搜索结果：design

In this work, a compact two-port MIMO antenna has been proposed for 5G millimeter wave applications. The antenna is fabricated on Rogers RT/Duroid 5880 substrate and occupies a footprint of only 12.5 × 4.8 × 0.65 mm³, which is among the smallest dimensions reported in the literature for comparable designs. Even with such tight physical constraints, the antenna is able to provide a gain ranging from 9.3 to 9.7 dBi across three distinct operating bands, namely 26.47-30.85 GHz, 31.29-32.86 GHz, and 35.8-38.48 GHz. The multiband behavior is mainly attributed to the meandered slot-loaded structure, which effectively increases the electrical length without any corresponding increase in the physical dimensions of the antenna. The MIMO performance parameters are also found to be satisfactory. Port isolation of - 20 dB has been achieved, while the envelope correlation coefficient (ECC) is maintained below 0.01. The channel capacity loss is recorded as 0.33 bps/Hz, which confirms that the mutual coupling between the ports is sufficiently low and that stable multi-stream data transmission can be supported. The measured results for S-parameters, gain, radiation patterns, and surface current distributions are in good agreement with the simulated values, thereby validating the design experimentally. SAR analysis has been carried out using a multi-layered human hand phantom model. The obtained SAR values are found to be within the permissible exposure limits for all three operating bands, which is of particular relevance since the antenna is intended for use in handheld and wearable 5G devices. On comparison with recently published works in the literature, the proposed design demonstrates competitive performance in terms of physical size, gain, bandwidth coverage, and electromagnetic safety compliance. Hence, the design can be considered as a suitable candidate for IoT and wearable 5G communication systems.

Accurate and timely assessment of consciousness is critical for triage, escalation of care, and patient safety in emergency and hospital settings. However, documentation using the AVPU scale (Alert, Verbal, Pain, Unresponsive) remains inconsistent owing to high workload, subjectivity, and fragmented workflows. This study developed and evaluated Consc.ia, a video-based clinical decision-support platform that automates AVPU inference while preserving clinician oversight and enabling seamless, interoperable documentation through HL7 FHIR. A simulated AVPU dataset comprising 136 videos from 58 healthcare professionals (physicians, nurses, paramedics, and first responders) was created under controlled conditions with ethics approval from the ISCTE - Instituto Universitário de Lisboa Ethics Commission (reference CE-ISTA/2025.08, July 2025). The system architecture combines edge-computing computer vision for real-time extraction of facial landmarks, eye state, arm movement, and verbal responses; a clinician-in-the-loop validation layer; and FHIR-mapped Observation resources for direct EHR integration. Three deployment scenarios (Emergency Medical Services, Emergency Departments, and Intermediate Care wards) were designed and compared. Technology adoption was modelled using Rogers' Innovation Adoption Curve and the Bass Diffusion Model (p = 0.01, q = 0.35, M = 111 Portuguese hospitals). The architecture achieves low-latency inference with privacy-by-design (local processing, no raw video storage). Stakeholder validation confirmed strong workflow fit and highlighted persistent documentation gaps during EMS-to-hospital transitions. Scenario analysis revealed distinct hardware and integration requirements (ambulance edge device versus ward multi-camera server). Bass modelling projects gradual adoption, reaching approximately 50% of Intermediate Care wards by 2037 in the realistic scenario, with the "chasm" phase occurring between 2030 and 2032. Sensitivity analysis identified early clinical evidence and FHIR integration support as the strongest accelerators of diffusion. As this constitutes a proof-of-concept study, no quantitative AVPU classification metrics (e.g., accuracy, sensitivity, specificity, or confusion matrix) are reported at this stage; empirical model evaluation against expert-annotated clinical recordings is identified as the primary prerequisite for future validation and clinical translation. As a proof-of-concept that has not yet undergone clinical validation, Consc.ia offers a feasible, interoperable solution for standardising AVPU documentation and strengthening early warning systems. By combining video analytics, edge computing, clinician validation, and FHIR integration, the platform addresses a longstanding gap in emergency-care digitalisation and provides a clear roadmap for real-world adoption.

Accurate soil sampling is fundamental for producing reliable data in soil science, particularly when addressing spatial variability and logistical constraints. Well-designed sampling strategies that balance sample size, spatial heterogeneity, and operational costs enable effective selection of sampling locations, thereby improving parameter estimation and spatial interpolation accuracy. Despite this, real-world challenges-including limited accessibility, dense vegetation, rugged terrain, and restrictions in time and budget-often hinder the implementation of optimal sampling plans. A thorough review of published studies reveals that stratified random sampling consistently offers the highest statistical accuracy and reliability, whereas spatial coverage-oriented methods provide superior geographic representativeness. Simple random, grid-based, and spatial coverage approaches remain the most straightforward to implement, while conditioned Latin hypercube sampling (cLHS) stands out as a robust, widely recommended advanced method. Integrating stratified random and spatial coverage-based techniques emerges as the most effective strategy, maximizing statistical soundness, spatial representativeness, and cost-efficiency in soil sampling design. This review synthesizes the advantages and limitations of major statistical and geometric sampling methods, offering practical guidance for planning soil sampling campaigns that support rigorous scientific analysis and geospatial applications.

The rational design of heterogeneous catalysts requires predictive tools that balance quantum-level accuracy with computational efficiency for estimating adsorbate energetics. Although recent machine learning approaches have shown strong performance, they mostly depend on total energy decomposition and/or force supervision, where forces are obtained as the gradient of the potential energy with respect to atomic positions. As a result, they tend to be computationally complex, often requiring the prediction of several components, including the surface, the adsorbate, and the combined system. To make a contribution in addressing this challenge, we introduce Q-CatNet, a partially quantum-informed crystal graph catalyst network designed for direct prediction of initial structure-to-adsorption energy correlation for both single-atom and multi-atom (molecular) adsorbates on bulk catalyst surfaces. The model extends the Crystal Graph Convolutional Neural Network (CGCNN) framework by incorporating electrostatic Hamiltonian-inspired edge descriptors together with global features derived from density of states (DOS) and X ray diffraction (XRD) signals, enabling the capture of both local bonding interactions and system-wide physical correlations. Benchmarking on a curated dataset confirms its robustness: Q-CatNet outperforms the image-based Fourier-Transformed Crystal Property (FTCP) representation by 46% and exceeds several invariant graph-based architectures by margins ranging from 8% to 38%. These results highlight Q-CatNet as a physically grounded framework that bypasses the need for total energy decomposition, enabling a faster and practical adsorption‑energy prediction towards accelerated catalyst discovery in materials science and engineering.

Sexual and gender minority youth (SGMY) experience a heightened risk of HIV acquisition due to barriers to HIV prevention, specifically connected to a lack of comfort in discussing sexual identity and practices with healthcare providers (HCPs). Decision-aid tools that support communication and shared decision-making may improve both access to and uptake of numerous HIV prevention modalities among SGMY. The study aimed to inform the decision-making process of HIV prevention modalities for HCPs and SGMY, providing key information about HIV prevention modalities with PrEPChoices, a web-based decision aid tool. Our study recruited two participant groups, HCPs (N = 15) and SGMY (N = 18). Eligible HCPs held a Doctor of Medicine (MD), Doctor of Osteopathic Medicine (DO), Nurse Practitioner (NP), or Physician Assistant/Associate (PA) degree, were licensed to prescribe medications in at least one United States state, had provided care to patients aged 15 to 24 within the past month, and were 18 years of age or older. Eligible SGMY were assigned male sex at birth, reported sexual attraction to and/or sexual behavior with cisgender men or transgender women in the past six months, and were 15 to 24 years of age. Participants completed a semi-structured Webex interview focused on the clarity and relevance, presentation and usability, and application of PrEPChoices, a web-based decision aid tool to support HIV prevention modalities selection among HCPs and SGMY. Interview transcripts were coded using an iteratively developed codebook. Findings were thematically analyzed within a Strengths, Weaknesses, Opportunities, and Threats (SWOT) analysis framework. Thirty-three participants enrolled (N = 33), including 15 HCPs and 18 SGMY. Strengths included: (1) support for HIV prevention-based decision making; (2) utilization of the filtering tool to select preferred HIV prevention modalities; and (3) intuitive website design. Weaknesses included: (1) gaps in needs and literacy levels between HCPs and SGMY; (2) limited visual design appeal and inclusive representation; and (3) limited in-tool features for comparing HIV prevention modalities. Opportunities included: (1) potential for multimodal dissemination; and (2) expanding external resources provided. Threats included: (1) challenge of integrating PrEPChoices into clinical practices; and (2) ability to stand out compared to other online HIV prevention education resources. HCPs and SGMY highlight the value of PrEPChoices as a web-based decision aid tool to enhance HIV prevention modality selection among SGMY. Our results emphasize the critical role PrEPChoices can play in reducing barriers to sexual health education among SGMY, improving the patient-provider relationship with the PrEP landscape, and strengthening HIV prevention among SGMY.

Subsequent to an intracerebral hemorrhage (ICH), a cascade of neuroinflammatory response drives the process of secondary brain injury. At present, no anti-inflammatory nor neuroprotective pharmacological interventions have been demonstrated to improve functional outcome after ICH. This Phase 2b study was designed to establish the safety and feasibility of CN-105, a neuroprotective and anti-inflammatory pentapeptide designed from the receptor binding region of apolipoprotein E, in patients with acute primary supratentorial ICH. The Singapore CN-105 in Participants with Acute Supratentorial ICH Trial (S-CATCH, NCT03711903) was a randomized, double-blind, placebo-controlled trial involving 60 patients (30 CN-105, 30 placebo) treated within 12 h of symptom onset. Safety was assessed through adverse events (AEs) and serious AEs (SAEs), while efficacy was evaluated using functional outcome measures, including the modified Rankin Scale (mRS) at 90 days. CN-105 was safe and well tolerated in patients with acute ICH, with no significant differences in incidence of SAEs between groups (30% SAEs in placebo vs. 26.7% in CN-105). Notably, fewer patients treated with CN-105 group experienced in-hospital neurological deterioration (0 vs. 10% in placebo). While treatment was not associated with a statistically significant improvement in 90-day mRS, higher proportion of patients treated with CN-105 achieved favorable mRS scores (≤ 3) compared with those in the placebo group (77.8 vs. 66.7%; p = 0.35). This Phase 2b trial confirmed the safety and feasibility of CN-105 administration in the acute setting of ICH. Although no statistically significant improvements in neurological outcomes were found, the observed trends warrant further investigation. Future Phase 3 trials should focus on refining patient selection and assessing the therapeutic efficacy of CN-105 in more targeted subgroups such as those with medium-sized subcortical ICH. Trial registration NCT03711903, https://clinicaltrials.gov/ https://clinicaltrials.gov/study/NCT03711903?term=NCT03711903&rank=1 . Registered 16 October 2018.

Vascular diseases, particularly atherosclerosis, represent a leading cause of global morbidity and mortality. Endovascular stenting has emerged as a cornerstone of therapy to restore vessel patency, yet conventional stents remain obstructed by significant clinical limitations, including in-stent restenosis, thrombosis, and mechanical failure. These adverse outcomes are intrinsically linked to their fundamental structural design, which is characterized by a positive Poisson's ratio, leading to foreshortening and a biomechanical mismatch with the native vasculature. This review critically examines auxetic stents as a next-generation solution, engineered with a structure possessing a negative Poisson's ratio. This unique property allows them to expand axially upon radial deployment, thereby eliminating foreshortening, enhancing conformability to tortuous vessels, and distributing mechanical stress more uniformly onto the arterial wall. This paper synthesizes the robust body of in-silico/bench-top evidence from computational modeling and in-vitro experimentation that validates these superior biomechanical characteristics. Furthermore, it explores the profound and favorable biological implications, arguing that the optimized mechanical environment and improved hemodynamics are hypothesized to attenuate the primary triggers for neointimal hyperplasia and foster rapid, complete endothelialization. The review concludes by outlining the translational pathway, including challenges in structure integration and discussing the vast future horizons for auxetic structured stents in complex peripheral, carotid, and non-vascular applications. Auxetic design represents a paradigm shift from material-centric iteration to structure-driven innovation, holding the promise to significantly improve the long-term safety and efficacy of endovascular stent implants.

The natural antifungal peptide Histatin 5 (Hst 5) is a histidine-rich cationic peptide secreted by human salivary glands and a key component of oral innate immunity, but its moderate activity limits clinical use. Hst 5 enters Candida albicans via the membrane receptor Ssa1/2. Here, we integrated artificial intelligence-assisted and computer-aided drug design to rationally modified the sequence structure of Hst 5. Truncated derivatives of Hst5 were screened for antimicrobial potential using ESM2-AFPpred, and high-probability candidates were docked with Ssa1/2. The Hst 5-22 was identified, then redesigned based on alanine scanning to yield the optimized derivative Hst 5-22-RW. Compared with Hst 5, Hst 5-22-RW has a shorter sequence, stronger Ssa1/2 binding, and improved activity against C. albicans. It also shows superior activity against fluconazole-resistant strains. RT-qPCR and transmembrane tracking confirmed higher cellular transport efficiency in C. albicans. The CADD/AIDD-driven optimization successfully generated the highly active antifungal peptide Hst 5-22-RW, providing a novel strategy for rational modification of antimicrobial peptides.

Tobacco 21 (T21) laws effectively reduce youth tobacco use by preventing initiation. This study examines their impact on body weight among young adults aged 18-20. Using 2009-2019 Behavioral Risk Factor Surveillance System data and a two-way fixed-effects difference-in-differences (DID) design, we find limited evidence of broad weight changes in either direction across the BMI distribution. Obesity declines due to modest weight reductions concentrated near the upper BMI threshold, with no significant changes in overweight status or average BMI. Event study shows that the obesity decline emerges in the first post-T21 year and attenuates afterward. Results are robust to alternative specifications, including an imputation DID approach addressing staggered adoption. Effects are driven by "never smokers", consistent with a prevention-based pathway, and are more pronounced among males and non-White individuals, with heterogeneity observed across education levels in the upper BMI tail. Supplemental analyses using Youth Risk Behavior Survey data show reduced adverse weight outcomes among high schoolers aged 18+. T21 laws increase exercise, improve diets, and reduce sedentary behavior, underage drinking, marijuana use, and mental distress. Overall, T21 laws avoid the typical cessation-related weight gain and modestly improve weight outcomes among at-risk young adults, suggesting broader public health benefits beyond tobacco prevention.

To address the dilemma of homogeneous talent training and the efficiency bottleneck of human resource management in universities, this study proposes an innovative personalized training framework integrating artificial intelligence, big data, and deep learning. Based on the 18-dimensional full-cycle behavior dataset of 5,000 students and OULAD dataset, a multimodal heterogeneous data fusion pipeline is constructed. This study adopts Generative Adversarial Network (GAN) for data imputation and bias optimization, designs Hierarchical Attention Graph Neural Network (HA-GNN) to capture hierarchical correlations among features, and uses Long Short-Term Memory (LSTM) to model temporal behavior patterns. The experimental results demonstrate that, under 10 independent repeated runs with random seed variation, the Hierarchical Attention Graph Neural Network-Long Short-Term Memory (HA-GNN-LSTM) model achieves lower prediction error on the academic performance prediction task, with a Mean Absolute Error (MAE) of 4.2 ± 0.3. Compared with the Temporal Fusion Transformer (TFT) baseline model, MAE is reduced by 31.1%. Welch's two-tailed t-tests based on independent run results remain statistically significant after Holm-Bonferroni multiple comparison correction [Formula: see text]. The Normalized Discontinued Cumulative Gain at Top 5 (NDCG @ 5) index of personalized recommendation system reaches 0.90, which verifies the effectiveness of spatio-temporal feature modeling. At the management application level, the improvements in advisor allocation response time and resource idle rate are derived from simulation experiments based on historical data replay, rather than online deployment in real campus management systems. The simulation results demonstrate that, under established constraints and historical sample distributions, advisor allocation response time could be reduced by 60% and resource idle rate could be decreased by 63.4%. These findings indicate the framework's potential for optimizing educational resource allocation. However, its managerial benefits require further validation through subsequent real-world deployment and long-term follow-up studies.

Brain tumors present a major global health concern, and a precise diagnosis is essential for proper treatment. Many existing MRI-based machine learning approaches focus solely on segmentation or classification, rather than addressing both tasks together. To bridge this gap, a unified deep learning model is designed to perform tumor segmentation and multiclass classification within a single architecture. By integrating a segmentation backbone with a dedicated classification head, the framework simultaneously captures anatomical details and tumor-specific features. On three public datasets, the proposed model achieved up to 99.6% classification accuracy and 0.935 Dice score, with average performance of 96.9% accuracy and 0.966 F1-score on Dataset 1, 99.4% accuracy and 0.984 F1-score on Dataset 2, and 98.2% accuracy and 0.982 F1-score on Dataset 3. Thus, evaluated on these publicly available brain MRI datasets, the proposed network outperforms CNN-based baselines and recent attention-based models, delivering improved tumor localization and classification accuracy within an integrated segmentation-classification framework, while maintaining computational efficiency. These results highlight its strong promise for supporting clinical decision-making in brain tumor diagnosis and treatment planning.

Bilayer MXene@Cu and trilayer MXene@Cu@PPy nanocomposites were synthesized for lightweight, thin X-band (8-12 GHz) electromagnetic wave absorbers. Ti3AlC2 MAX phase was prepared via molten-salt synthesis at 1050 °C, etched with HF to yield Ti3C2Tₓ MXene nanosheets, then loaded with Cu (0.5-4 mol%) via precipitation, and coated with PPy through in-situ polymerization. Low Cu loadings in MXene@Cu enabled optimal conductivity, strong interfacial polarization, and impedance matching (|Zin/Z0| ≈ 1), yielding high reflection loss (RL). High Cu disrupted matching, weakening absorption. PPy addition in ternary NCs increased polarization interfaces, stabilized conductivity, enhanced matching, and boosted dielectric/magnetic losses, improving RL across Cu ratios. Analysis of permittivity, permeability, attenuation, and Cole-Cole plots confirms balanced conductivity, polarization, magnetic loss, and matching as key to superior performance. This MXene-metal-polymer design advances broadband EMWA materials.

Anal fistula poses a significant clinical challenge with escalating research interest reflected by a more than 50-fold increase in publications from 2006 to 2026. This bibliometric study systematically analyzed 505 PubMed-indexed articles to elucidate the evolving knowledge structure and research paradigms in anal fistula diagnosis and treatment. Utilizing advanced visualization and clustering techniques via the R bibliometrix package, the analysis mapped global publication trends, geographic and institutional contributions, collaboration networks, journal impact, author influence, and keyword evolution. Results identified China and the United States as leading contributors with distinct international collaboration clusters, while Korean institutions demonstrated notable productivity and specialized research focuses. Key journals such as Diseases of the Colon & Rectum concentrated the majority of domain-specific publications, predominantly within moderate impact factor tiers. Authorship networks revealed diverse, multinational collaborative clusters emphasizing both traditional surgical and emerging minimally invasive approaches. Keyword co-occurrence and citation analyses indicated a thematic shift from basic pathophysiology and surgical techniques toward patient-centered outcomes, quality of life assessments, and prospective study designs. These findings highlight the field's transition toward precision medicine and interdisciplinary integration, underscoring the importance of evidence-based clinical decision-making and international cooperation. This comprehensive bibliometric mapping offers valuable insights to guide future research priorities, foster collaborative innovation, and improve therapeutic strategies aimed at enhancing patient outcomes in anal fistula management.

Extending the π-conjugated bridge in donor-π-acceptor dyes improves light harvesting but systematically erodes interfacial charge-transfer performance. Here, density functional theory (DFT) and time-dependent DFT (TD-DFT) were used to investigate triphenylamine-cyanoacrylic acid dyes with one to four thiophene bridge units (T1-T4) alongside a site-specific alkyl-chain engineering strategy on the tetra-thiophene scaffold. Bridge extension progressively red-shifts the absorption maximum from 445 nm (T1) to 508 nm (T4) but simultaneously delocalizes excited-state electron density over the π-bridge, reducing effective electron density at the acceptor-TiO₂ interface and weakening adsorption stability, dye regeneration thermodynamics, and injection efficiency. Site-specific alkyl substitution addresses this conflict: the alternating a, d-substitution pattern constructs a bilateral steric fence that suppresses face-to-face π-π stacking and relocates dye-electrolyte interactions away from the conjugated core, thereby suppressing charge recombination. Lorentzian fitting of projected density-of-states profiles confirms that all alkylated variants retain ultrafast electron injection and near-unity injection efficiency, demonstrating that the steric modification is electronically decoupled from the injection channel. These results establish site-specific alkyl-chain engineering as an effective strategy for mitigating the trade-off inherent to long-bridge dye design.

Aims To explore teachers' perspectives on incorporating oral health education within schools, since its statutory introduction in the National Curriculum in England in 2020.Methods An online survey including open and closed questions was designed and distributed via postal invitations to selected schools in North West England, and via social media between September 2024 and January 2025. Quantitative data were analysed descriptively, while directed content analysis was used for free-text responses.Results Fifty-four responses were received. Oral health education was reported as inconsistent with different methods and frequencies of delivery reported; 21% (n = 11) teachers taught the topic less than once a year. Challenges included lack of curriculum time (44%, n = 23), and resources to help deliver (27%, n = 14) or plan the teaching (19%, n = 10). Most teachers (93%, n = 50) expressed confidence in their oral health knowledge, but free-text responses highlighted that teachers valued dental professionals' involvement with schools, and parental engagement to reinforce oral health at home.Conclusions Oral health education in schools remains inconsistent despite statutory requirements. Sustainable resources and multi-agency partnerships can help embed oral health promotion within a whole-school framework.

Electrochemical chloride ion removal is essential for clean water and environmental protection, yet its practical application is hindered by the sluggish kinetics, especially using high-mass-loading electrodes. Conventional extrinsic modifications, such as conductive additives or structural design, exhibit constrained effectiveness. Here, we report an intrinsic enhancement strategy through heteroatom doping-induced dual defects engineering, demonstrated by the successful synthesis of fluorine-doped copper(I) phosphide with phosphorus vacancies (F-Cu3PV) via molten salt treatment. Based on density functional theory calculations and experimental results, F doping caused lattice distortion, generating P vacancies to form dual defects. These defects effectively modulated intrinsic electron redistribution, resulting in improved electrical conductivity, enhanced adsorption capability, and reduced chloride ion diffusion energy barriers. Therefore, electron transfer and ion diffusion kinetics were significantly accelerated, leading to superior electrochemical performance. Resultantly, the F-Cu3PV electrode performed exceptional electrochemical chloride ion removal performance with superior areal deionization capacity (3.16 ± 0.02 mg cm-2) and a remarkably rapid areal deionization rate (0.106 ± 0.001 mg cm-2 min-1), as well as outstanding cycling stability (95.65% retention after 70 cycles). This work elucidates electron redistribution via heteroatom doping-induced dual defects as a viable pathway to overcome the intrinsic kinetic bottleneck for high-performance electrochemical chloride ion removal.

Breast density is a breast cancer risk factor. The accurate quantification of breast density requires reliable segmentation of dense tissue in mammograms, but it is a challenging task due to large variations in tissue appearance across hospitals and imaging devices. We propose MammoDenseSegNet, a new deep encoder-decoder convolutional neural network designed to enhance segmentation performance through two complementary modules: a) Adaptive dual attention module, which captures long-range spatial and channel interdependencies to provide focused attention on relevant dense tissue areas regardless of their location; and b) Multi kernel receptive field module, which enlarges the network's receptive field at the bottleneck layer to aggregate multi-scale contextual features. Additionally, a multi-scale dice loss with deep supervision guides learning across decoder levels to improve robustness. We evaluated MammoDenseSegNet on two public digital mammogram datasets (VinDR-Mammo and EMBED) and one private dataset, spanning a variety of breast densities and imaging artifacts in a total of 1499 images from 606 women. Statistical analysis was done using generalized linear models accounting for correlation among images from the same women and adjusting for potential confounders (proc genmod, proc mixed, SAS v.9.4, SAS Institute, Cary, NC). MammoDenseSegNet demonstrated consistently high performance across various conditions (with Recall ranging from 0.64 to 0.90 and Dice from 0.63 to 0.91) and significantly (p&#x2009;<&#x2009;0.001) outperformed the publicly available state-of-the-art algorithm based on the VGG16 (with Recall from 0.04 to 0.91 and Dice from 0.06 to 0.82 across the same conditions). The improvement was largest for low-density tissue, where the baseline algorithm practically fails (with the mean Recall of 0.14 and Dice of 0.16) while MammoDenseSegNet remained clinically useful (with the mean Recall of 0.66 and Dice of 0.63).

To address the insufficient bearing capacity of roadside backfill bodies and the tilting or failure induced by uneven pressure relief of the coal seam during gob-side entry retaining in thick coal seams with hard roofs, combined compression-shear loading tests incorporating rapid resistance build-up and varying inclination angles were performed. A novel Compression Shear Coupling Test system (CSCT) was developed, and a fitted relationship between backfill width and roof subsidence was established. The strength degradation behavior of backfill specimens subjected to different shear stress components was systematically investigated. The results reveal that the peak strength of the specimens declines with increasing shear stress component, and the failure mode transitions progressively from compressive to shear-dominated failure. The high-resistance backfill material derived from this study was implemented at the N2302 gob-side entry retaining working face, accompanied by an anti-tilting design for the backfill wall. The measured roof subsidence was reduced by 59.4% relative to the theoretically predicted value, and no evident signs of failure or deterioration were observed in the backfill body. These findings provide both data support and theoretical reference for gob-side entry retaining under similar mining conditions.