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.

Objective To investigate the drug resistance phenotype/genotype and virulence gene carrying status, serotype, and sequence typing (ST) of Campylobacter jejuni in Fengtai District in 2024. Methods The drug resistance phenotype was detected by agar dilution method, and serotyping was performed by real-time fluorescence quantitative polymerase chain reaction (PCR). The whole genome sequencing was used to analyze drug resistance genes, virulence genes, and other characteristics. Results Among the 35 Campylobacter jejuni strains analyzed, the antibiotics with the highest resistance rates were nalidixic acid and ciprofloxacin (both 82.86%), followed by tetracycline (74.89%). All strains exhibited a 100.00% sensitivity rate to erythromycin. The multiple drug resistance rate was 77.14%. The predominant drug resistance genes were gyrA-T86I (94.29%) and tetO (74.29%). The most common serotype was HS:2, accounting for 25 strains. A total of 132 virulence genes were detected. There were 12 clonal complexes (CC) and 24 ST types. Conclusion The drug resistance situation of Campylobacter jejuni in Fengtai District is relatively severe, and the drug resistance phenotype is related to specific genetic mutations to some extent. The serotypes are diverse, the virulence genes are complex, and the pathogenic potential is significant, posing a notable risk for inducing Guillain-Barré syndrome (GBS). The high number of ST types indicates substantial genetic diversity.

ObjectiveTo construct B cell conditional knockout Spi1 gene mice and analysis their genotypes to provide an animal model basis for disease pathogenesis and drug target research. MethodsMb1-cre transgenic mice were crossed with Spi1^flox/flox mice. PCR combined with agarose gel electrophoresis was performed for genotyping， and Mb1-cre×Spi1^flox/flox homozygous knockout mice were screened out. Flow cytometry was used to detect the expression of PU.1 protein in various immune cell subsets to verify the specificity of gene knockout. ResultsPCR identification results showed that when using flox primers for detection， only a 220 bp band was amplified， indicating a genotype of Spi1^flox/flox in the mice. When using Mb1-Cre primers for detection， a 383 bp band was amplified， indicating a genotype of Mb1-cre×Spi1^flox/flox in the mice； Flow cytometry results showed that compared to Spi1^flox/flox mice， Mb1-cre×Spi1^flox/flox mice exhibited significantly reduced PU.1 expression levels in bone marrow-derived mononuclear cells， peripheral blood mononuclear cells， and spleen-derived B cells. Moreover， PU.1 exhibited normal expression in other immune cells， such as CD4⁺ T cells and macrophages. ConclusionBy utilizing the Cre/LoxP system and CRISPR/Cas9 technology， this study successfully generates mice with B cell-specific conditional knockout of the Spi1 gene， providing a reliable animal model for in-depth exploration of the specific role of PU.1 in B cell-related diseases.

Li-Fraumeni syndrome(LFS)is a rare autosomal dominant genetic disorder.Patients with LFS tend to develop tumors at a young age and are at risk of multiple types of cancer.The core pathogenic mechanism of LFS is germline mutation of the TP53 gene,which leads to loss of function of the p53 protein and an increase in the risk of tumor development.There are many tumor types closely related to LFS,including soft tissue sarcoma,osteosarcoma,brain tumors,breast cancer,and adrenocortical carcinoma.Although some common mutation sites of the TP53 gene in LFS patients have been identified,there are still differences in mutation sites among different patients,and the type of TP53 gene mutation may affect the clinical manifestations and prognosis of patients.Therefore,genetic testing for LFS patients to determine the specific mutation form of TP53 is of great significance.This article reviews the clinical patho-logical characteristics,treatment methods,and prognosis of LFS-related tumor patients,aiming to provide useful refer-ences for clinical practice.

Atypical teratoid/rhabdoid tumor(AT/RT)is a rare,highly malignant central nervous system tumor with a poor prognosis and often occurs in children under 3 years of age.In this article,the clinicopathological features,molecular subtypes,pathogenesis,treatment,prognosis,relevant clinical trials,and existing problems of AT/RT will be reviewed to deepen the understanding of AT/RT.

Objective:To analyze the clinicopathological factors affecting the prognosis of patients after curative resection of lung metastases (LMs) from colorectal cancer (CRC) and to construct a clinical risk scoring (CRS) model.Methods:This study retrospectively collected clinicopathological data and follow-up information on 132 patients who underwent radical resection of LMs from CRC at the Cancer Hospital of the Chinese Academy of Medical Sciences between January 2010 and December 2020. We analyzed the clinicopathological factors influencing patient prognosis using univariate and multivariate Cox proportional hazards regression models, and we developed a risk stratification model for prognostic prediction.Results:The median follow-up duration for the cohort of 132 patients was 54.2 months. During this period, 61 patients (46.2%) experienced recurrence or distant metastasis, resulting in a 5-year DFS rate of 54.1%. Additionally, 33 patients (25.0%) died, corresponding to a 5-year overall survival (OS) rate of 76.7%. Univariate Cox proportional hazards regression model analysis indicated that ten clinicopathological factors were significantly associated with OS (all P＜0.05). These factors include the total number of lymph nodes (LNs) dissected from the primary tumor (PT) <16, the number of negative LNs from the PT <13, pN(+) of the PT, logarithmic odds of positive lymph nodes (LODDS) of the PT ≥-1.1, lymph nodes ratio (LNR) of the PT ≥0.02, preoperative carcinoembryonic antigen (CEA) level before LMs resection ≥10 ng/ml, the presence of hilar/mediastinal LN metastasis, the number of LMs ≥2, the maximum diameter of LMs ≥2.5 cm, and the necessity for hilar/mediastinal lymphadenectomy. Multivariate Cox proportional hazards regression analysis identified the number of negative LNs <13 ( HR=3.01, 95% CI: 1.28-7.03, P=0.011), pN(+) of the PT ( HR=5.04, 95% CI: 1.51-16.84, P=0.009), preoperative CEA level before LMs resection ≥10 ng/ml ( HR=5.39, 95% CI: 1.80-16.19, P=0.003), the number of LMs ≥2 ( HR=2.47, 95% CI: 1.09-5.60, P=0.030), and the necessity for hilar/mediastinal lymphadenectomy ( HR=2.74, 95% CI: 1.15-6.52, P=0.023) as independent prognostic risk factors. Patients were categorized based on independent risk factors, revealing statistically significant differences in OS across the groups with CRS scores of ≤2, 3～4, and ≥5 ( P＜0.001). Conclusions:Independent risk factors associated with LMs from CRC patients include the number of negative LNs <13, pN(+) of the PT, preoperative CEA level before LMs resection ≥10 ng/ml, the number of LMs ≥2, and the necessity for hilar/mediastinal lymphadenectomy. Patients scoring 3 or higher on the CRS model may warrant cautious assessment for the appropriateness of direct surgical treatment.

Exosomes, serving as nanoscale molecular signals for intercellular communication, possess the ability to induce odontogenic differentiation and to promote pulp regeneration. Research on their roles in cell differentiation induction has gradually emerged as a research hotspot in the field of regenerative oral medicine. The diversity in cellular sources and pretreatment methods can influence the biological properties and functions of exosomes, while the selection of combined factors and optimization of scaffold delivery strategies further regulate the efficacy of exosomes during differentiation induction. In light of this, the present article aims to comprehensively review and delve into the multi-factorial regulatory mechanisms of exosomes in facilitating odontoblastic differentiation, aiming to provide valuable insights for the development of strategies in oral tissue engineering and regenerative therapies.

In recent years, peptide hydrogels have garnered significant attention in the biomedical field due to their excellent biocompatibility, biodegradability, and low cytotoxicity. They can mimic the microenvironment of the extracellular matrix, effectively encapsulate drugs, and achieve sustained or controlled drug release, thereby facilitate tissue regeneration. This article introduces the research progress in the application of peptide hydrogels for drug sustained/controlled release. It summarizes the research hotspots in the selection and design principles of peptides, optimization strategies for peptide concentrations, and hydrogel construction techniques. Furthermore, the principles and mechanisms underlying the sustained/controlled drug release from polypeptide hydrogels are elucidated, and the detection methods applicable to different sustained/controlled release systems are summarized. The article also prospects the application potential of peptide hydrogels in the field of drug delivery and offers insights for subsequent in-depth research.

Objective To explore the role of the brief visuospatial memory test-revised(BVMT-R)in predicting the clinical conversion to psychosis in subjects with clinical high-risk for psychosis(CHR-P).Methods A total of 217 CHR subjects were recruited and assessed using BVMT-R at baseline.Participants were followed up for three years to determine whether they converted to psychosis.The relationship between BVMT-R total score and CHR-P conversion probability was analyzed using generalized additive model,and the cutoff values of BVMT-R total score for predicting CHR-P conversion were calculated by maximally selected rank statistics.Then,the total BVMT-R score was stratified into different intervals based on the cutoff values obtained as previously described.Finally,the positive likelihood ratios and the conversion rates at different time points were calculated for each interval.Results A total of 168 subjects with CHR-P completed the 3-year follow-up assessment.According to the results of the generalized additive model,the relationship between the BVMT-R total score and CHR-P conversion probability exhibited the characteristics of a piecewise function model.The cutoff values identified using the maximally selected rank statistics method were 18 and 29,which divided the BVMT-R total scores into three intervals:0-18,19-29,and 30-36.The positive likelihood ratios of the three intervals for predicting CHR-P psychosis conversion were significantly different(all P<0.01).All three intervals had significantly different rates of psychosis conversion at different follow-up time points(all P<0.01).Conclusion The total BVMT-R score can be divided into three intervals,each associated with a distinct positive likelihood ratio for predicting psychosis conversion in CHR-P individuals.Accordingly,the BVMT-R total score may serve as a preliminary indicator for estimating the probability of psychosis conversion in the CHR-P population.

Objective·To investigate the longitudinal changes in amygdala and hippocampal subfield volumes before and after transcranial magnetic stimulation(TMS)treatment in patients with major depressive disorder(MDD)and explore their correlation with the antidepressant and anxiolytic efficacy of TMS.Methods·A total of 58 patients diagnosed with MDD at Shanghai Mental Health Center,Shanghai Jiao Tong University School of Medicine,were included in this study between January 2018 and August 2023.Clinical depressive and anxiety symptoms were assessed by using the Hamilton Depression Scale(HAMD),Montgomery-Asberg Depression Rating Scale(MADRS),and Hamilton Anxiety Scale(HAMA)at baseline and post-TMS treatment.Patients underwent a baseline magnetic resonance imaging(MRI)scan followed by TMS treatment targeting the left dorsolateral prefrontal cortex(DLPFC)at a frequency of 10 Hz,totaling 20 sessions.A follow-up MRI scan was conducted on the same day the TMS treatment concluded.Amygdala and hippocampal subfield volumes were segmented and calculated by using FreeSurfer v6.0.0 software.Longitudinal changes in the subfield volumes were analyzed with two-way analysis of variance.Controlling for age,sex,and intracranial volume,partial correlation analysis was conducted between subfield volumes and baseline clinical scores.The association between the rate of volume change in brain regions with significant volume changes and symptom improvement(reduction in HAMD,MADRS,and HAMA scores)was evaluated.Results·Following TMS treatment,a significant increase in the volume of the right amygdala central nucleus was observed(t=-2.441,P=0.018).While the volumes of bilateral hippocampal fimbria decreased,the volumes of most hippocampal subfield and the total hippocampus increased(P<0.05).No significant correlations were found between baseline amygdala or hippocampal subfield volumes and clinical depressive and anxiety symptoms.However,only in patients who responded effectively to TMS treatment,a positive correlation was found between the volume change rate of the left hippocampal tail and reductions in anxiety symptoms(HAMA:r=0.334,P=0.044).Conclusion·High-frequency TMS targeting the left DLPFC may induce volume increases in the right amygdala central nucleus and specific hippocampal subfields.Additionally,the volume change rate of the left hippocampal tail is associated with anti-anxiety effects in TMS responders,suggesting that high-frequency TMS targeting the left DLPFC may induce neuroplastic changes in the central nucleus of the right amygdala and key subfields of the hippocampus.