As a class of preconditioners, sparse approximate inverses (SAIs) have been proven to be effective in accelerating the convergence of iterative methods. However, given the wide variety of available preconditioners, determining how to select the most suitable one for a specific iterative solver remains a central challenge in scientific computing. In the past decade, the rapid development and widespread adoption of convolutional neural networks (CNNs) have inspired new perspectives for tackling this challenge. When applying CNNs to the parallel construction of SAIs, three key issues must be addressed: normalized representations of sparse matrices, network architectures that accommodate the unique characteristics of sparse data, and appropriate training datasets of sparse matrices tailored for SAI learning. To address these challenges, we propose an innovative CNN-based intelligent selection framework for SAIs. In this framework, three normalization methods for sparse matrices are first introduced, namely binary, density and column-histogram representation. Second, for each representation, a corresponding CNN architecture is specifically designed, and they are named BinaryNet, DensityNet and HistNet, respectively. Third, these three networks operate independently and adopt a late fusion strategy to integrate learned SAI features. This design effectively mitigates the interference that would arise from early fusion, given the heterogeneous nature of the different input representations. Finally, we present, for the first time, a method for constructing a training dataset of sparse matrices specifically designed for SAI learning. Experimental results demonstrate that the proposed framework is both effective and efficient, validating the feasibility and advantages of the approach.
To evaluate gynecologic cancer survivors' preferences regarding the organization of sexual health counseling, with a focus on timing, frequency, responsible healthcare professionals, and preferred counseling content and formats. This cross-sectional, questionnaire-based online study included cancer survivors aged 18 years or older, with Eastern Cooperative Oncology Group performance status of 2 or lower, who had completed treatment for ovarian, endometrial, or cervical neoplasia and were recruited nationwide in Poland. Participants were recruited through online patient forums, social media, and patient advocacy platforms; therefore, the sample was self-selected and may not be representative of all gynecologic cancer survivors. The survey addressed sexual activity status, sexual counseling, and preferences for when and by whom sexual health should be discussed. Sexual activity was defined as any genital stimulation with or without a partner for sexual pleasure. Comparative analyses were conducted between survivors who returned to sexual activity and those who did not. The final sample comprised 201 survivors. Sexual counseling was reported as having been offered to 15 of 201 respondents (7.4%), and 7 of 201 (3.4%) had visited a sexual medicine specialist. Among respondents who completed the sexual health counseling preference module and post-treatment sexual activity status (n = 156), 42 (26.9%) did not return to sexual activity. Most respondents indicated that sexual health should be addressed at diagnosis and reassessed during follow-up (55.1%, n = 87/156). Overall, 62.3% (n = 98/156) expressed a need to discuss sexual health at every follow-up visit, with a substantial preference for counseling delivered by gynecologic oncologists or gynecologists (63.4%, n = 99/156). Survivors most requested practical guidance on resuming sexual activity and managing treatment-related sexual problems (93.6%, n = 146/156), and most expressed interest in receiving written and/or digital educational materials. In this online cohort of gynecologic cancer survivors, sexual health counseling was infrequently reported, while many respondents expressed a preference for structured, repeated discussions from diagnosis through survivorship. Implementation of a structured, patient-centered counseling pathway may improve survivorship care quality and access. ClinicalTrials.gov NCT06958068.
Patients now arrive at the orthodontic consultation having already consulted a chatbot, yet the accuracy, quality and readability of these tools lack quantitative synthesis. Six databases and one trial register were searched on 24 April 2026 for LLM-based chatbots evaluated on patient-facing orthodontic content. RoB 2 was used for the RCT and an adapted JBI checklist for cross-sectional studies. Comparisons were pooled by instrument across accuracy, quality and readability with random-effects meta-analysis and the Hartung-Knapp-Sidik-Jonkman (HKSJ) small-sample correction; a ChatGPT-4-class primary analysis addressed mixed-generation pooling, and GRADE was applied to every pool. The review was single-reviewer with a pre-specified audit. Thirty-nine studies were included; risk of bias was low in 10, moderate in 10 and high in 19. ChatGPT-4-class and Gemini did not differ on Likert accuracy (MD -0.12 [-0.42 to +0.17]; I2=70.6%; p_HKSJ=0.533), and no accuracy or quality pool was significant under both the DerSimonian-Laird and HKSJ analyses. The largest pooled difference was in readability: across two studies, ChatGPT text was about 2.8 Flesch-Kincaid grades easier than Claude (MD -2.85 [-3.31 to -2.38]; p_DL<0.001), but only borderline under the conservative HKSJ (p_HKSJ=0.053) and therefore provisional. Five of 11 pools showed I2≥75%, and GRADE certainty was very low for every pool. No chatbot consistently dominates across accuracy, quality and readability; the largest difference, a provisional readability advantage of ChatGPT over Claude, rests on two studies. Chatbots cannot be relied upon as a safe complement to professional patient education and should never replace the information provided by the orthodontist, the more so because the same chatbot may not behave the same way within months.
To address the issue of missed detections caused by the similarity between weakly distinguishable objects and the background in both visible and infrared images, we utilize the correlation between local and global semantic features to find the difference clues between objects and the background. Specifically, we propose a novel Cross-Modality Fusion with Local and Global Semantic Feature Correlation Object Detection Network (CFLGNet), which effectively enhances the distinguishability. Firstly, we design a Semantic-Attention Mamba Fusion Block called SAMFB, which maps the cross-modal features into a hidden state space for interaction, and combines local and global semantic feature correlation to enhance the distinguishing features between the object and the background. SAMFB contains two branches: The Local Semantic Feature Correlation Module (LSCM) obtains an effective representation of local fine semantic feature through bilinear matrix computation, and this feature representation is embedded into another Global Semantic Feature Correlation Module with mamba (GSCM). GSCM captures the irregular spatial correlation through the scanning strategy of visual mamba, and then the local semantic feature extracted by the LSCM module are used as weights to suppress the spatially adjacent features with different representations. In addition, we propose a Lcal Feature Enhancement module based on Topological Invariants(LFETI) to reduce the problem of missed detection caused by incomplete object information, such as occlusion or truncation. Finally, the semantic feature correlation is combined with the original feature to enhance the distinguishing feature between the object and the background. Extensive experiments on the VEDAI and DroneVehicle benchmark datasets demonstrate that the proposed CFLGNet exhibits remarkable performance.
High-dimensional neuroimaging analyses for clinical diagnosis are often constrained by compromises in spatiotemporal fidelity and the limited adaptability of large-scale, general-purpose models. To address these challenges, we introduce Dynamic Curriculum Learning for Spatiotemporal Encoding (DCL-SE), an end-to-end framework centered on data-driven spatiotemporal encoding (DaSE). We leverage Approximate Rank Pooling (ARP) to efficiently encode three-dimensional volumetric brain data into information-rich, two-dimensional dynamic representations, and then decode this representation through a DGM-implemented architectural curriculum, where the fixed hierarchical decoder progressively refines features from global anatomical structures to fine pathological details without introducing a threshold-controlled stage-switch scheduler. Evaluated across six publicly available datasets, including Alzheimer's disease and brain tumor classification, cerebral artery segmentation, and brain age prediction, DCL-SE consistently outperforms existing methods in accuracy, robustness, and interpretability. These findings underscore the critical importance of compact, task-specific architectures in the era of large-scale pretrained networks.
Facial sketch synthesis is important for cross-modal face analysis and digital forensics, yet existing models often suffer from structural distortions and identity inconsistency under limited paired training data. Traditional methods, primarily based on generative adversarial networks, often suffer from training instability and insufficient detail reconstruction. To address these limitations, this study proposes a diffusion-based framework with a stage-wise multi-condition guidance mechanism that enhances both structural and textural fidelity. Specifically, (i) during the downsampling phase, semantic segmentation features are fused to guide the model with accurate structural information, such as facial part locations; and (ii) during upsampling, a hybrid cross-attention mechanism is employed to integrate coarse image textures with denoised noise, refining fine-grained details. Additionally, we incorporate the Vision Transformer within the U-Net backbone to better capture global contextual information in low-frequency regions, further enhancing image realism. Experiments on multiple benchmark datasets demonstrate strong overall performance in SSIM and FSIM, indicating improved structural similarity and feature-level fidelity under the evaluated settings. Moreover, our method obtains competitive LPIPS results, indicating improved perceptual similarity. We further compare with recent image-to-image translation and diffusion-based baselines and observe competitive performance in both visual coherence and identity preservation.
First-degree relatives (FDRs) of type 1 diabetes (T1D) patients are at high risk for T1D. Early screening enables timely intervention, yet participation remains suboptimal, missing opportunities in this high-risk group. Understanding determinants of screening participation and translating them into actionable, theory-informed, family-centered strategies is essential to optimize preventive care. In this qualitative study in China, semi-structured interviews were conducted with 21 adult T1D probands or parents of pediatric T1D patients who have unaffected children. The Theoretical Domains Framework (TDF) guided data collection and analysis. Barriers were mapped to Behavior Change Techniques (BCTs). We identified 18 facilitators, 11 barriers, and 2 neutral factors, mapped across 12 TDF domains. Key barriers included limited awareness, misconceptions, cost, restricted medical resources, insufficient publicity, and anxiety; key facilitators included perceived benefits, optimism, health beliefs, and social support. Through TDF-BCT mapping, 34 intervention strategies were categorized into 13 categories targeting knowledge enhancement, environmental adaptation, social and emotional support, and self-efficacy building. This study provides a systematic, theory-based analysis of factors influencing screening participation for T1D. The resulting strategy map offers a set of theory-informed, family-centered suggestions to guide nurses and educators in supporting parents and improving early detection in clinical practice, subject to future empirical validation. Nurses and educators should develop localized, family-centered screening promotion programs guided by theory-driven frameworks and BCTs. Concurrently, systemic enablers-such as resource allocation, financial support, and routine follow-up guidance-must be established to comprehensively address barriers to participation in screening.
Decentralized water disinfection in resource-limited regions urgently requires low-cost, solar-driven technologies that avoid chemical dosing and complex infrastructure. Photocatalytic H2O2 generation offers a clean alternative, but practical deployment is hindered by the poor processability, difficult recovery of powdered photocatalysts and high cost. Here, we address these barriers by developing a benzyl-integrated frameworks (BIFs) synthesized exclusively from benzene via a one-step Friedel-Crafts reaction. Despite its minimalist, non-conjugated structure, BIF exhibits broad-spectrum visible-light absorption and efficient H2O2 production (∼3.8 mmol g-1 h-1) via an unconventional interlayer π-π coupling mechanism. To enable continuous-flow operation, we encapsulated hydrophobic BIF into an agar hydrogel using a DMSO-mediated solvent-exchange strategy. The resulting composite hydrogel not only prevents aggregation but also enhances H2O2 production by 1.3-fold through improved mass transfer. When packed into a fixed-bed column, this solar biocidal gel achieves rapid and complete inactivation of E. coli (104 CFU mL-1, typical of treated wastewater) at flow rates up to 52 mL min-1, corresponding to a biocidal efficiency of 1300 CFU min-1 mg-1. With a daily treatment capacity of ∼75 L per gram of BIF, this work provides a practical, scalable, and energy-autonomous platform for solar-driven water disinfection, demonstrating that high performance need not come at the cost of complexity.
Continual multi-view clustering (CMVC) aims to discover cluster structures from multi-view data that arrive sequentially and exhibit distribution shifts. However, existing incremental anchor-based frameworks for CMVC still have some problems: 1. Anchors are learned in a purely reconstruction-driven way, resulting in ambiguous cluster semantics and higher sensitivity to noise. 2. Existing methods overlook the role of consistent anchor-level semantics, which limits their ability to suppress semantic drift. 3. It is tough to align the anchor graphs, weakening temporal consistency and degrading clustering stability. To address these limitations, we propose a novel continual multi-view clustering method with pseudo-label guided consistency anchor learning (CMVC-PCAL). In the initialization stage, we build a unified anchor graph representation by decomposing each anchor graph into a shared consensus component and a view-specific inconsistency component, and by introducing cluster-wise pseudo labels to align anchors with latent clusters, yielding a clearer and more noise-robust consensus anchor graph. In the continual stage, for each new data batch, a lightweight incremental scheme updates anchors and anchor graphs: an alignment matrix aligns historical and current anchors, and a consistency regularization term aligns their consensus parts, while a pseudo-label guided term jointly updates anchor pseudo labels and cluster centers to continuously refine anchor-level semantics. Finally, an alternating optimization algorithm is developed to efficiently train the model, and experiments on public multi-view benchmark datasets verify its effectiveness and efficiency in continual multi-view clustering scenarios.
Biological barriers remain a primary limitation to effective cancer therapy, restricting drug delivery, distribution, and therapeutic efficacy across tumor sites. These barriers arise from a complex and dynamic interplay between tumor- and host-derived factors, including abnormal vasculature, dense extracellular matrices, immune clearance, and spatiotemporal heterogeneity within the tumor microenvironment. Increasing evidence suggests that these components operate as an integrated and adaptive network. Here, we conceptualize this network as "tumor systemic intelligence," which enables tumors to sense and respond to therapeutic and physiological perturbations by reinforcing delivery barriers and limiting drug access. Conventional nanomedicine strategies, often designed to overcome individual barriers, have shown limited success in addressing this coordinated system. To overcome these limitations, we propose "adaptive intelligence" as a design paradigm for next-generation nanocarriers. In this framework, nanocarriers are engineered as responsive systems capable of sensing microenvironmental cues and dynamically modulating their physicochemical properties to navigate multiple, evolving barriers during systemic circulation and tumor penetration. We discuss key design strategies for adaptive nanocarriers, including programmable control over size, shape, surface charge, and bioactive interfaces, enabling improved circulation stability, enhanced transvascular transport, deeper tumor penetration, and reduced immune recognition. We further examine how these systems can modulate or bypass critical barriers such as extracellular matrix density, abnormal perfusion, and cellular uptake limitations. Integration with external physical stimuli is also considered to further enhance barrier penetration and delivery efficiency. Despite promising advances, clinical translation remains limited by challenges including nanocarrier safety, immunogenicity, and the complexity of tumor-host interactions. This review highlights emerging opportunities and design principles for adaptive nanocarriers that effectively navigate biological barriers, with the goal of improving drug delivery and therapeutic outcomes in cancer treatment. Importantly, the concepts of systemic intelligence and adaptive intelligence are intended as conceptual frameworks to guide the rational design of next-generation nanomedicines rather than as formal biological classifications.
Multi-view clustering has garnered significant research interest due to its ability to integrate complementary information from diverse perspectives. Graph-based multi-view clustering has emerged as a prominent research area in recent years, owing to its superior performance in handling non-convex datasets. Despite the progress made, graph-based multi-view clustering methods still face three key challenges: (1) high computational complexity due to graph construction and spectral analysis; (2) insufficient attention to balanced cluster sizes, which are crucial for practical applications such as resource allocation and load balancing; (3) reliance on a separate post-processing step (e.g., k-means) to discretize the continuous spectral embedding, which introduces bias. To address these issues, we propose Fast Balanced Multi-View Clustering (FBMVC), a fast and balanced graph-based multi-view clustering framework. FBMVC employs anchor and label transmission strategies, significantly reducing computational costs while maintaining performance. The adaptive view-weighting scheme autonomously optimizes view-specific weights to achieve effective multi-view integration. The resulting problem implicitly ensures a balanced clustering distribution. In particular, FBMVC obtains the discrete cluster labels within its optimization framework, thereby avoiding the need for a separate post-processing stage. Experiments on six benchmark datasets demonstrate that FBMVC outperforms state-of-the-art methods in terms of clustering performance and computational efficiency.
The power-law relationship between large language model (LLM) performance and the scale of pre-training corpora has driven remarkable advances, while raising serious ethical and legal concerns. Pre-training corpora often contain sensitive personal information, copyrighted content, or benchmark test data, and the opacity of pre-training corpora further exacerbates these concerns. Consequently, detecting pre-training data has become a significant research challenge. Most existing detection methods rely on intermediate outputs of LLMs (activations, token probabilities, or model loss), which are inaccessible in commercial LLMs, where only final outputs are available. Only a few early-stage attempts rely solely on final outputs by measuring lexical or semantic similarity between reproduced and original texts at the sentence or token level. Yet, these attempts overlook the dynamic nature of reproducibility: reproducibility fluctuates differently across reproduction positions, contextual spans, and expressions, between seen and unseen texts. This oversight narrows the gap of reproduction similarity, ultimately degrading detection performance. To address this limitation, we propose RADS-PDD (reproducibility-aware dynamic similarity-based pre-training data detection), which shifts black-box pre-training data detection from static similarity comparison to reproducibility-aware dynamic similarity modeling. RADS-PDD incorporates three reproducibility-aware mechanisms, positional gain weight, continual gain weight, and triplet occurrence probability, to quantify varying reproducibility by dynamic similarity, amplifying the reproduction similarity gap between seen and unseen texts. Extensive experiments across two representative datasets, multiple data domains, and diverse LLMs demonstrate that RADS-PDD consistently outperforms detection methods that rely solely on final outputs and achieves performance comparable to detection methods that require intermediate outputs.
Chagas disease and leishmaniasis are neglected diseases caused by Trypanosoma cruzi and Leishmania sp., respectively, leading to morbidity and mortality worldwide and affecting more than 6 million people per year. Despite the availability of treatments, their low efficacy and several adverse effects encourage the design and development of new drugs. The hybridization of natural products with heterocyclic scaffolds has been a molecular editing strategy for obtaining new drugs with antiparasitic activity. In this project, a new series of semisynthetic 4-carboxylate-1,2,3-triazine 1-oxide derivatives was designed, synthesized, and evaluated in vitro and in vivo against two strains of T. cruzi and L. mexicana. Compound TS-14 with the geranyl ester group at the 4-position and the ethyl group at the 5-position on the 1,2,3-triazine 1-oxide ring had better trypanocidal activity against Ninoa (IC50= 1.75 µM) and A1 (IC50=1.16 µM) strains than the reference drugs benznidazole (IC50 = 30.3 µM and 39.8 µM, respectively) and nifurtimox (IC50= 7.09 µM and 19.3 µM, respectively). Additionally, TS-14 reduced parasitemia by 40.4% against T. cruzi trypomastigotes in an in vivo model. Compound TS-16, featuring a geranyl ester group at the 4-position and an n-pentyl group at the 5-position on the 1,2,3-triazine 1-oxide ring, exhibited better leishmanicidal activity against FCQEPS (IC50 = 12.97 µM) and M379 (IC50 = 2.11 µM) strains than the reference drug glucantime (IC50 = 125.23 µM and 133.96 µM, respectively). Additionally, TS-21 and TS-30 cause the highest parasitemia reduction (60% and 64.1%, respectively) against L. mexicana in an in vivo assay. Finally, compounds TS-14 and TS-32 showed an inhibitory effect uncompetitive on trypanothione reductase of T. cruzi (TcTryR) and the compound TS-32 on trypanothione synthetase of T. cruzi (TcTryS). This study represents an innovative case of the use of fragments of natural products hybridized with 1,2,3-triazine 1-oxide scaffold as a new option for the development of drugs against parasites such as T. cruzi and L. mexicana.
This study investigates the potential of [47Sc]Sc-NOTA-trastuzumab as a novel targeted radionuclide agent in HER2 + breast cancer. Targeted radionuclide therapy offers a promising approach for enhanced delivery and internalization of radiation, minimizing off-target effects and treating advanced metastatic HER2 + breast cancers. By examining the synergistic relationship between [47Sc]Sc with trastuzumab, assessing [47Sc]Sc-NOTA-trastuzumab's impact on immune function, and monitoring its passage of a compromised blood-brain barrier for targeting HER2 + breast-to-brain metastases, this study provides a path towards synergistic and precision radiotherapy in breast cancer metastasis. The stability, specificity and reactivity of [47Sc]Sc-NOTA-trastuzumab was characterized with high performance liquid chromatography, Lindmo, and cell binding assays. In vivo localization and therapeutic efficacy were assessed with SPECT imaging, biodistribution, and tumor monitoring in models of HER2 + breast cancer. Immunostaining against M1 macrophages was assessed to determine mechanisms of enhanced anti-tumoral immune activity. [47Sc]Sc-NOTA-trastuzumab demonstrates high specificity and serum stability in HER2 + cells. SPECT shows [47Sc]Sc-NOTA-trastuzumab has high specificity for HER2 + breast cancer cells in vivo in both the primary and metastatic setting. Longitudinally, tumors treated with [47Sc]Sc-NOTA-trastuzumab show therapeutic benefit and are comparable to those treated with 16x dose of trastuzumab alone, with no significant toxicity. Exploratory results show radiopharmaceutical uptake in a model of HER2 + brain metastasis. Increased M1-like macrophages were seen in HER2 + tumors relative to HER2- tumors following treatment with [47Sc]Sc-NOTA-trastuzumab. [47Sc]Sc-NOTA-trastuzumab is serum stable and highly specific to HER2 + breast cancer. Developing and characterizing novel targeted radionuclide therapeutics has the potential to significantly improve radionuclide delivery in HER2 + metastatic disease.
Porcine reproductive and respiratory syndrome virus (PRRSV) is one of the important pathogens influencing global swine industry. PRRSV infection promotes the secretion and production of soluble triggering receptor expressed on myeloid cells 2 (sTREM2) in plasma. Whether sTREM2 regulated PRRSV replication remains unknown. In this study, we expressed and purified sTREM2 exogenously, and found that sTREM2 restrained PRRSV replication in vitro. In terms of mechanism, sTREM2 could interact with PRRSV envelope protein GP2a, wherein S112/S116/T120 amino acid sites in sTREM2 were essential for their binding. The interaction between sTREM2 and PRRSV GP2a interfered PRRSV binding stage, thus inhibiting PRRSV replication. In addition, sTREM2 promoted the activation of NF-ĸB signaling pathway and elevated the expression of proinflammatory cytokines and type I interferons, ultimately suppressing PRRSV replication. Our study demonstrated that sTREM2 possessed strong inhibitory effects against PRRSV and had the potential value for clinical application.
This work presents a single‑bubble model for dual‑frequency acoustic cavitation in an O2 environment. Dual-frequency acoustic cavitation generates extremely high temperatures in liquids, thereby driving complex sonochemical reactions. However, a direct quantitative link between operating parameters and sonochemical product yields has not been fully established. In this study, we systematically quantify the relationships between the maximum bubble temperature (Tmax) and a range of key operating parameters, including acoustic field settings (power allocation ratio, phase difference, frequency combination, and excitation amplitude) and initial bubble conditions (initial water vapor mole fraction and initial bubble radius). Using Tmax as an intermediate variable, we then establish quantitative correlations between these parameters and the yields of chemical products (total chemical species, H2, H2O2, and OH). Numerical simulations are performed for multiple dual-frequency combinations (20 + 40, 20 + 140, 20 + 200, 40 + 140, 40 + 200, and 140 + 200 kHz) while systematically varying each parameter. The results reveal that the sensitivity of Tmax to the power allocation ratio and phase difference is strongest under low-frequency combinations and weakens under high total acoustic pressure amplitude. A "high-frequency-dominant region" is identified, in which the specific low-frequency value no longer affects Tmax. The initial water vapor mole fraction exerts a noticeable modulating effect on Tmax at low acoustic pressures, but this effect is progressively suppressed as the excitation amplitude increases, except for certain frequency combinations at higher acoustic pressures. Tmax (approximately 3000-10190 K) determines the product yields, which follow distinct patterns. The overall trend of sonochemical yields as a function of Tmax remains consistent across different initial bubble radii, although the specific transition temperatures for each product vary with the initial radius; the initial radius of 2 μm is chosen as a representative case for the subsequent detailed analysis. H2 exhibits a three-stage pattern: "low plateau, sharp jump, slow saturation", with a threshold temperature of 5818 K and the sharp increase occurring between 5818 K and 7260 K; H2O2 shows a unimodal distribution with a peak at 6210 K; OH displays a multi-peak oscillation pattern with a peak at 6354 K. At 7237 K, the growth rate of the total yield of chemical species decelerates, marking a transition from reaction control to dissipation control. Based on these findings, we propose model-dependent optimal temperature windows for target products: approximately 7260 K for H2, 6210 K for H2O2, and 6354 K for OH, with the understanding that these values are derived from the present single-bubble model under an O2 atmosphere and serve as theoretical references rather than universal experimental guidelines. This study provides a theoretical basis for optimizing operating parameters to understand the selectivity of dual-frequency ultrasonic cavitation technology in the context of green hydrogen production and pollutant degradation.
暂无摘要（点击查看详情）
Atherosclerosis, a chronic inflammatory arteriopathy, constitutes a principal contributor to cardiovascular events. Its pathogenesis commences with endothelial dysfunction elicited by pro-inflammatory mediators like TNF-α. The exact involvement of BIRC3 (cIAP2) in this process remains largely uncharacterized. This study investigated whether BIRC3 contributes to TNF-α-induced endothelial injury via NF-κB signaling. Human aortic endothelial cells (HAECs) exposed to TNF-α exhibited markedly upregulated BIRC3 expression, concomitantly with NF-κB pathway activation. Silencing BIRC3 suppressed IκBα phosphorylation and p65 nuclear translocation and attenuated endothelial dysfunction, as evidenced by reduced levels of interleukin-6, interleukin-1β, VCAM-1, reactive oxygen species, and malondialdehyde, along with decreased apoptosis and a favorable Bax/Bcl-2 ratio. Conversely, BIRC3 overexpression exacerbated these pathological changes. Importantly, the deleterious effects of BIRC3 overexpression were largely reversed by co-treatment with the NF-κB inhibitor BAY 11-7082, indicating that BIRC3 drives TNF-α-mediated endothelial dysfunction in a strictly NF-κB-dependent manner. Collectively, our results identify BIRC3 as a positive regulator of TNF-α-induced endothelial injury and suggest that targeting BIRC3 may offer a novel avenue for alleviating endothelial dysfunction and hindering atherosclerotic progression.
To investigate the efficacy, optimal timing, procedure and influencing factors of non-invasive correction for various congenital auricular deformities in newborns. Retrospective study of 68 congenital auricular deformities in newborns (117 ears), including morphological and structural defects (2024-2025). Patients were divided into four groups: simple treated with self-made support strips, and others including those helical rim deformities with cartilage adhesions, microtia group and other deformities group corrected with EarWell system. Statistical analysis was performed using SPSS 23.0. The average age at initial correction was 22.38 days. No complications occurred in the self-made strip group, while mild adverse reactions in other groups were relieved after symptomatic treatment. The overall excellent rate reached 80.34%, with higher rates in helical rim groups and lower in microtia group. The average correction duration was 6.9 weeks. Notably, the self-made strip group had a significantly shorter correction duration (P = 0.0087). Satisfactory correction outcome can be achieved with a simple helical support strip for mild helical rim abnormalities without skin or cartilage adhesion. The EarWell Infant Ear Correction System effectively addresses external ear abnormalities and minor malformations. Earlier treatment initiation correlates with a shorter correction duration, superior clinical outcomes, and fewer complications.
Chronic low-grade inflammation has been implicated in the pathophysiology of cognitive dysfunction and neurodegenerative processes. Diet represents a key modifiable determinant of inflammatory status. We aimed to evaluate the associations between dietary inflammatory index (DII) scores and the serum levels of interleukin-6 (IL-6), leptin, and adiponectin in older adults with mild cognitive impairment (MCI), dementia, and normal cognition in Yucatan, Mexico. A case-control study was conducted between March 2022 and December 2025. Cases were defined as adults aged ≥ 55 years with a previous diagnosis of MCI or dementia, and controls were defined as adults aged ≥ 55 years with no prior history or current complaints of cognitive impairment and with normal performance on cognitive assessments. Dietary intake was assessed with a validated food frequency questionnaire, and the DII was calculated using 24 food parameters. Cytokine levels were determined using sandwich enzyme-linked immunosorbent assay kits. Participants with MCI and dementia exhibited a diet with higher pro‑inflammatory potential, characterized by lower intake of dietary fiber, vitamins, unsaturated fats and β-carotenes, as well as significantly higher adiponectin levels than controls. DII scores were inversely associated with adiponectin levels in patients with MCI and dementia (β = - 48, 95% CI: - 75 to - 22, p < 0.01). A pro‑inflammatory diet was associated with lower adiponectin in MCI and dementia, despite paradoxically elevated adiponectin. This association may be attenuated by excess adiposity. Adiponectin emerges as a promising biomarker, warranting longitudinal studies.