Abstract:

Leukocyte recruitment from blood into tissues involves sequential adhesive steps, including rolling and integrin-dependent arrest. The integrin VLA-4 is known to mediate firm adhesion, but can also support rolling. CD44–hyaluronan interactions have also been implicated in leukocyte rolling. Here, we used parallel-plate flow chamber assays to compare the contributions of CD44 and VLA-4 to monocyte rolling on different cellular monolayers. Monocytoid WEHI 78/24 cells rolled and adhered through CD44 on hyaluronan-presenting ECV304 monolayers, whereas VLA-4 dominated adhesion on endothelial monolayers expressing functional VCAM-1. Primary human monocytes showed similar CD44-dependent rolling on ECV304 monolayers. Blocking CD44, adding soluble hyaluronan, or removing surface hyaluronan with hyaluronidase reduced rolling and adhesion. These results show that CD44 can support monocyte rolling when VLA-4/VCAM-1 adhesion is not the dominant interaction. This cell-based flow model distinguishes CD44/hyaluronan-mediated rolling from VLA-4/VCAM-1-rolling and may help analyze monocyte rolling on hyaluronan, including tumor-derived monolayers.

Abstract: Stroke, a leading cause of global disability and mortality, exhibits significant spatiotemporal associations with environmental pollutants. Predicting daily stroke admissions becomes increasingly important as the population ages. Current prediction research on stroke-related medical services mainly relies on point prediction, which lacks the ability to quantify uncertainty. In this study, we try to develop parametric probability prediction models of stroke admissions based on machine learning and deep learning algorithms. We collected stroke data and environmental data from February 11, 2019 to May 26, 2023 in Chengdu, and employed prediction models encompass negative binomial regression, natural gradient boosting (NGBoost), long short-term memory networks (LSTM), and transformer. For performance assessment, mean absolute error (MAE) is used to evaluate point prediction accuracy, while continuous ranked probability score (CRPS) is applied to assess the quality of distribution fitting.We find that models with the ability to capture and process time-series information demonstrate greater advantages in probabilistic prediction, and among the four evaluated models, the transformer model proves to be the one that delivering more reliable and precise outcomes in both point prediction of admission counts and distribution fitting performance. This probabilistic forecasting approach provides robust evidence-based decision support for healthcare administrators to optimize resource allocation and staffing arrangements, and ultimately helps elevate the quality of medical care for stroke patients.

Abstract: Point-of-care testing (POCT) for cardiac troponin is increasingly used to support rapid clinical decision-making, particularly in resource-limited settings. However, while many central laboratories, including those in Sri Lanka, now use high-sensitivity cardiac troponin I (hs-cTnI) assays, commonly available POCT platforms continue to use conventional methodologies with different analytical characteristics and decision thresholds. We evaluated the analytical agreement and clinical concordance of two POCT troponin assays against a central laboratory hs-cTnI assay using paired samples obtained during routine clinical care in a Sri Lankan hospital. Both POCT devices demonstrated strong correlation with the laboratory assay (Pearson r ≈ 0.90). However, Bland–Altman analysis showed substantial positive bias and wide limits of agreement, indicating poor interchangeability at the individual sample level, with proportional bias observed in one device. Clinically relevant discordance was also identified, with 26.9% and 30.4% of samples classified as negative by POCT despite being positive by the reference assay. Regression-based recalibration did not significantly improve concordance. These findings highlight that strong correlation does not ensure diagnostic agreement, emphasizing the need for local validation before integrating POCT troponin assays into established hs-cTnI diagnostic pathways.

Abstract: Rare diseases, though individually uncommon, collectively represent a major global health challenge, affecting millions worldwide and increasingly recognized in India as a significant contributor to pediatric and adult morbidity. Cystic fibrosis (CF), a multisystem autosomal recessive disorder caused by pathogenic variants in the Cystic Fibrosis Transmembrane Conductance Regulator (\textit{CFTR}) gene, exemplifies this burden, with delayed diagnosis and diverse mutational spectra complicating clinical management in South Asian populations. To advance rare disease genomics, quantitative analysis of CFTR sequences across multiple species is essential, as evolutionary conservation highlights residues and motifs critical for channel function, while divergence reveals lineage-specific adaptations relevant to disease mechanisms. In the present study, we performed integrative analyses encompassing amino acid composition, sequence homology, frequency-dominant residue patterns, hydropathy-based n‑gram distributions, hydropathy profile continuity, and intrinsic disorder architectures across various CFTR sequences from multiple species. The quantitative signatures derived from amino acid composition, sequence homology, hydropathy-based n‑grams, hydropathy profiles, and intrinsic disorder analyses carry significant translational impact, as they provide a unified framework for identifying conserved motifs, resolving disorder-prone domains, and guiding the precise mapping of pathogenic mutations and their functional consequences. Collectively, our findings demonstrate how cross-species quantitative protein analysis of CFTR bridges evolutionary biology with clinical investigation, providing translational insights that strengthen rare disease research and therapeutic development in cystic fibrosis.

Abstract: Accurate fault location is essential for rapid service restoration in distribution networks. How-ever, modern active distribution networks (ADNs) with high penetration of distributed energy resources (DERs) challenge conventional methods through multi-source fault contributions, bi-directional power flows, and converter-limited fault currents. This paper presents a time-domain fault-location method for both passive distribution networks (PDNs) and ADNs, based on a three-sample apparent-inductance estimator that uses local voltage and current measurements. The estimator exploits the strong correlation between line inductance and fault distance, with reduced sensitivity to fault resistance compared to classical impedance approaches. Its performance is evaluated on a 22 kV, 20 km distribution feeder, covering three fault types, four fault resistance levels (5–500 Ω), four fault locations, EN 50160 standard-compliant harmonic distortion, and DER penetration levels from 0 to 80%. Under ideal sinusoidal conditions, relative location errors remain below 2% for low-resistance faults. In ADNs, the method achieves errors be-low 5% for low-resistance faults across all fault types, with accuracy decreasing for high-resistance faults at high DER penetration. A sensitivity analysis confirms practical robustness across SNR, load current, THD, and DER penetration.

Abstract: Most AI-for-science systems (agents4science) are evaluated as task-specific automation rather than persistent work environments. This leaves an important and pragmatic systems-question unresolved: what infrastructure enables scientific agents to be trustworthy, steerable, and reproducible? We present ApexClaw, a persistent workspace for human-supervised AI agents in scientific discovery. The system provisions isolated Linux environments with scientific computing libraries, a browser-based IDE, and a registry of reusable scientific skills. Rather than pursuing full autonomy, ApexClaw emphasizes human oversight through an interface that exposes agent reasoning, tool calls, and workspace files, allowing scientists to guide agents at key junctures. We validate this approach with telemetry from 81 users operating 243 workspaces across 252 conversations (March--May 2026): 87\% of users reconnected to existing workspaces, and humans intervened selectively on 3.4\% of interactions. These findings demonstrate that hybrid autonomy with persistent context is ordinary practice in real agentic science, and that durable workspaces, explicit artifacts, and auditable traces are necessary infrastructure for reproducible and steerable AI scientists.

Abstract: Background: Neuroinflammation contributes to etiopathology and symptom severity in neurodegenerative and neuropsychiatric disorders. Glial cells, especially microglia and astrocytes, play a crucial role in neuroinflammation. It has been reported that ginseng and its bioactive component ginsenoside Rg1 exhibit anti-inflammatory effects and improve cognitive performance in various models. However, the exact underlying mechanisms remain unclear. Methods: An astrocyte-microglia co-culture model simulating physiological (M5, 5-10% microglia) and pathological/inflammatory (M30, 30-40% microglia) conditions was treated with different concentrations of ginsenoside Rg1 (15, 30, 45 µM), ginseng extract (derived from Korean red ginseng) at low-dose (12.5, 25, 37.5 µg/ml) or high-dose (125, 250, 375 µg/ml) for 24 hours. Cell viability was assessed by MTT assay. Microglial reactivity was examined by immunocytochemistry. Astrocytic gap-junctional coupling was investigated using scrape loading method and connexin 43 (Cx43) expression was analyzed by immunocytochemistry and Western blot. Results: Both Rg1 and low-dose ginseng extract reduced microglial activation under inflammatory conditions by promoting a phenotypic shift from activated to homeostatic (resting) microglia. Rg1 preserved astrocytic gap-junctional function by preventing the inflammation-induced downregulation of Cx43 expression and enhancing Cx43-mediated gap-junctional intercellular communication. Rg1 caused a significant reduction of glial cell viability only at high concentrations (30 and 45 µM) under inflammatory conditions. High-dose ginseng extract showed significant concentration-dependent cytotoxicity, reducing glial cell viability under physiological and pathological conditions, without comparable anti-inflammatory benefits. Conclusions: This study suggests that low-dose ginseng and its active compound Rg1 exert anti-inflammatory effects by modulation of astrocytic coupling and microglial reactivity. These results provide a novel therapeutic perspective for ginseng in the treatment of neurodegenerative and neuropsychiatric diseases related to neuroinflammation.

Abstract: Hydrophilic surface modification is widely investigated as a strategy to improve the hemocompatibility of small-diameter vascular grafts. We previously developed a polyvinyl alcohol-coated poly-ε-caprolactone nanofiber graft (PVA–PCL graft) and showed that PVA coating improved graft hydrophilicity and mechanical properties. However, whether PVA coating provides an in vivo advantage over uncoated PCL grafts remains unclear. In addition, the influence of pre-implantation surface hydration on the function of hydrophilic grafts has not been fully examined. In this study, we first compared PVA–PCL and uncoated PCL grafts in a rat abdominal aorta implantation model and found no statistically significant difference in patency rate between the graft types. We then examined whether pre-wetting enhanced the anti-fouling function of the PVA coating. In vitro whole-blood flushing assays demonstrated that pre-wetting markedly reduced blood component deposition on PVA–PCL grafts. However, this effect did not translate into detectable improvements in patency or tissue regeneration in the rat model used in this study. These findings indicate that pre-wetting effectively enhances the in vitro anti-fouling behavior of PVA–PCL grafts and may serve as a simple strategy to optimize the functional surface state of hydrophilic coatings. Further studies are needed to determine whether this in vitro improvement can be translated into meaningful enhancement of graft performance in vivo.

Abstract: Reliable public-sector labour-market forecasting requires models that can be updated as data sources, AI tools, and labour-market signals evolve. This paper proposes a provider-independent multi-agent framework for dynamic predictive evaluation of national and regional labour markets in Bulgaria. Implemented as a Model Context Protocol (MCP) server, the system coordinates specialised agents for data ingestion, preprocessing, semantic extraction, AI-adjusted transformation modelling, automated model evaluation, and reporting through stable input-output contracts. The empirical application integrates Bulgarian Employment Agency administrative registered-unemployment indicators, Eurostat labour-market data, World Bank macroeconomic data, and text/audio/video evidence on AI, skills, and employment change. The study period covers 2015–2030, combining observed official inputs for 2015–2025 with forecast/scenario outputs for 2026–2030. For youth unemployment under 25, the semantic-enhanced model achieves the best predictive accuracy (RMSE = 0.2033; MAE = 0.1457), representing a small improvement over the structured baseline (RMSE = 0.2057; MAE = 0.1462) and a substantial RMSE reduction relative to the persistence benchmark (RMSE = 0.4750; MAE = 0.2891). Regional forecasts indicate persistent spatial inequality, with the Northwest remaining the highest-risk region and the Southwest the lowest-risk region.

Abstract: Localized lower respiratory tract infection including unilobar and round pneumonia can be associated with hypoxia and oxygen requirements. Previous explanations include shunting of deoxygenated blood, a systemic inflammatory response syndrome and vasoconstriction.This is unexplained.The alternative hypothesis is that spread of fluid absorption inhibiting cytokines in the alveolar spaces of the inflamed lung is cause of hypoxia in localized lower respiratory tract infection by spread of Cystic Fibrosis Transmembrane Conductance (CFTR) dysfunction in alveolar epithelial cells to more areas including those not infected. There is no evidence of pulmonary shunting to explain hypoxia in localized pneumonia. Systemic inflammatory response syndrome (SIRS) related generalized increase in alveolar capillary barrier or pulmonary vasoconstriction not visible on a chest x-ray cannot explain the hypoxia detected in most patients. Confirmation of the hypothesis could be achieved using pulmonary MRI or high resolution CT to confirm spread of alveolar fluid accumulation from the localized pneumonia focus as opposed to generalized SIRS related pulmonary oedema together with cytokine and chloride measurement in bronchoalveolar lavage samples from the lung segments near the affected lung segment and unaffected contralateral lung. Ventilation/perfusion scintigraphy could investigate for involvement of vasoconstriction or micro-emboli from intravascular coagulation.Should the posed hypothesis be confirmed adjuvant strategies including small molecule CFTR activators, CFTR activating combination of beta-agonists, phosphodiesterase inhibitors and steroids could be used to treat hypoxia and CFTR activating low-intensity ultrasound explored.

Abstract: The competitive ground of digital visibility has moved twice in eighteen months. The first move was from the ten blue links to the AI-generated answer; the second, still in progress, is from the answer to the autonomous action. This article develops a framework that integrates three optimization disciplines emerging in response to that shift—Answer Engine Optimization (AEO), Generative Engine Optimization (GEO), and Agentic Optimization (AgO)—under a single theoretical lens: delegated consumer–AI agency. Building on Puntoni et al.’s (2021) experiential perspective on consumer AI and Davenport et al.’s (2020) account of AI in marketing, we treat the AI assistant not as a channel but as a delegated decision-maker whose choices are governedby retrievability, attribution cost, and embedded brand associations. We argue that the central strategic risk of this transition is brand erasure—the systematic elimination of brand identity from synthesized answers and completed agentic actions—and we develop six propositions that link content and infrastructure choices to brand outcomes in this environment. We close with managerial implications for brands operating across the dual audience of human readers andmachine intermediaries, and a research agenda focused on measurement, audit methods, and governance.

Abstract: Poxviruses are large double-stranded DNA (dsDNA) viruses that cause important human and animal diseases, including smallpox and mpox. Poxviruses have also been identified in diverse bat populations; however, their potential for zoonotic transmission and adaptation to other mammalian hosts remains poorly understood. Poxviruses encode numerous immunomodulatory proteins that contribute to virulence, immune evasion, and host range. In this study, we performed a comparative genomic analysis of two bat-associated poxviruses belonging to the genus Vespertilionpoxvirus: hypsugopox virus (HYPV) and eptesipox virus (EPTV). We identified 24 and three previously unannotated putative open reading frames (ORFs) in HYPV, and EPTV, respectively, substantially expanding the predicted coding capacity of these viruses. Comparative analyses further revealed gene duplication and fragmentation events affecting several virulence and host range factors, as well as other unusual genomic features, including the presence of two divergent E3L homologs in EPTV. Together, our findings provide new insights into the genome evolution and potential host adaptation of bat-associated poxviruses and establish a foundation for future functional studies of Vespertilionpoxvirus biology, host-virus interactions, and zoonotic potential.

Abstract: In pregnancy physiologically based pharmacokinetic (PBPK) modelling within PK-Sim/Mobi, pregnancy stage is canonically parameterised on the fertilisation-age (FA) timescale, yet PK-Sim’s physiology database is indexed by chronological “Age”. The Open Systems Pharmacology (OSP) pregnancy framework therefore encodes FA on a dummy “Age” axis (30.00 years = FA 0 weeks; 30.75 years = FA 38 weeks) and generates pregnancy physiology vectors from FA 0 – 38 weeks discretised at 1-day intervals for database ingestion. Although the anchor points are publicly documented, an explicit closed-form transform and week-resolved lookup suitable for deterministic reproduction of daily physiology grids has not been routinely foregrounded in the literature or repository instructions. A unique affine mapping is derived, implied by the anchors, that provide forward and inverse equations in week- and day-space, quantifying rounding error in terms of FA-day misalignment, and supply a fertilization week (1-38) table for implementation-ready pregnancy virtual population creation in PK-Sim.

Abstract: Metabolomics is a powerful approach for monitoring metabolic effects in a particular situation by qualitatively or quantitatively analyzing metabolites related to specific physiological or pathological responses within a biological process. Rapeseed is a major source of vegetable oil and contains a wide variety of metabolites. Recent advances, particularly the integration of metabolomics with other omics approaches, now allow us not only to obtain a comprehensive overview but also to perform detailed analyses of key metabolites that respond to specific conditions. In this review, we summarize recent progresses in rapeseed metabolomics study, introduce the key metabolites uncovered by this approach, and discuss those associated with growth & development, and abiotic and biotic stresses, including macronutrient availability, temperature, water stress, salt stress, and cadmium toxicity. Future perspectives and current challenges in metabolomics are also discussed, along with its potential for breeding applications aimed at developing new rapeseed varieties with stable, high-yield, and high-quality traits.

Abstract: We introduce the Information Lattice Model (ILM), a theoretical framework in the braneworld tradition in which the brane–bulk system is represented as a stratified informational graph whose inter-layer link capacity is governed by a permeability function \mathrm{\Pi}\left(y\right)=1+\varepsilonexp\left(-\lambda\left|y\right|/\ell_{Pl}\right). The dimensionless parameter \varepsilon is identified with the ratio of bulk-to-brane entanglement entropy flux via the Ryu–Takayanagi formula, connecting the model to the Bekenstein–Hawking entropy bound and the AdS/CFT correspondence. This identification is presented as a conceptual proposal ahead of full formal derivation, in the tradition of framework papers from Kaluza–Klein to Verlinde’s entropic gravity. The permeability function modifies the Randall–Sundrum warp factor, introducing an additional entropic dilution of the effective gravitational coupling beyond geometric warping alone. Within this framework, the ILM suggests a unified phenomenological perspective on the hierarchy problem, the black hole information paradox, and the possible role of sub-brane entropic degrees of freedom as an effective dark matter component. The model also predicts frequency-dependent corrections to gravitational-wave propagation arising from bulk-mediated entropic coupling. The model generates two testable predictions. The primary prediction is a positive correlation between the dimensionless radiated energy excess \mathrm{\Delta A} and total merger mass M_{total} in binary black hole systems, arising from the mass dependence of the integrated horizon permeability. Under a quadratic area-law scaling, \mathrm{\Delta A} may approach \sim10-3 for M_{total}\sim{10}^6\thinsp M_\odot events accessible to LISA, while the sign and slope of the \mathrm{\Delta A}–M_{total} correlation is independently falsifiable via stacked regression analysis of the GWTC-3 catalogue; a full Bayesian treatment will be presented in future work. The secondary prediction is a non-linear deviation in quantum decoherence rates near lattice saturation, potentially testable with ion traps, superconducting qubits, and Bose–Einstein condensates.

Abstract: Background/Objectives: Artificial intelligence (AI) is integrated into diagnostic, thera-peutic, administrative, and communicative healthcare domains in Italy under regulations requiring human oversight. Empirical evidence on AI attitudes, acceptance, and per-ceptions in Italian healthcare is rapidly accumulating but not systematically mapped. This scoping review aimed to (i) map empirical evidence on AI attitudes, acceptance, and perceptions in Italy by population and domain; (ii) identify measurement instruments used in studies and their origins; and (iii) characterize determinants, themes, and methodological gaps in the Italian evidence base. Methods: The review used Joanna Briggs Institute methodology, reported via PRISMA-ScR (protocol Open Science Framework doi: 10.17605/OSF.IO/TZRVF). PubMed and Embase were searched on 27 April 2026 from January 2018 in English, Italian, or German, combining controlled vo-cabulary and free-text terms across AI, attitudes-acceptance, and healthcare delivery, with an Italian-context qualifier. Eligibility criteria used the Population–Concept–Context mnemonic. Results: Of 1,510 unique records screened, 35 empirical studies were retained, comprising seven studies of Italian patients and the general population, 22 studies of healthcare professionals, three psychometric validation studies of AI-acceptance instru-ments, one mixed-population study and two international comparator studies with sub-stantial Italian sub-samples. Acceptance was consistently positive but conditional on physician oversight, training and regulatory clarity. A recurrent optimism–knowledge gap and an absence of probabilistic, population-representative evidence were identified as principal gaps. Conclusions: Italian evidence on AI attitudes is expanding but methodologically narrow. Three Italian-validated acceptance instruments are now available. Population-representative, multilingual and longitudinal evidence is required.

Abstract: Maintaining food systems in the face of climate change (CC) is a major concern. Palm oil is included in many commodities and this food system will be affected detri-mentally by inclement future climate, when oil palm (OP) will experience increased disease and declining growth. The OP diseases considered are basal stem rot (BSR), bud rot (BR) and fusarium wilt (FW), where the attempts to control them have been un-successful. An approach may be to replace compromised OP with different crops better suited to future climate: These plants will have less disease because of the “Parasites Lost” phenomenon. Maintaining a vegetable oil product is an important advantage. How CC will affect OP and associated ailments has been determined previously by CLIMEX modelling. The modelling of future suitable climate (FSC) has also been car-ried out for soybean, maize and the common bean (CB) using the same modelling pa-rameters. This enables direct comparisons in the OP producing countries of Colombia, Nigeria and Papua New Guinea (PNG). Limited data for rapeseed are also discussed. The FSC for OP was much reduced in these countries and that for soybean was higher. Soybeans will have less disease as it would be an introduced and annual crop. Maize had much fewer advantages and the CB and rapeseed had none. Maize had potential advantages in Nigeria until 2050. A novel method for adapting to the serious diseases of OP and poor growth would be to grow soybeans in similar regions to where OP grows currently. Plans could be made for replacing OP with soybeans which could be modified when real time data becomes available. The paper provides a novel method for mitigating future diseases and poor growth of OP, which are otherwise unavailable, whilst maintaining a valuable oil product.

Abstract: Fourier Neural Operators have become a central tool for learning solution operators of partial differential equations, but their spectral layers remain entirely classical and rely on digital Fourier processing. In this work, we introduce the Continuous-Variable Quantum Fourier Neural Operator (CV-QFNO), a Gaussian photonic formulation of the FNO spectral layer. The proposed architecture maps the essential operations of Fourier-domain operator learning, Fourier transformation, mode selection, and channel mixing, onto native continuous-variable optical primitives. In this way, CV-QFNO provides a photonic quantum analogue of the truncated spectral mechanism underlying the classical FNO, while avoiding the compilation overhead and spectral mismatch that arise in qubit-based Quantum FNO constructions. We extend the framework to both one- and two-dimensional operator learning and validate it on standard PDE benchmarks, including Burgers’ equation, heat equation, Navier–Stokes dynamics, and Darcy flow. The results show that the proposed model preserves the predictive accuracy, resolution generalisation, and spectral inductive bias of Fourier neural operators while using a structurally constrained photonic parameterisation. Since all experiments are performed as classical simulations, the contribution should be understood as an architectural and algorithmic blueprint for photonic neural operators, rather than as a demonstration of quantum computational advantage.

Abstract: Background: Gastrectomy for gastric cancer is associated with substantial metabolic, nutritional, and immunological disturbances that may significantly influence postoperative recovery. Increasing evidence suggests that perioperative immunonutritional status, particularly as assessed by the Controlling Nutritional Status (CONUT) score, represents an important predictor of surgical outcomes. However, prospective data evaluating sex-related differences in postoperative nutritional recovery after gastrectomy remain limited. Methods: This prospective observational cohort study included 150 consecutive patients undergoing curative-intent gastrectomy for gastric adenocarcinoma at a tertiary referral center between 2021 and 2024. Nutritional and immune status were longitudinally assessed using the CONUT score at predefined perioperative timepoints: preoperatively (T0), early postoperatively (T1), and at 3-month follow-up (T3). Functional recovery outcomes, postoperative complications, and mid-term functional parameters were compared between male and female patients. Multivariable logistic regression analysis was performed to identify independent predictors of delayed postoperative recovery. Results: The study population included 91 male patients (60.7%) and 59 female patients (39.3%). Significant postoperative deterioration of albumin level, lymphocyte count, total cholesterol, and CONUT score was observed in the entire cohort (p-time < 0.001 for all comparisons), followed by partial recovery during follow-up. No significant sex-related differences were identified regarding longitudinal immunonutritional evolution, postoperative complications, gastrointestinal recovery, or functional outcomes (p > 0.05). Overall postoperative complications occurred in 31.3% of patients, while 90-day mortality was 2.7%. Elevated baseline CONUT score ≥5 (OR 2.74, 95% CI 1.48–5.09, p = 0.001), postoperative CONUT score T1 ≥5 (OR 3.36, 95% CI 1.82–6.19, p < 0.001), ASA class III (OR 2.08, 95% CI 1.19–3.63, p = 0.010), and anastomotic leakage (OR 4.91, 95% CI 1.74–13.88, p = 0.003) independently predicted delayed functional recovery. Male sex was not independently associated with adverse postoperative recovery (OR 1.18, 95% CI 0.74–1.89, p = 0.44). Conclusions: Gastrectomy induces significant postoperative immunonutritional deterioration irrespective of sex. Although biological sex did not independently influence postoperative recovery trajectories, impaired perioperative immunonutritional status—particularly elevated postoperative CONUT score—was strongly associated with delayed functional recovery. Serial perioperative CONUT assessment may represent a valuable tool for individualized postoperative risk stratification and nutritional management in gastric cancer patients undergoing gastrectomy.

Abstract: Chronic wounds are a persistent clinical and public health challenge. Natural polyphenols such as curcumin and resveratrol, alongside mesenchymal stem cell (MSC) secretome, have demonstrated complementary anti-inflammatory, antioxidant, and pro-angiogenic properties with potential for wound healing. This study reports two complementary in vitro investigations evaluating the release profiles of curcumin and resveratrol from two polymeric platforms: poly(vinyl alcohol)/sodium alginate/carboxymethylcellulose films (Study 1) and an acrylate copolymer-based hydrogel incorporating MSC secretome (Study 2). UV-Vis spectrophotometric analysis confirmed analytical selectivity with no interference from excipients. In Study 1, films containing curcumin alone exhibited low structural stability and early disintegration in aqueous medium, whereas resveratrol-only films (2% w/w) demonstrated sustained and reproducible release profiles. Combined formulations showed that curcumin compromised polymer matrix integrity and reduced resveratrol release efficiency. In Study 2, resveratrol exhibited progressive and consistent release from the hydrogel, reaching 14.31 µg/mL (isolated control) and 12.60 µg/mL (combined with curcumin) at 120 min. Curcumin showed unsatisfactory release in both systems, attributed to its low aqueous solubility. These results support resveratrol-loaded polymeric matrices as promising sustained-release platforms for bioactive wound dressings and highlight the need for nanoencapsulation strategies to improve curcumin bioavailability.