volumetric imaging by tracking a conventional ultrasound probe during freehand scanning, offering enriched spatial information that improves clinical diagnosis. However, the quality of reconstructed volumes is often compromised by tracking system noise and irregular probe movements, leading to artifacts in the final reconstruction. To address these challenges, we propose ImplicitCell, a novel framework that integrates Implicit Neural Representation (INR) with an ultrasound resolution cell model for joint optimization of volume reconstruction and pose refinement. Three distinct datasets are used for comprehensive validation, including phantom, common carotid artery, and carotid atherosclerosis. Experimental results demonstrate that ImplicitCell significantly reduces reconstruction artifacts and improves volume quality compared to existing methods, particularly in challenging scenarios with noisy tracking data. These improvements enhance the clinical utility of freehand 3D ultrasound by providing more reliable and precise diagnostic information.

The objective of this study is to develop a deep-learning-based detection and diagnosis technique for carotid atherosclerosis (CA) using a portable freehand 3-D ultrasound (US) imaging system. A total of 127 3-D carotid artery scans were acquired using a portable 3-D US system, which consisted of a handheld US scanner and an electromagnetic (EM) tracking system. A U-Net segmentation network was first applied to extract the carotid artery on 2-D transverse frame, and then, a novel 3-D reconstruction algorithm using fast dot projection (FDP) method with position regularization was proposed to reconstruct the carotid artery volume. Furthermore, a convolutional neural network (CNN) was used to classify healthy and diseased cases qualitatively. Three-dimensional volume analysis methods, including longitudinal image acquisition and stenosis grade measurement, were developed to obtain the clinical metrics quantitatively. The proposed system achieved a sensitivity of 0.71, a specificity of 0.85, and an accuracy of 0.80 for diagnosis of CA. The automatically measured stenosis grade illustrated a good correlation ( r = 0.76) with the experienced expert measurement. The developed technique based on 3-D US imaging can be applied to the automatic diagnosis of CA. The proposed deep-learning-based technique was specially designed for a portable 3-D freehand US system, which can provide a more convenient CA examination and decrease the dependence on the clinician's experience.

Vulnerable carotid plaques are extremely unstable and prone to rupture and fall off, which are closely related to transient ischemic attacks and ischemic strokes. Portable 3D ultrasound is a radiation-free and non-invasive technique that can conveniently provide more comprehensive dimensional information compared to 2D ultrasound. Five types of feature parameters including the carotid volumetric stenosis rate (CSR), low-intensity rate (LIR), grayscale median (GSM), fractal dimension (FD), and 3D gray level co-occurrence matrix (GLCM) properties were extracted from the 3D image volumes respectively and input into a support vector machine (SVM) classifier. The average accuracy of the SVM model was 0.73 +/- 0.05, with a sensitivity of 0.75 +/- 0.08 and a specificity of 0.74 +/- 0.05. The SVM classifier using extracted features as input performed acceptably in the classification of carotid plaque vulnerability since dimension, grayscale, spatial structure, and texture features were considered. It demonstrated that the proposed method has the potential to screen and diagnose carotid plaques based on portable 3D ultrasound.

Breast cancer remains a global health challenge, contributing significantly to mortality worldwide. Addressing this critical issue, we review existing machine-learning algorithms for accurate breast cancer diagnosis. Our research employs an ensemble of algorithms, including Gaussian Naive Bayes, XGBoost, Support Vector Machine, Logistic Regression, Principal Component Analysis, Linear Discriminant Analysis, k Nearest Neighbors, Random Forest, Decision Tree Classifier, and an ensemble classifier. This study relies on specialized data preprocessing and balancing techniques, ensuring the reliability of the analysis. Our approach utilizes two prominent datasets, the Wisconsin Breast Cancer Diagnosis (WBCD) and the Wisconsin Diagnostic Breast Cancer (WDBC) dataset, carefully partitioned with a 5-fold cross-validation strategy. The evaluation protocol is comprehensive, spanning diverse performance metrics. We explore confusion matrices, accuracy, precision, recall, F1 score, the area under the curve (AUC), and Receiver Operating Characteristic (ROC) curves. Notably, in classifying benign and malignant tumors using the WBCD dataset, the SVM model stands out with a detection accuracy of 99.27

OBJECTIVES:To investigate the added value of strain elastography (SE) by recategorizing ultrasound (US) breast imaging reporting and data system (BI-RADS) 3 and 4a lesions. METHODS:A total of 4371 patients underwent US and SE with BI-RADS 2-5 categories solid breast lesions were included from 32 hospitals. We evaluated the elastographic images according to elasticity scores (ES) and strain ratios (SR). Three combined methods (BI-RADS + ES, BI-RADS + SR, BI-RADS + ES + SR) and two reclassified methods were used (method one: upgrading BI-RADS 3 and downgrading BI-RADS 4a, method two: downgrading BI-RADS 4a alone). The diagnostic performance and the potential reduction of unnecessary biopsies were evaluated. RESULTS:Combining BI-RADS with SE had a higher area under the curve (AUC) than BI-RADS alone (0.822-0.898 vs 0.794, P < .01). For reclassified method one, the sensitivity, specificity, and accuracy were 99.36%, 66.70%, 78.36% for BI-RADS + ES and 98.01%, 66.45%, 77.72% for BI-RADS + SR, and 99.42%, 66.70%, 78.38% for BI-RADS + ES + SR, respectively. For reclassified method two, the sensitivity, specificity, and accuracy were 99.17%, 70.72%, 80.87% for BI-RADS + ES and 97.76%, 81.75%, 87.46% for BI-RADS + SR, and 99.23%, 69.83%, 80.32% for BI-RADS + ES + SR, respectively. Downgrading BI-RADS 4a alone had higher AUC, specificity, and accuracy (P < .01) and similar sensitivity (P > .05) to upgrading BI-RADS 3 and downgrading BI-RADS 4a. Combining SE with BI-RADS could help reduce unnecessary biopsies by 17.64%-55.20%. CONCLUSIONS:Combining BI-RADS with SE improved the diagnostic performance in distinguishing benign from malignant lesions and could decrease false-positive breast biopsy rates. Downgrading BI-RADS 4a lesions alone might be sufficient for achieving good diagnostic performance. ADVANCES IN KNOWLEDGE:Downgrading BI-RADS category 4a lesions alone had higher AUC, specificity, and accuracy, and similar sensitivity to upgrading or downgrading BI-RADS category 3 and 4a lesions.

Background Noninvasive evaluation of metabolic dysfunction-associated fatty liver disease (MAFLD) with multiparametric US is essential, but multicenter studies are lacking. Purpose To evaluate the ability of multiparametric US with attenuation imaging (ATI) and two-dimensional (2D) shear-wave elastography (SWE) for predicting metabolic dysfunction-associated steatohepatitis (MASH) in participants with MAFLD, regardless of hepatitis B virus infection status. Materials and Methods This prospective cross-sectional multicenter study of consecutive adults with MAFLD who underwent multiparametric US with ATI and 2D SWE, as well as liver biopsy, from September 2020 to June 2022 was conducted in 12 tertiary hospitals in China. Multivariable logistic regression was performed to assess risk factors associated with MASH. Area under the receiver operating characteristic curve (AUC) analysis was used to evaluate diagnostic performance in predicting MASH in training and validation groups (6:4 ratio of participants), and for a post hoc subgroup analysis of hepatitis B virus infection and diabetes. Results A total of 424 participants (median age, 47 years; IQR, 34-59 years; 244 male) were evaluated, including 332 participants (78%) with MASH and 92 (22%) without. Attenuation coefficient (AC) (odds ratio [OR], 3.32 [95% CI: 1.94, 5.71]; P < .001), alanine aminotransferase (ALT) level (OR, 4.42 [95% CI: 1.78, 10.94]; P = .001), and international normalized ratio (INR) (OR, 0.59 [95% CI: 0.37, 0.95]; P = .03) were independently associated with MASH. A combined model (AC, ALT, and INR) had AUCs of 0.85 (95% CI: 0.79, 0.91) and 0.77 (95% CI: 0.69, 0.85) for predicting MASH in the training and validation groups, respectively. AUC values for the subgroups with and without diabetes were 0.83 (95% CI: 0.72, 0.94) and 0.81 (95% CI: 0.75, 0.87) and for the subgroups with and without hepatitis B were 0.82 (95% CI: 0.74, 0.90) and 0.79 (95% CI: 0.71, 0.87), respectively. Conclusion A model combining AC, ALT level, and INR showed good discrimination ability for predicting MASH in participants with MAFLD. Clinical trial registration no. NCT04551716 © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Reuter in this issue.

Objectives:This study aimed to investigate the association between contrast-enhanced ultrasound-detected (CEUS) perfusion patterns of carotid plaque and the occurrence of stroke. Materials and methods:This prospective observational study finally enrolled 256 patients (151 of them having undergone CEUS) from 7851 patients who underwent carotid artery ultrasound from May 2019 to December 2019 with the endpoint being the occurrence of stroke. The risk factors and carotid ultrasound fetures was obtained from those patients. Analyze the relationship between these variables and stoke occurrence. Results:Patients in the recurrent stroke group and those in the no-recurrent stroke group were statistically different in stenosis rate, plaque echo, fibrous cap integrity, calcification of fibrous cap, hypoechoic area within the plaque, and the pattern of neovascularization perfusion from plaque surface to plaque interior (P < 0.05). Upon adjusting for variables, in all subjects, Cox regression analysis showed that symptoms experienced within the past 6 months (RR = 2.486, 95 % CI: 1.282-4.821), moderate-to-severe carotid stenosis (RR = 2.407, 95 % CI: 1.480-4.593), calcification of fibrous cap (RR = 1.599, 95 % CI: 0.727-3.516) and patchy hypoechoic areas within plaque (RR = 2.486, 95 % CI: 1.107-5.578)independently predicted stroke occurrence across all subjects.; Among subjects underwent CEUS, Cox regression analysis demonstrated that moderate-to-severe carotid artery stenosis (RR = 2.105, 95 % CI: 1.425-4.510) and microbubbles entering the interior of the plaque from its surface (RR = 2.323, 95 % CI: 1.175-4. 594) were independent predictors of stroke occurrence. Conclusions:The neovascularization perfusion pattern from the plaque surface to its interior serves as an independent predictor of stroke occurrence, thereby potentially enhancing clinical decision-making.

Carotid artery segmentation and atherosclerosis identification are important in the diagnosis of carotid atherosclerosis. Conventional 2D ultrasound devices may overlook some plaque information of carotid atherosclerosis due to angle issues. This study proposed a multi-task learning method for automatically diagnosing carotid atherosclerosis and simultaneously segmenting carotid arteries to extract the 3D volume of plaque for further analysis. A 3D U-net was firstly employed for carotid segmentation, followed by integrating the segmentation results and original data volume into the classification module composed by another adapted 3D U-net. The accuracy of atherosclerosis identification was 90.0% with a sensitivity of 91.7% and a specificity of 89.3%. Meanwhile, the DSC from our approach were 0.9189 for lumen-intima boundary (LIB) and 0.9444 for vessel-wall-volume (VWV). The result of the proposed algorithm showed the potential of clinical implications for the diagnosis of carotid atherosclerosis using the portable 3D ultrasound imaging technique.

ObjectivesThis study aims to develop 7×7 machine-learning cross-combinatorial methods for selecting and classifying radiomic features used to construct Radiomics Score (RadScore) of predicting the mid-term efficacy and prognosis in high-risk patients with diffuse large B-cell lymphoma (DLBCL).MethodsRetrospectively, we recruited 177 high-risk DLBCL patients from two medical centers between October 2012 and September 2022 and randomly divided them into a training cohort (n=123) and a validation cohort (n=54). We finally extracted 110 radiomic features along with SUVmax, MTV, and TLG from the baseline PET. The 49 features selection-classification pairs were used to obtain the optimal LASSO-LASSO model with 11 key radiomic features for RadScore. Logistic regression was employed to identify independent RadScore, clinical and PET factors. These models were evaluated using receiver operating characteristic (ROC) curves and calibration curves. Decision curve analysis (DCA) was conducted to assess the predictive power of the models. The prognostic power of RadScore was assessed using cox regression (COX) and Kaplan–Meier plots (KM).Results177 patients (mean age, 63 ± 13 years,129 men) were evaluated. Multivariate analyses showed that gender (OR,2.760; 95%CI:1.196,6.368); p=0.017), B symptoms (OR,4.065; 95%CI:1.837,8.955; p=0.001), SUVmax (OR,2.619; 95%CI:1.107,6.194; p=0.028), and RadScore (OR,7.167; 95%CI:2.815,18.248; p<0.001) independently contributed to the risk factors for predicting mid-term outcome. The AUC values of the combined models in the training and validation groups were 0.846 and 0.724 respectively, outperformed the clinical model (0.714;0.556), PET based model (0.664; 0.589), NCCN-IPI model (0.523;0.406) and IPI model (0.510;0.412) in predicting mid-term treatment outcome. DCA showed that the combined model incorporating RadScore, clinical risk factors, and PET metabolic metrics has optimal net clinical benefit. COX indicated that the high RadScore group had worse prognosis and survival in progression-free survival (PFS) (HR, 2.1737,95%CI: 1.2983, 3.6392) and overall survival (OS) (HR,2.1356,95%CI: 1.2561, 3.6309) compared to the low RadScore group. KM survival analysis also showed the same prognosis prediction as Cox results.ConclusionThe combined model incorporating RadScore, sex, B symptoms and SUVmax demonstrates a significant enhancement in predicting medium-term efficacy and prognosis in high-risk DLBCL patients. RadScore using 7×7 machine learning cross-combinatorial methods for selection and classification holds promise as a potential method for evaluating medium-term treatment outcome and prognosis in high-risk DLBCL patients.

Rationale and Objectives: To propose a novel deep learning method incorporating multiple regions based on ultrasound and grayscale ultrasound, evaluate its performance in reducing false positives for Breast Imaging Reporting (BI -RADS) category 4 lesions, and compare its diagnostic performance with that of ultrasound experts. Materials and Methods: This study enrolled 163 breast lesions in 161 women from November 2018 to March 2021 n t-enhan ultrasound and conventional ultrasound were performed before surgery or biopsy. A novel deep learning model incorporating multiple regions based on contrast -enhanced ultrasound and grayscale ultrasound was proposed for minimizing the number of false-positive biopsies. The area Linder the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy were compared between the deep learning model and ultrasound experts. Results: The AUC, sensitivity, specificity, and accuracy of the deep learning model in BI-RADS category 4 lesions were 0.910 91.5%, 90.5%, and 90.8%, respectively, compared with those of ultrasound experts were 0.869, 89.4%, 84.5%, and 85.9%, respectively. Conclusion: The novel deep learning model we proposed had a diagnostic accuracy comparable to that of ultrasound experts, showing the potential to be clinically useful in minimizing the number of false-positive biopsies.

Background The relationship between metabolic dysfunction-associated steatotic liver disease (MASLD) and atherosclerosis has been controversial, which has become a hit of recent research. The study aimed to explore the association between MASLD, cardiovascular and cerebrovascular diseases (CCVD), and the thickness of carotid plaque which was assessed by ultrasound. Methods From September 2018 to June 2019, 3543 patients were enrolled. We asked participants to complete questionnaires to obtain information. All patients underwent liver ultrasound and bilateral carotid ultrasound to obtain carotid intima-media thickness (IMT) and maximum carotid plaque thickness (CPT). Hepatic steatosis was quantified during examination according to Hamaguchi’s ultrasonographic score, from 0 to 6 points. A score < 2 was defined as without fatty liver, and a score ≥ 2 was defined as fatty liver. Information about blood lipids was collected based on the medical records. Results We found common risk factors for CCVD events, MASLD, and atherosclerosis. There was a significant correlation between MASLD and carotid plaque, but not with CPT. No association was found between MASLD and CCVD events. CPT and IMT were thicker in CCVD patients than in non-CCVD patients. No significant difference was found between IMT and CPT in MASLD patients and non-MASLD patients. CCVD was independently and consistently associated with higher IMT, and free fatty acid (FFA). Conclusions According to our results, we recommend carotid ultrasound examination of the patients when FFA is increased, regardless of the presence of risk factors and MASLD. Due to the distribution of CPT of both CCVD and MASLD patients in the CPT 2-4 mm group, contrast-enhanced ultrasound is necessary to assess the vulnerability of the plaque when CPT ≥ 2 mm. Timely treatment of vulnerable plaques may reduce the incidence of future CCVD events.