ObjectiveThe essence of syndrome manifestation recognition in traditional Chinese medicine (TCM) is to infer the body’s latent pathogenesis state from clinical observational information, rather than to perform simple label matching. However, previous studies have largely modeled this task as syndrome pattern classification within a fixed label space, which does not adequately reflect the cognition process of TCM syndrome differentiation centered on pathogenesis reasoning, and is also insufficient to capture the openness, semantic variability, and cross-disease reusability of syndrome manifestation expression. This study aimed to investigate whether introducing pathogenesis reasoning chain-of-thought (PR-CoT) supervision into large language models (LLMs) could improve the quality and cognitive consistency of syndrome manifestation recognition and support cross-disease transfer. MethodsSyndrome manifestation recognition was formulated as a conditional generation task under the framework of clinical observational information (X)→pathogenesis structure (Z)→syndrome pattern output (Y), where Z serves as an explicit intermediate structural variable linking the clinical evidence and syndrome judgment. Within this framework, a PR-CoT-supervised dataset for syndrome manifestation recognition was constructed based on medical case records of spleen-stomach disorders. After preprocessing, information extraction, manual proofreading, and data cleaning, the dataset comprised 4 800 training cases, 400 development cases, and 400 test cases. Each sample was annotated with a structured PR-CoT consisting of three progressive levels: clinical information summarization, comprehensive pathogenesis analysis, and syndrome pattern output. Supervised fine-tuning was conducted on open-source LLMs, with an end-to-end model serving as the baseline. Qwen3-32B was used as the primary experimental model, and Qwen3-14B as the scale comparison model. A progressive multidimensional evaluation framework was further established, comprising a structural parsing level, a semantic similarity level, and an expert blind review level. At the structural parsing level, syndrome pattern expressions were decomposed into structural elements and evaluated using Precision, Recall, F1 score, and Jaccard similarity. At the semantic similarity level, independent LLMs scored the theoretical proximity between predicted and reference syndrome patterns. At the expert blind review level, three TCM experts independently evaluated model outputs on two dimensions: syndrome differentiation consistency and terminology standardization of syndrome patterns. In addition, zero-shot cross-disease transfer evaluation was conducted on gynecological and heart-system disorder test sets. ResultsAt the structural parsing level, PR-CoT supervision did not lead to a stable improvement in the element-wise overlap of syndrome pattern structural components. Compared with the corresponding baselines, neither Qwen3-32B nor Qwen3-14B showed consistent advantages in structural matching metrics after the introduction of PR-CoT supervision. In contrast, at the semantic similarity level, PR-CoT supervision produced stable positive gains across different model scales and evaluation systems. The average semantic score of Qwen3-32B increased from 6.425 8 in the baseline model to 6.585 0 after PR-CoT supervision, and that of Qwen3-14B increased from 5.870 0 to 5.964 2. At the expert blind review level, the overall score of Qwen3-32B (PR-CoT) was 7.026 0±0.107 7, higher than 6.416 3±0.288 9 for its baseline. In zero-shot cross-disease testing, the PR-CoT model still showed advantages in semantic evaluation and expert evaluation on both gynecological and heart-system disorder test sets, indicating a certain degree of transferability. ConclusionThe benefits of PR-CoT supervision are mainly reflected in TCM semantic consistency and clinical plausibility, rather than in improved hard matching of structural elements. These findings support understanding syndrome manifestation recognition as a process of generating and expressing latent pathogenesis structures, rather than as a classification task within a traditional fixed label space. By introducing pathogenesis reasoning as an explicit intermediate structure into the modeling process and combining it with a progressive multidimensional evaluation framework, this study provides a methodological pathway for intelligent TCM syndrome differentiation that integrates theoretical alignment, interpretability, and multi-level evaluation.

ObjectiveThe essence of syndrome manifestation recognition in traditional Chinese medicine (TCM) is to infer the body’s latent pathogenesis state from clinical observational information, rather than to perform simple label matching. However, previous studies have largely modeled this task as syndrome pattern classification within a fixed label space, which does not adequately reflect the cognition process of TCM syndrome differentiation centered on pathogenesis reasoning, and is also insufficient to capture the openness, semantic variability, and cross-disease reusability of syndrome manifestation expression. This study aimed to investigate whether introducing pathogenesis reasoning chain-of-thought (PR-CoT) supervision into large language models (LLMs) could improve the quality and cognitive consistency of syndrome manifestation recognition and support cross-disease transfer. MethodsSyndrome manifestation recognition was formulated as a conditional generation task under the framework of clinical observational information (X)→pathogenesis structure (Z)→syndrome pattern output (Y), where Z serves as an explicit intermediate structural variable linking the clinical evidence and syndrome judgment. Within this framework, a PR-CoT-supervised dataset for syndrome manifestation recognition was constructed based on medical case records of spleen-stomach disorders. After preprocessing, information extraction, manual proofreading, and data cleaning, the dataset comprised 4 800 training cases, 400 development cases, and 400 test cases. Each sample was annotated with a structured PR-CoT consisting of three progressive levels: clinical information summarization, comprehensive pathogenesis analysis, and syndrome pattern output. Supervised fine-tuning was conducted on open-source LLMs, with an end-to-end model serving as the baseline. Qwen3-32B was used as the primary experimental model, and Qwen3-14B as the scale comparison model. A progressive multidimensional evaluation framework was further established, comprising a structural parsing level, a semantic similarity level, and an expert blind review level. At the structural parsing level, syndrome pattern expressions were decomposed into structural elements and evaluated using Precision, Recall, F1 score, and Jaccard similarity. At the semantic similarity level, independent LLMs scored the theoretical proximity between predicted and reference syndrome patterns. At the expert blind review level, three TCM experts independently evaluated model outputs on two dimensions: syndrome differentiation consistency and terminology standardization of syndrome patterns. In addition, zero-shot cross-disease transfer evaluation was conducted on gynecological and heart-system disorder test sets. ResultsAt the structural parsing level, PR-CoT supervision did not lead to a stable improvement in the element-wise overlap of syndrome pattern structural components. Compared with the corresponding baselines, neither Qwen3-32B nor Qwen3-14B showed consistent advantages in structural matching metrics after the introduction of PR-CoT supervision. In contrast, at the semantic similarity level, PR-CoT supervision produced stable positive gains across different model scales and evaluation systems. The average semantic score of Qwen3-32B increased from 6.425 8 in the baseline model to 6.585 0 after PR-CoT supervision, and that of Qwen3-14B increased from 5.870 0 to 5.964 2. At the expert blind review level, the overall score of Qwen3-32B (PR-CoT) was 7.026 0±0.107 7, higher than 6.416 3±0.288 9 for its baseline. In zero-shot cross-disease testing, the PR-CoT model still showed advantages in semantic evaluation and expert evaluation on both gynecological and heart-system disorder test sets, indicating a certain degree of transferability. ConclusionThe benefits of PR-CoT supervision are mainly reflected in TCM semantic consistency and clinical plausibility, rather than in improved hard matching of structural elements. These findings support understanding syndrome manifestation recognition as a process of generating and expressing latent pathogenesis structures, rather than as a classification task within a traditional fixed label space. By introducing pathogenesis reasoning as an explicit intermediate structure into the modeling process and combining it with a progressive multidimensional evaluation framework, this study provides a methodological pathway for intelligent TCM syndrome differentiation that integrates theoretical alignment, interpretability, and multi-level evaluation.

Objective The aim of this study was to investigate the correlation between the interaction of uric acid and inflammatory factors and hyperuricemia in overweight/obese patients. Methods The personnel with hyperuricemia who underwent physical examination in our hospital from September 2021 to September 2022 were selected as the study subjects, and they were divided into 100 cases of overweight group and 90 cases of obese group according to the BMI index; 120 cases of healthy and non-hyperuricemic personnel were randomly selected as the control group; venous blood of the three groups was collected in 5 mL after 8 h of fasting, and were tested respectively for serum uric acid, lipid indexes and inflammatory factors: IL-6, IL-2, IFN-γ, TNF-α, IL-4, IL-10. Results Glucose, triglycerides, total cholesterol, and LDL were significantly higher in the obese group versus the overweight group (P<0.001), while HDL was significantly lower than the control group (P<0.001), and these changes were more pronounced in the obese group (P<0.001).The Pearson correlation coefficient pointed out that the levels of serum uric acid in patients with hyperuricosuric acid were significantly associated with the pro-inflammatory factors IL- 6, IL-2, IFN-γ, and TNF-α were significantly positively correlated (P<0.001), whereas they were significantly negatively correlated with the anti-inflammatory factors IL-4, IL-10 (P<0.001). Conclusion High uric acid levels in overweight/obese patients can cause enhanced inflammatory responses and reduced expression levels of anti-inflammatory factors, and the interaction between uric acid and pro-inflammatory factors aggravates the condition of patients with hyperuricemia.

Objective: To investigate the characteristics and potential risk factors associated with HIV infection among newly reported HIV positive male student cases in Jiangsu Province from 2023 to 2024, so as to provide evidence for targeted intervention strategies. Methods: Data were obtained from the China CDC Surveillance System on newly reported HIV positive male student cases from 2023 to 2024. A survey was conducted to collect information on demographic characteristics, knowledge of AIDS prevention, education and training history, HIV testing history, behavioral and substance use patterns, and other relevant factors prior to HIV diagnosis among 343 newly reported HIV positive male student cases in Jiangsu Province. Multivariate Logistic regression analysis was used to assess the risk factors of HIV infection among male students. Results: Among the HIV positive male student cases, homosexual behavior accounted for 93.88% of transmission routes, while 10.20% involved heterosexual contact, as well as 4.08% for two sexual hehaviors. Awareness of HIV prevention knowledge was 97.08%, and 66.76% had previously undergone HIV testing. Among the respondents, 10.50% had used rush poppers as enhancers, and 72.30% had received HIV prevention education within the past year. Among students cases with homosexual behavior, the median time from first homosexual contact to HIV diagnosis was ≤2 years, with 54.66% of cases falling into this category; the most common way of finding same sex partners was through social software, accounting for 88.20% of cases, while the proportion of those via "Blued" app reached 87.07% ; the proportion of using condoms every time during sexual activity in the past six months was 12.27%. Among the student cases with homosexual behavior, the results of multiple Logistic regression analysis showed that student cases aged 18 to 24 ( OR =4.52) and >24 ( OR = 19.23 ), without receiving education on HIV prevention in the past year ( OR =1.86), having consistent condom use ( OR =2.73) and not using condoms ( OR =2.12) during the last sexual activity were more likely to had the first same sex sexual activity for more than 3 years before being diagnosed as HIV positive cases (all P <0.05). Student cases who were uncertain about their partner s sexual identity ( OR =0.33), and who primarily identified same sex partners through "other" means ( OR = 0.23 ) were more likely to avoid HIV testing; in contrast, student cases with consistent condom use during the last homosexual encounter ( OR =7.20) was significantly associated with increased likelihood of HIV testing (all P <0.05). Conclusions Newly reported HIV positive male student cases in Jiangsu Province exhibit serious discrepancies between knowledge and practice regarding HIV prevention. Measures are needed to accelerate the optimization of campus based HIV prevention education content and delivery methods. Simultaneously, enhanced management of extracurricular male populations is essential to effectively control the spread of HIV.

Objective:To explore the interrelationship between brain functional networks and features in functional magnetic resonance imaging(fMRI)of patients with Alzheimer's disease(AD),and to construct mixed-function networks(MFN),and apply them in machine learning classification models to improve the accuracy of AD classification.Methods:102 AD patients and 227 healthy subjects in the Alzheimer's Neuroimaging Initiative(ADNI)dataset were retrospectively analyzed.The partial correlation brain network of the blood oxygen level dependent(BOLD)signal was calculated and fused with low-frequency wave amplitude(ALFF),fractional low-frequency wave amplitude(fALFF)and local consistency(ReHo)features to construct MFN.Network topology parameters were extracted,and a variety of machine learning classification models were constructed based on MFN topological parameters,accuracy,precision,recall and area under the curve(AUC)were used to evaluate the predictive efficiency of the models.Results:By constructed MFN and calculated intra group to inter group ratio(IIGR),35 features could be obtained from ALFF,fALFF and ReHo feature topological parameter analysis,after rank sum test and FDR correction,there were statistical differences among 28 features(P＜0.05).The classification results show that,all the five classifiers have high classification performance on the test data set.The accuracy,precision and recall rates of random forest(RF),adaptive lifting algorithm(AdaBoost),guided aggregation algorithm(Bagging)and support vector machine(SVM)were all 99.7％,and the AUC values were up to 100％,99.5％,99.1％and 99.5％,respectively.The accuracy(98.5％),precision(98.5％),recall(98.5％),and AUC(99.1％)of the multi-layer perceptron(MLP)were slightly lower than other models,but remained excellent.It was worth noting that RF has the highest AUC value of all models at 100.0％,while Bagging has the lowest AUC value(99.1％)in the integrated approach.The results of performance comparison show that,MFN classification model can significantly improve the recognition and classification of AD disease,and greatly improve the performance of various indicators of the classifier.The results showed that,MFN classification model was superior to intelligent classification based fusion,DBN-based multitask learning,PVT-TSVM,unsupervised learning and clustering,SVM and SVM of degree 3 polynomial kernel function in key indicators such as accuracy(99.13％),AUC(99.42％),recall rate(99.46％)and specificity(99.42％)with plasma proteins,machine learning algorithms.It was further proved that MFN classification model has good generalization ability and robustness in AD disease classification.Conclusion:The AD classification model constructed based on brain mixed function network topology parameters and machine learning can improve the accuracy of AD classification.

Objective To explore the effectiveness of the OKR method in improving the efficiency of the whole process of critical value management,and to provide new ideas for the implementation of medical quality improvement in other medical insti-tutions.Methods A hospital in Tianjin was selected as the study object,which used the OKR method to reform the management of critical value since January 2024.The core goal of"improving the effectiveness of critical value management"was set,and the goal was broken down into quantitative key results such as"the rate of receiving the critical value system is over 95％"and"the rate of completing the standardised writing of medical records is up to 85％".Results After the reform,several quantitative key results,such as the completion rate of critical care medical record writing,the rate of standardised medical record writing,and the rate of overtime acceptance,were all better than before.Conclusion Through the OKR method to unify the whole hospital's strategic objectives,and the dynamic adjustment of the program based on data review,the hospital's critical value management efficiency has been significantly improved,effectively guaranteeing the safety of patients,and providing new perspectives and methods for the management of critical value in other medical institutions.

Objective:To gain a understanding of the occurrence of cognitive impairment among residents in drinking water-borne endemic fluorosis (drinking water-borne fluorosis) areas, and to study its influencing factors.Methods:In March 2023, a cluster sampling method was used to select local residents aged 18 and above from the drinking water-borne fluorosis areas in Jishan County, Shanxi Province as survey subjects. General demographic data were collected through face-to-face surveys, and a random urine sample was collected once to determine urinary fluoride level. Cognitive function was assessed using the mini-mental state examination (MMSE), and the survey subjects were divided into a cognitive impairment group ( < 27 points) and a control group (27 - 30 points) based on the MMSE scores. A multiple logistic regression model and a decision tree model based on chi-squared automatic interaction detector were constructed to analyze the factors affecting cognitive impairment, and the model fitting effect was evaluated using receiver operating characteristic (ROC) curve.Results:A total of 3 301 subjects were included in the survey, including 2 081 females and 1 220 males. There were 1 515 subjects < 60 years old and 1 786 subjects ≥60 years old, with urinary fluoride level [ M ( Q1, Q3)] of 2.92 (1.78, 4.54) mg/L. There were 1 939 cases in the cognitive impairment group and 1 362 cases in the control group, with a detection rate of 58.74% (1 939/3 301) for cognitive impairment; and the differences in gender, age, education level, marital status, annual household income, alcohol consumption, smoking distribution, and urinary fluoride level between the two groups were statistically significant ( P < 0.05). The results of multiple logistic regression analysis showed that female, ≥60 years, and urinary fluoride > 4.54 mg/L were risk factors for cognitive impairment [ OR (95% CI): 1.25 (1.01, 1.54), 2.66 (2.26, 3.14), 1.32 (1.06, 1.65), P < 0.05]. Education level of primary school or above, annual household income≥12 000 yuan, and mild alcohol consumption were protective factors for cognitive impairment [ OR (95% CI): 0.15 (0.09, 0.25), 0.58 (0.48, 0.68), 0.67 (0.51, 0.87), P < 0.05]. The analysis results of the decision tree model showed that age had the greatest impact on the occurrence of cognitive impairment, followed by annual household income, education level, and urinary fluoride. The areas under the ROC curves of the multiple logistic regression and decision tree model were 0.72 and 0.70 ( P < 0.001), respectively, indicating good model fitting performance. Conclusion:The detection rate of cognitive impairment in residents of drinking water-borne fluorosis areas is relatively high, and age, annual household income, education level, and urinary fluoride are all influencing factors for occurrence of cognitive impairment.

Aim To explore the effect of astragalosideⅣ(AS-Ⅳ)on lipopolysaccharide(LPS)-induced po-larization and migration of RAW 264.7 macrophages and the underlying mechanism.Methods 1 mg·L-1 LPS was used to construct cell migration model.Scratch assay was utilized to determine cell migration rate.Immunofluorescence staining was utilized to de-tect the expression and location of F4/80,iNOS and Arg-1.CCK-8 assay was used to determine the viabili-ty of RAW 264.7 cells.Griess assay was used to measure NO content.Molecular docking was used to analyze the interaction between AS-Ⅳ and the core tar-gets such as cGAS and STING protein.Western blot was employed to detect the expression of iNOS,Arg-1,cGAS,STING,NF-κB p65 and p-NF-κB p65 protein.Results AS-Ⅳ significantly inhibited the migration and M1 polarization of RAW 264.7 cells induced by LPS.Moreover,AS-Ⅳ could interact with cGAS and STING protein,especially cGAS.Further Western blot assay showed that AS-Ⅳ significantly downregulated the expression of iNOS,cGAS,STING and p-NF-κB p65 protein.Conclusions AS-Ⅳ could promote mac-rophage M1 to M2 polarization,thereby inhibited mac-rophage migration through restraining the cGAS/STING/NF-κB signaling pathway,which provides a new therapeutic target for AS-Ⅳ to improve the early inflammatory response of AS.

Purpose To investigate the change of network topology attribute of brain structure in obese adolescents using graph theory analysis.Materials and Methods A total of 86 obese adolescents in Children's Hospital of Nanjing Medical University from January 2023 to April 2024 were prospectively collected,including 60 patients in the simple obesity group(OB group),26 patients in the obese group with metabolic syndrome(MS group).Meanwhile,24 healthy volunteers were recruited as the control group.Diffusion tensor imaging scan of the head,anthropological characteristics examination and blood test were performed.The differences of global and local topology properties among the three groups were compared.Global or node attributes with statistical differences were correlated with clinical indicators.Results There were significant differences in body height,body weight,body mass index,waist circumference,diastolic blood pressure,systolic blood pressure and high density lipoprotein cholesterol among the three groups(χ2/F=3.208-5.085,all P<0.05).Compared with the control group,there were significant differences in total body fat,triglyceride,total cholesterol,alanine amino-transferase,aspartate aminotransferase and fasting insulin between the OB group and the MS group(all P<0.05).No statistically significant differences were observed in the small world attribute,shortest path length and global efficiency among the three groups(all P>0.05).A statistically significant difference was observed in the clustering coefficient and local efficiency among the three groups(F=2.534,3.012,both P<0.05),with the values of clustering coefficient and local efficiency of OB group were significantly lower than those of the control group(P<0.05).The node attribute differences occurred in the left frontal lobe,right insula,right anterior central gyrus,left anterior cingulate gyrus,left anterior cuneus and bilateral occipital lobes.The connection value between module Association and Limbic in control group was significantly lower smaller than those of the OB group(P=0.033).A negative correlation was observed between local efficiency and age(r=-0.239),body weight(r=-0.254),body mass index(r=-0.230)and waist circumference(r=-0.263,all P<0.05).The modularity was negatively correlated with fasting insulin(r=-0.325,P<0.05).Conclusion Adolescent obesity leads to the abnormal topological properties of structural network in some brain regions,which may be associated with the abnormal cognitive function and could provide neuroimaging evidence for clinical intervention.

Objective:To summarize the current situation and shortcomings of China's health technology innovation policies,and provide reference for policy formulation and improvement.Methods:Text mining was used to sort out 24 policy documents related to health technology innovation issued by the national and provincial levels since the 13th Five Year Plan period.A PMC index evaluation model for health technology innovation policies was established,and a quantitative analysis of health technology innovation policies was conducted through 9 primary indicators and 43 secondary indicators.Results:Among the 24 policies,2 were rated as perfect,8 were rated as excellent,and 14 were rated as acceptable.Conclusions and Suggestions:China's policies on health and medical science and technology innovation have been basically improved.They can be further refined by focusing on core and key technologies,emphasizing clinical research and transformation,and advancing digital and intelligent strategies.