Abstract: Sports injury management is undergoing a paradigm shift from ‘reactive post-event intervention’ to ‘proactive pre-event prevention’. Although traditional machine learning (ML) and deep learning (DL) have demonstrated some efficacy in identifying injury risk factors, they still face analytical bottlenecks when dealing with complex, heterogeneous, and multimodal sports data. This paper aims to systematically explore the current applications of large language models (LLMs) in competitive sports injury prevention, the optimisation pathways for core technologies, and the empirical challenges they face in complex clinical decision-making. It examines how Retrieval-Augmented Generation (RAG), Parameter-Efficient Fine-Tuning (PEFT), and Multimodal Large Language Models (MLLMs) can bridge the general language barrier to adapt to specialised knowledge in sports medicine. Although LLMs possess powerful natural language processing and multimodal data integration capabilities, issues such as data privacy, model hallucinations, out-of-distribution (OOD) robustness, and insufficient explainability limit their independent application in high-risk medical decision-making. Future research needs to be deepened in the directions of edge computing deployment and explainable artificial intelligence (XAI). Read More

Abstract: Hepatocellular carcinoma (HCC) is still one of the top causes of cancer-related mortalities worldwide, and immunosuppressive tumour microenvironment (TME) is a key obstacle for effective therapy. Tumour-associated macrophages (TAMs) are one of the most abundant immune cells in the HCC TME and display a pro-tumourigenic character, which is related with immunosuppression, angiogenesis and resistance to systemic therapy. Reprogramming TAMs from M2-like pro-tumourigenic to an anti-tumour M1-like state has appeared as a viable therapeutic strategy. Herein we discuss the molecular pathways involved in TAM polarisation in HCC, including the CSF1R, PI3Kγ and NF-κB/STAT3 signalling pathways and review the therapeutic approaches targeting TAM re-programming including CSF1R inhibition, TLR agonists, combination with immune checkpoint blockade, CAR-macrophage engineering and nanoparticle-mediated delivery.  We also highlight the clinical challenges of TAM-targeted therapy such as phenotypic plasticity and off-target toxicity, and markers for patient stratification. Read More

Abstract: Methylmalonic acid (MMA), a sensitive marker of vitamin B12 deficiency, exerts multifaceted adverse effects on male reproductive function when accumulated. This systematic review examines how MMA leads to reduced sperm count, decreased motility, and DNA damage by inhibiting mitochondrial complex II, inducing oxidative stress, and disrupting one-carbon metabolism; causing Leydig cell atrophy, impaired testosterone synthesis, and compromised Sertoli cell function; and disrupting endocrine regulation within the hypothalamic-pituitary-testicular axis. Mechanistically, MMA induces reactive oxygen species (ROS) production, ATP synthesis inhibition, inflammatory responses, and epigenetic modifications, forming a pathological cascade of “energy crisis–oxidative damage–apoptosis.” Clinical observations indicate that MMA levels are negatively correlated with sperm quality and testosterone levels, and supplementation with vitamin B12 can effectively reverse the associated damage. It is recommended that MMA monitoring be incorporated into the screening system for male infertility etiology, and that early nutritional intervention be implemented for high-risk populations to improve male reproductive health. Read More

Abstract: This paper examines the Health Star Rating (HSR) system in Australia and New Zealand as a front-of-pack (FOP) nutrition labelling policy, with a focus on its role in improving obesity-related dietary health and its potential applicability in other contexts such as China. While the HSR aims to simplify nutritional information and support healthier food choices, this study suggests that its effectiveness is constrained by both behavioral and structural limitations. The analysis shows that the HSR enhances consumer understanding by providing a simplified summary of nutritional quality and facilitates product comparison within categories. It also creates incentives for manufacturers to reformulate products. However, these benefits do not necessarily translate into meaningful improvements in dietary behavior or weight management outcomes. Consumer misunderstandings, including inappropriate cross-category comparisons and the health halo effect, may distort perceptions of product healthfulness. In addition, the system does not account for portion size or total energy intake, limiting its relevance to obesity prevention. At the structural level, the voluntary nature of the HSR leads to selective labelling, with higher-rated products more likely to display the rating, resulting in a biased information environment. Incomplete market adoption further reduces its population-level impact. The paper suggests that while HSR-type systems may offer useful insights for countries such as China, their effectiveness would depend on adaptation to local regulatory frameworks. Emerging digital and AI-based tools may provide a potential pathway to enhance the system by enabling more personalized and context-specific dietary guidance. Read More

Abstract: Objective: This study aims to investigate the correlation between ultrasonic myocardial work parameters and triglyceride-glucose-body mass index (TyG-BMI) with the degree of coronary stenosis, and to evaluate the diagnostic value of each parameter individually and in combination for significant coronary stenosis. Methods: Data were collected from 155 patients who underwent coronary angiography for chest pain at the Department of Cardiology of the Affiliated Hospital of Youjiang Medical University for Nationalities between September 2025 and April 2026. Based on angiographic findings, patients were classified into a group with significant coronary stenosis (n=89) and a group without significant stenosis (n=66). Differences between groups were compared using independent-sample t-test or Mann-Whitney U test. Correlation analyses were performed to calculate the coefficients of correlation between coronary stenosis severity and ultrasound parameters as well as clinical indicators. Binary logistic regression was employed to identify independent influencing factors, while the receiver operating characteristic (ROC) curve was used to evaluate the diagnostic performance of individual indicators and combined models. Results: Correlation analysis revealed an extremely strong positive correlation between the degree of coronary stenosis and the global myocardial work index (GCW) (r=0.898), an extremely strong negative correlation between global wasted work (GWW) and global myocardial work efficiency (GWE) (r=-0.904), moderate positive correlations between systolic blood pressure (SBP) and interventricular septal thickness (IVST) (r=0.525) and between total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) (r=0.597), as well as a moderate positive correlation between TyG-BMI. Compared with the group with significant stenosis, the group with non-significant stenosis exhibited significantly reduced IVST, GWW, and TyG-BMI, while GWI, GCW, and GWE were significantly … Read More