Search Results (656)

Non-stationarity is increasingly recognized as a defining feature of contemporary hydroclimatic regimes, challenging the statistical assumptions that underpin inter-annual discharge deficits analysis and water-resources design. This study investigates how shifts in first- and second-order statistical moments (mean and variance, respectively) alter the perceived rarity and persistence of inter-annual drought events in low-memory, rapid-response mountain systems. I develop a stochastic Monte Carlo framework to explore changes in inter-annual discharge deficit frequency and multi-year drought clustering across successive climatic regimes, using the Northern Apennines (Italy) as a representative case study. The model is explicitly exploratory: it does not aim to reproduce observed discharge distributions, but to quantify how regime shifts in mean and variability propagate into tail exceedances and drought spells under stationarity-based metrics. Results show a pronounced amplification of annual hydrological drought exceedances and the emergence of persistent multi-year drought spells under contemporary conditions, which are strongly underestimated when historical baselines are assumed stationary. A comparison with long-term regional discharge trends—while acknowledging the distinct hydro-climatic response of high-memory versus low-memory basins—serves to contextualize the systemic nature of the observed drought amplification. The findings highlight the structural vulnerability of low-memory catchments to non-stationary forcing and underscore the limitations of traditional design thresholds for drought-risk assessment under the evolving climate. Full article

Archaeological sites are fragile, non-renewable heritage environments where visitors’ inappropriate behavior may cause irreversible damage. This study designed and evaluated a VR serious game, Guarding the Terracotta Warriors, to promote visitors’ responsible visiting behavior intention at archaeological sites. Examining the Terracotta Warriors Museum as a case study, this study identified conservation needs through literature and case analysis, expert interviews, and preliminary visitor interviews, and translated them into four design requirements: path guidance for visiting boundaries, reflection on visiting order, contextualized risk decisions, and concrete feedback on behavioral consequences. Based on the MDA framework, these requirements were further transformed into interactive game mechanisms. A within-subject experiment with 46 participants compared the VR serious game with conservation-content-matched video instruction, followed by semi-structured interviews with 20 participants. Quantitative results showed that the VR serious game produced significantly higher scores than video instruction in presence, experience satisfaction, conservation-related attitude, awareness of consequences, personal norms, and responsible visiting behavior intention. Task load was slightly higher in the VR condition, but its mean score remained below the midpoint of the seven-point scale, indicating a relatively modest level of perceived task load. Qualitative findings further indicated that embodied exploration, safe-to-fail risk decisions, visitor-order feedback, and irreversible restoration feedback helped participants understand conservation boundaries and behavioral consequences. This study provides a design framework and empirical evidence for using VR serious games to support responsible visitor education at archaeological sites. Full article

Modern energy systems, including those supporting transportation electrification, are increasingly exposed to volatile market conditions and external events. Effective decision-making therefore requires the integration of structured operational data with unstructured contextual information. Existing studies on Large Language Model (LLM)-assisted energy systems have mainly applied LLMs to individual tasks such as forecasting, scheduling, or decision support, while forecasting and control are typically treated separately. As a result, semantic information extracted from external events is not consistently propagated from market prediction to operational decision-making. This paper proposes a unified dual-stage framework in which the LLM functions as a shared semantic information processor, converting raw event data into structured representations used by both forecasting and control modules. In the forecasting stage, these representations improve price prediction under non-stationary conditions. In the control stage, the same information provides an event-aware contextual action prior for reinforcement learning-based energy management. This design allows external event information to inform both future-state estimation and subsequent control decisions, establishing a consistent connection between prediction and decision-making. The framework is evaluated using real-world electricity market data and a battery energy management environment. The results show that the proposed framework achieves the highest average cumulative reward among the evaluated methods while maintaining greater robustness than the forecasting-only LLM configuration. Overall, this work demonstrates the benefit of consistently propagating structured semantic information across forecasting and control and provides a viable approach to event-aware intelligent energy management, with potential extensions to multi-energy transportation systems and electrified mobility applications. Full article

Architecture design studios are complex learning environments in which knowledge develops through critique, collaboration, individual reflection, material exploration, and increasingly digital forms of practice. Although the physical organization of these spaces influences how such activities unfold, research on furniture arrangement remains fragmented across ergonomics, learning environments, educational research, technology-enhanced education, and participatory design. This structured integrative review examined the mechanisms reported between furniture arrangement, related physical spatial configurations, and learning processes in architecture design studios and relevant adjacent educational settings. It also investigated the pedagogical, ergonomic, technological, and cultural conditions under which these mechanisms were reported as supportive or restrictive. Scopus, Web of Science All Databases, and ScienceDirect were searched on 7 August 2026. The searches identified 64 source-level records: 26 from Scopus, 19 from Web of Science, and 19 from ScienceDirect. Eight duplicate records were identified within the ScienceDirect results. Following cross-source deduplication, title and abstract screening, and full-text eligibility assessment, 15 publications were included in the integrative synthesis. Study characteristics, methodological quality, contextual relevance, furniture-related conditions, pedagogical activities, reported outcomes, explanatory mechanisms, evidentiary directness, and transferability were examined through an integrative, mechanism-oriented synthesis. The resulting relationships were subsequently organized through an abductive framework-development process into the provisional Adaptive Studio Learning Ecosystem (ASLE) framework. ASLE comprises five interrelated dimensions: spatial flexibility, ergonomic responsiveness, pedagogical mediation, technological support, and cultural–participatory fit. The synthesis supports activity–layout alignment rather than a universally optimal furniture configuration and indicates that spatial adaptability becomes educationally meaningful only when supported by appropriate pedagogical practices, ergonomic conditions, technological integration, and patterns of user participation. Because the evidence base includes a limited number of direct architecture-studio studies and relies partly on mechanisms transferred from adjacent settings, ASLE should be treated as a provisional, review-derived framework requiring empirical validation. Full article

Object detection in remote sensing imagery remains challenging due to vast scale variations, complex backgrounds, and the prevalence of small, densely packed targets. Existing CNN-based detectors are limited by restricted receptive fields, while Transformer-based methods incur prohibitive computational overhead for high-resolution inputs. In this paper, we propose SFSMamba-DETR, a detection framework that integrates state space models with Dual-Scale Window Attention for efficient and accurate remote sensing object detection. Specifically, we design a Selective Feature Scanning (SFS) module that uses the Mamba-based 2D Selective Scan mechanism to model long-range spatial dependencies with linear computational complexity. To capture both fine-grained local patterns and broader contextual cues simultaneously, we introduce a Dual-Scale Window Attention (DSWA) mechanism that operates at two complementary window scales with multi-kernel convolution bridging. These modules are orchestrated within a Cross-scale Feature Aggregation Module (CFAM) that performs hierarchical multi-scale fusion in a hybrid encoder. Extensive experiments on three primary benchmarks (MAR20, UCAS-AOD, and the Jilin-1 Satellite Aircraft Detection Dataset), together with supplementary results on DOTA and DIOR, demonstrate that SFSMamba-DETR achieves strong detection accuracy while maintaining competitive inference speed. Full article

Background/Objectives: Ageing populations face complex challenges, yet many older adults maintain well-being despite adversity. Understanding resilience offers critical direction for health promotion. This integrative review takes a conceptual and methodological approach to resilience in older adults, with clear implications for geriatric care. The review addresses: (1) how resilience is conceptualised and measured; (2) individual, social, and contextual factors linked to resilience; and (3) the effectiveness of interventions that aim to promote resilience. Methods: Following Whittemore and Knafl’s framework, a systematic search of six databases (PubMed, CINAHL, MEDLINE, Scopus, Web of Science, and Cochrane CENTRAL; 2016–2025) identified 42 empirical studies across 18 countries of diverse designs. Quality was assessed using the Mixed Methods Appraisal Tool, and thematic analysis identified five convergent themes. Results: Most studies explicitly defined resilience using trait-, process-, or outcome-based frameworks and employed validated instruments, most commonly the CD-RISC. Self-efficacy, spirituality, sense of coherence, and gratitude were consistently associated with resilience. Family relationships, social networks, and community participation were important relational correlates. Health status, education, economic security, and digital literacy were key contextual factors. Six longitudinal studies linked higher resilience to reduced mortality, slower cognitive decline, and attenuated frailty progression. Seven intervention studies, including five randomised controlled trials, supported physical activity, mindfulness, positive thinking training, Tai Chi, and technology-supported exercise as promising intervention modalities, though one web-based programme yielded null findings. Conclusions: Resilience in older adults is a multifaceted and potentially modifiable construct. Growing longitudinal evidence extends its relevance beyond psychological well-being to physical health outcomes, though cross-sectional designs predominate and limit causal inference. Measurement heterogeneity across more than twenty instruments underscores the need for international consensus and culturally valid tools. Larger longitudinal studies and adequately powered trials are needed to guide definitive practice and policy for ageing populations worldwide. Full article

Contemporary approaches to urban water management increasingly rely on concepts such as Nature-Based Solutions (NBSs), Green Infrastructure, and Blue-Green Infrastructure. Although these frameworks have gained broad acceptance, their typological character provides limited guidance for project-oriented decision-making, as they primarily describe infrastructure types rather than their functions within integrated hydrological systems. To address this gap, this study proposes a complementary functional-operational framework for interpreting ecological infrastructures in urban flood mitigation. Employing a qualitative comparative case study methodology, we analysed four diverse international models—the Dutch Water Squares (Rotterdam), Tokyo’s underground flood control system, Copenhagen’s Cloudburst Management Plan, and Singapore’s ABC Waters Programme—to examine the systemic interaction between grey, green, and blue infrastructures at different watershed scales. The results indicate that flood mitigation effectiveness depends less on the predominance of a single infrastructure type and more on the functional coupling among them. Specifically, three primary functions were identified: rapid conveyance (grey infrastructure), infiltration and thermal regulation (green infrastructure), and dynamic storage and biodiversity support (blue infrastructure). Despite the contextual limitations and varying scales of the selected cases, blue infrastructure universally emerges as a systemic buffer that enhances urban resilience by regulating excess volumetric flows. Ultimately, the proposed framework introduces an actionable interpretative layer that complements existing typological classifications, providing planners and urban designers with a robust, scalable basis for implementing integrated ecological infrastructures. Full article

Accurate morphological analysis of blood smears is vital for hematological diagnosis, yet manual examination is labor-intensive and subjective. While deep learning offers automation, its black-box nature and computational demands often hinder clinical trust and deployment. We propose XHIC-Net, an Explainable Hybrid Involution–Convolution Network designed for efficient and transparent cell classification. By integrating spatially adaptive involution operations with convolutional layers within a residual framework, XHIC-Net captures both contextual and fine-grained features efficiently. To enhance interpretability, a Grad-CAM-based explainable AI (XAI) module visualizes the cellular regions driving model predictions. The proposed framework was evaluated on a dataset comprising 12,879 microscopic blood smear images belonging to 12 morphological cell categories. Experimental results demonstrate that XHIC-Net achieves an overall accuracy of 98.88%, precision of 98.89%, recall of 98.87%, F1-score of 0.9887, and Cohen’s Kappa score of 0.9887. It outperformed established models, including DL models such as EfficientNetV2S, MobileNet family, DenseNet family, and VGG16, while using fewer parameters and requiring shorter training times. Furthermore, the XAI maps consistently highlighted biologically relevant structures, validating the model’s decision-making process. XHIC-Net is a strong, effective, and clear research model for automated hematology. With future clinical validation, it has the potential to be modified for point-of-care diagnostics in healthcare settings with limited resources. Full article

Background and Objectives: Free flap reconstruction is the gold standard for complex head and neck defects following oncological ablation. While surgical techniques have improved flap success rates, donor site sensory morbidity remains relatively understudied. The radial forearm free flap (RFFF) and fibula free flap (FFF) are the two most commonly used free flaps for head and neck reconstruction and sensory abnormalities at their donor sites represent a significant but underrecognized source of postoperative morbidity. Materials and Methods: The aim of this study is to systematically review and synthesize the existing evidence on the prevalence, characteristics and assessment methods of sensory abnormalities and pain at the donor sites of RFFF and FFF harvests. A systematic review was conducted according to PRISMA 2020 guidance. PubMed/MEDLINE, Scopus, Web of Science, the Cochrane Library, and Google Scholar were searched from inception to 1 October 2025, with additional gray-literature and reference-list screening. Primary studies reporting sensory abnormalities, donor-site pain, neuropathic pain, or cold intolerance after RFFF and/or FFF harvest were included; prior reviews were used for citation tracking and contextual comparison only. Risk of bias was assessed with study-design-specific Joanna Briggs Institute (JBI) checklists, and the level of evidence was classified using the Oxford Centre for Evidence-Based Medicine (OCEBM) framework. Descriptive prevalence summaries and exploratory random-effects pooled estimates were calculated when numerators and denominators could be extracted. Results: From 1247 initially identified records, 73 studies met the inclusion criteria (36 RFFF-specific, 35 FFF-specific, and 2 addressing both flap types). Most studies were observational, and sensory outcomes were frequently secondary endpoints. For RFFF, reported sensory abnormalities ranged from 0% to 100% across extractable studies (median 31.3%; IQR 13.1–55.0%), reflecting differences in definitions, follow-up, and testing; the exploratory pooled prevalence was 34.4% (95% CI 23.6–47.1%; I2 = 88.4%). General donor-site pain ranged from 0% to 36.8% (median 12.0%; pooled 16.0%, 95% CI 9.9–24.8%), whereas Douleur Neuropathique en 4 Questions (DN4)-defined neuropathic pain was 18% in the largest dedicated RFFF study and 18% in a smaller pain-focused study. Cold intolerance ranged from 0% to 40.0% (median 16.2%). For FFF, sensory abnormalities ranged from 0% to 76.3% (median 21.4%; IQR 8.2–46.8%; pooled 24.0%, 95% CI 16.5–33.6%; I2 = 88.6%). General pain ranged from 0% to 72.7% (median 17.5%; pooled 18.8%, 95% CI 13.9–24.9%), and DN4-defined neuropathic pain was 21% in the only dedicated FFF study. Objective testing (quantitative sensory testing [QST], Semmes-Weinstein monofilaments, two-point discrimination, or standardized neurological examination) was used inconsistently, and no included study systematically used nerve conduction studies or electromyography (EMG). Conclusions: Sensory abnormalities and pain after RFFF and FFF harvest appear common, but available estimates remain uncertain because of heterogeneous definitions, follow-up intervals, assessment instruments, and predominantly Level 2b-4 evidence. The literature shows a persistent gap between patient-reported symptoms and objective sensory testing. Future prospective studies would benefit from standardized neuropathic pain screening, structured neurological examination, and QST; however, current data do not support strong causal inferences or a validated treatment algorithm. Full article

With the accelerated progress of computational aesthetics and digital interior design, the demand for real-time and precise detection of aesthetic objects on edge devices has become increasingly pressing in applications such as intelligent design assistance, domestic aesthetic assessment, and augmented reality-based interior staging. As a core task in digital home aesthetics governance, virtual interior furnishing, household cultural archive development, and automated aesthetic evaluation, multi-view aesthetic object detection plays an essential role. However, this task still faces substantial difficulties arising from pronounced viewpoint variation, scale inconsistency, reflective materials, intricate decorative patterns, and cluttered indoor scenes. To address these issues, this study presents GhostNetV2-YOLO, a lightweight yet robust detection framework designed for accurate localization of aesthetic objects under unconstrained multi-view acquisition settings. The task is formally defined as closed-set detection of 10 pre-selected home aesthetic decorative items, including both planar decorative pieces and three-dimensional ornamental objects, and all performance claims are bounded within the horizontal bounding box detection paradigm. The framework incorporates three complementary components tailored to the target task. First, a task-adapted GhostNetV2 backbone is employed to enable efficient multi-scale feature extraction and long-range dependency modeling, with optimization specifically oriented toward structured aesthetic objects with stable global contours under viewpoint variation. Second, an improved Attention-based Intra-scale Feature Interaction (AIFI) module is introduced, integrating compressed QKV projection, linear attention, depthwise spatial refinement, and channel gating so that reflection-induced noise and background disturbance can be effectively reduced. Third, an enhanced Distance-IoU regression loss is adopted, in which explicit edge alignment and dynamic sample weighting are incorporated to improve boundary regression accuracy for rectangular and regularly contoured aesthetic objects. These designs jointly enhance contextual representation, boundary localization, and computational efficiency. Extensive experiments on two newly constructed multi-view aesthetic object datasets (AestheticHome-12K and AestheticHome-2K) demonstrate that the proposed detector achieves 94.80 ± 0.32%/94.20 ± 0.37% mAP@0.5, 96.30 ± 0.28%/95.60 ± 0.31% precision, and 94.70 ± 0.35%/93.80 ± 0.39% recall across two datasets (reported as mean ± standard deviation of 5 independent training runs with distinct random seeds), with only 2.89 M parameters and 6.0 GFLOPs. Statistical significance is verified via paired two-tailed t-tests with Bonferroni correction (adjusted p < 0.05) for all performance comparisons against baseline models. Compared with the YOLOv11n baseline, the method improves mAP@0.5 by 1.87–2.09 percentage points and recall by 3.27–3.48 percentage points while reducing computational cost. Notably, it also achieves 79.2–80.5% mAP@0.5:0.95, outperforming the baseline by 4.7–4.9 percentage points, indicating significantly superior localization accuracy under stricter criteria. The proposed model achieves a remarkable balance between accuracy and efficiency, making it highly suitable for deployment on resource-constrained edge devices commonly used in digital design and home aesthetic monitoring systems. The results indicate that combining lightweight long-range feature extraction optimized for rigid aesthetic objects, compact attention-based feature interaction for interference suppression, and geometry-aware regression tailored for aesthetic targets provides an effective and efficient solution for robust aesthetic object detection in real-world computational aesthetics and digital interior design applications. Full article

Decolonising assessment remains an underdeveloped yet essential component of higher education transformation. In research, much attention has been assigned to decolonising content knowledge and pedagogy, with limited attention to assessment. Hence, this systematic literature review examines empirical studies published between 2018 and 2025 to explore how decolonial assessment is enacted in higher education. Using the PRISMA 2020 framework, 17 studies were analysed to identify theoretical foundations, methodological trends, dominant themes, key findings, and literature gaps to make recommendations for future research. The findings reveal six interconnected practices characterising decolonial assessment: humanising learning, contextualising assessment within students’ cultural and community realities, disrupting power hierarchies through co-creation, collaborative and peer-based approaches, diversifying modes of expression, and enhancing transparency and cultural responsiveness. These practices validate students’ lived experiences, foreground multiple epistemologies, and challenge traditional assessment norms, thereby promoting deeper engagement, agency, and belonging. However, the scholarship is constrained by a reliance on small-scale qualitative designs, limited theoretical development, and regional concentration, particularly in South Africa and the United States. The review concludes that while decolonial assessment offers transformative potential, its broader institutional uptake requires methodological pluralism, stronger theoretical frameworks, and policy-level support. The paper recommends embedding decolonial principles in assessment design, expanding research across diverse contexts, and strengthening lecturer capacity to develop equitable, inclusive, and contextually grounded assessment practices that advance epistemic justice in higher education. Full article

Rice leaf diseases significantly affect crop health and yield potential, creating a need for accurate and timely disease diagnosis. Many existing approaches still rely on single-stream feature extraction architectures, which may limit the ability to simultaneously capture global contextual information and fine-grained disease characteristics. Moreover, limited interpretability and decision-support capability hinder their practical deployment in real-world rice farming. To address these limitations, we employed a framework consists of two parallel feature extraction streams designed to capture different characteristics of disease patterns. The first stream uses a Swin Transformer to learn global contextual information and long-range spatial relationships across the leaf image. The second stream employs ConvNeXt to extract local texture features, including lesion details, spots, and color variations. By combining these complementary representations, the proposed framework effectively integrates global semantic information with local disease-specific features for improved classification performance. The extracted features from the two streams are fused through concatenation followed by an attention-based feature refinement module, enabling adaptive weighting of discriminative features. The refined representation is then used by a fully connected classifier for disease prediction. To enhance model interpretability, Grad-CAM visualization is incorporated to highlight disease-relevant regions and provide visual explanations for the model decisions. Furthermore, an LLM-based advisory module is integrated as a post-diagnosis decision-support component to provide contextualized disease management information and suggestions based on the predicted disease category. The generated suggestions are intended to support, rather than replace, expert agronomic recommendations and should be validated by agricultural professionals before practical application. The proposed framework was evaluated on two rice leaf disease datasets, achieving accuracies of 99.55% and 97.06% on Dataset-1 and Dataset-2, respectively, which are higher than those reported in previous studies. Additionally, five-fold cross-validation on Dataset-1 achieved an average accuracy of 98.91% ± 0.47, demonstrating the stability of the proposed approach. Cross-dataset evaluation using nine common disease classes across both datasets achieved 90.80% accuracy, indicating improved generalization across different data distributions. The proposed framework provides an accurate and explainable approach for rice leaf disease diagnosis in smart agriculture applications. Full article

Women’s concerns about personal safety significantly affect their use of public spaces, particularly semi-enclosed micro-public environments such as public toilets. Existing studies have not fully clarified how different spatial elements influence women’s perceived safety or how perceived-safety-oriented design priorities should be determined. This study develops and preliminarily examines an evaluation framework for women’s perceived safety in public toilets, which are treated as a highly private type of micro-public space. Grounded in Crime Prevention Through Environmental Design theory, the framework integrates visual scenario simulation, the Kano model, and the Best–Worst Method. A total of 257 female participants evaluated simulated spatial scenarios, and 10 experts assessed the relative importance of key indicators. The Kano results show that Facility condition is a must-be requirement; Lighting, Degree of enclosure, Material transparency, Signage clarity, Entrance buffer zone, Surveillance cameras, Hygiene conditions, and Graffiti and advertisements are one-dimensional requirements; and Women’s care area, Auditory conditions, Olfactory conditions, Loitering detection, and Emergency support are attractive requirements. The weighting results identify Degree of enclosure, Facility condition, and Hygiene conditions as the most important indicators. Based on these findings, the study proposes an integrated framework of “evaluation dimensions–demand attributes–relative weights–design priorities” and a design path of “basic guarantee–priority improvement–support enhancement.” The findings provide design guidance for public toilets and a methodological reference for other micro-public spaces, although the specific Kano classifications and weights require contextual recalibration. Full article

Automatic speech recognition (ASR) systems based on Conformer architectures achieve strong performance by jointly modeling local acoustic patterns and global contextual dependencies. However, their interleaved convolution–attention design leads to progressive entanglement of fine-grained acoustic features and global semantic representations, which weakens monotonic alignment in speech recognition and degrades performance in long utterances. To address this limitation, we propose an AS-Split Conformer–Mamba framework that decouples local and global modeling into two explicit stages. First, a stage-wise encoder is introduced, where a dedicated local modeling stage extracts phonetic-level acoustic features using SE-enhanced convolution, followed by a global modeling stage that captures long-range dependencies via multi-head self-attention and temporal convolution. Second, a Transition Fusion Block (TFB) is designed as an adaptive transition module that transforms local acoustic representations before they enter the global modeling stage. Third, intermediate CTC supervision is introduced to explicitly strengthen monotonic alignment at shallow representations. Finally, a hybrid Transformer–Mamba decoder is adopted, in which the Mamba block provides O(N) state-space computation within the replaced FFN sublayer while retaining Transformer attention mechanisms for acoustic–text alignment. Experiments conducted on AISHELL-1, THCHS-30, and ST-CMDS demonstrate that the proposed method achieves consistent improvements over strong baselines. On AISHELL-1, our model reduces Character Error Rate (CER) from 5.7% to 4.8% and Sentence Error Rate (SER) from 24.8% to 20.5%, while maintaining competitive computational efficiency. Full article

Background: MSC–hydrogel composite systems were developed to overcome the poor cell retention and limited durability of conventional marrow stimulation and suspension-based MSC delivery. A rigorous synthesis of the clinical evidence is lacking. Objectives: To evaluate the safety and efficacy of MSC–hydrogel composite therapy for focal knee cartilage defects and knee osteoarthritis, and to assess the certainty of evidence using the GRADE framework. Methods: PROSPERO-registered systematic review (CRD420261393525), conducted and reported per PRISMA 2020 and SWiM. Five databases (Embase, PubMed/MEDLINE, Cochrane CENTRAL, Scopus, Web of Science) were searched in May 2026 without date, language, or design restrictions. Adults receiving MSCs co-delivered in a hydrogel carrier for knee cartilage pathology were eligible. Risk of bias was assessed using RoB 2 (RCTs) and ROBINS-I (non-randomized studies). Narrative synthesis following Popay et al. was the primary method; GRADE certainty was assessed per outcome domain. Results: Ten studies (N = 521) were identified. Eight studies fulfilled the predefined eligibility criteria for MSC–hydrogel composite interventions and formed the primary evidence synthesis. Two additional studies were retained as contextual comparators because they evaluated either hydrogel-based therapy without MSC administration or MSC therapy without a structured hydrogel carrier. Surgical MSC–hydrogel implantation was associated with improvements in cartilage repair. Intra-articular injection without a hydrogel scaffold produced synovitis reduction but no detectable structural regeneration at six months. No serious treatment-related adverse events were recorded. Risk of bias was serious or critical in seven of nine assessable studies; GRADE certainty was low to very low across all outcome domains. Conclusions: MSC–hydrogel composite implantation may provide favorable safety signals and directionally positive effects on cartilage repair, pain, and function, with seven-year follow-up data suggesting a possible durability advantage over marrow stimulation. However, adverse-event reporting was inconsistent, and certainty of evidence remains low or very low because most studies were non-randomized, single-center, and concentrated around one commercial platform. These findings are relevant to international cartilage-regeneration research because they identify key methodological limitations and trial-design priorities for translating MSC–hydrogel systems across different clinical and regulatory settings. Full article