Simulating federated learning for steatosis detection using ultrasound images.

We aimed to implement four data partitioning strategies evaluated with four federated learning (FL) algorithms and investigate the impact of data distribution on FL model performance in detecting steatosis using B-mode US images. A private dataset (153 patients; 1530 images) and a public dataset (55 patient; 550 images) were included in this retrospective study. The datasets contained patients with metabolic dysfunction-associated fatty liver disease (MAFLD) with biopsy-proven steatosis grades and control individuals without steatosis. We employed four data partitioning strategies to simulate FL scenarios and we assessed four FL algorithms. We investigated the impact of class imbalance and the mismatch between the global and local data distributions on the learning outcome. Classification performance was assessed with area under the receiver operating characteristic curve (AUC) on a separate test set. AUCs were 0.93 (95% CI 0.92, 0.94) for source-based partitioning scenario with FedAvg, 0.90 (95% CI 0.89, 0.91) for a centralized model, and 0.83 (95% CI 0.81, 0.85) for a model trained in a single-center scenario. When data was perfectly balanced on the global level and each site had an identical data distribution, the model yielded an AUC of 0.90 (95% CI 0.88, 0.92). When each site contained data exclusively from one single class, irrespective of the global data distribution, the AUC fell in the range of 0.34-0.70. FL applied to B-mode US images provide performance comparable to a centralized model and higher than single-center scenario. Global data imbalance and local data heterogeneity influenced the learning outcome.

Automated analysis of pectoralis major thickness in pec-fly exercises: evolving from manual measurement to deep learning techniques.

This study addresses a limitation of prior research on pectoralis major (PMaj) thickness changes during the pectoralis fly exercise using a wearable ultrasound imaging setup. Although previous studies used manual measurement and subjective evaluation, it is important to acknowledge the subsequent limitations of automating widespread applications. We then employed a deep learning model for image segmentation and automated measurement to solve the problem and study the additional quantitative supplementary information that could be provided. Our results revealed increased PMaj thickness changes in the coronal plane within the probe detection region when real-time ultrasound imaging (RUSI) visual biofeedback was incorporated, regardless of load intensity (50% or 80% of one-repetition maximum). Additionally, participants showed uniform thickness changes in the PMaj in response to enhanced RUSI biofeedback. Notably, the differences in PMaj thickness changes between load intensities were reduced by RUSI biofeedback, suggesting altered muscle activation strategies. We identified the optimal measurement location for the maximal PMaj thickness close to the rib end and emphasized the lightweight applicability of our model for fitness training and muscle assessment. Further studies can refine load intensities, investigate diverse parameters, and employ different network models to enhance accuracy. This study contributes to our understanding of the effects of muscle physiology and exercise training.

Sonographic methods to predict type 2 diabetes patients with sarcopenia: B mode ultrasound and shear wave elastography.

BACKGROUND: Type 2 diabetes accelerates the loss of muscle mass and strength. Sarcopenia is also one of the chronic complications of diabetes. OBJECTIVE: To investigate the clinical value of B mode ultrasound (BMUS) and shear wave elastography (SWE) for predicting type 2 diabetic sarcopenia. METHODS: We recorded Skeletal Muscle Mass Index (ASMI), grip strength, muscle thickness (MT), pinna angle (PA), fascicle length (FL), and the difference of Young's modulus in the relaxed states and tense states (ΔSWE). The correlations between clinical indicators and ultrasound characteristics were compared. A diagnostic model of sarcopenia was developed to assess the independent correlates and evaluate the diagnostic efficacy of sarcopenia. RESULTS: ASMI was significantly and positively correlated with MT and ΔSWE (r = 0.826, 0.765, P < 0.01), and grip strength was significantly and positively correlated with MT and ΔSWE (r = 0.797, 0.818, P < 0.01). MT was the most significant predictor of sarcopenia (OR = 4.576, P < 0.001), and the cut-off value of MT was 11.4 mm (AUC: 0.952). CONCLUSION: BMUS and SWE can quantitatively assess muscle mass and strength, and are effective methods to predict the occurrence of sarcopenia in elderly patients with type 2 diabetes.

B-mode ultrasound to elastography synthesis using multiscale learning.

Elastography is a promising diagnostic tool that measures the hardness of tissues, and it has been used in clinics for detecting lesion progress, such as benign and malignant tumors. However, due to the high cost of examination and limited availability of elastic ultrasound devices, elastography is not widely used in primary medical facilities in rural areas. To address this issue, a deep learning approach called the multiscale elastic image synthesis network (MEIS-Net) was proposed, which utilized the multiscale learning to synthesize elastic images from ultrasound data instead of traditional ultrasound elastography in virtue of elastic deformation. The method integrates multi-scale features of the prostate in an innovative way and enhances the elastic synthesis effect through a fusion module. The module obtains B-mode ultrasound and elastography feature maps, which are used to generate local and global elastic ultrasound images through their correspondence. Finally, the two-channel images are synthesized into output elastic images. To evaluate the approach, quantitative assessments and diagnostic tests were conducted, comparing the results of MEIS-Net with several deep learning-based methods. The experiments showed that MEIS-Net was effective in synthesizing elastic images from B-mode ultrasound data acquired from two different devices, with a structural similarity index of 0.74 ± 0.04. This outperformed other methods such as Pix2Pix (0.69 ± 0.09), CycleGAN (0.11 ± 0.27), and StarGANv2 (0.02 ± 0.01). Furthermore, the diagnostic tests demonstrated that the classification performance of the synthetic elastic image was comparable to that of real elastic images, with only a 3 % decrease in the area under the curve (AUC), indicating the clinical effectiveness of the proposed method.

Ex-vivo validation of spatial gain sonography for the quantification of echo intensity in fascicle-aligned ultrasound images in ten anatomical muscles in Bos taurus.

This study aimed to validate the concept of spatial gain sonography for quantifying texture-related echo intensity in B-mode ultrasound of skeletal muscle. Fifty-one bovine muscles were scanned postmortem using B-mode ultrasonography at varying fascicle probe angles (FPA). The relationship between mean gray values (MGV) and FPA was fitted with a sinusoidal and a linear function, the slope of which was defined as tilt echo gain (TEG). Macroscopic muscle cross sections were optically analyzed for intramuscular connective tissue (IMCT) content which was plotted against MGV at 0° FPA (MGV_00). MGV peaked at FPA 0°. Sine fits were superior to linear fits (adjusted r2-values 0.647 vs. 0.613), especially for larger FPAs. In mixed models, the pennation angle was related to TEG (P < 0.001) and MGV_00 (P = 0.035). Age was relevant for MGV_00 (P < 0.001), but not TEG (P > 0.10). The correlation between the IMCT percentage and MGV_00 was significant but weak (P = 0.026; adjusted r2 = 0.103). The relationship between fascicle probe angle and echo intensity in B-mode ultrasound can be modeled more accurately with a sinusoidal but more practically for clinical use with a linear fit. The peak mean gray value MGV_00 can be used to compare echo intensity across muscles without the bias of pennation angle.

Hepatic steatosis using ultrasound-derived fat fraction: First technical and clinical evaluation.

OBJECTIVES: To explore the technical and clinical evaluation of ultrasound-derived fat fraction (UDFF) measurement in adult patients in whom fatty liver was suspected. MATERIALS AND METHODS: In this prospective study, 41 participants were initially enrolled in our hospital between October 2022 and December 2022 and received UDFF assessment using Siemens ACUSON Sequoia system equipped with DAX transducer. UDFF measurement was performed three times to obtain UDFF values from each imaging location (V hepatic segment and VIII hepatic segment) per participant, and the depth (skin-to-capsule distance) was automatically measured. The echogenicity of liver tissue in B mode ultrasound (BMUS) was compared to the normal kidney tissue, and fatty liver was graded as mild (Grade 1), moderate (Grade 2), and severe (Grade 3). The median of the acquired overall median UDFF values was used for statistical analysis. All ultrasound examinations were performed by one of two radiologists (with 20 and 10 years of liver ultrasound imaging experience). RESULTS: Finally, UDFF measurement was successfully performed on 38 participants to obtain valid values, including 21 men with a median age of 40.0 years (interquartile range [IQR]: 23.0 - 58.5) and 17 women with a median age of 60.0 years (IQR: 29.5 - 67.0). Fatty liver was diagnosed by BMUS features in 47.4% (18/38) participants. Among all participants, the median UDFF value was 7.0% (IQR: 4.0 - 15.6). A significant difference in UDFF values was found between participants with fatty liver and without fatty liver (U = 7.0, P < 0.001), and UDFF values elevated as the grade of the fatty liver increased (P < 0.001). The median UDFF values from the three UDFF measurements obtained during each ultrasound examination showed excellent agreement (ICC = 0.882 [95% confidence interval: 0.833 - 0.919]). The Spearman correlation of UDFF values in different depths was moderate, with a rs value of 0.546 (P < 0.001). No significant differences in UDFF values were found between V hepatic segment and VIII hepatic segment (U = 684.5, P = 0.697). CONCLUSIONS: UDFF provides a novel non-invasive imaging tool for hepatic steatosis assessment with excellent feasibility.

Role of Shear Wave Elastography in the Diagnosis of Peyronie Disease.

OBJECTIVES: The present study aims to explore the role of shear wave elastography (SWE) in the diagnosis of Peyronie disease (PD). METHODS: A total of 59 PD patients and 59 age-matched healthy adult men were included in this study. The B-mode ultrasound (US) and SWE were performed for all subjects, and the Young modulus (YM) values of the corresponding regions of the penis in the PD and control groups were recorded and compared. RESULTS: The mean age of the included PD patients and age-matched controls was 53.81 years (SD 9.52, range 32-73). On B-mode US evaluation, 41 (69.5%) of 59 included PD patients were found to have penile plaques, and the remaining 18 (30.5%) patients had no evidence of penile plaque. After evaluation using SWE, the YM values in the penile plaque region of these 41 patients with penile dysplasia were found to be significantly higher (60.29 kPa ± 19.95) than those outside the plaque (in the same patient) (21.05 kPa ± 4.58) and in the same penile region of the control group (20.59 kPa ± 4.65) (P < .001). In the remaining 18 PD patients, the results showed that the YM value of the abnormal penile region in the PD patients (56.67 kPa ± 13.52) was significantly higher than the YM value outside the abnormal penile region in the same patients (22.79 kPa ± 4.31) and in the same penile region in the control group (19.87 kPa ± 3.48) (P < .001; P < .001). CONCLUSIONS: In conclusion, this study showed that SWE as a non-invasive technique is useful in identifying and differentiating penile plaques in PD patients and is a simple, rapid and complementary method to B-mode US.

Intra- and Interobserver Reliability of Shear Wave Elastography in Breast Cancer Diagnosis.

OBJECTIVES: Shear wave elastography (SWE) is increasingly used in breast cancer diagnostics. However, large, prospective, multicenter data evaluating the reliability of SWE is missing. We evaluated the intra- and interobserver reliability of SWE in patients with breast lesions categorized as BIRADS 3 or 4. METHODS: We used data of 1288 women at 12 institutions in 7 countries with breast lesions categorized as BIRADS 3 to 4 who underwent conventional B-mode ultrasound and SWE. 1243 (96.5%) women had three repetitive conventional B-mode ultrasounds as well as SWE measurements performed by a board-certified senior physician. 375 of 1288 (29.1%) women received an additional ultrasound examination with B-mode and SWE by a second physician. Intraclass correlation coefficients (ICC) were calculated to examine intra- and interobserver reliability. RESULTS: ICC for intraobserver reliability showed an excellent correlation with ICC >0.9, while interobserver reliability was moderate with ICC of 0.7. There were no clinically significant differences in intraobserver reliability when SWE was performed in lesions categorized as BI-RADS 3 or 4 as well as in histopathologically benign or malignant lesions. CONCLUSION: Reliability of additional SWE was evaluated on a study cohort consisting of 1288 breast lesions categorized as BI-RADS 3 and 4. SWE shows an excellent intraobserver reliability and a moderate interobserver reliability in the evaluation of solid breast masses.

The relationship between biological parameters and facial soft tissue thickness measured by ultrasound and its forensic implications.

Facial soft tissue thickness (FSTT) data are currently widely used in forensic and medical science. In the forensic sciences, they form the basis for craniofacial reconstruction and identification methods. Since there are few FSTT data in the Slovak population, this study aims to enrich the data in well-defined age categories, taking into account differences between sexes and body mass index (BMI). The sample consisted of 127 participants aged 17 to 86 years from Slovakia. In addition to biological sex and age information, stature and body weight were recorded to calculate BMI. Subsequently, 17 facial anthropometric landmarks were used to measure FSTT using a noninvasive General Electric LOGIQe R7 ultrasound device. The mean values of FSTT were greater in the mouth region in males and in the zygomatic and eye regions in females. Differences between males and females, regardless of sex and BMI, were significant only at two landmarks. When BMI and age were taken into account, there were differences in 12 of 17 landmarks. Linear regression results showed the strongest correlation of most landmarks with BMI, followed by age and sex. When the FSTT was estimated in association with sex/age/BMI, landmarks in the zygomatic, mandibular, and frontal regions were the best regressors. The results of the present study demonstrate that B-mode ultrasound measurements of FSTT can be used in facial reconstruction as a function of BMI, age, and sex of the subject. Furthermore, the present regression equations can help practitioners in the forensic/medical field to calculate individual tissue thickness.

A model for predicting lymph node metastasis of thyroid carcinoma: a multimodality convolutional neural network study.

Background: Early preoperative evaluation of cervical lymph node metastasis (LNM) in papillary thyroid carcinoma (PTC) is critical for further surgical treatment. However, insufficient accuracy in predicting LNM status for PTC based on ultrasound images is a problem that needs to be urgently resolved. This study aimed to clarify the role of convolutional neural networks (CNNs) in predicting LNM for PTC based on multimodality ultrasound. Methods: In this study, the data of 308 patients who were clinically diagnosed with PTC and had confirmed LNM status via postoperative pathology at Beijing Tiantan Hospital, Capital Medical University, from August 2018 to April 2022 were incorporated into CNN algorithm development and evaluation. Of these patients, 80% were randomly included into the training set and 20% into the test set. The ultrasound examination of cervical LNM was performed to assess possible metastasis. Residual network 50 (Resnet50) was employed for feature extraction from the B-mode and contrast-enhanced ultrasound (CEUS) images. For each case, all of features were extracted from B-mode ultrasound images and CEUS images separately, and the ultrasound examination data of cervical LNM information were concatenated together to produce a final multimodality LNM prediction. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) were used to evaluate the performance of the predictive model. Heatmaps were further developed for visualizing the attention region of the images of the best-working model. Results: Of the 308 patients with PTC included in the analysis, 158 (51.3%) were diagnosed as LNM and 150 (48.7%) as non-LNM. In the test set, when a triple-modality method (i.e., B-mode image, CEUS image, and ultrasound examination of cervical LNM) was used, accuracy was maximized at 80.65% (AUC =0.831; sensitivity =80.65%; specificity =82.26%), which showed an expected increased performance over B-mode alone (accuracy =69.00%; AUC =0.720; sensitivity =70.00%; specificity =73.00%) and a dual-modality method (B-mode image plus CEUS image: accuracy =75.81%; AUC =0.742; sensitivity =74.19%; specificity =77.42%). The heatmaps of our triple-modality model demonstrated a possible focus area and revealed the model's flaws. Conclusions: The PTC lymph node prediction model based on the triple-modality features significantly outperformed all the other feature configurations. This deep learning model mimics the workflow of a human expert and leverages multimodal data from patients with PTC, thus further supporting clinical decision-making.

Application and prospects of AI-based radiomics in ultrasound diagnosis.

Artificial intelligence (AI)-based radiomics has attracted considerable research attention in the field of medical imaging, including ultrasound diagnosis. Ultrasound imaging has unique advantages such as high temporal resolution, low cost, and no radiation exposure. This renders it a preferred imaging modality for several clinical scenarios. This review includes a detailed introduction to imaging modalities, including Brightness-mode ultrasound, color Doppler flow imaging, ultrasound elastography, contrast-enhanced ultrasound, and multi-modal fusion analysis. It provides an overview of the current status and prospects of AI-based radiomics in ultrasound diagnosis, highlighting the application of AI-based radiomics to static ultrasound images, dynamic ultrasound videos, and multi-modal ultrasound fusion analysis.

Development of a Clinical-Radiomics Nomogram That Used Contrast-Enhanced Ultrasound Images to Anticipate the Occurrence of Preoperative Cervical Lymph Node Metastasis in Papillary Thyroid Carcinoma Patients.

Background and Objectives: Papillary thyroid carcinoma (PTC) is a prevalent histological type of thyroid cancer; however, noninvasive assessment of cervical lymph node metastasis (LNM) poses a challenge. This study aims to develop a novel clinical-radiomics nomogram that utilizes ultrasound (US) images to predict the presence of cervical LNM metastasis in patients with PTC. Methods: A total of 423 patients with PTC were recruited to participate in this study between January 2020 and December 2022, of which 282 were classified into the training group and 141 patients were classified into the validation set. Contrast-enhanced ultrasound (CEUS) and B-mode ultrasound (BMUS) images were subjected to radiomic analysis, leading to the extraction of 912 radiomic features. Thereafter, a radiomics score (Radscore) was developed to effectively integrate the information derived from BMUS and CEUS modalities. Univariate and multivariate backward stepwise logistic regression analysis techniques were used to construct the clinical and clinical-radiomics models, respectively. Results: The findings revealed that the clinical-radiomics nomogram incorporated age, sex, CEUS Radscore, and US-reported LNM as risk factors. The nomogram demonstrated good performance using data from the training (AUC = 0.891) and validation (AUC = 0.870) sets. The decision curve analysis implied that this nomogram exhibited good clinical utility, which was further supported by the results of the calibration curves and Hosmer-Lemeshow test. Conclusion: The CEUS Radscore-based clinical radiomics nomogram could serve as a valuable tool for predicting cervical LNM metastasis in patients with PTC, thereby tailoring individualized treatment strategies for them.

Domain knowledge-guided adversarial adaptive fusion of hybrid breast ultrasound data.

Contrast-enhanced ultrasound (CEUS), which provides more detailed microvascular information about the tumor, is always taken by radiologists in clinic diagnosis along with B-mode ultrasound (B-mode US). However, automatically analyzing breast CEUS is challenging due to the difference between the CEUS video and the natural video, e.g., sports or action videos, where the CEUS video has no positional displacements. Additionally, most existing methods rarely use the Time Intensity Curve (TIC) information of CEUS and non-imaging clinical (NIC) data. To address these issues, we propose a novel breast cancer diagnosis framework that learns the complementarity and correlation across hybrid modal data, including CEUS, B-mode US, and NIC data, by an adversarial adaptive fusion method. Furthermore, to fully exploit the CEUS information, the proposed method, inspired by the clinical processing of radiologists, first extracts the TIC parameters of CEUS. Then, we select a clip from CEUS using a frame screening strategy and finally get spatio-temporal features from these clips through a critical frame attention network. To our knowledge, this is the first AI system to use TIC parameters, NIC data, and ultrasound imaging in diagnoses. We have validated our method on a dataset collected from 554 patients. The experimental results demonstrate the excellent performance of the proposed method. The result shows that our method can achieve an accuracy of 87.73%, which is higher than that of uni-modal approaches by nearly 5%.

Contrast-enhanced ultrasound-guided biopsy of suspicious breast lesions on contrast-enhanced mammography and contrast-enhanced MRI: a case series.

PURPOSE: To assess the effectiveness of contrast-enhanced ultrasound (CEUS) in guiding biopsies of breast lesions that were detected on contrast-enhanced mammography (CEM) or contrast-enhanced breast MRI (CE-MRI) but were not clearly visible on B-mode ultrasound (B-US). METHODS: In this study, 23 lesions in 16 patients were selected for CEUS-guided biopsy due to poor visualization on B-US despite being detected on CEM (n = 20) or CE-MRI (n = 3). B-US, color Doppler ultrasound (CDUS), and CEUS were used to visualize the suspicious lesions, followed by a CEUS-guided core needle biopsy using Sonazoid as the contrast agent. The accuracy of the biopsy was assessed based on pathology-radiology concordance and 12-month imaging follow-up. The conspicuity scores for lesion visualization were evaluated using a 5-point conspicuity scale agreed upon by two breast radiologists. RESULTS: The enhancing lesions detected on CEM/CE-MRI had an average size of 1.6 ± 1.3 cm and appeared as mass-enhancing (61%) or non-mass-enhancing (39%). The lesions had mean conspicuity scores of 2.30 on B-US, 2.78 on CDUS, and 4.61 on CEUS, with 96% of the lesions showing contrast enhancement on CEUS. CEUS-guided biopsy showed increased visibility in 96% and 91% of the lesions compared to B-US and CDUS, respectively. The overall accuracy of CEUS-guided biopsy was 100% based on concordance with histology and 12-month follow-up. CONCLUSIONS: CEUS enhances the visibility of suspicious CEM/CE-MRI lesions that are poorly visible on B-US during biopsy procedures.

The diagnostic value of Doppler Resistive Index in the differential diagnosis of focal liver lesions.

Aim: To investigate the diagnostic value of resistance index (RI) in differentiating focal liver lesions. Patients and methods: In this retrospective study, a total of 576 patients with histologically confirmed focal liver lesions were included. Each patient underwent B-mode ultrasound examination and color Doppler ultrasound examination. The RI values of different focal liver lesions were recorded and compared. Results: The mean RI value of benign lesions was significantly lower than that of malignant lesions (0.54 ± 0.10 vs. 0.71 ± 0.12) (p <0.05). In malignant lesions, the RI value of intrahepatic cholangiocarcinoma was significantly lower than that of hepatocellular carcinoma lesions. Furthermore, in hepatocellular carcinoma lesions, the RI of large lesions (group 4: >10 cm) was significantly lower than that of small lesions (group 1: ≤2 cm, group 2: 2-5 cm) (p <0.05). Taken RI of 0.615 as a cutoff value to differentiate malignant and benign lesions, the sensitivity, specificity, positive predictive value and negative predictive value were 82.80%, 81.00%, 81.34% and 82.48%, respectively. Conclusion: Color Doppler ultrasound examination is a valuable imaging method in detecting blood flow signal within liver lesions. The RI parameter should be helpful in differentiating malignant and benign liver tumors.

Diagnostic Performance of Greyscale Ultrasound in Detecting Fatty Liver Disease in a Type 2 Diabetes Population Using FibroScan as the Reference Standard.

Introduction Brightness mode ultrasound (B-mode US) and FibroScan (Echosens, Paris, France) are the two ultrasound methods often recommended for screening non-alcoholic fatty liver disease (NAFLD) in persons with type 2 diabetes mellitus (T2DM). This study assessed the diagnostic performance of B-mode US using FibroScan as the reference standard. Methods Persons with a known history of T2DM were invited to screen for NAFLD using B-mode US and FibroScan on separate days within a one-month period. Assessors of B-mode US and FibroScan were blinded to each other's findings. Both B-mode US and FibroScan independently assessed and graded each participant for the presence of NAFLD. Using the diagnostic test findings of FibroScan as a reference standard, the sensitivity and specificity of B-mode US were analyzed. The area under the receiver operating characteristic curve (AUROC) was analyzed using Jamovi (version 2.3.21). A multinomial logistic regression of the B-mode US and FibroScan in predicting NAFLD grade was also analyzed. Results A total of 171 participants were assessed. B-mode US detected NAFLD in T2DM patients with 63.6% sensitivity, 65.6% specificity, and 0.646 AUROC. Sensitivity and specificity in overweight and obese participants were 36-43% and 76-85%, respectively. Multinomial logistic regression demonstrated an insignificant statistical relationship between FibroScan and B-mode US in predicting grade 1 steatosis (p-value = 0.397), which was significantly affected by a higher BMI (p-value = 0.034) rather than a higher liver fibrosis level (p-value = 0.941). The logistic regression further showed a significant relationship between B-mode US and FibroScan in predicting steatosis grade 2 (p-value = 0.045) and grade 3 (p-value < 0.001), which was not significantly affected by BMI (p-value = 0.091). Conclusion B-mode US can replace FibroScan for severe steatosis; however, it cannot be used to screen for NAFLD in T2DM patients due to lower sensitivity for early detection in the overweight.

Prediction of clinical response to neoadjuvant therapy in advanced breast cancer by baseline B-mode ultrasound, shear-wave elastography, and pathological information.

Background: Neoadjuvant therapy (NAT) is the preferred treatment for advanced breast cancer nowadays. The early prediction of its responses is important for personalized treatment. This study aimed at using baseline shear wave elastography (SWE) ultrasound combined with clinical and pathological information to predict the clinical response to therapy in advanced breast cancer. Methods: This retrospective study included 217 patients with advanced breast cancer who were treated in West China Hospital of Sichuan University from April 2020 to June 2022. The features of ultrasonic images were collected according to the Breast imaging reporting and data system (BI-RADS), and the stiffness value was measured at the same time. The changes were measured according to the Response evaluation criteria in solid tumors (RECIST1.1) by MRI and clinical situation. The relevant indicators of clinical response were obtained through univariate analysis and incorporated into a logistic regression analysis to establish the prediction model. The receiver operating characteristic (ROC) curve was used to evaluate the performance of the prediction models. Results: All patients were divided into a test set and a validation set in a 7:3 ratio. A total of 152 patients in the test set, with 41 patients (27.00%) in the non-responders group and 111 patients (73.00%) in the responders group, were finally included in this study. Among all unitary and combined mode models, the Pathology + B-mode + SWE model performed best, with the highest AUC of 0.808 (accuracy 72.37%, sensitivity 68.47%, specificity 82.93%, P<0.001). HER2+, Skin invasion, Post mammary space invasion, Myometrial invasion and Emax were the factors with a significant predictive value (P<0.05). 65 patients were used as an external validation set. There was no statistical difference in ROC between the test set and the validation set (P>0.05). Conclusion: As the non-invasive imaging biomarkers, baseline SWE ultrasound combined with clinical and pathological information can be used to predict the clinical response to therapy in advanced breast cancer.

B-mode ultrasound based CAD for liver cancers via multi-view privileged information learning.

B-mode ultrasound-based computer-aided diagnosis model can help sonologists improve the diagnostic performance for liver cancers, but it generally suffers from the bottleneck due to the limited structure and internal echogenicity information in B-mode ultrasound images. Contrast-enhanced ultrasound images provide additional diagnostic information on dynamic blood perfusion of liver lesions for B-mode ultrasound images with improved diagnostic accuracy. Since transfer learning has indicated its effectiveness in promoting the performance of target computer-aided diagnosis model by transferring knowledge from related imaging modalities, a multi-view privileged information learning framework is proposed to improve the diagnostic accuracy of the single-modal B-mode ultrasound-based diagnosis for liver cancers. This framework can make full use of the shared label information between the paired B-mode ultrasound images and contrast-enhanced ultrasound images to guide knowledge transfer It consists of a novel supervised dual-view deep Boltzmann machine and a new deep multi-view SVM algorithm. The former is developed to implement knowledge transfer from the multi-phase contrast-enhanced ultrasound images to the B-mode ultrasound-based diagnosis model via a feature-level learning using privileged information paradigm, which is totally different from the existing learning using privileged information paradigm that performs knowledge transfer in the classifier. The latter further fuses and enhances feature representation learned from three pre-trained supervised dual-view deep Boltzmann machine networks for the classification task. An experiment is conducted on a bimodal ultrasound liver cancer dataset. The experimental results show that the proposed framework outperforms all the compared algorithms with the best classification accuracy of 88.91 ± 1.52%, sensitivity of 88.31 ± 2.02%, and specificity of 89.50 ± 3.12%. It suggests the effectiveness of our proposed MPIL framework for the BUS-based CAD of liver cancers.

Predicting Extrathyroidal Extension in Papillary Thyroid Carcinoma Using a Clinical-Radiomics Nomogram Based on B-Mode and Contrast-Enhanced Ultrasound.

Papillary thyroid carcinoma (PTC) is the most common pathological type of thyroid cancer. PTC patients with extrathyroidal extension (ETE) are associated with poor prognoses. The preoperative accurate prediction of ETE is crucial for helping the surgeon decide on the surgical plan. This study aimed to establish a novel clinical-radiomics nomogram based on B-mode ultrasound (BMUS) and contrast-enhanced ultrasound (CEUS) for the prediction of ETE in PTC. A total of 216 patients with PTC between January 2018 and June 2020 were collected and divided into the training set (n = 152) and the validation set (n = 64). The least absolute shrinkage and selection operator (LASSO) algorithm was applied for radiomics feature selection. Univariate analysis was performed to find clinical risk factors for predicting ETE. The BMUS Radscore, CEUS Radscore, clinical model, and clinical-radiomics model were established using multivariate backward stepwise logistic regression (LR) based on BMUS radiomics features, CEUS radiomics features, clinical risk factors, and the combination of those features, respectively. The diagnostic efficacy of the models was assessed using receiver operating characteristic (ROC) curves and the DeLong test. The model with the best performance was then selected to develop a nomogram. The results show that the clinical-radiomics model, which is constructed by age, CEUS-reported ETE, BMUS Radscore, and CEUS Radscore, showed the best diagnostic efficiency in both the training set (AUC = 0.843) and validation set (AUC = 0.792). Moreover, a clinical-radiomics nomogram was established for easier clinical practices. The Hosmer-Lemeshow test and the calibration curves demonstrated satisfactory calibration. The decision curve analysis (DCA) showed that the clinical-radiomics nomogram had substantial clinical benefits. The clinical-radiomics nomogram constructed from the dual-modal ultrasound can be exploited as a promising tool for the pre-operative prediction of ETE in PTC.

USC-ENet: a high-efficiency model for the diagnosis of liver tumors combining B-mode ultrasound and clinical data.

Purpose: Ultrasound image acquisition has the advantages of being low cost, rapid, and non-invasive, and it does not produce radiation. Currently, ultrasound is widely used in the diagnosis of liver tumors. However, owing to the complex presentation and diverse features of benign and malignant liver tumors, accurate diagnosis of liver tumors using ultrasound is difficult even for experienced radiologists. In recent years, artificial intelligence-assisted diagnosis has proven to provide effective support to radiologists. However, there is room for further improvement in the existing ultrasound artificial intelligence diagnostic model of liver tumor. First, the image diagnostic model may not fully consider relevant clinical data in the decision-making process. Second, owing to the difficulty in collecting biopsy pathology and physician-labeled ultrasound data of liver tumors, training datasets are usually small, and commonly used large neural networks tend to overfit on small datasets, which seriously affects the generalization of the model. Methods: In this study, we propose a deep learning-assisted diagnosis model called USC-ENet, which integrates B-mode ultrasound features of liver tumors and clinical data of patients, and we design a small neural network specifically for small-scale medical images combined with an attention mechanism. Results and conclusion: Real data from 542 patients with liver tumors (N = 2168 images) are used during model training and validation. Experiments show that USC-ENet can achieve a good classification effect (area under the curve = 0.956, sensitivity = 0.915, and specificity = 0.880) after small-scale data training, and it has certain interpretability, showing good potential for clinical adoption. In conclusion, our model provides not only a reliable second opinion for radiologists but also a reference for junior radiologists who lack clinical experience.