OBJECTIVES: To construct an intelligent analysis model for classifying fetal orientation during intrapartum ultrasound videos based on multi-feature fusion. METHODS: The proposed model consists of the Input, Backbone Network and Classification Head modules. The Input module carries out data augmentation to improve the sample quality and generalization ability of the model. The Backbone Network was responsible for feature extraction based on Yolov8 combined with CBAM, ECA, PSA attention mechanism and AIFI feature interaction module. The Classification Head consists of a convolutional layer and a softmax function to output the final probability value of each class. The images of the key structures (the eyes, face, head, thalamus, and spine) were annotated with frames by physicians for model training to improve the classification accuracy of the anterior occipital, posterior occipital, and transverse occipital orientations. RESULTS: The experimental results showed that the proposed model had excellent performance in the tire orientation classification task with the classification accuracy reaching 0.984, an area under the PR curve (average accuracy) of 0.993, and area under the ROC curve of 0.984, and a kappa consistency test score of 0.974. The prediction results by the deep learning model were highly consistent with the actual classification results. CONCLUSIONS The multi-feature fusion model proposed in this study can efficiently and accurately classify fetal orientation in intrapartum ultrasound videos.
Humans ; Female ; Ultrasonography, Prenatal/methods* ; Pregnancy ; Fetus/diagnostic imaging* ; Neural Networks, Computer ; Video Recording

The five-year survival rate of gastric cancer in China is close to that in European and American countries but far lower than that in the Republic of Korea and Japan,which have established national gastric cancer screening systems.It is of great significance to build a high-quality gastric cancer screening system adaptive to China's national conditions.Due to the large number of people at risk of gastric cancer and uneven distribution of medical resources,it is still difficult for China to carry out a nationwide gastroscopy screening plan for gastric cancer.Gastric ultrasound,with painlessness,no radiation,and easy acceptance and popularization,could be used as one of the alternative methods for initial screening of gastric cancer.Based on two gastric ultrasound-related consensuses published in 2020,this consensus elaborates on the necessity,feasibility,and existing problems of conducting preliminary gastric cancer ultrasound screening in China by analyzing the gastric cancer screening strategies and the difficulties faced by nationwide gastric cancer screening.Furthermore,this consensus introduces the indications and contraindications of gastric ultrasound examination,requirements for the operator and the contrast agent,ultrasound standard section,essentials of scanning operations,and stomach ultrasound report and data system (Su-RADS) and proposes the relevant consensus opinions accordingly.After multiple rounds of discussions and voting by experts from multiple societies,a total of 17 consensus opinions have been formed on gastric ultrasound as a preliminary screening technique for gastric cancer,with the aim of standardizing the popularization of gastric ultrasound.In addition,the consensus calls for conducting nationwide multicenter prospective studies to improve the level of evidence and provide data support for the construction of a preliminary gastric cancer ultrasound screening system that is in line with China's national conditions.
Humans ; China ; Early Detection of Cancer/methods* ; Mass Screening ; Stomach/diagnostic imaging* ; Stomach Neoplasms/diagnostic imaging* ; Ultrasonography

Gastric contrast-enhanced ultrasound is a non-invasive imaging method that uses oral gastrointestinal ultrasound contrast agents to fill the stomach cavity and display the structure and lesions of the stomach wall.In recent years,the development of contrast agents and the technological innovations of ultrasound equipment have boosted the unique advantages of this examination technique in the diagnosis of gastrointestinal diseases.Gastric contrast-enhanced ultrasound is becoming an important complementary examination means to gastroscopy and X-ray barium meal examination.In this paper,we summarize the clinical applications of gastric contrast-enhanced ultrasound at home and abroad in recent years and systematically analyze its clinical application value and limitations in six aspects:screening and staging of gastric cancer,differentiation and diagnosis of gastric tumors,diagnosis and follow-up of gastritis and gastric ulcers,assessment of gastric contents and gastric volume,evaluation of gastric emptying and gastric motility,and other special applications.
Humans ; Contrast Media ; Ultrasonography/methods* ; Stomach/diagnostic imaging* ; Stomach Neoplasms/diagnostic imaging*

Gastric contrast-enhanced ultrasound includes oral contrast-enhanced ultrasound (OCUS) and double contrast-enhanced ultrasound (DCEUS),which can provide valuable clinical information about tumor morphology,vascular characteristics,and treatment responses.OCUS can clearly identify the gastric wall structure and the extent and depth of lesions by applying oral contrast agents.DCEUS,based on OCUS combined with venography,can display the anatomical and perfusion characteristics of lesions.In recent years,gastric contrast agents and imaging techniques have developed rapidly.However,the clinical application of gastric contrast-enhanced ultrasound is still in the developmental stage.This article reviews the clinical status of OCUS and DCEUS in the screening,diagnosis,staging,pathological typing,and treatment evaluation of gastric cancer.Studies have shown that gastric contrast-enhanced ultrasound has high sensitivity and specificity in the assessment of diagnosis and T-staging of gastric cancer.Furthermore,gastric contrast-enhanced ultrasound has the advantages of being cost-effective,convenient,non-invasive,free from radiation exposure,real-time,and easy to repeat.In the diagnosis and treatment of gastric cancer,it is expected to become one of the important imaging assessment tools.
Humans ; Stomach Neoplasms/diagnostic imaging* ; Contrast Media ; Ultrasonography/methods*

Objective To compare the diagnostic value of transabdominal intestinal ultrasound (IUS) and magnetic resonance enterography (MRE) for intestinal stenosis in inflammatory bowel disease (IBD). Methods A retrospective analysis was conducted on the imaging features of 51 IBD patients who underwent both IUS and MRE at Beijing Friendship Hospital,Capital Medical University,between January 2021 and February 2025.With endoscopy as the gold standard,the diagnostic performance of the two methods was compared. Results The sensitivity (84.2% vs. 52.6%,P=0.008) and accuracy (66.7% vs. 45.1%,P=0.035) of IUS for stenosis were higher than those of MRE.In the localization of stenosis,IUS demonstrated higher sensitivity than MRE for detecting stenosis in the terminal ileum (78.6% vs. 35.7%,P=0.070) and colorectum (86.7% vs. 53.3%,P=0.060).Furthermore,IUS showed higher diagnostic accuracy than MRE for terminal ileum stenosis (70.6% vs. 29.4%,P=0.039).The intestinal wall thickness[(8.2±2.7) mm vs. (10.3±3.8) mm;t=3.20,P=0.002)] and stenosis inner diameter[(3.0±1.6) mm vs. (4.3±1.8) mm;t=2.15,P=0.035] measured by IUS were lower than those measured by MRE,with a moderate level of consistency (ICC:0.19-0.53).In addition,IUS demonstrated a higher detection rate for mesenteric fat hypertrophy (70.6% vs. 27.5%,Kappa=0.27,P=0.005),whereas MRE was more sensitive in detecting lymphadenopathy (90.2% vs. 56.9%,Kappa=0.16,P=0.080). Conclusions IUS is superior to MRE in the diagnosis and localization sensitivity for intestinal stenosis in IBD.However,the two methods showcase poor consistency in detecting and quantitating some inflammatory signs.IUS can be used as a first-line screening method for diagnosing intestinal stenosis in IBD patients,while its clinical application should be combined with specific needs to optimize diagnosis.
Humans ; Retrospective Studies ; Constriction, Pathologic/diagnostic imaging* ; Ultrasonography/methods* ; Magnetic Resonance Imaging/methods* ; Inflammatory Bowel Diseases/diagnostic imaging* ; Male ; Female ; Adult ; Middle Aged ; Intestines/diagnostic imaging* ; Sensitivity and Specificity

The accurate segmentation of breast ultrasound images is an important precondition for the lesion determination. The existing segmentation approaches embrace massive parameters, sluggish inference speed, and huge memory consumption. To tackle this problem, we propose T ²KD Attention U-Net (dual-Teacher Knowledge Distillation Attention U-Net), a lightweight semantic segmentation method combined double-path joint distillation in breast ultrasound images. Primarily, we designed two teacher models to learn the fine-grained features from each class of images according to different feature representation and semantic information of benign and malignant breast lesions. Then we leveraged the joint distillation to train a lightweight student model. Finally, we constructed a novel weight balance loss to focus on the semantic feature of small objection, solving the unbalance problem of tumor and background. Specifically, the extensive experiments conducted on Dataset BUSI and Dataset B demonstrated that the T ²KD Attention U-Net outperformed various knowledge distillation counterparts. Concretely, the accuracy, recall, precision, Dice, and mIoU of proposed method were 95.26%, 86.23%, 85.09%, 83.59%and 77.78% on Dataset BUSI, respectively. And these performance indexes were 97.95%, 92.80%, 88.33%, 88.40% and 82.42% on Dataset B, respectively. Compared with other models, the performance of this model was significantly improved. Meanwhile, compared with the teacher model, the number, size, and complexity of student model were significantly reduced (2.2×10 ⁶ vs. 106.1×10 ⁶, 8.4 MB vs. 414 MB, 16.59 GFLOPs vs. 205.98 GFLOPs, respectively). Indeedy, the proposed model guarantees the performances while greatly decreasing the amount of computation, which provides a new method for the deployment of clinical medical scenarios.
Humans ; Breast Neoplasms/diagnostic imaging* ; Female ; Ultrasonography, Mammary/methods* ; Image Processing, Computer-Assisted/methods* ; Algorithms ; Neural Networks, Computer ; Breast/diagnostic imaging*

Medical cross-modal retrieval aims to achieve semantic similarity search between different modalities of medical cases, such as quickly locating relevant ultrasound images through ultrasound reports, or using ultrasound images to retrieve matching reports. However, existing medical cross-modal hash retrieval methods face significant challenges, including semantic and visual differences between modalities and the scalability issues of hash algorithms in handling large-scale data. To address these challenges, this paper proposes a Medical image Semantic Alignment Cross-modal Hashing based on Transformer (MSACH). The algorithm employed a segmented training strategy, combining modality feature extraction and hash function learning, effectively extracting low-dimensional features containing important semantic information. A Transformer encoder was used for cross-modal semantic learning. By introducing manifold similarity constraints, balance constraints, and a linear classification network constraint, the algorithm enhanced the discriminability of the hash codes. Experimental results demonstrated that the MSACH algorithm improved the mean average precision (MAP) by 11.8% and 12.8% on two datasets compared to traditional methods. The algorithm exhibits outstanding performance in enhancing retrieval accuracy and handling large-scale medical data, showing promising potential for practical applications.
Algorithms ; Semantics ; Humans ; Ultrasonography ; Information Storage and Retrieval/methods* ; Image Processing, Computer-Assisted/methods*

To address the high computational complexity of the Transformer in the segmentation of ultrasound thyroid nodules and the loss of image details or omission of key spatial information caused by traditional image sampling techniques when dealing with high-resolution, complex texture or uneven density two-dimensional ultrasound images, this paper proposes a thyroid nodule segmentation method that integrates the receiving weighted key-value (RWKV) architecture and spherical geometry feature (SGF) sampling technology. This method effectively captures the details of adjacent regions through two-dimensional offset prediction and pixel-level sampling position adjustment, achieving precise segmentation. Additionally, this study introduces a patch attention module (PAM) to optimize the decoder feature map using a regional cross-attention mechanism, enabling it to focus more precisely on the high-resolution features of the encoder. Experiments on the thyroid nodule segmentation dataset (TN3K) and the digital database for thyroid images (DDTI) show that the proposed method achieves dice similarity coefficients (DSC) of 87.24% and 80.79% respectively, outperforming existing models while maintaining a lower computational complexity. This approach may provide an efficient solution for the precise segmentation of thyroid nodules.
Thyroid Nodule/diagnostic imaging* ; Humans ; Ultrasonography/methods* ; Algorithms ; Image Processing, Computer-Assisted/methods* ; Thyroid Gland/diagnostic imaging*

Thyroid nodules are a common endocrine disorder, and their early detection and accurate diagnosis are crucial for the prevention of thyroid cancer. However, the highly heterogeneous morphology and boundaries of thyroid nodules pose significant challenges to their precise identification and classification. Traditional diagnostic approaches rely heavily on physicians' experience, which increases the risk of misdiagnosis and missed diagnoses. With the rapid advancement of computer-aided diagnosis (CAD) technologies, applying deep learning algorithms to the analysis of thyroid nodule ultrasound images has shown great potential. This paper reviews the latest research progress on deep learning-based CAD methods for thyroid nodules, with a focus on their applications in image preprocessing, segmentation and classification. The advantages and limitations of current techniques are analyzed, and potential future directions are discussed. This review aims to highlight the potential of deep learning in thyroid nodule diagnosis and to provide a foundation for selecting feasible pathways for future clinical applications.
Humans ; Thyroid Nodule/diagnostic imaging* ; Deep Learning ; Ultrasonography/methods* ; Diagnosis, Computer-Assisted/methods* ; Algorithms ; Thyroid Neoplasms/diagnostic imaging* ; Image Processing, Computer-Assisted/methods*

OBJECTIVE: To investigate the accuracy of color Doppler ultrasound (CDU) and CT angiography (CTA) in the preoperative evaluation of perforator vessels in free posterior interosseous artery perforator (PIAP) flaps. METHODS: Between January 2020 and December 2023, 19 patients with hand skin and soft tissue defects caused by trauma were admitted. There were 11 males and 8 females, with a median age of 45 years (range, 26-54 years). The interval between injury and admission was 5-11 days (mean, 7.2 days). The skin and soft tissue defects were located on the dorsum of the hand in 8 cases and on the fingers in 11 cases. The size of defect ranged from 4.0 cm×2.5 cm to 7.5 cm×3.5 cm. After locating the perforator vessels through CDU and CTA before operation, the free PIAP flaps were designed to repair hand defects, with the size of 4.5 cm×3.0 cm-7.5 cm×4.0 cm. The defects of donor sites were directly sutured. The number and diameter of perforator vessels in the posterior interosseous artery detected by CDU and CTA were compared. The differences in localization of perforator vessels using CDU and CTA and their clinical effects were also compared to calculate the accuracy and recognition rate. During follow-up, the survival of the skin flap was observed, and the Vancouver scar scale (VSS) score was used to evaluate the healing of the donor site, while the visual analogue scale (VAS) score was used to evaluate the patient's satisfaction with the appearance of the skin flap. RESULTS: The number and the diameter of PIAP vessels was 5.8±1.2 and (0.62±0.08) mm assessed by CDU and 5.2±1.0 and (0.60±0.07) mm by CTA, showing no significant difference between the two methods ( P>0.05). The number, course, and distribution of perforator vessels of the PIAP vessels observed during operation were basically consistent with those detected by preoperative CDU and CTA. Compared with intraoperative observation results, the recognition rates of dominant perforating vessels by CDU and CTA were 95.0% (18/19) and 89.5% (17/19), respectively, and the accuracy rates were 100% (19/19) and 84.2% (16/19), with no significant difference between the two methods ( P>0.05). All flaps survived after operation, and all wounds and incisions at donor sites healed by first intention. All patients were followed up 6-13 months (mean, 8.2 months). At last follow-up, the skin flaps had elasticity and soft texture，with the patient satisfaction VAS score of 9.2±0.8. The donor sites had no obvious scar hyperplasia with the VSS score of 11.7±0.9. CONCLUSION CDU and CTA accurately identify the dominant perforator vessels and provide reliable information for vessel localization, facilitating precise flap harvesting and minimizing donor site injury. However, CDU offers superior visualization of distal end of perforator vessels in the forearm compared to CTA.
Humans ; Female ; Male ; Adult ; Perforator Flap/blood supply* ; Middle Aged ; Ultrasonography, Doppler, Color/methods* ; Computed Tomography Angiography/methods* ; Soft Tissue Injuries/diagnostic imaging* ; Hand Injuries/diagnostic imaging* ; Plastic Surgery Procedures/methods* ; Hand/surgery* ; Preoperative Care ; Arteries/diagnostic imaging*