Objective: To develop QingNangTCM, a specialized large language model (LLM) tailored for expert-level traditional Chinese medicine (TCM) question-answering and clinical reasoning, addressing the scarcity of domain-specific corpora and specialized alignment. Methods: We constructed QnTCM_Dataset, a corpus of 100 000 entries, by integrating data from ShenNong_TCM_Dataset and SymMap v2.0, and synthesizing additional samples via retrieval-augmented generation (RAG) and persona-driven generation. The dataset comprehensively covers diagnostic inquiries, prescriptions, and herbal knowledge. Utilizing P-Tuning v2, we fine-tuned the GLM-4-9B-Chat backbone to develop QingNangTCM. A multi-dimensional evaluation framework, assessing accuracy, coverage, consistency, safety, professionalism, and fluency, was established using metrics such as bilingual evaluation understudy (BLEU), recall-oriented understudy for gisting evaluation (ROUGE), metric for evaluation of translation with explicit ordering (METEOR), and LLM-as-a-Judge with expert review. Qualitative analysis was conducted across four simulated clinical scenarios: symptom analysis, disease treatment, herb inquiry, and failure cases. Baseline models included GLM-4-9B-Chat, DeepSeek-V2, HuatuoGPT-II (7B), and GLM-4-9B-Chat (freeze-tuning). Results: QingNangTCM achieved the highest scores in BLEU-1/2/3/4 (0.425/0.298/0.137/0.064), ROUGE-1/2 (0.368/0.157), and METEOR (0.218), demonstrating a balanced and superior normalized performance profile of 0.900 across the dimensions of accuracy, coverage, and consistency. Although its ROUGE-L score (0.299) was lower than that of HuatuoGPT-II (7B) (0.351), it significantly outperformed domain-specific models in expert-validated win rates for professionalism (86%) and safety (73%). Qualitative analysis confirmed that the model strictly adheres to the “symptom-syndrome-pathogenesis-treatment” reasoning chain, though occasional misclassifications and hallucinations persisted when dealing with rare medicinal materials and uncommon syndromes. Conclusion Combining domain-specific corpus construction with parameter-efficient prompt tuning enhances the reasoning behavior and domain adaptation of LLMs for TCM-related tasks. This work provides a technical framework for the digital organization and intelligent utilization of TCM knowledge, with potential value for supporting diagnostic reasoning and medical education.

ObjectiveThis study systematically analyzed the arginine metabolic dysregulation in the rat model of heart failure （HF）， providing a modern scientific basis for elucidating the pathogenesis of HF and offering new insights for the prevention and treatment of HF with traditional Chinese medicine （TCM）. MethodsA thoracotomy was performed to ligate the left anterior descending coronary artery of rats， which induced acute myocardial ischemia and thus led to the development of post-myocardial infarction heart failure. The rats were divided into a sham surgery group and a model group， with eight rats in each group. Serum targeted metabolomics analysis was performed using ultra-performance liquid chromatography-triple quadrupole mass spectrometry （UPLC-TQ-S）， and the spatial distribution of metabolites in cardiac tissue was observed using airflow-assisted desorption electrospray ionizationmass spectrometry imaging （AFADESI-MSI）. Targets associated with HF and arginine metabolism were screened from databases including GeneCards and the Gene Expression Omnibus （GEO）， a protein-protein interaction （PPI） network was constructed， and enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes （KEGG） pathway and Gene Ontology （GO） was performed. Finally， molecular docking was conducted to verify the binding between core metabolic components and key targets， and potential TCMs were predicted based on the core pathways and targets. ResultsCompared with the sham surgery group， the levels of arginine and citrulline in the serum of model rats were significantly decreased （P<0.01）， while those of proline， ornithine， creatine， creatinine and glutamate were significantly increased （P<0.05， P<0.01）. Cardiac mass spectrometry imaging showed a decreased abundance of arginine in the local myocardial tissue. Bioinformatics analysis identified 24 core functional targets， such as the angiotensin-converting enzyme （ACE）， neuronal nitric oxide synthase （NOS1）， 5-hydroxytryptamine receptor 2A （HTR2A）， and epidermal growth factor receptor （EGFR）， and enrichment analysis indicated that these targets were significantly involved in the calcium signaling pathway， neuroactive ligand-receptor interactions， and phosphatidylinositol signaling pathway. Molecular docking confirmed strong binding activities between arginine， citrulline and HTR2A， as well as between creatine， creatinine and EGFR. Based on pathway-target prediction， potential TCM interventions， such as ginseng and magnolia， were identified. ConclusionThis study revealed characteristic arginine metabolic disorder in HF， and the core targets of HF were closely associated with the phosphatidylinositol signaling pathway. It provides a modern biological interpretation of the pathogenesis of HF in TCM from the perspectives of metabolites and signaling pathways， and offers valuable insights for targeted therapy of HF and the development of TCM.

ObjectiveThis study systematically analyzed the arginine metabolic dysregulation in the rat model of heart failure （HF）， providing a modern scientific basis for elucidating the pathogenesis of HF and offering new insights for the prevention and treatment of HF with traditional Chinese medicine （TCM）. MethodsA thoracotomy was performed to ligate the left anterior descending coronary artery of rats， which induced acute myocardial ischemia and thus led to the development of post-myocardial infarction heart failure. The rats were divided into a sham surgery group and a model group， with eight rats in each group. Serum targeted metabolomics analysis was performed using ultra-performance liquid chromatography-triple quadrupole mass spectrometry （UPLC-TQ-S）， and the spatial distribution of metabolites in cardiac tissue was observed using airflow-assisted desorption electrospray ionizationmass spectrometry imaging （AFADESI-MSI）. Targets associated with HF and arginine metabolism were screened from databases including GeneCards and the Gene Expression Omnibus （GEO）， a protein-protein interaction （PPI） network was constructed， and enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes （KEGG） pathway and Gene Ontology （GO） was performed. Finally， molecular docking was conducted to verify the binding between core metabolic components and key targets， and potential TCMs were predicted based on the core pathways and targets. ResultsCompared with the sham surgery group， the levels of arginine and citrulline in the serum of model rats were significantly decreased （P<0.01）， while those of proline， ornithine， creatine， creatinine and glutamate were significantly increased （P<0.05， P<0.01）. Cardiac mass spectrometry imaging showed a decreased abundance of arginine in the local myocardial tissue. Bioinformatics analysis identified 24 core functional targets， such as the angiotensin-converting enzyme （ACE）， neuronal nitric oxide synthase （NOS1）， 5-hydroxytryptamine receptor 2A （HTR2A）， and epidermal growth factor receptor （EGFR）， and enrichment analysis indicated that these targets were significantly involved in the calcium signaling pathway， neuroactive ligand-receptor interactions， and phosphatidylinositol signaling pathway. Molecular docking confirmed strong binding activities between arginine， citrulline and HTR2A， as well as between creatine， creatinine and EGFR. Based on pathway-target prediction， potential TCM interventions， such as ginseng and magnolia， were identified. ConclusionThis study revealed characteristic arginine metabolic disorder in HF， and the core targets of HF were closely associated with the phosphatidylinositol signaling pathway. It provides a modern biological interpretation of the pathogenesis of HF in TCM from the perspectives of metabolites and signaling pathways， and offers valuable insights for targeted therapy of HF and the development of TCM.

Objective: To investigate the current status of benefit finding among young and middle-aged patients with type 2 diabetes mellitus (T2DM) and analyze its influencing factors, so as to provide a reference for improving the level of benefit finding in this population. Methods: From November 2022 to May 2023, young and middle-aged patients with T2DM aged 18-59 years hospitalized in the endocrinology departments of 2 tertiary hospitals in Hengyang City, Hunan Province were selected as survey subjects by a convenience sampling method. Basic demographic information was collected using a general questionnaire survey. Benefit finding, resourcefulness, and stigma were evaluated using the Benefit Finding Scale, the Chinese Version of the Resourcefulness Scale, and the Type 2 Diabetes Stigma Assessment Scale, respectively. A multiple linear regression model was used to analyze the influencing factors of benefit finding among young and middle-aged patients with T2DM. Results: A total of 305 young and middle-aged patients with T2DM were investigated, including 222 males (72.79%) and 83 females (27.21%). There were 231 cases aged 45-59 years, accounting for 75.74%. The scores for benefit finding, resourcefulness, and stigma were (42.86±6.06), (75.12±11.30), and (41.20±10.10), respectively. Multiple linear regression analysis showed that young and middle-aged patients with T2DM who were male (β′=0.088), aged 18-<45 years (β′=0.083), absence of diabetes complications (β′=0.124), and had higher resourcefulness scores (β′=0.679) had higher levels of benefit finding, while patients with higher stigma scores (β′=-0.097) had lower levels of benefit finding. Conclusion The level of benefit finding among young and middle-aged patients with T2DM was moderate, and was related to gender, age, diabetes complications, resourcefulness, and stigma.

Objective To investigate serum HCV-RNA detection rate in patients with positive hepatitis C virus (HCV) immunoglobulin G (IgG) antibody, and to analyze the related factors of HCV-RNA detection in patients with positive HCV IgG antibody by logistic regression analysis. Methods A retrospective study was conducted on 500 patients with positive HCV IgG antibody. The positive rate of HCV-RNA was counted. Logistic regression analysis was adopted to analyze the related factors of HCV-RNA detection in HCV IgG antibody positive population. Results A total of 317 patients with positive HCV-RNA detection were included in serum HCV-RNA positive group (n=317), and 183 patients with negative HCV-RNA detection were enrolled as serum HCV-RNA negative group (n=183). The diagnosis rate of HCV infection was 63.40% (317/500). Logistic regression analysis showed that abnormal levels of serum ALT, AST, GGT, TBIL, HCV IgG antibody, prothrombin time, and the presence of clinical symptoms were independently correlated with the detection of HCV-RNA in HCV IgG antibody positive individuals (P<0.05). Established 7-factor prediction models and created a nomogram. The chi-square value was 0.586, and P = 0.89 > 0.05. The ROC curve was plotted using pROC, and the area under the ROC curve was 0.912. Conclusion Among the population positive for HCV IgG antibody, the detection rate of serum HCV-RNA is high. Abnormal levels of ALT, AST, GGT, TBIL, IgG antibody, prothrombin time and clinical symptoms are the related factors for the detection of HCV-RNA. These indicators provide a reference for the clinical judgment of the real infection risk of HCV IgG positive individuals.

OBJECTIVE To evaluate the cost-effectiveness of fluzoparib monotherapy versus combination with apatinib as second-line or later-line treatment for patients with germline BRCA1 / 2 -mutated human epidermal growth factor receptor 2-negative （HER2－） metastatic breast cancer， from the perspective of the Chinese healthcare system. METHODS Based on the FABULOUS trial， a partitioned survival model was developed to simulate disease progression， with a cycle length of 21 days and a time horizon of 10 years. A discount rate of 4.5% per annum was applied. Model outputs included total costs and quality-adjusted life years （QALYs）. Cost-utility analysis was conducted by comparing the incremental cost-effectiveness ratio （ICER） between the two regimens against a willingness-to-pay （WTP） threshold set at twice China’s per capita gross domestic product in 2025 （199 330 yuan/QALY）. Robustness of the base-case results was tested through one-way sensitivity analysis， probabilistic sensitivity analysis， and scenario analyses. RESULTS For patients with germline BRCA1/2 -mutated HER2－ metastatic breast cancer， fluzoparib combination with apatinib provided greater health benefits （1.91 QALYs vs. 1.69 QALYs） and incurred lower total costs （1 634 119.3 yuan vs. 1 855 264.2 yuan） compared with fluzoparib monotherapy， thus representing a dominant strategy. One-way sensitivity and probabilistic sensitivity analyses confirmed the robustness of the base-case findings. Scenario analyses showed that the conclusion remained robust when the time horizon was shortened to 5 years or when the discount rate was changed to 3%； however， the results reversed when the extrapolation distribution model was altered. CONCLUSIONS Under the current WTP threshold， compared with fluzoparib monotherapy， the combination of fluzoparib and apatinib is more cost-effective as a second-line or later-line treatment for Chinese patients with germline BRCA1 / 2 -mutated HER2－ metastatic breast cancer.

Objective To investigate the efficacy of lacosamide combined with ketogenic diet in the treatment of pediatric epilepsy and its impacts on the levels of serum glial fibrillary acidic protein(GFAP)and specific protein of the central nervous system(S-100β).Methods A total of 64 chil-dren with epilepsy from July 2022 to April 2024 were selected and randomly divided into conventional group(n=32)receiving oral lacosamide tablets and combined group(n=32)receiving ketogenic di-et therapy based on the conventional group.The efficacy,cognitive function,and intelligence of chil-dren in both groups after treatment,as well as the levels of serum GFAP and S-100β,the frequency of seizures,and electroencephalogram(EEG)findings before and after treatment were evaluated.Results After treatment,the number of alpha waves,the Full-scale Intelligence Quotient(FIQ)of the Wechsler Intelligence Scale for Children,and the Montreal Cognitive Assessment(MoCA)scores increased in both groups,with a greater increase in the combined group compared to the conventional group,and the differences were statistically significant(P＜0.05).The levels of serum S-100β and GFAP,the frequency of seizures,and the number of cases with slow waves and epileptiform waves decreased after treatmen,and these indicators were lower in the combined group than in the conventional group,with sta-tistically significant differences(P＜0.05).The total effective rate was 90.63％in the combined group and 68.75％in the conventional group,with a statistically significant difference(P＜0.05).The incidence of adverse reactions was 12.50％in the combined group and 9.38％in the conventional group,with no statistically significant difference(P＞0.05).Conclusion Lacosamide combined with ketogenic diet shows good efficacy in the treatment of pediatric epilepsy,and the combined therapy improves neurological function by synergistically reducing the levels of GFAP/S-100β.

Objective To conduct a unique and pioneering molecular epidemiological investigation of a case of Eurasian avian-like H1N1 swine influenza identified in Yunnan Province in 2022(the fourth such case in the province)and to understand its genetic characteristics so as to reveal its potential impact on human health.Methods Real-time fluorescent quantitative PCR detection technology was used for the nucleic acid testing of the case's pharyngeal swab samples,close contacts,and environmental samples from the living area.Positive samples were subjected to virus isolation using MDCK cells.Cell cultures were authenticated using erythrocyte agglutination assay with guinea pig blood and real-time fluorescence quantitative RT-PCR.Whole genome sequencing was performed using the Illumina MiseqNext-generation sequencing platform,and a phylogenetic tree was constructed using MEGA 7.0 software to analyze the genetic molecular characteristics.Results The first G5 genotype Eurasian avian-like H1N1 swine influenza virus in Yunnan Province was successfully isolated,and the whole genome sequence of the virus was obtained.This virus possessed the molecular characteristics associated with increased adaptability,virulence,or transmissibility in mammals and had a nucleotide consistency of 99.2％～99.7％with a porcine strain isolated in Jiangsu province.These findings underscored the potential threat this virus poses to human health.Conclusion The study underscores the importance of further monitoring swine influenza in preventing new influenza virus subtypes that can infect humans.

Objective To evaluate the relationship between nutritional parameters and the risk of venous thromboembolicism(VTE)in patients with tuberculosis so as to identify the risk factors and predictors of thrombosis and assist in the early identification of high-risk factors for VTE in patients with the pulmonary tuberculosis.Methods A total of 323 patients diagnosed with the pulmonary tuberculosis and hospitalized in Kunming Third People's Hospital from August 2021 to August 2023 were collected.According to the VTE risk assessment of non-operative patients,they were divided into the high-risk group and the low-risk group respectively with 116 and 207 in each group.The nutritional indicators with statistically significant differences between the two groups were screened by Lasso regression.Multivariate Logistic regression was used to screen the independent risk factors for high VTE risk in pulmonary tuberculosis patients,and a nomogram prediction model was constructed.The prediction model was evaluated by receiver operating characteristic curve(ROC),calibration curve,decision curve,and influence curve.Results Patients in the high-risk group were significantly older than those in the low-risk group(59 vs.41,P<0.001),hypertension,gender,and Type 2 diabetes did not differ significantly(P values were 0.084,0.724 and 0.488,respectively).9 variables were selected from the inter-group comparison and Lasso regression,including ALB,HCT,NRS2002 scores,HBDH,RDW-SD,RDW-CV,TG,CONUT scores,and NEFA.Multivariate Logistic regression analysis showed that ALB,NRS2002 scores,RDW-SD,and CONUT scores were independent influencing factors for the high risk of VTE scores in patients with tuberculosis(P<0.005).Area under the ROC curve showed that the AUC(0.892)for high-risk VTE scores in patients with the pulmonary tuberculosis was greater than that of ALB(0.803),NRS2002 score(0.735),RDW-SD(0.685),and CONUT score(0.774).Fitting prediction model:Logit(P):Y=0.433×NRS-0.136×ALB+0.411×CONUT score+0.072×RDW-SD-1.770,P=1/(1+e-Y)(Y:prediction index,P:prediction probability).Calibration curve showed that the model prediction tended to be consistent with the actual results(U:>0.05),and the decision curve and influence curve showed that the model can bring clinical benefits.Conclusion ALB,NRS2002 scores,RDW-SD,and CONUT scores are independent influencing factors for the high risk of VTE scores in patients with tuberculosis.They can guide the clinical practice,improve these indicators as soon as possible,reduce VTE scores,and reduce the thrombosis risk.At the same time,the prediction model performs well in the verification cohort,with its discrimination ability,calibration accuracy and clinical utility(decision curve analysis)all reaching a satisfactory level.

Objective: To understand current state of physical health levels of first year students in different body mass index (BMI) categories in Ningxia universities, and to explore the correlation between BMI and physical fitness index (PFI), so as to provide a reference for enhancing physical health levels of university students. Methods: In November 2024, physical fitness test data from 16 631 first year students across four universities in Yinchuan City, Ningxia from 2019 to 2023 were collected by adopting convenience and stratified cluster random sampling methods. The PFI was calculated using the Z score of the physical fitness test results, and a nonlinear quadratic model was established via least squares regression to examine the relationship between BMI and PFI among university students. Results: The BMI for males was (21.69±3.53)kg/m 2, while for females was (20.78±2.94)kg/m 2. The composite score for males physical fitness (69.86±9.25) was lower than that for females (72.24± 8.15 ), with a statistically significant difference ( t =-17.54, P <0.01). Moreover, the failure rates of various physical fitness indicators (vital capacity, sit and reach, standing long jump, pull ups/1 minute sit ups, 1 000 m/800 m run) were higher among males than females ( χ 2=103.48, 72.45, 14.38, 5 134.85, 188.89, all P <0.01). Comparisons across BMI categories revealed that among males, the normal weight group outperformed other groups in the 50 m sprint, standing long jump, 1 000 m sprint, composite score, and PFI ( F =89.17, 113.90, 179.02, 573.35, 593.08); among female students, the normal weight group outperformed other groups in the 50 m sprint, sit and reach, 800 m run, composite score, and PFI ( F =10.67, 19.58 , 96.45, 294.05, 183.45) (all P <0.01). The relationship between BMI and PFI among first year students exhibited a parabolic change trend, students with a moderate BMI demonstrated higher PFI, and as BMI increased, PFI decreased (all P <0.01). Conclusions The physical health level of male students in Ningxia universities is lower than that of female students. There is a correlation between BMI classification and PFI. Tailored intervention measures should be implemented according to the physical characteristics of students across different genders and BMI classifications to enhance university students physical health.