Testicular histology based on testicular biopsy is an important factor for determining appropriate testicular sperm extraction surgery and predicting sperm retrieval outcomes in patients with azoospermia. Therefore, we developed a deep learning (DL) model to establish the associations between testicular grayscale ultrasound images and testicular histology. We retrospectively included two-dimensional testicular grayscale ultrasound from patients with azoospermia (353 men with 4357 images between July 2017 and December 2021 in The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China) to develop a DL model. We obtained testicular histology during conventional testicular sperm extraction. Our DL model was trained based on ultrasound images or fusion data (ultrasound images fused with the corresponding testicular volume) to distinguish spermatozoa presence in pathology (SPP) and spermatozoa absence in pathology (SAP) and to classify maturation arrest (MA) and Sertoli cell-only syndrome (SCOS) in patients with SAP. Areas under the receiver operating characteristic curve (AUCs), accuracy, sensitivity, and specificity were used to analyze model performance. DL based on images achieved an AUC of 0.922 (95% confidence interval [CI]: 0.908-0.935), a sensitivity of 80.9%, a specificity of 84.6%, and an accuracy of 83.5% in predicting SPP (including normal spermatogenesis and hypospermatogenesis) and SAP (including MA and SCOS). In the identification of SCOS and MA, DL on fusion data yielded better diagnostic performance with an AUC of 0.979 (95% CI: 0.969-0.989), a sensitivity of 89.7%, a specificity of 97.1%, and an accuracy of 92.1%. Our study provides a noninvasive method to predict testicular histology for patients with azoospermia, which would avoid unnecessary testicular biopsy.
Humans ; Male ; Azoospermia/diagnostic imaging* ; Deep Learning ; Testis/pathology* ; Retrospective Studies ; Adult ; Ultrasonography/methods* ; Sperm Retrieval ; Sertoli Cell-Only Syndrome/diagnostic imaging*

In the field of dental aesthetics,digital aesthetic design plays a crucial role in helping dentists to predict treatment outcomes vis-ually,as well as in enhancing the consistency of knowledge and understanding of aesthetic goals between dentists and patients.It serves as the foundation for achieving ideal aesthetic effects.However,there is no clear standard for this digital process currently in China and abroad.Many dentists lack of systematic understanding of how to carry out digital aesthetic design for treatment.To establish standardized processes for dental aesthetic design and to improve the homogeneity of treatment outcomes,Chinese Society of Digital Dental Industry(CSD-DI)convened domestic experts in related field to compile this consensus.This article elaborates on the key aspects of digital aesthetic data collection,integration steps,and the digital aesthetic design process.It also formulates a decision tree for dental aesthetics at macro level and outlines corresponding workflows for various clinical scenarios,serving as a reference for clinicians.

Objective: This study aims to overcome challenges in lumbar spine imaging, particularly lumbar spinal stenosis, by developing an automated segmentation model using advanced techniques. Traditional manual measurement and lesion detection methods are limited by subjectivity and inefficiency. The objective is to create an accurate and automated segmentation model that identifies anatomical structures in lumbar spine magnetic resonance imaging scans. Methods: Leveraging a dataset of 539 lumbar spinal stenosis patients, the study utilizes the residual U-Net for semantic segmentation in sagittal and axial lumbar spine magnetic resonance images. The model, trained to recognize specific tissue categories, employs a geometry algorithm for anatomical structure quantification. Validation metrics, like Intersection over Union (IOU) and Dice coefficients, validate the residual U-Net’s segmentation accuracy. A novel rotation matrix approach is introduced for detecting bulging discs, assessing dural sac compression, and measuring yellow ligament thickness. Results: The residual U-Net achieves high precision in segmenting lumbar spine structures, with mean IOU values ranging from 0.82 to 0.93 across various tissue categories and views. The automated quantification system provides measurements for intervertebral disc dimensions, dural sac diameter, yellow ligament thickness, and disc hydration. Consistency between training and testing datasets assures the robustness of automated measurements. Conclusion Automated lumbar spine segmentation with residual U-Net and deep learning exhibits high precision in identifying anatomical structures, facilitating efficient quantification in lumbar spinal stenosis cases. The introduction of a rotation matrix enhances lesion detection, promising improved diagnostic accuracy, and supporting treatment decisions for lumbar spinal stenosis patients.

Objective: This study aims to overcome challenges in lumbar spine imaging, particularly lumbar spinal stenosis, by developing an automated segmentation model using advanced techniques. Traditional manual measurement and lesion detection methods are limited by subjectivity and inefficiency. The objective is to create an accurate and automated segmentation model that identifies anatomical structures in lumbar spine magnetic resonance imaging scans. Methods: Leveraging a dataset of 539 lumbar spinal stenosis patients, the study utilizes the residual U-Net for semantic segmentation in sagittal and axial lumbar spine magnetic resonance images. The model, trained to recognize specific tissue categories, employs a geometry algorithm for anatomical structure quantification. Validation metrics, like Intersection over Union (IOU) and Dice coefficients, validate the residual U-Net’s segmentation accuracy. A novel rotation matrix approach is introduced for detecting bulging discs, assessing dural sac compression, and measuring yellow ligament thickness. Results: The residual U-Net achieves high precision in segmenting lumbar spine structures, with mean IOU values ranging from 0.82 to 0.93 across various tissue categories and views. The automated quantification system provides measurements for intervertebral disc dimensions, dural sac diameter, yellow ligament thickness, and disc hydration. Consistency between training and testing datasets assures the robustness of automated measurements. Conclusion Automated lumbar spine segmentation with residual U-Net and deep learning exhibits high precision in identifying anatomical structures, facilitating efficient quantification in lumbar spinal stenosis cases. The introduction of a rotation matrix enhances lesion detection, promising improved diagnostic accuracy, and supporting treatment decisions for lumbar spinal stenosis patients.

Aim To observe the effects of Wenfei Jiangzhuo formula(WFJZF)on rats with vascular de-mentia and investigate its possible mechanism of ac-tion.Methods Thirty-six healthy male SD rats were randomly divided into the sham group,model group,donepezil group,and low-dose,medium-dose and high-dose groups of Wenfei Jiangzhuo formula,with six rats per group.Except for the sham group,the other groups were prepared as VaD models,and each group was gavaged with the corresponding drugs after suc-cessful modeling,and tests were performed after three weeks of treatment.Behavioral,hippocampal CA1 area morphology,neural dendrites and mitochondrial chan-ges were observed in all groups of rats,and phospho-glycerate mutase 5(PGAM5),dynamics-related pro-teins1(Drp1),opticatrophyprotein-1(OPA1),and other proteins were detected in each group.Results Compared with the sham group,rats in the model group and each intervention group had prolonged es-cape latency(P＜0.05),a shorter number of travers-als across the platforms(P＜0.05),a sparse morphol-ogy of hippocampal neurons,a reduction in the number of secondary dendritic spines,and a rupture of the out-er membrane of the mitochondria;the expression of the PGAM5 and Drp1 proteins in hippocampal tissues was elevated(P＜0.05),and the expression of the OPA1 and Mfn1/2 protein expression decreased(P＜0.05);compared with the model group,donepezil group and Wenfei Jiangzhuo formula high-dose group of rats had shorter evasion latency(P＜0.05),increased number of times to traverse the platform(P＜0.05),increased number of hippocampal neurons,tightly packed,more secondary dendritic structures,and reduced mitochon-drial damage;the expression of PGAM5 and Drp1 pro-teins was reduced(P＜0.05),and the expression of OPA1 and Mfn1/2 proteins was elevated(P＜0.05).Conclusions Wenfei Jiangzhuo formula can regulate the PGAM5-Drp1 signaling axis to improve the balance of mitochondrial homeostasis,thus improving the cog-nitive condition of the brain and exerting cerebroprotec-tive effects.

Objective: This study aims to overcome challenges in lumbar spine imaging, particularly lumbar spinal stenosis, by developing an automated segmentation model using advanced techniques. Traditional manual measurement and lesion detection methods are limited by subjectivity and inefficiency. The objective is to create an accurate and automated segmentation model that identifies anatomical structures in lumbar spine magnetic resonance imaging scans. Methods: Leveraging a dataset of 539 lumbar spinal stenosis patients, the study utilizes the residual U-Net for semantic segmentation in sagittal and axial lumbar spine magnetic resonance images. The model, trained to recognize specific tissue categories, employs a geometry algorithm for anatomical structure quantification. Validation metrics, like Intersection over Union (IOU) and Dice coefficients, validate the residual U-Net’s segmentation accuracy. A novel rotation matrix approach is introduced for detecting bulging discs, assessing dural sac compression, and measuring yellow ligament thickness. Results: The residual U-Net achieves high precision in segmenting lumbar spine structures, with mean IOU values ranging from 0.82 to 0.93 across various tissue categories and views. The automated quantification system provides measurements for intervertebral disc dimensions, dural sac diameter, yellow ligament thickness, and disc hydration. Consistency between training and testing datasets assures the robustness of automated measurements. Conclusion Automated lumbar spine segmentation with residual U-Net and deep learning exhibits high precision in identifying anatomical structures, facilitating efficient quantification in lumbar spinal stenosis cases. The introduction of a rotation matrix enhances lesion detection, promising improved diagnostic accuracy, and supporting treatment decisions for lumbar spinal stenosis patients.

Objective: This study aims to overcome challenges in lumbar spine imaging, particularly lumbar spinal stenosis, by developing an automated segmentation model using advanced techniques. Traditional manual measurement and lesion detection methods are limited by subjectivity and inefficiency. The objective is to create an accurate and automated segmentation model that identifies anatomical structures in lumbar spine magnetic resonance imaging scans. Methods: Leveraging a dataset of 539 lumbar spinal stenosis patients, the study utilizes the residual U-Net for semantic segmentation in sagittal and axial lumbar spine magnetic resonance images. The model, trained to recognize specific tissue categories, employs a geometry algorithm for anatomical structure quantification. Validation metrics, like Intersection over Union (IOU) and Dice coefficients, validate the residual U-Net’s segmentation accuracy. A novel rotation matrix approach is introduced for detecting bulging discs, assessing dural sac compression, and measuring yellow ligament thickness. Results: The residual U-Net achieves high precision in segmenting lumbar spine structures, with mean IOU values ranging from 0.82 to 0.93 across various tissue categories and views. The automated quantification system provides measurements for intervertebral disc dimensions, dural sac diameter, yellow ligament thickness, and disc hydration. Consistency between training and testing datasets assures the robustness of automated measurements. Conclusion Automated lumbar spine segmentation with residual U-Net and deep learning exhibits high precision in identifying anatomical structures, facilitating efficient quantification in lumbar spinal stenosis cases. The introduction of a rotation matrix enhances lesion detection, promising improved diagnostic accuracy, and supporting treatment decisions for lumbar spinal stenosis patients.

Objective: This study aims to overcome challenges in lumbar spine imaging, particularly lumbar spinal stenosis, by developing an automated segmentation model using advanced techniques. Traditional manual measurement and lesion detection methods are limited by subjectivity and inefficiency. The objective is to create an accurate and automated segmentation model that identifies anatomical structures in lumbar spine magnetic resonance imaging scans. Methods: Leveraging a dataset of 539 lumbar spinal stenosis patients, the study utilizes the residual U-Net for semantic segmentation in sagittal and axial lumbar spine magnetic resonance images. The model, trained to recognize specific tissue categories, employs a geometry algorithm for anatomical structure quantification. Validation metrics, like Intersection over Union (IOU) and Dice coefficients, validate the residual U-Net’s segmentation accuracy. A novel rotation matrix approach is introduced for detecting bulging discs, assessing dural sac compression, and measuring yellow ligament thickness. Results: The residual U-Net achieves high precision in segmenting lumbar spine structures, with mean IOU values ranging from 0.82 to 0.93 across various tissue categories and views. The automated quantification system provides measurements for intervertebral disc dimensions, dural sac diameter, yellow ligament thickness, and disc hydration. Consistency between training and testing datasets assures the robustness of automated measurements. Conclusion Automated lumbar spine segmentation with residual U-Net and deep learning exhibits high precision in identifying anatomical structures, facilitating efficient quantification in lumbar spinal stenosis cases. The introduction of a rotation matrix enhances lesion detection, promising improved diagnostic accuracy, and supporting treatment decisions for lumbar spinal stenosis patients.

Background/Aims: Despite the high efficacy of direct-acting antivirals (DAAs), approximately 1–3% of hepatitis C virus (HCV) patients fail to achieve a sustained virological response. We conducted a nationwide study to investigate risk factors associated with DAA treatment failure. Machine-learning algorithms have been applied to discriminate subjects who may fail to respond to DAA therapy. Methods: We analyzed the Taiwan HCV Registry Program database to explore predictors of DAA failure in HCV patients. Fifty-five host and virological features were assessed using multivariate logistic regression, decision tree, random forest, eXtreme Gradient Boosting (XGBoost), and artificial neural network. The primary outcome was undetectable HCV RNA at 12 weeks after the end of treatment. Results: The training (n=23,955) and validation (n=10,346) datasets had similar baseline demographics, with an overall DAA failure rate of 1.6% (n=538). Multivariate logistic regression analysis revealed that liver cirrhosis, hepatocellular carcinoma, poor DAA adherence, and higher hemoglobin A1c were significantly associated with virological failure. XGBoost outperformed the other algorithms and logistic regression models, with an area under the receiver operating characteristic curve of 1.000 in the training dataset and 0.803 in the validation dataset. The top five predictors of treatment failure were HCV RNA, body mass index, α-fetoprotein, platelets, and FIB-4 index. The accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of the XGBoost model (cutoff value=0.5) were 99.5%, 69.7%, 99.9%, 97.4%, and 99.5%, respectively, for the entire dataset. Conclusions Machine learning algorithms effectively provide risk stratification for DAA failure and additional information on the factors associated with DAA failure.

Objective: To analyze the prevalence and the risk factors of fungal sepsis in 25 neonatal intensive care units (NICU) among preterm infants in China, and to provide a basis for preventive strategies of fungal sepsis. Methods: This was a second-analysis of the data from the "reduction of infection in neonatal intensive care units using the evidence-based practice for improving quality" study. The current status of fungal sepsis of the 24 731 preterm infants with the gestational age of <34⁺⁰ weeks, who were admitted to 25 participating NICU within 7 days of birth between May 2015 and April 2018 were retrospectively analyzed. These preterm infants were divided into the fungal sepsis group and the without fungal sepsis group according to whether they developed fungal sepsis to analyze the incidences and the microbiology of fungal sepsis. Chi-square test was used to compare the incidences of fungal sepsis in preterm infants with different gestational ages and birth weights and in different NICU. Multivariate Logistic regression analysis was used to study the outcomes of preterm infants with fungal sepsis, which were further compared with those of preterm infants without fungal sepsis. The 144 preterm infants in the fungal sepsis group were matched with 288 preterm infants in the non-fungal sepsis group by propensity score-matched method. Univariate and multivariate Logistic regression analysis were used to analyze the risk factors of fungal sepsis. Results: In all, 166 (0.7%) of the 24 731 preterm infants developed fungal sepsis, with the gestational age of (29.7±2.0) weeks and the birth weight of (1 300±293) g. The incidence of fungal sepsis increased with decreasing gestational age and birth weight (both P<0.001). The preterm infants with gestational age of <32 weeks accounted for 87.3% (145/166). The incidence of fungal sepsis was 1.0% (117/11 438) in very preterm infants and 2.0% (28/1 401) in extremely preterm infants, and was 1.3% (103/8 060) in very low birth weight infants and 1.7% (21/1 211) in extremely low birth weight infants, respectively. There was no fungal sepsis in 3 NICU, and the incidences in the other 22 NICU ranged from 0.7% (10/1 397) to 2.9% (21/724), with significant statistical difference (P<0.001). The pathogens were mainly Candida (150/166, 90.4%), including 59 cases of Candida albicans and 91 cases of non-Candida albicans, of which Candida parapsilosis was the most common (41 cases). Fungal sepsis was independently associated with increased risk of moderate to severe bronchopulmonary dysplasia (BPD) (adjusted OR 1.52, 95%CI 1.04-2.22, P=0.030) and severe retinopathy of prematurity (ROP) (adjusted OR 2.55, 95%CI 1.12-5.80, P=0.025). Previous broad spectrum antibiotics exposure (adjusted OR=2.50, 95%CI 1.50-4.17, P<0.001), prolonged use of central line (adjusted OR=1.05, 95%CI 1.03-1.08, P<0.001) and previous total parenteral nutrition (TPN) duration (adjusted OR=1.04, 95%CI 1.02-1.06, P<0.001) were all independently associated with increasing risk of fungal sepsis. Conclusions: Candida albicans and Candida parapsilosis are the main pathogens of fungal sepsis among preterm infants in Chinese NICU. Preterm infants with fungal sepsis are at increased risk of moderate to severe BPD and severe ROP. Previous broad spectrum antibiotics exposure, prolonged use of central line and prolonged duration of TPN will increase the risk of fungal sepsis. Ongoing initiatives are needed to reduce fungal sepsis based on these risk factors.
Infant ; Infant, Newborn ; Humans ; Birth Weight ; Intensive Care Units, Neonatal ; Retrospective Studies ; Tertiary Care Centers ; Infant, Extremely Low Birth Weight ; Gestational Age ; Infant, Extremely Premature ; Sepsis/epidemiology* ; Retinopathy of Prematurity/epidemiology* ; Bronchopulmonary Dysplasia/epidemiology*