Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Cryoablation therapy with explicit anti-tumor mechanisms and histopathological manifestations has a long history.A large number of clinical practice has shown that cryoablation therapy is safe and effective,making it an ideal tumor treatment method in theory.Previously,its efficacy and clinical application were constrained by the limitations of refrigerants and refrigeration equipment.With the development of the new generation of cryoablation equipment represented by argon helium knives,significant progress has been made in refrigeration efficien-cy,ablation range,and precise temperature measurement,greatly promoting the progression of tumor cryoablation technology.This consensus systematically summarizes the mechanism of cryoablation technology,indications for oral mucosal melanoma(OMM)cryotherapy,clinical treatment process,adverse reactions and management,cryotherapy combination therapy,etc.,aiming to provide reference for carrying out the standardized cryoablation therapy of OMM.

Objective:To investigate the predictive value of early thyroid function changes on the efficacy of patients with Graves′ disease (GD) after 131I therapy. Methods:Data of patients with GD (59 males, 214 females; age (37.4±11.4) years) who underwent single therapy of 131I in Tianjin Medical University General Hospital from November 2017 to January 2019 were retrospectively analyzed. Symptoms, signs and laboratory tests (serum free triiodothyronine (FT 3) and serum free thyroxine (FT 4)) of patients were observed to assess the efficacy of 131I treatment. Efficacy was divided into complete remission (CR), partial remission (PR), non-remission (NR) or relapse. The changes of thyroid function (ΔFT 3=FT 3 before treatment-FT 3 after treatment)/FT 3 before treatment×100%; ΔFT 4=FT 4 before treatment-FT 4 after treatment)/FT 4 before treatment×100%) 1 month after 131I therapy in each efficacy group and differences among them were compared by using independent-sample t test, χ2 test, one-way analysis of variance and the least significant difference t test. ROC curves were drawn to analyze the predictive values of early thyroid function changes on the efficacy of 131I treatment for GD. Logistic regression analyses were performed to identify the influencing factors for the efficacy of 131I therapy. Results:CR rate and total effective rate of 273 GD patients after single therapy of 131I were 67.03%(183/273) and 92.67%(253/273), respectively. After 1 month, CR rate of euthyroidism group ( n=95) was significantly higher than that of hyperthyroidism group ( n=178; 81.05%(77/95) vs 59.55%(106/178); χ2=4.60, P=0.032). ΔFT 3 and ΔFT 4 at the first month were statistically significant and decreased sequentially in the CR group ( n=183), PR group ( n=70), NR or relapse groups ( n=20; F values: 15.40, 12.54, both P<0.001). ROC curve analysis showed that patients with ΔFT 3≥73.64% and (or) ΔFT 4≥59.03% had a higher probability of achieving CR, with sensitivities of 84.3% and 86.7%, and specificities of 62.6% and 62.6%, respectively. Logistic regression analysis showed that 24 h radioactive iodine uptake (odds ratio ( OR)=1.095, 95% CI: 1.031-1.139), dose of 131I given per gram of thyroid tissue ( OR=1.562, 95% CI: 1.321-1.694), ΔFT 3 ( OR=1.354, 95% CI: 1.295-1.482), ΔFT 4 ( OR=1.498, 95% CI: 1.384-1.608) were factors affecting the outcome of patients with GD treated with 131I treatment (all P<0.05). Conclusion:Effects of 131I treatment can be predicted based on the change of the thyroid function at the first month after 131I treatment in patients with GD.

Objective To investigate the mechanism by which G protein-coupled estrogen receptor(GPER)reduces retinal neovascularisation in oxygen-induced retinopathy(OIR)in newborn mice.Methods A total of 42 newborn mice were randomly divided into normoxic control group(n=11),OIR group(n=11),G-1(GPER agonist)group(n=10),and solvent control group(n=10).On postnatal day 17(P17),the distribution of GPER in the retina of mice in the normoxic control group was observed by immunofluorescence staining on frozen sections of the eyeballs.The mice of the G-1 group and solvent control group were given by intraperitoneal injection 50 μg/(kg·d)G-1 or corn oil solvent from P12 to P15.Immunofluorescence staining of retinal spreads at P17 was performed to observe the expression of retinal vascular marker IB4 and astrocyte marker glia fibrilary acidic protein(GFAP).Western blotting was used to quantify the expression of retinal vascular endothelial growth factor A(VEGFA),GFAP,and inflammatory factors TNF-α,IGF-1,and IL1-β.Results GPER was present throughout the retina and co-stained with the astrocyte marker GFAP in the ganglion cell layer(GCL).Compared with the normoxic control group,the retinas of mice in the OIR group showed neovascular and avascular areas,and the expression of VEGFA and GFAP was significantly increased(P＜0.05).Compared with the solvent control group,the retinal neovascularisation was reduced in the G-l group,and the expression levels of VEGFA,GFAP,TNF-α,IGF-1,and ILl-β proteins were decreased significantly(P＜0.05).Conclusion GPER inhibits astrocyte activity,reduces VEGFA expression and release of inflammatory factors,and then decreases retinal neovascularisation in OIR mice.

Objective To explore the influential factors and predictive value of early femoral head collapse progression following allogeneic fibula grafting(AFG)surgery.Methods Clinical and radiological data of 68 patients(75 hips)with osteonecrosis of the femoral head(ONFH)who underwent AFG between January 2008 and December 2022 at the Orthopedics and Traumatology Department,Affiliated Hospital of Nanjing University of Chinese Medicine were retrospectively analyzed.Seventy-five hips were divided into stable(n=40)and progressive(n=35)groups based on the presence or absence of postoperative collapse progression.Age,gender,etiology,location of the lesion,Association Research Circulation Osseous(ARCO)stage,Japanese Committee of Osteonecrosis Investigation(JIC)classification,China-Japan Friendship Hospital(CJFH)classification,and Hounsfield units(HU)value of anterolateral sclerosis rim(ⅠSHU)were collected.Univariate and multivariate logistic regression analyses were used to identify the factors influencing early collapse progression after AFG.Receiver operating characteristic(ROC)curve was used to analyze the predictive value of the identified factors influencing postoperative early collapse progression.Results Of the 75 hips,35(46.7%)had postoperative collapse progression.Univariate logistic regression analysis showed that age,ARCO stage,JIC classification,and ⅠSHU were in fluencing factors for early femoral head collapse progression after AFG(P<0.05).Multivariate logistic regression analysis showed that ARCO stage ⅢA and JIC classification C2 were independent risk factors for early femoral head collapse progression after AFG,while ⅠSHU was identified as an independent protective factor(P<0.05).The ROC curve analysis showed that the sensitivities of ARCO stage,JIC classification,ⅠSHU,and the combined predictive model were 0.850,0.725,0.800,and 0.775,the specificities were 0.486,0.657,0.743,and 0.914,and the area under the ROC curve(AUC)were 0.668,0.725,0.811,and 0.896,respectively.Conclusions ⅠSHU is associated with early collapse progression after AFG in patients with ONFH.ARCO stage ⅢA,JIC classification C2,and ⅠSHU are independent factors influencing postoperative early collapse progression and have a certain predictive value.

Objective To investigate how maternal MTR gene polymorphisms and their interactions with periconceptional folic acid supplementation are associated with the incidence of ventricular septal defects(VSD)in offspring.Methods A case-control study was conducted,recruiting 426 mothers of infants with VSD under one year old and 740 mothers of age-matched healthy infants.A questionnaire survey collected data on maternal exposures,and blood samples were analyzed for genetic polymorphisms.Multivariable logistic regression analysis and inverse probability of treatment weighting were used to analyze the associations between genetic loci and VSD.Crossover analysis and logistic regression were utilized to examine the additive and multiplicative interactions between the loci and folic acid intake.Results The CT and TT genotypes of the maternal MTR gene at rs6668344 increased the susceptibility of offspring to VSD(P<0.05).The GC and CC genotypes at rs3768139,AG and GG at rs1050993,AT and TT at rs4659743,GG at rs3768142,and GT and TT at rs3820571 were associated with a decreased risk of VSD(P<0.05).The variations at rs6668344 demonstrated an antagonistic multiplicative interaction with folic acid supplementation in relation to VSD(P<0.05).Conclusions Maternal MTR gene polymorphisms significantly correlate with the incidence of VSD in offspring.Mothers with variations at rs6668344 can decrease the susceptibility to VSD in their offspring by supplementing with folic acid during the periconceptional period,suggesting the importance of periconceptional folic acid supplementation in genetically at-risk populations to prevent VSD in offspring.