Objective: To investigate methods for constructing a high-quality instructional dataset for traditional Chinese medicine (TCM) mental disorders and to validate its efficacy. Methods: We proposed the Fine-Med-Mental-T&P methodology for constructing high-quality instruction datasets in TCM mental disorders. This approach integrates theoretical knowledge and practical case studies through a dual-track strategy. (i) Theoretical track: textbooks and guidelines on TCM mental disorders were manually segmented. Initial responses were generated using DeepSeek-V3, followed by refinement by the Qwen3-32B model to align the expression with human preferences. A screening algorithm was then applied to select 16 000 high-quality instruction pairs. (ii) Practical track: starting from over 600 real clinical case seeds, diagnostic and therapeutic instruction pairs were generated using DeepSeek-V3 and subsequently screened through manual evaluation, resulting in 4 000 high-quality practice-oriented instruction pairs. The integration of both tracks yielded the Med-Mental-Instruct-T&P dataset, comprising a total of 20 000 instruction pairs. To validate the dataset’s effectiveness, three experimental evaluations (both manual and automated) were conducted: (i) comparative studies to compare the performance of models fine-tuned on different datasets; (ii) benchmarking to compare against mainstream TCM-specific large language models (LLMs); (iii) data ablation study to investigate the relationship between data volume and model performance. Results: Experimental results demonstrate the superior performance of T&P-model fine-tuned on the Med-Mental-Instruct-T&P dataset. In the comparative study, the T&P-model significantly outperformed the baseline models trained solely on self-generated or purely human-curated baseline data. This superiority was evident in both automated metrics (ROUGE-L > 0.55) and expert manual evaluations (scoring above 7/10 across accuracy). In benchmark comparisons, the T&P-model also excelled against existing mainstream TCM LLMs (e.g., HuatuoGPT and ZuoyiGPT). It showed particularly strong capabilities in handling diverse clinical presentations, including challenging disorders such as insomnia and coma, showcasing its robustness and versatility. Data ablation studies showed that T&P-model performance had an overall upward trend with minor fluctuations when training data increased from 10% to 50%; beyond 50%, performance improvement slowed significantly, with metrics plateauing and approaching a saturation point. Conclusion This study has successfully constructed the specialized Med-Mental-Instruct-T&P instruction dataset for TCM mental disorders proposed the systematic Fine-Med-Mental-T&P methodology for its development, effectively addressing the critical challenge of high-quality, domain-specific data scarcity in TCM, and providing essential data support for developing intelligent TCM diagnostic and therapeutic systems.

Lung cancer constitutes the most prevalent and lethal malignant tumor in China. Approximately 85% of lung cancer diagnoses correspond to the non-small cell histological subtype [non-small cell lung cancer (NSCLC)]. Despite surgery being the mainstay for early-stage disease, postoperative recurrence remains high and adjuvant chemotherapy offers limited benefit. In recent years, targeted therapy has demonstrated substantial advantages in driver mutation-positive NSCLC. To this end, the Lung Cancer Medical Education Committee of the Chinese Medical Education Association developed guidelines based on a systematic review of evidence through November 2025, using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach and a modified Delphi method. Focusing on epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK), and addressing ROS proto-oncogene 1 (ROS1), B-Raf proto-oncogene serine/threonine kinase (BRAF) V600E mutation, and mesenchymal-epithelial transition factor (MET) exon 14 (METex14) skipping, the guideline covers molecular testing, neoadjuvant/adjuvant therapy, perioperative strategies, minimal residual disease monitoring, and postoperative surveillance. It defines testing requirements, specifies stage-directed and subtype-specific treatments, and standardizes minimal residual disease monitoring. These recommendations emphasize precision and feasibility to improve survival and quality of life.

Abstract In the wave of digital and intelligent applications, artificial intelligence (AI) is transforming the development trajectories of industries across the globe. Traditional Chinese medicine (TCM), as a cultural treasure of the Chinese nation, carries thousands of years of wisdom and practical experience. However, in the context of the rapid advancements in modern medicine and technology, TCM faces dual challenges: preserving its heritage while innovating. DeepSeek, a major achievement in the field of AI, offers a new opportunity for the development of TCM with its powerful technological capabilities. Exploring the integration of DeepSeek with TCM not only helps modernize the practice but also promises unique contributions to global health.

Objective: To investigate the trends in incidence and mortality of lung cancer in Huangpu District, Shanghai Municipality from 2002 to 2019, so as to provide the evidence for formulating lung cancer prevention and control measures. Methods: Data of lung cancer incidence and mortality among residents in Huangpu District from 2002 to 2019 were collected through the Shanghai Cancer Registration and Reporting Management System. The crude incidence and mortality of lung cancer was calculated, and standardized by the data from the Chinese Fifth National Population Census in 2000 (Chinese-standardized rate) and the Segi's world standard population in 1960 (world-standardized rate). The trends in incidence and mortality of lung cancer among residents by age and gender were evaluated using annual percent change (APC). Results: A total of 12 965 cases of lung cancer were reported in Huangpu District from 2002 to 2019, and the crude incidence rate was 80.66/105, the Chinese-standardized incidence rate was 34.54/105, and the world-standardized incidence rate was 31.30/105, all showing upward trends (APC=4.588%, 2.933% and 3.247%, all P<0.05). A total of 10 102 deaths of lung cancer were reported, and the crude mortality rate was 62.30/105, showing an upward trend (APC=0.959%, P<0.05); the Chinese-standardized mortality was 25.93/105, and the world-standardized mortality was 22.05/105, both showing downward trends (APC=-1.282% and -1.263%, both P<0.05). The crude incidence and mortality rates of lung cancer in males were higher than those in females (101.39/105 vs. 60.52/105, 85.45/105 vs. 39.87/105, both P<0.05). The crude incidence and mortality rates of lung cancer showed upward trends with age (both P<0.05), reaching their peaks in the age groups of 80-<85 years (341.37/105) and 85 years or above (355.97/105), respectively. Conclusions The incidence of lung cancer showed an upward trend, while the mortality showed a downward trend in Huangpu District from 2002 to 2019. Elderly men were the high-risk group for lung cancer incidence and mortality.

[Objective] To determine the correlation between traditional Chinese medicine (TCM) inspection of spirit classification and the severity grade of depression based on facial features, offering insights for intelligent intergrated TCM and western medicine diagnosis of depression. [Methods] Using the Audio-Visual Emotion Challenge and Workshop (AVEC 2014) public dataset on depression, which conclude 150 interview videos, the samples were classified according to the TCM inspection of spirit classification: Deshen (得神, presence of spirit), Shaoshen (少神, insufficiency of spirit), and Shenluan (神乱, confusion of spirit). Meanwhile, based on Beck Depression Inventory-II (BDI-II) score for the severity grade of depression, the samples were divided into minimal (0 – 13, Q1), mild (14 – 19, Q2), moderate (20 – 28, Q3), and severe (29 – 63, Q4). Sixty-eight landmarks were extracted with a ResNet-50 network, and the feature extracion mode was stadardized. Random forest and support vectior machine (SVM) classifiers were used to predict TCM inspection of spirit classification and the severity grade of depression, respectively. A Chi-square test and Apriori association rule mining were then applied to quantify and explore the relationships. [Results] The analysis revealed a statistically significant and moderately strong association between TCM spirit classification and the severity grade of depression, as confirmed by a Chi-square test (χ2 = 14.04, P = 0.029) with a Cramer’s V effect size of 0.243. Further exploration using association rule mining identified the most compelling rule: “moderate depression (Q3) → Shenluan”. This rule demonstrated a support level of 5%, indicating this specific co-occurrence was present in 5% of the cohort. Crucially, it achieved a high Confidence of 86%, meaning that among patients diagnosed with Q3, 86% exhibited the Shenluan pattern according to TCM assessment. The substantial Lift of 2.37 signifies that the observed likelihood of Shenluan manifesting in Q3 patients is 2.37 times higher than would be expected by chance if these states were independent—compelling evidence of a highly non-random association. Consequently, Shenluan emerges as a distinct and core TCM diagnostic manifestation strongly linked to Q3, forming a clinically significant phenotype within this patient subgroup. [Conclusion] Automated facial analysis can serve as a common lens for TCM and western psychological assessments align in the diagnosis of depression. The inspection of spirit decline trajectory parallels worsening depression, supporting early screening and stratified intervention, and providing a reference for the intelligent assistance of integrated TCM and western medicine in the diagnosis of depression.

Objective:To analyze the clinical features of patients with traumatic optic neuropathy (TON) and to explore its clinical patterns and treatment outcomes.Methods:A retrospective analysis was conducted on data from 323 patients (334 eyes) with TON, who were treated in the Department of Otorhinolaryngology Head and Neck Surgery, the Affiliated Hospital of Qingdao University from April 1999 to October 2024. Among these patients, 288 were male and 35 were female, with ages ranging from 4 to 70 years. All patients were followed up for a period of 6 to 24 months, with the final follow-up visual acuity recorded as the ultimate visual outcome. The visual acuity evaluation criteria were classified into five levels: no light perception, light perception, hand movement in front of the eye, counting fingers at 1 meter, and "chart-visible acuity". A treatment outcome was deemed effective if the post-treatment visual acuity improved by one level or more compared to pre-treatment, or if the chart-visible acuity improved by two lines or more on the logMAR chart. The clinical characteristics of patients, causes of injury, complications, treatment methods, and changes in visual acuity before and after treatment were summarized. Logistic regression analyses were performed to identify the influencing factors affecting treatment efficacy.Results:TON occurred mostly in young (215/323, 66.56%) males (288/323, 89.16%), the majority of patients came from villages and towns (236/323, 73.07%). Traffic accidents (232/323, 71.83%) remained the main etiology. Most patients had craniofacial injuries and other bodily injuries. The effective rate of vision improvement was 50.30% (168/334). Multiple logistic regression analyses identified that residual vision (light perception or better) at presentation ( OR=3.26, P<0.001) and receiving treatment within 7 days after injury ( OR=2.04, P=0.008) were protective factors on visual acuity recovery, while the presence of orbital wall fracture was a risk factor for visual acuity recovery ( OR=0.26, P<0.001). Additionally, undergoing surgical treatment was a protective factor for visual improvement in patients with no light perception ( OR=2.94, P=0.007). For patients with residual vision at presentation, orbital wall fracture was a significant risk factor ( OR=0.28, P=0.009). Conclusions:TON is more prevalent in young males and is primarily caused by traffic accidents, leading to a poor prognosis. Timely medical intervention following injury significantly influences prognostic outcomes. Early surgical intervention (within 7 days) is recommended, particularly for patients with no light perception at presentation.

This article is based on modern psychology and TCM emotional theory,combined with facial expression recognition technology,to apply deep learning methods to the digital research of TCM emotions to more accurately capture and analyze patients'emotional states.A cross-disciplinary framework was established by synthesizing facial expression-emotion correlations from psychological and TCM perspectives.The methodology included:Data annotation of TCM-defined emotional expressions using standardized coding systems;Facial expression acquisition,spatiotemporal feature extraction and emotion classification through a 3D convolutional neural network(3D CNN).The framework achieved 91.43%accuracy in video-based emotion classification.High-arousal emotional states demonstrated superior recognition performance,with anger showing optimal recall(1.000 0)and F1-score(0.946 3),while surprise attained the highest precision(0.976 0).These findings aligned with TCM pathological characteristic of"anger induces qi ascending,surprise disrupts qi flow".The digital recognition framework for TCM emotional quantification based on facial expression recognition technology exhibits strong alignment with TCM observation diagnosis,providing clinicians with an objective tool to assess the"seven emotions"and elucidating emotion-facial correlations in classical TCM theory.Future research should focus on longitudinal validation across diverse populations and establish development pathways for AI-assisted TCM diagnostic systems.

Objective:To investigate the clinical features and pathogenesis of a child with Stargardt disease caused by variants of ABCA4 gene. Methods:A child presented at Shenzhen Eye Hospital between September 5, 2020, and April 3, 2023 was selected as the study subject. Clinical data of the child were collected. Whole exome sequencing was performed on peripheral blood samples from the child and his parents. Candidate variants were validated by Sanger sequencing and bioinformatic analysis. This study was approved by the Medical Ethics Committee of Shenzhen Eye Hospital (Ethics No.: 2022KYPJ072).Results:The child was a 10-year-old male presenting with uncorrected visual acuity of 0.1 in both eyes without improvement with refractive correction. Fundus photography showed diffusely distributed yellow-white flecks in the macular region. FAF revealing central hypofluorescence surrounded by a hyperfluorescent ring, and OCT demonstrating significant foveal thinning (right eye: 45 μm; left eye: 50 μm) with ellipsoid zone disruption. Whole exome sequencing and Sanger sequencing revealed that the child has harbored compound heterozygous variants of the ABCA4 gene, namely c. 2384G>T (p.Gly795Val) and c. 2903G>A (p.Arg968Glu), which were inherited from his phenotypically normal parents and consistent with an autosomal recessive inheritance. This specific combination of the variants was previously unreported. According to the guidelines from the American College of Medical Genetics and Genomics (ACMG) guidelines, both variants were classified as likely pathogenic (PM2_Supporting+ PM3+ PP3+ PP4; PM1+ PM2_Supporting+ PP3+ PP4). Conclusion:The novel compound heterozygous variants of the ABCA4 gene probably underlay the genetic etiology of Stargardt disease type 1 in this child. Above finding has expanded the mutational spectrum of the ABCA4 gene among the Chinese population and provided further evidence for understanding the genetic heterogeneity and genotype-phenotype correlation of the Stargardt disease.

Objective:To investigate the protective effect of myrislignan（MRL）on concanavalin A（Con A）-induced autoimmune hepatitis（AIH）.Methods:C57BL/6J mice were divided into the following groups using a random number table，with five mice in each group：control group，MRL group，model group（Con A group），and MRL pretreatment group（MRL+Con A group）. MRL was injected intraperitoneally at a dose of 30 μg/g；3 h after pretreatment，Con A（18 μg/g）was administrated by intravenous injection；mouse livers and serum samples were collected 12 h after injection for measuring serum transaminase levels and liver cell apoptosis. The mRNA and protein expression levels of IL-6，IL-12，and TNF-α were measured using qRT-PCR and ELISA. Flow cytometry was performed to detect the proportion and activation status of macrophages in liver tissues. Bone marrow-derived macrophages（BMDMs）were isolated and induced in vitro to analyze the regulatory effect of MRL on macrophages. One-way analysis of variance was used to compare the differences in various indicators among groups. Results:Compared with the Con A group，MRL（30 μg/g）pretreatment significantly reduced alanine aminotransferase（ P<0.05）and aspartate transaminase（ P<0.01）levels，attenuated liver oxidative stress（increased superoxide dismutase activity，while decreased levels of malondialdehyde and myeloperoxidase；all P<0.05），and suppressed hepatocyte apoptosis（ P<0.01）. Both in vivo and in vitro experiments confirmed that MRL（30 μg/g）could reduce the proportion of M1 macrophages（ in vivo： P<0.05； in vitro：all P<0.001）and inhibit macrophage activation（ in vivo： P<0.01； in vitro：all P<0.05）. Conclusion:MRL effectively prevents Con A-induced autoimmune hepatitis by inhibiting liver cell apoptosis，attenuating liver oxidative stress，suppressing M1 macrophage polarization，and reducing inflammatory cytokine expression.

With the continuous progress of modern science and technology, the research of TCM inspection is gradually moving towards a new stage of objectification and intelligence. By systematically sorting out the key technologies involved in the objectification of inspection and diagnosis, and summarizing its clinical application progress in the three core areas of face diagnosis, tongue diagnosis and visual diagnosis, it is found that the inspection and diagnosis of TCM can not only significantly improve the efficiency and accuracy of clinical diagnosis with the help of modern technology, but also achieve more results in disease differentiation, constitution identification and efficacy evaluation. The key technologies include digital image processing, infrared thermal imaging, spectroscopy, photoelectric volume and artificial intelligence (AI) technology. In terms of clinical application, facial diagnosis focuses on image and infrared thermography feature analysis, tongue diagnosis involves image, spectral feature research and AI model construction, and eye diagnosis focuses on white eye collaterals extraction, fundus sign analysis and eye movement information research. However, there are still some problems, such as imperfect standardization system, bottleneck of multi-modal fusion and interpretation, insufficient clinical verification and practicability. In the future, the deep integration of TCM observation and modern science and technology should be promoted by means of technological integration and innovation and strengthening interdisciplinary cooperation, so as to help the modernization of TCM diagnosis.