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 establish the UPLC fingerprint and the method for multi-component content determination in Jingangteng capsules， and to evaluate its quality by combining chemical pattern recognition analysis. METHODS An UPLC method was established. Separation was performed on a Zorbax SB-C 18 Rapid Resolution HD column， with acetonitrile-0.1% formic acid as the mobile phase for gradient elution.Using the Similarity Evaluation System for Chromatographic Fingerprints of Traditional Chinese Medicines （2012 edition）， UPLC fi ngerprints were established for 10 batches of Jingangteng capsules， and similarity was evaluated. SPSS 22.0 and SIMCA 14.1 software were used to perform hierarchial-cluster analysis and orthogonal partial least squares discriminant analysis （OPLS-DA）， respectively. The same UPLC method was employed to determine the contents of chlorogenic acid， 3，5-dihydroxy-2-methylbenzoic acid-3- O -glucoside （M1）， caffeic acid， astilbin， oxyresveratrol， quercitrin and resveratrol in the 10 batches of samples. RESULTS A total of 17 common peaks were identified in UPLC fingerprints of the 10 batches of samples， of which 7 were identified as chlorogenic acid， M1， caffeic acid， astilbin， oxyresveratrol， quercitrin， and resveratrol. The similarities of 10 batches of samples ranged from 0.820 to 0.985. The results of hierarchial-cluster analysis showed that 10 batches of samples were grouped into four categories： S1-S4 formed one group， S5 and S6 formed another， S7， S8 and S10 formed a third， and S9 formed a fourth， consistent with the OPLS-DA results； the variable importance projection values for peaks 7， 10， 2， 16 （resveratrol）， 13 （oxyresveratrol）， 11， 6 （caffeic acid）， 5 （M1） and 15 （quercitrin） were ＞1. Quantitative analysis results showed that the contents of chlorogenic acid， M1， caffeic acid， astilbin， oxyresveratrol， quercitrin， and resveratrol were 1.650 8-4.213 7， 0.636 2-2.161 7， 0.031 0-0.086 5， 0.239 1-1.069 3， 0.211 9-1.104 0， 0.488 8-2.399 2， and 0.164 0-0.699 8 mg/g， respectively. CONCLUSIONS UPLC fingerprint and content determination methods established in this study are simple to operate， accurate， reliable and reproducible； when combined with chemical pattern recognition analysis， they can be used to evaluate the quality of Jingangteng capsules. Nine components， such as resveratrol， oxyresveratrol， caffeic acid， M1 and quercitrin， may serve as markers of quality variation.

Objective To investigate the association between school bullying and comorbidity of overweight/obesity and depressive symptoms in middle school students, and to provide a basis for formulating intervention strategies for the prevention and control of common diseases and comorbidities among students. Methods In September 2023, a stratified cluster random sampling method was employed to select junior high school, high school, and vocational high school students in 103 districts and counties across 17 cities (prefectures) in Hubei Province. A questionnaire survey was conducted on these students regarding school bullying and depressive symptoms, and physical measurements of height and weight were taken. The association between school bullying and comorbidity of overweight/obesity and depressive symptoms was analyzed using a logistic regression model. Results A total of 99 921 middle school students aged 11-19 years old were included in the study. The detection rate of overweight/obesity was 29.39%, the detection rate of depressive symptoms was 31.15%, and the detection rate of comorbidity of overweight/obesity and depressive symptoms was 9.15%. Among all students, 24.12% reported experiencing school bullying. Multivariate logistic regression analysis showed that experiencing school bullying was positively associated with the comorbidity of overweight/obesity and depressive symptoms among middle school students (OR = 2.692, 95% CI: 2.571-2.817, P < 0.001). Stratified analysis results showed that both male and female students who experienced school bullying were more likely to have comorbidity of overweight/obesity and depressive symptoms (OR values were 2.756 for males and 2.594 for females; P values were both < 0.001). Students in all educational stages who experienced school bullying had an increased risk of comorbidity of overweight/obesity and depressive symptoms (OR values were 2.739 for junior high, 2.676 for senior high, and 2.378 for vocational high; P values were all < 0.001). Conclusion The comorbidity of overweight/obesity and depressive symptoms among middle school students is widespread, and there is a significant association between school bullying and comorbidity of overweight/obesity and depressive symptoms. It is necessary to strengthen the prevention and control of school bullying to promote the healthy physical and mental development of adolescents.

OBJECTIVE To investigate the antidepressant effect and underlying mechanism of total triterpenes from Poriae Cutis （PCTT）. METHODS UPLC-Q-TOF-MS technology was adopted to analyze the chemical constituents of PCTT. Network pharmacology was applied to screen the antidepressant active ingredients， key targets and signaling pathways of PCTT， and molecular docking was performed to verify the binding capacity between potential active ingredients and core targets. In animal experiment， mice were intragastrically administered PCTT at doses of 80， 160 and 320 mg/kg for 14 consecutive days. Lipopolysaccharide was used to establish an acute depression mouse model. Subsequently， behavioral assays were conducted. The levels of serum inflammatory factors， hippocampal neurotransmitters and brain-derived neurotrophic factor （BDNF） were detected. Western blot and qPCR assays were adopted to validate the antidepressant mechanism of PCTT. RESULTS A total of 101 chemical constituents were identified in PCTT. Network pharmacology screened out 10 potential active ingredients （including poriacosone B， poriacosone A， pachymic acid， etc.）， six core targets [including tumor necrosis factor （TNF）， peroxisome proliferator-activated receptor， caspase-3， etc.]， as well as multiple signaling pathways including TNF and interleukin-17 （IL-17）. All 10 potential active ingredients exhibited strong binding affinity to the six core targets. Animal experimental results demonstrated that 320 mg/kg PCTT markedly increased the movement distance and movement speed of depression model mice （P＜0.05）， and shortened immobility time in tail suspension test and forced swimming test （P＜0.05）. PCTT reduced serum levels of IL-6， IL-1β and TNF-α， decreased hippocampal glutamate level； it also downregulated the relative mRNA expression of TNF-α， tumor necrosis factor receptor-2 （TNFR2）， phosphoinositide 3-kinase （PI3K） and protein kinase B （AKT）， as well as TNFR2 protein expression and the ratios of phosphorylated-PI3K/PI3K，phosphorylated-AKT/AKT in the prefrontal cortex. Meanwhile， PCTT elevated the levels of 5- hydroxytryptamine， dopamine and BDNF in the hippocampus. CONCLUSIONS PCTT exerts antidepressant effects possibly by regulating the TNF-α/PI3K/AKT signaling pathway to downregulate the expression of pro-inflammatory factors， thereby modulating the release of neurotransmitters and BDNF.

OBJECTIVE To explore the pharmacokinetic characteristics of 3 blood-absorbed components of Xiangshao sanjie oral liquid in rats with hyperplasia of mammary gland （HMG）. METHODS Female SD rats were divided into control group and HMG group according to body weight， with 6 rats in each group. The HMG group was given estrogen+progesterone to construct HMG model. After modeling， two groups were given 1.485 g/kg of Xiangshao sanjie oral liquid （calculated by crude drug） intragastrically， once a day， for 7 consecutive days. Blood samples were collected before the first administration （0 h）， and at 5， 15， 30 minutes and 1， 2， 4， 8， 12， 24 hours after the last administration， respectively. Using chlorzoxazone as the internal standard， the plasma concentrations of ferulic acid， paeoniflorin and rosmarinic acid in rats were detected by UPLC-Q/TOF-MS. The pharmacokinetic parameters [area under the drug time curve （AUC0－24 h， AUC0－∞）， mean residence time （MRT0－∞）， half-life （t1/2）， peak time （tmax）， peak concentration （cmax）] were calculated by the non-atrioventricular model using Phoenix WinNonlin 8.1 software. RESULTS Compared with the control group， the AUC0－24 h， AUC0－∞ and cmax of ferulic acid in the HMG group were significantly increased （P＜0.05）； the AUC0－24 h， AUC0－∞ ， MRT0－∞ ， t1/2 and cmax of paeoniflorin increased， but there was no significant difference between 2 groups （P＞0.05）； the AUC0－24 h and MRT0-∞ of rosmarinic acid were significantly increased or prolonged （P＜0.05）. C ONCLUSIONS In HMG model rats， the exposure of ferulic acid， paeoniflorin and rosmarinic acid in Xiangshao sanjie oral liquid all increase， and the retention time of rosmarinic acid is significantly prolonged.

ObjectiveBased on modern analytical techniques and a depressed mouse model established by chronic unpredictable mild stress（CUMS）， to evaluate the quality marker（Q-Marker） and pharmacodynamic difference of Baihe Dihuangtang prepared by different processes. MethodsHigh performance liquid chromatography（HPLC） was used to establish the characteristic profiles of Baihe Dihuangtang， and determine the content of Q-Marker in the samples prepared by ancient and modern processes. Seventy C57BL/6J mice were randomly divided into the normal group， model group， fluoxetine group（3 mg·kg^-1）， low and high dose groups of ancient process（6.5， 26 g·kg^-1）， and low and high dose groups of modern process（6.5， 26 g·kg^-1）， with 10 mice in each group. Except for the normal group， CUMS was used to induce depression in mice from the other groups for 28 d. After successful modeling， administration groups were given the corresponding drugs by gavage every day， and the normal and model groups were given an equal volume of pure water by gavage for 21 consecutive days. Change in body mass of mice was recorded， tail suspension test and open field test were used to evaluate the depressive behavior of mice， and enzyme-linked immunosorbent assay（ELISA） was employed to determine the contents of tumor necrosis factor-α（TNF-α）， interleukin（IL）-1β and IL-6 in serum. ResultsCharacteristic profiles of Baihe Dihuangtang prepared by the two processes were established， the similarity between the two was 0.951， and 8 characteristic peaks were recognized with the reference peak of regaloside A. The results of quantitative analysis showed that the Q-Marker content was similar in Baihe Dihuangtang prepared by ancient and modern processes. The results of pharmacodynamics showed that， compared with the normal group， the model group showed increased immobility time in the tail suspension test， reduced total movement distance in the open field test， and elevated IL-1β， IL-6 and TNF-α levels in the serum（P<0.01）. Compared with the model group， the behavioral indicators of mice in the Baihe Dihuangtang treatment group were significantly improved in terms of tail suspension time and open field exercise， and the levels of IL-1β， IL-6 and TNF-α in serum were significantly reduced（P<0.05， P<0.01）. Baihe Dihuangtang prepared by the two processes both had antidepressant effects， and the difference between the two was not statistically significant in improving depressive symptoms. ConclusionQ-Marker of Baihe Dihuangtang prepared by modern and ancient methods are equivalent in content， and the pharmacological effects are consistent， indicating that dried Lilii Bulbus can replace fresh products in the preparation of Baihe Dihuangtang. This study provides a scientific basis for the development of new drugs of Baihe Dihuangtang and a reference for its rational application and clinical use.

ObjectiveBased on modern analytical techniques and a depressed mouse model established by chronic unpredictable mild stress（CUMS）， to evaluate the quality marker（Q-Marker） and pharmacodynamic difference of Baihe Dihuangtang prepared by different processes. MethodsHigh performance liquid chromatography（HPLC） was used to establish the characteristic profiles of Baihe Dihuangtang， and determine the content of Q-Marker in the samples prepared by ancient and modern processes. Seventy C57BL/6J mice were randomly divided into the normal group， model group， fluoxetine group（3 mg·kg^-1）， low and high dose groups of ancient process（6.5， 26 g·kg^-1）， and low and high dose groups of modern process（6.5， 26 g·kg^-1）， with 10 mice in each group. Except for the normal group， CUMS was used to induce depression in mice from the other groups for 28 d. After successful modeling， administration groups were given the corresponding drugs by gavage every day， and the normal and model groups were given an equal volume of pure water by gavage for 21 consecutive days. Change in body mass of mice was recorded， tail suspension test and open field test were used to evaluate the depressive behavior of mice， and enzyme-linked immunosorbent assay（ELISA） was employed to determine the contents of tumor necrosis factor-α（TNF-α）， interleukin（IL）-1β and IL-6 in serum. ResultsCharacteristic profiles of Baihe Dihuangtang prepared by the two processes were established， the similarity between the two was 0.951， and 8 characteristic peaks were recognized with the reference peak of regaloside A. The results of quantitative analysis showed that the Q-Marker content was similar in Baihe Dihuangtang prepared by ancient and modern processes. The results of pharmacodynamics showed that， compared with the normal group， the model group showed increased immobility time in the tail suspension test， reduced total movement distance in the open field test， and elevated IL-1β， IL-6 and TNF-α levels in the serum（P<0.01）. Compared with the model group， the behavioral indicators of mice in the Baihe Dihuangtang treatment group were significantly improved in terms of tail suspension time and open field exercise， and the levels of IL-1β， IL-6 and TNF-α in serum were significantly reduced（P<0.05， P<0.01）. Baihe Dihuangtang prepared by the two processes both had antidepressant effects， and the difference between the two was not statistically significant in improving depressive symptoms. ConclusionQ-Marker of Baihe Dihuangtang prepared by modern and ancient methods are equivalent in content， and the pharmacological effects are consistent， indicating that dried Lilii Bulbus can replace fresh products in the preparation of Baihe Dihuangtang. This study provides a scientific basis for the development of new drugs of Baihe Dihuangtang and a reference for its rational application and clinical use.

Objective To investigate the therapeutic effects and underlying mechanisms of Ganluqingwen formula on lipopolysaccharide(LPS)-induced acute lung injury/acute respiratory distress syndrome(ALI/ARDS)in mice.Methods Fifty ICR mice were randomly divided into five groups:control,model,and Ganluqingwen formula(GLQW)low dose(7.10 g/kg),medium dose(15.21 g/kg),and high dose(30.42 g/kg)groups,with 10 mice per group.On days 1-3,mice in GLQW groups were daily gavaged with the corresponding dose of GLQW,while control and model groups received equal volumes of saline.On day 4,ALI/ARDS was induced in model and GLQW groups using intraperitoneal injection of LPS(20 mg/kg),while control group received an equal volume of PBS.At 24 h post-treatment,survival rate,wet-to-dry weight ratio(W/D)and lung histological changes(HE staining)were observed.Serum levels of tumor necrosis factor(TNF)-α,interferon gamma(IFN-γ),interleukin(IL)-4,IL-10,IL-12,as well as lung tissue levels of TNF-α,IFN-γ,IL-1β,IL-4,IL-6,IL-10 were measured by ELISA.Western blotting was used to determine the expression levels of NOD-like receptor thermal protein domain associated protein 3(NLRP3),cystatinase-1(Caspase-1),apoptosis-associated speck-like protein(ASC),and membrane perforating protein Gasdermin D(GSDMD)in lung tissue.Results No significant differences in survival rates were observed among the groups(P>0.05).Compared with control group,ELISA and Western blotting results showed that lung tissue W/D,IFN-γ,TNF-α,IL-4,IL-12,IL-1β,IL-6,NLRP3,ASC,and Caspase-1,GSDMD and serum IFN-γ,TNF-α,IL-4,IL-12 levels were significantly higher(P<0.05),and IL-10 levels in lung tissue and serum were significantly lower in mice of model group(P<0.05).Compared with model group,lung tissue W/D,IFN-γ,TNF-α,IL-1β,IL-4,IL-6,IL-12,NLRP3,ASC,Caspase-1,and GSDMD,and serum IFN-γ,TNF-α,IL-4,and IL-12 levels were significantly lower(P<0.05),and lung tissue IL-10 levels were significantly higher(P<0.05)in GLQW low,medium,and high dose groups,with high-dose group showing significantly higher level in serum IL-10(P<0.05).Compared with GLQW low-dose group,the lung tissue levels of IFN-γ,IL-6,NLRP3,ASC,Caspase-1,and GSDMD,and serum TNF-α were significantly lower(P<0.05),and lung and serum IL-10 levels were significantly higher in GLQW high-dose group(P<0.05).HE staining results showed that lung structure was clear and normal in control group;part of the lung interstitium was congested and hemorrhagic,and some of the fine bronchial periphery was infiltrated with inflammatory cells in model group;the phenomena of lung interstitial congestion and hemorrhage were reduced,and the degree of infiltration of inflammatory cells was alleviated in GLQW low-,medium-,and high-dose groups.Conclusion Ganluqingwen formula can delay the development of ALI/ARDS in mice by inhibiting NLRP3/Caspase-1/GSDMD pathway,thereby suppressing cellular pyroptosis.

Lung cancer has the highest incidence and mortality among malignant tumors in China.Persistent subsolid nodules(SSNs)are closely associated with early-stage lung adenocarcinoma.Artificial intelligence(AI),as an emerging technology,is capable of performing in-depth analysis of large-scale imaging data through autonomous learning and possesses the ability to predict outcomes from new data,demonstrating great potential and application prospects in the assessment of SSNs.AI can not only effectively assist radiologists in diagnosis and treatment,but also improve work efficiency while reducing misdiagnosis and missed diagnosis rates.This review summarizes the recent applications and research progress of AI in the assessment of SSNs,to provide new insights for the diagnosis and treatment of SSNs.

Objective To explore the correlation between the rs7296288 single nucleotide polymorphism(SNP)of the DHH gene and the onset of depression.Methods From March 2018 to December 2019,480 patients with depression in Sichuan Provincial People's Hospital,Jining Mental Hospital,Yunnan Mental Health Center were collected as the case group,and 329 patients without depression were collected as the control group.Peripheral blood samples were collected and DNA was extracted.After multiple amplification and high-throughput sequencing,statistical software was used to analyze the relevant data.Results The rs7296288 polymorphism of the DHH gene,the AA genotype was 32.9％in the control group and 36.7％in the case group,with no statistical difference;the AC genotype was 52.8％in the control group and 47.7％in the case group,with no statistical difference;the CC genotype was 14.3％in the control group and 15.6％in the case group,with no statistical difference.The co-dominant,dominant,and recessive statistical models were not statistically significant.Subgroup analysis of the degree of depressive episodes,suicide attempts,and first-episode patients did not find an association between the rs7296288 SNP locus of the DHH gene and the clinical features of depression,that is,the SNP rs7296288 of the DHH gene had no significant correlation with depression.Conclusion The rs7296288 polymorphism of DHH gene is not correlated with the pathogenesis of depression.