共找到 20 条结果
暂无摘要（点击查看详情）
To compare transfer accuracy between a three-dimensionally (3D) printed transfer jig and a conventional silicone tray in an in vitro setting using 3D linear measurements. 3D intraoral scans were obtained from four patients. Sixteen 3D-printed models (eight maxillary, eight mandibular) were prepared for each group. In Group 1, brackets were digitally positioned using 3Shape Ortho Analyzer software and transferred from working to posttransfer models using eight 3D-printed jigs. In Group 2, eight silicone trays were fabricated for bracket transfer. Working and posttransfer models were superimposed using Geomagic software, and divergence was summarized as the median and interquartile range. Tooth-level deviations between methods were compared using the Wilcoxon signed-rank test, stratified by arch and tooth type. Statistical significance was set at P < .05. Although a statistically significant difference between methods was observed (P = .024), median deviations for both methods were within the American Board of Orthodontics clinically acceptable limit (≤0.5 mm). In the mandible, the 3D-printed jig group showed smaller positional deviations than the conventional method (P < .001). Conversely, in the maxilla, the conventional method showed smaller positional deviations (P = .003). Deviations occurred most often in the vertical and horizontal directions in the 3D-printed jig group and in the transverse and vertical directions in the conventional group. The 3D-printed jig demonstrated greater accuracy in the mandible, while the conventional silicone tray was more accurate in the maxilla. However, the differences were not clinically significant, and both techniques achieved clinically acceptable bracket transfer accuracy.
To investigate the effects of turbulence intensity on wake velocity evolution in groove-flap vertical-axis wind turbines (VAWTs) during operation, this study proposes the Dynamic Mode Decomposition (DMD) method for multi-modal visualization of wake flow fields in both baseline and groove-flap VAWTs. The spatiotemporal wake characteristics were analyzed under turbulence intensities of 0.1%, 8%, and 15%, and application scenarios for both VAWT configurations were recommended. The results reveal that spatiotemporal variations in wake velocity become increasingly unstable in both axial and normal directions with rising turbulence intensity. The baseline VAWT demonstrates simpler and more stable wake structures with periodically arranged vortices, making it suitable for low-turbulence environments such as suburban areas. In contrast, the groove-flap VAWT enhances wake mixing through localized perturbations, showing superior adaptability to urban deployment scenarios including street canyons and skyscraper rooftops.
Evaluating large language models (LLMs) in the mental health domain presents distinct challenges due to the subtle, context-dependent, and subjective nature of psychological symptoms. We introduce PsyEval, a benchmark specifically designed to evaluate LLMs in mental health-related tasks across three core dimensions: knowledge, diagnosis, and emotional support. PsyEval is constructed to reflect the complexity of mental health scenarios and provides a structured framework for assessing model performance within this sensitive domain. Using PsyEval, we evaluate eleven advanced LLMs with different prompting strategies to investigate how prompting affects their responses. The results reveal considerable gaps in LLMs' current ability to reason accurately and respond appropriately in mental health contexts, while also indicating promising directions for future model enhancement.
Frequently evolving into a highly immunosuppressive niche that facilitates tumor growth and metastasis, the tumor microenvironment (TME) serves as the primary site for interactions between tumor cells and the host immune system. Reversing this immunosuppressive state therefore represents a pivotal strategy for antitumor therapy. Natural products possess the distinct advantages of multi-targeted and multi-pathway mechanisms, as well as the capacity to enhance therapeutic efficacy while attenuating systemic toxicity. These characteristics perfectly align with the core principles of antitumor immunotherapy. Accumulating evidence demonstrates that natural products including Ganoderma lucidum polysaccharides, Astragalus polysaccharides, and Lycium barbarum polysaccharides, as well as bioactive single compounds such as curcumin, resveratrol, and bufalin, exert potent antitumor immunotherapeutic effects by remodeling the suppressive TME. Consequently, they are increasingly being utilized as synergists or adjuvants in advanced therapeutic modalities, which include adoptive immune cell therapy, cancer vaccines, and immune checkpoint inhibitors. In addition, the discovery of novel potential regulatory factors for tumor immunotherapy from natural products has attracted significant attention. This review provides a comprehensive overview of the latest progress on how natural products reshape the immunosuppressive landscape and enhance the systemic antitumor response. Ultimately, this review both underscores the immense translational potential of natural products in tumor immunotherapy and highlights critical directions for future standardized research.
Prostate cancer (PC) is the fourth most common cancer worldwide and is recognised as one of the major men's health challenges of the twenty-first century. There has been a steady increase in PC incidence, particularly for advanced-stage PC cases, which have increased by 5% since 2014. Despite advances in diagnostic and treatment techniques, understanding of the genetic mechanisms underlying PC remains limited, especially in families with multiple cases of the disease, where genetic predispositions significantly increase the risk of developing cancer. This chapter provides a comprehensive and up-to-date review of PC, covering key aspects such as its epidemiology, aetiology and associated risk factors. From the hereditary point of view, this chapter covers the main genes associated with the susceptibility to develop PC and related pathologies. Additionally, the chapter evaluates the current paradigm of detection and treatment protocols, while also exploring future research directions.
Pedestrian safety remains a major public health concern in urban environments, particularly in low- and middle-income countries. While individual risk behaviours have been widely studied, less is known about how the built environment influences their accumulation during real-world crossing events. To analyse the association between roadway characteristics and the accumulation of pedestrian risk behaviours at urban intersections in Guadalajara, Mexico. An observational cross-sectional study was conducted based on direct observation of pedestrian crossings at multiple urban intersections between March and June 2024. The unit of analysis was the individual crossing event. Seven risk behaviours including crossing during vehicular green signals, crossing outside the crosswalk and diagonal crossing were used to construct a composite risk index (range 0-7). Associations were evaluated using multivariable Poisson regression models with robust standard errors clustered at the intersection level. A total of 1018 pedestrian crossings were analysed. Most pedestrians exhibited none (39.8%) or one (45.3%) risk behaviour while 14.0% accumulated two or more. The most frequent behaviours were crossing with vehicular green (28.1%), crossing outside the crosswalk (19.6%) and diagonal crossing (19.2%). In multivariable models, a higher number of traffic directions (incidence rate ratio 1.50, 95% CI 1.30 to 1.80), curb presence (incidence rate ratio 1.30, 95% CI 1.10 to 1.50) and unsignalised intersections (incidence rate ratio 1.30, 95% CI 1.10 to 1.5) were associated with greater accumulation of risk behaviours. In contrast, arterial and higher-level roadway types were associated with lower risk accumulation (incidence rate ratio 0.84, 95% CI 0.79 to 0.90). Pedestrian risk behaviours tend to accumulate and vary according to roadway characteristics. These findings underscore the role of the built environment in shaping pedestrian behaviour and support infrastructure-based interventions to improve urban safety.
Genome-wide association studies (GWAS) of asthma have identified nearly 200 independent loci, yet the mechanisms through which individual loci influence asthma risk remain largely unknown. A growing array of computational and experimental tools has begun to fill these gaps by identifying causal variants and effector genes and characterizing their functions. In parallel, emerging studies are exploring the translational applications of genetic and multi-omics data in asthma, including defining molecular endotypes and predicting disease risk. In this manuscript, we review the strengths and limitations of current approaches for addressing the post-GWAS challenges and discuss the next tier of questions and directions for the field.
Artificial intelligence (AI) is rapidly revolutionizing biomedical research. Empowered by enhanced object recognition and modern machine learning protocols, AI tools detect subtle patterns in human and animal behavior, efficiently quantifying and classifying them using supervised and unsupervised learning approaches. Complementing rodent studies, the zebrafish (Danio rerio) represents a crucial model organism in neuroscience research with well-characterized quantifiable behaviors, high-throughput potential and high genetic, neurochemical, and neuroanatomical homology to humans. The integration of AI strategies into zebrafish neuroscience research enhances behavioral endpoint monitoring, efficient processing and interpretation of data, establishing high-throughput screens and finding critical connections among behavioral and molecular endpoints. Here, we discuss the current status of the application of AI methods in zebrafish neurobehavioral research, as well as its limitations, future research directions, and remaining open questions in the field, with a particular focus on the use of AI in behavioral analyses and neuroactive drug discovery.
A 14-year-old girl (the victim) was reported missing. Call-data records were obtained for the mobile telephone of the victim and for the mobile telephone of the suspected abductor. The telephones were registered with different networks. Each telephone connected to a pair of antennas that were approximately 640 m apart. The antennas of the different networks were adjacent to one another. Survey data mapping the service areas (the cells) for the four antennas were not available. We assisted the investigation by, without using survey data, estimating the area within which the two telephones were most likely to be located. We estimated areas for each telephone separately and an area assuming that the telephones were co-located. The latter area was approximately a rectangle about 125 m long and less than 50 m wide. The present paper describes the methods that we used. Those methods made use of the locations of the antennas, the point-directions (azimuths) of the antennas, and timing-advance information. We discuss limitations of the methods used and discuss desirable properties for improved methods.
Clustered regularly interspaced short palindromic repeats (CRISPR) systems have revolutionized microbial genome editing. However, their reliance on DNA double-strand breaks (DSBs) and homologous recombination limits applications such as large DNA integration and engineering nonmodel microorganisms. CRISPR-associated transposase (CAST) systems provide a promising alternative, enabling DSB-free, recombination-independent, and programmable integration of large DNA fragments. This review summarizes the discovery and characterization of diverse CAST systems, the engineering strategies to enhance integration efficiency and specificity, and their applications in microbial engineering. Current limitations and future directions are also discussed. Overall, CAST systems hold great potential for advancing microbial genome editing, particularly in multi-copy DNA integration and genetic reprogramming of nonmodel microorganisms and complex microbial communities.
Redox-responsive lipid nanoparticles (LNPs) are emerging as a powerful platform for precision nanomedicine by exploiting disease-associated redox imbalances, such as elevated glutathione and reactive oxygen species, to trigger controlled cargo release and structural activation. This strategy is of great importance for developing gene-based therapeutics, where efficient cytosolic delivery is essential. In addition to nucleic acids, redox-responsive LNPs have been also explored in delivering small molecules, proteins, and theranostic agents, broadening their potential in both cancer and non-cancer diseases. This review summarizes the biological basis of redox responsiveness, key design principles for responsive chemical structures, and major advances in payload delivery and targeting capability of LNP carriers. Moreover, some major bottlenecks, including redox heterogeneity, stability, responsiveness trade-offs, and translational complexity, are critically discussed. Future directions are also highlighted, particularly for organ-selective delivery, multifunctional theranostics, and clinically translatable LNP-based medicines.
Automated segmental volumetry is effective for aortic dissection surveillance but lacks standardization. This study aimed to identify the optimal deep learning workflow by comparing a direct "end-to-end" approach with a "hybrid" strategy integrating vertebral-based partitioning. We analyzed 112 computed tomography scans (57 type A, 55 type B). The aorta was subdivided into five segments using vertebral landmarks. We evaluated two strategies across three architectures (nnU-Net, SwinUNETR, and U-Mamba): (1) end-to-end (simultaneous multi-class segmentation) and (2) hybrid (true and false-lumen segmentation followed by vertebral-based partitioning). Performance was assessed using the dice similarity coefficient, 95% Hausdorff distance, and relative volume error. The hybrid strategy improved segmental accuracy over the end-to-end approach across architectures. Using nnU-Net, segment-averaged true lumen dice similarity coefficient improved from 0.877 to 0.945 (p < 0.001), and 95% Hausdorff distance decreased from 16.87 mm to 4.41 mm. False-lumen accuracy also improved but was lower than the true lumen and declined in the distal segments. The hybrid strategy also reduced false-positive false-lumen detections in negative cases (nnU-Net, from 50.0% to 29.2%). Bland‒Altman analysis showed low bias but wide limits of agreement (approximately ± 40%), narrower for the hybrid strategy in the false lumen. For automated segmental volumetry in aortic dissection, a hybrid workflow-combining lumen segmentation with vertebral-based partitioning-improved segmental boundary accuracy over an end-to-end approach and provided a standardized, reproducible partitioning across same-patient scans. Further improvement of the underlying segmentation will be needed before routine clinical surveillance, toward which this workflow offers a reproducible foundation.
Standard k-mer methods treat amino acids as categorical tokens without directly encoding physicochemical properties. Although physicochemical properties have been incorporated into various bioinformatics tasks, their potential as a direct, systematic alternative for the k-mer counting paradigm has not been fully evaluated. We present PhysioChem-K-mer, a framework that transforms protein sequences into physicochemical property-based feature spaces, serving as an alternative to conventional amino-acid-identity k-mer representations. Our main hypothesis is that property-based representations capture functional constraints more effectively than traditional amino acid-based methods. To test this hypothesis, we created a controlled benchmark comprising 1500 synthetic sequences spanning 10 diverse protein families. The dataset retained core functional motifs while deliberately excluding evolutionary patterns typically found in natural biological sequences. Notably, our hydropathy-based PhysioChem-K-mer achieved a classification accuracy of 81.33% on a controlled synthetic benchmark, representing an absolute gain of 44.33% points over standard 3-mer methods (37.00%). The framework was further evaluated using real UniProt/Swiss-Prot data, comprising 11,620 sequences across 10 families, to ensure practical generalizability. Based on real data, PhysioChem-Hydropathy achieved 64.63%, an absolute gain of 47.68% points over the standard 3-mer baseline (16.95%), while reducing features by 73.9% and training time by 81.6%. By directly integrating biochemical knowledge into feature representations as a primary design principle, PhysioChem-K-mer combines interpretability with computational efficiency. These results suggest that physicochemical properties offer a vital source of information for protein classification, validated here on both synthetic and real-world data.
The immunosuppressive tumor microenvironment (TME) poses a significant challenge to effective cancer immunotherapy, as it enables tumor escape through redundant checkpoint pathways, metabolic constraints, and direct inhibition of effector cells. To overcome these barriers, we developed a next-generation NK cell platform using a tri-cistronic retroviral vector that enhances NK cell activation, recruitment, survival, and metabolic fitness. Pan-cancer transcriptomic analyses reveal consistent co-expression of PD-L1 and HLA-E across tumors, which correlates with immune infiltration accompanied by strong immunosuppression, highlighting these molecules as key targets for immune evasion. To improve NK cell recruitment, activation, cytolytic function, survival, and metabolic fitness in the TME, we developed a next-generation NK cell platform using a tri-cistronic retroviral vector that encodes an extracellular PD-1 domain (exPD1) fused to the intracellular portion of NKG2D with the costimulatory molecule 4-1BB, and expressing soluble IL15 and NKG2A single-chain variable fragments (scFv). Thus, the exPD1 allows the recognition of cells expressing PD-L1, while the intracellular costimulatory signaling transforms the inhibitory interaction PD-1/PD-L1 into an activating one. This strategy effectively targets PD-L1-positive tumor cells and induces de novo PD-L1 expression in otherwise negative tumors. To further enhance anti-tumor activity, we incorporated a module encoding soluble NKG2A-scFv to mask the NKG2A receptor on NK cells and hinder NKG2A/HLA-E inhibitory interaction. Additionally, we improved NK cell survival and expansion by delivering controlled low doses of IL15, which prevents NK cell exhaustion and extends their presence in vivo. This integrated strategy may provide a novel, ready-to-use allogenic mature NK cell therapy that could overcome checkpoint-mediated inhibition, metabolic suppression, and immune escape, offering a promising model for treating high-risk and treatment-resistant tumors.
Otitis media is a major cause of hearing loss, particularly in children. However, nonspecific symptoms and subjective evaluations make its diagnosis challenging. To address this, we developed transformer-based models to classify tympanic membrane conditions from otoscopic images. This approach aims to enhance diagnostic transparency and reliability in clinical settings. We trained vision transformer (ViT) and Data-efficient Image Transformer (DeiT) models on 454 pediatric and adult otoscopic images. These models performed multi-class classification to distinguish between normal, effusion, and tube conditions. For explainability, we utilized Gradient-weighted Class Activation Map (Grad-CAM), Layer-wise Relevance Propagation (LRP), and Attention Rollout (AR). Furthermore, we introduced a hybrid fusion strategy based on Canonical Correlation Analysis. The framework's effectiveness was then evaluated using insertion and deletion causal metrics. The ViT model achieved an accuracy of 97.78% (AUC: 0.998), outperforming DeiT, which reached 93.33% (AUC: 0.994). Notably, ViT attained an F1-score of 97.30% for the effusion class. Among the Explainable Artificial Intelligence (AI) methods, the hybrid LRP and AR approach provided the highest explainability. It yielded an average deletion score of 0.3008 and an insertion score of 0.8918, precisely highlighting critical image features for model predictions. In conclusion, integrating transformer-based models with hybrid explainability methods significantly enhances diagnostic transparency. These advancements foster clinician trust and lay a strong foundation for reliable clinical decision support systems.
Estimating the population-level impact of gambling harm has been historically challenged by inconsistent methods for counting affected individuals and quantifying their health loss. This study integrates two recent methodological advances, a novel reconciliation framework (Browne et al., 2026), and synthesised health disutility weights (Tulloch et al., 2026), and applies them to the 2025 Australian population. The model estimates that approximately 3.29 million Australians aged 5 years and over are affected annually by their own or another person's gambling, including approximately 1.38 million adult affected others and a reference-case estimate of a further 280,000 affected children. The model also indicates a concentration of harm among individuals experiencing dual exposure from both their own and another person's gambling; this group carries 25.9% of the total adult health burden despite representing only 11.2% of affected adults. The annual health loss associated with gambling-related harm in Australian adults is estimated at over 400,000 Years Lived with Disability (YLDs). These estimates should be interpreted as model-derived approximations conditional on the input assumptions and data sources. Requiring only the adult population size as input, the framework generates this full suite of estimates in a single transparent workflow, enabling direct replication by researchers or policymakers seeking a model-based estimate of the scope of gambling harm in the community.
Industrial digital radiography images are degraded by scattering that obscures defects, and conventional enhancement methods often sacrifice either noise suppression or detail preservation. This work presents a physics-inspired, annotation-free framework that explicitly separates the scattering component from the detected intensity through a radiation-matter interaction model. A multistage pipeline consisting of attenuation estimation, spatially adaptive scatter modeling, residual scatter removal, edge sharpening, and local contrast enhancement recovers the direct transmission signal. Evaluated on 60 industrial weld radiographs (ship plates, boilers, oil pipelines), the proposed method performs favorably against global histogram equalization, contrast-limited adaptive histogram equalization, a discrete wavelet transform approach, and a generic convolutional neural network baseline (serving as a lower-bound reference). It achieves the highest average contrast-to-noise ratio among the compared methods (up to 2.94), the lowest naturalness image quality evaluator scores (down to 4.67), and the lowest blind/referenceless image spatial quality evaluator scores (down to 22.37). Ablation studies verify incremental contributions of each stage, while sensitivity analyses on an independent test set and cross-condition validation confirm stable performance under parameter perturbations and across weld categories. The empirically calibrated coefficients are not derived from first principles, and cross-dataset generalization remains a necessary future investigation. The proposed framework provides a new detail-enhancement solution that significantly improves defect visibility and structural fidelity for industrial nondestructive evaluation using X-ray imaging. The C# source code developed for this study is freely available at https://doi.org/10.5281/zenodo.20711760.
Exosomal miRNAs mediate intracellular communication between the tumor microenvironment and cancer cells in non-small cell lung cancer (NSCLC). However, effects of exosomal miRNAs on NSCLC and its mechanisms have not been completely clarified. Here, exosome miRNA profiling and patient serum analysis revealed that the expression level of exosomal miR-664b-5p is significantly elevated in NSCLC, with particularly high levels in cancer-associated fibroblasts (CAFs) and their secreted exosomes. Functionally, miR-664b-5p promotes cell proliferation, migration, and invasion and inhibits apoptosis, thereby promoting malignant progression and metastasis in NSCLC. Moreover, experiments in a CAF-organoid coculture system and a CAF-NSCLC cell coinjection animal model demonstrated that exosomal miR-664b-5p is derived primarily from CAFs and is transferred to NSCLC cells via exosomes, contributing to the malignant phenotype of NSCLC cells, whereas miR-664b-5p knockdown in CAFs attenuated their tumor-promoting ability. Further exploration revealed that G protein γ subunit 11 (GNG11) is a direct functional target gene of miR-664b-5p. GNG11 expression is negatively correlated with miR-664b-5p expression, and ectopic expression of GNG11 partially abrogates the malignant phenotypes induced by miR-664b-5p overexpression in NSCLC. Mechanistically, CAF-derived exosomal miR-664b-5p facilitates NSCLC progression and metastasis by downregulating GNG11, which correlated with activation of the CXCL12/CXCR4 chemokine pathway. Serum exosomal miR-664b-5p levels were positively correlated with tumor burden. Taken together, the results of our study reveal that exosomal miR-664b-5p derived from CAFs can be transferred to NSCLC cells to promote NSCLC progression and metastasis via the CXCL12/CXCR4 axis, supporting the potential of exosomal miR-664b-5p as a biomarker and therapeutic target for NSCLC.
To quantify gendered patterns of professional misidentification among hospital doctors. An anonymised cross-sectional survey of hospital doctors was conducted in February 2025 at Basingstoke and North Hampshire Hospital, a UK district general hospital. Respondents reported the frequency, source and type of misidentification, alongside perceived impact on workflow. Responses were analysed by gender using Fisher's exact test. 52 doctors responded (32 female, 20 male). All female respondents reported misidentification at least once (vs 75.0% of men, p=0.006) and were more often misidentified weekly to daily (81.2% vs 25.0%, p<0.001). All women reported being mistaken for nurses (vs 30.0%, p<0.001) and misidentified downward (vs 50.0%, p<0.001); men were more often misidentified upward (65.0% vs 25.0%, p=0.008). Elevating nicknames were more often directed at men (60.0% vs 18.8%, p=0.003). Misidentification disrupted workflow at least sometimes for 40.6% of women versus 5.0% of men (p=0.008). In this single-centre cohort, misidentification showed consistent, statistically significant gendered patterns-women downward, men upward-suggesting a broader pattern warranting investigation in larger, multicentre studies. Visible role-labelling, inclusive communication training and leadership modelling merit consideration and future evaluation.