Search Results (753)

Background: Increasing regulatory pressure for maritime decarbonization (e.g., IMO CII, FuelEU) drives adoption of low-carbon fuels and prompts reassessment of regional ports’ competitiveness. This study aims to evaluate the economic and environmental viability of rerouting deep-sea container services to regional ports in a decarbonizing world. Methods: A scenario-based analysis is used to evaluate total costs and CO₂ emissions across the entire container shipping supply chain, incorporating deep-sea shipping, port operations, feeder services, and inland rail/road transport. The Port of Liverpool serves as the primary case study for rerouting Asia–Europe services from major ports. Results: Analysis indicates Liverpool’s competitiveness improves with shipping lines’ slow steaming, growth in hinterland shipment volume, reductions in the emission factors of alternative low-carbon fuels, and an increased modal shift to rail matching that of competitor ports (e.g., Southampton). A dual-port strategy, rerouting services to call at both Liverpool and Southampton, shows potential for both economic and environmental benefits. Conclusions: The study concludes that rerouting deep-sea services to regional ports can offer cost and emission advantages under specific operational and market conditions. Findings on factors and conditions influencing competitiveness and the dual-port strategy provide insights for shippers, ports, shipping lines, logistics agents, and policymakers navigating maritime decarbonization. Full article

Adversarial attacks in Natural Language Processing (NLP) present a critical challenge, particularly in sentiment analysis, where subtle input modifications can significantly alter model predictions. In search of more robust defenses against adversarial attacks on sentimental analysis, this research work introduces two novel defense mechanisms: the Lexicon-Based Random Substitute Model (LRSM) and the Word-Variant Voting Model (WVVM). LRSM employs randomized substitutions from a dataset-specific lexicon to generate diverse input variations, disrupting adversarial strategies by introducing unpredictability. Unlike traditional defenses requiring synonym dictionaries or precomputed semantic relationships, LRSM directly substitutes words with random lexicon alternatives, reducing overhead while maintaining robustness. Notably, LRSM not only neutralizes adversarial perturbations but occasionally surpasses the original accuracy by correcting inherent model misclassifications. Building on LRSM, WVVM integrates LRSM, Frequency-Guided Word Substitution (FGWS), and Synonym Random Substitution and Voting (RS&V) in an ensemble framework that adaptively combines their outputs. Logistic Regression (LR) emerged as the optimal ensemble configuration, leveraging its regularization parameters to balance the contributions of individual defenses. WVVM consistently outperformed standalone defenses, demonstrating superior restored accuracy and F1 scores across adversarial scenarios. The proposed defenses were evaluated on two well-known sentiment analysis benchmarks: the IMDB Sentiment Dataset and the Yelp Polarity Dataset. The IMDB dataset, comprising 50,000 labeled movie reviews, and the Yelp Polarity dataset, containing labeled business reviews, provided diverse linguistic challenges for assessing adversarial robustness. Both datasets were tested using 4000 adversarial examples generated by established attacks, including Probability Weighted Word Saliency, TextFooler, and BERT-based Adversarial Examples. WVVM and LRSM demonstrated superior performance in restoring accuracy and F1 scores across both datasets, with WVVM excelling through its ensemble learning framework. LRSM improved restored accuracy from 75.66% to 83.7% when compared to the second-best individual model, RS&V, while the Support Vector Classifier WVVM variation further improved restored accuracy to 93.17%. Logistic Regression WVVM achieved an F1 score of 86.26% compared to 76.80% for RS&V. These findings establish LRSM and WVVM as robust frameworks for defending against adversarial text attacks in sentiment analysis. Full article

Background/Objectives: Conventional ketogenic diets, although effective for weight loss, often contain high total and saturated fat intake, which leads to increased low-density lipoprotein cholesterol (LDL-C). Thus, the Healthy Ketogenic Diet (HKD) was developed to address these concerns. It emphasizes calorie restriction, limiting net carbohydrate intake to 50 g per day, prioritizing unsaturated fats, and reducing saturated fat intake. However, adherence to the HKD remains a challenge in urban, time-constrained environments. Therefore, this pilot randomized controlled trial aimed to investigate the effects of Healthy Ketogenic Diet Ready-To-Eat (HKD-RTE) meals (provided for the first month only) versus HKD alone on weight loss and metabolic parameters among overweight adults. Methods: Multi-ethnic Asian adults (n = 50) with a body mass index (BMI) ≥ 27.5 kg/m² were randomized into the HKD-RTE group (n = 24) and the HKD group (n = 26). Both groups followed the HKD for six months, with the HKD-RTE group receiving HKD-RTE meals during the first month. Five in-person workshops and mobile health coaching through the Nutritionist Buddy Keto app helped to facilitate dietary adherence. The primary outcome was the change in body weight at 6 months. Linear regression was performed on the change from baseline for each continuous outcome, adjusting for demographics and relevant covariates. Logistic regression was performed on binary weight loss ≥ 5%, adjusting for demographics and relevant covariates. Results: In the HKD group, participants’ adherence to the 50 g net carbohydrate target was 15 days, while that in the HKD-RTE group was 19 days over a period of 30 days. Participants’ adherence to calorie targets was 21 days in the HKD group and 23 days in the HKD-RTE. The average compliance with the HKD-RTE meals provided in the HKD-RTE group was 55%. The HKD-RTE group experienced a greater percentage weight loss at 1 month (−4.8 ± 3.0% vs. −1.8 ± 6.2%), although this was not statistically significant. This trend continued up to 6 months, with the HKD-RTE group showing a greater percentage weight reduction (−8.6 ± 6.8% vs. −3.9 ± 8.6%; p = 0.092). At 6 months, the HKD-RTE group had a greater reduction in total cholesterol (−0.54 ± 0.76 mmol/L vs. −0.05 ± 0.56 mmol/L; p = 0.283) and LDL-C (−0.43 ± 0.67 mmol/L vs. −0.03 ± 0.52 mmol/L; p = 0.374) compared to the HKD group. Additionally, the HKD-RTE group exhibited greater reductions in systolic blood pressure (−8.3 ± 9.7 mmHg vs. −5.3 ± 11.0 mmHg), diastolic blood pressure (−7.7 ± 8.8 mmHg vs. −2.0 ± 7.0 mmHg), and HbA1c (−0.3 ± 0.5% vs. −0.1 ± 0.4%) than the HKD group (not statistically significant for any). Conclusions: Both HKD-RTE and HKD led to weight loss and improved metabolic profiles. The HKD-RTE group tended to show more favorable outcomes. Short-term HKD-RTE meal provision may enhance initial weight loss, with sustained long-term effects. Full article

Seafood processing byproducts (SB) such as bones and skin can be safely used as food ingredients to increase profitability for the seafood sector and provide nutritional value. An online survey of 716 US adult seafood consumers was conducted to explore SB trial intent, responsiveness to health and safety information, and associated elicited emotions (nine-point Likert scale). Consumers’ SB-elicited emotions were defined as those changing in reported intensity (from a baseline condition) after the delivery of SB-related information (dependent t-tests). As criteria for practical significance, a raw mean difference of >0.2 units was used, and Cohen’s d values were used to classify effect sizes as small, medium, or large. Differences in willingness-to-try, responsiveness to safety and health information, and SB-elicited emotions were found based on self-reported gender and race, with males and Hispanics expressing more openness to SB consumption. SB-elicited emotions were then used to model consumers’ willingness-to-try foods containing SB via logistic regression modeling. Traditional stepwise variable selection was compared to variable selection using raw mean difference > 0.2 units and Cohen’s d > 0.50 constraints for SB-elicited emotions. Resulting models indicated that extrinsic information considered at the point of decision-making determined which emotions were relevant to the response. These new approaches yielded models with increased Akaike Information Criterion (AIC) values (lower values indicate better model fit) but could provide simpler and more practically meaningful models for understanding which emotions drive consumption decisions. Full article

The estimation of runout distances for long return period avalanches is vital in zoning schemes for mountainous countries. There are two broad methods to estimate snow avalanche runout distance. One involves the use of a physical model to calculate speeds along the incline, with runout distance determined when the speed drops to zero. The second method, which is used here, is based on empirical or statistical models from databases of extreme runout for a given mountain range or area. The second method has been used for more than 40 years with diverse datasets collected from North America and Europe. The primary reason for choosing the method used here is that it is independent of physical models such as avalanche dynamics, which allows comparisons between methods. In this paper, data from diverse datasets are analyzed to explain the relation between them to give an overall view of the meaning of the data. Runout is formulated from nine different datasets and 738 values of extreme runout, mostly with average return periods of about 100 years. Each dataset was initially fit to 65 probability density functions (pdf) using five goodness-of-fit tests. Detailed discussion and analysis are presented for a set of extreme value distributions (Gumbel, Frechet, Weibull). Two distributions had exemplary results in terms of goodness of fit: the generalized logistic (GLO) and the generalized extreme value (GEV) distributions. Considerations included both the goodness-of-fit and the heaviness of the tail, of which the latter is important in engineering decisions. The results showed that, generally, the GLO has a heavier tail. Our paper is the first to compare median extreme runout distances, the first to compare exceedance probability of extreme runout, and the first to analyze many probability distributions for a diverse set of datasets rigorously using five goodness-of-fit tests. Previous papers contained analysis mostly for the Gumbel distribution using only one goodness-of-fit test. Given that climate change is in effect, consideration of stationarity of the distributions is considered. Based on studies of climate change and avalanches, thus far, it has been suggested that stationarity should be a reasonable assumption for the extreme avalanche data considered. Full article

Background/Objectives: With the rising trend of out-of-home dining in China, the use of iodized salt (IS) in eating-out venues plays a key role in preventing iodine deficiency disorders (IDDs). However, the coverage rate of iodized salt (CRIS) and the utilization rate of adequately iodized salt (URAIS) in these venues in China remain underexplored, potentially undermining IDD prevention strategies. This study aims to assess the CRIS and URAIS in such venues across China and identify the factors influencing their prevalence. Methods: From 2021 to 2024, a nationwide cross-sectional study was conducted in China, involving 19,346 venues. A 50 g sample of cooking salt was collected from each venue, and the iodine content was measured. The CRIS and URAIS were calculated, and associations with various factors were assessed using Chi-square tests, the Cochran–Armitage trend test, and multivariate logistic regression. Results: Of the 19,346 samples, 18,519 tested positive for IS, and 17,588 contained adequately iodized salt (AIS), resulting in a CRIS of 95.7% and a URAIS of 90.9%. Significant regional differences were found, with coastal areas showing a lower CRIS and URAIS than inland areas (87.0% vs. 97.8%; 81.0% vs. 93.2%) and urbanized areas having lower rates compared to less urbanized areas (94.1% vs. 97.3%; 88.9% vs. 92.9%). Higher per capita income was associated with a lower CRIS and URAIS (Z = −19.72, p < 0.0001; Z = −13.85, p < 0.0001). Lower per capita income (OR = 3.24, OR = 1.36, p < 0.0001), inland areas (OR = 4.14, OR = 2.68, p < 0.0001), and mountainous areas (OR = 2.48, OR = 1.27, p < 0.0001) were associated with a higher likelihood of IS and AIS use. Conclusions: While the CRIS and URAIS in dining venues meet national standards, regional disparities persist, particularly in coastal, plain, and economically advanced areas. Strengthening regulatory oversight and public education on iodized salt’s health benefits is essential. Full article

Inefficient inland repositioning of empty containers between depots remains a persistent challenge in container logistics, contributing significantly to unnecessary truck movements, elevated operational costs, and increased CO₂ emissions. Acknowledging the importance of this problem, a large amount of relevant literature has appeared. The objective of this paper is to track the empty container flow between ports, empty depots, inland terminals, and customer premises. Additionally, it aims to simulate and assess CO₂ emissions, capturing the dynamic interactions between different agents. In this study, agent-based modeling (ABM) was proposed to simulate the empty container movements with an emphasis on inland transportation. ABM is an emerging approach that is increasingly used to simulate complex economic systems and artificial market behaviours. NetLogo was used to incorporate real-world geographic data and quantify CO₂ emissions based on truckload status and to evaluate the other operational aspects. Behavior Space was also utilized to systematically conduct multiple simulation experiments, varying parameters to analyze different scenarios. The results of the study show that customer demand frequency plays a crucial role in system efficiency, affecting container availability and logistical tension. Full article

Introduction: As in other crises, during COVID-19 pandemic, journalists were under immense pressure to report precise scientific information in a timely manner, which may have had a negative influence on their mental health. There could be an association between the digital health literacy of journalists and their mental health. The aim of this article was to explore the association between digital health literacy and burnout and depression among journalists in Serbia. Methods: A cross-sectional study was conducted involving a total of 180 journalists working on television with national coverage in Serbia. The main research instrument used was a questionnaire with four sections containing personal demographic information, the Digital Health Literacy Instrument, the Maslach Burnout Inventory-Human Services Survey, and the Beck Depression Inventory. Results: A total of 30% participants were found to have high levels of burnout on the emotional exhaustion (EE) subscale. On the depersonalization (DP) subscale, 10.6% experienced high levels of burnout. On the personal accomplishment (PA) subscale, 38.3% of participants faced high levels of burnout. Multivariate logistic regression analyses showed the association between high burnout on the EE scale and health status (OR: 0.597, 95% CI: 0.375–0.952) and protecting privacy (OR: 0.522, 95% CI: 0.311–0.875). Multivariate logistic regression analysis showed the association between high burnout on the PA scale and information searching (OR: 0.255, 95% CI: 0.124–0.526), sex (OR: 2.594, 95% CI: 1.007–6.68), socioeconomic status (OR: 2.282, 95% CI: 1.133–4.595), and alcohol consumption (OR: 2.188, 95% CI: 1.004–4.769). Multivariate logistic regression analysis showed associations between depression and sex (OR: 0.180, 95% CI: 0.059–0.548), health status (OR: 0.316, 95% CI: 0.160–0.626), the use of anti-anxiety medications (OR: 7.303, 95% CI: 3.167–16.840), information searching (OR: 0.432, 95% CI: 0.191–0.981), and protecting privacy (OR: 0.443, 95% CI: 0.233–0.841). Conclusions: Our study showed a negative association between different domains of burnout, depression, and scores on protecting privacy and information searching scales. Full article

Introduction: Delirium is a common and debilitating clinical complication among ICU patients. Despite the prevalence of this condition, there are insufficient data to support or refute the routine use of atypical antipsychotics since the existing evidence remains sparse and inconclusive. The objective of the present study was to evaluate whether pre-ICU administration of the atypical antipsychotic olanzapine is associated with a differential time to delirium resolution relative to the control condition. Methods: In this emulated clinical trial, we utilized the MIMIC-IV v3.1 database, which contains deidentified health records from approximately 65,000 ICU patients, to derive a cohort of patients with a positive delirium screening within 24 h of ICU admission. We exluded patients who received any antipsychotic other than olanzapine prior to ICU admission. We performed propensity score matching using logistic regression and nearest-neighbor matching (1:1, caliper = 0.2) to balance covariates between the olanzapine and control groups. The primary outcome was time to delirium resolution, defined as the first negative delirium assessment. A Cox proportional hazards model, adjusted for multiple covariates and incorporating age as a time-dependent variable, was used to examine the association between olanzapine use and delirium resolution. Interaction terms were included to evaluate effect modification by age and gender. Results: A total of 5070 patients with a positive delirium screening within 24 h and no exposure to other antipsychotics met the eligibility criteria; 421 olanzapine users were matched to 421 controls using propensity score matching. Covariate balance was achieved (all standardized mean differences < 0.1), and no multicollinearity was detected (all VIFs < 2). Pre-ICU olanzapine use was associated with a 27% decrease in the likelihood of delirium resolution (HR = 0.73; 95% CI: 0.63–0.86; p < 0.001). A significant interaction with age indicated that the negative impact of olanzapine on delirium resolution increased with advancing age (HR = 1.0024 per unit of age × log(time), p = 0.023), translating to a 2.4% increase in the risk of prolonged delirium resolution for every 10-year increase in age per log(time). There was no modification of the association according to gender. Discussion: The negative effect of olanzapine on ICU delirium resolution, particularly among the elderly, presented in this study is in line with the results of our earlier study showing a negative effect (i.e., prolonged ICU stay) among patients receiving quetiapine relative to both control and haloperidol conditions. Distinctly strong anticholinergic effects of both olanzapine and quetiapine relative to the other antipsychotic agents may be driving the identified negative outcomes. Conclusions: Results of this emulated clinical trial do not support the use of olanzapine for the treatment of ICU delirium because the agent prolongs time to resolution of the condition. Full article

Perinatal depression is one of the most common mental illnesses in women. The aim of this study was to assess the association of life stressors, perceived stress, obstetric and neonatal complications, and depressive symptoms in the early postpartum period and to compare these variables in two groups of women (preterm and term deliveries). Methods: A case–control study was conducted among 300 women who gave birth in 2019 at the University of Szeged. Cases included women with preterm deliveries (<37 weeks, n = 100), and the controls included women with term deliveries (≥37 weeks, n = 200). Data were collected during postpartum hospital stays through a self-administered questionnaire (containing validated questionnaires: the Holmes–Rahe Life Stress Inventory, the Perceived Stress Scale (PSS-14), and the Edinburgh Postnatal Depression Scale (EPDS)) and the medical records of women and newborns. A descriptive statistical analysis and logistic regression were used to identify predictors of high EPDS scores (≥10). Results: Perceived stress levels were significantly higher among cases than controls (p < 0.001). Higher perceived stress was associated with a higher risk of depression in cases (OR: 1.31, 95% CI: 1.17–1.48, p < 0.001) and controls (OR: 1.33, 95% CI: 1.21–1.45, p < 0.001), too. Newborn complications were associated with an increased perinatal depression risk in the controls (OR: 2.48, 95% CI: 1.05–5.91; p = 0.039) but not in the cases (OR: 2.79, 95% CI: 0.79–9.85; p = 0.111). It is supposed that premature birth was stressful itself, and women with preterm babies were less sensitive to any complications occurring in their newborns compared to women with term newborns. Neither maternal age, education, nor obstetric complications predicted depressive symptoms. Conclusions: Our findings highlight the impact of maternal perceived stress and newborns’ health status on the risk of developing depression during the early postpartum period. These results emphasize the need for ongoing screening and follow-up measures, especially for women with higher EPDS scores. Full article

Fruit is typically transported in stacked packaging units, where external packaging constraints play a critical role in influencing mechanical damage during transit. This study primarily investigated the effects of external packaging constraints on vibration-induced damage and response vibration in ‘Huangguan’ pears (Pyrus bretschneideri Rehd. ‘Huangguan’). Three external packaging constraint types—free constraint, elastic constraint, and fixed constraint—were applied to a two-layer stacked packaging system to limit vertical movement. The pears inside the containers were divided by a corrugated paperboard. Vibration excitation was simulated using the ASTM D4169 spectrum at three vibration levels. Damage indicators, including damage area, flesh firmness, respiratory rate, weight loss, titratable acidity, ascorbic acid, and tissue microstructure, were analyzed after vibration experiments. The results demonstrated that external packaging constraint type significantly affects the mechanical damage of ‘Huangguan’ pears, with damage severity being closely related to constraint strength. Comprehensive analysis revealed that the most severe damage occurred under free constraint, while the least damage was observed under fixed constraint. Stacking position also influenced damage, as pears on the top layer exhibited more severe damage compared to those on the bottom layer. The response vibration results aligned with the observed damage patterns. SEM analysis further revealed that vibration disrupted the tissue microstructure and damaged stone cells, which decreased in number and even disappeared at higher vibration levels. This study provides valuable insights for improving postharvest transport packaging designs and minimizing fruit loss during logistics. Full article

The global port and maritime industry is experiencing a new paradigm shift known as the artificial intelligence transformation (AX). Thus, domestic container-terminal companies should focus beyond mere automation to a paradigm shift in AI that encompasses operational strategy, organizational structure, system, and human resource management. This study proposes a resilience-based AX strategy and implementation system that allows domestic container-terminal companies to proactively respond to the upcoming changes in the global supply chain, thus securing sustainable competitiveness. In particular, we aim to design an AI-based governance model to establish a trust-based logistics supply chain (trust value chain). As a research method, the core risk factors of AX processes were scientifically identified via text-mining and fault-tree analysis, and a step-by-step execution strategy was established by applying a backcasting technique based on scenario planning. Additionally, by integrating social control theory with new governance theory, we designed a flexible, adaptable, and resilience-oriented AI governance system. The results of this study suggest that the AI paradigm shift should be promoted by enhancing the risk resilience, trust, and recovery of organizations. By suggesting AX strategies and policy as well as institutional improvement directions that embed resilience to secure the sustainable competitiveness of AI-based smart ports in Korea, this study serves as a basis for establishing strategies for the domestic container-terminal industry and for constructing a global leading model. Full article

Background: Gliomas represent the most prevalent and aggressive primary brain tumors, requiring precise classification to guide treatment strategies and improve patient outcomes. Purpose: This study aimed to develop and evaluate a machine learning-driven approach for glioma classification by identifying the most relevant genetic and clinical biomarkers while demonstrating clinical utility. Methods: A dataset from The Cancer Genome Atlas (TCGA) containing 23 features was analyzed using an integrative approach combining Boruta, Least Absolute Shrinkage and Selection Operator (LASSO), and SHapley Additive exPlanations (SHAP) for feature selection. The refined feature set was used to train four machine learning models: Random Forest, Support Vector Machine, XGBoost, and Logistic Regression. Comprehensive evaluation included class distribution analysis, calibration assessment, and decision curve analysis. Results: The feature selection approach identified 13 key predictors, including IDH1, TP53, ATRX, PTEN, NF1, EGFR, NOTCH1, PIK3R1, MUC16, CIC mutations, along with Age at Diagnosis and race. XGBoost achieved the highest AUC (0.93), while Logistic Regression recorded the highest testing accuracy (88.09%). Class distribution analysis revealed excellent GBM detection (Average Precision 0.840–0.880) with minimal false negatives (5–7 cases). Calibration analysis demonstrated reliable probability estimates (Brier scores 0.103–0.124), and decision curve analysis confirmed substantial clinical utility with net benefit values of 0.36–0.39 across clinically relevant thresholds. Conclusions: The integration of feature selection techniques with machine learning models enhances diagnostic precision, interpretability, and clinical utility in glioma classification, providing a clinically ready framework that bridges computational predictions with evidence-based medical decision-making. Full article

Background: The Port of Santa Marta, located on Colombia’s northern Caribbean coast, plays a vital role in the country’s maritime trade, particularly in the export of agricultural and perishable goods. This raises the question: how competitive is Santa Marta’s container terminal compared to national and regional ports, and what strategic factors shape its performance within the Colombia and Latin American maritime logistics system? Methods: This study evaluates the port’s competitiveness by applying Porter’s Extended Diamond Model. A mixed-methods ap-proach was employed, combining structured surveys and interviews with port stakeholders and operational data analysis. A competitiveness matrix was developed and examined using standardized residuals and L1 regression to identify critical performance gaps and strengths. Results: The analysis reveals several competitive advantages, including the port’s strategic location, natural deep-water access, and advanced infrastructure for refrigerated cargo. It also benefits from skilled labour and proximity to global shipping routes, such as the Panama Canal. Nonetheless, challenges remain in storage capacity, limited road connectivity, and insufficient public investment in hinterland infrastructure. Conclusions: While the Port of Santa Marta shows strong maritime capabilities and spe-cialized services, addressing its land-side and institutional constraints is essential for positioning it as a resilient, competitive logistics hub in the Latin American and Caribbean region. Full article

Background: Opportunistic screening is one major screening approach for esophageal squamous cell carcinoma (ESCC). We aimed to develop a score-based risk stratification model to assess the risk of ESCC and precancerous lesions in opportunistic screening and to validate it in an external population. Methods: The study was a secondary analysis of a published esophageal cancer screening trial. The trial was conducted in 39 secondary or tertiary hospitals in China, with 14,597 individuals including 71 high-grade intraepithelial neoplasia (HGIN) and 182 ESCC, enrolled for opportunistic screening. Additionally, questionnaires and endoscopy were performed. The primary outcome was histology-confirmed high-grade esophageal lesions, including HGIN and ESCC. The predictors were selected using univariable and multivariable logistic regression. Model performance was primarily measured with the area under the receiver operating characteristic curve (AUROC). Results: The score-based prediction model contained 8 variables on a 21-point scale. The model demonstrated an AUROC of 0.833 (95% CI, 0.803–0.862) and 0.828 (95% CI, 0.793–0.864) for detecting high-grade lesions in the training and validation cohorts, respectively. Using the cut-off score determined in the training cohort (≥9), the sensitivity reached 70.0% (95% CI, 50.6–85.3%), 81.3% (95% CI, 63.6–92.8%), and 81.1% (95% CI, 64.9–92.0%) in the validation cohort for detecting HGIN, early ESCC, and advanced ESCC, respectively, at a specificity of 76.4% (95%CI, 75.4–77.4%). The score-based model exhibited satisfactory calibration in the calibration plots. The model could result in 75.6% fewer individuals subjected to endoscopy. Conclusions: This score-based model demonstrated superior discrimination for esophageal high-grade lesions. It has the potential to inform referral decisions in an opportunistic screening setting. Full article