ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

Diabetes is a very complex endocrine disease whose common feature is the increase in blood glucose concentration. Persistent hyperglycemia can lead to blindness, kidney and heart disease, neurodegeneration, and many other serious complications that have a significant impact on human health and quality of life. The number of people with diabetes is increasing yearly. The global diabetes prevalence in 20-79 year olds in 2021 was estimated to be 10.5% (536.6 million), and it will rise to 12.2% (783.2 million) in 2045. The main modes of intervention for diabetes include medication, dietary management, and exercise conditioning. Medication is the mainstay of treatment. Marketed diabetes drugs such as metformin and insulin, as well as GLP-1 receptor agonists, are effective in controlling blood sugar levels to some extent, but the preventive and therapeutic effects are still unsatisfactory. Peptide drugs have many advantages such as low toxicity, high target specificity, and good biocompatibility, which opens up new avenues for the treatment of diabetes and other diseases. Currently, insulin and its analogs are by far the main life-saving drugs in clinical diabetes treatment, enabling effective control of blood glucose levels, but the risk of hypoglycemia is relatively high and treatment is limited by the route of delivery. New and oral anti-diabetic drugs have always been a market demand and research hotspot. Inhibitor cystine knot (ICK) peptides are a class of multifunctional cyclic peptides. In structure, they contain three conserved disulfide bonds (C3-C20, C7-C22, and C15-C32) form a compact “knot” structure, which can resist degradation of digestive protease. Recent studies have shown that ICK peptides derived from legume, such as PA1b, Aglycin, Vglycin, Iglycin, Dglycin, and aM₁, exhibit excellent regulatory activities on glucose and lipid metabolism at the cellular and animal levels. Mechanistically, ICK peptides promote glucose utilization by muscle and liver through activation of IR/AKT signaling pathway, which also improves insulin resistance. They can repair the damaged pancrease through activation of PI3K/AKT/Erk signaling pathway, thus lowering blood glucose. The biostability and hypoglycemic efficacy of the ICK peptides meet the requirements for commercialization of oral drugs, and in theory, they can be developed into natural oral anti-diabetes peptide drugs. In this review, the structural properties, activity and mechanism of ICK pattern peptides in regulating glucose and lipid metabolism were summaried, which provided a reference for the development of new oral peptides for diabetes.

Turning to critical illness is a common stage of various diseases and injuries before death. Patients usually have complex health conditions, while the treatment process involves a wide range of content, along with high requirements for doctor′s professionalism and multi-specialty teamwork, as well as a great demand for time-sensitive treatments. However, this is not matched with critical care professionals and the current state of medical care in China. Telemedicine, which shortens the distance of medical professionals and the gap of disease diagnosis and treatments in various regions through electronic information, can effectively solve the current problem. Therefore, there is an urgent need to develop a standardized, high-quality visualization telemedicine round system .Therefore, experts have been organized to search domestic and foreign literature on telemedicine round for critically ill patients and to form this consensus based on clinical experiences so as to further improve the level of critical care treatments in regions.

ObjectivePrimary liver cancer, predominantly hepatocellular carcinoma (HCC), is a significant global health issue, ranking as the sixth most diagnosed cancer and the third leading cause of cancer-related mortality. Accurate and early diagnosis of HCC is crucial for effective treatment, as HCC and non-HCC malignancies like intrahepatic cholangiocarcinoma (ICC) exhibit different prognoses and treatment responses. Traditional diagnostic methods, including liver biopsy and contrast-enhanced ultrasound (CEUS), face limitations in applicability and objectivity. The primary objective of this study was to develop an advanced, light-weighted classification network capable of distinguishing HCC from other non-HCC malignancies by leveraging the automatic analysis of brightness changes in CEUS images. The ultimate goal was to create a user-friendly and cost-efficient computer-aided diagnostic tool that could assist radiologists in making more accurate and efficient clinical decisions. MethodsThis retrospective study encompassed a total of 161 patients, comprising 131 diagnosed with HCC and 30 with non-HCC malignancies. To achieve accurate tumor detection, the YOLOX network was employed to identify the region of interest (ROI) on both B-mode ultrasound and CEUS images. A custom-developed algorithm was then utilized to extract brightness change curves from the tumor and adjacent liver parenchyma regions within the CEUS images. These curves provided critical data for the subsequent analysis and classification process. To analyze the extracted brightness change curves and classify the malignancies, we developed and compared several models. These included one-dimensional convolutional neural networks (1D-ResNet, 1D-ConvNeXt, and 1D-CNN), as well as traditional machine-learning methods such as support vector machine (SVM), ensemble learning (EL), k-nearest neighbor (KNN), and decision tree (DT). The diagnostic performance of each method in distinguishing HCC from non-HCC malignancies was rigorously evaluated using four key metrics: area under the receiver operating characteristic (AUC), accuracy (ACC), sensitivity (SE), and specificity (SP). ResultsThe evaluation of the machine-learning methods revealed AUC values of 0.70 for SVM, 0.56 for ensemble learning, 0.63 for KNN, and 0.72 for the decision tree. These results indicated moderate to fair performance in classifying the malignancies based on the brightness change curves. In contrast, the deep learning models demonstrated significantly higher AUCs, with 1D-ResNet achieving an AUC of 0.72, 1D-ConvNeXt reaching 0.82, and 1D-CNN obtaining the highest AUC of 0.84. Moreover, under the five-fold cross-validation scheme, the 1D-CNN model outperformed other models in both accuracy and specificity. Specifically, it achieved accuracy improvements of 3.8% to 10.0% and specificity enhancements of 6.6% to 43.3% over competing approaches. The superior performance of the 1D-CNN model highlighted its potential as a powerful tool for accurate classification. ConclusionThe 1D-CNN model proved to be the most effective in differentiating HCC from non-HCC malignancies, surpassing both traditional machine-learning methods and other deep learning models. This study successfully developed a user-friendly and cost-efficient computer-aided diagnostic solution that would significantly enhances radiologists’ diagnostic capabilities. By improving the accuracy and efficiency of clinical decision-making, this tool has the potential to positively impact patient care and outcomes. Future work may focus on further refining the model and exploring its integration with multimodal ultrasound data to maximize its accuracy and applicability.

Accurate pain assessment is a prerequisite for pain management.However,pain is a relatively subjective feeling greatly affected by the environment,education,and emotional state.Therefore,in diagnosing and treating pain,it is necessary not only to assess the subjective feelings of patients,but also to have stable and reliable objective indicators to evaluate the pain status of different patients.At present,there are many clinical tools for assessing and monitoring pain.This paper introduces the commonly used subjective assessment tools(single-dimensional pain assessment scale,multi-dimensional pain assessment scale)and objective assessment tools(surgical pleth index,analgesia and nociception index,nociceptive level index,and biomarkers),and analyzes their advantages and disadvantages,so as to provide reference for future clinical pain assessment.

Accurate pain assessment is a prerequisite for pain management.However,pain is a relatively subjective feeling greatly affected by the environment,education,and emotional state.Therefore,in diagnosing and treating pain,it is necessary not only to assess the subjective feelings of patients,but also to have stable and reliable objective indicators to evaluate the pain status of different patients.At present,there are many clinical tools for assessing and monitoring pain.This paper introduces the commonly used subjective assessment tools(single-dimensional pain assessment scale,multi-dimensional pain assessment scale)and objective assessment tools(surgical pleth index,analgesia and nociception index,nociceptive level index,and biomarkers),and analyzes their advantages and disadvantages,so as to provide reference for future clinical pain assessment.

Turning to critical illness is a common stage of various diseases and injuries before death. Patients usually have complex health conditions, while the treatment process involves a wide range of content, along with high requirements for doctor′s professionalism and multi-specialty teamwork, as well as a great demand for time-sensitive treatments. However, this is not matched with critical care professionals and the current state of medical care in China. Telemedicine, which shortens the distance of medical professionals and the gap of disease diagnosis and treatments in various regions through electronic information, can effectively solve the current problem. Therefore, there is an urgent need to develop a standardized, high-quality visualization telemedicine round system .Therefore, experts have been organized to search domestic and foreign literature on telemedicine round for critically ill patients and to form this consensus based on clinical experiences so as to further improve the level of critical care treatments in regions.

Hypoxia-inducible factor (HIF-1α), a core transcription factor responding to changes in cellular oxygen levels, is closely associated with a wide range of physiological and pathological conditions. However, its differential impacts on vascular cell types and molecular programs modulating human vascular homeostasis and regeneration remain largely elusive. Here, we applied CRISPR/Cas9-mediated gene editing of human embryonic stem cells and directed differentiation to generate HIF-1α-deficient human vascular cells including vascular endothelial cells, vascular smooth muscle cells, and mesenchymal stem cells (MSCs), as a platform for discovering cell type-specific hypoxia-induced response mechanisms. Through comparative molecular profiling across cell types under normoxic and hypoxic conditions, we provide insight into the indispensable role of HIF-1α in the promotion of ischemic vascular regeneration. We found human MSCs to be the vascular cell type most susceptible to HIF-1α deficiency, and that transcriptional inactivation of ANKZF1, an effector of HIF-1α, impaired pro-angiogenic processes. Altogether, our findings deepen the understanding of HIF-1α in human angiogenesis and support further explorations of novel therapeutic strategies of vascular regeneration against ischemic damage.
Humans ; Vascular Endothelial Growth Factor A/metabolism* ; Endothelial Cells/metabolism* ; Transcription Factors/metabolism* ; Gene Expression Regulation ; Hypoxia/metabolism* ; Cell Hypoxia/physiology*

Objective To explore the efficacy and safety of recombinant human anti-severe acute respiratory syn-drome coronavirus 2(anti-SARS-CoV-2)monoclonal antibody injection(F61 injection)in the treatment of patients with coronavirus disease 2019(COVID-19)combined with renal damage.Methods Patients with COVID-19 and renal damage who visited the PLA General Hospital from January to February 2023 were selected.Subjects were randomly divided into two groups.Control group was treated with conventional anti-COVID-19 therapy,while trial group was treated with conventional anti-COVID-19 therapy combined with F61 injection.A 15-day follow-up was conducted after drug administration.Clinical symptoms,laboratory tests,electrocardiogram,and chest CT of pa-tients were performed to analyze the efficacy and safety of F61 injection.Results Twelve subjects(7 in trial group and 5 in control group)were included in study.Neither group had any clinical progression or death cases.The ave-rage time for negative conversion of nucleic acid of SARS-CoV-2 in control group and trial group were 3.2 days and 1.57 days(P=0.046),respectively.The scores of COVID-19 related target symptom in the trial group on the 3rd and 5th day after medication were both lower than those of the control group(both P＜0.05).According to the clinical staging and World Health Organization 10-point graded disease progression scale,both groups of subjects improved but didn't show statistical differences(P＞0.05).For safety,trial group didn't present any infusion-re-lated adverse event.Subjects in both groups demonstrated varying degrees of elevated blood glucose,elevated urine glucose,elevated urobilinogen,positive urine casts,and cardiac arrhythmia,but the differences were not statistica-lly significant(all P＞0.05).Conclusion F61 injection has initially demonstrated safety and clinical benefit in trea-ting patients with COVID-19 combined with renal damage.As the domestically produced drug,it has good clinical accessibility and may provide more options for clinical practice.