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.

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.

The auxin/indole-3-acetic acid (Aux/IAA) gene family is an important regulator for plant growth hormone signaling, involved in plant growth, development, as well as response to environmental stresses. In the present study, we identified SmIAA7 which is potentially associated with Salvia miltiorrhiza leaf development through comparatively analyzed the transcriptome data from different pinnate leaves. SmIAA7 was successfully isolated from S. miltiorrhiza using the specific primers. Then subsequent bioinformatic analysis, prokaryotic expression and purification, subcellular localization, and induction expression analysis under auxin and abiotic stress were performed. The full-length of SmIAA7 contained an ORF of 684 bp encoding a protein of 227 amino acid with a molecular weight of 25.3 kD. Conserved domain analysis showed that SmIAA7 contains the conserved Aux_IAA domain (pfam02309). Sequence analysis and phylogenetic tree analysis results indicated SmIAA7 was phylogenetically close to IAA7 and IAA14 from other plants, suggesting SmIAA7 involved in plant growth and development as well as response to environmental stresses. The prokaryotic expression vector pET28a-SmIAA7 was constructed and SmIAA7 recombinant protein was successfully expressed in E. coli Rosetta (DE3) strain. Subcellular localization experiment demonstrated that SmIAA7 localized in the nucleus of plant cells. Real-time fluorescence quantitative PCR results showed that the expression level of SmIAA7 was upregulated in response to auxin. Drought, low temperature, and salt stress significantly increased the transcript level of SmIAA7 gene. This study lays a foundation for further elucidating the role of SmIAA7 in leaf development, signal transduction, and stress defense in S. miltiorrhiza.

AIM To study the chemical constituents from the water fraction of rhizoma of Smilax trinervula Miq.and their biological activities.METHODS Polyamide,silica gel,Sephadex LH-20,ODS and semi-preparative HPLC were used for isolation and purification,then the structures of obtained compounds were identified by physicochemical properties and spectral data.The antitumor activities were determined by MTT mothod,and the inhibitory activities on α-glucosidase were determined by PNPG method.RESULTS Eleven compounds were isolated and identified as tyrosine(1),uridine(2),2-(2',3',4'-trihydroxybutyl)-6-(2",3",4"-trihydroxybutyl)-pyrazine(3),2-(1',2',3',4'-tetrahydroxybutyl)-6-(2",3",4"-trihydroxybutyl)-pyrazine(4),2-(1',2',3',4'-tetrahydroxybutyl)-5-(2",3",4"-trihydroxybutyl)-pyrazine(5),uracil(6),2-(1',2',3',4'-tetrahydroxybutyl)-5-(1",2",3",4"-tetrahydroxybutyl)-pyrazine(7),dioscin(8),shikimic acid(9),pyrazine(10),3,4-dihydroxyphenyethyl alcohol 8-O-β-D-glycopyranoside(11).The IC50 values of compounds 8 to human breast cancer cell MCF-7 was(2.36±0.26)μg/mL,and the IC50 values of compounds 3-5 and 7 to α-glucosidase were(1.54±0.15)-(10.53±0.38)μg/mL.CONCLUSION Compounds 1-7,10 are isolated from Smilax genus for the first time,and compound 9,11 are first isolated from this plant.Compound 8 has anti-tumor activity,and compounds 3-5,7 have α-glucosidase inhibitory activities.

Objective To investigate the distribution and antimicrobial resistance profiles of clinically isolated Haemophilus influenzae and Moraxella catarrhalis in hospitals across China from 2015 to 2021,and provide evidence for rational use of antimicrobial agents.Methods Data of H.influenzae and M.catarrhalis strains isolated from 2015 to 2021 in CHINET program were collected for analysis,and antimicrobial susceptibility testing was performed by disc diffusion method or automated systems according to the uniform protocol of CHINET.The results were interpreted according to the CLSI breakpoints in 2022.Beta-lactamases was detected by using nitrocefin disk.Results From 2015 to 2021,a total of 43 642 strains of Haemophilus species were isolated,accounting for 2.91％of the total clinical isolates and 4.07％of Gram-negative bacteria in CHINET program.Among the 40 437 strains of H.influenzae,66.89％were isolated from children and 33.11％were isolated from adults.More than 90％of the H.influenzae strains were isolated from respiratory tract specimens.The prevalence of β-lactamase was 53.79％in H.influenzae strains.The H.influenzae strains isolated from children showed higher resistance rate than the strains isolated from adults.Overall,779 strains of H.influenzae did not produce β-lactamase but were resistant to ampicillin(BLNAR).Beta-lactamase-producing strains showed significantly higher resistance rates to these antimicrobial agents than the β-lactamase-nonproducing strains.Of the 16 191 M.catarrhalis strains,80.06％were isolated from children and 19.94％isolated from adults.M.catarrhalis strains were mostly susceptible to both amoxicillin-clavulanic acid and cefuroxime,evidenced by resistance rate lower than 2.0％.Conclusions The emergence of antibiotic-resistant H.influenzae due to β-lactamase production poses a challenge for clinical anti-infective treatment.Therefore,it is very important to implement antibiotic resistance surveillance for H.influenzae and guide rational antibiotic use.All local clinical microbiology laboratories should actively improve antibiotic susceptibility testing and strengthen antibiotic resistance surveillance for H.influenzae.

Objective:This study aimed to explore the molecular mechanisms of the maladaptation of H3N2 influenza virus in chicken embryos, provide a theoretical basis for the restoration of H3N2 influenza vaccine production in chicken embryos.Methods:Samples of respiratory secretions from patients with influenza-like symptoms (Influenza-like Illness, ILI) caused by H3N2 influenza virus were inoculated into chicken embryos and Madin-Darby Canine Kidney cells (MDCK), respectively. After isolating the virus, hemagglutination experiments were conducted to detect hemagglutination titers and hemagglutination inhibition experiments were used to compare antigenic differences; further, whole-genome sequencing of H3N2 influenza virus was performed using second-generation high-throughput gene sequencing (Next Generation High-Throughput Gene Sequencing, NGS), and key amino acid sites of mutations were identified through sequence alignment; combined with sialic acid receptor binding experiments, the differences in the binding of wild-type and mutant receptor binding sites (RBS) to sialic acid receptors were compared; finally, molecular docking and molecular dynamics simulation method were used to explore the specific molecular mechanisms of how mutation sites affect the differences in the affinity of the RBS pocket for sialic acid receptors.Results:The hemagglutination assay result indicated that both chicken embryos and MDCK cells could isolate the influenza virus, and the hemagglutination inhibition test showed that no antigenic differences were produced in the isolated strains. NGS analysis revealed that the H3N2 virus underwent an F195Y mutation in the (RBS) region of the hemagglutinin (HA) protein after adaptation through chicken embryo passages. Receptor-binding experiments demonstrated that the F195Y mutation enhanced the virus′s binding ability to α2, 3-linked sialic acid glycan (Neu5Acα2-3Galβ1-4GlcNAcβ-PAA, 3′SLN), while the mutation did not affect the affinity of the RBS pocket for α2, 6-linked sialic acid glycan (Neu5Acα2-6Galβ1-4GlcNAcβ-PAA, 6′SLN). Molecular docking and molecular dynamics simulation result indicate that the F195Y mutation, by replacing a hydrophobic amino acid with a hydrophilic one, leads to a significant decrease in the structure of the RBS pocket, enhancing the binding stability of the H3N2 influenza virus with α2, 3-sln. This is specifically manifested by an increase in binding time and an increase in the number of hydrogen bonds at the RBS site with the receptor. Furthermore, the F195Y mutation does not alter the binding of the virus to other receptors.Conclusions:The F195Y mutation in the RBS pocket of H3N2 influenza virus is a key site affecting the viral chicken embryo inadaptability.

Amber and subfossil resins are subjects of interdisciplinary research across multiple fields. However, due to their diverse origins and complex compositions, different disciplines vary in their definitions and functional interpretations. In traditional Chinese medicine(TCM), amber has been utilized as a medicinal material since ancient time, with extensive historical documentation. However, its classification, provenance, and nomenclature remain ambiguous, and authentic medicinal amber artifacts are exceedingly rare. This study employed Fourier-transform infrared spectroscopy(FTIR) to characterize amber and subfossil resins from various geological sources and commercially "medicinal amber". Additionally, historical literature and market surveys were analyzed to explore their provenance, composition, and functional attributes. The results indicate that amber and subfossil resins from different sources and with different compositions exhibit distinct fingerprint characteristics in the FTIR spectral range of 1 800-700 cm~(-1). "Medicinal amber" available in the market primarily consists of subfossil or modern resins, significantly differing in composition and structure from geological amber. This study highlights the importance of interdisciplinary research on amber identification and resource management. It is essential to establish a systematic database of amber and subfossil resin characteristics and integrate modern analytical techniques to enhance research on their composition, pharmacological mechanisms, and potential therapeutic effects, thereby promoting the standardized utilization of amber resources and advancing the modernization of TCM.
Amber/history* ; Archaeology ; Spectroscopy, Fourier Transform Infrared ; Medicine, Chinese Traditional

OBJECTIVES: To analyze the clinical characteristics of diffuse panbronchiolitis (DPB) in children and to enhance the clinical diagnosis and treatment of this disease. METHODS: A retrospective analysis was conducted on the clinical data of 6 children diagnosed with DPB who were hospitalized at The First Affiliated Hospital of Guangzhou Medical University from January 2011 to December 2019. RESULTS: Among the 6 patients, there were 2 males and 4 females; the age at diagnosis ranged from 7 to 12 years. All patients presented with cough, sputum production, and exertional dyspnea, and all had a history of sinusitis. Two cases showed positive serum cold agglutinin tests, and 5 cases exhibited pathological changes consistent with chronic bronchiolitis. High-resolution chest CT in all patients revealed centrilobular nodules diffusely distributed throughout both lungs with a tree-in-bud appearance. Five patients received low-dose azithromycin maintenance therapy, but 3 showed inadequate treatment response. After empirical anti-tuberculosis treatment, non-tuberculous Mycobacteria were found in the bronchoalveolar lavage fluid. Follow-up over 2 years showed 1 case cured, 3 cases significantly improved, and 2 cases partially improved. CONCLUSIONS The clinical presentation of DPB is non-specific and can easily lead to misdiagnosis. In cases where DPB is clinically diagnosed but does not show improvement with low-dose azithromycin treatment, special infections should be considered.
Humans ; Male ; Female ; Bronchiolitis/drug therapy* ; Retrospective Studies ; Child ; Haemophilus Infections/diagnosis*