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.

Objective:To analyze the clinical effect of XinjiaHuanglong Decoction add and subtract combined with auricular point sticking on delirium after hip fracture in elderly patients.Methods:Select 80 elderly patients with postoperative delirium after hip fracture admitted to our hospital from January 2022 to June 2024,and divide them into a matched group and an observation group,with 40 cases in each group.The matched group was treated with olanzapine,and the observation group was treated with XinjiaHuanglong Decoction add and subtract combined with auricular point sticking.Inflammatory response indexes(IL-6,TNF-α,hs-CRP)and stress response indexes(dopamine(DA),5-hydroxytryptamine(5-HT)were detected before and Post-treatment in both groups.Compare the Delirium Rating Scale 98(DRS-R-98)scores,Mini Mental State Examination(MMSE)scores,and duration of delirium between two groups before and Post-treatment,comprehensively evaluate the efficacy,and record the occurrence of adverse reactions.Results:Post-treatment,the DRS-R-98 score in the observation group was lower than that in the matched group,and the MMSE score was higher than that in the matched group(P＜0.05);The duration of delirium in the observation group was shorter than that in the matched group,and the overall effective rate was higher than that in the matched group(P＜0.05).The levels of inflammatory response indicators in the observation group were lower than those in the matched group Post-treatment(P＜0.05);The level of stress response indicators in the observation group was lower than that in the matched group Post-treatment(P＜0.05);There was no difference in the incidence of adverse reactions between the two groups(P＞0.05).Conclusion:XinjiaHuanglong Decoction add and subtract combined with auricular point sticking has a certain effect on improving delirium after hip fracture in the elderly,and can shorten the duration of delirium,which may be related to reducing inflammation and stress response,and it is safe and worthy of clinical application.

Objective:To analyze the clinical effect of XinjiaHuanglong Decoction add and subtract combined with auricular point sticking on delirium after hip fracture in elderly patients.Methods:Select 80 elderly patients with postoperative delirium after hip fracture admitted to our hospital from January 2022 to June 2024,and divide them into a matched group and an observation group,with 40 cases in each group.The matched group was treated with olanzapine,and the observation group was treated with XinjiaHuanglong Decoction add and subtract combined with auricular point sticking.Inflammatory response indexes(IL-6,TNF-α,hs-CRP)and stress response indexes(dopamine(DA),5-hydroxytryptamine(5-HT)were detected before and Post-treatment in both groups.Compare the Delirium Rating Scale 98(DRS-R-98)scores,Mini Mental State Examination(MMSE)scores,and duration of delirium between two groups before and Post-treatment,comprehensively evaluate the efficacy,and record the occurrence of adverse reactions.Results:Post-treatment,the DRS-R-98 score in the observation group was lower than that in the matched group,and the MMSE score was higher than that in the matched group(P＜0.05);The duration of delirium in the observation group was shorter than that in the matched group,and the overall effective rate was higher than that in the matched group(P＜0.05).The levels of inflammatory response indicators in the observation group were lower than those in the matched group Post-treatment(P＜0.05);The level of stress response indicators in the observation group was lower than that in the matched group Post-treatment(P＜0.05);There was no difference in the incidence of adverse reactions between the two groups(P＞0.05).Conclusion:XinjiaHuanglong Decoction add and subtract combined with auricular point sticking has a certain effect on improving delirium after hip fracture in the elderly,and can shorten the duration of delirium,which may be related to reducing inflammation and stress response,and it is safe and worthy of clinical application.

Objective:To investigate the effects of traditional Chinese curve-correcting and rotation-reducing spinal manipulation on biomechanical parameters and factors influencing herniated disc resorption in lumbar disc herniation(LDH).Methods:A retrospective analysis of 51 LDH patients treated between January 2022 and May 2024 was conducted.Lumbosacral parameters(vertebral rotation angle[α],disc angle[β],sacral slope[SS],lumbar lordosis[LL])were measured via MRI before treatment and at final follow-up.Disc resorption was assessed using Michigan State University(MSU)classification.Multivariate logistic regression identified factors associated with resorption.Results:Post-treatment α angle significantly decreased(3.02°→1.86°,P=0.002),while SS(28.4°→30.0°,P＜0.001)and LL angles(31.0°→35.12°,P＜0.001)increased;Disc resorption occurred in 56.86％(29/51)of patients.Longer disease course(OR=0.79,95％CI:0.69-0.91)and disc calcification(OR=0.03,95％CI:0.00-0.25)were independent inhibitors of resorption(P＜0.001).Conclusion:Spinal manipulation restores lumbosacral biomechanics by reducing vertebral rotation and increasing lumbar curvature,with higher resorption rates in patients with short duration(≤6 months),non-calcified discs,and MSU type 2-3 herniations.

Objective:To investigate the effects of traditional Chinese curve-correcting and rotation-reducing spinal manipulation on biomechanical parameters and factors influencing herniated disc resorption in lumbar disc herniation(LDH).Methods:A retrospective analysis of 51 LDH patients treated between January 2022 and May 2024 was conducted.Lumbosacral parameters(vertebral rotation angle[α],disc angle[β],sacral slope[SS],lumbar lordosis[LL])were measured via MRI before treatment and at final follow-up.Disc resorption was assessed using Michigan State University(MSU)classification.Multivariate logistic regression identified factors associated with resorption.Results:Post-treatment α angle significantly decreased(3.02°→1.86°,P=0.002),while SS(28.4°→30.0°,P＜0.001)and LL angles(31.0°→35.12°,P＜0.001)increased;Disc resorption occurred in 56.86％(29/51)of patients.Longer disease course(OR=0.79,95％CI:0.69-0.91)and disc calcification(OR=0.03,95％CI:0.00-0.25)were independent inhibitors of resorption(P＜0.001).Conclusion:Spinal manipulation restores lumbosacral biomechanics by reducing vertebral rotation and increasing lumbar curvature,with higher resorption rates in patients with short duration(≤6 months),non-calcified discs,and MSU type 2-3 herniations.

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.

Bacterial biofilms gave rise to persistent infections and multi-organ failure, thereby posing a serious threat to human health. Biofilms were formed by cross-linking of hydrophobic extracellular polymeric substances (EPS), such as proteins, polysaccharides, and eDNA, which were synthesized by bacteria themselves after adhesion and colonization on biological surfaces. They had the characteristics of dense structure, high adhesiveness and low drug permeability, and had been found in many human organs or tissues, such as the brain, heart, liver, spleen, lungs, kidneys, gastrointestinal tract, and skeleton. By releasing pro-inflammatory bacterial metabolites including endotoxins, exotoxins and interleukin, biofilms stimulated the body’s immune system to secrete inflammatory factors. These factors triggered local inflammation and chronic infections. Those were the key reason for the failure of traditional clinical drug therapy for infectious diseases.In order to cope with the increasingly severe drug-resistant infections, it was urgent to develop new therapeutic strategies for bacterial-biofilm eradication and anti-bacterial infections. Based on the nanoscale structure and biocompatible activity, nanobiomaterials had the advantages of specific targeting, intelligent delivery, high drug loading and low toxicity, which could realize efficient intervention and precise treatment of drug-resistant bacterial biofilms. This paper highlighted multiple strategies of biofilms eradication based on nanobiomaterials. For example, nanobiomaterials combined with EPS degrading enzymes could be used for targeted hydrolysis of bacterial biofilms, and effectively increased the drug enrichment within biofilms. By loading quorum sensing inhibitors, nanotechnology was also an effective strategy for eradicating bacterial biofilms and recovering the infectious symptoms. Nanobiomaterials could intervene the bacterial metabolism and break the bacterial survival homeostasis by blocking the uptake of nutrients. Moreover, energy-driven micro-nano robotics had shown excellent performance in active delivery and biofilm eradication. Micro-nano robots could penetrate physiological barriers by exogenous or endogenous driving modes such as by biological or chemical methods, ultrasound, and magnetic field, and deliver drugs to the infection sites accurately. Achieving this using conventional drugs was difficult. Overall, the paper described the biological properties and drug-resistant molecular mechanisms of bacterial biofilms, and highlighted therapeutic strategies from different perspectives by nanobiomaterials, such as dispersing bacterial mature biofilms, blocking quorum sensing, inhibiting bacterial metabolism, and energy driving penetration. In addition, we presented the key challenges still faced by nanobiomaterials in combating bacterial biofilm infections. Firstly, the dense structure of EPS caused biofilms spatial heterogeneity and metabolic heterogeneity, which created exacting requirements for the design, construction and preparation process of nanobiomaterials. Secondly, biofilm disruption carried the risk of spread and infection the pathogenic bacteria, which might lead to other infections. Finally, we emphasized the role of nanobiomaterials in the development trends and translational prospects in biofilm treatment.

Objective To analyze the therapeutic effects and survival benefits of sintilimab combined with albumin-bound paclitaxel chemotherapy in the treatment of patients with advanced gastric cancer.Methods Patients with advanced gastric cancer were divided into the treatment group and the control group by cohort method.The control group was treated with albumin-bound paclitaxel-based chemotherapy[intravenous infusion of albumin-bound paclitaxel at 125 mg·m-2 from day 1 to day 8,for a cycle(21 days as a cycle);Tiggio capsule 40 mg·m-2·d-1 was taken orally for 1-14 days for 1 consecutive cycle;Trastuzumab was administered once every 3 weeks at an initial loading dose of 8 mg·kg-1,followed by maintenance treatment at a dose of 6 mg·kg-1 every 3 weeks].On this basis,the treatment group was treated with intravenous infusion of sintilimab injection at a dose of 200 mg·time-1 on the first day of each cycle,with 21 d as a cycle.After 6 cycles of continuous treatment,both groups were given maintenance treatment and were followed up for 8 months.The two groups were compared in terms of clinical efficacy,the levels of serum tumor markers[carbohydrate antigen 242(CA242),carbohydrate antigen 724(CA724),carcinoembryonic antigen(CEA),tissue polypeptide-specific antigen(TPS),soluble intercellular cell adhesion molecule-1(sICAM-1)and E-cadherin],survival and evaluated the safety.Results In this study,39 and 41 patients were enrolled in the control group and the treatment group,respectively.At the end of treatment,the objective response rates(ORR)in the treatment group and the control group were 56.10％and 33.33％;the disease control rates(DCR)were 78.05％and 48.71％.The differences were statistically significant(all P＜0.05).After treatment,serum CA242 levels in the treatment group and the control group were(57.64±5.82)and(68.95±7.23)mg·L-1;CA724 levels were(36.58±3.79)and(43.65±4.48)U·mL-1;CEA levels were(17.33±1.78)and(20.16±2.35)ng·mL-1;TPS levels were(21.35±2.44)and(37.65±3.84)U·L-1;sICAM-1 levels were(216.77±22.53)and(275.34±28.63)ng·mL-1;E-cadherin levels were(12.15±1.36)and(9.87±1.45)ng·mL-1.The differences were statistically significant(all P＜0.05).The average progression free survival(PFS)of the treatment group and the control group was 7.55 months and 7.17 months;PFS rates were 65.78％and 56.42％.The differences were statistically significant(P＜0.05).The adverse drug reactions in the treatment group and the control group were mainly bone marrow suppression,nausea and vomiting,liver function damage,peripheral nerve paresthesia,and hypothyroidism.There was no statistically significant difference in the above adverse drug reactions between the treatment group and the control group(all P＞0.05).Conclusion Sintilimab combined with albumin-bound paclitaxel chemotherapy is effective in the treatment of patients with advanced gastric cancer,which can significantly improve serum tumor markers and prolong PFS,with good safety.