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.

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 analyze chemical components of Yao ethnic medicine Pittosporum pauciflorum， establish its fingerprint and investigate the spectrum-effect relationship of its anti-inflammatory effect. METHODS UHPLC-Q-Orbitrap-MS technology was used to analyze the chemical components of P. pauciflorum （batch S6）. The fingerprints for 10 batches of P. pauciflorum from different producing areas in Guangxi Province （batches S1-S10） were established by HPLC， and similarity assessment and chemometric pattern recognition analysis were conducted. RAW264.7 inflammatory cell model was induced by lipopolysaccharide， and the anti-inflammatory activity of P. pauciflorum was investigated. Using inhibition rates of nitric oxide （NO）， tumor necrosis factor-α （TNF-α）， interleukin-6 （IL-6） and IL-1β as efficacy indicators， grey relational analysis and partial least squares regression analysis were adopted to evaluate the spectrum-effect relationship of the anti-inflammatory effect of P. pauciflorum. RESULTS There were 60 chemical components， including flavonoids， phenolic acids， lipids， etc.， identified in P. pauciflorum. The fingerprints for 10 batches of P. pauciflorum showed 14 common peaks，with similarity values ranging from 0.883 to 0.991. Three common peaks were assigned neochlorogenic acid （peak 5）， chlorogenic acid （peak 7）， and syringaldehyde （peak 10）. The classification results of the systematic clustering analysis and the principal component analysis were basically consistent. Batches S1 to S10 of P. pauciflorum significantly reduced the levels of NO， TNF-α， IL-6 （except for batch S5） and IL-1β in the cell supernatant （P＜0.05 or P＜0.01）. Inhibition rates of above inflammatory indexes were 10.26%-39.96%， 14.96%-31.36%， 1.38%-21.27%， 18.54%-28.00%， respectively. The contents of neochlorogenic acid， syringaldehyde， as well as the components corresponding to peaks 1， 3， 9， 12 and 14，exhibited a strong correlation with the anti-inflammatory effects of P. pauciflorum. CONCLUSIONS The present study has analyzed the chemical components of P. pauciflorum and established HPLC fingerprints for 10 batches of P. pauciflorum. Each batch of medicinal herbs demonstrates certain anti- inflammatory activities， among which neochlorogenic acid， syringaldehyde， and the components corresponding to peaks 1， 3， 9， 12 and 14 are likely to be the active anti-inflammatory components.

Plants of the Isodon genus are an important source of terpenoids， with their constituents exhibiting rich structural diversity and remarkable biological activities （such as anticancer， antimicrobial， and anti-inflammatory properties）， demonstrating significant potential in the field of immunomodulation. This review summarizes recent advances in the immunomodulatory mechanisms， development and application of terpenoid compounds from the Isodon genus. It has been found that these compounds can modulate key inflammatory signaling pathways， such as nuclear factor-kappa B （NF-κB） and mitogen-activated protein kinases （MAPKs）， thereby blocking the cascade amplification of inflammatory factors， alleviating chronic inflammatory responses， and correcting immune dysregulation. Additionally， they can influence the polarization direction of macrophages and dynamically regulate the balance among different functional subsets of T cells， restoring immune homeostasis. Their clinical translation faces multiple challenges， including poor druggability， a lack of systematic safety data， the absence of precise pharmacodynamic biomarkers， complexities in clinical trial design， and unclear industrialization pathways.

OBJECTIVE: A new three-dimensional(3D) classification of posterior cruciate ligament (PCL) tibial avulsion fracture based on computed tomography(CT) features was established and the significance in clinical treatment was explored in this study. METHODS: From May 2013 to November 2023, 43 cases of PCL tibial avulsion fracture in the Second Hospital of Jilin University were analyzed retrospectively, including 29 males and 14 females, aged (34.3±8.5) years. According to traditional Meyers and McKeever classification, 3 cases were typeⅠ;2 cases of typeⅡ;38 cases were type Ⅲ. Based on the characteristics of CT images, 43 patients were given specific treatment strategies and followed up to evaluate the curative effect. According to the degree of fracture displacement, involved range and the integrity of fracture block demonstrated by CT images, the new three-dimensional classification of PCL avulsion fracture was established. Kappa coefficient was used for consistency test. RESULTS: A new 3D classification of PCL tibial avulsion fracture was established. TypeⅠwas the non-displaced fracture (displacement degree ≤3 mm), in which typeⅠa was the avulsion range limited in the posterior intercondylar fossa, and Ib was the avulsion range beyond the posterior intercondylar fossa. TypeⅡrepresented the displaced fracture in the posterior intercondylar fossa (avulsion limited to the posterior intercondylar fossa and fracture displacement>3 mm), in which typeⅡa represented a slight displacement with a intact broken block and the posterior elevation of the avulsion (hinge mechanism), typeⅡb represented the complete separation of fracture ends with a intact fracture block, and typeⅡc was the comminuted fracture. Type Ⅲ was the displaced fracture beyond the posterior intercondylar fossa (avulsion involving the articular surface of the tibial plateau or the intercondylar ridge and the degree of displacement > 3 mm), among which type Ⅲa was the simple fracture with intact broken block, type Ⅲb represented the comminuted fracture, and type Ⅲc was the complex fracture with tibial plateau fracture. According to this new 3D classification, 43 patients were classified as type Ia in 2 cases and typeⅠb in 1 case;typeⅡa in 2 cases, typeⅡb in 15 cases and typeⅡc in 7 cases;type Ⅲa in 2 cases, type Ⅲb in 5 cases and type Ⅲc in 9 cases. All the 43 cases in this study achieved bone union. At the last follow-up, according to the hospital for special surgery knee score(HSS)evaluation system for the knee joint function, 27 cases were excellent, 11 cases were good, 5 cases were fair. The average Kappa value of inter-observer reliability in the first stage was 0.793, and the second stage was 0.855. The average Kappa value of the whole stage was 0.839, indicating high level of consistency. The average Kappa value of intra-observer reliability was 0.893, indicating high level of consistency. CONCLUSION The 3D classification of PCL tibial avulsion fracture is intuitive, demonstrating a high level of reliability. It has a certain guiding significance for the selection of clinical treatment methods, and it is suggested to be promoted and applied as a new classification system in clinical practice.
Humans ; Male ; Female ; Posterior Cruciate Ligament/surgery* ; Adult ; Tibial Fractures/classification* ; Tomography, X-Ray Computed ; Middle Aged ; Retrospective Studies ; Fractures, Avulsion/classification* ; Imaging, Three-Dimensional ; Young Adult

OBJECTIVES: To study the incidence and disease burden of congenital birth defects in children under five in China from 1990 to 2021 and to predict the incidence of congenital birth defects in this population from 2022 to 2036, providing a reference for the prevention of congenital birth defects in children. METHODS: Using the Global Burden of Disease Study 2021 (GBD 2021) database, the incidence and disability-adjusted life years (DALY) were employed to describe the disease burden. The Joinpoint regression model was used to analyze the trends in incidence and DALY rates of congenital birth defects in children under five. A grey prediction model GM(1,1) was applied to fit the trend of incidence rates of congenital birth defects in this age group and to predict the incidence from 2022 to 2036. RESULTS: In 2021, the incidence rate of congenital birth defects among children under five in China was 737.28 per 100 000. Among these, congenital musculoskeletal and limb deformities had the highest incidence rate at 307.15 per 100 000, followed by congenital heart defects (223.53 per 100 000), congenital urinary and genital tract malformations (74.99 per 100 000), and congenital gastrointestinal malformations (62.61 per 100 000). From 1990 to 2021, the incidence rate and DALY rate of congenital birth defects in children under five in China decreased at an average annual rate of 1.73% and 5.42%, respectively. The prediction analysis indicated a decreasing trend in the incidence of congenital birth defects among children under five in China from 2022 to 2036, with the incidence rate dropping from 892.36 per 100 000 in 2022 to 783.35 per 100 000 in 2036. CONCLUSIONS The incidence and disease burden of congenital birth defects in children under five in China showed a significant declining trend from 1990 to 2021. It is predicted that this incidence will continue to decrease until 2036.
Humans ; Congenital Abnormalities/epidemiology* ; China/epidemiology* ; Incidence ; Infant ; Infant, Newborn ; Child, Preschool ; Female ; Male ; Forecasting ; Disability-Adjusted Life Years