Intravitreal anti-vascular endothelial growth factor(anti-VEGF)therapy has been widely used, but the variability in its therapeutic efficacy limits individualized treatment. In recent years, the application of artificial intelligence(AI)has opened up new avenues for personalized treatment response prediction, and its core branches include machine learning(ML)and deep learning(DL). This review systematically retrieved and analyzed 41 relevant studies published up to April 2025. Comprehensive analysis reveals that AI predictive models are evolving from forecasting single endpoints(such as visual acuity or central retinal thickness)to integrating multi-dimensional endpoints(encompassing anatomical, functional, and treatment demand parameters)and generating predictive imaging outputs. In terms of technical approaches, DL models(28 studies, accounting for 68.3%)dominate this field due to their robust image interpretation capabilities, while ML models(10 studies, 24.4%)retain significant value in the analysis of structured clinical data. Cross-disease comparisons indicate that research efforts are most concentrated on age-related macular degeneration(ARMD)and diabetic macular edema(DME), with shared conceptual frameworks for model construction, yet distinct anatomical and functional indicators are prioritized for each disease. Currently, the field confronts several key challenges, including insufficient prospective clinical validation, limited model interpretability(the “black box problem”), and a scarcity of high-quality multi-center datasets. Moving forward, it is imperative to advance real-world validation and develop explainable AI techniques to expedite the clinical translation of these predictive models.

Objective:To develop and validate clinical predictive models for identifying poor short-term response to recombinant human growth hormone(rhGH) treatment in children with short stature.Methods:A retrospective analysis was conducted on 118 children diagnosed with growth hormone deficiency or idiopathic short stature who were treated at the First Affiliated Hospital of Zhengzhou University and two other hospitals between January 1, 2020, and January 1, 2024. A poor response to rhGH was defined as a height increase of less than 0.2 standard deviation score(SDS) after 6 months of rhGH treatment. LASSO regression was used to identify predictive variables from baseline and follow-up data. Two logistic regression models were conducted: Model A(incorporating baseline variables only) and model B(incorporating both baseline and follow-up variables), and nomograms were created for visualization. External data and internal resampling were used for dual validation of the models, and their performance was compared.Results:A total of 118 children with short stature were included. Six baseline predictive variables(diagnosis, initial height SDS, bone age, bone age-chronological age difference, rhGH dose, and gender) and one follow-up variable(height SDS after 3 months of rhGH treatment) were identified. Area under the curve values for Model A and Model B were 0.753(95% CI 0.696-0.811) and 0.930(95% CI 0.891-0.975), respectively. Calibration curves, decision curve analysis, and other evaluation metrics demonstrated good discrimination and clinical utility for both models. Model B, incorporating the 3-month follow-up variable, showed superior predictive performance compared to Model A. Conclusions:The clinical prediction models developed in this study(Model A and Model B) are practical and reliable tools for quantitatively, conveniently, and intuitively identifying children with short stature at risk of poor response to rhGH treatment.

Objective To preliminarily explore the potential efficacy of Xuesaitong Soft Capsule(XST)against post-stroke depression(PSD),and to investigate the material basis of XST's anti-PSD effect based on the metabolomics results to analyze its related pharmacokinetic(PK)characteristics and further analyze the pharmacodynamic(PD)equation of representative ingredients.Methods The initial evaluation of drug effica-cy was conducted by detecting the depressive-like behavior and neurotransmitter levels in rats.The Pearson correlation analysis was employed to analyze the correlation between the main metabolites regulated by XST and the saponin components entering the bloodstream.At various time points after drug administration,the blood concentration of ginsenoside Re and the concentration of norepinephrine(NE)in the serum of PSD rats were measured,and the compartment model was fitted accordingly.Furthermore,the liquid chromatography-mass spectrometry was utilized to determine the content of ginsenoside Re in the liver,spleen,kidney,prefron-tal cortex,hippocampus and striatum of PSD rats.Results Ginsenoside Re showed the optimal correlation by the Pearson correlation analysis.Based on its pharmacokinetic parameters,the pharmacodynamic equation with NE was E=160.462 × Ce/(38.663+Ce).The contents of ginsenoside Re in the liver,spleen,kidney,prefron-tal cortex,hippocampus and striatum of rats were(17.23+11.90),(19.05+5.67),(1.95+0.79),(70.13+6.75),(57.03+3.11),and(72.45+5.45)ng/g,respectively.Conclusion XST could improve the depressive-like behaviors in PSD rats by regulating the expression levels of neurotransmitter NE and 5-HT.Ginsenoside Re may be the pharmacodynamical material foundation for XST's preventative treatment of PSD.

Objective To construct a risk prediction model for post-stroke dysphagia(PSD)based on clinical and laboratory indicators of elderly stroke patients with explainable machine learning.Methods A retrospective analysis was conducted on 3994 stroke patients hospitalized in Depart-ment of Neurology of First Medical Center of Chinese PLA General Hospital from October 2010 to December 2021.Among them,the 1390 cases admitted during January 2019 and December 2021 were assigned into an external validation set,and the 2604 cases admitted during October 2010 to January 2019 were into a training group.Those from the training group were further divided into a training set(1823 cases)and an internal validation set(781 cases)in a 7∶3 ratio,and also grouped into a PSD subgroup(773 cases)and a non-PSD group(1831 cases).With occurrence of swallowing difficulties as an endpoint,risk prediction models were constructed using random for-est(RF),eXtreme Gradient Boosting(XGBoost),Support Vector Machine(SVM),and logistic regression.ROC curve analysis was employed to evaluate the performance of our models.After the optimal model was selected,SHAP was employed to interpret feature contributions.Results There were significant differences in muscle strength,right/left-sided stroke,and area of brain in-jury between the PSD and the non-PSD groups(P＜0.01).The PSD group had obviously larger proportions of hypertension,diabetes,and drinking history,increased neutrophil counts,and de-creased levels of potassium and albumin when compared with the non-PSD group(P＜0.05,P＜0.01).Multivariate logistic regression analysis showed that age,drinking history,diabetes,hyper-tension,muscle strength grade,area of brain injury,hemispheric stroke,neutrophil count,and al-bumin and potassium levels were risk factors for PSD(P＜0.05,P＜0.01).The external validation results showed that the area under curve value of the RF model,XGBoost model,SVM model,and our logistic model was 0.883,0.902,0.877,and 0.868,respectively.The distribution of SHAP value showed that drinking history,hypertension and diabetes were positively correlated with PSD risk;Muscle strength was negatively correlated with the risk;Age growth was positively correlated with the risk;Subtentorial lesions showed stronger predictive efficacy than supratentorial lesions and entire lesions;The bilateral and right-sided stroke had higher risk for PSD than the left-sided stroke.Conclusion The model based on the XGBoost model shows best performance in predicting the risk for swallowing disorders in elderly patients after stroke.

Objective To construct a risk prediction model for post-stroke dysphagia(PSD)based on clinical and laboratory indicators of elderly stroke patients with explainable machine learning.Methods A retrospective analysis was conducted on 3994 stroke patients hospitalized in Depart-ment of Neurology of First Medical Center of Chinese PLA General Hospital from October 2010 to December 2021.Among them,the 1390 cases admitted during January 2019 and December 2021 were assigned into an external validation set,and the 2604 cases admitted during October 2010 to January 2019 were into a training group.Those from the training group were further divided into a training set(1823 cases)and an internal validation set(781 cases)in a 7∶3 ratio,and also grouped into a PSD subgroup(773 cases)and a non-PSD group(1831 cases).With occurrence of swallowing difficulties as an endpoint,risk prediction models were constructed using random for-est(RF),eXtreme Gradient Boosting(XGBoost),Support Vector Machine(SVM),and logistic regression.ROC curve analysis was employed to evaluate the performance of our models.After the optimal model was selected,SHAP was employed to interpret feature contributions.Results There were significant differences in muscle strength,right/left-sided stroke,and area of brain in-jury between the PSD and the non-PSD groups(P＜0.01).The PSD group had obviously larger proportions of hypertension,diabetes,and drinking history,increased neutrophil counts,and de-creased levels of potassium and albumin when compared with the non-PSD group(P＜0.05,P＜0.01).Multivariate logistic regression analysis showed that age,drinking history,diabetes,hyper-tension,muscle strength grade,area of brain injury,hemispheric stroke,neutrophil count,and al-bumin and potassium levels were risk factors for PSD(P＜0.05,P＜0.01).The external validation results showed that the area under curve value of the RF model,XGBoost model,SVM model,and our logistic model was 0.883,0.902,0.877,and 0.868,respectively.The distribution of SHAP value showed that drinking history,hypertension and diabetes were positively correlated with PSD risk;Muscle strength was negatively correlated with the risk;Age growth was positively correlated with the risk;Subtentorial lesions showed stronger predictive efficacy than supratentorial lesions and entire lesions;The bilateral and right-sided stroke had higher risk for PSD than the left-sided stroke.Conclusion The model based on the XGBoost model shows best performance in predicting the risk for swallowing disorders in elderly patients after stroke.

Objective:To develop and validate clinical predictive models for identifying poor short-term response to recombinant human growth hormone(rhGH) treatment in children with short stature.Methods:A retrospective analysis was conducted on 118 children diagnosed with growth hormone deficiency or idiopathic short stature who were treated at the First Affiliated Hospital of Zhengzhou University and two other hospitals between January 1, 2020, and January 1, 2024. A poor response to rhGH was defined as a height increase of less than 0.2 standard deviation score(SDS) after 6 months of rhGH treatment. LASSO regression was used to identify predictive variables from baseline and follow-up data. Two logistic regression models were conducted: Model A(incorporating baseline variables only) and model B(incorporating both baseline and follow-up variables), and nomograms were created for visualization. External data and internal resampling were used for dual validation of the models, and their performance was compared.Results:A total of 118 children with short stature were included. Six baseline predictive variables(diagnosis, initial height SDS, bone age, bone age-chronological age difference, rhGH dose, and gender) and one follow-up variable(height SDS after 3 months of rhGH treatment) were identified. Area under the curve values for Model A and Model B were 0.753(95% CI 0.696-0.811) and 0.930(95% CI 0.891-0.975), respectively. Calibration curves, decision curve analysis, and other evaluation metrics demonstrated good discrimination and clinical utility for both models. Model B, incorporating the 3-month follow-up variable, showed superior predictive performance compared to Model A. Conclusions:The clinical prediction models developed in this study(Model A and Model B) are practical and reliable tools for quantitatively, conveniently, and intuitively identifying children with short stature at risk of poor response to rhGH treatment.

Objective:To study the surveillance results of soil-transmitted nematode (STN) disease at the national surveillance site in Nanchang City, so as to scientifically formulate prevention and control strategies for parasitic diseases in Nanchang City.Methods:From 2016 to 2020, in Nanchang County, the national surveillance site of STN disease in Nanchang City, no less than 1 000 fecal samples and 25 soil samples of permanent residents over 3 years of age were investigated annually. The improved Kato-Katz thick smear method was used to detect the eggs of STN in the population, species identification of hookworm was carried out by test-tube filter paper incubation, the Enterobius vermicularis eggs of 3 - 9 years old children were detected by cellophane anal swab method, and the soil samples were used to identify hookworm larvae and of Ascaris lumbricoides eggs. Results:The total infection rate of STN was 0.93% (47/5 054) from 2016 to 2020 in the national surveillance site of Nanchang City. The annual infection rates were 0.10% (1/1 000), 2.94% (30/1 021), 0.79% (8/1 008), 0.50% (5/1 002) and 0.29% (3/1 023), respectively; the infection rates of hookworm, Ascaris lumbricoides and Trichuris trichiura were 0.42% (21/5 054), 0.02% (1/5 054) and 0.49% (25/5 054), respectively. Totally 97.87% (46/47) of the infected persons were mild infection. The population distribution characteristics showed that there was significant difference in STN infection rate among different education levels (χ 2 = 18.06, P = 0.001), but there was no significant difference in infection rate among different sex, age and occupation (χ 2 = 0.92, 2.01, 13.60, P > 0.05). Among them, the infection rate of junior high school cultural was the highest (1.84%, 24/1 301). The culture of hookworm larvae showed that 90.48% (19/21) were Necator americanus; the infection rate of Enterobius vermicularis in 3 - 9 years old children, the contamination rate of soil hookworm larvae and Ascaris lumbricoides eggs, were all 0. Conclusion:STN disease shows a low prevalence trend at the national monitoring site in Nanchang City, Trichuris trichiura and hookworm are the key insect species for STN disease control in Nanchang City in the future.

Objective:To explore the effect of pulmonary rehabilitation program based on Delphi method in patients with overlap syndrome (OS) .Methods:From March 2019 to March 2021, a total of 107 OS patients admitted to Qinghai Provincial People's Hospital were enrolled using the convenience sampling method and divided into the observation group ( n=54) and the control group ( n=53) by the random number table method. The control group received conventional nursing, and the observation group conducted pulmonary rehabilitation program based on Delphi method on the basis of conventional nursing. The COPD Assessment Test (CAT) , the 6-Minute Walk Test (6MWT) , and the St George's Respiratory Questionnaire (SGRQ) were used to assess the improvement of the conditions, exercise endurance, and quality of life of the two groups before and after intervention. Results:After intervention, the CAT score of the observation group was lower than that of the control group, 6MWD was longer than that of the control group, and the clinical symptoms, disease impact, and activity scores of SGRQ were lower than those of the control group, and the differences were statistically significant ( P<0.05) . Conclusions:The pulmonary rehabilitation program based on the Delphi method can reduce patients' condition, improve exercise endurance and quality of life.

Background: Insulin regulates glucose homeostasis and has important effects on metabolism, cell growth, and differentiation. Depending on the cell type and physiological context, insulin signal has specific pathways and biological outcomes in different tissues and cells. For studying the signal pathway of insulin on glycolipid metabolism in porcine embryonic fibroblast (PEF), we used high-throughput sequencing to monitor gene expression patterns regulated by insulin. Objectives: The goal of our research was to see how insulin affected glucose and lipid metabolism in PEFs. Methods: We cultured the PEFs with the addition of insulin and sampled them at 0, 48, and 72 h for RNA-Seq analysis in triplicate for each time point. Results: At 48 and 72 h, 801 and 1,176 genes were differentially expressed, respectively. Of these, 272 up-regulated genes and 264 down-regulated genes were common to both time points. Gene Ontology analysis was used to annotate the functions of the differentially expressed genes (DEGs), the biological processes related to lipid metabolism and cell cycle were dominant. And the DEGs were significantly enriched in interleukin-17 signaling pathway, phosphatidylinositol-3-kinase-protein kinase B signaling pathway, pyruvate metabolism, and others pathways related to lipid metabolism by Kyoto Encyclopedia of Genes and Genomes enrichment analysis. Conclusions These results elucidate the transcriptomic response to insulin in PEF. The genes and pathways involved in the transcriptome mechanisms provide useful information for further research into the complicated molecular processes of insulin in PEF.

Objective Summarize and share the working practice of Investigator Research Ethics and Research Integrity Training conducted at Peking University Human Research Protection Program (PKU HRPP) ,to further explore continuing quality improvement of investigator ethical training at university level .Methods Conduct systematic review of the archiving files of PKU HRPP investigator training activities during 2012-2018 ,summarize feedback information from investigators to i-dentify possible experiences for sharing and space for improvement .Results There are some positive experiences for sharing a-bout the training mechanism and practices at PKU HRPP .Conclusions Based on the previous ethical training work and experi-ences at PKU HRPP ,possible proposals for continuing quality improvement may including :strengthening the requirements of ethical training of investigators ,encouraging and recognizing ethical training conducted by research teams ,emphasizing training Quality and the promotion of sharing and mutual recognition mechanisms for ethical training .