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.

Objective To analyze the differences in bone serum protein between patients with osteoporosis accompanied by obstructive sleep apnea syndrome(OSAS)and those with osteoporosis only using proteomics.Methods A total of 80 osteoporosis patients who attended our hospital between June 2022 and June 2024 were enrolled.Based on their polysomnography results,the participants were divided into an OSAS and osteoporosis comorbidity(OSAS-osteoporosis)group(n=42)and an osteoporosis only group(n=38).Propensity score matching was applied to incorporate covariates in logistic regression so that the individual characteristics of the two groups of patients were generally balanced.Following the matching procedure,a final cohort of 20 matched pairs was obtained and subsequently utilized for further analysis.The mass spectrum was obtained using laser desorption ionization mass spectrometry.Principal component analysis(PCA)was performed to assess differences in metabolic patterns between groups.Partial least squares discriminant analysis(PLS-DA)and orthogonal PLS-DA(OPLS-DA)were employed for further data analysis.Variable importance in projection(VIP)scores of each substance were calculated with OPLS-DA to screen the metabolites showing inter-group differences.Heatmaps were generated to visualize metabolic profile differences between the OSAS-osteoporosis group and the osteoporosis group.Enrichment pathway analysis was conducted on the differential identified metabolites.Results After propensity score matching,individual characteristics between the groups were well balanced.Mass spectrometry revealed significant differences between the OSAS-osteoporosis and osteoporosis groups.In the PCA score plot,the separation trend of the two groups was not significant.The PLS-DA score plot showed a discernible separation trend,with R2 and Q2 lower than those of the corresponding results of the real model,confirming the reliability of the model.OPLS-DA showed that the total R2X of the model was 0.635,R2Y was 0.879,and O2Y was 0.728,showing obvious separation trends between the two groups.A total of 16 differential metabolites were identified,including stearyl-oleyl-glycerol phosphate choline,phosphate choline,L-histidine,erucamide,2'-deoxyuridine,1-palmitoyl glycerol,thymine,tyramine,L-pyroglutamic acid,L-glutamic acid,myristate,glycerol-3-phosphate,caprylic acid,pregnenolone,L-arginine,D-4-hydroxyphenylglycine,and isobutyric acid.Heatmaps showed significant differences in metabolic profiles between the OSAS-osteoporosis group and the osteoporosis group.Pathway enrichment analysis showed that 27 metabolic pathways were involved.27 metabolic pathways.Under the conditions of P<0.05 and pathway impact>0.2,the three most significant metabolic pathways identified included mainly alanine,aspartate,and glutamate metabolism,arginine biosynthesis,and histidine metabolism.Conclusion Significant differences were observed in the metabolic profiles between patients with both OSAS and osteoporosis and those with osteoporosis alone.

OBJECTIVE: Anemoside B4 (AB4), the most abundant triterpenoidal saponin isolated from Pulsatilla chinensis, inhibited influenza virus FM1 or Klebsiella pneumoniae-induced pneumonia. However, the anti-SARS-CoV-2 effect of AB4 has not been unraveled. Therefore, this study aimed to determine the antiviral activity and potential mechanism of AB4 in inhibiting human coronavirus SARS-CoV-2 in vivo and in vitro. METHODS: The cytotoxicity of AB4 was evaluated using the Cell Counting Kit-8 (CCK8) assay. SARS-CoV-2 infected HEK293T, HPAEpiC, and Vero E6 cells were used for in vitro assays. The antiviral effect of AB4 in vivo was evaluated by SARS-CoV-2-infected hACE2-IRES-luc transgenic mouse model. Furthermore, label-free quantitative proteomics and bioinformatic analysis were performed to explore the potential antiviral mechanism of action of AB4. Type I IFN signaling-associated proteins were assessed using Western blotting or immumohistochemical staining. RESULTS: The data showed that AB4 reduced the propagation of SARS-CoV-2 along with the decreased Nucleocapsid protein (N), Spike protein (S), and 3C-like protease (3CLpro) in HEK293T cells. In vivo antiviral activity data revealed that AB4 inhibited viral replication and relieved pneumonia in a SARS-CoV-2 infected mouse model. We further disclosed that the antiviral activity of AB4 was associated with the enhanced interferon (IFN)-β response via the activation of retinoic acid-inducible gene I (RIG-1) like receptor (RLP) pathways. Additionally, label-free quantitative proteomic analyses discovered that 17 proteins were significantly altered by AB4 in the SARS-CoV-2 coronavirus infections cells. These proteins mainly clustered in RNA metabolism. CONCLUSION Our results indicated that AB4 inhibited SARS-CoV-2 replication through the RLR pathways and moderated the RNA metabolism, suggesting that it would be a potential lead compound for the development of anti-SARS-CoV-2 drugs.

Objective To evaluate the relationship between diabetic nephropathy(DN)and diabetic retinopathy(DR)in patients with type 2 diabetes mellitus(T2DM)based on imaging and clinical testing data.Methods Totally 600 T2DM patients who visited the First People's Hospital of Ziyang from March 2021 to December 2022 were included.The fundus photography and fundus fluorescein angiography were performed on all these patients and their age,gender,T2DM duration,cardiovascular diseases,cerebrovascular disease,hypertension,smoking history,drinking history,body mass in-dex,systolic blood pressure,diastolic blood pressure and other clinical data were collected.The levels of fasting blood glu-cose(FPG),triglyceride(TG),total cholesterol(TC),high-density lipoprotein cholesterol(HDL-C),low-density lipo-protein cholesterol(LDL-C),glycosylated hemoglobin(HbA1c),24 h urinary albumin(UAlb),urinary albumin to creati-nine ratio(ACR),serum creatinine(Scr)and blood urea nitrogen(BUN)were measured.Logistic regression was used to analyze the risk factors associated with DR.DR staging was performed according to fundus images,and the convolutional neural network(CNN)algorithm was used as an image analysis method to explore the correlation between DR and DN based on emission computed tomography(ECT)and clinical testing data.Results The average lesion area rates of DR and DN detected by the CNN in the non-DR,mild-non-proliferative DR(NPDR),moderate-NPDR,severe-NPDR and pro-liferative DR(PDR)groups were higher than those obtained by the traditional algorithm(TCM).As DR worsened,the Scr,BUN,24 h UAlb and ACR gradually increased.Besides,the incidence of DN in the non-DR,mild-NPDR,moderate-NPDR,severe-NPDR and PDR groups was 1.67％,8.83％,16.16％,22.16％and 30.83％,respectively.Logistic regression analysis showed that the duration of T2DM,smoking history,HbA1c,TC,TG,HDL-C,LDL-C,24 h UAlb,Scr,BUN,ACR and glomerular filtration rate(GFR)were independent risk factors for DR.Renal dynamic ECT analysis demonstrated that with the aggravation of DR,renal blood flow perfusion gradually decreased,resulting in diminished renal filtration.Conclusion The application of CCN in the early stage DR and DN image analysis of T2DM patients will improve the diag-nosis accuracy of DR and DN lesion area.The DN is worsening as the aggravation of DR.

Methods: A total of 10 specific pathogen free (SPF) grade female DBA/2J mice were randomly divided into model group and QGA II group (n = 5 for each group), while additional 5 SPF-grade female C57BL/6J mice were assigned to control group. Mice presented spontaneous high IOP and showed elevated approximately at the age of seven months. The high IOP was maintained until week 38, when gavage was initiated. Mice in control group underwent the same intragastric treatment, while those in QGA II group were gavaged with QGA II (9.67 g/kg), once a day for four weeks. Retinal morphology was examined using hematoxylin and eosin (HE) staining, with the number of retinal ganglion cells (RGCs) counted. The expression level of Brn3a protein, a specific marker for RGCs, was detected by immunofluorescence, with the mean optical density (OD) measured for quantitative analysis. In addition, 16S rDNA sequencing was leveraged to analyze changes in the diversity of gut microbiota, including their α-diversity (Chao1, Shannon, Pielou’s evenness, and observed species index) and β-diversity. Venn diagrams and linear discriminant analysis effect size (LEfSe) analysis was employed to investigate the number of amplicon sequence variants (ASVs), the abundance of differential gut microbiota species, and the classification of species at both the phylum and genus levels within the three groups of mice. Results: HE staining revealed that compared with control group, model group showed significant reduction in the number of RGCs (P < 0.01), with intracellular vacuolar degeneration and nuclear pyknosis. After QGA II treatment, the number of RGCs was significantly increased compared with model group (P < 0.01), with notable improvements in intracellular vacuolar degeneration. Immunofluorescence analysis showed that the mean OD of Brn3a protein was significantly decreased in model group compared with control group (P < 0.01), while QGA II treatment significantly elevated its expression level (P < 0.01). Analysis of α-diversity showed that after QGA II intervention, the Chao1, Shannon, and Pielou’s evenness indices were significantly increased (P < 0.01), and the observed species index was elevated (P < 0.05). β-Diversity analysis demonstrated distinct clustering among the three groups, indicating relatively low similarity in bacterial community structures. ASV clustering identified a total of 14 061 ASVs across all groups, with 9 514 ASVs shared between model and QGA II groups. At the phylum level, the abundance of Bacteroidetes was significantly decreased in model group compared with control group (P < 0.01), while Firmicutes and the Firmicutes/Bacteroidetes (F/B) ratio were significantly increased (P < 0.01). QGA II treatment significantly reduced both Firmicutes abundance and the F/B ratio (P < 0.01). At the genus level, Lactobacillus was dominant across all groups, with its abundance significantly increased in model group (P < 0.01) and subsequently decreased following QGA II intervention (P < 0.05). Conclusion QGA II restructured the gut microbiota of DBA/2J mice with chronic high IOP, bringing changes in their diversity and abundance of components. Firmicutes, Bacteroidetes, Lactobacillus, along with their associated microorganisms, are likely critical components of the gut microbiota that contribute to the optic neuroprotective effects of QGA II on chronic high IOP mice.

Objective Clinical medical record data associated with hepatitis B-related acute-on-chronic liver failure(HBV-ACLF)generally have small sample sizes and a class imbalance.However,most machine learning models are designed based on balanced data and lack interpretability.This study aimed to propose a traditional Chinese medicine(TCM)diagnostic model for HBV-ACLF based on the TCM syndrome differentiation and treatment theory,which is clini-cally interpretable and highly accurate. Methods We collected medical records from 261 patients diagnosed with HBV-ACLF,includ-ing three syndromes:Yang jaundice(214 cases),Yang-Yin jaundice(41 cases),and Yin jaun-dice(6 cases).To avoid overfitting of the machine learning model,we excluded the cases of Yin jaundice.After data standardization and cleaning,we obtained 255 relevant medical records of Yang jaundice and Yang-Yin jaundice.To address the class imbalance issue,we employed the oversampling method and five machine learning methods,including logistic regression(LR),support vector machine(SVM),decision tree(DT),random forest(RF),and extreme gradient boosting(XGBoost)to construct the syndrome diagnosis models.This study used precision,F1 score,the area under the receiver operating characteristic(ROC)curve(AUC),and accuracy as model evaluation metrics.The model with the best classification per-formance was selected to extract the diagnostic rule,and its clinical significance was thor-oughly analyzed.Furthermore,we proposed a novel multiple-round stable rule extraction(MRSRE)method to obtain a stable rule set of features that can exhibit the model's clinical in-terpretability. Results The precision of the five machine learning models built using oversampled balanced data exceeded 0.90.Among these models,the accuracy of RF classification of syndrome types was 0.92,and the mean F1 scores of the two categories of Yang jaundice and Yang-Yin jaun-dice were 0.93 and 0.94,respectively.Additionally,the AUC was 0.98.The extraction rules of the RF syndrome differentiation model based on the MRSRE method revealed that the com-mon features of Yang jaundice and Yang-Yin jaundice were wiry pulse,yellowing of the urine,skin,and eyes,normal tongue body,healthy sublingual vessel,nausea,oil loathing,and poor appetite.The main features of Yang jaundice were a red tongue body and thickened sublin-gual vessels,whereas those of Yang-Yin jaundice were a dark tongue body,pale white tongue body,white tongue coating,lack of strength,slippery pulse,light red tongue body,slimy tongue coating,and abdominal distension.This is aligned with the classifications made by TCM experts based on TCM syndrome differentiation and treatment theory. Conclusion Our model can be utilized for differentiating HBV-ACLF syndromes,which has the potential to be applied to generate other clinically interpretable models with high accura-cy on clinical data characterized by small sample sizes and a class imbalance.

Objectives:To analyze the status and temporal changes of disability-adjusted life year (DALY) for stomach and colorectal cancers among registered permanent residents in Changning District of Shanghai Municipality, and provide scientific basis for the prevention and treatment of stomach and colorectal cancers in this district.Methods:Using the cancer registration data of stomach and colorectal cancers from 2002 to 2019, we estimated the indices such as the DALYs, the DALY crude rates, the age-standardized DALY rates, etc. Then we used the Joinpoint regression model to calculate the average annual percent change (AAPC) and annual percent change (APC) to explore the temporal variations in different periods.Results:The DALYs of stomach and colorectal cancers in Changning District from 2002 to 2019 were 55 931 person years and 65 252 person years, respectively. The crude rates of DALY were 512.16/10 5 and 597.51/10 5, respectively. We observed a higher disease burden in men than in women, and the peak rate of DALY in stomach cancer was in the 75-79 years age group, while in colorectal cancer the rate was in the 85-years-or-older age group. Joinpoint regression analysis showed that from 2002 to 2019, the age-standardized DALY rate of stomach cancer showed a downward trend (AAPC=-3.86%, P＜0.05), while the trend of colorectal cancer was not statistically significant(AAPC=-0.08%, P＞0.05). However, the trends in the age-standardized DALY rates of colorectal cancer were different between males and females, with males showing an upward trend (AAPC=1.24%, P＜0.05) and females showing a downward trend (AAPC=-1.67%, P＜0.05). Conclusions:The DALY of stomach and colorectal cancers in Changning District of Shanghai showed a decreasing trend. Males and the middle-aged and elderly populations are still the key targets for disease prevention and control in this district.

Objectives:To analyze the status and temporal changes of disability-adjusted life year (DALY) for stomach and colorectal cancers among registered permanent residents in Changning District of Shanghai Municipality, and provide scientific basis for the prevention and treatment of stomach and colorectal cancers in this district.Methods:Using the cancer registration data of stomach and colorectal cancers from 2002 to 2019, we estimated the indices such as the DALYs, the DALY crude rates, the age-standardized DALY rates, etc. Then we used the Joinpoint regression model to calculate the average annual percent change (AAPC) and annual percent change (APC) to explore the temporal variations in different periods.Results:The DALYs of stomach and colorectal cancers in Changning District from 2002 to 2019 were 55 931 person years and 65 252 person years, respectively. The crude rates of DALY were 512.16/10 5 and 597.51/10 5, respectively. We observed a higher disease burden in men than in women, and the peak rate of DALY in stomach cancer was in the 75-79 years age group, while in colorectal cancer the rate was in the 85-years-or-older age group. Joinpoint regression analysis showed that from 2002 to 2019, the age-standardized DALY rate of stomach cancer showed a downward trend (AAPC=-3.86%, P＜0.05), while the trend of colorectal cancer was not statistically significant(AAPC=-0.08%, P＞0.05). However, the trends in the age-standardized DALY rates of colorectal cancer were different between males and females, with males showing an upward trend (AAPC=1.24%, P＜0.05) and females showing a downward trend (AAPC=-1.67%, P＜0.05). Conclusions:The DALY of stomach and colorectal cancers in Changning District of Shanghai showed a decreasing trend. Males and the middle-aged and elderly populations are still the key targets for disease prevention and control in this district.

Objective To study the clinical application value of abdominal pressure band in CT-guided puncture biopsy of lower lung nodules.Methods A retrospective analysis was conducted on 100 patients who underwent lower lung nodule punc-ture biopsy at the Cancer and Radiotherapy Center of Tongji Hospital affiliated to Huazhong University of Science and Technol-ogy from January to December 2022.They were divided into abdominal pressure band group and control group.There were 40 patients in abdominal pressure band group,among which abdominal pressure bands were applied during the puncture process.There were 60 patients in control group who did not use abdominal pressure bands during the puncture process.The CT scan frequency,puncture depth,success rate,intraoperative time,incidence of complications,and radiation dose for puncture biopsy were compared between the two groups.Results The median value of CT scan frequency during puncture biopsy in ab-dominal pressure band group was 5 times,with a success rate of 100％(40/40).The median value of intraoperative time was 7.9 min,and the incidence of complications was 10％(4/40).The median value of radiation dose was 210.4 mGy·cm.The median value of CT scan frequency during puncture biopsy in control group was 8 times,with a success rate of 86.7％(52/60).The me-dian value of intraoperative time was 13 min,and the incidence of complications was 40％(24/60).The median value of radiation dose was 287.7 mGy·cm.Differences in five indicators were statistically significant between the two groups(all P＜0.05).Conclusion Application of the abdominal pressure bands in lower lung nodule puncture biopsy can significantly reduce the number of scans,shorten surgical time,reduce the radiation dose received by patients,and have a higher success rate and lower incidence of complications.Utilization of pressure band should be promoted in clinical practice.

Objective To analyze the allocation efficiency of medical and health resources in 26 tertiary hospitals in Suzhou from 2017 to 2022 and perform quantitative analysis in order to provide suggestions for relevant departments to rationally coordinate the health resources allocation,regional health planning and hospital management.Methods The number of health technicians,the number of beds in health institutions and the total health expenditure were selected as input indicators,while the number of diagnostic and treatment visits and the number of discharged patients were selected as output indicators.The efficiency was measured by the SBM(Slack-Based Measure)model and the SBM window model respectively.Results Influenced by public health emergencies,the allocation efficiency of medical resources in tertiary hospitals in Suzhou city decreased first and then started to increase Under the two models,the average efficiency scores of 26 hospitals were 0.687 and 0.707,respectively.Notably,under the SBM window model,19 hospitals(73％)achieved efficiency scores near or above the average.Conclusion The two models present a conclusion that the overall efficiency of the tertiary hospitals in Suzhou is generally effective.To further enhance the allo-cation and utilization of medical resources,it is suggested that a comprehensive consideration of health needs guide the planning of medical resource distribution.Leveraging information technology to innovate medical service models,and strengthening internal de-velopment and management practices are essential strategies for promoting high-quality development in tertiary hospitals.