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.

Background/Aims: Lenvatinib resistance remains a critical barrier in advanced hepatocellular carcinoma (HCC) therapy. However, the underlying mechanisms and strategies for reversing resistance remain incompletely understood. Methods: Integrated transcriptomics of lenvatinib-resistant patient tumors and an acquired-resistance murine model identified a novel macrophage subpopulation. Functional validation employed CRISPR-SAM screening, conditioned medium (CM) assays, subcutaneous/orthotopic xenografts, patient-derived organoids (PDOs), and patient-derived xenografts (PDXs). Mechanistic studies included ChIP-qPCR, co-immunoprecipitation, and pharmacologic targeting. Clinical relevance was assessed in a retrospective cohort. Results: Resistant HCC exhibited significant enrichment of a COLEC12high TAM subset , which correlated with poor survival and treatment response. These TAMs secreted neuregulin-1 (NRG1) , activating HER2/HER3-AKT signaling in tumor cells to drive cancer stemness and lenvatinib resistance. Mechanistically, in TAMs COLEC12 sequestered STAT1 in the cytoplasm, preventing its phosphorylation, and thereby derepressing STAT3-mediated NRG1 transcription. Depletion of NRG1 reversed the stemness phenotypes and resensitized tumors to lenvatinib both in vitro and in vivo. Clinically, high NRG1 expression predicted an inferior lenvatinib response and shorter survival. Crucially, the bispecific anti-HER2/HER3 antibody zenocutuzumab restored lenvatinib efficacy in PDOs, PDXs, and murine models. Conclusions Our work establishes the COLEC12high TAM/NRG1 axis as a master regulator of therapeutic resistance and identifies NRG1 as a predictive biomarker, providing a clinically actionable strategy to overcome lenvatinib resistance in HCC.

Alzheimer's disease (AD) is the major form of dementia in the elderly and is closely related to the toxic effects of microglia sustained activation. In AD, sustained microglial activation triggers impaired synaptic pruning, neuroinflammation, neurotoxicity, and cognitive deficits. Accumulating evidence has demonstrated that aberrant expression of deubiquitinating enzymes is associated with regulating microglia function. Here, we use RNA sequencing to identify a deubiquitinase YOD1 as a regulator of microglial function and AD pathology. Further study showed that YOD1 knockout significantly improved the migration, phagocytosis, and inflammatory response of microglia, thereby improving the cognitive impairment of AD model mice. Through LC-MS/MS analysis combined with Co-IP, we found that Myosin heavy chain 9 (MYH9), a key regulator maintaining microglia homeostasis, is an interacting protein of YOD1. Mechanistically, YOD1 binds to MYH9 and maintains its stability by removing the K48 ubiquitin chain from MYH9, thereby mediating the microglia polarization signaling pathway to mediate microglia homeostasis. Taken together, our study reveals a specific role of microglial YOD1 in mediating microglia homeostasis and AD pathology, which provides a potential strategy for targeting microglia to treat AD.

OBJECTIVES: To investigate the therapeutic mechanism of 2,6-dimethoxy-1,4-benzoquinone (DMQ) for alleviating dextran sulfate sodium (DSS)-induced ulcerative colitis (UC) in mice. METHODS: Eighteen male C57BL/6J mice were equally randomized into control group, DSS group and DMQ treatment group. In DSS and DMQ groups, the mice were treated with DSS in drinking water to induce UC, and received intraperitoneal injections of sterile PBS or DMQ (20 mg/kg) during modeling. The changes in body weight, disease activity index (DAI), colon length, spleen weight, and colon histological scores of the mice were examined, and the percentages of Th17 and IFN-γ⁺ CD8⁺ T cells in the mesenteric lymph nodes and spleen were analyzed using flow cytometry. The expressions of tight junction proteins (Occludin and ZO-1), proteins associated with inflammasome activation (caspase-1 and p20), IL-1β and TNF-α in the colon tissues were detected using Western blotting or ELISA. In the cell experiment, mouse bone marrow-derived macrophages (BMDMs) primed with lipopolysaccharide (LPS) were treated with DMQ, followed by stmulation with nigericin to activate the classical NLRP3 inflammasome pathway. In cultured human peripheral blood mononuclear cells (PBMCs) treated with either LPS alone or LPS plus nigericin, the effects of DMQ on inflammasome activation, pyroptosis, and cytokine release were evaluated via Western blotting, ELISA, and flow cytometry. RESULTS: In DSS-treated mice, DMQ treatment significantly alleviated DSS-induced body weight loss, colon shortening, spleen enlargement, and colon inflammation. The DMQ-treated mice showed significantly reduced percentages of Th17 cells and IFN-γ⁺ CD8⁺ T cells in the mesenteric lymph nodes and spleen, with increased occludin and ZO-1 expressions and decreased caspase-1 expression in the colon tissue. DMQ obviously inhibited classical NLRP3 inflammasome activation in mouse BMDMs and both the classical and alternative pathways of NLRP3 activation in human PBMCs, causing also suppression of caspase-1-dependent pyroptosis. CONCLUSIONS DMQ ameliorates DSS-induced UC in mice by inhibiting NLRP3 inflammasome activation.
Animals ; NLR Family, Pyrin Domain-Containing 3 Protein/metabolism* ; Mice, Inbred C57BL ; Colitis, Ulcerative/metabolism* ; Dextran Sulfate/adverse effects* ; Male ; Inflammasomes/metabolism* ; Mice ; Benzoquinones/therapeutic use* ; Th17 Cells ; Caspase 1/metabolism*

Approximately 300 million tons of plastic are produced globally each year,which has a serious impact on human health,marine life and the livestock industry.Microplastics have also been detected in meat and milk samples.Research has shown that nanoplastics(NP)(＜1 μm)and mi-croplastics(MP)(1 μm-5 mm)can affect the digestive,immune and reproductive systems of ani-mals.This experiment aims to investigate whether NP and MP regulate autophagy and damage the nervous system and spatial memory of animals.This experiment was divided into control group,nanoplastic group(PS-NP group,0.1 μm)and microplastic group(PS-MP group,1 μm),with 20 mice in each group.The mice were given 0.5 mL of PS-NP and PS-MP every day for 35 consecutive days,followed by neck amputation and brain analysis.The results showed that NPs and MPs of dif-ferent diameters caused varying degrees of damage to the brains of mice.In the behavioral tests of new object recognition,barnes maze and Y-shaped maze spatial memory,compared with the control group,the PS-NP group and PS-MP group showed a significant decrease in spatial memory ability of mice.HE staining results showed that neuronal cells in the PS-NP and PS-MP groups of mice exhibited shrinkage,decreased cell volume and deepened staining.The number of Nissl bodies de-creased,leading to dissolution and disappearance.RT-PCR and Western blot results showed that compared with the control group,the expression of glutamate receptors NR1,NR2A and NR2B in-creased in mice administered NP and MP orally,while the expression of autophagy related proteins Parkin,LC3B and Beclin1 was inhibited.In summary,this study suggests that nanoplastics and mi-croplastics stimulate glutamate receptors in mice by inhibiting the autophagy pathway,leading to impaired spatial memory.

Purpose To develop an integrate model combining deep learning features from contrast-enhanced lung ultrasonography with clinical characteristics for predicting epidermal growth factor receptor mutation status in peripheral non-small cell lung cancer.Materials and Methods This retrospective study included 117 patients with pathologically confirmed non-small cell lung cancer from the First Affiliated Hospital of Guangxi Medical University(July 2021 to February 2024).Patients were randomly divided into training(n=93)and test(n=24)sets at an 8∶2 ratio.Regions of interest were delineated at the peak enhancement phase of contrast-enhanced lung ultrasonography.Various deep learning convolutional neural networks were pretrained,with ResNet18 selected as optimal for feature extraction.Deep learning,clinical,and integrated models were constructed using naive Bayesian algorithm.Performance was evaluated via receiver operating characteristic and calibration curves,while class activation mapping and Shapley additive explanation values provided model interpretability.Results In the training set,the deep learning,clinical and integrated models achieved area under the curve of 0.93(95%CI 0.88-0.98),0.86(95%CI 0.68-1.00),and 0.91(95%CI 0.85-0.97),respectively.Corresponding test set area under the curve were 0.81(95%CI 0.72-0.90),0.56(95%CI 0.33-0.80),and 0.87(95%CI 0.72-1.00).Both deep learning and integrated models significantly outperformed the clinical model in training(Z=2.380,P=0.017;Z=2.597,P=0.009)and test sets(Z=2.034,P=0.042;Z=2.577,P=0.010).The integrated model demonstrated excellent calibration and predictive performance.Conclusion The integrated model combining deep learning features from contrast-enhanced lung ultrasonography with clinical characteristics effectively predicts epidermal growth factor receptor mutation status in peripheral non-small cell lung cancer.

Objective To explore the targeted regulatory relationship of miR-330-5p on OY-TES-1 in glioblastoma and the effect of miR-330-5p/OY-TES-1 axis on the migration ability of glioblastoma.Methods Bioinformatics analysis was performed to analyze the expression level of miR-330-5p in patients with glioblastoma and its influence on prognosis and survival of patients.The glioblastoma cells U251 were divided into miR-330-5p minics group,minics-NC group,and miR-330-5p+OY-TES-1 overexpression group(miR-330-5p minics+pcDNA3.1-OY-TES-1).The effect of miR-330-5p on the activity of OY-TES-1 3'UTR region was detected by double luciferase reporter gene experiment.The expression of OY-TES-1 mRNA was detected by qRT-PCR.The effect of miR-330-5p/OY-TES-1 axis on the migration ability of glioblastoma cells was detected by Transwell migration assay.Results The expression of miR-330-5p in glioblastoma tissue was significantly lower than those in non-tumor brain tissue and low-grade glioma tissue(P<0.05).The survival time of glioblastoma patients with high expression of miR-330-5p was significantly longer than that of patients with low expression of miR-330-5p(P<0.05).After overexpression of miR-330-5p,the activity of OY-TES-1 3'UTR region was decreased(P<0.05).Compared with minics-NC group,the expression levels of OY-TES-1 mRNA of U251 and U87MG cells in miR-330-5p minics group were significantly decreased(P<0.01).Compared with minics-NC group,the numbers of migrating cells in miR-330-5p minics group and miR-330-5p+OY-TES-1 overexpression group were significantly decreased(P<0.05).Compared with miR-330-5p minics group,the number of migrating cells in miR-330-5p+OY-TES-1 overexpression group was significantly increased(P<0.01).Conclusion MiR-330-5p targets OY-TES-1 to inhibit the migration of glioblastoma.