ObjectiveTo construct a large language model （LLM）-based intelligent pre-consultation system for traditional Chinese medicine （TCM） to improve efficacy of clinical practice. MethodsA TCM large language model was fine-tuned using DeepSpeed ZeRO-3 distributed training strategy based on YAYI 2-30B. A weighted undirected graph network was designed and an agent-based syndrome differentiation model was established based on relationship data extracted from TCM literature and clinical records. An agent collaboration framework was developed to integrate the TCM LLM with the syndrome differentiation model. Model performance was comprehensively evaluated by Loss function， BLEU-4， and ROUGE-L metrics， through which training convergence， text generation quality， and language understanding capability were assessed. Professional knowledge test sets were developed to evaluate system proficiency in TCM physician licensure content， TCM pharmacist licensure content， TCM symptom terminology recognition， and meridian identification. Clinical tests were conducted to compare the system with attending physicians in terms of diagnostic accuracy， consultation rounds， and consultation duration. ResultsAfter 100 000 iterations， the training loss value was gradually stabilized at about 0.7±0.08， indicating that the TCM-LLM has been trained and has good generalization ability. The TCM-LLM scored 0.38 in BLEU-4 and 0.62 in ROUGE-L， suggesting that its natural language processing ability meets the standard. We obtained 2715 symptom terms， 505 relationships between diseases and syndromes， 1011 relationships between diseases and main symptoms， and 1 303 600 relationships among different symptoms， and constructed the Agent of syndrome differentiation model. The accuracy rates in the simulated tests for TCM practitioners， licensed pharmacists of Chinese materia medica， recognition of TCM symptom terminology， and meridian recognition were 94.09%， 78.00%， 87.50%， and 68.80%， respectively. In clinical tests， the syndrome differentiation accuracy of the system reached 88.33%， with fewer consultation rounds and shorter consultation time compared to the attending physicians （P＜0.01）， suggesting that the system has a certain pre- consultation ability. ConclusionThe LLM-based intelligent TCM pre-diagnosis system could simulate diagnostic thinking of TCM physicians to a certain extent. After understanding the patients' natural language， it collects all the patient's symptom through guided questioning， thereby enhancing the diagnostic and treatment efficiency of physicians as well as the consultation experience of the patients.

ObjectiveTo investigate the immunomodulatory effect of polysaccharides from Astragali Radix-Angelicae Sinensis Radix herb pair（Qi-gui polysaccharides） on lipopolysaccharide（LPS）-induced RAW264.7 macrophages and to characterize the structure of the active component Qi-gui homogeneous polysaccharide（AAPS-4a）， and evaluate its protective effect on ulcerative colitis（UC）. MethodsThe effects of six Qi-gui polysaccharides（0.01-100 mg·L^-1） on the proliferation of RAW264.7 cells were assessed by cell proliferation and activity assay（CCK-8）， and enzyme-linked immunosorbent assay（ELISA） was used to investigate the effects of the six polysaccharides（3， 10 mg·L^-1） on the secretion levels of tumor necrosis factor（TNF）-α， interferon（IFN）-β， and nitric oxide（NO） in LPS-induced RAW264.7 cells. After screening for active polysaccharides， high-performance size-exclusion chromatography（HPSEC） was used to determine its homogeneity and relative molecular weight， then its characteristic functional groups were identified by Fourier transform infrared spectroscopy（FT-IR）， monosaccharide composition was analyzed by high performance liquid chromatography（HPLC）， methylation analysis combined with gas chromatography-mass spectrometry（GC-MS） was performed to determine the types and linkage modes of sugar residues， and one- and two-dimensional nuclear magnetic resonance（NMR） were used to identify the sugar residue composition and configuration of the active polysaccharide. Finally， experimental animals were divided into the normal group， model group， AAPS-4a low-dose group（50 mg·kg^-1）， AAPS-4a high-dose group（100 mg·kg^-1）， and sulfasalazine（SASP） group （75 mg·kg^-1）. Except for the normal group， the acute UC mouse model was induced using 3.5% dextran sulfate sodium salt（DSS）. Each treatment group was administered the corresponding dose via oral gavage for 7 days， and changes in body weight were recorded. After treatment， the spleen index and disease activity index（DAI） score were calculated， TNF-α and interleukin-6（IL-6） levels in the serum were detected by ELISA， and histopathological changes in colon tissue were observed by hematoxylin-eosin（HE） staining. ResultsAt the cellular level， AAPS-4a exhibited a dose-dependent inhibition of LPS-induced increases in TNF-α， IFN-β， and NO levels（P<0.01）. Structural characterization of AAPS-4a revealed that it was a homogeneous polysaccharide with a relative molecular weight of 7.6×10³ Da， consisting of mannose（Man）， glucose（Glc）， and galactose（Gal） in a molar ratio of 1.3∶23.9∶1.0. It was primarily composed of five sugar residues of 1，6-α-D-Glcp， T-α-D-Glcp， 1，3-β-D-Galp， 1，4-α-D-Manp， and 1，2-α-D-Galp. In vivo experiments showed that compared with the normal group， the model group demonstrated markedly increased DAI score and spleen index， significantly reduced colon length， and significantly elevated levels of TNF-α and IL-6（P<0.01）. Compared with the model group， the AAPS-4a high-dose group significantly reduced the DAI score and spleen index， as well as TNF-α and IL-6 levels， and improved colonic atrophy（P<0.05， P<0.01）. Pathological observations showed that AAPS-4a significantly inhibited inflammatory cell infiltration in colon tissue and alleviated pathological damage. ConclusionAAPS-4a， a neutral homogeneous polysaccharide composed of 1，6-α-D-Glcp， T-α-D-Glcp， 1，3-β-D-Galp， 1，4-α-D-Manp and 1，2-α-D-Galp， is identified as a key bioactive component contributing to the anti-UC effect of the Qi-gui herb pair. Its immunoregulatory and anti-UC properties suggest its potential as a therapeutic agent for UC.

OBJECTIVE To evaluate the efficacy， safety and cost-effectiveness of toripalimab （Tor） combined with chemotherapy （CT） in the treatment of locally advanced or metastatic non-small cell lung cancer （NSCLC）. METHODS PubMed， the Cochrane Library， Embase， Web of Science， CBM， CNKI， Wanfang Data， and Health Technology Assessment （HTA） related websites were searched to collect the HTA reports， systematic reviews/meta-analyses and pharmacoeconomic studies of Tor+CT in the treatment of locally advanced or metastatic NSCLC from database/website inception to March 31， 2025. After data extraction and quality evaluation， the results of the included studies were analyzed descriptively. RESULTS A total of eleven studies were included， involving five systematic reviews/meta-analyses， and six pharmacoeconomic studies. Among the five systematic reviews/ meta-analyses， two were of high quality， while there was one each of moderate， low， and very low quality. All six pharmacoeconomic studies were of good quality. In terms of efficacy， compared with CT， Tor+CT significantly improved patients’ progression-free survival （PFS） and overall survival （P＜0.05）. In addition， compared with ipilimumab+CT， durvalumab， durvalumab+tremelimumab and sugemalimab+CT， Tor+CT could also improve the PFS （P＜0.05）. In terms of safety， there was no significant difference in the incidence of grade≥3 adverse events between patients receiving Tor+CT and CT （P＞0.05）； while Tor+CT had a lower incidence of grade≥3 adverse E-mail： events， compared with camrelizumab+CT， pembrolizumab+ 3233255290@qq.com ipilimumab， nivolumab+CT and atezolizumab+CT （P＜0.05）.In terms of cost-effectiveness， Tor+CT treatment had certain cost-effectiveness advantages， compared with CT. CONCLUSIONS Compared with CT， other programmed death-1/programmed death-ligand 1 inhibitors alone， or their combination with CT， Tor+CT for the treatment of locally advanced or metastatic NSCLC has good efficacy， safety and cost-effectiveness.

OBJECTIVE: To investigate the midterm clinical efficacy of medial patellofemoral complex (MPFC) reconstruction for recurrent patellar dislocation with high-grade trochlear dysplasia. METHODS: A retrospective analysis was carried out among adult patients who underwent arthroscopically assisted MPFC reconstruction between January 2014 and December 2020. Dejour classification was evaluated to grade trochlear dysplasia; tibial tubercle-trochlear groove (TT-TG) distance and Insall-Salvati index were measured. Preoperative and postoperative patient-reported outcome measures (PROMs) were compared, including International Knee Documentation Committee (IKDC) score, Kujala score, Lysholm score and Tegner score. Information regarding returning-to-sport rate, re-instability events and complications was collected. Patellar tilt (PT), lateral patellar displacement (LPD) and bisect offset (BSO) ratio were measured based on axial computed tomography before and after surgery to assess the patellofemoral congruence. RESULTS: A total of 46 MPFC reconstructions in 43 patients were enrolled, including 16 male and 27 female. Mean age at surgery was (22.2±7.6) years (range: 14-44 years). Mean follow-up was (49.9±22.6) months (range: 18-102 months). The percentages of Dejour B, C and D dysplasia were 37.0% (17/46), 43.5% (20/46), and 19.6% (9/46), respectively. Mean Insall-Salvati index was 1.2±0.2 (range: 0.85-1.44), and mean TT-TG distance was (19.6±3.5) mm (range: 10.6-28.7 mm). At latest follow-up, there were significant improvements in all PROMs (P < 0.001): IKDC score, from 56.3±15.1 to 86.2±8.1; Kujala score, from 58.9±15.6 to 92.6±5.4; Lysholm score, from 63.7±15.0 to 94.0±5.7; Tegner score, from 3.1±1.4 to 4.7±1.4, and there were no significant differences in the improvements of the scores between the patients with Dejour B, C and D dysplasia. Overall, ninety percent of the patients returned to their preoperative sports level. One patient reported a postoperative subluxation, while no cases of infection, limited range of motion or patella fracture were observed. PT, LPD and BSO ratio were all significant altered (P < 0.001) after MPFC reconstruction. CONCLUSION Arthroscopically assisted MPFC reconstruction yielded satisfactory midterm clinical results for recurrent patellar dislocation with high-grade trochlear dysplasia. No significant differences of improvements in knee function were observed among the three types of high-grade trochlear dysplasia.
Humans ; Patellar Dislocation/surgery* ; Male ; Female ; Adult ; Arthroscopy/methods* ; Retrospective Studies ; Adolescent ; Young Adult ; Patellofemoral Joint/surgery* ; Recurrence ; Plastic Surgery Procedures/methods* ; Patella/surgery* ; Treatment Outcome

Schizophrenia (SZ) stands as a severe psychiatric disorder. This study applied diffusion tensor imaging (DTI) data in conjunction with graph neural networks to distinguish SZ patients from normal controls (NCs) and showcases the superior performance of a graph neural network integrating combined fractional anisotropy and fiber number brain network features, achieving an accuracy of 73.79% in distinguishing SZ patients from NCs. Beyond mere discrimination, our study delved deeper into the advantages of utilizing white matter brain network features for identifying SZ patients through interpretable model analysis and gene expression analysis. These analyses uncovered intricate interrelationships between brain imaging markers and genetic biomarkers, providing novel insights into the neuropathological basis of SZ. In summary, our findings underscore the potential of graph neural networks applied to multimodal DTI data for enhancing SZ detection through an integrated analysis of neuroimaging and genetic features.
Humans ; Schizophrenia/pathology* ; Diffusion Tensor Imaging/methods* ; Male ; Female ; Adult ; Brain/metabolism* ; Young Adult ; Middle Aged ; White Matter/pathology* ; Gene Expression ; Nerve Net/diagnostic imaging* ; Graph Neural Networks

Objective:To provide references for the application of finite element model in the study of cervical vertebra by statistically analysing the frequency, numerical value, properties, and boundary setting of the finite element model and the corresponding material features as well as boundary settings in the literature.Methods:The literature on cervical vertebra-related finite element models was collected from CNKI, Wanfang, VIP, PubMed, Web of Science, and Embase databases from January 2013 to December 2023. The quality assessment was followed by manual screening. The data sources, application classification, material properties (Young’s modulus and Poisson’s ratio), and boundary conditions of cervical vertebra, cervical intervertebral, and cervical ligaments were statistically analyzed.Results:A total of 102 papers were included. The finite element models of the cervical vertebra were derived from medical image reconstruction modeling techniques, predominantly CT plain scan and magnetic resonance imaging. Among the 102 cervical vertebra models, the C3-C7 (lower cervical segment) model appeared with the highest frequency (19). The Young’s modulus of the cortical bone, cancellous bone, and posterior structure of cervical vertebrae were set at about 12 000 or 10 000, 440, and 3 600 MPa, respectively, and the Poisson’s ratios were mainly set at about 0.29 or 0.30, 0.29, and 0.29. The Young’s modulus of the cervical intervertebral disc endplate, nucleus pulposus, and annulus fibrosus were concentrated around 500 or 2 000, 1, and 100 MPa, respectively, and the Poisson’s ratios were set at about 0.40, 0.50, and 0.40, respectively. The Young’s modulus of the anterior longitudinal ligament, posterior longitudinal ligament, transverse ligament, ligamentum flavum, interspinous ligament, capsular ligament, and articular cartilage of the cervical spine were set around 30, 20, 20, 6-10, 4-8, 10 or 20, 10 MPa, and the Poisson’s ratios were set at aoubt 0.30, 0.30, 0.30, 0.30, 0.30, 0.40, and 0.30, respectively. The Young’s modulus of the upper cervical interdental ligament, lamina, cruciate ligament, nuchal ligament, and pterygoid ligament were set at about 10, 10, 10 or 20, 20, and 5 MPa, respectively, and the Poisson’s ratios were set at about 0.30. Head weight settings were more common at 50, 74, and 100 N.Conclusions:The finite element model of the cervical vertebra has great value in the study of cervical spondylosis, but further optimization is still needed in the assignment of material properties, mesh division, and model verification to improve the accuracy and clinical applicability of the model.

Objective To identify the characteristics of nutritional problems in patients with Parkinsonism who par-ticipated in multi-disciplinary team(MDT)and to investigate the correlation between nutritional status and dyspha-gia in these patients.The predictive value of video fluoroscopic swallowing study(VFSS)and EAT-10 scale for malnutrition was compared.Methods Subjects were patients with Parkinsonism participated in joint MDT consulta-tion project in Peking Union Medical College Hospital from November 2020 to January 2023.Subjective and objec-tive dysphagia were evaluated by EAT-10 scale score and VFSS dysphagia score.Nutritional status of the patients was evaluated by geriatric nutrition risk index(GNRI),albumin,prealbumin,serum folic acid,vitamin B12 and temporal muscle thickness.Results A total of 30 participants met the criteria and were included in the study.The age was 45-82(66.1±9.0)years old.Six(20%)were at risk of malnutrition shown b a comprehensive nutri-tional status assessment using GNRI.The areas under receiver operating characteristic(ROC)curves(AUC)of VFSS dysphagia score and EAT-10 score to predict malnutrition were 0.781(0.568-0.995)and 0.927(0.827-1.000),respectively.EAT-10 score was correlated with GNRI(r=-0.524,P＜0.01),BMI(r=-0.618,P＜0.001),prealbumin(r=-0.616,P＜0.001).The VFSS dysphagia score was only correlated with BMI(r=-0.446,P＜0.05)and prealbumin(r=-0.387,P＜0.05).Conclusions Patients with Parkin-sonism requiring MDT often have multiple micronutrient imbalance.Patients' subjective perception of dysphagia has a greater impact on their nutritional status than objective assessment of dysphagia.

Angiosarcoma(AS)is a rare soft tissue sarcoma originating from vascular or lymphatic endothelial cells,accounting for less than 1% of soft tissue sarcomas.It is most common in skin and surface of soft tissue.The diagno-sis of AS is based on pathology,and the expression of CD31,ERG and CD34 is commonly positive in immuno-his-tochemistry.Staging is performed according to the American Joint Committee on Cancer(AJCC)criteria for soft tis-sue sarcoma.Radical surgery is the primary therapy for locally confined disease.For un-resectable or metastatic an-giosarcoma,chemotherapy is the main treatment method.Targeted therapy and immunotherapy have developed rap-idly in recent years,and the combined therapy of different types of chemotherapy drug has been the focus of optimi-zation in the future.

Objective:To explore the impact of nutrition education on the vitamin D nutritional status of pregnant women.Methods:It was a randomized controlled trial. A total of 96 women with vitamin D deficiency at 8-13 weeks of gestation who were regularly followed-up in the pregnancy nutrition clinic of the Department of Clinical Nutrition of Peking Union Medical College Hospital from January to October 2021 were selected by convenience sampling. The women were divided into a control group and an experimental group (48 cases in each group) according to the random number table method. The control group received routine outpatient treatment and was given different doses of vitamin D supplements according to the degree of vitamin D deficiency. On this basis, the experimental group were given additional nutritional education, including the distribution of vitamin D education manuals at 8-13 weeks of pregnancy, the assessment of vitamin D intake and cognition of vitamin D, the provision of individualized vitamin D nutritional status improvement plans, and the one-time assessment and guidance of vitamin D nutritional status at 24-28 weeks and 36-38 weeks of pregnancy, to continuously promote vitamin D health. Due to factors such as miscarriage or incomplete data acquisition during the study, 44 cases were finally included in the control group and 42 cases in the experimental group. The non-parametric test Mann-Whitney U was used to compare the vitamin D nutritional status of the two groups at 24-28 weeks and 36-38 weeks of pregnancy, and to analyze the effect of nutritional education on the vitamin D nutritional status of pregnant women. Results:At 8-13 weeks of pregnancy, there was no significant difference in serum 25-hydroxyvitamin D [25(OH)D] concentration between the two groups ( P>0.05). At 24-28 weeks of pregnancy, the serum 25(OH)D concentration and the proportion of pregnant women with normal vitamin D nutritional status in the experimental group were both significantly higher than those in the control group [29.90 (24.63, 34.28) vs 25.65 (21.50, 31.28) μg/L, 50.0% vs 29.5%] (both P<0.05). At 36-38 weeks of pregnancy, there was no significant difference in serum 25(OH)D level between the two groups ( P>0.05); however, the proportion of pregnant women with normal vitamin D nutritional status in the control group and the experimental group was 80.5% and 70.7%, respectively, which were both significantly higher than those in 8-13 and 24-28 weeks of pregnancy. The main effect of time was significant in both groups ( F=201.040, P<0.001). Conclusion:Nutrition education can enhance pregnant women′s knowledge of vitamin D, promote behavioral changes, and improve vitamin D nutritional status.