ObjectiveTo propose an X-Gait motor rehabilitation assessment system based on a perception-cognition decoupling architecture to jointly achieve high-accuracy gait recognition and clinical interpretability for low clinical credibility and opaque decision-making workflows in existing intelligent gait assessment models. MethodsIn the perception layer, YOLOv11-Pose and PoseFormerV2 were integrated for 2D-to-3D human pose reconstruction. A long short-term memory network was adopted for temporal classification of normal, abnormal and pathological gaits. The layer-wise relevance propagation algorithm was used to identify discriminative joints and automatically quantify biomechanical indices including trunk forward inclination angle, stance phase percentage and stride symmetry. In the cognition layer, rehabilitation-oriented prompt engineering and prior knowledge graphs were combined to drive large language models, converting quantitative biomechanical measurements into clinically reasonable structured reports. Eight subjects were enrolled for bedside clinical validation to evaluate classification performance and clinical consistency. ResultsThe overall gait classification accuracy of X-Gait reached 94.1%. The 2D pose detection achieved 100% valid detection rate with an average confidence of 0.819. The mean per-joint position error of 3D pose reconstruction was 105.97 mm. In clinical validation of eight cases, outputs generated by explainable artificial intelligence combined with large language models showed high consistency with manual assessments from experienced rehabilitation physicians. ConclusionThe proposed X-Gait system based on perception-cognition decoupling realizes high-precision gait identification and clinically interpretable outputs.

Objective To explore the safety and feasibility of 5G remote robot-assisted thoracoscopic anatomic lobectomy and segmentectomy. Methods A retrospective analysis was conducted on patients who underwent 5G remote robot-assisted thoracoscopic anatomic lobectomy or segmentectomy between July and September 2024. The surgeries were performed collaboratively by Shanghai Chest Hospital, Affiliated Hospital of Yangzhou University, and The First Affiliated Hospital of Nanchang University. Perioperative outcomes were recorded and analyzed. Results Ten patients were included, comprising 4 anatomic lobectomies and 6 segmentectomies. The median age was 61.0 (56.0, 72.0) years. The median intraoperative bi-directional network latency was 52.5 (39.0, 54.0) ms. There were no instances of network interruption, robot-related adverse events, or conversions to open thoracotomy. The median postoperative chest tube drainage volume was 500.0 (375.0, 600.0) mL, and the median hospital stay was 5.5 (4.0, 6.0) days. No complications of Clavien-Dindo grade Ⅱor higher occurred. All patients were recurrence-free and alive during the 6-month postoperative follow-up period. Conclusion 5G remote robot-assisted thoracoscopic surgery demonstrates high safety and operational stability. This technology shows promising potential for clinical application and is worthy of further development and utilization.

Both type 2 diabetes mellitus (T2DM) and metabolic associated fatty liver disease (MAFLD) fall within the spectrum of metabolic diseases, and they exhibit a bi-directional causal relationship and robust reciprocal association. Their shared pathological cornerstone is insulin resistance (IR), which involves the interplay of mitochondrial dysfunction and chronic inflammation, forming a cascading pathological process of “IR-mitochondrial dysfunction-inflammation”. This largely explains the notable upward trend in T2DM-MAFLD co-occurrence observed over recent years. Exercise intervention, as a safe and effective non-pharmacological approach, can improve the pathological progression of these patients at multiple levels. Following the logical framework of “pathogenesis-efficacy comparison-molecular mechanisms-clinical translation”, this article systematically compares the efficacy and molecular mechanisms of moderate-intensity continuous training (MICT), resistance exercise, high-intensity interval training (HIIT), and combined training. MICT reduces intrahepatic triglycerides by promoting lipolysis and improving cardiorespiratory fitness; resistance exercise increases muscle mass and basal metabolic rate, offering unique advantages in preserving muscle while reducing fat and improving insulin sensitivity; HIIT is a time-efficient exercise modality that enhances patients’ cardiorespiratory fitness and insulin sensitivity by alternating brief periods of vigorous exertion with recovery periods, with a prominent short-term triglyceride-lowering effect; combined training produces synergistic effects, comprehensively improving glucolipid metabolism and showing the best long-term adherence. Mechanistically, exercise exerts its beneficial effects through three common pathways: (1) AMPK-mediated lipid oxidation and mitochondrial biogenesis; (2) IRS/PI3K/Akt-mediated insulin signaling sensitization; and (3) Nrf2/ARE anti-oxidation and TGF-β/Smads anti-fibrosis regulation. Different exercise modalities activate these pathways with distinct emphases: MICT most directly and persistently activates the AMPK pathway; resistance exercise uniquely improves IRS/PI3K/Akt signaling through muscle mass gain; HIIT induces the highest AMPK activation intensity and triggers unique lactate-mediated signaling regulation; combined training integrates the above multiple mechanistic advantages. For clinical translation, multidisciplinary team collaboration is essential to ensure safety and adherence. Individualized prescriptions should be formulated according to the FITT-VP principle and patient phenotypes—frequency of 3-5 sessions/week of aerobic exercise combined with 2-3 sessions/week of resistance exercise; intensity of moderate-intensity (40%-60% heart rate reserve (HRR))aerobic exercise and 60%-80% of one-repetition maximum (1-RM) for resistance exercise; time of at least 150 min/week of moderate-intensity aerobic exercise, 30-60 min per session; type of combined training as the preferred modality; total volume of≥500-1 000 MET-min/week; and progression adjusted every 4-6 weeks—with real-time adjustments supported by wearable devices, ultimately forming a closed-loop management system from initial assessment to long-term follow-up. Notably, current studies have limitations such as small sample sizes and short intervention periods. Future research should focus on long-term follow-up, multi-omics biomarkers, and combined exercise-drug strategies. In conclusion, the systematic integration of structured, individualized, and sustainable exercise interventions into the multidisciplinary management pathway for patients with T2DM complicated by MAFLD is an urgent need in current clinical practice.

Adipose tissue is a critical energy reservoir in animals and humans, with multifaceted roles in endocrine regulation, immune response, and providing mechanical protection. Based on anatomical location and functional characteristics, adipose tissue can be categorized into distinct types, including white adipose tissue (WAT), brown adipose tissue (BAT), beige adipose tissue, and pink adipose tissue. Traditionally, adipose tissue research has centered on its morphological and functional properties as a whole. However, with the advent of single-cell transcriptomics, a new level of complexity in adipose tissue has been unveiled, showing that even under identical conditions, cells of the same type may exhibit significant variation in morphology, structure, function, and gene expression——phenomena collectively referred to as cellular heterogeneity. Single-cell transcriptomics, including techniques like single-cell RNA sequencing (scRNA-seq) and single-nucleus RNA sequencing (snRNA-seq), enables in-depth analysis of the diversity and heterogeneity of adipocytes at the single-cell level. This high-resolution approach has not only deepened our understanding of adipocyte functionality but also facilitated the discovery of previously unidentified cell types and gene expression patterns that may play key roles in adipose tissue function. This review delves into the latest advances in the application of single-cell transcriptomics in elucidating the heterogeneity and diversity within adipose tissue, highlighting how these findings have redefined the understanding of cell subpopulations within different adipose depots. Moreover, the review explores how single-cell transcriptomic technologies have enabled the study of cellular communication pathways and differentiation trajectories among adipose cell subgroups. By mapping these interactions and differentiation processes, researchers gain insights into how distinct cellular subpopulations coordinate within adipose tissues, which is crucial for maintaining tissue homeostasis and function. Understanding these mechanisms is essential, as dysregulation in adipose cell interactions and differentiation underlies a range of metabolic disorders, including obesity and diabetes mellitus type 2. Furthermore, single-cell transcriptomics holds promising implications for identifying therapeutic targets; by pinpointing specific cell types and gene pathways involved in adipose tissue dysfunction, these technologies pave the way for developing targeted interventions aimed at modulating specific adipose subpopulations. In summary, this review provides a comprehensive analysis of the role of single-cell transcriptomic technologies in uncovering the heterogeneity and functional diversity of adipose tissues.

Objective To describe the significance of the pretreatment inflammatory indicators in predicting the prognosis of patients with esophageal squamous cell carcinoma (ESCC) after undergoing radical radiotherapy. Methods The data of 246 ESCC patients who underwent radical radiotherapy were retrospectively collected. Receiver operating characteristic (ROC) curves were drawn to determine the optimal cutoff values for platelet-lymphocyte ratio (PLR), neutrophil-lymphocyte ratio (NLR), and systemic immune-inflammation index (SII). The Kaplan-Meier method was used for survival analysis. We conducted univariate and multivariate analyses by using the Cox proportional risk regression model. Software R (version 4.2.0) was used to create the nomogram of prognostic factors. Results The results of the ROC curve analysis showed that the optimal cutoff values of PLR, NLR, and SII were 146.06, 2.67, and 493.97, respectively. The overall response rates were 77.6% and 64.5% in the low and high NLR groups, respectively (P<0.05). The results of the Kaplan-Meier survival analysis revealed that the prognosis of patients in the low PLR, NLR, and SII group was better than that of patients in the high PLR, NLR, and SII group (all P<0.05). The results of the multivariate Cox regression analysis showed that gender, treatment modalities, T stage, and NLR were independent factors affecting the overall survival (OS). In addition, T stage and NLR were independent factors affecting the progression-free survival (PFS) (all P<0.05). The nomogram models of OS and PFS prediction were established based on multivariate analysis. The C-index values were 0.703 and 0.668. The calibration curves showed excellent consistency between the predicted and observed OS and PFS. Conclusion The pretreatment values of PLR, NLR, and SII are correlated with the prognosis of patients with ESCC who underwent radical radiotherapy. Moreover, NLR is an independent factor affecting the OS and PFS of ESCC patients. The NLR-based nomogram model has a good predictive ability.

Mycophenolic acid (MPA), the active moiety of both mycophenolate mofetil (MMF) and enteric-coated mycophenolate sodium (EC-MPS), serves as a primary immunosuppressant for maintaining solid organ transplants. Therapeutic drug monitoring (TDM) enhances treatment outcomes through tailored approaches. This study aimed to develop an evidence-based guideline for MPA TDM, facilitating its rational application in clinical settings. The guideline plan was drawn from the Institute of Medicine and World Health Organization (WHO) guidelines. Using the Delphi method, clinical questions and outcome indicators were generated. Systematic reviews, Grading of Recommendations Assessment, Development, and Evaluation (GRADE) evidence quality evaluations, expert opinions, and patient values guided evidence-based suggestions for the guideline. External reviews further refined the recommendations. The guideline for the TDM of MPA (IPGRP-2020CN099) consists of four sections and 16 recommendations encompassing target populations, monitoring strategies, dosage regimens, and influencing factors. High-risk populations, timing of TDM, area under the curve (AUC) versus trough concentration (C₀), target concentration ranges, monitoring frequency, and analytical methods are addressed. Formulation-specific recommendations, initial dosage regimens, populations with unique considerations, pharmacokinetic-informed dosing, body weight factors, pharmacogenetics, and drug‍-‍drug interactions are covered. The evidence-based guideline offers a comprehensive recommendation for solid organ transplant recipients undergoing MPA therapy, promoting standardization of MPA TDM, and enhancing treatment efficacy and safety.
Mycophenolic Acid/administration & dosage* ; Drug Monitoring/methods* ; Humans ; Organ Transplantation ; Immunosuppressive Agents/administration & dosage* ; Delphi Technique

In recent years, the application of artificial intelligence (AI) in the field of biology has witnessed remarkable advancements. Among these, the most notable achievements have emerged in the domain of protein structure prediction and design, with AlphaFold and related innovations earning the 2024 Nobel Prize in Chemistry. These breakthroughs have transformed our ability to understand protein folding and molecular interactions, marking a pivotal milestone in computational biology. Looking ahead, it is foreseeable that the accurate prediction of various physicochemical properties of proteins—beyond static structure—will become the next critical frontier in this rapidly evolving field. One of the most important protein properties is thermodynamic stability, which refers to a protein’s ability to maintain its native conformation under physiological or stress conditions. Accurate prediction of protein stability, especially upon single-point mutations, plays a vital role in numerous scientific and industrial domains. These include understanding the molecular basis of disease, rational drug design, development of therapeutic proteins, design of more robust industrial enzymes, and engineering of biosensors. Consequently, the ability to reliably forecast the stability changes caused by mutations has broad and transformative implications across biomedical and biotechnological applications. Historically, protein stability was assessed via experimental methods such as differential scanning calorimetry (DSC) and circular dichroism (CD), which, while precise, are time-consuming and resource-intensive. This prompted the development of computational approaches, including empirical energy functions and physics-based simulations. However, these traditional models often fall short in capturing the complex, high-dimensional nature of protein conformational landscapes and mutational effects. Recent advances in machine learning (ML) have significantly improved predictive performance in this area. Early ML models used handcrafted features derived from sequence and structure, whereas modern deep learning models leverage massive datasets and learn representations directly from data. Deep neural networks (DNNs), graph neural networks (GNNs), and attention-based architectures such as transformers have shown particular promise. GNNs, in particular, excel at modeling spatial and topological relationships in molecular structures, making them well-suited for protein modeling tasks. Furthermore, attention mechanisms enable models to dynamically weigh the contribution of specific residues or regions, capturing long-range interactions and allosteric effects. Nevertheless, several key challenges remain. These include the imbalance and scarcity of high-quality experimental datasets, particularly for rare or functionally significant mutations, which can lead to biased or overfitted models. Additionally, the inherently dynamic nature of proteins—their conformational flexibility and context-dependent behavior—is difficult to encode in static structural representations. Current models often rely on a single structure or average conformation, which may overlook important aspects of stability modulation. Efforts are ongoing to incorporate multi-conformational ensembles, molecular dynamics simulations, and physics-informed learning frameworks into predictive models. This paper presents a comprehensive review of the evolution of protein thermodynamic stability prediction techniques, with emphasis on the recent progress enabled by machine learning. It highlights representative datasets, modeling strategies, evaluation benchmarks, and the integration of structural and biochemical features. The aim is to provide researchers with a structured and up-to-date reference, guiding the development of more robust, generalizable, and interpretable models for predicting protein stability changes upon mutation. As the field moves forward, the synergy between data-driven AI methods and domain-specific biological knowledge will be key to unlocking deeper understanding and broader applications of protein engineering.

Objective To analyze the clinical characteristics,treatment,and prognosis of immune checkpoint inhibitor-related diabetes mellitus(ICI-DM).Methods The clinical characteristics,laboratory examinations,treatment regimens,and follow-up outcomes of 10 ICI-DM patients who were diagnosed and treated in the First Medical Center of Chinese PLA General Hospital between July 2019 and December 2024 were retrospectively analyzed.Relevant literatures were retrieved from domestic and foreign databases such as PubMed,CNKI,and VIP.The clinical characteristics of ICI-DM were summarized based on the literature results.Results All 10 patients were PD-1 inhibitor users,including 5 males and 5 females,with a median age of 54.5(51.3,64.0)years and a body mass index(BMI)of(22.0±2.15)kg/m2.Among them,9 cases(90.0%)were fulminant type 1 diabetes mellitus(FT1DM);9 cases(90.0%)had a severity of adverse events reaching grade 3-4 according to the Common Terminology Criteria for adverse events(CTCAE).The median time from PD-1 inhibitor treatment to the occurrence of the classic diabetes symptoms referred to as"three more and one less"(polyuria,polydipsia,polyphagia,and weight loss)in all patients was 145.5(110.5,204.8)days,and the medication duration was 6.0(4.3,7.8)cycles.The average blood glucose level of the 10 patients at the time of consultation was 25.3(10.0-41.4)mmol/L,and the glycated hemoglobin(HbA1c)level was 8.0%(6.6%-10.9%).Eight patients had fasting and 2-hour C-peptide levels<0.1 ng/ml(fasting C-peptide from<0.010 to 0.067 ng/ml,2-hour C-peptide from<0.010 to 0.077 ng/ml).Nine of the 10 patients were negative for diabetes autoantibodies,while 1 was not tested.All 10 patients were successfully treated with insulin and other therapies.During the follow-up after discharge,all patients still relied on insulin treatment,and no significant recovery of pancreatic islet β cell function was observed compared with that at discharge.Literature review revealed that ICI-DM was more common in PD-1 inhibitor users,with clinical mainly manifested as diabetic ketoacidosis(DKA)(65.4%)and diabetic ketosis(13.1%).Patients had severely impaired pancreatic islet function and required long-term insulin treatment,and some cases were complicated by thyroid or pituitary dysfunction.Conclusions ICI-DM typically presents as FT1DM,often manifesting with DKA or diabetic ketosis at onset.It is characterized by severe and irreversible loss of pancreatic islet function,necessitating lifelong insulin therapy.To enable early detection and prompt treatment,close monitoring of blood glucose is essential during ICI treatment.

Drug-induced liver injury(DILI)is a common adverse drug reaction in clinical practice,which can lead to acute liver failure and even death in severe cases.In recent years,with the continuous introduction of new drugs and the expansion of their usage,the incidence and mortality rates of DILI have shown an upward trend,posing significant challenges to public health and clinical treatment.Macrophages,as a crucial component of the innate immune system,exhibit high plasticity and heterogeneity.They can polarize into pro-inflammatory M1 type or anti-inflammatory M2 type in response to microenvironmental signals.Research has demonstrated that macrophage polarization plays a central regulatory role in the occurrence and progression of DILI by influencing various processes such as inflammatory responses,cell apoptosis,and tissue repair.This review focuses on elucidating the regulatory mechanisms and roles of macrophage polarization in DILI,providing a theoretical framework for developing precise immunotherapeutic strategies.

Objective:To construct a nomogram model for predicting the incidence of hepatocellular carcinoma based on serum abnormal prothrombin and alpha-fetoprotein and evaluate the predictive effect.Methods:Retrospective analysis of data from 351 patients with liver disease who received treatment at the First Affiliated Hospital of Zhengzhou University from January 2021 to December 2023, including 285 males and 66 females, aged (52.9±11.9) years. Among the 351 patients, there were 229 cases (65.2%) of hepatocellular carcinoma, 87 cases (24.8%) of liver cirrhosis, and 35 cases (10.0%) of chronic hepatitis B. All patients were randomly divided into a training set ( n=245) and a testing set ( n=106) in a 7∶3 ratio without replacement sampling. The training set was used to construct the model, and the testing set was used to evaluate the model. At the same time, gender, age, disease type, and other indicators were compared between the two sets. The risk factors of hepatocellular carcinoma were analyzed by univariate and multivariate logistic regression based on the training set, and a nomogram was constructed to predict the incidence of hepatocellular carcinoma based on the multivariate results. Receiver operating characteristic (ROC) curve and calibration curve were used to evaluate the predictive performance of nomogram, and decision curve analysis was used to evaluate the clinical applicability of the model. Results:There was no statistically significant difference in age, gender, disease type, etc. between the training and testing sets of patients (all P>0.05). Univariate logistic regression analysis showed that age, abnormal prothrombin logarithm (LnPIVKA-Ⅱ), alpha-fetoprotein logarithm (LnAFP), and diabetes were associated with hepatocellular carcinoma (all P<0.05). Multivariate logistic regression analysis showed that older age ( OR=1.07, 95% CI: 1.03-1.12), higher LnPIVKA-Ⅱ ( OR=2.97, 95% CI: 1.97-4.46), higher LnAFP ( OR=1.43, 95% CI: 1.11-1.84) and diabetes ( OR=5.17, 95% CI: 1.02-26.17) were risk factors for hepatocellular carcinoma (all P<0.05). Based on the above variables, a nomogram model for predicting the incidence of hepatocellular carcinoma was constructed. The area under the ROC curve analysis of the nomogram for predicting the incidence of hepatocellular carcinoma was 0.920 (95% CI: 0.886-0.953) in the training set and 0.934 (95% CI: 0.891-0.977) in the testing set. The calibration curve fit well with the standard curve, and the prediction was basically consistent with the actual situation. The decision curve analysis showed that the net benefit of the model was greater than 0 under most thresholds (0.1-1.0). Conclusion:The nomogram constructed based on age, LnPIVKA-Ⅱ, LnAFP and diabetes can effectively predict the incidence of hepatocellular carcinoma and has clinical applicability.