Browse Articles

Multimodal deep learning model for AI-based functional prognostic risk stratification in patients undergoing radical nephrectomy

Burden and patterns of dyslipidaemia among adult Ghanaians: A systematic review

Background Dyslipidaemia is a major modifiable risk factor for cardiovascular disease and has become an increasing public health concern in sub-Saharan Africa. Rapid urbanisation, dietary transitions, and sedentary lifestyles have contributed to the rising incidence of cardiometabolic diseases in Ghana. However, evidence on the overall burden and patterns of dyslipidaemia in adults remains unclear. This systematic review aimed to synthesise the evidence on the prevalence and lipid profile patterns of adult Ghanaians. Methods A comprehensive search of PubMed, Scopus, Web of Science, African Journals Online, Africa-Wide Information and African Index Medicus databases was conducted for studies published between January 1980 and November 2025. Eligible studies reported the prevalence or distribution of dyslipidaemia in Ghanaian adults (≥ 18 years). Two reviewers independently screened, extracted, and assessed the quality of included studies. Findings were synthesised narratively and summarised in descriptive tables and figures, due to heterogeneity in study designs and data reporting. Results Twenty-four studies published between 2003 and 2023, comprising approximately 11,400 participants, met the inclusion criteria. Across the included studies, the reported prevalence of dyslipidaemia ranged from 3.0% to 72.4%, reflecting differences in study populations, diagnostic criteria and study settings. Low high-density lipoprotein cholesterol (HDL-C) was the most frequent abnormality, followed by elevated total and low-density lipoprotein cholesterol (LDL-C), whereas hypertriglyceridaemia was the least common. Studies from southern and middle regions—particularly Ashanti, Brong-Ahafo, and Greater Accra—reported a higher dyslipidaemia prevalence than those from northern Ghana. Considerable variation in diagnostic criteria, study populations and sampling strategies limited comparability and precluded meta-analysis. Conclusion Dyslipidaemia is common among adult Ghanaians, with low HDL-C emerging as the predominant abnormality. Methodological differences and sampling biases limit the precise estimation of the national burden, but the available evidence indicates a growing cardiometabolic risk. Future studies should prioritise community-based sampling and context-appropriate lipid thresholds to support evidence-based cardiovascular risk-reduction strategies.

Retraction: A study of the economic growth effects of market integration: An examination of 27 cities in the Yangtze River Delta city cluster

Retraction: Optimum 3D Matrix Stiffness for Maintenance of Cancer Stem Cells Is Dependent on Tissue Origin of Cancer Cells

Bee-inspired navigation robot pinpoints its home using a neural network

A scoring system for euthanasia decisions in rescued raccoon dogs (Nyctereutes procyonoides): Standardized criteria in wildlife rehabilitation

Wild animal rescue and control centers routinely encounter situations that require evaluating a rescued animal’s potential for recovery and making decisions on appropriate treatment, rehabilitation, or euthanasia. However, currently, scientific standards to guide these decisions are lacking, and frequent reliance on empirical judgment leads to application of subjective criteria for euthanasia. Therefore, this study aimed to analyze factors affecting rescue outcomes in 1,376 raccoon dogs rescued by the Chungnam Wild Animal Rescue Center over 13 years, to predict the likelihood of release, and to develop a system to inform euthanasia decisions. The model that best predicted the release of rescued raccoon dogs included the cause of the accident, season, and level of injury. Based on this, a scored assessment support system was constructed, and raccoon dogs were classified into a ‘low-risk group,’ requiring active treatment and rehabilitation owing to the high likelihood of live release, a ‘medium-risk group,’ requiring mid-to-long-term euthanasia decisions through close monitoring and re-assessment, and a ‘high-risk’ group, for which immediate euthanasia may be considered because of the extremely low probability of release. The proposed scoring-based euthanasia decision support system integrates clinical and situational factors and remains practical by relying on data that can be readily obtained in the field immediately after rescue. Therefore, by supporting rapid, consistent decision-making and using basic data to prepare standardized euthanasia guidelines, this system can improve animal welfare and operational efficiency.

Exploring the impact of specialist and generalist stars on organizational performance

We analyze the impact of specialist and generalist stars on organizational performance by linking the literature on star employees and specialization. Our adaptive role framework allows stars to switch roles over time. We develop a game-theoretical model to derive hypotheses and to empirically test them using basketball data. Consistent with theoretical predictions, the results show that teams with generalist stars outperform those with specialists. While a generalist switching to a specialist role always worsens performance, a specialist switching to a generalist role may improve performance. Additional estimations suggest that these findings hold for teams relying on a unique star rather than multiple stars.

Who Leads, What Matters? Machine Learning and the Complexity of University Performance

The role of leadership in public institutions, particularly universities, is often linked to goal-setting and decision-making processes that impact efficiency. In Spain, public university rectors are directly elected by academics, staff, and students, offering a unique context for studying leadership influence. This study uses a unique database to analyze Spanish public universities across five categories: academic and research performance, social objectives, internationalization, university characteristics, and rector profiles. Using a K-Means unsupervised machine learning algorithm, we identify five distinct clusters of Spanish public universities, each characterised by a specific combination of institutional performance indicators and management characteristics.

Expression of Concern: Soft computing models to predict the compressive strength of GGBS/FA- geopolymer concrete

Correction: Soil microbial communities in the face of changing farming practices: A case study in an agricultural landscape in France

Machine learning-based prediction of hepatocellular carcinoma risk in steatotic liver disease: A nationwide cohort study

Background and aims Steatotic liver disease (SLD) has emerged as an important risk factor for hepatocellular carcinoma (HCC), often in the absence of cirrhosis. We aimed to develop explainable machine learning (ML) models to predict HCC risk in individuals with SLD using routinely collected screening data. Methods Using the Korean National Health Insurance Service database, we included adults aged 20–79 years who underwent national health screening in 2017. SLD was defined as a fatty liver index (FLI) ≥ 60. Multiple ML algorithms, including deep learning models, were trained using a 7:3 train–test split with repeated non-replacement undersampling at a 1:3 case-to-control ratio to address extreme class imbalance. Results Among 1,241,560 adults with SLD, 2,152 (0.17%) developed HCC during a 6-year follow-up period. In the internal validation cohort, the final weighted multi-head attention deep neural network ensemble achieved an area under the receiver operating characteristic curve of 0.923, with a sensitivity of 71.36% and specificity of 93.65%. SHapley Additive exPlanations consistently identified age, sex, triglycerides, total cholesterol, aminotransferases, gamma-glutamyl transferase (GGT), Charlson Comorbidity Index, and FLI as key contributors to HCC risk. In multivariable Cox models, older age, male sex, elevated GGT, higher aspartate aminotransferase and FLI, and greater comorbidity burden were positively associated with HCC risk, whereas higher triglyceride and total cholesterol levels were inversely associated. Model-based risk stratification identified four groups with distinct HCC-free survival curves; the extremely high-risk group had an approximately 74.9-fold higher hazard of HCC than the low-risk group (95% CI, 55.3–101.5). Conclusions Overall, this explainable ML model based on routine health screening variables enables robust HCC risk stratification and may help inform future targeted surveillance strategies in SLD populations after external validation.

A comparison of capturing speech breathing using respiratory inductive plethysmography and electromagnetic articulography

This study introduces a novel approach to speech breathing analysis by evaluating Electromagnetic Articulography (EMA) for tracking speech breathing patterns and thus providing an alternative to the commonly used Respiratory Inductive Plethysmography (RIP). We recorded speech production data from 18 speakers (9 F, 9 M), aged between 23 and 54 years, producing sentences that differ in length and syntactic complexity. Multiple EMA sensor configurations were evaluated to determine which measures best reflect the RIP signal. For the purposes of this study, the analysis was restricted exclusively to thoracic (chest) movements for both EMA and RIP. The results demonstrate that EMA reliably captures key respiratory events in speech, including inhalation and exhalation duration as well as full breathing cycle trajectories. Horizontal EMA distance measures, particularly those derived from a sternum or middle rib cage sensor combined with a sensor on the lowest cervical vertebra, showed the highest correspondence with RIP across temporal and spatial dimensions. Distance measures of these EMA sensors exhibited strong correlations with RIP signals and reproduced comparable speech breathing patterns in the temporal and spatial domain. Vertical EMA measures were more variable, but did not compromise the overall detection of speech breathing intervals, such as inhalation and exhalation phases. In sum, these findings highlight that EMA can be an effective method for the kinematic analysis of articulatory and speech breathing data. The proposed two-sensor setup is easy to apply and provides robust speech breathing data without additional instrumentation. This approach offers a practical path for incorporating speech breathing assessment into articulatory research and has the potential to advance our understanding of speech production in linguistic and clinical contexts.

Genomic and functional characterization of a novel halophilic bacteriophage targeting carbapenem-resistant Klebsiella pneumoniae

Carbapenem-resistant Klebsiella pneumoniae (CRKP) is a multidrug-resistant (MDR) pathogen causing severe infections in immunocompromised patients, prompting the exploration of alternative therapies like bacteriophage therapy. In this study, we isolated and characterized a novel halophilic lytic bacteriophage, Halo KS-7, targeting K. pneumoniae , and used an AI-driven annotation pipeline in Python to analyze its genome and therapeutic potential. Bacteriophages were isolated from Hospital wastewater, purified through plaque isolation, and confirmed using the double-layer agar method. Morphological analysis via transmission electron microscopy (TEM) and plaque assays assessed lytic activity. In vitro assays, including one‑step growth curve and MOI determination, were performed to evaluate the replication kinetics and lytic activity of bacteriophage Halo KS‑7 against carbapenem‑resistant Klebsiella pneumoniae. In vivo efficacy was assessed using a BALB/c mouse wound infection model by monitoring wound contraction and performing blinded histopathological analysis following phage treatment. DNA sequencing was done using Illumina HiSeq 2000, followed by genome assembly, AI-guided annotation, gene prediction, protein function classification, and comparative genomics using CLC Genomics Workbench. We also evaluated host range, temperature stability, pH sensitivity, and salt stress tolerance to assess therapeutic potential. Halo KS-7 exhibited strong lytic activity against CRKP and was classified as a Myoviridae bacteriophage by TEM. Phenotypic assays demonstrated optimal activity at 37 °C and neutral pH, effective activity from pH 4–10, and enhanced performance in high-salinity conditions. Bacteriophage Halo KS-7 exhibited a short latent period (~20 min), a modest burst size (5.73 PFU/cell), and optimal antibacterial activity at MOI 0.1, resulting in sustained suppression of K. pneumoniae growth in vitro . In vivo , Halo KS-7 treatment significantly enhanced wound healing in infected BALB/c mice, achieving near-complete wound closure, effective infection control, and improved histopathological regeneration comparable to uninfected controls. Halo KS-7 have 58.716 kb linear dsDNA genome (44.4% G + C), contains 49 predicted ORFs, lacks integrase, lysogeny, or antibiotic-resistance genes, and includes three tRNA genes (tRNATyr, tRNAPro, and tRNAAsn). It also includes a toxin gene and auxiliary factors like MazG, pyrophosphatase, and HNH endonucleases that enhance bacterial killing without promoting horizontal gene transfer or resistance. Functional annotation assigned ~65% of ORFs to structural, replication, and packaging roles. Comparative genomics showed moderate similarity to other Myoviridae but with distinct accessory features, emphasizing its novelty and therapeutic value. Halo KS-7 is a novel, strictly lytic bacteriophage with strong antibacterial activity and stress resilience, supporting its use as a promising biocontrol agent against CRKP and its potential for clinical development in managing MDR infections.

Research on an islanding detection method suitable for distributed generation grid-connection complex system

With the increase in penetration rate of distributed generation under dual-carbon goals, hazards of unplanned islanding caused by grid-connection of distributed generation systems have attracted attention. Due to the dilution effect of characteristic parameters in multi-machine grid-connected systems, active and passive hybrid islanding detection is prone to have islanding detection blind zones. Therefore, an islanding detection method suitable for complex multi-machine grid-connected systems of distributed generation is proposed. To verify the feasibility of the proposed islanding detection method, a simulation model of the distributed generation grid-connected system was built in MATLAB/Simulink based on IEC 61850-7-420 standard. Two different operating conditions of islanding and grid-connection, were simulated to obtain voltage and current waveforms of each node, which were converted into time series. Then, time-frequency spectrum analysis and calculation were performed based on Short-Time Fourier Transform (STFT) to generate time-frequency spectrum. After feature extraction, training iterations were carried out using Convolutional Neural Network (CNN) algorithm to obtain the optimal islanding detection model. Simulation results show that the proposed method achieves a detection accuracy of 99.84% on the independent test set, with a missed detection rate of 3.4% under five-fold cross-validation, which significantly reduces the non-detection zone.

System immunoinformatics–based design of a multi-epitope vaccine candidate against La Crosse virus

The emergence of La Crosse virus (LACV) infection has been attributed to the global expansion of mosquito habitats and climate change. This pressing issue underscores the urgency of preemptive actions to combat the infection. Currently, no vaccines or targeted antivirals are available for LACV infection. The present study utilizes advanced immunoinformatics approaches for the design of a multiepitope vaccine (LACV-mVax01) targeting the viral proteins G1, G2, and N. The structural assessment suggested favorable stability characteristics, reflected by an instability score of 17.12 and an aliphatic index of 81.88. Based on the two-dimensional structure, the LACV-mVax01 had an alpha-helical content of ~45%, with favorable beta sheets and coils. The refinement of the LACV-mVax01 resulted in an accurate model (TM-score 0.49, C-score −1.86), supported by high-quality metrics (RMSD 0.489, MolProbity 2.230, ERRAT 85.539). Docking analyses indicated potential binding interactions with TLR4, showing a binding energy of −1000.0 compared to −939.2 for TLR2. Molecular dynamics simulations were consistent with these observations, suggesting relatively stable interactions and maintenance of structural integrity in both complexes. Codon optimization yielded a codon adaptation index of 0.9848, and the predicted mRNA structure had a free energy of −342.16 kcal/mol, consistent with a stable conformation. Subsequent in silico immunological analyses suggested humoral and cellular responses, including probable memory B- and T-cell responses. However, sustained immunity was also supported by consecutive cytokine and antibody synthesis. The comprehensive structural, molecular, and immunological evaluations suggest that LACV-mVax01 is a promising vaccine candidate capable of eliciting sustained protective immunity against LACV infection. Though future in vitro , in vivo , and clinical validations are required to perceive its effectiveness and safety.

Cost-effectiveness of various referral pathways to identify advanced fibrosis among type 2 diabetes mellitus patients with metabolic dysfunction-associated steatotic liver disease in primary care setting in Malaysia

Background and aim Most international guidelines recommend a two-step approach using the Fibrosis-4 index (FIB-4) and vibration-controlled transient elastography (VCTE) to identify advanced fibrosis, a key predictor of all-cause and liver-related mortality in patients with metabolic dysfunction-associated steatotic liver disease (MASLD). However, VCTE is not available in most primary care settings in Malaysia, and there is scarce data on the cost-effectiveness of different approaches. This study evaluated the cost-effectiveness of three referral pathways for identifying advanced fibrosis among type 2 diabetes mellitus(T2DM) patients with MASLD. Methods We developed a decision-analytical model from the healthcare provider’s perspective, using 1,000 simulated patients to compare: (i) Current Practice (direct referral based on elevated alanine transaminase), (ii) Clinical Practice Guidelines (CPG) Pathway using FIB-4 single-cutoff 1.3, and (iii) FIB-4 dual-cutoffs (1.3,3.25) followed by a gamma-glutamyl transferase (GGT) test for indeterminate cases (Sequential FIB-4/GGT Pathway). Current practice served as the reference comparator. The primary outcomes were the average cost-effectiveness ratio (ACER) and the incremental cost-effectiveness ratio (ICER). Model parameters were mainly derived from local studies. Direct medical costs were reported in 2024 Malaysian Ringgit (MYR). Results Sequential FIB-4/GGT pathway had the lowest ACER at MYR930 per advanced fibrosis case identified, compared to MYR1,299 for current practice and MYR1,581 for the CPG pathway. Sequential FIB-4/GGT pathway was potentially more effective and less costly, demonstrating dominance over current practice with a cost savings of MYR2,911/additional advanced fibrosis case identified. CPG pathway was more effective and more costly than current practice, with an ICER of MYR3,785. Conclusions Sequential FIB-4/GGT pathway was cost-effective for identifying advanced fibrosis in T2DM patients with MASLD. This pragmatic approach could reduce tertiary care referrals, lower healthcare resource use and costs compared to current practice. CPG pathway was more effective than current practice, but incurred higher costs and required increased availability of VCTE within clinical practice.

Yaws active case detection surveys in 15 districts of Cote d’Ivoire

Background Yaws is one of the skin-related neglected tropical disease (skin NTD). It is closely linked to poverty and occurs mainly in poor communities in hot, humid, tropical regions. Children in disadvantaged communities in the tropics are most affected. Following the recommendations of the WHO NTD Road map 2021–2030, Côte d’Ivoire planned an evaluation (clinical and serological investigation) to assess the true extent of yaws in the country. Côte d’Ivoire is one of the nine countries currently endemic for yaws in the WHO African Region. The aim of this investigation was to establish the endemicity of yaws and to gather information that will inform planning for yaws eradication in Côte d’Ivoire. Method A public health intervention involving active case-finding was conducted in 345 health areas of 15 targeted health districts. All children aged 2–15 years in village primary schools and in households were systematically recruited and examined. All enrolled children with suspicious lesions were initially tested with Standard Diagnostic (SD) serological treponemal test. Cases positive for treponemal SD test were subsequently tested using the dual treponemal and non-treponemal point-of-care test (DPP), a rapid serological test that is used to confirm yaws. All confirmed cases were treated with a single dose of azithromycin according to age. Additionally, other skin NTD and skin conditions were detected. Results A total of 3870 villages and hamlets were investigated in which 486 975 children were targeted. Overall, 65 918 (14.66%) children with various skin lesions were identified. Two hundred and ninety-one (291) RDT (Rapid Diagnostic Test) positive cases of yaws were documented. Half of the RDT positive cases (146/291) were confirmed as DPP positive. Children aged 5–14 years ( n  = 129) represented 88% of DPP positive cases. The four most common skin conditions identified were traumatic wounds (13 855; 21.07%), dermatophytosis (15 640; 23.8%), pityriasis versicolor (6998; 10.64%) and scabies (5955; 9.05%). Conclusion This rapid investigation confirmed the endemicity of yaws in Côte d’Ivoire at low and variable prevalence. Treatment interventions will be adapted to the level of endemicity that we have found. Analysis of the results opens avenues of research to better understand the endemicity of yaws in Côte d’Ivoire.

Effectiveness of virtual clinical simulation in palliative care education for nursing students: Protocol of a randomized controlled trial

Background Palliative care often involves nursing for complex patient groups at the end of life, which requires specialized knowledge and ethical considerations. Therefore, it is full of challenges when carrying out practical teaching. Nursing students, as the next generation of healthcare professionals, will assume the important responsibility of providing specialized health services related to end-of-life care (EOLC). A deficit in direct clinical practice may impede the acquisition of essential knowledge and ability in palliative care for nursing students.Virtual clinical simulation (VCS) is acknowledged as an effective educational tool in nursing education. However, its application in palliative care education remains underexplored, particularly through rigorous study designs. While VCS has been tested in areas such as acute care or clinical deterioration, there is a global scarcity of randomized controlled trials (RCTs) evaluating its effectiveness in palliative care education, and relevant research in China remains limited. In this study protocol, we hypothesize that VCS will improve nursing students’ knowledge, ability, and attitudes regarding palliative care, compared to conventional teaching alone. Methods This study protocol outlines a prospective RCT aimed at evaluating the educational benefits of incorporating VCS into the curriculum for nursing students at a medical university in southwestern China.These students, registered for the elective Cultural Perspectives on Death and End-of-Life Education will be systematically invited to join this research. The study will randomly assign 96 students into two distinct groups: a control group and an intervention group, each comprising 48 students. The control group will participate solely in conventional classroom teaching, whereas the intervention group will receive both conventional classroom teaching and enhanced learning through an additional VCS system, aimed at improving their knowledge, ability and attitudes regarding palliative care. The intervention is grounded in experiential learning theory, which supports the use of simulated environments to foster reflective practice and skill acquisition. Assessments will occur at baseline and post-intervention of the VCS, with the primary metrics for evaluation encompassing variations in students’ knowledge, ability, and attitudes concerning palliative care, as well as their satisfaction with VCS pedagogy and confidence in applying their newly acquired skills and knowledge in practical medical settings. As this is a protocol paper, no outcome data are yet available. Discussion VCS may be an effective tool to enhance nursing students’ palliative care knowledge, ability, and attitudes, and it provides scientific evidence to inform nursing students in selecting the most appropriate approach for their palliative care education. Trial registration Chinese Clinical Trial Registry: ChiCTR2500096377, Registered on January 22, 2025.

Efficient robot navigation inspired by honeybee learning flights

Abstract Navigation is a crucial capability for both animals and robots. Although tiny flying insects can robustly navigate over long distances 1 , state-of-the-art robot navigation methods are computationally expensive and therefore restricted to large robots 2,3 . Here we propose ‘Bee-Nav’, a highly efficient navigation strategy inspired by the visual learning flights of honeybees 4–6 . In equivalent robotic learning flights, a tiny neural network is trained to map omnidirectional images to a home vector based on path integration. After learning, the robot can fly far away from home, come straight back using path integration and cancel integration drift using the visual homing network. Simulations showed that, for realistic path integration accuracies, the neural network requires training on only approximately 0.25–10.00% of the total flight area. In real-world indoor and outdoor experiments, a small drone successfully returned to within 0.5 m of home for 100% of 30–110-m flights and 70% of 200–600-m flights in windy conditions, using 3.4-kB and 42-kB neural networks, respectively. The proposed navigation strategy will be vital for resource-constrained robots that perform tasks while travelling from and to a home location. Furthermore, it provides new perspectives on the neuroethology of insect navigation, from how visual learning shapes homing trajectories to the nature of cognitive maps.

Cost-effectiveness of inclisiran in patients with atherosclerotic cardiovascular disease from Chinese healthcare perspective

Background Atherosclerotic cardiovascular disease (ASCVD) imposes a substantial clinical and economic burden in China. Despite maximally tolerated statin therapy, many patients fail to achieve recommended low-density lipoprotein cholesterol (LDL-C) targets. Inclisiran, a novel small interfering ribonucleic acid (siRNA) therapy, provides sustained LDL-C reduction, but its economic value under the Chinese healthcare system remains uncertain. Objective To evaluate the cost-effectiveness of inclisiran when added to standard lipid-lowering therapy for patients with ASCVD in China. Methods A Markov cohort multistate-transition model was developed from the perspective of the Chinese healthcare system, using a one-year cycle length and a lifetime horizon. Clinical inputs, costs, and utilities were derived from published literature and clinical trials. Both costs and outcomes were discounted at 4.5% annually. The primary result of the economic evaluation was the incremental cost-effectiveness ratio (ICER), with inclisiran considered cost-effective if the ICER was below the willingness-to-pay (WTP) threshold of CNY 191,498 per quality-adjusted life year (QALY), equivalent to two times China’s per capita gross domestic product in 2024. One-way and probabilistic sensitivity analyses were conducted to assess model uncertainty. Results Compared with statin therapy alone, inclisiran added to statin therapy yielded an additional 0.41 QALYs and incurred incremental costs of CNY 58,321.35, resulting in an ICER of CNY 143,044.57/QALY, which is well below the WTP threshold. One-way sensitivity analyses showed that the mortality reduction effect, discount rate, and inclisiran price were the main drivers of model outcomes; however, inclisiran remained cost-effective across all tested scenarios. Probabilistic sensitivity analysis showed a 95.1% probability that inclisiran is cost-effective at the current price. Conclusions At its current negotiated price of CNY 2,790 per dose, inclisiran added to standard therapy is cost-effective for patients with ASCVD in China.