Search Results (455)

Flocs, visible particles formed in sugar-sweetened beverages, reduce clarity and consumer acceptance of products. Their presence can be caused not only by different types of trace impurities in the sugar but also by interactions among beverage components. In this review, scientific reports on acid beverage flocs (ABFs) and alcohol flocs are summarized, the main pathways for their formation are described, and practical options for detecting them and preventing their formation in beverages are compiled. Using Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 and related guidance, literature searches of Scopus, Web of Science (WoS), PubMed, Food Science and Technology Abstracts (FSTA), CAB Abstracts, and International Commission for Uniform Methods of Sugar Analysis (ICUMSA) resulted in the inclusion of 56 studies. In various types of beverages, complexes formed between proteins (Ps) and polyphenols (PPs) often initiate haze and floc formation, while polysaccharides (dextran, pectin, and starch), silica or silicates, and inorganic ions influence charge balance, particle bridging, and floc growth rate. Ethanol in alcohol beverages can further destabilize colloids and promote aggregation. For beet sugars, saponin–protein interactions are a likely pathway for the formation of ABF, but the available evidence is not consistent. In cane sugars, the reported roles of proteins, polysaccharides, silica, and starch in floc formation vary considerably between studies. For quality assurance, ICUMSA floc tests (GS2-40 and GS2-44) should be complemented by turbidity or haze measurement and colloid characterization such as light scattering, ζ–potential, and infrared IR-based analytical methods supported by chemometrics. Risk mitigation works best as a two-level strategy that combines impurity removal during sugar production and stabilization steps in beverage formulation and storage, including the use of clarification agents and control of pH, temperature, ionic strength, and oxygen exposure. Standardized reporting and validation of rapid predictors against ICUMSA benchmarks remain essential. Full article

Heat stress is a major challenge for broiler production worldwide and is expected to intensify with more frequent heatwaves. This study focuses on mechanically ventilated broiler houses in South Korea, where heatwaves have become increasingly frequent. Three regression meta-models were developed to predict the indoor temperature–humidity index (THI) directly from weather forecast data, using simulated results from a validated building energy simulation (BES) model. A TRNSYS-based BES model was validated against field measurements from four rearing cycles in a commercial broiler house (RMSE 1.31–2.16; MAPE < 2.00%). Using 3072 simulation cases that combined multiple sites, thermal-transmittance levels, cooling conditions, building sizes, and broiler body weights, three regression meta-model approaches were evaluated: a condition-specific regression meta-model for each condition set, a unified regression meta-model with categorical predictors, and a single variable meta-model using only external THI as a predictor. All three showed strong predictive performance, and the unified regression meta-model achieved R² = 0.978, RMSE = 0.817, and MAPE = 0.829, providing the best balance between accuracy and simplicity. This unified model offers a practical tool to link weather forecasts with broiler-house design and environmental-control decisions for heat-stress risk management. Full article

Background and Objectives: Endometrial polyps are common in women presenting with subfertility, yet uncertainty persists regarding which lesions warrant removal and how best to integrate hysteroscopic management with contemporary fertility treatment pathways. This narrative review synthesizes current evidence on pathophysiological mechanisms, diagnostic approaches, fertility outcomes, and practical clinical management for women under 40 years of age. Materials and Methods: PubMed/MEDLINE, Embase, Scopus, Web of Science, and the Cochrane Library were searched for English-language human studies published between January 2005 and December 2025. From 2352 records identified, 83 studies were included after screening of 1517 unique records (7 randomized controlled trials, 12 systematic reviews/meta-analyses, 14 prospective cohort studies, 31 retrospective cohort studies, 5 case–control and other study designs, 11 narrative reviews and supporting evidence studies, 1 clinical guideline, and 2 targeted 2025 additions). This structured narrative review employed a systematic search strategy to ensure comprehensive coverage, with evidence synthesized thematically in accordance with the SANRA guidelines. No formal risk-of-bias assessment or pre-registered protocol was used. Results: Across treatment modalities, hysteroscopic polypectomy was consistently associated with improved fertility outcomes. The landmark Pérez-Medina randomized trial reported a relative risk of 2.1 (95% CI 1.5–2.9) for pregnancy after polypectomy before intrauterine insemination. For IVF/ICSI, reported clinical pregnancy rates after polypectomy range from 53–72% and live birth rates from 43–66%. Proposed mechanisms include mechanical interference, chronic inflammation with cytokine dysregulation, altered endometrial receptivity (including dysregulation of HOXA10/HOXA11), and impaired decidualization. Conclusions: Current evidence supports hysteroscopic polypectomy as an effective intervention to improve fertility outcomes in subfertile women with endometrial polyps, particularly prior to intrauterine insemination. For IVF/ICSI, polypectomy of documented polyps appears beneficial, though evidence quality is moderate and heterogeneity exists across studies. It is critical to distinguish routine screening hysteroscopy before IVF from targeted polypectomy when a polyp has been documented. Contemporary guidance (including the 2024 SOGC guideline) favors polypectomy for symptomatic polyps and those that meet specific clinical criteria; for small asymptomatic polyps (<10 mm), individualized decision-making is appropriate, given limited direct evidence and the potential for spontaneous regression. Future research should clarify molecular predictors of polyp-associated infertility, optimal timing relative to fertility treatment, and long-term reproductive outcomes. Full article

Introduction: University students are vulnerable to psychological distress due to the academic and social demands of this life stage. Mindfulness and self-compassion are effective and adaptable strategies in an academic environment that promote emotional regulation and psychological well-being. This study aims to conduct a systematic review and meta-analysis to evaluate the combined impact of mindfulness- and self-compassion-based interventions (MBSCIs) on psychological distress. It will also analyze their role as predictors of therapeutic change, as well as the moderating influence of sociodemographic and contextual factors. Method: We systematically searched PubMed, Scopus and Web of Science for randomized controlled trials (RCTs) and single-group pre-post trials investigating the effect of MBSCI on anxiety, depression and stress in college students. Studies were combined using the inverse variance method in a random effects model. Additional subgroup and meta-regression analyses were performed, and risk of bias was assessed. The review was pre-registered (PROSPERO registration number: CRD420251003822). Results: Our review included 49 studies with a total of 5043 participants (3721 in the intervention group, and 1322 in the control group). The results provide relevant evidence on the efficacy of MBSCI in the university population, especially in reducing symptoms of stress, anxiety, and depression. The effect sizes observed were moderate-to-large for stress and small-to-moderate for anxiety and depression, supporting their clinical usefulness in university educational settings. However, these findings should be interpreted with caution, as no included study achieved low risk of bias, and heterogeneity was moderate-to-high across most outcomes. Conclusions: The results suggest that MBSCI could alleviate psychological distress in university students. However, these results are limited by some methodological issues (risk of bias, heterogeneity, lack of follow-ups, poor standardization). It would be advisable to integrate these practices into the university curriculum as workshops or complementary activities. Further studies are needed to confirm their effectiveness and explore sustained effects and differences according to individual characteristics. Full article

Background: Carotid intima-media thickness (IMT) is a well-established surrogate marker of subclinical atherosclerosis and a predictor of cardiovascular risk. Carotenoids, particularly lycopene and β-carotene, have been proposed as protective antioxidants against vascular damage, but evidence from population-based studies is inconsistent. Objective: We aim to perform a systematic review and meta-analysis of the associations between circulating levels of lycopene and β-carotene and carotid IMT in the general adult population, including potential sex-specific effects. Methods: A systematic search was conducted in PubMed, Scopus, and Web of Science up to March 2025, following PRISMA guidelines (PROSPERO registration: CRD420251003810). Observational and experimental studies reporting cross-sectional associations between plasma carotenoids and IMT were included. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated via random effects models. Subgroup and meta-regression analyses explored potential modifiers, including sex and cardiovascular risk factors. Results: Thirteen studies (n = 9131; mean age 46.4–71.6 years) met the inclusion criteria, eight of which were eligible for meta-analysis. High circulating lycopene levels were significantly associated with low IMT (pooled OR = 0.70; 95% CI: 0.59–0.84; I² = 65.7%). The association was stronger in men (OR = 0.62; 95% CI: 0.45–0.84) than in women (OR = 0.74; 95% CI: 0.58–0.95). In contrast, β-carotene was only marginally associated with IMT (pooled OR = 0.96; 95% CI: 0.92–0.99; I² = 72.6%). Meta-regression suggested that systolic blood pressure modified the lycopene-IMT relationship, whereas body mass index and low-density lipoprotein cholesterol influenced the β-carotene-IMT association. No evidence of publication bias was found. Conclusions: Increased serum lycopene concentrations, and to a lesser extent β-carotene concentrations, are inversely associated with carotid IMT, suggesting a protective role of lycopene in vascular health. The effect appears more pronounced in men, highlighting potential sex-specific differences in carotenoid metabolism and cardiovascular risk modulation. Full article

Missense single nucleotide variants (SNVs) represent one of the most common forms of genetic variation and account for a substantial proportion of variants of uncertain significance in clinical databases. Accurate computational classification of these variants remains an important challenge in precision medicine and genomic research. In this study, we present PathoPredictor, an interpretable machine-learning framework designed to distinguish pathogenic from benign missense variants using curated clinical variant data and functional annotations. High-confidence variants were obtained from the November 2023 ClinVar release and annotated using dbNSFP v5.1 (GRCh37). After data filtering, imputation, and normalization, 59,302 expert-reviewed missense variants were retained for model development. Six machine-learning algorithms were evaluated under identical cross-validation conditions applied to the training set. Among the evaluated models, LightGBM demonstrated the strongest overall performance and was selected as the final PathoPredictor classifier, achieving a mean ROC–AUC of 0.93 ± 0.004, accuracy of 0.90 ± 0.006, and Matthew’s correlation coefficient of 0.80 ± 0.008 across five cross-validation folds. Model interpretability was examined using SHAP (SHapley Additive exPlanations), enabling both global feature ranking and variant-level explanation of predictions. Temporal validation using ClinVar variants submitted after November 2023 showed consistent predictive performance on previously unseen submissions within the same database ecosystem (ROC–AUC = 0.91). While the framework demonstrates strong discrimination and structured interpretability, potential limitations include training data bias and partial circularity associated with the inclusion of existing meta-predictors. Overall, PathoPredictor provides a reproducible and interpretable computational framework for integrating functional annotations in missense variant prioritization, supporting research and genomic analysis workflows. Full article

Background: Diabetes mellitus is a global health challenge, especially among homeless people. Early prediction of diabetes can reduce treatment costs and improve interventions. This study aimed to identify predictors of diabetes among homeless adults by utilizing artificial intelligence and providing recommendations for diabetes prevention. Methods: A case-control study of 150 homeless adults in Giza, Egypt (99 diabetes cases and 51 controls), analyzed 43 variables collected through interviews and physiological measures, with missing data imputed. Feature selection using recursive feature elimination and univariate and correlation analyses reduced the predictors to 13 variables. The class imbalance was addressed using synthetic minority over-sampling on the training set. Six models and a stacking ensemble with XGBoost as a meta-learner were evaluated using 5-fold cross-validation and performance metrics, including the accuracy, precision, recall, F1-score, and AUC-ROC. Results: The key predictors included BMI, systolic blood pressure, triceps skinfold thickness, waist circumference, lifestyle factors, comorbidities, diastolic blood pressure, age, medication adherence, educational level, marital status, duration of residence, and diabetes knowledge. Individual classifiers achieved a moderate performance (accuracy: 56.7–70.0%, F1-score: 0.686–0.781). The stacking ensemble substantially outperformed individual models, achieving a 95.45% accuracy, a 100% precision, a 93.75% recall, a 0.968 F1-score, and a 0.979 AUC-ROC on the test set. Conclusions: Machine learning models can reliably predict diabetes. The proposed hybrid stacking model outperformed conventional classifiers in terms of the prediction performance, highlighting the benefits of ensemble learning and sophisticated resampling strategies in dealing with imbalanced medical data. It is recommended that healthcare institutions integrate AI-powered diagnostic assistance technology into clinical processes to aid in the early detection and treatment of diabetes. Full article

Suicidal behavior among adolescents and young adults represents a growing public health concern due to its high prevalence and its negative impact on psychological well-being. The aim of this study was to examine the associations between emotional regulation, attachment styles, cyberviolence, and suicidal ideation among Chilean university students. A descriptive cross-sectional design was employed with a sample of 1083 participants, using the Suicidal Ideation Frequency Inventory, the Close Relationship Experience Questionnaire (ECR-R), the Spanish Modified Version of the Trait Meta-Mood Scale (TMMS-24) and the Cyber Dating Violence Instrument for Teens (CyDAV-T). Bivariate analyses and binary logistic regression were conducted to identify significant predictors of suicidal ideation. The results revealed a high prevalence of suicidal ideation, particularly among women (19.06%; p < 0.001). Difficulties in emotion regulation were strongly associated with a higher likelihood of suicidal ideation (p < 0.001), whereas adequate (p < 0.001) or excellent (p < 0.01) regulation functioned as a significant protective factor. In addition, a disorganized attachment style was identified as a risk factor (p < 0.05), especially among women (p < 0.01). In conclusion, emotion regulation emerges as a key protective factor against suicidal ideation, underscoring the importance of implementing socioemotional training programs within university settings. Full article

Objective: Chronic subdural hematoma (cSDH) is a common neurosurgical condition with rising incidence in the aging population. Recurrence after surgical evacuation remains frequent, affecting up to one third of patients. Prior studies have proposed an association between ABO blood type and recurrence risk, though the findings are inconsistent. This study aimed to determine whether ABO blood group independently predicts cSDH recurrence after adjusting for clinical and radiological risk factors and to contextualize these findings in the context of previously published studies. Methods: We conducted a retrospective cohort study of all consecutive patients surgically treated for cSDH at University Hospital RWTH Aachen between 2015 and 2023. Clinical, laboratory, and imaging data, including hematoma volume, laterality, and architecture, were extracted from medical records. The primary outcome was recurrence requiring reintervention. Associations between ABO blood type and recurrence were assessed using chi-square tests and multivariable logistic regression. A random-effects meta-analysis was performed to integrate our findings with all identified prior studies reporting recurrence stratified by blood type. Results: Of 630 treated patients, 488 with documented ABO blood type and recurrence status were included. Recurrence occurred in 31.1% of these patients. ABO distribution matched European population frequencies. Univariate analysis showed no association between blood type and recurrence (p = 0.434). Adjusted models likewise showed no significant differences between type A and type B (OR 1.43, 95% CI 0.67–3.00), type AB (OR 2.35, 95% CI 0.74–7.24), and type O (OR 0.95, 95% CI 0.57–1.58). Hematoma architecture remained strongly associated with recurrence (p < 0.001). A meta-analysis of available studies similarly demonstrated no association between any ABO blood type and recurrence, with pooled odds ratios near unity across comparisons. Conclusions: ABO blood type was not associated with cSDH recurrence in our cohort, and pooled evidence from previously published studies confirms the absence of a meaningful effect. Hematoma architecture and volume remain the most important predictors of recurrence. Based on these results, blood type should not influence postoperative surveillance or counseling, and future work should focus on modifiable biological and imaging-based determinants of recurrence. Full article

Background/Objectives: Recent meta-analyses have established that factors such as sarcopenia, male sex, and low serum albumin significantly correlate with increased postoperative complications in head and neck surgery, with routine neck computed tomography (CT) at the third cervical vertebra (C3) serving as a practical tool for muscle mass assessment. This study aimed to confirm the prognostic value of C3-based sarcopenia and specific clinical comorbidities in predicting early postoperative complications in patients with oral squamous cell carcinoma (OSCC). Methods: A retrospective cohort study was conducted on 167 patients undergoing primary surgery for OSCC. Sarcopenia was assessed using the C3-vertebra skeletal muscle index (SMI) derived from routine preoperative neck CT scans. Clinical and surgical variables, including preoperative serum albumin levels, comorbidities, and flap reconstruction types, were evaluated. A priori multivariate logistic regression models were utilized to identify independent predictors of surgical site and pulmonary and total complications (Clavien–Dindo classification) within 30 days. Results: The overall complication rate was 51%. Multivariate analysis revealed that sarcopenia (aOR: 3.26; 95% CI: 1.11–9.56), male sex (aOR: 3.48; 95% CI: 1.11–10.85), coronary artery disease (CAD) (aOR: 4.30; 95% CI: 1.21–15.36), and free-flap reconstruction (aOR: 15.06; 95% CI: 2.47–92.01) were robust independent predictors of total complications. Male sex (aOR: 4.17; 95% CI: 1.51–11.58) and preoperative hypoalbuminemia (<3.5 g/dL) (aOR: 3.43; 95% CI: 1.20–9.82) were independent predictors of surgical site complications, while regional flap reconstruction was independently associated with pulmonary complications (aOR: 5.97; 95% CI: 1.38–25.97). Conclusions: Sarcopenia, male sex, CAD, and flap reconstruction type are strong independent predictors of postoperative morbidity in OSCC. These findings advocate for “opportunistic screening” of muscle mass via routine preoperative neck CT, alongside rigorous cardiovascular profiling, to identify high-risk phenotypes for targeted perioperative optimization. Full article

Peri-implantitis is a chronic inflammatory condition driven by dysregulated host immune responses, yet clinical risk assessment continues to rely on routinely collected clinical indicators. Clinical prediction models, including machine learning-based and conventional approaches, have been proposed to integrate these indicators for peri-implantitis risk stratification, but their conceptualization of immunopathological risk has not been systematically examined. This systematic review and functional meta-synthesis were conducted according to PRISMA 2020. Six eligible studies were included, comprising 1316 patients and 2438 dental implants. Four studies employed machine learning-based models, and two used conventional clinical prediction approaches. A functional meta-synthesis was performed to interpret how models integrate clinical predictors as surrogate manifestations of immune dysregulation. Additionally, an exploratory random-effects meta-analysis of area under the receiver operating characteristic curve (AUC) values was conducted where applicable. Discriminative performance ranged from moderate to high across studies, with overlapping AUC estimates between modeling paradigms. Despite methodological differences, both machine learning and conventional models converged on shared immunopathological constructs related to inflammatory burden, prior periodontal disease, plaque-related factors, and host systemic conditions. These findings support the clinical utility of immunopathologically informed prediction models for peri-implantitis and highlight the need for future studies incorporating external validation. Full article

Introduction: Immune checkpoint inhibitors (ICIs) have transformed the treatment of metastatic melanoma; however, predictive markers of therapeutic response remain poorly defined. This study systematically assesses clinical, histological, and molecular predictors associated with survival outcomes in melanoma patients treated with ICIs. Methods: Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines, a systematic search was conducted in MEDLINE, Web of Science, and the Cochrane Central Register of Controlled Trials (CENTRAL) for studies published between January 2018 and October 2025. Eligible studies reported associations between predictive factors and overall survival (OS) or progression-free survival (PFS) in adult melanoma patients receiving ICIs. Pooled hazard ratios (HRs) with corresponding 95% confidence intervals (CIs) from univariate (UVA) and multivariate analyses (MVA) were synthesized using random-effects meta-analyses. Results: Sex was not a consistent predictor (contradictory effects; PFS heterogeneity I² ≈ 90%), whereas older age predicted worse OS (MVA continuous: HR 1.05, 95% CI 1.02–1.08; UVA ≥ 65 vs. <65: HR 1.70, 95% CI 1.36–2.12). Poor performance status, assessed using the Eastern Cooperative Oncology Group (ECOG) scale, strongly predicted inferior outcomes (ECOG ≥ 1 vs. 0: MVA OS HR 2.01, 95% CI 1.61–2.51; MVA PFS HR 1.49, 95% CI 1.18–1.88; ECOG ≥ 2 vs. <2: MVA OS HR 2.24, 95% CI 1.79–2.81). Elevated lactate dehydrogenase (LDH) was consistently associated with poorer survival (MVA OS HR 1.71, 95% CI 1.53–1.91; MVA PFS HR 1.61, 95% CI 1.41–1.85), whereas body mass index (BMI) > 25 kg/m² was associated with improved OS (HR 0.82, 95% CI 0.68–0.98). Higher disease burden predicted worse prognosis (Stage IV vs. III: MVA OS HR 1.57, 95% CI 1.16–2.13; >2 metastatic sites vs. ≤2: MVA OS HR 2.38, 95% CI 1.40–4.07; brain metastases: MVA OS HR 1.69, 95% CI 1.30–2.20; MVA PFS HR 1.52, 95% CI 1.00–2.33). Histologic and molecular factors showed prognostic value: ulceration worsened OS (UVA HR 2.08, 95% CI 1.25–3.44) and PFS (UVA HR 2.97, 95% CI 1.39–6.32); acral subtype had poorer OS than cutaneous melanoma (MVA HR 2.99, 95% CI 1.63–5.48); high tumor mutational burden (TMB) improved PFS (UVA HR 0.47, 95% CI 0.33–0.70); and cutaneous immune-related adverse events (irAEs) were associated with favorable outcomes (skin disorders: UVA OS HR 0.26, 95% CI 0.14–0.47; UVA PFS HR 0.50, 95% CI 0.34–0.74). In contrast, detectable circulating tumor DNA (ctDNA) predicted markedly worse PFS (MVA HR 4.72, 95% CI 2.31–9.65) and a non-significant trend toward worse OS (MVA HR 3.34, 95% CI 0.96–11.67). Liver metastases and programmed death-ligand 1 (PD-L1) expression were not significantly associated with survival. Discussion: This meta-analysis synthesizes evidence on clinicopathologic, laboratory, and histopathologic predictors of immunotherapy outcomes in metastatic melanoma. Performance status, age, LDH, BMI, and metastatic burden consistently correlated with prognosis, while ulceration, disease stage, and TMB emerged as key histologic determinants. Conversely, PD-L1 and gender showed no consistent predictive value, whereas cutaneous immune-related adverse events and ctDNA reflected favorable and poor outcomes, respectively. These findings highlight the multifactorial nature of immunotherapy response and support the further development of integrated prognostic models to refine patient stratification and optimize treatment outcomes. Full article

This study develops a hybrid framework integrating ensemble learning with explainable artificial intelligence to address the methodological challenge of balancing predictive accuracy and interpretability in credit risk model comparison. Using the German Credit Dataset, we implemented a comprehensive preprocessing pipeline, including feature encoding, scaling, and SMOTE for class imbalance handling. Four base models, logistic regression, Random Forest, XGBoost, and Multilayer Perceptron, were combined through a Stacked Ensemble with a logistic regression meta learner. The ensemble demonstrated strong performance, achieving an AUC of 0.761, precision of 0.783, recall of 0.806, and an F1 score of 0.794, which represented the highest scores among all models tested. Notably, Random Forest (AUC = 0.749) surpassed XGBoost (AUC = 0.733), challenging conventional algorithmic hierarchies. SHAP analysis provided transparent global and local interpretability, identifying Current Account status (SHAP = 0.153), Loan Duration (0.064), and Savings Account (0.063) as dominant predictor variables. Class-imbalance handling and threshold optimisation enhanced practical utility by reducing false positives from 39 to 16, thereby aligning with financial risk priorities. The framework provides a reproducible methodological pipeline for systematically comparing credit scoring approaches, demonstrating how predictive performance can be evaluated alongside interpretability considerations within a benchmark dataset context. Full article

Background and Objectives: Diabetes mellitus represents one of the most prevalent chronic metabolic disorders worldwide, necessitating precise insulin dose management to prevent both acute and long-term complications. The optimization of insulin dosing remains a significant clinical challenge, as inappropriate dosing can lead to hypoglycemia or hyperglycemia, each carrying substantial morbidity risks. Machine learning approaches have emerged as promising tools for developing clinical decision support systems; however, their practical implementation requires both high predictive accuracy and model interpretability. This study aimed to develop and evaluate an explainable machine learning framework for predicting insulin dose adjustments in diabetic patients. We sought to compare multiple ensemble learning approaches and identify the optimal model configuration that balances predictive performance with clinical interpretability through comprehensive SHAP and LIME analyses. Materials and Methods: A comprehensive dataset comprising 10,000 patient records with 12 clinical and demographic features was utilized. We implemented and compared nine machine learning models, including gradient boosting variants (XGBoost, LightGBM, CatBoost, GradientBoosting), AdaBoost, and four ensemble strategies (Voting, Stacking, Blending, and Meta-Learning). Model interpretability was achieved through SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) analyses. Performance was evaluated using accuracy, weighted F1-score, area under the receiver operating characteristic curve (AUC-ROC), precision-recall AUC (PR-AUC), sensitivity, specificity, and cross-entropy loss. Results: The Meta-Learning Ensemble achieved superior performance across all evaluation metrics, attaining an accuracy of 81.35%, weighted F1-score of 0.8121, macro-averaged AUC-ROC of 0.9637, and PR-AUC of 0.9317. The model demonstrated exceptional sensitivity (86.61%) and specificity (91.79%), with particularly high performance in detecting dose reduction requirements (100% sensitivity for the ‘down’ class). SHAP analysis revealed insulin sensitivity, previous medications, sleep hours, weight, and body mass index as the most influential predictors across different insulin adjustment categories. The meta-model feature importance analysis indicated that LightGBM probability estimates contributed most significantly to the ensemble predictions. Conclusions: The proposed explainable Meta-Learning Ensemble framework demonstrates robust predictive capability for insulin dose adjustment recommendations while maintaining clinical interpretability. The integration of SHAP-based explanations facilitates clinician understanding of model predictions, supporting transparent and informed decision-making in diabetes management. This approach represents a significant advancement toward the clinical implementation of artificial intelligence in personalized insulin therapy. Full article

Green FinTech involves facilitating sustainable payments, banking, and investment; nevertheless, it is subject to consumer trust and perceptions of ‘green’ value. The literature on this topic is fragmented, with information systems literature typically considering trust as a broad acceptance construct, while sustainable literature considers it as a risk of ‘greenwashing’ without integrating credibility into adoption models. This systematic review aggregates 15 empirical studies and addresses five research questions. RQ1 examines the theoretical models applied to examine trust in green/sustainable FinTech adoption. RQ2 examines the conceptualization and measurement of trust across different contexts, distinguishing institutional/provider trust, platform/tech trust, and sustainability claim credibility trust. RQ3 examines the function of trust within behavioral models (predictor, mediator, moderator). RQ4 examines methodological characteristics and quality indicators (research design, sampling frame, reliability, and bias). RQ5 examines the direct relationship between trust and adoption intention using meta-analysis. The systematic review follows a set of PRISMA guidelines, where we searched Scopus and Web of Science (2015–2026) and applied an RQ-based coding scheme to peer-reviewed articles. Measures of trust varied significantly (unidimensional, integrity–competence–benevolence, and technology-specific scales), limiting cross-study comparability. Using random effects, we found a significant positive relationship between trust and intention (pooled standardized direct path coefficient β = 0.27, 95% CI [0.14, 0.41]) with considerable heterogeneity (I² = 88%) and a wide prediction interval including near-zero effects. Literature essentially endorses trust as a significant yet context-dependent construct, emphasizing the necessity for measurement standardization, a more distinct differentiation between sustainability trust and general platform trust, regular reporting of reliability and bias assessments, and focused evaluations of boundary conditions (e.g., environmental skepticism, regulatory framework, and FinTech type). Full article