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.

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:A two-dimensional (2D) in-house-built scintillator detection system (SDS) was utilized for quality assurance of the active spot scanning proton and heavy ion accelerator, aiming to establish a rapid detection method and provide reference for the quality of proton and heavy ion beam (spot position, spot size, virtual source-to-axis distance, profile depth dose distribution and beam range).Methods:The SDS consisted of a ceramic gadolinium-sulfoxylate phosphor-scintillating screen, a mirror and a commercial digital camera. The dose distribution image was obtained based on scintillator, mirror reflector and optical signal acquisition device to transform the proton and heavy ion beam into visible light through sulfur gadolinium oxide scintillator and collect visible light information to meet the clinical requirements for the quality of proton and heavy ion beam.Results:The deviation of spot position measured by multifilament proportional chamber and the SDS was less than 1mm. The differences of beam spot size measured by multifilament proportional chamber and the SDS were (1.40±0.59)mm for protons, and (0.5±0.08)mm for carbon ions. For 429.25MeV/u carbon, the virtual source-to-axis distance (V SAD) at the x-and y-axes was 751.8cm and 805.6cm. And difference between physical distance and virtual source-to-axis distance was less than 1%. The range of 287.5MeV/u carbon measured by SDS was 160mm. Conclusions:The in-house-built scintillator detector can measure beam spot position and size, virtual source, depth distribution curve and range, which can be used as an effective tool for quality assurance control of proton and heavy ion therapy.

Human pluripotent stem cells (hPSCs), including embryonic stem cells (ESCs) and induced PSCs (iPSCs), represent potentially unlimited cell sources for clinical applications. Previous studies have suggested that hPSCs may benefit from immune privilege and limited immunogenicity, as reflected by the reduced expression of major histocompatibility complex class-related molecules. Here we investigated the global immune-related gene expression profiles of human ESCs, hiPSCs and somatic cells and identified candidate immune-related genes that may alter their immunogenicity. The expression levels of global immune-related genes were determined by comparing undifferentiated and differentiated stem cells and three types of human somatic cells: dermal papilla cells, ovarian granulosa cells and foreskin fibroblast cells. We identified the differentially expressed genes CD24, GATA3, PROM1, THBS2, LY96, IFIT3, CXCR4, IL1R1, FGFR3, IDO1 and KDR, which overlapped with selected immune-related gene lists. In further analyses, mammalian target of rapamycin complex (mTOR) signaling was investigated in the differentiated stem cells following treatment with rapamycin and lentiviral transduction with specific short-hairpin RNAs. We found that the inhibition of mTOR signal pathways significantly downregulated the immunogenicity of differentiated stem cells. We also tested the immune responses induced in differentiated stem cells by mixed lymphocyte reactions. We found that CD24- and GATA3-deficient differentiated stem cells including neural lineage cells had limited abilities to activate human lymphocytes. By analyzing the transcriptome signature of immune-related genes, we observed a tendency of the hPSCs to differentiate toward an immune cell phenotype. Taken together, these data identify candidate immune-related genes that might constitute valuable targets for clinical applications.
Embryonic Stem Cells ; Female ; Fibroblasts ; Foreskin ; Granulosa Cells ; Humans* ; Induced Pluripotent Stem Cells ; Lymphocyte Culture Test, Mixed ; Lymphocytes* ; Major Histocompatibility Complex ; Phenotype ; Pluripotent Stem Cells* ; RNA ; Signal Transduction ; Sirolimus ; Stem Cells ; Transcriptome

BACKGROUND/AIMS: Spontaneous bacterial peritonitis (SBP) contributes to poorer short-term mortality in cirrhotic patients with ascites. However, it is unknown how long the effect of the first SBP event persists in these patients. METHODS: The National Health Insurance Database, derived from the Taiwan National Health Insurance Program, was used to identify and enroll 7,892 cirrhotic patients with ascites who were hospitalized between January 1 and December 31, 2007. All patients were free from episodes of SBP from 1996 to 2006. RESULTS: The study included 1,176 patients with SBP. The overall 30-day, 90-day, 1-year, and 3-year mortality rates in this group were 21.8%, 38.9%, 57.5%, and 73.4%, respectively. The overall 30-day, 90-day, 1-year, and 3-year mortality rates in the non-SBP group were 15.7%, 32.5%, 53.3%, and 72.5%, respectively. After adjusting for gender, age, and other medical comorbidities, the adjusted hazard ratios of SBP for 30-day, 30- to 90-day, 90-day to 1-year, and 1- to 3-year mortality were 1.49 (95% confidence interval [CI], 1.30 to 1.71), 1.19 (95% CI, 1.02 to 1.38), 1.04 (95% CI, 0.90 to 1.20), and 0.90 (95% CI, 0.77 to 1.05), respectively, compared with the non-SBP group. CONCLUSIONS: The effect of SBP on the mortality of cirrhotic patients with ascites disappeared in those surviving more than 90 days after the first SBP event.
Ascites* ; Comorbidity ; Fibrosis ; Humans ; Mortality* ; National Health Programs ; Peritonitis* ; Taiwan*

The use of pneumococcal vaccine plays an important role for prevention of invasive pneumococcal disease (IPD). However, introducing the pneumococcal vaccine into the national immunization program (NIP) is complex and costly. The strategy of progressively integrating the pneumococcal conjugate vaccine (PCV) into the NIP in Taiwan provides valuable experience for policy makers. The 7-valent PCV (PCV7) was first available in Taiwan in late 2005. PCV7 was first provided free to children with underlying diseases, those in vulnerable socioeconomic status, and those with inadequate health care resources. The catch-up immunization program with the 13-valent PCV was launched in 2013 and the national pneumococcal immunization program was implemented in 2015. Children aged 2-5 years had the highest incidence of IPD among pediatric population in Taiwan. Although the incidence of IPD caused by PCV7 serotypes has declined, the overall incidence of IPD remained high in the context of PCV7 use in the private sector. A surge of IPD caused by serotype 19A occurred, accounting for 53.6% of IPD cases among children aged < or = 5 years in 2011-2012. After the implementation of the national pneumococcal immunization program, serogroup 15 has become the leading serogroup for IPD in children. Continued surveillance is necessary to monitor the serotype epidemiology in Taiwan.
Administrative Personnel ; Child ; Delivery of Health Care ; Epidemiology ; Humans ; Immunization Programs ; Incidence ; Pneumococcal Vaccines* ; Private Sector ; Social Class ; Taiwan*