OBJECTIVE: To explore the application value of artificial intelligence (AI)-assisted chromosomal karyotype analysis in the diagnosis of prenatal chromosomal mosaicism. METHODS: A retrospective analysis was conducted on 172 pregnant women who underwent amniocentesis at the Department of Medical Genetics and Prenatal Diagnosis, the Third Affiliated Hospital of Zhengzhou University between January 2019 and December 2024. All cases whose fetuses were diagnosed with chromosomal mosaicism via karyotype analysis and stratified into two groups based on the analytical software employed: the conventional analysis group (n = 70), which utilized Leica analysis software for karyotype image recognition and cell counting; and the AI-assisted analysis group (n = 102), which utilized AI-assisted software for the same procedures. The clinical performance of AI-assisted karyotype analysis in diagnosing chromosomal mosaicism was comprehensively evaluated by comparing the types of mosaic karyotypes, distribution of mosaic ratios, and verification outcomes of different detection modalities between the two groups. This study was approved by the Medical Ethics Committee of the Third Affiliated Hospital of Zhengzhou University (Ethics No.: 2024-406-01). RESULTS: No statistically significant difference was observed in baseline characteristics (maternal age, gestational week, and indications for prenatal diagnosis) between the two groups. Regarding the detection efficacy for numerical and structural mosaicisms, no significant difference was found in the detection of numerical mosaicism. However, the conventional analysis group exhibited a significantly higher detection rate of autosomal structural mosaicism compared to the AI-assisted group (11.43% vs. 0.98%, P < 0.05). Numerical mosaicism cases were further verified using copy number variation sequencing (CNV-seq) and/or fluorescence in situ hybridization (FISH). The AI-assisted group demonstrated a significantly lower inconsistency rate (5.56% vs. 20.41%, P < 0.05) compared to the conventional group. For low-proportion (< 10%) chromosomal mosaicism, the AI-assisted group had a significantly lower detection rate (13.25% vs. 29.69%, P < 0.05). Subsequent validation of low-proportion mosaicism by CNV-seq and/or FISH showed a higher consistency rate in the AI-assisted group (81.82% vs. 54.55%), though the difference did not reach statistical significance (P = 0.360). CONCLUSION For the karyotyping analysis of prenatal chromosomal mosaicism, AI-assisted karyotype analysis shows high accuracy and consistency in identifying numerical chromosomal mosaicism, particularly in reducing the detection of low-proportion (< 10%) mosaicism while improving verification accuracy. AI-assisted analysis can significantly improve the detection accuracy of numerical mosaicism and mitigate the risk of misclassification for low-proportion (< 10%) mosaicism, thereby providing more precise clinical evidence for the prenatal diagnosis of chromosomal mosaicisms.
Humans ; Female ; Mosaicism ; Pregnancy ; Karyotyping/methods* ; Artificial Intelligence ; Prenatal Diagnosis/methods* ; Adult ; Retrospective Studies ; Chromosome Disorders/genetics* ; Amniocentesis

Dry eye （DE） is a prevalent multifactorial disease of the ocular surface， clinically characterized by tear film homeostasis imbalance accompanied by related ocular surface symptoms. Specifically， the tear film is a thin liquid layer of tears covering the cornea and conjunctiva through blinking， while tear film homeostasis serves as the foundation for maintaining normal ocular surface structure and function. Insufficient tear secretion and excessive tear film evaporation lead to tear hyperosmolarity and the production of inflammatory mediators， disrupting tear film homeostasis and subsequently forming DE. Additionally， cascade reactions are triggered， resulting in a "vicious cycle of DE" that exacerbates the disease severity and prolongs its duration. Therefore， for DE treatment， it is crucial to restore tear film homeostasis and terminate this vicious cycle. Traditional Chinese medicine （TCM）， which differentiates and treats DE based on systemic conditions， often achieves favorable therapeutic outcomes， offering additional treatment options for DE. Studies have demonstrated that TCM can alleviate DE by regulating tear film homeostasis and terminating the vicious cycle. This review systematically summarizes recent basic experimental research in China and abroad on TCM in alleviating DE by regulating tear film homeostasis， aiming to provide a theoretical basis for clinical treatment and an insight for research design.

ObjectiveTo establish a model of dry eye with lung Yin deficiency syndrome in mice. MethodsA total of 40 SPF C57BL/6J mice were assigned via the random number table method into 5 groups （n=8）： Normal control， model control， and high-， medium-， and low-dose （11.7， 5.85， and 2.925 g·kg^-1， respectively） Yangyin Qingfeitang. Mice in the normal control group were fed normally without any intervention. Mice in Yangyin Qingfeitang group and model control group were treated with 0.2% benzalkonium chloride eye drops （5 μL） twice a day and fed in a controlled drying system in a dry environment for 28 days. At the same time， the mice were administrated with thyroxine tablet solution by gavage and placed in a glass fumigation tank （SO₂ concentration： 0.5 g·m^-3） for 14 days. After 4 weeks， mice in Yangyin Qingfeitang groups were treated with Yangyin Qingfeitang by gavage and those in the normal control group and model control group were administrated with deionized water at 0.01 mL·g^-1. The body mass， anal temperature， four examination information （claw and nail appearance）， basic tear secretion test， tear film rupture time， corneal fluorescein staining， and lacrimal gland HE staining were compared among groups. Compound Yangyin Qingfeitang granules were used to measure the syndrome to verify the success of modeling. ResultsAfter 28 days of continuous modeling， compared with the normal control group， the model group exhibited listless and emaciated status， coughing， drowsiness， dry and dull hair， dry and hard stool， reduced food intake and water intake， red lip circumference， red tongue with reduced fluid， dry nose and teeth， red claws and nails， body mass gain， decreased anal mild tear secretion （P<0.05）， and shortened tear film rupture time （P<0.05）. After 28 days of modeling， the mice showed large corneal fluorescein staining range， severe corneal injury， and increased content of interleukin （IL）-18， IL-β， and tumor necrosis factor （TNF）-α in lacrimal gland， compared with those in the normal control group （P<0.05）. After the treatment with Yangyin Qingfeitang， the mice had good drinking and eating conditions， with lighter redness of the tongue， moist nose， moist and shiny teeth， and the claw and nail color close to that in the normal group. Compared with the model control group， Yangyin Qingfeitang groups showed increases in body mass and anal temperature （P<0.05）， tear secretion （P<0.05）， and tear film rupture time （P<0.05）， narrowed range of corneal fluorescein staining， and declined levels of IL-18， IL-β， and TNF-α in lacrimal glands （P<0.01）. The high-dose group had the best effect， with the indicators close to the levels in the normal control group. ConclusionThe animal model of dry eye with lung Yin deficiency syndrome can be established by culture in a controlled drying system， treatment with benzalkonium chloride eye drops for 28 days， and administration of thyroxine tablet solution combined with SO₂ fumigation for 14 days.

ObjectiveTo establish a mouse model of particulate matter 2.5 （PM_2.5）-induced dry eye and investigate whether Chufeng Yisuntang can ameliorate the PM_2.5-induced ocular surface damage by regulating the reactive oxygen species （ROS）/p38 mitogen-activated protein kinase （p38 MAPK） signaling pathway. MethodsSixty 8-week-old male C57BL/6J mice were used. Ten were randomly selected as the control group. The remaining 50 mice received topical instillation of 1 drop （0.1 mL） of 5 g·L^-1 PM_2.5 suspension in both eyes， four times daily. Successfully modeled mice were randomized into four groups （n=10）： Model， p38 MAPK inhibitor， Chufeng Yisuntang， and combination （Chufeng Yisuntang at 7.3 g·kg^-1 + p38 MAPK inhibitor SB203580 at 5 mg·kg^-1）. Chufeng Yisuntang was administered via gavage， and the inhibitor group via intraperitoneal injection. The control and model groups received equal volumes of distilled water by gavage. All treatments lasted for 4 weeks. General conditions were dynamically observed. Tear secretion， tear film break-up time， and corneal fluorescein staining were assessed. After intervention for 4 weeks， hematoxylin and eosin （HE） staining was used to examine the histopathological changes. Enzyme-linked immunosorbent assay （ELISA） was adopted to measure serum levels of ROS， malondialdehyde （MDA）， superoxide dismutase （SOD） 1， and SOD2. Western blot and Real-time PCR were employed to determine the protein and gene levels， respectively， of p38 MAPK， B-cell lymphoma-2 （Bcl-2）， Bcl-2-associated X protein （Bax）， and cysteinyl aspartate-specific proteinase-3 （Caspase-3） in the corneal tissue. ResultsCompared with the control group， the model group exhibited reduced tear secretion volume and tear film breakup time， along with increased corneal fluorescein staining scores （P<0.01）. Compared with the model group， the Chufeng Yisuntang group， p38 MAPK inhibitor group， and combination group demonstrated increased tear secretion volume and tear film breakup time， along with decreased corneal fluorescein staining scores （P<0.01）. HE staining revealed that compared with the control group， the model group exhibited marked increases in corneal epithelial cell layers and epithelial thickness， along with reduced meibomian gland acini and intensely stained， densely packed nuclei around the acini. Compared with the model group， the Chufeng Yisuntang group， p38 MAPK inhibitor group， and combination group showed intact corneal structure， improved cell morphology， and reduced damage severity. ELISA revealed elevated ROS and MDA levels （P<0.01） and decreased SOD1 and SOD2 levels （P<0.01） in the model group compared with the control group. Compared with the model group， Chufeng Yisuntang， p38 MAPK inhibitor， and the combination lowered ROS and MDA levels （P<0.01）， while raising SOD1 and SOD2 levels （P<0.05， P<0.01）. Western blot revealed that compared with the control group， the model group exhibited increased protein levels of p38 MAPK， Bax， and Caspase-3 （P<0.01） and reduced protein level of Bcl-2 （P<0.01）. Compared with the model group， Chufeng Yisuntang， p38 MAPK inhibitor， and the combination down-regulated the protein levels of p38 MAPK， Bax， and Caspase-3 （P<0.01）， while up-regulating the protein level of Bcl-2 （P<0.01）. Compared with the Chufeng Yisuntang group， the combination group exhibited decreased protein levels of p38 MAPK， Bax， and Caspase-3 （P<0.01） and increased protein level of Bcl-2 （P<0.01）. Real-time PCR revealed that compared with the control group， the model group exhibited upregulated mRNA levels of p38 MAPK， Bax， and Caspase-3 （P<0.01）， and downregulated mRNA level of Bcl-2 （P<0.01）. Compared with the model group， Chufeng Yisuntang， p38 MAPK inhibitor， and the combination down-regulated the mRNA levels of p38 MAPK， Bax， and Caspase-3 （P<0.01）， while up-regulating the mRNA level of Bcl-2 （P<0.05， P<0.01）. Compared with the Chufeng Yisuntang group， the combination group exhibited decreased mRNA levels of p38 MAPK， Bax， and Caspase-3 expression （P<0.05， P<0.01） and increased mRNA level of Bcl-2 （P<0.01）. ConclusionChufeng Yisuntang may partially protect against PM_2.5-induced corneal injury by inhibiting the ROS/p38 MAPK pathway， enhancing antioxidant defense， and reducing epithelial apoptosis.

ObjectiveTo investigate the mechanisms of Runmu Dihuang decoction （RMDHD） in treating perimenopausal dry eye with liver-kidney Yin deficiency syndrome based on the silent information regulator 3 （SIRT3）/hypoxia-inducible factor-1α （HIF-1α）/nuclear factor-κB （NF-κB） signaling pathway. MethodsSixty female Sprague-Dawley rats were randomly divided into six groups （n=10 per group）： Sham operation group， model group， sodium hyaluronate eye drop group， and low-， medium-， and high-dose RMDHD groups （5.625， 11.25， 22.50 g·kg^-1）. Except for the sham operation group， all rats underwent bilateral ovariectomy and were administered 0.1% benzalkonium chloride eye drops combined with long-term chronic irritation to establish a perimenopausal dry eye model with liver-kidney Yin deficiency syndrome. Drug administration began in the 11th week after modeling and continued for 21 days. General conditions， screen-grip test scores， tear secretion volume， tear film breakup time （TFBUT）， and corneal fluorescein staining were recorded. Serum levels of reactive oxygen species （ROS）， follicle-stimulating hormone （FSH）， estradiol （E₂）， and progesterone （PROG） were measured by enzyme-linked immunosorbent assay （ELISA）. Pathological changes in the lacrimal glands， corneas， and uteri were observed using hematoxylin-eosin （HE） staining. Protein expression levels of SIRT3， HIF-1α， phosphorylated NF-κB p65 （p-NF-κB p65）， and total NF-κB p65 in the lacrimal glands were detected by Western blot. The expression of inflammatory cytokines interleukin-1β （IL-1β） and tumor necrosis factor-α （TNF-α） in the lacrimal glands was assessed by immunohistochemistry （IHC）. ResultsAfter model establishment， no significant differences were observed among the groups except the sham operation group. Compared with the sham operation group， the other groups exhibited slowed movement， dull responses， increased irritability， reduced body weight， elevated rectal temperature， decreased screen-grip test scores， reduced tear secretion， and significantly shortened TFBUT （P<0.05）. After treatment， compared with the model group， the sodium hyaluronate eye drop group and all RMDHD groups showed improved general conditions， significantly increased tear secretion （P<0.05）， prolonged TFBUT （P<0.05）， and elevated screen-grip test scores （P<0.05）. Serum ROS and FSH levels were significantly decreased， while E₂ and PROG levels were significantly increased （P<0.05）. Pathological damage to the cornea， lacrimal glands， and uterus was ameliorated. In addition， protein expression levels of SIRT3 and HIF-1α in the lacrimal glands were significantly upregulated （P<0.05）， whereas the expression of p-NF-κB p65， IL-1β， and TNF-α was significantly downregulated （P<0.05）. ConclusionRMDHD increases tear secretion and TFBUT， improves lacrimal gland and corneal injury， and alleviates dry eye symptoms in a perimenopausal dry eye rat model with liver-kidney Yin deficiency syndrome. The underlying mechanism may be related to regulation of the SIRT3/HIF-1α/NF-κB signaling pathway， inhibition of oxidative stress and inflammatory responses， and reduction of ocular surface tissue damage.

ObjectiveThis paper aims to investigate the effects and mechanism of Mimenghua prescription in modulating the mitogen-activated protein kinase kinase （MEK）/rat sarcoma viral oncogene homolog （Ras）/rapidly accelerated fibrosarcoma kinase （Raf）/extracellular signal-regulated kinase （ERK） signaling pathway to inhibit inflammatory responses in a dry eye animal model. MethodsA total of 60 C57BL/6J mice （eight weeks old， half male and half female） were used in the experiment. Ten mice were randomly selected as the blank control group， while the remaining 50 were exposed to a controlled dry system and received instillation of 0.2% benzalkonium chloride （BAC） into the eyes for four weeks to establish a dry eye mouse model. After successful modeling， the mice were randomly divided into five groups： Model group， sodium hyaluronate group， and Mimenghua prescription groups with low dose （4.83 g·kg^-1）， medium dose （9.67 g·kg^-1）， and high dose （19.34 g·kg^-1）. The mice in the model group received an equal volume of normal saline via gavage for four weeks. The mice in the sodium hyaluronate group received instillation of sodium hyaluronate eye drops twice daily for 14 consecutive days. The tear secretion volume， tear film break-up time （TBUT）， and corneal fluorescein staining were evaluated once every two weeks. After four weeks of administration， mice were euthanized， and their lacrimal gland tissues and corneas were harvested. Hematoxylin-eosin （HE） staining was used to assess histopathological morphology. Western blot was performed to detect the protein expression levels of MEK， Ras， Raf， and ERK. Enzyme-linked immunosorbent assay （ELISA） was used to measure the contents and expressions of MEK， Ras， Raf， ERK， and interleukin （IL）-1β in lacrimal gland and corneal tissues of the mice in each group. Quantitative real-time polymerase chain reaction （Real-time PCR） was employed to determine mRNA expression levels of MEK， Ras， Raf， and ERK. ResultsThe Mimenghua prescription groups and the sodium hyaluronate group exhibited significantly increased tear secretion volume （P<0.05） and prolonged TBUT （P<0.05） after treatment. Ocular surface damage of mice was visibly recovered. Western blot results indicated that protein expression levels of MEK， Ras， Raf， and ERK in the lacrimal gland and corneal tissues were significantly downregulated in the sodium hyaluronate group and Mimenghua prescription group with high dose （P<0.05）. ELISA results showed that IL-1β levels were highest in the model group but significantly reduced in the sodium hyaluronate group and Mimenghua prescription groups （P<0.05）. Both ELISA and Real-time PCR results demonstrated that the expression levels of MEK， Ras， Raf， and ERK in the lacrimal glands and corneal tissues were significantly elevated in the model group （P<0.05）， but markedly downregulated in the sodium hyaluronate group and Mimenghua prescription groups （P<0.05）， suggesting that Mimenghua prescription can decrease the expressions of MEK， Ras， Raf， and ERK in the lacrimal glands and corneal tissues. ConclusionMimenghua prescription can reduce inflammatory responses， increase tear secretion， prolong TBUT， and promote corneal recovery by inhibiting the MEK， Ras， Raf， and ERK signaling pathways in lacrimal gland and corneal tissues.

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.

ObjectiveTo investigate the risk factors for bleeding within 5 days and 2 weeks after endoscopic variceal ligation （EVL） or endoscopic cyanoacrylate injection （ECI） for the treatment of esophagogastric varices in patients with liver cirrhosis， as well as the safety of EVL/ECI in patients with thrombocytopenia. MethodsA total of 489 patients with liver cirrhosis and esophagogastric varices who underwent EVL/ECI in Guangzhou Eighth People’s Hospital， Guangzhou Medical University， from January 2018 to December 2023 were enrolled as subjects， and according to the presence or absence of bleeding after surgery， they were divided into bleeding group and non-bleeding group. The risk factors for bleeding within 5 days and 2 weeks after surgery were analyzed. The independent-samples t test or the Mann-Whitney U test was used for comparison of continuous data between groups， and the chi-square test or the continuity-corrected chi-square test was used for comparison of categorical data between groups； the receiver operating characteristic （ROC） curve was plotted to determine the cut-off value of MELD score； a multivariate logistic regression analysis was used to identify the independent risk factors for postoperative bleeding. ResultsThere were no significant differences in the bleeding rates within 5 days and 2 weeks after EVL/ECI between the 386 patients with a platelet count of ≥50×10⁹/L and the 103 patients with a platelet count of （25 — 49）×10⁹/L （5 days： 1.94% vs 2.85%， P=0.870； 2 weeks： 2.91% vs 4.92%， P=0.544）. The overall bleeding rate was 2.66% （13/489） and 4.50% （22/489）， respectively， within 5 days and 2 weeks after EVL/ECI. The multivariate logistic regression analysis showed that MELD score was an independent risk factor for bleeding within 5 days （odds ratio ［OR］=3.726， 95% confidence interval ［CI］： 1.214 — 11.429， P=0.021） and 2 weeks （OR=5.760， 95%CI： 1.779 — 18.651， P=0.003） after EVL/ECI， while hemoglobin （Hb） was a protective factor against bleeding within 5 days （OR=0.972， 95%CI： 0.948 — 0.996， P=0.025） and 2 weeks （OR=0.976， 95%CI： 0.957 — 0.995， P=0.016） after surgery； portal vein tumor thrombus （OR=2.667， 95%CI： 1.000 — 7.117， P=0.050） was an independent risk factor for bleeding within 2 weeks after surgery， while platelet count ［（25 — 49）×10⁹/L］ was not a risk factor for postoperative bleeding （P>0.05）. ConclusionBoth EVL and ECI have good safety in patients with liver diseases and grade 3 thrombocytopenia. MELD score is an independent risk factor for bleeding within 5 days and 2 weeks after EVL/ECI， while Hb is a protective factor； portal vein tumor thrombus is an independent risk factor for bleeding within 2 weeks after surgery.

ObjectiveThis paper aims to investigate the effect and mechanism of Runmu Xiaoyao powder on mice with dry eye and the syndrome of liver depression-heat syndrome based on the toll-like receptor 4 （TLR4）/myeloid differentiation primary response protein 88 （MyD88）/nuclear factor-κB （NF-κB） signaling pathway. MethodsSixty-six C57BL/6J female mice were randomly divided into a normal group， a model group， a sodium hyaluronate group， and high-， medium-， and low-dose Runmu Xiaoyao powder groups， with 11 mice per group. Except for those in the normal group， mice in the other groups were subjected to mouse models with dry eye and liver depression-heat syndrome by instilling a benzalkonium chloride solution and applying chronic pain stimulation in a dry environment. After modeling， mice in the high-， medium-， and low-dose Runmu Xiaoyao powder groups were administered intragastrically with 29.7， 14.85， 7.43 g·kg^-1， respectively， twice a day for 14 consecutive days. The mice in the sodium hyaluronate group received 5 μL of sodium hyaluronate eye drops in each eye twice daily. The mice in the normal and model groups were administered intragastrically with an equal volume of deionized water. Measurements were taken of tear secretion in mice， irritability scores， corneal fluorescein staining， and histopathological changes in the cornea， lacrimal glands， meibomian glands， and liver tissue. Enzyme-linked immunosorbent assay （ELISA） was used to detect the levels of interleukin-1β （IL-1β） and tumour necrosis factor-α （TNF-α） in serum. Immunohistochemistry （IHC） was used to detect the protein expression of IL-1β and TNF-α in corneal and lacrimal gland tissues. Real-time quantitative polymerase chain reaction（Real-time PCR） and Western blot were employed to detect the mRNA and protein expressions of TLR4， MyD88， and NF-κB p65 in corneal， lacrimal gland， and meibomian gland tissues. ResultsCompared with those in the normal group， mice in the model group exhibited significantly reduced tear secretion， significantly higher irritability scores， and more pronounced corneal fluorescence staining， with marked pathological damage observed in the cornea， lacrimal glands， meibomian glands， and liver tissue. IL-1β and TNF-α levels in serum were significantly elevated. The protein expressions of IL-1β and TNF-α in corneal and lacrimal gland tissues， as well as the mRNA and protein expressions of TLR4， MyD88， and NF-κB p65 in corneal， lacrimal gland， and meibomian gland tissues， were all significantly increased （P<0.01）. Compared with the model group， the sodium hyaluronate group and Runmu Xiaoyao powder groups with different doses exhibited increased tear secretion to varying degrees， alleviated corneal fluorescence staining and histopathological damage， reduced IL-1β and TNF-α levels in serum， and downregulated protein expressions of IL-1β and TNF-α in corneal and lacrimal gland tissues， as well as the mRNA and protein expressions of TLR4， MyD88， and NF-κB p65 in corneal， lacrimal gland， and meibomian gland tissues. The high-dose Runmu Xiaoyao powder group demonstrated a more pronounced effect， with multiple indicators showing superior results compared to those in the sodium hyaluronate group （P<0.05， P<0.01）. ConclusionRunmu Xiaoyao powder down-regulates the TLR4/MyD88/NF-κB signaling pathway activity in the cornea， lacrimal glands， and meibomian glands of model mice with dry eye and liver depression-heat syndrome， thereby suppressing inflammatory responses and mitigating ocular surface tissue damage. The therapeutic effect is dose-dependent， and the high-dose group exerts the most prominent effect.

The philosophy of “treating disease before its onset” is a fundamental concept of traditional Chinese medicine (TCM), permeating its diagnostic and therapeutic framework, and is central to clinical practice. However, current TCM diagnostic and treatment models for the “pre-disease to disease” window period face several limitations, including the lack of comprehensive clinical parameters, difficulties in characterizing and integrating heterogeneous multimodal data, and insufficient dynamic precision in interventions and efficacy evaluations. To address these issues, guided by Professor Candong Li’s theory of TCM stateology, this study focuses on integrating objective multimodal data. It proposes a new model for personalized TCM diagnosis and treatment targeting the “pre-disease to disease” window period. This approach first proposes the idea of restructuring the conceptual framework of “symptom” and integrating multi-source heterogeneous data at macroscopic, mesoscopic, and microscopic levels to form a three-dimensional assessment indicator system. By integrating graph neural networks, convolutional neural networks, attention mechanisms, and knowledge graph-guided weight allocation, this approach enables collaborative representation, alignment, and fusion of multi-source data. Subsequently, it plans to construct a multimodal fusion model at both feature and decision levels, in order to establish mappings between indicators and TCM state elements, and to screen key indicators characterizing pathological evolution during the window period. Furthermore, it proposes a technical path for enhancing model interpretability using methods such as SHapley Additive exPlanations (SHAP) and Ablation-CAM++. Finally, with state assessment as the core, it proposes the concept of constructing a dynamic evaluation method for individualized diagnosis and treatment based on time-series data analysis using algorithms such as long short-term memory (LSTM) networks and gated recurrent units (GRUs). Moreover, a causal inference framework and semi-supervised learning strategies are introduced to enable quantitative evaluation of individual intervention effects and to provide interpretable therapeutic feedback, forming a complete technical path from data representation and fusion, weight adjustment, and interpretability analysis, to dynamic diagnosis feedback. This study aims to address deficiencies in the current TCM diagnosis and treatment model during the “pre-disease to disease” window period and to provide an operational framework for the clinical practice of TCM’s “treating disease before its onset”.