Objective: To develop a dual-branch deep learning framework for accurate multi-label classification of fundus diseases, addressing the key limitations of insufficient complementary feature extraction and inadequate cross-modal feature fusion in existing automated diagnostic methods. Methods: The fundus multi-label classification dataset with 12 disease categories (FMLC-12) dataset was constructed by integrating complementary samples from Ocular Disease Intelligent Recognition (ODIR) and Retinal Fundus Multi-Disease Image Dataset (RFMiD), yielding 6 936 fundus images across 12 retinal pathology categories, and the framework was validated on both FMLC-12 and ODIR. Inspired by the holistic multi-regional assessment principle of the Five Wheels theory in traditional Chinese medicine (TCM) ophthalmology, the dual-branch multi-label network (DBMNet) was developed as a novel framework integrating complementary visual feature extraction with pathological correlation modeling. The architecture employed a TransNeXt backbone within a dual-branch design: one branch processed red-green-blue (RGB) images to capture color-dependent features, such as vascular patterns and lesion morphology, while the other processed grayscale-converted images to enhance subtle textural details and contrast variations. A feature interaction module (FIM) effectively integrated the multi-scale features from both branches. Comprehensive ablation studies were conducted to evaluate the contributions of the dual-branch architecture and the FIM. The performance of DBMNet was compared against four state-of-the-art methods, including EfficientNet Ensemble, transfer learning-based convolutional neural network (CNN), BFENet, and EyeDeep-Net, using mean average precision (mAP), F1-score, and Cohen's kappa coefficient. Results: The dual-branch architecture improved mAP by 15.44 percentage points over the single-branch TransNeXt baseline, increasing from 34.41% to 44.24%, and the addition of FIM further boosted mAP to 49.85%. On FMLC-12, DBMNet achieved an mAP of 49.85%, a Cohen’s kappa coefficient of 62.14%, and an F1-score of 70.21%. Compared with BFENet (mAP: 45.42%, kappa: 46.64%, F1-score: 71.34%), DBMNet outperformed it by 4.43 percentage points in mAP and 15.50 percentage points in kappa, while BFENet achieved a marginally higher F1-score. On ODIR, DBMNet achieved an F1-score of 85.50%, comparable to state-of-the-art methods. Conclusion DBMNet effectively integrates RGB and grayscale visual modalities through a dual-branch architecture, significantly improving multi-label fundus disease classification. The framework not only addresses the issue of insufficient feature fusion in existing methods but also demonstrates outstanding performance in balancing detection across both common and rare diseases, providing a promising and clinically applicable pathway for standardized, intelligent fundus disease classification.

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.

The scientific research and innovation capabilities of medical students are intrinsically linked to the sustained and high-quality development of national healthcare initiatives.Cultivating outstanding medi-cal students with independent scientific capabilities and innovative consciousness is a critical component in the education and training of high-level medical professionals.Our investigation revealed that within the imperfections of the cultivating model,some faculty and students at medical schools have an insufficient understanding of scientific research and innovation and lack motivation for engaging in such activities,which hinder the progression of scientific research activities.Consequently,we initiated a teaching practice and exploratory study on the"tutorial system"aimed at fostering medical students'scientific research and innovation abilities.Based on the principle of"research informing teaching,teaching and research advan-cing together,"this study implements a"tutorial system"coordinated by tutors,supplemented by graduate and undergraduate student mentors,to cultivate innovative thinking,stimulate interest in scientific re-search,and enhance practical and research skills among medical students.Through collaborative efforts within"scientific research innovation teams,"various educational methods—including preliminary re-search,in-class and extracurricular activities,intra-group and inter-group interactions,and theoretical and practical applications—are employed to improve and strengthen the cultivation of medical students'scientif-ic research and innovation abilities.This study aims to provide valuable references for optimizing medical education management systems and enhancing the quality of medical student training.

The scientific research and innovation capabilities of medical students are intrinsically linked to the sustained and high-quality development of national healthcare initiatives.Cultivating outstanding medi-cal students with independent scientific capabilities and innovative consciousness is a critical component in the education and training of high-level medical professionals.Our investigation revealed that within the imperfections of the cultivating model,some faculty and students at medical schools have an insufficient understanding of scientific research and innovation and lack motivation for engaging in such activities,which hinder the progression of scientific research activities.Consequently,we initiated a teaching practice and exploratory study on the"tutorial system"aimed at fostering medical students'scientific research and innovation abilities.Based on the principle of"research informing teaching,teaching and research advan-cing together,"this study implements a"tutorial system"coordinated by tutors,supplemented by graduate and undergraduate student mentors,to cultivate innovative thinking,stimulate interest in scientific re-search,and enhance practical and research skills among medical students.Through collaborative efforts within"scientific research innovation teams,"various educational methods—including preliminary re-search,in-class and extracurricular activities,intra-group and inter-group interactions,and theoretical and practical applications—are employed to improve and strengthen the cultivation of medical students'scientif-ic research and innovation abilities.This study aims to provide valuable references for optimizing medical education management systems and enhancing the quality of medical student training.

Objective To explore the factors that influence major adverse cardiovascular events(MACE)in atrial fibrillation(AF)patients undergoing radiofrequency ablation(RFA),as well as to compare the prognosis of patients with advanced AF to that of the general population.Methods We prospectively recruited AF patients who underwent RFA treatment at Peking University Third Hospital between January 2021 and March 2023.General patient data were collected through the hospital's inpatient system,and MACE were tracked through outpatient visits and telephone follow-ups.Patients were categorized into three age groups:Group 1(under 65 years),Group 2(65 to 75 years),and Group 3(over 75 years).In this study,MACE was defined as include cardiovascular death,all-cause death,readmission for heart failure,acute coronary syndrome(ACS),grade 2 or higher bleeding and stroke.Results A total of 431 patients were included in this study,with an average age of(66.17±12.22)years.Among these patients,259 were male(60.09％),and the mean of CHA2DS2-VASc score was(1.79±1.30).The median follow-up period was 16.0(11.3,21.3)months,during which 28(6.50％)patients experienced MACE,with ACS and stroke being the most common events.Variables were selected using LASSO regression,and a LASSO-Cox regression model was constructed.Age(HR 1.06,95％CI 1.02-1.10,P=0.006)and hypertrophic cardiomyopathy(HR 3.70,95％CI 1.27-8.68,P=0.008)were identified as independent predictors of MACE after RFA for AF.Subgroup analysis revealed that patients under 65 had significantly better prognoses compared to older AF patients(P=0.030 compared with group 2;P=0.021 compared with group 3).Conclusions Age and hypertrophic cardiomyopathy are independent risk factors for MACE in AF patients undergoing RFA.The prognosis for younger patients is better than that for older patients,while the prognosis for advanced patients is comparable to that of patients aged 65-75 years.

Aim To observe the mechanism of salvian-olic acid B(SalB)against oxygen glucose deprivation/re-oxygenation(OGD/R)injury in H9c2 cardiomyo-cytes.Methods The protective concentration of SalB against OGD/R-injured H9c2 cardiomyocytes was screened by CCK-8 assay.The levels of lactate dehy-drogenase(LDH),aspartate transaminase(AST)and creatine kinase(CK)were detected by ELISA kit.The mechanism of action of SalB on OGD/R-injured H9c2 cardiomyocytes was explored using high-through-put sequencing of the transcriptome.The binding of SalB to differential proteins was assessed using molecu-lar docking assays.Fatty acid content was determined using free fatty acid kits.The relative expressions of mRNA and protein of differential genes were verified by RT-qPCR and Western blot.The causal relationship between the target of action of SalB and heart failure was examined by Mendelian randomization experiment.Results SalB protected OGD/R-injured H9c2 cardio-myocytes and significantly reduced the levels of CK,LDH and AST compared with the blank control group.One hundred differential genes were screened by tran-scriptome sequencing,which were mainly involved in fatty acid elongation,central carbon metabolism of cancer,tryptophan metabolism pathways.Molecular docking showed that SalB had good binding energy to differential proteins.The mRNA and protein expression of core differential genes Elovl6,Echs1 and Acot1 were consistent with the transcriptome sequencing results.SalB reduced fatty acidsafter OGD/R injury.Mende-lian randomization experiments suggested that SalB might reduce the risk of heart failure through fatty acid metabolism,thereby reducing the risk of heart failure.Conclusion SalB can protect H9c2 cardiomyocytes after OGD/R injury by down-regulating Elovl6,Echs1 and Acot1 expression through the fatty acid metabolism pathway.

Depression is a prevalent mental and emotional disor-der that often results in significant emotional disturbances,cog-nitive dysfunction,and memory impairments.It is characterized by a high incidence rate,a substantial disability burden,and limited therapeutic efficacy.Currently,the long-term use of medications for the treatment of depression can result in a range of adverse reactions,highlighting the urgent need to explore no-vel approaches that can effectively alleviate depressive symptoms while minimizing side effects.Curcumin,a natural polyphenolic compound derived from the rhizome of turmeric,demonstrates considerable potential in the prevention and treatment of depres-sion,owing to its diverse array of biological activities.In recent years,numerous studies have investigated the use of curcumin for the treatment of depression.This article aims to provide a comprehensive review of the mechanisms of action underlying curcumin's efficacy in treating depression.Specifically,it focu-ses on its ability to improve neurotransmitter imbalances,restore neural plasticity,alleviate neural damage,mitigate dysfunction of the hypothalamic-pituitary-adrenal(HPA)axis,regulate in-flammatory factors and neuroinflammatory signaling pathways,and inhibit oxidative stress.This review is intended to offer in-sights and methodological references for basic research on curcu-min,as well as for the development of novel therapeutic agents for the treatment of depression.

Background and Aims:Endovascular repair of aortic arch diseases poses a major challenge in vascular surgery due to the need to both effectively exclude the lesion and preserve perfusion of supra-aortic branch vessels.The Castor branched aortic covered stent,with its integrated design and ability to maintain left subclavian artery(LSA)patency,offers potential advantages.When combined with the chimney technique for the left common carotid artery(LCCA),it may provide a minimally invasive and feasible solution for patients with insufficient proximal landing zones.This study aims to evaluate the preliminary feasibility and safety of this combined approach and provide clinical reference for the endovascular management of complex aortic arch pathologies.Methods:A retrospective analysis was conducted on 15 patients with aortic arch diseases who underwent treatment with the Castor branched stent-graft combined with LCCA chimney stenting at the Second Xiangya Hospital of Central South University between February 2023 and December 2024.Baseline characteristics,surgical procedures,perioperative complications,and follow-up outcomes were analyzed to assess technical success,complication rates,and branch vessel patency.Results:Among the 15 patients(11 males,average age 63.8 years),primary diagnoses included aortic dissection(33.4%),aortic arch aneurysm(53.3%),and penetrating aortic ulcer(13.3%).The technical success rate was 100%,with no perioperative deaths or major complications.During the follow-up period(4-26 months,mean 12.9 months),no adverse events such as stroke,paralysis,endoleak,or stent migration occurred.The patency rate of both the LCCA and LSA remained 100%.Conclusion:The Castor branched aortic stent-graft combined with LCCA chimney technique appears to be a technically feasible and safe short-term option for treating aortic arch diseases with insufficient proximal landing zones.It may serve as a promising alternative for complex aortic arch repair;however,large-scale,multicenter studies with long-term follow-up are needed to further validate its efficacy and safety.

Objective:To investigate the effects of Japan tallow(JT)on liver injury and gut microbiota regulation in type 2 diabetes mellitus(T2DM)mice,thereby providing a theoretical basis for developing therapeutic edible oils for diabetes treatment.Methods:T2DM animal model was established through a combined approach of nutritional intervention and chemical induction.Experimental animals were first fed a high-fat diet(HFD)for 6 weeks,followed by intraperitoneal injection of freshly prepared streptozotocin(STZ).After successful model establishment,the mice were divided into five groups(n=6 per group):control group without any intervention;T2DM group;HFD reversion to standard chow group;metformin group;and JT intervention group,which received respective treatments for 4 weeks.At the endpoint,fresh fecal samples were collected from all groups,and the gut microbiota composition was analyzed using 16s rDNA high-throughput sequencing.Liver histopathological changes were examined using histological methods.Results:Compared with the normal control(ND),T2DM mice showed significantly increased fasting blood glucose(FBG)levels,with evident hepatocyte lipid accumulation,steatosis,inflammatory cell infiltration,and widespread vacuolar and fatty degeneration.After Japan tallow(JT)intervention,FBG levels decreased significantly,liver color approximated normal appearance,and pathological morphology improved noticeably.16s rDNA sequencing analysis demonstrated that JT treatment could effectively regulate the intestinal microbiota structure in T2DM mice.Increased microbial α-diversity indices(Chao1,observed_species,Faith_pd,Simpson);At the phylum level,increased Verrucomicrobia abundance while decreased Proteobacteria were detected;At the family level,elevated Bifidobacteriaceae and reduced Porphyromonadaceae were seen;At the genus level,Bifidobacterium was promoted and Akkermansia proliferat while Escherichia and Klebsiella were downregulated.Conclusions:Japan tallow exhibits significant effects in alleviating liver tissue damage and regulating intestinal microbiota disorders in T2DM mice,our study indicated new theoretical basis for the research and development of potential strategy for diabetes.