ObjectiveTo address the challenges of unstructured classical Chinese expressions， nested entity relationships， and limited annotated data in famous traditional Chinese medicine（TCM） case records， this study proposes a joint relation extraction framework that integrates data augmentation and entity mapping， aiming to support the construction of TCM diagnostic knowledge graphs and clinical pattern mining. MethodsWe developed an annotation structure for entities and their relationships in TCM case texts and applied a data augmentation strategy by incorporating multiple ancient texts to expand the relation extraction dataset. A cascade binary tagging framework for relation triple extraction（CasRel） model for TCM semantics was designed， integrating a pre-trained bidirectional encoder representations from transformers（BERT） layer for classical TCM texts to enhance semantic representation， and using a head entity-relation-tail entity mapping mechanism to address entity nesting and relation overlapping issues. ResultsExperimental results showed that the CasRel model， combining data augmentation and entity mapping， outperformed the pipeline-based Bert-Radical-Lexicon（BRL）-bidirectional long short-term memory（BiLSTM）-Attention model. The overall precision， recall， and F₁-score across 12 relation types reached 65.73%， 64.03%， and 64.87%， which represent improvements of 14.26%， 7.98%， and 11.21% compared to the BRL-BiLSTM-Attention model， respectively. Notably， the F₁-score for tongue syndrome relations increased by 22.68%（69.32%）， and the prescription-syndrome relations performed the best with the F₁-score of 70.10%. ConclusionThe proposed framework significantly improves the semantic representation and complex dependencies in TCM texts， offering a reusable technical framework for structured mining of TCM case records. The constructed knowledge graph can support clinical syndrome differentiation， prescription optimization， and drug compatibility， providing a methodological reference for TCM artificial intelligence research.

ObjectiveTo explore the mechanism by which Sini San （SNS） inhibits ferroptosis， alleviates inflammation and myocardial injury， and improves myocardial infarction （MI）. MethodsThe active ingredients of SNS were obtained by searching the Traditional Chinese Medicine System Pharmacology Platform （TCMSP） database， its target sites were predicted using the SwissTargetPrediction Database， and the core components were screened out using the CytoNCA plug-in. The targets of MI and ferroptosis were obtained by using GeneCards， Online Mendelian Inheritance in Man （OMIM） database， DrugBank， Therapeutic Target Database （TTD）， FerrDb database and literature review， respectively. The intersection of these targets of SNS-MI-ferroptosis was plotted as a Venn diagram. The protein-protein interaction （PPI） network was constructed using the STRING database， and the visualization graph was prepared using Cytoscape. The core targets were screened out using the CytoNCA plug-in， and the biological functions were clustered by the MCODE plug-in. Gene Ontology （GO） functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes （KEGG） pathway enrichment analysis were performed using the David database. Molecular docking was performed using AutoDock and visualized with PyMOL2.5.2. The Kunming mice were randomly divided into the control group， the model group， the SNS group， and the trimetazidine （TMZ） group. The mice were subcutaneously injected with isoprenaline （ISO， 5 mg·kg^-1·d^-1） to establish an MI model. The drug was continuously intervened for 7 days. The ST-segment changes were recorded by electrocardiogram （ECG）， and the tissue morphology changes were observed by hematoxylin-eosin （HE） staining. Cardiomyocyte ferroptosis was investigated by transmission electron microscopy. Serum creatine kinase （CK）， creatine kinase isoenzyme （CK-MB）， lactate dehydrogenase （LDH）， reduced glutathione （GSH）， and malondialdehyde （MDA） levels were detected by biochemical assay. Enzyme-linked immunosorbent assay （ELISA） was used to detect serum levels of interleukin （IL）-6 and 4-hydroxynonenal （4-HNE）. Immunohistochemical staining was employed to detect IL-6 and phosphorylated signal transducer and transcription activator 3 （p-STAT3） in cardiac tissues. Western blot was used to detect STAT3 and p-STAT3 in cardiac tissues. Real-time PCR was used to detect the levels of IL-6， IL-18， solute carrier family 7 member 11 （SLC7A11）， arachidonic acid 15-lipoxygenase （ALOX15）， and glutathione peroxidase 4 （GPx4） in cardiac tissues. ResultsA total of 121 active ingredients of SNS were obtained， and 58 potential targets of SNS in the treatment of MI by regulating ferroptosis were screened. The three protein modules with a score5 were mainly related to the inflammatory response. The GO function was mainly related to inflammation， and KEGG enrichment analysis showed that SNS mainly regulated ferroptosis- and inflammation- related signaling pathways. Molecular docking indicated that the core component had a higher binding force to the target site. Animal experiments confirmed that SNS reduced the level of p-STAT3 （P0.01）， down-regulated the expression of ALOX15 mRNA （P0.01）， up-regulated the level of serum GSH， and the expressions of SLC7A11 and GPx4 mRNA， reduced MDA and 4-HNE levels （P0.05， P0.01）. Additionally， SNS improved the mitochondrial injury induced by cardiomyocyte ferroptosis， reduced the area of MI， alleviated inflammation and myocardial injury， lowered the levels of serum CK， CK-MB， LDH， IL-6， and the mRNA expression levels of IL-16 and IL-18 （P0.05）， and improved ST segment elevation. ConclusionSNS can reduce ISO-induced STAT3 phosphorylation levels， inhibit ferroptosis in cardiomyocytes， alleviate inflammation and myocardial injury， thereby improving MI.

ObjectiveTo explore the mechanism by which Sini San （SNS） inhibits ferroptosis， alleviates inflammation and myocardial injury， and improves myocardial infarction （MI）. MethodsThe active ingredients of SNS were obtained by searching the Traditional Chinese Medicine System Pharmacology Platform （TCMSP） database， its target sites were predicted using the SwissTargetPrediction Database， and the core components were screened out using the CytoNCA plug-in. The targets of MI and ferroptosis were obtained by using GeneCards， Online Mendelian Inheritance in Man （OMIM） database， DrugBank， Therapeutic Target Database （TTD）， FerrDb database and literature review， respectively. The intersection of these targets of SNS-MI-ferroptosis was plotted as a Venn diagram. The protein-protein interaction （PPI） network was constructed using the STRING database， and the visualization graph was prepared using Cytoscape. The core targets were screened out using the CytoNCA plug-in， and the biological functions were clustered by the MCODE plug-in. Gene Ontology （GO） functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes （KEGG） pathway enrichment analysis were performed using the David database. Molecular docking was performed using AutoDock and visualized with PyMOL2.5.2. The Kunming mice were randomly divided into the control group， the model group， the SNS group， and the trimetazidine （TMZ） group. The mice were subcutaneously injected with isoprenaline （ISO， 5 mg·kg^-1·d^-1） to establish an MI model. The drug was continuously intervened for 7 days. The ST-segment changes were recorded by electrocardiogram （ECG）， and the tissue morphology changes were observed by hematoxylin-eosin （HE） staining. Cardiomyocyte ferroptosis was investigated by transmission electron microscopy. Serum creatine kinase （CK）， creatine kinase isoenzyme （CK-MB）， lactate dehydrogenase （LDH）， reduced glutathione （GSH）， and malondialdehyde （MDA） levels were detected by biochemical assay. Enzyme-linked immunosorbent assay （ELISA） was used to detect serum levels of interleukin （IL）-6 and 4-hydroxynonenal （4-HNE）. Immunohistochemical staining was employed to detect IL-6 and phosphorylated signal transducer and transcription activator 3 （p-STAT3） in cardiac tissues. Western blot was used to detect STAT3 and p-STAT3 in cardiac tissues. Real-time PCR was used to detect the levels of IL-6， IL-18， solute carrier family 7 member 11 （SLC7A11）， arachidonic acid 15-lipoxygenase （ALOX15）， and glutathione peroxidase 4 （GPx4） in cardiac tissues. ResultsA total of 121 active ingredients of SNS were obtained， and 58 potential targets of SNS in the treatment of MI by regulating ferroptosis were screened. The three protein modules with a score5 were mainly related to the inflammatory response. The GO function was mainly related to inflammation， and KEGG enrichment analysis showed that SNS mainly regulated ferroptosis- and inflammation- related signaling pathways. Molecular docking indicated that the core component had a higher binding force to the target site. Animal experiments confirmed that SNS reduced the level of p-STAT3 （P0.01）， down-regulated the expression of ALOX15 mRNA （P0.01）， up-regulated the level of serum GSH， and the expressions of SLC7A11 and GPx4 mRNA， reduced MDA and 4-HNE levels （P0.05， P0.01）. Additionally， SNS improved the mitochondrial injury induced by cardiomyocyte ferroptosis， reduced the area of MI， alleviated inflammation and myocardial injury， lowered the levels of serum CK， CK-MB， LDH， IL-6， and the mRNA expression levels of IL-16 and IL-18 （P0.05）， and improved ST segment elevation. ConclusionSNS can reduce ISO-induced STAT3 phosphorylation levels， inhibit ferroptosis in cardiomyocytes， alleviate inflammation and myocardial injury， thereby improving MI.

To summarize the clinical experience of treating idiopathic olfactory disorder from the perspective of heart and lungs. It is believed that the sense of olfaction is based on the nose， rooted in the heart， functioning through the lungs， and conveyed by pectoral qi. The key pathogenesis of idiopathic olfactory disorder lies in the accumulation of pathogenic factors in the heart and lungs， blood vessel obstruction， the failure of the lungs to disperse and descend， and the loss of control of pectoral qi. In treatment， internal and external therapies could be combined. The internal therapy can correct the imbalance of the zang-fu organs， using the self-prescribed Bilong Formula （鼻聋方） to dispel wind and diffuse the lung， invigorate blood and relieving stuffy orifices； the external therapy can clear nasal obstruction， supplemented by intradermal needle embedding at three acupoints， bilateral Yingxiang （LI 20）， bilateral Shangyingxiang （EX-HN 8） and Yintang （GV 29）， and integrated olfactory training for comprehensive treatment.

To explore the advantages of traditional Chinese medicine （TCM） and integrative TCM-Western medicine approaches in the treatment of diabetic microvascular complications （DMC）， refine key pathophysiological insights and treatment principles， and promote academic innovation and strategic research planning in the prevention and treatment of DMC. The 38th session of the Expert Salon on Diseases Responding Specifically to Traditional Chinese Medicine， hosted by the China Association of Chinese Medicine， was held in Beijing， 2024. Experts in TCM， Western medicine， and interdisciplinary fields convened to conduct a systematic discussion on the pathogenesis， diagnostic and treatment challenges， and mechanism research related to DMC， ultimately forming a consensus on key directions. Four major research recommendations were proposed. The first is addressing clinical bottlenecks in the prevention and control of DMC by optimizing TCM-based evidence evaluation systems. The second is refining TCM core pathogenesis across DMC stages and establishing corresponding "disease-pattern-time" framework. The third is innovating mechanism research strategies to facilitate a shift from holistic regulation to targeted intervention in TCM. The fourth is advancing interdisciplinary collaboration to enhance the role of TCM in new drug development， research prioritization， and guideline formulation. TCM and integrative approaches offer distinct advantages in managing DMC. With a focus on the diseases responding specifically to TCM， strengthening evidence-based support and mechanism interpretation and promoting the integration of clinical care and research innovation will provide strong momentum for the modernization of TCM and the advancement of national health strategies.

BackgroundLow medication compliance among patients with severe mental disorders increases the disease burden on both the patients' families and the society. Medication adherence is influenced by numerous factors. Traditional methods such as Logistic regression struggle to quantify the importance of these factors. By introducing Extreme Gradient Boosting （XGBoost） combined with Shapley Additive Explanations （SHAP）， enables the quantification of the relative contribution weights of each factor， providing support for identifying the core influencing factors. ObjectiveTo explore the influencing factors of medication adherence among patients with severe mental disorders in Zhuhai， aiming to provide references for optimizing patient management strategies. MethodsExtract the data of patients with severe mental disorders who were registered on the mental health system platform in Zhuhai City from January 1， 2023 to March 31， 2025. A total of 9 329 patients were finally included for analysis. Influencing factors were screened using univariate analysis and multivariate logistic regression analysis， and an XGBoost model combined with the SHAP algorithm was constructed to quantify the importance of each influencing factor. ResultsAmong 9 329 patients， 8 446 demonstrated medication adherence， yielding an adherence rate of 90.53%. Multivariable analysis identified several risk factors significantly associated with medication non-adherence， being unmarried （OR=1.237， 95% CI： 1.019–1.502） or divorced （OR=1.389， 95% CI： 1.038–1.832）， a diagnosis of mental retardation with psychiatric disorders （OR=3.025， 95% CI： 2.402–3.796） or paranoid psychosis （OR=5.117， 95% CI： 3.086–8.299）， a disease duration of 2–4 years （OR=1.355， 95% CI： 1.085–1.696）， 4–6 years （OR=2.143， 95% CI： 1.671–2.747）， or >6 years （OR=1.681， 95% CI： 1.365–2.079）， lack of guardian subsidies （OR=1.412， 95% CI： 1.099–1.801）， absence of a disability certificate （OR=1.900， 95% CI： 1.588–2.282）， not being enrolled in care and support groups （OR=1.384， 95% CI： 1.183–1.617） or community services （OR=1.313， 95% CI： 1.042–1.645）， and not cohabiting with a guardian （OR=1.257， 95% CI： 1.048–1.501）. Conversely， the enrollment in special outpatient disease programs （OR=0.716， 95% CI： 0.609–0.842） and a family history of mental illness （OR=0.713， 95% CI： 0.503–0.982） were identified as protective factors. The XGBoost model exhibited robust predictive performance， with a sensitivity of 0.433， specificity of 0.944， accuracy of 0.891， Area Under the Curve （AUC） of 0.837， and F1 value of 0.449. Feature importance ranking indicated that the top three factors were disease duration， diagnosis， and the acquisition of disability certificates. ConclusionPolicy-based support （acquisition of disability certificates， special outpatient disease enrollment） and clinical disease characteristics （disease duration， diagnosis type） are key factors affecting medication adherence among patients with severe mental disorders in Zhuhai City. ［Funded by Zhuhai Medical Research Project （number， 2220009000281）］

BACKGROUND:Traumatic brain injury is a condition in which the normal function of the brain is disrupted by a bump or impact to the head.It is necessary to find effective treatments and objective targets that can help doctors diagnose the injury status and restore the brain function of patients. OBJECTIVE:To explore the effect of electroacupuncture combined with low-frequency transcranial ultrasound stimulation on the electroencephalographic signals of rats with traumatic brain injury. METHODS:Forty 6-week-old SPF male Sprague-Dawley rats were randomly divided into five groups:sham group,model group,electroacupuncture group,low-frequency transcranial ultrasound stimulation group and combined group(electroacupuncture+low-frequency transcranial ultrasound stimulation),with eight rats in each group.Feeney weight-drop method was used to establish the animal model of traumatic brain injury.In the sham group,the bone window was only opened without impact.Interventions were started at 1 day after modeling.Electroacupuncture in the electroacupuncture group,low-frequency transcranial ultrasound stimulation in the low-frequency transcranial ultrasound stimulation group,and electroacupuncture+low-frequency transcranial ultrasound stimulation in the combined group were performed for days in total.The modified neurological severity scale score for assessing rats'neurological deficits was performed at 8 hours after modeling.The percentage of spontaneous alternation behavior in the Y-maze was measured at 7 days after modeling.Then,the electroencephalographic signals were collected and electroencephalographic data of α,β,θ,and δ waves were extracted by fast Fourier transform,and the value of oscillation amplitude and energy ratio were calculated in α,β,θ,and δ waves,as well as the Lempel-Ziv complexity and sample entropy. RESULTS AND CONCLUSION:Compared with the sham group,the modified neurological severity scale scores in the model group,electroacupuncture group,low-frequency transcranial ultrasound stimulation group and combined group were significantly increased at 8 hours after modeling(P<0.05).Compared with the sham group,the value of oscillation amplitude in δ wave and the value of δ energy ratio were significantly increased in the model group at 7 days after modeling,meanwhile the percentage of spontaneous alternation behavior in Y-maze,and the value of α/β energy ratio,Lempel-Ziv complexity,and sample entropy were significantly decreased(P<0.05).Compared with the model group,the value of oscillation amplitude in α and δ waves was significantly decreased in the combined group(P<0.05),while the value of α/β energy ratio was significantly increased(P<0.05)and the value of δ energy ratio was significantly decreased(P<0.05)in the electroacupuncture group,low-frequency transcranial ultrasound stimulation group and combined group.Compared with the electroacupuncture group and low-frequency transcranial ultrasound stimulation group,the value of δ energy ratio was significantly decreased in the combined group(P<0.05),while the percentage of spontaneous alternation behavior,the value of α/β energy ratio,the Lempel-Ziv complexity,and the sample entropy were significantly increased(P<0.05).To conclude,abnormal electroencephalographic signals can appear in rats with traumatic brain injury,while the electroacupuncture combined with low-frequency transcranial ultrasound stimulation can alleviate the abnormal electroencephalographic signals in rats,which suggests the electroencephalographic frequency domain value and nonlinear features can be used to assess the severity of traumatic brain injury.

BACKGROUND:Platelet-rich fibrin(PRF)has many advantages,such as simple preparation,low production cost,and high safety,and has been widely used in the study of bone defect repair in oral and maxillofacial surgery,but there are problems such as too fast degradation rate and short release time of growth factors. OBJECTIVE:PRF was loaded into gelatin methacryloyl(GelMA)hydrogel and its osteogenic properties were analyzed by in vivo and in vitro experiments. METHODS:(1)New Zealand white rabbit venous blood was extracted to prepare PRF.GelMA hydrogels containing 0,0.05,0.075,and 0.1 g PRF were prepared,respectively,and were recorded as GelMA,GelMA/PRF-0.05,GelMA/PRF-0.075,and GelMA/PRF-0.1,respectively,to characterize the micromorphology and in vitro slow-release properties of the hydrogels.(2)Four kinds of hydrogels were co-cultured with MC3T3-E1 cells,respectively,and the cell proliferation activity was detected with the single cultured cells as the control.After osteogenic induction,alkaline phosphatase activity,mineralization ability,mRNA and protein expression levels of osteogenic genes(osteocalcin,osteopontin,RUNX2),ERK1/2-p38 MAPK pathway protein mRNA and protein expression levels were detected.(3)Fifteen New Zealand white rabbits were taken.Four full-layer bone defects of 8 mm diameter were prepared in the skull of each rabbit,one of which was implanted without any material(blank control group),and the other three were implanted with GelMA hydrogel,PRF,and GelMA/PRF-0.1 hydrogel,respectively.The bone defect was detected by Micro-CT and bone morphology was observed at 4,8,and 12 weeks after operation. RESULTS AND CONCLUSION:(1)Scanning electron microscopy observed that all the hydrogels of the four groups had honeycomb pore structure,and the pore size of the hydrogels decreased slightly with the increase of PRF content,but there was no significant difference between the groups.The three groups of GelMA/PRF hydrogel could release transforming growth factor β1 and insulin-like growth factor 1 at a certain rate,and the cumulative release of transforming growth factor β1 and insulin-like growth factor 1 increased significantly with the extension of time.(2)CCK-8 assay and live/dead staining showed that GelMA/PRF hydrogel could promote the proliferation of MC3T3-E1 cells.The results of alkaline phosphatase staining,alizarin red staining,and osteogenic gene detection showed that GelMA/PRF hydrogel could promote the osteogenic differentiation of MC3T3-E1 cells,and inhibit the expression of ERK1/2-p38 MAPK pathway protein,and showed a PRF content dependence.(3)Micro-CT scan showed that the bone mineral density and bone volume fraction in the bone defect of GelMA/PRF-0.1 hydrogel group were higher than those in the other three groups(P<0.05).Hematoxylin-eosin staining showed that compared with the other three groups,GelMA/PRF-0.1 hydrogel group had faster and more mature new bone formation at the bone defect.(4)These findings indicate that GelMA/PRF hydrogel has good osteogenic activity both in vivo and in vitro,which may be related to inhibiting the expression of ERK1/2-p38 MAPK pathway protein.

ObjectiveTo investigate the features of serum HBV RNA in different stages of the natural history of chronic hepatitis B virus （HBV） infection without antiviral treatment， as well as its correlation with serum HBV DNA and HBsAg. MethodsA total of 306 treatment-naïve patients with chronic HBV infection who attended Department of Infections Diseases and Hepatoloty， the Second Hospital of Shandong University from January 2023 to June 2024 were divided into six groups based on the different stages of natural history， i.e.， HBeAg-positive chronic HBV infection group with 29 patients， HBeAg-positive chronic hepatitis B （CHB） group with 107 patients， HBeAg-negative chronic HBV infection group with 18 patients， HBeAg-negative CHB group with 60 patients， HBeAg-positive indeterminate-phase chronic HBV infection group with 7 patients， and HBeAg-negative indeterminate-phase chronic HBV infection group with 85 patients. Real-time isothermal RNA amplification was used to measure serum high-sensitivity HBV RNA. The Kruskal-Wallis H test was used for comparison between multiple groups of continuous data， while the Mann-Whitney U test was used for comparison between two groups. The Spearman method was used to investigate the correlation of HBV RNA with HBV DNA and HBsAg. ResultsThe HBeAg-positive chronic HBV infection group showed the highest level of serum HBV RNA ［7.5 （7.4‍ ‍—‍ ‍7.9） log₁₀ copies/mL］， followed by the HBeAg-positive CHB group ［7.4 （6.4‍ ‍—‍ ‍7.9） log₁₀ copies/mL］， the HBeAg-negative CHB group ［4.5 （3.0‍ ‍—‍ ‍5.7） log₁₀ copies/mL］， and the HBeAg-negative chronic HBV infection group ［1.0 （1.0‍ ‍—‍ ‍2.0） log₁₀ copies/mL］； the HBeAg-positive indeterminate-phase chronic HBV infection group had a serum HBV RNA level of 3.9 （3.7‍ ‍—‍ ‍5.7） log₁₀ copies/mL， and the HBeAg-negative indeterminate-phase chronic HBV infection group had a serum HBV RNA level of 2.0 （1.0‍ ‍—‍ ‍3.0） log₁₀ copies/mL； there was a significant difference in serum HBV RNA level between the six groups （H=830.770， P<0.001）. There was a significant difference in HBV RNA level between the HBeAg-positive chronic HBV infection group and all the other groups except the HBeAg-positive CHB group （all P<0.001）. In the 306 patients with HBV infection， HBV RNA was strongly correlated with HBV DNA （r=0.92， P<0.001） and was moderately correlated with HBsAg （r=0.67， P<0.001）. The correlation between serum HBV RNA and HBsAg in HBeAg-positive patients （r=0.61， P<0.001） was stronger than that in HBeAg-negative patients （r=0.31， P<0.001）. For the patients with HBeAg-positive chronic HBV infection， the male patients with ALT>30 U/L and the female patients with ALT>19 U/L had a significantly lower serum HBV RNA level than the male patients with ALT≤30 U/L and the female patients with ALT≤19 U/L （P<0.001）， and there was no significant difference in serum HBV RNA level between the latter group of patients and the HBeAg-positive CHB group （P>0.05）. ConclusionIn patients with chronic HBV infection who do not receive antiviral therapy， there is a difference in serum HBV RNA level in different stages of natural history， and serum HBV RNA level has the strongest correlation with HBV DNA and a relatively weak correlation with HBsAg. In patients with HBeAg-positive chronic HBV infection， serum HBV RNA level in male patients with ALT>30 U/L and female patients with ALT>19 U/L are in the transition stage between HBeAg-positive chronic HBV infection and HBeAg-positive CHB.

Objective To improve the efficiency and accuracy of videonystagmography calibration test results while enabling effective recognition of saccadic undershoot waveform by developing a dual-stream architecture-based deep learning model. Methods A vestibular function calibration test recognition model with cross-modal feature fusion was constructed by integrating vision transformer (ViT) and a modified ConvNeXt convolutional network. The model utilized trajectory pictures and spatial distribution maps as inputs, employed a multi-task learning framework to classify calibration data, and to directly evaluate undershoot waveform. Results The model showed outstanding performance in assessing calibration compliance. The accuracy, sensitivity, specificity of the model in left side, middle, and right side were all greater than 90%, and AUC values were all greater than 0.99, with 97.66% of optimal accuracy (middle), 98.98% of optimal sensitivity (middle), 96.87% of optimal specificity (right side), and 0.997 of AUC (right side). The model also showed promising performance in undershoot waveform recognition with 87.50% of accuracy, 89.66% of sensitivity, 85.71% of specificity, 86.67% of F1 score, and 0.931 of AUC. Conclusions The proposed method not only significantly enhances the efficiency and accuracy of calibration test results, but also provides a novel solution for undershoot waveform recognition.