Search Results (625)

In recent years, data augmentation techniques have become the predominant approach for addressing highly imbalanced classification problems in machine learning. Algorithms such as the Synthetic Minority Over-sampling Technique (SMOTE) and Conditional Tabular Generative Adversarial Network (CTGAN) have proven effective in synthesizing minority class samples. However, these methods often introduce distributional bias and noise, potentially leading to model overfitting, reduced predictive performance, increased computational costs, and elevated cybersecurity risks. To overcome these limitations, we propose a novel architecture, FADEL, which integrates feature-type awareness with a supervised discretization strategy. FADEL introduces a unique feature augmentation ensemble framework that preserves the original data distribution by concurrently processing continuous and discretized features. It dynamically routes these feature sets to their most compatible base models, thereby improving minority class recognition without the need for data-level balancing or augmentation techniques. Experimental results demonstrate that FADEL, solely leveraging feature augmentation without any data augmentation, achieves a recall of 90.8% and a G-mean of 94.5% on the internal test set from Kaohsiung Chang Gung Memorial Hospital in Taiwan. On the external validation set from Kaohsiung Medical University Chung-Ho Memorial Hospital, it maintains a recall of 91.9% and a G-mean of 86.7%. These results outperform conventional ensemble methods trained on CTGAN-balanced datasets, confirming the superior stability, computational efficiency, and cross-institutional generalizability of the FADEL architecture. Altogether, FADEL uses feature augmentation to offer a robust and practical solution to extreme class imbalance, outperforming mainstream data augmentation-based approaches. Full article

This research aims to offer a deep learning-based diagnostic approach for hemorrhagic complications linked to patent ductus arteriosus (PDA) in preterm infants. Utilizing the You Only Look Once (YOLO) algorithm, this research analyzed five key cardiac parameters derived from echocardiographic ultrasonic waves: the left ventricular ejection time (LVET), left ventricular internal dimension at diastole (LVIDd), left ventricular internal dimension at systole (LVIDs), posterior wall thickness at end-systole (HES), and RR interval between two successive R-waves. The proposed ensemble model achieved best-in-class detection accuracies for each parameter, with rates of 97.56% (LVET), 88.69% (LVIDd), 99.50% (LVIDs), 82.29% (HES), and 81.15% (RR interval). Furthermore, assessment of cardiac function using derived indices—end-systolic wall stress (ESWS) and rate-corrected mean velocity of circumferential fiber shortening (mVcfc)—achieved mean accuracy rates of 82.33% and 90.16%, respectively. This approach enables physicians to accurately evaluate cardiac function in preterm infants and facilitates the diagnosis of PDA-related hemorrhagic complications. Full article

As semiconductor devices demand higher input–output density and faster signal transmission, fan-out panel-level packaging has emerged as a promising solution for next-generation electronic systems. However, the hygroscopic nature of epoxy molding compounds raises critical reliability concerns under high-temperature and high-humidity conditions. This study investigates the hygrothermal stress of a single fan-out panel-level package unit through experimental characterization and numerical simulation. Thermal–mechanical analysis was conducted at 100 °C and 260 °C to evaluate the strain behavior of two commercial epoxy molding compounds in granule form after moisture saturation. The coefficient of moisture expansion was calculated by correlating strain variation with moisture uptake obtained under 85 °C and 85% relative humidity, corresponding to moisture sensitivity level 1 conditions. These values were directly considered into a moisture -thermal coupled finite element analysis. The simulation results under reflow conditions demonstrate accurate principal stress and failure location predictions, with stress concentrations primarily observed at the die corners. The results confirm that thermal effects influence stress development more than moisture effects. Finally, a structural sensitivity analysis of the single-package configuration showed that optimizing the thickness of the dies and epoxy molding compound can reduce maximum principal stress by up to 12.4%, providing design insights for improving package-level reliability. Full article

Background/Objectives: Depression is frequently comorbid with obesity. We previously showed that astrocyte-mediated hyperactive ventral hippocampal glutamatergic afferents to the nucleus accumbens determined the exhibition of depression-like behaviors in obese murine models. However, it remains unclear if the metabolic disorder-induced depressive phenotypes and astrocytic maladaptation in the ventral hippocampus (vHPC) could be reversed following the amelioration of key metabolic impairments such as insulin resistance and dyslipidemia. Method: Male mice were fed a high-fat diet (HFD) for 12 weeks, followed by either continued HFD feeding (HFD/HFD group) or a switch to a standard diet for 4 weeks (HFD/SD group). Results: Results showed that HFD/HFD mice displayed not only glucose/lipid metabolic dysfunction, but also depression-like behaviors. In contrast, HFD/SD mice showed improvements in metabolic disorders and depressive phenotypes. Mechanistically, dietary fat reduction restored astrocyte morphology and glutamate transporter expression (GLT-1, GLAST) in the vHPC and suppressed neuroinflammatory signaling, as evidenced by reduced levels of phospho-IKK, TNF-α, IL-1β, and IL-6 in the vHPC. Conclusions: These findings suggest that dietary fat reduction reverses obesity-induced depressive phenotypes, astrocytic deficits, at least in part via suppression of neuroinflammation through the NF-κB signaling pathway. Full article

Background: Acute complicated cholecystitis (ACC) is associated with prolonged hospitalization, increased morbidity, and higher mortality. However, objective imaging-based criteria to guide early clinical decision-making remain limited. This study aimed to develop a predictive scoring system integrating clinical characteristics, laboratory biomarkers, and computed tomography (CT) findings to facilitate the early identification of ACC in the emergency department (ED). Methods: We conducted a retrospective study at an urban tertiary care center in Taiwan, screening 729 patients who presented to the ED with suspected cholecystitis between 1 January 2018 and 31 December 2020. Eligible patients included adults (≥18 years) with a confirmed diagnosis of acute cholecystitis based on the Tokyo Guidelines 2018 (TG18) and who were subsequently admitted for further management. Exclusion criteria included (a) the absence of contrast-enhanced CT imaging, (b) no hospital admission, (c) alternative final diagnosis, and (d) incomplete clinical data. A total of 390 patients met the inclusion criteria. Demographic data, laboratory results, and CT imaging features were analyzed. Logistic regression and decision tree analyses were used to construct predictive models. Results: Among the 390 included patients, 170 had mild, 170 had moderate, and 50 had severe cholecystitis. Key predictors of ACC included gangrenous changes, gallbladder wall attenuation > 80 Hounsfield units, CRP > 3 mg/dL, and WBC > 11,000/μL. A novel scoring system incorporating these variables demonstrated good diagnostic performance, with an area under the curve (AUC) of 0.775 and an optimal cutoff score of ≥2 points. Decision tree analysis similarly identified these four predictors as critical determinants in stratifying disease severity. Conclusions: This CT- and biomarker-based scoring system, alongside a decision tree model, provides a practical and robust tool for the early identification of complicated cholecystitis in the ED. Its implementation may enhance diagnostic accuracy and support timely clinical intervention. Full article

We report a 26-year-old male presenting with a chronic cough and hemoptysis. Imaging revealed a large hypermetabolic mass in the left lower lung with the invasion of adjacent great vessels. A biopsy confirmed sarcomatoid carcinoma, a rare and aggressive form of primary pulmonary sarcoma. Due to vascular involvement, the patient underwent preoperative bronchial artery embolization followed by left pneumonectomy with pulmonary arterioplasty via median sternotomy. Postoperative recovery was uneventful. A two-year follow-up CT showed no recurrence. Primary pulmonary sarcomas are extremely rare, accounting for only 0.013–0.4% of lung malignancies, and are often diagnosed late due to nonspecific symptoms. This case highlights the importance of timely imaging, multidisciplinary planning, and aggressive surgical management in achieving long-term disease control, even in cases with extensive vascular invasion. Full article

Monitoring fetal growth throughout pregnancy is essential for early detection of developmental abnormalities. This study developed a Taiwan-specific fetal growth reference using a web-based data collection platform and polynomial regression modeling. We analyzed ultrasound data from 980 pregnant women, encompassing 8350 prenatal scans, to model six key fetal biometric parameters: abdominal circumference, crown–rump length, estimated fetal weight, head circumference, biparietal diameter, and femur length. Quadratic regression was selected based on a balance of performance and simplicity, with R² values exceeding 0.95 for most parameters. Confidence intervals and real-time anomaly detection were implemented through the platform. The results demonstrate the potential for efficient, population-specific fetal growth monitoring in clinical settings. Full article

Waardenburg syndrome (WS) is a rare genetic disorder that affects both hearing and pigmentation. The wide divergence of WS poses significant diagnostic and management challenges. This study investigated type 4 WS within an underrepresented Mongolian population. Whole-exome sequencing revealed that two unique heterozygous variants were identified in the SOX10 gene: c.393C>G (p.Asn131Lys) in a five-year-old female patient presenting with profound sensorineural hearing loss (SNHL), dystopia canthorum, and a white forelock; and c.535A>T (p.Lys179Ter) in a nine-year-old male patient presenting with profound SNHL, dystopia canthorum, and Hirschsprung’s disease. Temporal bone imaging revealed abnormalities in the inner ear structure in both patients. The genotypic and phenotypic characteristics were meticulously delineated, incorporating the deleterious effects of these variants, as evaluated by multiple predictive tools and the American College of Medical Genetics and Genomics (ACMG) criteria. In addition, structural characterizations were also presented using AlphaFold. The findings of this study contribute valuable genetic data to the limited literature on type 4 WS within this ethnic group and highlight the importance of genetic testing and multidisciplinary care for this rare disorder in settings with limited resources. Full article

Glioblastoma multiforme (GBM) is the most aggressive and common primary brain tumor. Magnetic resonance imaging (MRI) provides detailed visualization of tumor morphology, edema, and necrosis. However, manually segmenting GBM from MRI scans is time-consuming, subjective, and prone to inter-observer variability. Therefore, automated and reliable segmentation methods are crucial for improving diagnostic accuracy. This study employs an image semantic segmentation model to segment brain tumors in MRI scans of GBM patients. The MRI recall images include T1-weighted imaging (T1WI) and fluid-attenuated inversion recovery (FLAIR) sequences. To enhance the performance of the semantic segmentation model, image preprocessing techniques were applied before analyzing and comparing commonly used segmentation models. Additionally, a survival model was constructed using discrete genotype attributes of GBM patients. The results indicate that the DeepLabV3+ model achieved the highest accuracy for semantic segmentation, with an accuracy of 77.9% on T1WI image sequences, while the U-Net model achieved 80.1% accuracy on FLAIR image sequences. Furthermore, in constructing the survival model using a discrete attribute dataset, the dataset was divided into three subsets based on different missing value handling strategies. This study found that replacing missing values with 1 resulted in the highest accuracy, with the Bernoulli Bayesian model and the multinomial Bayesian model achieving an accuracy of 94.74%. This study integrates image preprocessing techniques and semantic segmentation models to improve the accuracy and efficiency of brain tumor segmentation while also developing a highly accurate survival model. The findings aim to assist physicians in saving time and facilitating preliminary diagnosis and analysis. Full article

This study prepared epoxy–clay nanocomposites (ECNs) by incorporating organophilic clays modified with either non-redox cetyltrimethylammonium bromide (CTAB) or redox-active aniline pentamer (AP), then compared their anticorrosion performance on metal substrates in saline environments. The test solution contained 2 wt% alkaline copper quaternary (ACQ) wood preservatives. Cold-rolled steel (CRS) panels coated with the ECNs were evaluated via electrochemical impedance spectroscopy (EIS) in saline media both with and without ACQ. For CRS coated with unmodified epoxy, the Nyquist plot showed impedance dropping from 255 kΩ to 121 kΩ upon adding 2 wt% ACQ—indicating that Cu²⁺ ions accelerate iron oxidation. Introducing 1 wt% CTAB–clay into the epoxy increased impedance from 121 kΩ to 271 kΩ, while 1 wt% AP–clay raised it to 702 kΩ. This improvement arises because the organophilic clay platelets create a more tortuous path for Cu²⁺ and O₂ diffusion, as confirmed by ICP–MS measurements of Cu²⁺ after EIS and oxygen permeability tests (GPA), thereby slowing iron oxidation. Moreover, ECN coatings containing AP–clay outperformed those with CTAB–clay in corrosion resistance, suggesting that AP not only enhances platelet dispersion but also promotes formation of a dense, passive metal oxide layer at the coating–metal interface, as shown by TEM, GPA, and XRD analyses. Finally, accelerated salt-spray exposure following ASTM B-117 yielded corrosion behavior consistent with the EIS results. Full article

Breast cancer is a leading cause of cancer-related mortality worldwide, requiring efficient diagnostic tools for early detection and monitoring. Human epidermal growth factor receptor 2 (HER2) is a key biomarker for breast cancer classification, typically assessed using immunohistochemistry (IHC). However, IHC requires invasive biopsies and time-intensive laboratory procedures. In this study, we present a biosensor integrated with a reusable printed circuit board (PCB) and functionalized glucose test strips designed for rapid and non-invasive HER2 detection in saliva. The biosensor achieved a limit of detection of 10⁻¹⁵ g/mL, 4 to 5 orders of magnitude more sensitive than the enzyme-linked immunosorbent assay (ELISA), with a sensitivity of 95/dec and a response time of 1 s. In addition to HER2, the biosensor also detects cancer antigen 15-3 (CA15-3), another clinically relevant breast cancer biomarker. The CA15-3 test demonstrated an equally low limit of detection, 10⁻¹⁵ g/mL, and a higher sensitivity, 190/dec, further validated using human saliva samples. Clinical validation using 29 saliva samples confirmed our biosensor’s ability to distinguish between healthy, in situ breast cancer, and invasive breast cancer patients. The system, which integrates a Bluetooth Low-Energy (BLE) module, enables remote monitoring, reduces hospital visits, and enhances accessibility for point-of-care and mobile screening applications. This ultra-sensitive, rapid, and portable biosensor can serve as a promising alternative for breast cancer detection and monitoring, particularly in rural and underserved communities. Full article

With the soaring demand for electric power and the limited spinning reserve in the power system in Taiwan, the comprehensive management of both thermal power generation and load demand turns out to be a key to achieving the robustness and sustainability of the power system. This paper aims to design a demand bidding (DB) mechanism to collaborate between customers and suppliers on demand response (DR) to prevent the risks of energy shortage and realize energy conservation. The concurrent integration of the energy, transmission, and reserve capacity markets necessitates a new formulation for determining schedules and marginal prices, which is expected to enhance economic efficiency and reduce transaction costs. To dispatch energy and reserve markets concurrently, a hybrid approach of combining dynamic queuing dispatch (DQD) with direct search method (DSM) is developed to solve the extended economic dispatch (ED) problem. The effectiveness of the proposed approach is validated through three case studies of varying system scales. The impacts of tie-line congestion and area spinning reserve are fully reflected in the area marginal price, thereby facilitating the determination of optimal load reduction and spinning reserve allocation for demand-side management units. The results demonstrated that the multi-area bidding platform proposed in this paper can be used to address issues of congestion between areas, thus improving the economic efficiency and reliability of the day-ahead market system operation. Consequently, this research can serve as a valuable reference for the design of the demand bidding mechanism. Full article

Recessive variants of SLC26A4 are a common cause of hereditary hearing impairment and are responsible for non-syndromic enlarged vestibular aqueducts and Pendred syndrome. Patients with bi-allelic SLC26A4 variants often suffer from fluctuating hearing loss and recurrent vertigo, ultimately leading to severe to profound hearing impairment. However, there are currently no satisfactory prevention or treatment options for this condition. The CRISPR/Cas9 genome-editing technique is a well-known tool for correcting point mutations or manipulating genes and shows potential therapeutic applications for hereditary disorders. In this study, we used the homology-independent targeted integration (HITI) strategy to correct the SLC26A4 c.919-2A>G variant, the most common SLC26A4 variant in the Han Chinese population. Next-generation sequencing was performed to evaluate the editing efficiency of the HITI strategy. The results showed that only 0.15% of the reads successfully exhibited HITI integration, indicating that the c.919-2 region may not be a suitable region for HITI selection. This suggests that other site selection or insertion strategies may be needed to improve the efficiency of correcting the SLC26A4 c.919-2A>G variant. This experience may serve as a valuable reference for other researchers considering CRISPR target design in this region. Full article

(1) Background: The role of air pollutants as risk factors for tinnitus remains unclear. To address this gap in research, we conducted a nationwide retrospective cohort study in Taiwan by integrating patients’ clinical data with daily air quality data to evaluate the environmental risk factors associated with tinnitus. (2) Methods: The Taiwan National Health Research Database (NHIRD) includes medical records for nearly all residents of Taiwan. To assess pollution levels, we used daily air quality data from the Taiwan Environmental Protection Agency regarding SO₂, CO, NO, NO_X, and particulate matter (PM_2.5 and PM₁₀). We merged the NHIRD data with air quality information based on the residents’ locations and the positions of air quality monitoring stations. Pollutant levels were then categorized into quartiles (Q1, Q2, Q3, and Q4). (3) Results: This study included 284,318 subjects. After controlling for covariates, the adjusted HR (95 CI%) for tinnitus increased with increasing SO₂, CO, NO, NO_X, PM_2.5, and PM₁₀ exposure levels, specifically from 1.24 (95 CI% = 1.18, 1.30) to 1.35 (95 CI% = 1.28–1.41); from 1.15 (95 CI% = 1.09, 1.21) to 1.90 (95 CI% = 1.81, 2.00); from 0.86 (95 CI% = 0.82, 0.91) to 1.69 (95 CI% = 1.62, 1.77); from 1.62 (95 CI% = 1.54, 1.71) to 1.69 (95 CI% = 1.60, 1.77); from 0.16 (95 CI% = 0.15, 0.18) to 2.70 (95 CI% = 2.57, 2.84); and from 2.53 (95 CI% = 2.38, 2.69) to 3.58 (95 CI% = 3.39, 3.78), respectively, compared to the Q1 concentrations for all air pollutants. (4) Conclusions: During the 15-year follow-up period, we found a significant positive correlation between air pollutant exposure and the risk of tinnitus. Full article

Water pesticide contamination represents a major threat to ecological systems and public health, particularly in agricultural regions. Although conventional detection methods such as liquid chromatography and electrochemical analysis are highly accurate, they are expensive, require skilled operators, and cannot provide real-time results. This study developed a portable miniaturized electrochemical analysis platform based on cyclic voltammetry (CV) for rapid pesticide detection. The platform was compared with a commercial electrochemical analyzer and yielded similar performance in detecting chlorpyrifos at different concentrations. When ultrapure water was used as the background solution, the total area under the CV curve exhibited a linear correlation (R² = 0.89) with the pesticide concentration, indicating its potential as a characteristic index. When molecularly imprinted polymers were added, the platform achieved a limit of detection of 50 ppm, with the area under the CV curve maintaining a logarithmic linear relationship (R² = 0.98) with the pesticide concentration. These findings confirm the total area under the CV curve as the most reliable characteristic index for pesticide quantification. Overall, the proposed platform offers portability, straightforward operation, cost-effectiveness, and expandability, making it promising for on-site environmental monitoring. By incorporating GPS functionality, the platform can provide real-time pesticide concentration mapping, supporting its use in precision agriculture and water quality management. Full article