ObjectiveTo investigate the effect and mechanism of transplantation of human umbilical cord mesenchymal stem cell （hUC-MSC） with overexpression of the Numb gene in the treatment of cholestatic liver fibrosis （CLF）. MethodsThe technique of lentiviral transfection was used to induce the overexpression of the Numb gene in hUC-MSC （hUC-MSC^Numb-OE）， and hUC-MSC transfected with empty vector （hUC-MSC^OE-EV） was used as negative control. Bile duct ligation （BDL） was performed to establish a rat model of CLF， and then the rats were randomly divided into BDL group， hUC-MSC group， hUC-MSC^OE-EV group， and hUC-MSC^Numb-OE group， while a sham-operation group was also established. The rats in the intervention groups were given a single splenic injection of the corresponding cells after BDL， and samples were collected at the end of week 4. Related indicators were measured， including serum biochemistry， liver histopathology， the content of hydroxyproline （Hyp） in the liver， hepatic stellate cell activation， ductular reaction， liver regeneration， and the expression levels of key molecules in the Numb-p53 signaling axis. A one-way analysis of variance was used for comparison of continuous data between multiple groups， and the least significant difference t-test was used for further comparison between two groups. ResultsCompared with the BDL group， the hUC-MSC group and the hUC-MSC^OE-EV group had significant reductions in the levels of serum biochemical parameters （aspartate aminotransferase， gamma-glutamyl transpeptidase， total bile acid， total bilirubin， and direct bilirubin）， liver fibrosis markers （the content of Hyp and the expression levels of alpha-smooth muscle actin， tumor necrosis factor-α， and transforming growth factor-beta 1）， and ductular reaction markers （the expression levels of CK7 and CK19） （all P <0.05）， and compared with the hUC-MSC^OE-EV group， the hUC-MSC^Numb-OE group had significantly greater improvements in the above indicators （all P <0.05）. In addition， compared with the hUC-MSC^OE-EV group， the hUC-MSC^Numb-OE group had significant improvements in the expression levels of liver regeneration-related markers （albumin and hepatocyte nuclear factor 4α） and the molecules associated with the Numb-p53 signaling axis （Numb， pNumb， Mdm2， and p53） （all P <0.05）. ConclusionOverexpression of the Numb gene can enhance the therapeutic effect of hUC-MSC on CLF， possibly by activating the Numb-PTB^L-p53-HNF4α axis， promoting the hepatic differentiation of hUC-MSCs and subsequently enhancing liver regeneration.

Tumors are the second leading cause of death worldwide. Exosomes are a type of extracellular vesicle secreted from multivesicular bodies, with particle sizes ranging from 40 to 160 nm. They regulate the tumor microenvironment, proliferation, and progression by transporting proteins, nucleic acids, and other biomolecules. Compared with other drug delivery systems, exosomes derived from different cells possess unique cellular tropism, enabling them to selectively target specific tissues and organs. This homing ability allows them to cross biological barriers that are otherwise difficult for conventional drug delivery systems to penetrate. Due to their biocompatibility and unique biological properties, exosomes can serve as drug delivery systems capable of loading various anti-tumor drugs. They can traverse biological barriers, evade immune responses, and specifically target tumor tissues, making them ideal carriers for anti-tumor therapeutics. This article systematically summarizes the methods for exosome isolation, including ultracentrifugation, ultrafiltration, size-exclusion chromatography (SEC), immunoaffinity capture, and microfluidics. However, these methods have certain limitations. A combination of multiple isolation techniques can improve isolation efficiency. For instance, combining ultrafiltration with SEC can achieve both high purity and high yield while reducing processing time. Exosome drug loading methods can be classified into post-loading and pre-loading approaches. Pre-loading is further categorized into active and passive loading. Active loading methods, including electroporation, sonication, extrusion, and freeze-thaw cycles, involve physical or chemical disruption of the exosome membrane to facilitate drug encapsulation. Passive loading relies on drug concentration gradients or hydrophobic interactions between drugs and exosomes for encapsulation. Pre-loading strategies also include genetic engineering and co-incubation methods. Additionally, we review approaches to enhance the targeting, retention, and permeability of exosomes. Genetic engineering and chemical modifications can improve their tumor-targeting capabilities. Magnetic fields can also be employed to promote the accumulation of exosomes at tumor sites. Retention time can be prolonged by inhibiting monocyte-mediated clearance or by combining exosomes with hydrogels. Engineered exosomes can also reshape the tumor microenvironment to enhance permeability. This review further discusses the current applications of exosomes in delivering various anti-tumor drugs. Specifically, exosomes can encapsulate chemotherapeutic agents such as paclitaxel to reduce side effects and increase drug concentration within tumor tissues. For instance, exosomes loaded with doxorubicin can mitigate cardiotoxicity and minimize adverse effects on healthy tissues. Furthermore, exosomes can encapsulate proteins to enhance protein stability and bioavailability or carry immunogenic cell death inducers for tumor vaccines. In addition to these applications, exosomes can deliver nucleic acids such as siRNA and miRNA to regulate gene expression, inhibit tumor proliferation, and suppress invasion. Beyond their therapeutic applications, exosomes also serve as tumor biomarkers for early cancer diagnosis. The detection of exosomal miRNA can improve the sensitivity and specificity of diagnosing prostate and pancreatic cancers. Despite their promising potential as drug delivery systems, challenges remain in the standardization and large-scale production of exosomes. This article explores the future development of engineered exosomes for targeted tumor therapy. Plant-derived exosomes hold potential due to their superior biocompatibility, lower toxicity, and abundant availability. Furthermore, the integration of exosomes with artificial intelligence may offer novel applications in diagnostics, therapeutics, and personalized medicine.

BACKGROUND: Based on the China-VHD database, this study sought to develop and validate a Valvular Heart Disease- specific Age-adjusted Comorbidity Index (VHD-ACI) for predicting mortality risk in patients with VHD. METHODS & RESULTS: The China-VHD study was a nationwide, multi-centre multi-centre cohort study enrolling 13,917 patients with moderate or severe VHD across 46 medical centres in China between April-June 2018. After excluding cases with missing key variables, 11,459 patients were retained for final analysis. The primary endpoint was 2-year all-cause mortality, with 941 deaths (10.0%) observed during follow-up. The VHD-ACI was derived after identifying 13 independent mortality predictors: cardiomyopathy, myocardial infarction, chronic obstructive pulmonary disease, pulmonary artery hypertension, low body weight, anaemia, hypoalbuminaemia, renal insufficiency, moderate/severe hepatic dysfunction, heart failure, cancer, NYHA functional class and age. The index exhibited good discrimination (AUC, 0.79) and calibration (Brier score, 0.062) in the total cohort, outperforming both EuroSCORE II and ACCI (P < 0.001 for comparison). Internal validation through 100 bootstrap iterations yielded a C statistic of 0.694 (95% CI: 0.665-0.723) for 2-year mortality prediction. VHD-ACI scores, as a continuous variable (VHD-ACI score: adjusted HR (95% CI): 1.263 (1.245-1.282), P < 0.001) or categorized using thresholds determined by the Yoden index (VHD-ACI ≥ 9 vs. < 9, adjusted HR (95% CI): 6.216 (5.378-7.184), P < 0.001), were independently associated with mortality. The prognostic performance remained consistent across all VHD subtypes (aortic stenosis, aortic regurgitation, mitral stenosis, mitral regurgitation, tricuspid valve disease, mixed aortic/mitral valve disease and multiple VHD), and clinical subgroups stratified by therapeutic strategy, LVEF status (preserved vs. reduced), disease severity and etiology. CONCLUSION The VHD-ACI is a simple 13-comorbidity algorithm for the prediction of mortality in VHD patients and providing a simple and rapid tool for risk stratification.

Objective To explore the effect of gait biomechanics on sports injuries in middle-and long-distance teenage runners.Methods Eighteen healthy middle-and long-distance runners were select-ed as subjects,and the biomechanics of their gait during mid-speed running was tested using a three-dimensional infrared motion capture system,a three-dimensional force measurement platform,and a la-ser speed measurement system.After the test,data on sports injuries within one year were collected.Based on the severity of their injuries,the athletes were divided into an injury group,and a non-inju-ry group(healthy group).Their gait spatiotemporal parameters,ground reaction forces,and independent sample t-test for trunk and lower limb joint angles,and lower limb joint power were all analyzed.Re-sults There was no significant difference in the spatiotemporal parameters of the mid-speed running gait and ground reaction force indicators between the two groups(P>0.05).The maximum trunk exten-sion angle of the injury group was significantly greater than the healthy group(P<0.05),while the max-imum trunk flexion angle was significantly smaller than the latter(P<0.05).Moreover,the maximum right flexion angle,hip adduction angle,maximum hip adduction angle,ankle varus angle and ankle internal rotation angle of the injury group were significantly smaller than that of the healthy group,with the ankle internal rotation angle significantly bigger than the latter(P<0.05 for all).Meanwhile,the maximum hip joint extension power of the injury group was significantly lower than the healthy group(P<0.05),while the maximum knee joint abduction power was significantly higher than the latter group(P<0.01).Conclusion When running at medium speed,the spatiotemporal parameters of the sup-porting leg and ground reaction force have a relatively smaller impact on lower limb sports injuries of adolescent middle-and long-distance runners,while a greater degree of trunk tilt back,a smaller right flexion angle,a smaller hip joint adduction of the supporting leg,a smaller internal rotation an-gle when the ankle joint is on the ground,and a higher internal rotation angle when the ankle joint is off the ground make the lower limbs more prone to sports injuries.Moreover,such runners with low-er hip joint extension power and higher knee joint abduction power in their support legs are more like-ly to suffer from lower limb injuries.

Objective To investigate the association between four single nucleotide polymorphisms(SNP)(rs9340 in MAPK1,rs14804 in NRAS,rs712 and rs7973450 in KRAS)in the 3'UTR of ERK1/2 signaling pathway-related genes and non-small cell lung cancer(NSCLC).Methods A total of 478 NSCLC patients and 480 healthy controls were enrolled in this study.Four SNPs were genotyped by using TaqMan assays.The association between the four SNPs and NSCLC was analyzed.Results The distribution frequency difference of the allele of rs9340 was statistically significant between the control group and the non-small cell squamous cell carcinoma(SCC)group(P = 0.009),suggesting that the G allele of rs9340 may be a protective factor for non-small cell lung squamous cell carcinoma(OR = 0.67,95%CI:0.50～0.91).In addition,in the<50 years age group,the distribution frequency difference of the allele of rs9340 was statistically significant between the control group and the NSCLC group(P = 5.07×10-4),indicating that the G allele of rs9340 may be a protective factor for NSCLC(OR = 0.46,95%CI:0.29～0.72).Conclusion The SNP rs9340 in MAPK1 may be associated with the risk of NSCLC.

BACKGROUND:Nowadays,there are increasing reports on the digitization and visualization system of acupuncture points for adults in traditional Chinese medicine,and the digitization and visualization system of acupuncture points for children in traditional Chinese medicine and the simulation system of acupuncture manipulation for Mongolian medicine training have been reported.However,there are no reports on relevant systems for children in Mongolian medicine. OBJECTIVE:To develop a simulation system of acupuncture points for children in Mongolian medicine,in the hope that it can be used for clinical teaching,manipulation practice and research on acupuncture safety. METHODS:Based on the tomographic anatomical dataset of preschool boys,a three-dimensional(3D)digital virtual anatomical model of children with multiple internal organs and tissues was constructed by using PhotoShop.2021 and Digihuman Reconstruction System software.The relevant annotation information database of 27 acupoints such as Dinghui acupoint of Mongolian medicine was compiled by the Unity database language.The Mongolian gold needle and silver needle were selected to record the acupuncture point teaching video on the 3D printed head and neck resin model of children.In Unity3D software,children's anatomical model,acupoint annotation information database and acupuncture operation video were integrated and coded,and a 3D digital children's Mongolian acupuncture acupoint visualization system integrating simulation acupuncture training,clinical teaching and acupuncture safety research was successfully created. RESULTS AND CONCLUSION:This study was based on real children's specimens.In order to reduce the error of two-dimensional segmentation,the manual layer-by-layer segmentation section image method was used to ensure the accuracy of the 3D model to the greatest extent.The Digihuman Reconstruction System was used to extract and save the individual segmentation data while maximizing the accuracy of the 3D model.PhotoShop.2021 software was used to complete the 3D reconstruction of the outer skin of the head and neck of children and the internal bony structure,cervical spinal cord,blood vessels and nerves,muscles and ligaments.After 3D reconstruction,the basic morphology of each independent structure and the integrity of the overall contour were verified in MeshLab software and the final fine adjustment and anatomical position confirmation were conducted using 3-matic research 13.0 software.The real anatomical morphology of the head and neck of preschool children was successfully simulated and restored.Unity3D software was used to integrate the 3D model of children,acupuncture operation video and acupoint annotation database,and the 3D digital Mongolian acupuncture acupoint visualization system for children was successfully constructed.Based on the real continuous fault ultra-thin dataset of preschool boys in China,China's first 3D digitization and visualization system of acupuncture points in the head and neck of children in Mongolian medicine was developed.Compared with the previous acupuncture soft works,this system is more suitable for the anatomical morphological development characteristics of Asian children,and has a high value of application in the fields of research on the safety of acupuncture in Mongolian medicine,clinical teaching and acupuncture simulation training.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.