Search Results (34,060)

Cuproptosis is a novel form of metabolism-associated cell death. Cervical cancer (CC) exhibits elevated serum copper levels and mitochondrial metabolic reprogramming, making cuproptosis-related genes (CRGs) potentially critical for prognosis prediction and therapeutic targeting. However, studies on CRGs in CC remain limited. This study aimed to construct prognostic and cancer staging models for CC using machine learning (ML) algorithms. Gene expression profiles of patients with CC were obtained from the TCGA and GEO databases. Five ML algorithms were employed to identify significant factors, including random forest (RF), support vector machine (SVM), Gaussian mixture model (GMM), Bayesian, and StepCox. A prognostic model was subsequently constructed using LASSO–Cox regression based on the selected genes. Concurrently, a cancer staging model was built using ML algorithms incorporating three distinct gene categories. Finally, qRT-PCR and Western blotting were conducted to validate the expression of signature genes at both the tissue and cellular levels. Additionally, CTD-based screening and in vitro functional assays were performed to evaluate the effects of DDP on CC cells. Through integrated bioinformatics and ML approaches, a prognostic model comprising nine CRGs was successfully established (GMM = 0.72). The derived risk score served as an independent prognostic indicator for CC (p < 0.001, 95% CI: 3.681 [1.785–7.591]). Calibration curves confirmed that the nomogram accurately predicted overall survival (OS) at 1, 3, and 5 years. Additionally, a cancer staging model was effectively constructed using the GMM algorithm (AUC = 0.74). DDP dose-dependently inhibited CC proliferation/migration and down-regulated CRG expression. In this study, we developed two different models—a cuproptosis-related prognostic model and a cancer staging model—that highlight promising biomarkers for predicting patient prognosis and cancer progression in patients with CC. Full article

Background: CKD remains substantially underdiagnosed and inadequately monitored. Objectives: To evaluate whether a laboratory-integrated CDSS enables opportunistic CKD detection and longitudinal surveillance in Primary Care via automated KDIGO stratification and progression detection. Methods: We implemented a CDSS embedded within the laboratory information system at Hospital Universitari i Politècnic La Fe, a tertiary university hospital in València, Spain. A one-time baseline profile screened patients with diabetes, hypertension, or age 60–80 years without documented advanced CKD, while a follow-up profile targeted patients with advanced CKD or those previously assessed by baseline screening. Both profiles applied KDIGO risk stratification and automated detection of renal and albuminuric progression based on predefined criteria, classifying each patient into one of three action categories: no CKD or low-risk requiring no action; CKD requiring Primary Care monitoring; or CKD meeting criteria for automated nephrology referral. Missed opportunities for albuminuria monitoring and economic impact of the intervention were evaluated. Results: During a 4-month period, the CDSS processed 10,733 requests, of which 10,414 were evaluable. No further action was required in 8661 patients (83.2%), monitoring was recommended in 1591 (15.3%), and nephrology referral was triggered in 162 (1.6%). Among 82 patients with albuminuric progression, 39% had not undergone ACR testing for more than two years despite ongoing Primary Care contact. Following implementation, ACR testing increased by 15.0% compared with the previous year, resulting in an additional analytical cost of EUR 1550.54. Conclusions: The CDSS enabled systematic CKD detection and surveillance in Primary Care, identifying silent renal and albuminuric progression and supporting nephrology referral at minimal cost. Full article

Post-earthquake fire, also referred to as fire following earthquake (FFE), can cause severe urban damage, particularly where detailed empirical fire-spread records are unavailable. This study derives a simulation-informed regional regression relationship between covering volume fraction (CVF) and fire-burned ratio for screening-level assessment in Pohang, South Korea, and integrates it into an existing static FFE risk-assessment framework. A total of 6604 dynamic fire-spread simulations were generated by varying the composition and assignment of four fire-resistance categories within the fixed GIS geometry of Region H. Building spacing, street form, wind direction, wind speed, relative humidity, and fire-spread duration were not independently varied; all calibration simulations used an NNW wind direction, a wind speed of 6.4 m/s, a relative humidity of 40%, and a fire-spread duration of 3 h. For the matched benchmark zones A, H, and J, the proposed relationship yielded relative differences of 0.21%, 19.31%, and 6.35%, respectively, relative to the TFD-based benchmark simulations. These comparisons are reported as simulation-informed matched-zone benchmarking within the same modeling framework and should not be interpreted as independent real-event or cross-city validation. From a sustainability perspective, the framework can support disaster-risk-informed urban planning by prioritizing districts for detailed analysis and mitigation, although environmental, social, and economic benefits were not quantified directly. Full article

Background and Objectives: Rare diseases collectively affect large populations and require coordinated, patient- and caregiver-informed health and social support. This study had an overarching descriptive objective to characterize patient- and caregiver-reported experiences of living with rare diseases in Türkiye and a secondary exploratory objective to assess differences across broad disease categories. Materials and Methods: This cross-sectional, anonymous, voluntary non-probability online survey recruited affected individuals and relatives/caregivers through patient-organization networks. The 44-item questionnaire covered sociodemographic and caregiving context, diagnostic experiences, psychosocial support, healthcare access, and service-improvement priorities. Descriptive analyses, exploratory disease-group comparisons, Cramér’s V, and selected multivariable logistic regression models were used. Results: A total of 1642 responses were analyzed; 76.7% were completed by relatives/caregivers, and 66.4% concerned patients younger than 18 years. Among participants diagnosed after symptom onset (n = 1388), 34.4% reported an incorrect or alternative diagnosis and 64.0% sought at least one additional specialist opinion. Psychosocial support was needed by 74.2%, whereas 34.5% reported receiving professional support. Access to current treatment services was perceived as moderately adequate by 29.3% and adequate by 13.0%. The leading service-improvement priorities were increasing healthcare professionals’ knowledge (82.9%) and establishing specialized rare-disease centers (79.5%). Conclusions: This patient-organization-recruited survey provides a large descriptive account of the experiences and priorities reported by participating patients and caregivers. The findings identify perceived priorities, particularly clinician education and specialized centers, while not constituting nationally representative estimates. Full article

The availability of large, structured legal datasets is essential for advancing legal natural language processing, information retrieval, and the development of reliable language technologies. However, such resources remain scarce for Kazakh, a morphologically rich and comparatively low-resource language, limiting the development and evaluation of domain-specific NLP methods. This data descriptor presents KazakhLawCorpus, a curated corpus of 215,889 document-level legislative metadata records from the Republic of Kazakhstan. The dataset was derived from a PostgreSQL export containing 223,245 source records and constructed through Kazakh-language field selection, essential-field filtering, source-preserving normalization, schema reconstruction, and systematic quality validation, resulting in a retention rate of 96.705%. The final release comprises 25 fields covering document identifiers, temporal attributes, legal status, registry information, Kazakh-language titles and descriptions, legal and institutional categories, geographical information, and source URLs. The records cover legislative acts adopted between 30 October 1947 and 20 July 2026. Quality-control procedures confirmed unique law_id values, complete title and description fields, and parseable adoption dates, while 4635 records with potential source-level or historical anomalies were retained and transparently documented through diagnostic flags. KazakhLawCorpus provides a large-scale, traceable resource for metadata-based legal information retrieval, document classification, temporal and institutional analysis, provenance tracking, and the development of downstream Kazakh legal NLP resources. Full article

Urban green spaces play an important role in supporting ecosystem functions, landscape connectivity, and urban resilience; however, their potential contribution may be constrained by fragmentation, inadequate maintenance, and limited integration among spatial planning instruments. This study develops a preliminary green infrastructure (GI) planning framework for the municipality of Noicattaro (Bari, Apulia, Italy) by integrating planning provisions with the observed condition and spatial distribution of existing public green spaces. A GIS-based spatial analysis, complemented by field surveys, assessed 89 public green spaces for compliance with planning requirements and maintenance status. The analysis also considered the spatial structure of the surrounding peri-urban landscape, including the karstic dry valleys (Lame), agricultural land uses, and relevant landscape and hydrogeological planning constraints. The results show that only 20.2% of the surveyed green spaces met both planning and maintenance criteria, whereas 78.4% fell into categories characterised by planning non-compliance, inadequate maintenance, or both. These findings highlight potential gaps between formal planning provisions and the observed condition and management of the municipal green system. The analysis of the Lame further revealed a predominance of agricultural land over natural vegetation, emphasising the importance of considering the wider agricultural and peri-urban landscape when identifying potential ecological connections. Based on the combined spatial and field assessment, a preliminary GI network framework was developed around core ecological areas, buffer zones, and strategically located green spaces, with the aim of identifying potential spatial connections and areas for further investigation. Overall, the study demonstrates the potential of integrating planning analysis, GIS-based assessment, and field observations as a screening and decision-support approach for more context-sensitive GI planning in peri-urban Mediterranean landscapes. The proposed framework should be regarded as a preliminary planning proposal requiring further validation through biodiversity, ecosystem-service, accessibility, and implementation assessments. Full article

Blockchain networks have become a major target for fraudulent activities, including phishing and scamming attacks, due to the irreversible and pseudonymous nature of cryptocurrency transactions. Existing Ethereum fraud detection studies commonly formulate the problem as a binary classification task, which limits their ability to distinguish among different malicious behaviors. This paper proposes an explainable Graph Neural Network-based framework for multi-class Ethereum address classification using an enhanced version of the Benchmark Labeled Transactions Ethereum (BLTE) dataset. The original binary BLTE dataset is extended into a multi-class dataset by integrating external attack-category information and labeling addresses as Normal, Phishing, or Scamming. The resulting BLTE multi-class dataset contains 73,033 Ethereum addresses, 23 behavioral node features, and 6588 directed transaction edges. Three GNN architectures, namely Graph Attention Network (GAT), Graph Convolutional Network (GCN), and GraphSAGE, are evaluated under identical experimental settings. Across ten stratified random seeds, GraphSAGE achieves the best overall performance among the implemented GNNs, with $98.84\pm 0.06\%$ accuracy, $80.42\pm 2.67\%$ macro recall, $51.08\pm 1.54\%$ macro F1-score, and $99.27\pm 0.03\%$ weighted F1-score. To improve model transparency, GNNExplainer is used to identify influential transaction subgraphs and important node features that contribute to each prediction. The results indicate that neighborhood-aware graph learning is useful for multi-class Ethereum address classification, while also showing that minority-class precision remains challenging under severe class imbalance. A feature-only multilayer perceptron is found to be a strong baseline whose macro F1-score is statistically indistinguishable from GraphSAGE after multiple-comparison correction; the improvement attributable to graph structure is therefore modest and concentrated on the small connected portion of the network, and external validation on independent datasets is still required before the framework is used in real investigations. Full article

Objectives: Oral frailty is associated with sarcopenia and underweight; however, its relationship with sarcopenic obesity remains unclear. This cross-sectional study investigated the associations between oral frailty and body mass index (BMI) categories (underweight/obesity), stratified by sarcopenia status. Methods: Data were obtained from dental checkups for 3829 older adults in Tottori, Japan (1489 men and 2340 women, mean age: 80.1 ± 4.3 years). Participants were divided into sarcopenia risk and non-sarcopenia groups by using the Yubi–Wakka test and then further subdivided by BMI into underweight, normal weight, or obese. Oral frailty was defined as meeting three or more of six oral frailty assessment criteria adapted from previous studies. Multinomial logistic regression analysis was performed to examine the association between oral frailty and each BMI category, with normal weight used as the reference category. Results: In the sarcopenia risk group, both obesity and underweight were significantly associated with oral frailty. The odds ratio (OR) for obesity was 1.727, with a 95% confidence interval (CI) of 1.092–2.732, and the OR for underweight was 1.443, with a 95% CI of 1.066–1.952. In contrast, no significant association between BMI category and oral frailty was observed in the non-sarcopenia group. Conclusions: Among late-stage older adults at risk of sarcopenia, oral frailty was associated with both obesity and underweight, with different contributing factors identified for each BMI category. Full article

Conditional imitation learning (CIL) for CARLA depends on diverse expert data spanning map, weather, traffic, and control-perturbation axes, yet it remains unclear which axes actually drive closed-loop behavior, and whether ablation conclusions survive deployment guardrails. We train a matched baseline and four equal-budget leave-one-axis-out (LOO) variants of a fixed CIL architecture (v14) and evaluate each under three nested tiers: pure policy rollout, minimal traffic-rule shields, and a fully deployed stack with route blending and recovery. The factorial design comprises $18\times 5\times 3$ scenario-variant-tier cells, each repeated under n = 5 traffic-seed replicates (1350 closed-loop episodes) to estimate NPC-seed variance on every eval stack. No single withheld axis dominates pooled outcomes. LOO effects are tag- (scenario-category) and spawn- (vehicle starting location) specific: removing perturbation-labeled recovery data costs $-285$ m on geometry_stress but can gain distance on in-distribution spawns; removing multi-town data changes held-out Town05 mobility on some routes while depressing others. Axis-importance rankings reorder across tiers; Kendall $\tau$ between pure and full rankings is $0.0$, and guardrails compress or erase pure-tier gaps (e.g., drop_perturbation pooled distance $\Delta$ from $-93$ m to 0 m). Pure-tier seed replicates show that geometry-driven lane-tracking degradation under drop_perturbation is seed-stable; traffic-axis and full-tier drop_traffic loads are more seed-sensitive. We release the evaluation ledgers, parsing scripts, and analysis tooling with the paper. Training-data ablation claims should report per-scenario or tag-stratified LOO metrics under pure evaluation; guardrailed tiers are supplementary deployment checks, not substitutes for isolating what the policy learned. Full article

Deep visual recognition can turn ordinary video into interpretable motor knowledge, yet coordination among the joints of a single body during social interaction remains largely unexplored. We present an interactive playback visualizer that couples markerless pose estimation with the cross-wavelet transform to quantify amplitude co-modulation between all joint pairs. Amplitude co-modulation proved anatomically structured: bilateral homologous pairs exceeded cross-limb and head–body pairs even among pairs sharing no keypoint, where correlated tracking error cannot produce it. Within-limb pairs also scored high but share keypoints, so their elevation is confounded with measurement error and not treated as established. Co-modulation concentrated at low postural frequencies and showed no systematic temporal trend; each profile remained temporally consistent within the session, and equivalence to a flat trend was not established. Vocal activity correlated with upper-limb movement at gesture frequencies. Each participant’s 78-dimensional profile was individually distinctive: split-half identification reached 28.6% (95% CI 19.3–40.1%) against 1.4% chance, demonstrating within-session identifiability rather than a cross-session trait. No sex differences were detected in overall or category-level co-modulation, and equivalence was not established for any measure; women exceeded men in the micro-movement band, the narrowest and most attenuated by preprocessing, so that difference is reported but not interpreted. Full article

Cirsium japonicum Fisch. ex DC. has long been used in traditional East Asian medicine and has attracted increasing attention because of its diverse pharmacological activities, including antioxidant, anti-inflammatory, antifibrotic, metabolic regulatory, and anticancer effects. Although numerous studies have investigated its phytochemical composition and biological activities, current evidence has largely been organized according to individual compounds or disease categories, providing limited insight into the shared molecular mechanisms underlying its pleiotropic actions. In this review, we critically re-evaluate the molecular pharmacology of C. japonicum using an evidence-oriented framework that distinguishes experimentally supported mechanisms from pharmacological associations and emerging hypotheses. Rather than accepting changes in signaling proteins or downstream biomarkers as sufficient evidence of mechanism, we assess the strength of evidence based on reproducibility, pathway-specific interventions, genetic or pharmacological validation, and direct target engagement. Current evidence indicates that nuclear factor erythroid 2-related factor 2 (Nrf2)-mediated antioxidant responses and nuclear factor kappa B (NF-κB)-associated inflammatory responses represent the most consistently observed pathway associations, although direct molecular targets and causal pathway dependency remain insufficiently established. Evidence for AMPK/PI3K-Akt-associated metabolic regulation and TGF-β/Smad-associated antifibrotic responses is comparatively more limited. In contrast, modulation of apoptosis and autophagy is currently supported primarily by indirect or context-dependent observations. We further discuss how multiple phytochemicals converge on interconnected signaling networks regulating oxidative stress, inflammation, metabolism, tissue remodeling, and cell fate, thereby providing a systems-level explanation for the broad therapeutic potential of C. japonicum. Finally, we highlight the need for standardized phytochemical characterization, rigorous target validation, multi-omics integration, and artificial intelligence-assisted systems biology to establish causal molecular mechanisms and accelerate the translational development of evidence-based phytopharmaceuticals. Full article

Background and Objectives: Deciding which pigmented or non-pigmented skin lesions require surgical excision, as opposed to continued observation, is a common challenge in everyday dermatologic practice. Although dermoscopy is well established as a tool that improves melanoma detection accuracy compared with naked-eye examination, real-world evidence on how demographic profile, patient history, and dermoscopic impression jointly influence the surgical decision remains limited. This study examined the clinical and dermoscopic factors associated with a positive surgical decision (PSD) in a large single-centre screening cohort. Materials and Methods: This retrospective cross-sectional analysis included 643 consecutive patients (415 female (64.5%), 228 male (35.5%); mean age 51.3 ± 16.8 years) undergoing dermoscopic skin cancer screening. Data recorded included family history of melanoma (FHM), personal history of skin cancer (HSC), childhood sunburn (CS), clinically atypical lesions (CAL), suspicion of basal/squamous cell carcinoma (BCC/SCC), nevus count (NC, 0–4), and the surgical decision (PSD). Free-text dermoscopic descriptions were grouped into seven diagnostic categories. Group comparisons used the Mann–Whitney U, χ², and Fisher’s exact tests; predictors of PSD were characterised using multivariable logistic regression (with Hosmer–Lemeshow and Nagelkerke R² statistics), multiple correspondence analysis (MCA), Gower-distance k-medoids clustering, and receiver operating characteristic (ROC) analysis, including the area under the curve (AUC). Missing NC data (13.2%) were assessed using Little's test for data missing completely at random (MCAR test). Results: A positive surgical decision was recorded in 133/643 patients (20.7%) and was significantly more common in men than women (25.9% vs. 17.9%; p = 0.017; corresponding to an odds ratio (OR) of 1.61). In multivariable analysis, suspicion of BCC/SCC (OR = 24.7, 95% confidence interval (CI): 9.26–66.0) and clinically atypical lesions (OR = 5.53, 95% CI 3.50–8.74) were the strongest independent predictors of PSD (Nagelkerke R² = 0.31; Hosmer–Lemeshow p = 0.977), far outweighing family or personal history. A significant interaction between childhood sunburn and nevus count (OR = 1.39, 95% CI 1.16–1.66; p < 0.001) predicted clinical atypia. Missing NC data were consistent with a completely random pattern (Little’s χ² = 2.80, df = 2, p = 0.247). Dermoscopic impression showed strong discriminative value for PSD: “no concern” had the highest negative predictive value (NPV = 0.927; AUC = 0.762), while “atypical lesion” combined high specificity (0.916) and positive predictive value (0.632; AUC = 0.736). On average, 4.8 patients were examined for every surgical referral generated (a screening-to-referral ratio, not a number needed to screen for a histologically confirmed cancer). Conclusions: Surgical decision-making in this cohort was driven almost exclusively by the dermoscopic impression—suspected BCC/SCC and clinical atypia—rather than by family or personal oncological history, while male sex was independently associated with a higher surgical rate. These findings support dermoscopy-led triage protocols and highlight the potential value of standardised risk stratification to optimise referral pathways in routine skin cancer screening. Full article

Background and Objectives: Caries experience, plaque accumulation, gingival inflammation, and periodontal screening findings may coexist as manifestations of a shared oral-health burden. We examined whether these routinely recorded indicators form distinct multivariable profiles or primarily represent ordered levels along a common burden continuum in children and adolescents. Materials and Methods: This secondary exploratory analysis included 638 participants from a multicenter cross-sectional clinical database: 407 aged 6–12 years and 231 aged 13–19 years. Permanent-dentition Decayed, Missing, and Filled Teeth score, Plaque Index, Gingival Index, and Community Periodontal Index were standardized and analyzed separately by age stratum. Unsupervised K-means clustering compared solutions containing two to five groups to summarize participant-level patterns without imposing predefined clinical categories. Hierarchical Ward classification assessed algorithmic concordance; principal component analysis and a composite standardized burden score assessed dimensionality. Sensitivity analyses excluded the periodontal index or treated it as ordinal with Gower-distance partitioning around medoids, and 1000 repeated 80% subsamples assessed sampling stability using the adjusted Rand index. Results: A single principal component explained 91.0% of variance at 6–12 years and 91.8% at 13–19 years, with all four indicators loading strongly on the same dimension. Ordered cluster membership correlated closely with the composite burden score (Spearman ρ = 0.942 and 0.939; both p < 0.001). The retained low-, intermediate-, and high-burden groupings were highly consistent when the periodontal index was excluded or treated as ordinal and under repeated subsampling. In adolescents, a two-group alternative preserved the low- and high-burden extremes while dividing the intermediate group, indicating that the number of groups changes descriptive granularity rather than revealing a separate phenotype. Conclusions: The principal new finding is that four commonly used oral-health indicators converge on one dominant participant-level burden dimension rather than defining distinct clinical phenotypes. This supports an integrated epidemiological description of cumulative oral-health burden while cautioning against interpreting data-driven groups as diagnostic or treatment categories. The numerical group distributions are sample-specific and require external and longitudinal validation before being transferred to other populations or clinical decision-making. Full article

To address the overlapping risks and weak agricultural resilience in karst underdeveloped mountainous areas, this study examines 110 new-type rural collective economy projects in Hechi City, Guangxi. Grounded theory is used to qualitatively identify risk factors. The analytic hierarchy process and fuzzy comprehensive evaluation are then combined to quantify risk indicator weights and overall risk intensity, clarify how various risks constrain agricultural resilience, and construct risk control pathways to enhance resilience. Findings reveal that risks fall into four categories—systemic, external environmental, operational and management, and financial and capital risks—encompassing ten secondary dimensions. The overall risk level is moderate but clearly stratified. Market price, capital recovery, operational and sales, investment decision-making, and internal governance are relatively high-risk dimensions. Specifically, policy support risk carries the greatest relative importance weight, while natural disaster risk has the lowest current intensity. Guided by prioritizing relatively high-risk areas for targeted reinforcement and routinely consolidating moderate risks, this paper proposes a three-dimensional strategy: targeted risk control, routine foundation consolidation, and multi-dimensional safeguards. This provides theoretical support and practical reference for precise risk governance and synergistic agricultural resilience enhancement in ecologically fragile karst regions. Full article

Recent image diffusion models have enabled automatic 3D object creation from text or image guidance, but their 2D generative priors often bake illumination and shadow into textures, making relighting and physically based rendering (PBR) difficult. To address this issue, we propose MaterialSeg3D, a framework that predicts surface materials for 3D assets by leveraging 2D material semantics. Given a mesh and its albedo UV map, MaterialSeg3D renders multi-view images, performs material segmentation using a 2D prior model, projects the predictions back to UV space, and fuses them through weighted voting and region unification to obtain coherent material maps. To train the prior model, we construct MIO++, a large-scale single-object material segmentation dataset containing 115,542 images, 12 object themes, and 30 fine-grained material categories, substantially extending the previous MIO dataset. Each MIO++ material category is associated with an independent pair of roughness and metallic values for PBR assignment. We further observe that poor topology in AI-generated assets can degrade PBR quality even when plausible materials are assigned, and introduce a plane-simplification strategy as an auxiliary preprocessing step for such meshes. Experiments show that MIO++ improves material segmentation and that the resulting material maps support more consistent relightable renderings for both human-crafted and AI-generated 3D assets. Full article