Objective:To investigate the local hemodynamic spatial distribution of internal carotid artery stenosis and its correlation with plaque characteristics using high-resolution vessel wall imaging (HR-VWI) combined with computational fluid dynamics.Methods:This was a cross-sectional study. A retrospective analysis was conducted on the clinical and imaging data of 70 patients with moderate to severe stenosis at the initiation of the internal carotid artery in First Teaching Hospital of Tianjin University of Traditional Chinese Medicine and Tianjin First Central Hospital from March 2018 to June 2020. All patients underwent HR-VWI and CT angiography examinations. The parameters related to plaque characteristics, such as plaque length, maximum wall thickness, plaque volume, wall volume percentage and intraplaque hemorrhage (IPH) were measured and evaluated on HR-VWI images. CT angiography images were used to construct a local hemodynamic vascular model to measure various wall shear stress (WSS) derived parameters, such as time-averaged wall shear stress (TAWSS), oscillatory shear index (OSI), relative residence time (RRT), and transverse wall shear stress (transWSS), at the narrowest, proximal, and distal parts of the lesion. The Friedman test was used to analyze the difference of hemodynamic parameters in different parts of the lesion. Spearman correlation analysis was performed to assess the correlation between plaque burden and local hemodynamic parameters. Univariate and multivariate logistic regression analyses were used to explore the independent risk factors for predicting IPH.Results:Among the 70 patients, 25 patients with IPH and 45 patients without IPH. The overall differences in TAWSS, OSI, RRT and transWSS at the narrowest, proximal, and distal parts of the lesion in 70 patients were statistically significant ( P<0.05). The TAWSS and transWSS at the narrowest parts were significantly higher than those at the proximal and distal parts ( P<0.05). The OSI at the distal part was significantly higher than that at the narrowest and the proximal parts ( P<0.05). The RRT at the proximal part was significantly lower than that at the narrowest and the distal parts ( P<0.05). Spearman correlation analysis showed RRT at the distal part was correlated with plaque volume ( r s=0.249, P=0.044) and wall volume percentage ( r s=0.286, P=0.016), respectively. In a multivariate logistic regression showed plaque length ( OR=1.315, 95% CI 1.073-1.612, P=0.008) and TAWSS at the narrowest part ( OR=1.631, 95% CI 1.308-1.854, P=0.008) were independent risk factors for predicting IPH. Conclusions:The spatial distribution of local hemodynamics of moderate to severe stenosis at the initiation of the internal carotid artery is different, and the WSS parameters in different parts of the lesion have different effects on plaque volume, wall volume percentage and IPH.

Objective To investigate the feasibility and accuracy of an ultrasound-based transfer learning artificial intelligence model in predicting the malignancy probability of partially cystic thyroid nodules(PCTN).Methods A retrospective analysis was conducted on 246 patients with PCTN who had definitive pathological results and were admitted to Weihai Municipal Hospital,Cheeloo College of Medicine,Shandong University from January 2021 to December 2023.Patients were randomly divided into training and test cohorts at a ratio of 7:3.Ultrasonic image features of PCTN were evaluated,and independent risk factors were identified using multivariate logistic regression analysis,with the area under the curve(AUC)subsequently calculated.Additionally,five different pre-trained models-Inception_v3,EfficientNet,VGG19,ResNet50,and DenseNet121-were selected for transfer learning after data preprocessing using the PyTorch framework in Python.The AUC values of these models were calculated and compared.Results Solid portion greater than 50%,eccentric acute angle,ill-defined margin,spiculated or microlobulated margin,rim calcification,and microcalcification exhibited statistically significant differences(P<0.05)in distinguishing between benign and malignant PCTN.The AUC value derived from these independent risk factors was 0.843.Furthermore,among the five transfer learning models evaluated,the ResNet50 model demonstrated the highest diagnostic efficiency,achieving an AUC value of 0.903 2.Conclusion The ultrasound-based transfer learning artificial intelligence model demonstrated superior performance compared to traditional ultrasound image evaluation methods,enabling accurate prediction of the nature of PCTN and thereby reducing unnecessary ultrasound-guided fine needle biopsies.

Objective To investigate the feasibility and accuracy of an ultrasound-based transfer learning artificial intelligence model in predicting the malignancy probability of partially cystic thyroid nodules(PCTN).Methods A retrospective analysis was conducted on 246 patients with PCTN who had definitive pathological results and were admitted to Weihai Municipal Hospital,Cheeloo College of Medicine,Shandong University from January 2021 to December 2023.Patients were randomly divided into training and test cohorts at a ratio of 7:3.Ultrasonic image features of PCTN were evaluated,and independent risk factors were identified using multivariate logistic regression analysis,with the area under the curve(AUC)subsequently calculated.Additionally,five different pre-trained models-Inception_v3,EfficientNet,VGG19,ResNet50,and DenseNet121-were selected for transfer learning after data preprocessing using the PyTorch framework in Python.The AUC values of these models were calculated and compared.Results Solid portion greater than 50%,eccentric acute angle,ill-defined margin,spiculated or microlobulated margin,rim calcification,and microcalcification exhibited statistically significant differences(P<0.05)in distinguishing between benign and malignant PCTN.The AUC value derived from these independent risk factors was 0.843.Furthermore,among the five transfer learning models evaluated,the ResNet50 model demonstrated the highest diagnostic efficiency,achieving an AUC value of 0.903 2.Conclusion The ultrasound-based transfer learning artificial intelligence model demonstrated superior performance compared to traditional ultrasound image evaluation methods,enabling accurate prediction of the nature of PCTN and thereby reducing unnecessary ultrasound-guided fine needle biopsies.

Objective:To systematic evaluate the correlation between arsenic and Alzheimer's disease (AD).Methods:Literature search was conducted using China National Knowledge Infrastructure, Wanfang Data Knowledge Service Platform, PubMed, Science Direct, Web of Science, and Cochrane Library to collect relevant literature on the correlation between arsenic and AD published domestically and internationally. The time limit was from the establishment of the database to February 1, 2024. Meta-analysis was conducted using RevMan 5.3 software, heterogeneity test was conducted using I2 statistic and Q-test, and random effects model or fixed effects model were used for comprehensive quantitative analysis based on heterogeneity results. Standard mean difference ( SMD) was used as the effect indicator and subgroup analysis was conducted based on detection markers and disease progression. Results:The final results included 10 articles (including 1 235 AD patients and 1 293 healthy controls), all of which were in English. After heterogeneity testing, significant heterogeneity was found in the results of each study ( I2 = 78%, P < 0.001). A random effects model was used for meta-analysis, and the combined SMD value (95% CI) was 0.58 (0.45, 0.70). The sub group analysis of detection markers showed that compared with the healthy controls the SMD values (95% CI) of urine arsenic, blood arsenic, ventricular fluid arsenic, frontal cortex arsenic, hair arsenic, nail arsenic, groundwater arsenic, and brain arsenic in AD patients were 5.02 (- 2.09, 12.14), 0.40 (- 0.02, 0.83), 5.00 (- 0.35, 10.35), - 14.00 (- 27.92, - 0.08), 0.75 (0.21, 1.29), 1.42 (0.48, 2.35), - 0.10 (- 0.48, 0.28), and 0.57 (0.36, 0.78), respectively. The sub group analysis of disease progression showed that that compared with the healthy controls the SMD values (95% CI) of arsenic content in mild, moderate, and severe AD patients were 0.88 (0.52, 1.24), 0.57 (0.24, 0.91), and 2.13 (0.26, 4.00), respectively. Conclusion:There is a correlation between arsenic and AD, and hair arsenic and nail arsenic can serve as potential important indicators for evaluating the correlation between arsenic and AD.

Objective:To construct a prediction model for the risk of sarcopenia in middle-aged and elderly patients with knee osteoarthritis (KOA) based on machine learning, and to provide a basis for carrying out the prevention of sarcopenia in patients with KOA.Methods:Clinical data of KOA patients from three tertiary hospitals in Guangdong Province were collected between December 2023 and September 2024 using a convenience sampling method. The data were randomly split into training and test sets at an 8:2 ratio, with the occurrence of sarcopenia as the outcome variable. Risk prediction models for sarcopenia were constructed using eight machine learning algorithms: logistic regression, K-nearest neighbors, support vector machine, decision tree, neural network, random forest, gradient boosting machine (GBM), and eXtreme gradient boosting. Model performance was evaluated based on metrics including the area under the receiver operating characteristic curve (AUC), accuracy, precision, sensitivity, specificity, and F1 score. The optimal model was selected, and feature importance was visualized using the Shapley Additive exPlanations (SHAP) method.Results:Data from 640 KOA patients were analyzed, 143 males and 497 females, (67.51± 7.72) years, with 136 cases (21.25%) developing sarcopenia. All eight prediction models showed high AUC values, with the GBM model demonstrating the best performance. Its metrics included an AUC of 0.926 (95% CI 0.874 - 0.965), accuracy of 0.852, precision of 0.611, sensitivity of 0.815, specificity of 0.861, and F1 score of 0.698. SHAP analysis identified body mass index, calf circumference, body fat percentage, WOMAC score, and age as the most important predictive features. Conclusions:The GBM-based risk prediction model for sarcopenia in middle- aged and elderly KOA patients demonstrated optimal performance, enabling healthcare professionals to accurately and promptly identify high-risk groups among these patients and to develop effective, evidence-based intervention strategies.

Objective:To systematic evaluate the correlation between arsenic and Alzheimer's disease (AD).Methods:Literature search was conducted using China National Knowledge Infrastructure, Wanfang Data Knowledge Service Platform, PubMed, Science Direct, Web of Science, and Cochrane Library to collect relevant literature on the correlation between arsenic and AD published domestically and internationally. The time limit was from the establishment of the database to February 1, 2024. Meta-analysis was conducted using RevMan 5.3 software, heterogeneity test was conducted using I2 statistic and Q-test, and random effects model or fixed effects model were used for comprehensive quantitative analysis based on heterogeneity results. Standard mean difference ( SMD) was used as the effect indicator and subgroup analysis was conducted based on detection markers and disease progression. Results:The final results included 10 articles (including 1 235 AD patients and 1 293 healthy controls), all of which were in English. After heterogeneity testing, significant heterogeneity was found in the results of each study ( I2 = 78%, P < 0.001). A random effects model was used for meta-analysis, and the combined SMD value (95% CI) was 0.58 (0.45, 0.70). The sub group analysis of detection markers showed that compared with the healthy controls the SMD values (95% CI) of urine arsenic, blood arsenic, ventricular fluid arsenic, frontal cortex arsenic, hair arsenic, nail arsenic, groundwater arsenic, and brain arsenic in AD patients were 5.02 (- 2.09, 12.14), 0.40 (- 0.02, 0.83), 5.00 (- 0.35, 10.35), - 14.00 (- 27.92, - 0.08), 0.75 (0.21, 1.29), 1.42 (0.48, 2.35), - 0.10 (- 0.48, 0.28), and 0.57 (0.36, 0.78), respectively. The sub group analysis of disease progression showed that that compared with the healthy controls the SMD values (95% CI) of arsenic content in mild, moderate, and severe AD patients were 0.88 (0.52, 1.24), 0.57 (0.24, 0.91), and 2.13 (0.26, 4.00), respectively. Conclusion:There is a correlation between arsenic and AD, and hair arsenic and nail arsenic can serve as potential important indicators for evaluating the correlation between arsenic and AD.

Objective:To investigate the local hemodynamic spatial distribution of internal carotid artery stenosis and its correlation with plaque characteristics using high-resolution vessel wall imaging (HR-VWI) combined with computational fluid dynamics.Methods:This was a cross-sectional study. A retrospective analysis was conducted on the clinical and imaging data of 70 patients with moderate to severe stenosis at the initiation of the internal carotid artery in First Teaching Hospital of Tianjin University of Traditional Chinese Medicine and Tianjin First Central Hospital from March 2018 to June 2020. All patients underwent HR-VWI and CT angiography examinations. The parameters related to plaque characteristics, such as plaque length, maximum wall thickness, plaque volume, wall volume percentage and intraplaque hemorrhage (IPH) were measured and evaluated on HR-VWI images. CT angiography images were used to construct a local hemodynamic vascular model to measure various wall shear stress (WSS) derived parameters, such as time-averaged wall shear stress (TAWSS), oscillatory shear index (OSI), relative residence time (RRT), and transverse wall shear stress (transWSS), at the narrowest, proximal, and distal parts of the lesion. The Friedman test was used to analyze the difference of hemodynamic parameters in different parts of the lesion. Spearman correlation analysis was performed to assess the correlation between plaque burden and local hemodynamic parameters. Univariate and multivariate logistic regression analyses were used to explore the independent risk factors for predicting IPH.Results:Among the 70 patients, 25 patients with IPH and 45 patients without IPH. The overall differences in TAWSS, OSI, RRT and transWSS at the narrowest, proximal, and distal parts of the lesion in 70 patients were statistically significant ( P<0.05). The TAWSS and transWSS at the narrowest parts were significantly higher than those at the proximal and distal parts ( P<0.05). The OSI at the distal part was significantly higher than that at the narrowest and the proximal parts ( P<0.05). The RRT at the proximal part was significantly lower than that at the narrowest and the distal parts ( P<0.05). Spearman correlation analysis showed RRT at the distal part was correlated with plaque volume ( r s=0.249, P=0.044) and wall volume percentage ( r s=0.286, P=0.016), respectively. In a multivariate logistic regression showed plaque length ( OR=1.315, 95% CI 1.073-1.612, P=0.008) and TAWSS at the narrowest part ( OR=1.631, 95% CI 1.308-1.854, P=0.008) were independent risk factors for predicting IPH. Conclusions:The spatial distribution of local hemodynamics of moderate to severe stenosis at the initiation of the internal carotid artery is different, and the WSS parameters in different parts of the lesion have different effects on plaque volume, wall volume percentage and IPH.

Objective:To construct a prediction model for the risk of sarcopenia in middle-aged and elderly patients with knee osteoarthritis (KOA) based on machine learning, and to provide a basis for carrying out the prevention of sarcopenia in patients with KOA.Methods:Clinical data of KOA patients from three tertiary hospitals in Guangdong Province were collected between December 2023 and September 2024 using a convenience sampling method. The data were randomly split into training and test sets at an 8:2 ratio, with the occurrence of sarcopenia as the outcome variable. Risk prediction models for sarcopenia were constructed using eight machine learning algorithms: logistic regression, K-nearest neighbors, support vector machine, decision tree, neural network, random forest, gradient boosting machine (GBM), and eXtreme gradient boosting. Model performance was evaluated based on metrics including the area under the receiver operating characteristic curve (AUC), accuracy, precision, sensitivity, specificity, and F1 score. The optimal model was selected, and feature importance was visualized using the Shapley Additive exPlanations (SHAP) method.Results:Data from 640 KOA patients were analyzed, 143 males and 497 females, (67.51± 7.72) years, with 136 cases (21.25%) developing sarcopenia. All eight prediction models showed high AUC values, with the GBM model demonstrating the best performance. Its metrics included an AUC of 0.926 (95% CI 0.874 - 0.965), accuracy of 0.852, precision of 0.611, sensitivity of 0.815, specificity of 0.861, and F1 score of 0.698. SHAP analysis identified body mass index, calf circumference, body fat percentage, WOMAC score, and age as the most important predictive features. Conclusions:The GBM-based risk prediction model for sarcopenia in middle- aged and elderly KOA patients demonstrated optimal performance, enabling healthcare professionals to accurately and promptly identify high-risk groups among these patients and to develop effective, evidence-based intervention strategies.

Objective To analyze the expression level of YEATS2 in hepatocellular carcinoma(HCC)about its clinical prognosis and therapeutic value based on biological information from the cancer genome atlas(TCGA)and human protein atlas(HPA)databases.Methods The mRNA expression data and clinical information of HCC were downloaded from the TCGA database,the expression of YEATS2 between HCC tissues and normal tissues was analyzed by using the R software,and the protein expression differences were preliminary verified by the HPA database.The expression differences of YEATS2 between various clinical features of HCC were compared,and their effects on the survival of HCC patients by Kaplan-Meier method and COX regression analysis were then evaluated.Receiver operating characteristic(ROC)curves were plotted to evaluate their diagnostic values.The biological functions of YEATS2 in HCC were analyzed using gene ontology(GO)and Kyoto encyclopedia of genes and genomes(KEGG)enrichment analysis.The relationship between YEATS2 expression and tumor microenvironment(TME)was analyzed by the"ESTIMATE"algorithm,and its relationship with tumor-infiltrating immune cells(TIICs)was assessed by CIBERSORT.Analysis of YEATS2 expression levels to immune checkpoints and drug sensitivity was performed using the R package.Results The expression of YEATS2 was increased in HCC tissues(P=4.96e-21),and its expression level was correlated with age,clinical stage,pathological grade and T stage(all P＜0.05).Overall survival(OS)(P＜0.001)and progression-free survival(FPS)(P=0.016)were decreased in HCC patients with high expression of YEATS2,COX regression results showed that the expression level ofYEATS2 was associated with poor prognosis in HCC patients(OS:HR=2.167,95％CI:1.441～3.261,P=2.06e-04),and it was an independent risk factor for predicting poor prognosis in HCC patients(OS:HR=1.891,95％CI:1.243～2.877,P=0.003).The ROC curve suggested the AUCs for 1,3 and 5 years were 0.677,0.622 and 0.612,respectively,indicating good predictive ability.The TCGA database screened a total of 6 764 differential genes in the YEATS2 high and low expression groups,of which 4 094 genes were up-regulated and 2 670 genes were down-regulated in the YEATS2 high expression group.The results of GO and KEGG enrichment analyses showed that the differentially differentiated genes in the YEATS2 high expression group were mainly enriched in immunoregulation,and cell cycle regulation drug resistance pathway.The results of the TME score showed that the YEATS2 high expression group caused a decrease in immunity score(P＜0.01).The correlation between YEATS2 and TIICs showed that YEATS2 expression was positively correlated with the level of M0-type macrophage infiltration levels(r=0.48,P＜0.001)and 23 immune checkpoint genes(r=0.20～0.46,all P＜0.05),and was negatively correlated with the CD8+T-cells,plasma cells and monocyte(r=-0.26,-0.29,-0.30,P=0.021,0.011,0.008).Drug sensitivity analysis showed that the half maximal inhibitory concentration(IC50)of cabozantinib,lincitinib,doxorubicin,and cyclobenzaprine in patients with high expression of YEATS2 was higher than those in patients with low expression(all P＜0.01).Conclusion YEATS2 was highly expressed in HCC,and the expression level was associated with poor prognosis in HCC patients.YEATS2 can be used as a biomarker for the clinical early diagnosis,prognosis and immunotherapy of HCC,which may provide new ideas for clinical diagnosis and treatment.

Objective:To investigate the effects of breast milk to total milk intake ratio during hospitalization on the duration of antibiotic therapy in preterm infants less than 34 weeks of gestation.Methods:Clinical data of preterm infants ( n=1 792) less than 34 gestational weeks were retrospectively collected in 16 hospitals of Jiangsu Province Neonatal-Perinatal Cooperation Network from January 1, 2019, to December 31, 2021. The days of therapy (DOT) were used to evaluate the duration of antibiotic administration. The median DOT was 15.0 d (7.0-27.0 d). The patients were divided into four groups based on the quartiles of DOT: Q 1 (DOT≤7.0 d), Q 2 (7.0 d27.0 d) groups. According to the breast milk intake ratio (breast milk intake to total milk intake during hospitalization×100%), they were also divided into four groups: very-low-ratio breastfeeding group (breast milk intake ratio≤25%), low-ratio breastfeeding group (25%75%). Univariate analysis ( Chi-square test and Kruskal-Wallis rank-sum test) was used to analyze the factors influencing DOT. Spearman correlation analysis and trend Chi-square test were used to explore the relationship between breast milk intake ratio and DOT. After using multiple imputations to address missing data, two models were constructed after adjusting for different factors, and multinomial logistic regression model was applied to evaluate the effects of the breast milk intake ratio on DOT. Finally, sensitivity analysis was conducted to assess the stability of the models. Results:(1) Of the 1 792 preterm infants, there were 507 (28.3%) in the Q 1 group, 422 (23.5%) in the Q 2 group, 438 (24.4%) in the Q 3 group and 425 (23.7%) in the Q 4 group. (2) The median values of DOT in the very-low-ratio, low-ratio, medium-ratio and high-ratio breastfeeding groups were 20.0 d (11.0-31.0 d), 20.0 d (11.0-32.0 d), 13.0 d (6.0-25.8 d) and 10.0 d (4.0-21.0 d), respectively. Compared with the very-low-ratio and low-ratio breastfeeding groups, the medium-ratio and high-ratio breastfeeding groups had shorter DOT (all P<0.05). (3) After adjusting for factors with P<0.1 (prenatal glucocorticoid exposure, antimicrobial use within 24 h before delivery, gestational age at delivery, birth weight, Apgar score≤7 at 1 min, neonatal respiratory distress syndrome, infectious pneumonia and early-onset neonatal sepsis) between the DOT quartile groups, it showed that medium-ratio and high-ratio breastfeeding were protective factors in contrast to very-low-ratio breastfeeding in the Q 2, Q 3 and Q 4 groups as compared with the Q 1 group [Q 2 group: OR=0.50 (95% CI: 0.30-0.85) and OR=0.36 (95% CI: 0.26-0.51); Q 3 group: OR=0.31 (95% CI: 0.18-0.55) and OR=0.20 (95% CI: 0.14-0.29); Q 4 group: OR=0.22 (95% CI: 0.12-0.42) and OR=0.17 (95% CI: 0.12-0.26)]. Conclusion:Breast milk intake accounting for over 50% of total milk intake has a positive impact on reducing DOT in premature infants requiring antibiotics, which suggests that breastfeeding should be actively encouraged.