Lipid nanoparticles(LNPs),composed of ionizable lipids,are currently the most promis-ing non-viral nucleic acid drug delivery vectors in clinical practice,and have great potential in gene ther-apy drug delivery and vaccine delivery.Ionizable lipids,the main components of LNPs,play a decisive role in the endosome escape rate,transfection efficiency,organ targeting and safety of LNPs.Among them,the hydrophilic head group of ionizable lipids contains tertiary amine groups,which can improve the buffering capacity of LNPs and thus change the pKa value,and imidazole can enhance the stability and transfection activity of mRNA-LNP.The ligand contains ester groups,which can induce gene silencing efficiently and improve the degradation rate and safety.When the number of hydrophobic tails is 3-4,with 1-2 unsaturation and 8-18 carbon chain length,LNPs can effectively induce gene silencing.Meanwhile,the presence of branching or asymmetric hydrophobic tails can improve the transfection efficiency of LNPs.Based on the chemical structure of ionizable lipids,this review summarizes the influ-ence of the structure of ionizable lipids on the transfection efficiency and safety of nucleic acid carrying drugs LNPs,and the structure-activity relationship of ionizable lipids so as to provide reference for studies on novel ionizable lipids.

Objective:To investigate the interactive effect of mismatch repair (MMR) protein status and adverse clinicopathological features on prognosis of stage Ⅰ-Ⅲ colon cancer.Methods:The retrospective cohort study was conducted. The clinicopathological data of 1 650 patients with colon cancer of stage Ⅰ-Ⅲ who were admitted to 7 hospitals in China from January 2016 to December 2017 were collected. There were 963 males and 687 females, aged 62(53,71)years. Patients were classified as 230 cases of MMR deficiency (dMMR) and 1 420 cases of MMR proficiency (pMMR) based on their MMR protein status. Observation indicators: (1) comparison of clinicopathological characteristics between patients of different MMR protein status; (2) analysis of factors affecting the survival outcomes of patients of dMMR; (3) analysis of factors affecting the survival outcomes of patients of pMMR; (4) interaction analysis of MMR and adverse clinicopathological features on survival outcomes. Measurement data with normal distribution were represented as Mean± SD, and comparison between groups was conducted using the independent t test. Measurement data with skewed distribution were represented as M( Q1, Q3), and comparison between groups was conducted using the Mann-Whitney U test. Count data were described as absolute numbers, and comparison between groups was conducted using the chi-square test or Fisher exact probability. Comparison of ordinal data was conducted using the Mann-Whitney U test. The random forest interpolation method was used for missing values in data interpolation. Univariate analysis was conducted using the COX proportional risk regression model, and multivariate analysis was conducted using the COX stepwise regression with forward method. The coefficient of multiplication interaction effect was obtained using the interaction term coefficient of COX proportional risk regression model. Evaluation of additive interaction effects was conducted using the relative excess risk due to interaction ( RERI). Results:(1) Comparison of clinicopathological characteristics between patients of different MMR protein status. There were significant differences in age, T staging, the number of lymph node harvest, the number of lymph node harvest <12, high grade tumor between patients of dMMR and pMMR ( P<0.05). (2) Analysis of factors affecting the survival outcomes of patients of dMMR. Results of multivariate analysis showed that T staging, N staging, the number of lymph node harvest <12 were independent factors affecting the disease-free survival (DFS) of colon cancer patients of dMMR ( hazard ratio=3.548, 2.589, 6.702, 95% confidence interval as 1.460-8.620, 1.064-6.301, 1.886-23.813, P<0.05). Age and N staging were independent factors affecting the overall survival (OS) of colon cancer patients of dMMR ( hazard ratio=1.073, 10.684, 95% confidence interval as 1.021-1.126, 2.311-49.404, P<0.05). (3) Analysis of factors affecting the survival outcomes of patients of pMMR. Results of multivariate analysis showed that age, T staging, N staging, vascular tumor thrombus were independent factors affecting the DFS of colon cancer patients of pMMR ( hazard ratio=1.018, 2.214, 2.598, 1.549, 95% confidence interval as 1.006-1.030, 1.618-3.030, 1.921-3.513, 1.118-2.147, P<0.05). Age, T staging, N staging, high grade tumor were independent factors affecting the OS of colon cancer patients of pMMR ( hazard ratio=1.036, 2.080, 2.591, 1.615, 95% confidence interval as 1.020-1.052, 1.407-3.075, 1.791-3.748, 1.114-2.341, P<0.05). (4) Interaction analysis of MMR and adverse clinicopathological features on survival outcomes. Results of interaction analysis showed that the multiplication interaction effect between the number of lymph node harvest <12 and MMR protein status was significant on DFS of colon cancer patients ( hazard ratio=3.923, 95% confidence interval as 1.057-14.555, P<0.05). The additive interaction effects between age and MMR protein status, between high grade tumor and MMR protein status were significant on OS of colon cancer patients ( RERI=-0.033, -1.304, 95% confidence interval as -0.049 to -0.018, -2.462 to -0.146). Conclusions:There is an interaction between the MMR protein status and the adverse clinicopathological features (the number of lymph node harvest <12, high grade tumor) on prognosis of colon cancer patients of stage Ⅰ-Ⅲ. In patients of dMMR, the number of lymph node harvest <12 has a stronger predictive effect on poor prognosis. In patients of pMMR, the high grade tumor has a stronger predictive effect on poor prognosis.

Objective To present our initial experience of suturing uterine blood vessels with barbed polydioxanone cog thread by single-port-access large laparoscopic hysterectomy.Methods A respective non-randomized study with two parallel groups was performed from June 2021 to June 2022.The medical records of a total of 41 patients with enlarged uterus were divided into two groups:The single-port access large laparoscopic hysterectomy group by suturing uterine blood vessels with barbed cog thread(experimental group,n = 20),and the four-port assess laparoscopic hysterectomy group(control group,n = 21).The age,body mass index(BMI),a history of abdominal and pelvic surgery,pre-and post-operative hemoglobin(HGB),operative duration,intraoperative bleeding volume,uterine weight,postoperative pain visual analogue scale(VAS)at 24 h,the first postoperative exhaust time and postoperative length of hospital stay were compared between the two groups.Results 41 patients completed the entire procedure successfully without serious complications.Experimental group had a longer median operative time(P<0.01).There were no significant differences in intraoperative bleeding volume,uterine weight,HGB on postoperative day 1,postoperative hospital stays,postoperative exhaust time and postoperative 24-hour pain VAS(P>0.05).Conclusion Our study shows that suturing uterine blood vessels with barbed polydioxanone cog thread in experimental group is safe and feasible for large uterus,but the surgical time is significantly longer than the control group,as the technology becomes more proficient,this situation will be improved.

Objective To explore the effect of different treatment methods on prognosis in elderly patients with lower extremity arterial occlusive disease.Methods A total of 352 elderly patients with lower extremity arterial occlusive disease admitted in our hospital from May 2020 to May 2022 were enrolled,and according to their willingness and characteristics of lower extremity le-sions,they were divided into balloon dilation group(142 patients),stent implantation group(145 patients)and conservative treatment group(65 patients).All patients were followed up for 13-24 months.The incidences of major adverse cardiovascular events(MACE),including all-cause death,acute myocardial infarction,acute ischemic stroke,and major adverse lower limb events(MALE),including lower extremity pain at rest,ulcers or skin defects,gangrene,reocclusion,and amputation were observed and recorded.The clinical data and prognosis were compared and ana-lyzed of the three groups.Kaplan-Meier survival curves were drawn.Results The incidence of all-cause mortality was significantly lower in the stent implantation group than the conservative treatment group(9.7％vs 23.1％,P＜0.01).The incidence of MALE was obviously lower in the stent implantation group and the balloon dilatation group than the conservative treatment group(4.8％and 9.2％vs 24.6％,P＜0.01).Conclusion Endovascular therapy can reduce the risk of all-cause death and MALE occurrence in elderly patients with lower extremity arterial occlusive disease who are suitable for interventional therapy.

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.

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.