Ultrasound in Medicine and Biology最新文献

Objective: This study aimed to develop a machine learning model based on ultrasonography (US) and clinicopathological features to predict pathological complete response (pCR) following neoadjuvant chemotherapy (NAC) in patients with breast cancer. The goal was to establish a non-invasive prediction tool to facilitate individualized treatment planning.

Methods: A retrospective analysis was conducted on data from 463 patients with breast cancer who underwent NAC at Shanxi Bethune Hospital between January 2018 and December 2024. Patients were randomly allocated into a training set (n = 277) and a test set (n = 118). To address class imbalance, the Synthetic Minority Over-sampling Technique algorithm was applied. Ten key features, including tumor short diameter, maximum elasticity, and age group, were selected through Least Absolute Shrinkage and Selection Operator regression. Seven machine learning models were constructed, including Random Forest, Logistic Regression, and Extreme Gradient Boosting (XGBoost). Model parameters were optimized through ten-fold cross-validation. Performance evaluation involved receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and calibration curves.

Results: Among the developed models, XGBoost demonstrated superior performance, achieving an area under the ROC curve of 0.8955 (95% confidence interval: 0.8409-0.9601), sensitivity of 0.8095, and specificity of 0.8026 in the test set. Shapley Additive Explanations analysis identified ER-negative, PR-negative, tumor short diameter, and HER2-positive as significant predictors of pCR (contribution > 15%). DCA indicated that XGBoost provided the highest net benefit within clinical decision thresholds (10%-90%), and the calibration curve demonstrated good consistency between predicted and observed outcomes, with a slope approaching 1 (Brier score = 0.11).

Conclusion: The XGBoost model, incorporating US imaging and clinicopathological features, demonstrated high accuracy in predicting pCR following NAC in patients with breast cancer. These findings indicate that the model may serve as a valuable tool for efficacy evaluation. Further validation with multi-center data is necessary to confirm generalizability and support clinical application.

Background: It has been suggested that ultrasound (US) imaging can be used to assess cortical bone health, which is of particular interest owing to its major role in bone mechanical stability. Intra-cortical US imaging extends B-mode imaging into bone using a dedicated image reconstruction algorithm that corrects for refraction at the bone-soft tissue interfaces. It has shown promising results in a few healthy, predominantly young adults, providing anatomical images of the cortex (periosteal and endosteal surfaces) along with estimations of US wave speed. However, its reliability in older or osteoporotic bones remains uncertain.

Objective: In this study, we critically assessed the performance of intra-cortical US imaging ex vivo in bones with various microstructural patterns, including bones exhibiting signs of unbalanced intra-cortical remodeling.

Methods: We analyzed factors influencing US image quality, particularly endosteal surface reconstruction, as well as the accuracy of wave speed estimation and its relationship with porosity. We imaged 20 regions of interest from the femoral diaphysis of 5 elderly donors using a 2.5 MHz US transducer. The reconstructed US images were compared to site-matched high-resolution micro-computed tomography images.

Results: In samples with moderate porosity, the endosteal surface was accurately identified, and thickness estimates from US and high-resolution micro-computed tomography differed by less than 10%. In highly remodeled bones with increased porosity, pore size and an heterogeneous distribution of pores, the reconstructed endosteal surface appeared less bright and was located above the trabecularized cortex region. We observed a decrease in US wave speed with increasing cortical porosity, aligning well with literature data, suggesting that, based on wave speed value the method could discriminate between bones with low porosity (<5%) and those with moderate to high porosity (>10%).

Conclusion: This study paves the way for the application of US imaging in diagnosing cortical bone health, particularly for detecting increased cortical porosity and reduced cortical thickness.

Objective: It is essential to determine the location and evaluate the status of the sentinel lymph node (SLN) in breast cancer patients before surgery. However, the diagnostic performance of the conventional ultrasound (US) examination and lymphatic contrast-enhanced US (LCEUS) was not satisfied. This study aimed to explore the US diagnostic value of the relative p value method for SLN in breast cancer, to provide reference for precise clinical diagnosis.

Methods: A retrospective collection of 157 breast cancer patients who were admitted to the Affiliated Hospital of Nanjing University of Chinese Medicine from July 2018 to December 2022 was conducted. All patients underwent US and LCEUS examinations before surgery and confirmed by pathology. Because pathology was considered the gold standard, the diagnostic efficacies of US, LCEUS, US + LCEUS, the p value method and the p value method + LCEUS were analyzed.

Results: Using pathology as the gold standard, there were 68 cases (43.31%) in the SLN metastasis-positive group including macro-metastasis and micro-metastasis, whereas the SLN metastasis-negative group included 89 cases (56.69%), including no metastasis and isolated tumor cell clusters. Among 157 SLNs of breast cancer patients, five methods (ie, US, LCEUS, US + LCEUS, p value and p value + LCEUS) in diagnosing SLNs achieved, respectively, an accuracy of 80.25%, 81.53%, 79.62%, 84.08% and 83.44%; sensitivity of 76.47%, 77.94%, 88.24%, 85.39% and 92.65%; specificity of 83.14%, 86.52%, 73.03%, 82.35% and 76.40%; positive predictive value (of 77.61%, 81.54%, 71.43%, 86.36% and 75.00%; and negative predictive value of 82.22%, 83.70%, 89.04%, 81.16% and 93.15%. Additionally, the sensitivity of the p value + LCEUS method showed statistically significant differences when compared with that of the US and the LCEUS method, respectively (p < 0.05). A statistically significant difference in negative predictive value was observed between the US and the p value + LCEUS method (p < 0.05). Receiver operating characteristic curves were plotted for the diagnostic sensitivity of five groups (i.e., US, LCEUS, US + LCEUS p value and p value + LCEUS) in diagnosing SLNs of the breast. The areas under the curve (AUC) were 0.798, 0.822, 0.806, 0.839 and 0.845, respectively. No statistically significant differences were found in the pairwise comparisons.

Conclusion: The method of predicting the metastasis status of SLN based on pre-operative LCEUS and the p value method can assist clinicians in assessing the risk of SLN metastasis before surgery. It is possible to decrease unnecessary SLN biopsies in low-risk patients and lower the incidence of complications.

Objective: To develop a radiomics model that uses grayscale ultrasound (GSUS) and contrast-enhanced ultrasound (CEUS) images, integrated with clinical and radiological characteristics, to forecast ablation zone absorption after microwave ablation in patients with T1aN0M0 papillary thyroid carcinoma (PTC).

Methods: This research involved a total of 163 patients with 203 T1aN0M0 PTCs. The cohort was randomly divided into training (n = 142) and validation (n = 61) cohorts with a distribution ratio of 7:3. Radiomics features were obtained from GSUS and CEUS images 1 month after ablation. Feature selection used Pearson and Spearman correlation analyses, as well as the least absolute shrinkage and selection operator regression model. Radiomics models based on GSUS, CEUS and their combination were developed using Cox regression. A nomogram was created to forecast the absorption of the ablation zone, incorporating radiomics scores with clinical and radiological factors. Model performance was assessed and compared using concordance index (C-index) values, net reclassification improvement and integrated discrimination improvement metrics.

Results: The radiomics score from GSUS and CEUS emerged as an independent predictor of ablation zone absorption. The combined model, integrating multimodal radiomics scores with clinical-radiological data, demonstrated optimal performance in the training cohort (C-index = 0.891) and the validation cohort (C-index = 0.906), with significant clinical benefits demonstrated via calibration and decision curves.

Conclusion: The combined model integrating radiomic features with clinical and radiological characteristics can effectively predict the absorption of the ablation zone after microwave ablation in patients with T1aN0M0 PTC.

Objective: The safety of active surveillance (AS) in highly suspicious thyroid nodules classified as ACR TI-RADS 5, particularly those located close to the thyroid capsule, has not been fully established. This study aims to assess the feasibility and safety of AS in patients with subcapsular ACR TI-RADS 5 nodules compared with non-subcapsular nodules.

Methods: This retrospective study included 675 patients with 763 ACR TI-RADS 5 nodules who underwent AS at Ruijin Hospital between 2015 and 2024. Nodules were categorized according to the distance from the thyroid capsule as subcapsular (≤2 mm) or non-subcapsular (>2 mm). Disease progression was defined as an increase in maximum diameter of at least 3 mm, a volume increase of at least 50%, the appearance of new suspicious thyroid lesions, or lymph node metastasis. Clinical and ultrasound characteristics were compared between groups, and progression-free survival was analyzed using the Kaplan-Meier method.

Results: Among the 763 nodules, 175 (22.9%) were subcapsular. The median follow-up duration was 55 months. Patients with subcapsular nodules were younger (33 ± 11 years vs 43 ± 12 years, p < 0.001) and more frequently had multifocal disease (21.6% vs 12.5%, p = 0.004). No significant differences were observed between the subcapsular and non-subcapsular groups in nodule enlargement (2.9% vs 2.7%), volume increase (25.7% vs 24.5%), new lesion development (5.1% vs 2.4%), or lymph node metastasis (2.3% vs 2.0%). No distant metastasis occurred. Progression-free survival did not differ significantly between groups (log-rank p = 0.78).

Conclusions: Subcapsular location was not associated with a higher risk of disease progression during AS. With appropriate selection and regular follow-up, AS appears to be a safe management option for ACR TI-RADS 5 nodules, including those near the thyroid capsule.

Objective: To investigate whether sonocavitation, induced by low-intensity focused ultrasound combined with microbubbles, can overcome paclitaxel resistance in ovarian cancer by promoting apoptosis through reactive oxygen species (ROS)-mediated mitochondrial dysfunction.

Methods: Paclitaxel-resistant ovarian cancer tissues and cell lines were compared with chemotherapy-sensitive counterparts for the expression of apoptosis-related proteins. Sonocavitation treatment was applied to resistant cells using optimized ultrasound parameters. Apoptosis, ROS production, mitochondrial morphology, oxygen consumption, mitochondrial membrane potential and mitochondrial membrane proteins were evaluated by flow cytometry, transmission electron microscopy, oxygen consumption assays, adenosine triphosphate (ATP) measurements, mitochondrial membrane potential assay kit staining and Western blotting. In vivo antitumor efficacy and biosafety were examined in paclitaxel-resistant xenograft mouse models, with tumor growth curves, survival analysis, and hematological/organ histology assessments.

Results: Paclitaxel-resistant ovarian cancer tissues exhibited elevated Bcl-2 and reduced Bax and Caspase-3, indicating impaired intrinsic apoptosis. Sonocavitation significantly increased apoptosis in resistant ovarian cancer cells and induced marked mitochondrial dysfunction, including reduced mitochondrial size, disrupted oxygen consumption, decreased ATP levels, collapse of mitochondrial membrane potential and destruction of mitochondrial membrane proteins. Cytochrome c release and activation of cleaved Caspase-3 confirmed mitochondrial-dependent apoptosis. In vivo, sonocavitation suppressed tumor growth and prolonged survival without causing systemic toxicity. ROS scavengers partially reversed these effects, confirming that ROS accumulation is a key mediator of the therapeutic mechanism.

Conclusion: Sonocavitation induces apoptosis in paclitaxel-resistant ovarian cancer through ROS-mediated mitochondrial dysfunction and demonstrates effective tumor-suppressive activity with a favorable safety profile. These findings support sonocavitation as a promising adjuvant strategy to overcome chemoresistance and enhance ovarian cancer treatment outcomes.

Population aging has driven a rise in heart failure cases, increasing the clinical burden on cardiac diagnostics. As a first-line imaging method, transthoracic echocardiography (TTE) faces limitations due to operator dependence, patient variability, and workflow inefficiencies. Meanwhile, advances in artificial intelligence (AI) and robotic ultrasound systems offer new potential pathways toward automated diagnosis. This review examines the current landscape of AI-based image analysis and robotic-assisted echocardiography. It presents a detailed analysis of advancements in artificial intelligence (AI) applied to echocardiography and the evolution of robotic ultrasound systems, aiming to introduce a discussion on semantic-to-motion mapping. By synthesizing recent progress and outlining future directions, we can correctly recognize the current maturity level of artificial intelligence development in the field of ultrasound examination and prepare well for the subsequent work.

Objective: This study aims to develop an advanced deep learning framework to overcome the challenges associated with real-time ultrasound monitoring of the median nerve.

Method: We propose UltraMN, a novel multitask learning model integrating standard plane classification (UltraCLS) and tissue segmentation (UltraSEG) for comprehensive analysis. The dataset includes 446 cases, with 8 videos collected bilaterally for each case, resulting in a total of 3568 videos and 249,985 images across four standardized imaging planes (4-SIP). The classification accuracy of the median nerve was compared among UltraCLS, MedMamba, and FPT models, while precision, recall, F1 scores, and mean Intersection over Union (mIoU) for 4-SIP segmentation were evaluated. Statistical analyses were conducted using Python 3.12.9.

Results: UltraMN significantly outperformed MedMamba and FPT. The UltraCLS module achieved a classification accuracy of 95.6%, with precision, recall, and F1 scores exceeding 95.0% across all standard planes. The UltraSEG module achieved an mIoU of 97.6%, demonstrating superior segmentation performance across all imaging planes.

Conclusion: UltraMN offers a robust and efficient solution for real-time assessment, achieving high classification accuracy and precise segmentation. As a preliminary feasibility study on healthy subjects, this work is based exclusively on ultrasound data of healthy median nerves-its generalizability to pathological scenarios (e.g., carpal tunnel syndrome) requires further validation. It lays the foundation for enhancing clinical workflows in median nerve disorder management, pending subsequent testing on pathological cases.