Search Results (230)

Objectives: To assess the feasibility and reproducibility of predicting antimicrobial resistance (AMR) in Escherichia coli from MALDI-TOF mass spectrometry data using a standardized, open-source machine learning (ML) workflow, we systematically compared four ML algorithms, evaluated the impact of culture conditions, extract storage, and spectral preprocessing on model performance, and validated results through nested cross-validation with statistical significance testing. Methods: A total of 282 clinical E. coli isolates were analyzed. Two MALDI-TOF MS datasets were generated from freshly cultured extracts (T₁) and recultured isolates one year later (T₃), yielding 4468 spectra. A third dataset from the T₁ extracts stored at −20 °C for one year (T₂) was evaluated for spectral stability but excluded from primary modeling likely due to storage-induced degradation. Protein spectra (m/z 2000–15,000) were preprocessed using an in-house developed MALDI-TOF preprocessing pipeline (MTPP) comprising variance stabilization, Savitzky–Golay smoothing, SNIP baseline correction, TIC normalization, LOWESS alignment, and MAD-based peak detection (SNR ≥ 3), yielding 121 m/z features. Four classifiers—Random Forest (RF), Logistic Regression, Support Vector Machine, and Gradient Boosting—were trained to predict resistance to 11 antibiotics using nested cross-validation: outer GroupShuffleSplit (5-fold, isolate-level) for evaluation and inner GroupKFold for recursive feature elimination (RFECV) and hyperparameter tuning (RandomizedSearchCV). Classification thresholds were optimized via the precision–recall curve. Model performance was assessed using AUROC, AUPRC, F1-score, Matthews Correlation Coefficient (MCC), and bootstrap 95% confidence intervals (1000 replicates). Pairwise model comparisons were tested with McNemar’s chi-squared test. Results: Among the 12 antibiotics included in the analysis (meropenem excluded for absence of resistance), resistance prevalence ranged from 1.1% (colistin) to 59.9% (amoxicillin). Colistin was subsequently also excluded from ML modeling due to insufficient resistant isolates (n = 3), leaving 11 antibiotics for prediction. The best predictive performance was observed for ciprofloxacin (AUROC 0.76 [95% CI 0.74–0.77]; F1 0.54; MCC 0.38) and ceftazidime (AUROC 0.68 [0.65–0.71]; F1 0.36; MCC 0.29), using 13 and 37 RFECV-selected features, respectively. Amoxicillin achieved the highest F1-score (0.76), driven by high recall (0.98) but modest AUROC (0.58). No meaningful predictive signal was detected for amikacin, cefepime, or tigecycline (AUROC ≤ 0.57, F1 ≤ 0.17), attributable to extreme class imbalance, and no robust multi-peak resistance signature was detected in this dataset. McNemar’s test confirmed that RF significantly outperformed Logistic Regression for all antibiotics (p < 0.01), while Gradient Boosting performed comparably to RF for ciprofloxacin (p = 0.17) and ceftazidime (p = 0.28). Frozen extracts (T₂) produced lower spectral similarity and were excluded from model training; the aligned T₁₊₃ dataset yielded the most stable performance across metrics. Conclusions: Machine learning analysis of MALDI-TOF spectra enables reproducible AMR prediction for selected antibiotics in E. coli, with ciprofloxacin and ceftazidime showing the strongest signal. Nested isolate-level cross-validation, multi-model comparison with statistical testing, and open-source code provide a transparent, reproducible foundation for integrating ML-assisted MALDI-TOF analysis into diagnostic AMR surveillance. Extract storage at −20 °C degrades spectral quality and should be avoided in ML training workflows. Full article

Airport turnaround is an important operational process that directly affects flight punctuality, airport capacity, and ground-handling efficiency. However, many turnaround activities are still monitored manually or through fragmented operational records, which can limit real-time visibility and delay identification. This study proposes a computer vision-based airport turnaround monitoring pipeline that integrates YOLOv11 object detection, Norfair multi-object tracking, and frame differencing-based motion analysis to extract key operational events from airport video footage. Publicly available turnaround footage from Shinshu Matsumoto Airport, Japan, was collected under different environmental conditions, including daytime, nighttime, rainy, after-rain, and transition lighting conditions. From selected videos, 1446 images were labeled into 11 airport turnaround object classes, including tow tug, aerobridge, airplane, baggage container, belt loader, belt loader roof, fuel line, fuel tanker, fuel tube, tractor, and window. The dataset was divided into training, validation, and testing sets using a 70:20:10 ratio. The trained YOLOv11 model achieved strong detection performance, with overall test an precision of 0.9609, recall of 0.9445, and mAP50 of 0.9617. To support activity-level interpretation beyond object detection, the proposed pipeline applies frame differencing within specific regions of interest, including the aerobridge window region for passenger deboarding and boarding detection, and the belt loader roof region for baggage unloading and loading detection. The extracted object detections, motion spikes, and temporal logs are then converted into a Gantt chart that summarizes major turnaround activities, including airplane parking, deboarding, baggage unloading, refueling, baggage loading, boarding, and pushback. The results demonstrate that the proposed modified YOLO-based pipeline can transform ordinary airport video footage into structured operational timelines, supporting more transparent, data-driven, and automated monitoring of airport turnaround processes. Full article

Industrial defect generation has long been challenged by the scarcity of real anomaly samples and the imbalance of defect categories, particularly in complex industrial scenarios involving transparent containers. Taking vials as an example, glass reflection, specular highlights, and fine-grained defects make continuous defect acquisition difficult, thereby making the realism and controllability of augmented samples critical to downstream detection performance. Although existing diffusion-based generation methods can improve synthetic image quality, they often require additional training or lightweight fine-tuning, which limits their efficiency in sample-limited industrial scenarios. To address this issue, this paper builds upon the TF-IDG framework and proposes a training-free industrial defect generation method based on multi-domain consistency and geometric-semantic constraints. To alleviate the unnatural texture details, boundary transitions, and background blending commonly observed in generated defects, a multi-domain consistency constraint is introduced to enhance generation realism from both frequency-domain structures and cross-domain contextual representations, thereby improving anomaly texture expression and overall visual coherence. To further mitigate unstable defect contours, spatial deviation, and structural mismatch with target objects, a geometric-semantic constraint is designed to regulate the generation process through elastic shape constraints and semantic region-anchored attention, enhancing the rationality of defect morphology evolution and spatial localization. Experimental results on both the MVTec AD dataset and a self-built vial defect dataset demonstrate that the proposed method outperforms comparative approaches. Specifically, when YOLOv11 is used as the downstream detector, the mAP@50 on the MVTec AD dataset and the self-built vial defect dataset is improved from 88.5% and 98.0% for the TF-IDG baseline to 89.6% and 98.8%, respectively. Full article

Background/Objectives: Artificial intelligence is increasingly applied to dental imaging, yet favorable internal performance does not necessarily indicate clinical transferability. This systematic review evaluated whether imaging-based dental artificial intelligence models have progressed beyond internal algorithmic development toward external validation, generalizability, reproducibility, privacy-preserving learning, and clinical implementation readiness. Methods: Searches were conducted in PubMed/MEDLINE, Scopus, and Embase up to May 2026. Eligible studies were primary empirical investigations based on human dental or oral imaging data that assessed at least one translational-validation dimension beyond internal development, including external testing, multicenter or multi-device validation, cross-dataset reproducibility, or privacy-preserving learning. Evidence was synthesized using a structured narrative synthesis reported according to the Synthesis Without Meta-analysis framework. Results: Fifteen studies published between 2023 and 2026 were included. They addressed caries detection, periodontal bone loss, gingival inflammation, root morphology, palatal radicular grooves, radiographic quality control, tooth-width estimation, and dental-structure segmentation. Translational-readiness domains included external validation, generalizability, reproducibility, privacy-preserving learning, transparency, and workflow relevance. Validation varied across cohorts, repositories, centers, devices, cross-dataset benchmarks, and federated-learning settings. Reproducibility, annotation harmonization, uncertainty reporting, explainability, workflow evaluation, and code or model availability were inconsistent. Quantitative pooling was not performed because tasks, modalities, units of analysis, reference standards, validation designs, and metrics were highly heterogeneous. Conclusions: Within this selected subset of externally tested studies, translational progress is emerging but remains uneven. Implementation readiness requires stronger reproducibility, clinically meaningful validation, workflow evaluation, and attention to regulatory, organizational, and human-factor barriers. Full article

Background/Objectives: Chronic kidney disease (CKD) remains underdiagnosed until advanced stages, motivating reliable, clinically deployable screening models that pair high discrimination with an explicit operating threshold and transparent explanations. Methods: In this study, we propose a CKD detection framework that integrates structured preprocessing, class imbalance handling, stratified 10-fold cross-validation with out-of-fold (OOF) prediction, and clinically oriented threshold selection via the Youden index, followed by explainability using SHAP and LIME. Experiments were conducted on two datasets. Across a broad panel of ten machine learning models, gradient boosting methods consistently dominated. Results: LightGBM achieved the best overall clinical composite performance on both datasets. On Dataset 1, LightGBM delivered near-ceiling OOF discrimination (ROC-AUC = 99.98, PR-AUC = 99.98) and an excellent clinically balanced performance at the best Youden threshold (0.41), reaching sensitivity = 99.20, specificity = 99.60, accuracy = 99.40, F1 = 99.40, and MCC = 98.80, with robust cross-validation stability (CV AUC = 99.99 ± 0.04; CV sensitivity = 99.10 ± 1.81; CV specificity = 99.46 ± 1.42; CV MCC = 98.59 ± 2.19), strong calibration (Brier = 0.006), and fast training (0.078 ± 0.019 s/fold). On Dataset 2, LightGBM maintained high generalization (ROC-AUC = 99.72, PR-AUC = 99.64) and clinically deployable balance at the best Youden threshold (0.35), achieving sensitivity = 98.10, specificity = 98.03, accuracy = 98.06, F1 = 98.06, and MCC = 96.13, with consistent fold-wise performance (CV AUC = 99.69 ± 0.25; CV sensitivity = 97.25 ± 1.25; CV specificity = 98.11 ± 1.02; CV MCC = 95.37 ± 1.56), acceptable calibration (Brier = 0.0173), and practical training time (0.742 ± 0.144 s/fold). Conclusions: Finally, SHAP and LIME explanations confirmed that model decisions align with clinically meaningful renal function and symptom/biomarker patterns at both population and patient levels, supporting safer translation of the proposed framework into CKD screening and decision-support workflows. Full article

Background/Objectives: Artificial intelligence (AI) is increasingly transforming cardiology through advanced analytical tools capable of identifying complex patterns across cardiovascular imaging, electrophysiology, and clinical datasets. Machine learning (ML) and deep learning (DL) algorithms are being integrated into echocardiography, cardiac computed tomography (CT), cardiac magnetic resonance imaging (MRI), and electrocardiography (ECG), enabling earlier diagnosis and more personalized cardiovascular care. This narrative review summarizes current clinical and organizational applications of AI in cardiology and discusses emerging concepts related to explainable and trustworthy AI. Methods: A narrative review was conducted according to SANRA recommendations using the PubMed, MEDLINE, Web of Science, and Scopus databases, including peer-reviewed publications from 2015 to 2026 addressing clinical, organizational, and ethical applications of AI in cardiology, with particular emphasis on cardiovascular imaging, electrocardiography, heart failure, digital health, and explainable AI frameworks. Results: Substantial evidence demonstrates that AI-based tools can achieve expert-level performance in cardiovascular imaging interpretation, automated electrocardiographic analysis, and clinical risk prediction. Across multiple cardiovascular settings, AI has been associated with improved diagnostic accuracy, enhanced workflow efficiency, and earlier detection of cardiovascular disease. Predictive models support risk stratification in heart failure and ischemic heart disease, while chatbots and digital health platforms may facilitate patient engagement, remote monitoring, and continuity of care. Despite these advances, important challenges remain, including algorithmic bias, limited transparency, insufficient external validation, data heterogeneity, and barriers to routine clinical implementation. Emerging explainable AI approaches may improve model interpretability, clinician confidence, and the safe adoption of AI-driven decision support systems. Conclusions: Artificial intelligence is rapidly evolving from a research-oriented technology into a clinically relevant component of cardiovascular care. Current evidence indicates that AI can enhance diagnostic performance, improve risk prediction, streamline clinical workflows, and facilitate more personalized management across multiple cardiovascular domains. However, the successful translation of AI into routine practice will depend on robust external validation, transparent decision-making mechanisms, regulatory oversight, and clinician acceptance. The development of explainable and trustworthy AI frameworks represents a critical step toward the safe, ethical, and sustainable integration of AI into modern cardiology. Full article

Background: The management of diabetic foot disease and knee osteoarthritis (OA) with smart orthotics holds significant importance during the early stages of these conditions, given their potential consequences, including functional impairment, chronic pain, and economic burden. Real-time monitoring of plantar foot pressure enables early detection of abnormal force distribution and gait biomechanics, allowing for the redirection of forces away from affected ulcers or arthritic joints. This is the first systematic review to synthesise clinical evidence for smart orthotics technology with real-time plantar pressure sensor biofeedback across both diabetic foot ulcer prevention and knee osteoarthritis management simultaneously. A search of the PROSPERO register confirmed no existing registration covers this specific combination. Objectives: To examine the clinical evidence for the use of standard and smart orthotics in the prevention and management of diabetic foot ulcers (DFUs) and knee OA, and to evaluate their impact on plantar pressure redistribution, ulcer recurrence, pain, biomechanics, and economic burden. Eligibility criteria: Studies published in English involving human adult participants (≥18 years) with a clinical diagnosis of diabetes mellitus (at risk of DFU or with peripheral neuropathy) or knee OA, where the intervention involved any orthotic device or smart/intelligent insole with clinical outcomes reported, were included. Studies on healthy individuals only, those not reporting participant age, and non-weight-bearing protocols not differentiated from weight-bearing were excluded. Information sources: Five databases were searched: CINAHL (EBSCO Information Services, Ipswich, MA, USA), PubMed Advanced (National Library of Medicine, Bethesda, MD, USA), Wiley Online Library (John Wiley & Sons, Hoboken, NJ, USA), Cochrane Library (Cochrane Collaboration, London, UK), and Google Scholar (Google LLC, Mountain View, CA, USA). Searches were completed in May 2026. Methods: We conducted a comprehensive literature review. This review was structured and reported with reference to the PRISMA 2020 statement (Preferred Reporting Items for Systematic Reviews and Meta-Analysis; University of Ottawa, Ottawa, ON, Canada) to guide transparency of reporting. It does not constitute a full Cochrane-style systematic review; risk of bias assessment was applied to key included studies and GRADE (Grading of Recommendations Assessment, Development and Evaluation; McMaster University, Hamilton, ON, Canada) certainty ratings were applied informally and narratively rather than as formal per-outcome evidence profiles. Five databases were searched yielding 92,637 records. After removal of 398 duplicates by Rayyan, 92,239 records remained. A subsequent automated keyword-based relevance filter applied within Rayyan (Rayyan AI, Doha, Qatar), prior to human screening, excluded 84,572 records that did not contain any terms related to orthotics, diabetic foot, or knee osteoarthritis, yielding 7667 records for human title/abstract screening. A narrative synthesis approach was adopted owing to the heterogeneity of study designs and outcome measures across included studies, which precluded meta-analysis. This review was not prospectively registered. A complete list of all 78 included studies, including those not individually discussed in the results and discussion. Results: The available clinical studies report promising findings for orthotics and smart orthotics in pain reduction, ulcer prevention, and potential reduction in economic burden, though conclusions are limited by small sample sizes, heterogeneity, and predominantly open-label designs. Recent research found that orthotics can be used to alter the gait pattern that influences knee OA by reducing excessive force on the affected joint. A randomised controlled trial demonstrated an 80% relative risk reduction in DFU recurrence (RR = 0.20; 95% CI: 0.06–0.79; p = 0.022), with absolute event rates of 6.3% in the intervention group versus 30.8% in controls (ARR = 24.5%); a second trial reported a 71% reduction in ulcer incidence over 18 months; and a third randomised controlled trial demonstrated statistically significant plantar pressure reduction (p < 0.01) in patients with diabetic neuropathy. Conclusions: The available evidence suggests that orthotics may be associated with improved pressure redistribution, reduced ulcer incidence, and benefit in the management of knee OA. Although the number of studies directly comparing smart orthotics with standard orthotics remains limited, the limited comparative studies suggested that smart orthotics showed promising results in reducing ulcer incidence, providing the patient with real-time feedback to offload via their electronic devices. These findings, while preliminary, highlight the potential of smart orthotic technology as an adjunct to standard orthotic care in reducing the overall burden of diabetic foot disease and knee osteoarthritis. Limitations: The primary methodological limitation of this review is the open-label design of all included smart orthotic trials, which precludes participant blinding and introduces performance bias. However, this limitation is structural and inherent to the wearable technology field—analogous to surgical trials—and is substantially mitigated by the use of objective primary outcome measures (plantar pressure and ulcer recurrence) across the three included RCTs, the consistency of effect direction across independent RCTs conducted in different countries, and a narrative sensitivity analysis confirming robustness of findings (Risk of Bias Across Studies Section). Formal per-outcome GRADE evidence profiles were not produced; overall certainty of evidence was assessed narratively with reference to GRADE domains and is judged to be low to moderate for smart orthotics in DFU prevention and low for knee OA management, consistent with the Level 2–3 evidence base and open-label study designs. Future adequately powered, multi-site RCTs with standardised outcome reporting, minimum 24-month follow-up, and integrated health economic modelling are the highest priority to extend these preliminary findings. Registration: This review was not prospectively registered. Full article

Electric Vehicle (EV) charging infrastructure plays a critical role in modern energy systems, affecting energy load distribution, demand-response programs, and grid stability. As EV adoption accelerates globally, the varied charging habits and concurrent interactions among users, stations, and shared infrastructure create operational inefficiencies that existing machine learning and optimization approaches cannot fully diagnose, because these methods rely on aggregated or single-entity representations that discard cross-object process dependencies. To address this gap, we propose EnOptiMine (Energy Optimization Framework for Electric Vehicles through Object-Centric Process Mining), a novel four-phase analytical framework that applies Object-Centric Process Mining (OCPM) to EV charging infrastructure. EnOptiMine operates by transforming raw EV charging data into an Object-Centric Event Log (OCEL 2.0), discovering the complete charging lifecycle as a structured multi-object process through Object-Centric Directly-Follows Graphs (OC-DFGs), performing conformance analysis to detect and quantify process deviations across object-type lifecycles, and proposing process improvement interventions. Applied to the EV charging dataset, EnOptiMine identifies sessions that exhibit post-charge station idle-blocking, departure mismatch, and carry lifecycle ordering violations. In the present work, the real-world simulation confirms that a graduated idle fee policy recovers 22.9% of wasted station-hours, and a departure reconfirmation protocol reduces mismatch sessions by 54.0%. These results demonstrate that OCPM provides process-transparent diagnostic capabilities for EV charging infrastructure that are inaccessible to existing prediction- and optimization-based methods. Full article

Background and Objectives: Assessing knee extensor (KE) strength is important for detecting muscle weakness in older adults, yet dynamometry is often impractical in community settings. This study examined whether smartphone-derived kinematics during the Five Times Sit-to-Stand Test (FTSST) could predict seated isometric KE strength. Materials and Methods: A cross-sectional study included 105 community-dwelling older adults (68.19 ± 5.85 years). A smartphone application extracted rising time, vertical velocity, and smartphone-estimated relative vertical power during the FTSST. KE strength was measured as maximum voluntary isometric contraction (MVIC) using fixed-frame dynamometry with a Lafayette dynamometer head. Bioelectrical impedance-derived body composition variables were reported descriptively but excluded from the primary prediction models to maintain a transparent movement-based model independent of device-specific body-composition estimates. Hierarchical regression models used smartphone-derived variables and transparent non-BIA covariates. Agreement was examined using Bland–Altman analysis. Results: Smartphone-estimated relative vertical power showed the strongest correlation with MVIC (r = 0.787, p < 0.001). The combined model including sex, age, femur length, and smartphone-estimated relative vertical power explained 71.6% of MVIC variance (adjusted R² = 0.716, SEE = 3.276 kg), outperforming vertical velocity, rising time, and total FTSST time models. Internal validation using repeated 10-fold cross-validation showed CV-R² = 0.701, CV-adjusted R² = 0.689, CV-RMSE = 3.343 kg, and CV-MAE = 2.739 kg. Bland–Altman analysis showed minimal mean bias (0.00 kg), 95% limits of agreement from −6.296 to 6.296 kg, and significant proportional bias (slope = −0.172, p = 0.002), indicating overestimation in weaker individuals and underestimation in stronger individuals. Conclusions: Consistent with our hypothesis, smartphone-estimated relative vertical power was the strongest kinematic predictor of seated isometric KE strength among the evaluated FTSST-derived variables. This approach may support community screening and monitoring, but it should not replace standardized dynamometry for precise individual-level strength quantification. Full article

Paralympic water polo requires classification systems that reflect sport-specific functional performance under ecologically valid conditions. This pilot study proposes a task-specific kinematic profiling framework for deriving objective, biomechanically interpretable descriptors of residual motor function. Five male national-level water polo athletes—three with eligible motor impairments and two able-bodied reference participants—performed standardized sport-specific tasks comprising upright floating, vertical propulsion, unilateral passing, non-contested shooting, and contested shooting under physical opposition. Stereoscopic video, OpenPose-based three-dimensional reconstruction, and phase-based analysis were used to extract features and composite indices of postural control, propulsion capacity, upper-limb residual function, and resistance to perturbation. Automatic ball-release detection matched manual frame-level verification in all 128 analyzed ball-related trials. Within the task-specific indices, where higher scores indicate greater functional burden, core values ranged from 0.05–0.15 for upright floating, 0.29–0.68 for combined arm-and-leg vertical propulsion, and 0.040–0.148 for contested shooting across the available subject–side combinations. The profiles showed task- and side-specific differences in stabilization, propulsion, and post-contact motor reorganization. The framework uses pose estimation as a quantitative measurement tool and treats visibility interruptions as functionally meaningful events rather than noise. It is not intended to replace official classification procedures, but to provide transparent and interpretable candidate descriptors for future evidence-based classification research in Paralympic water polo. Full article

Introduction: The physical characteristics of microplastics (MPs), particularly their size and color, closely resemble natural food prey for several marine organisms, leading to active or accidental ingestion by marine species, including fish larvae. Despite growing concern, the occurrence of MPs in wild fish during early developmental stages remains insufficiently documented, and laboratory studies report inconsistent results. Given their key ecological role in marine food webs and their economic relevance, the health and survival of fish larvae are critical for maintaining fish populations. Objective: This study aimed to investigate MPs’ presence throughout the larval developmental stages and assess whether MP contamination profiles (concentration, color, type, and size) differ between species. Methodology: MPs were analyzed in the larval stages of two fish species with distinct ecological niches: the European sardine (Sardina pilchardus), a marine migratory species, and the common goby (Pomatoschistus microps), an estuarine resident species. Samples were collected from the Douro Estuary (NW Portugal) over one year, covering different developmental stages. Results: MPs were detected in both species at all developmental stages observed, including the yolk-sac stage (where the feeding of larvae is endogenous), indicating contamination at a stage when the mouth is not yet functional. Sardina pilchardus showed a higher abundance of transparent nylon fibers of 0.5 mm, and Pomatoschistus microps transparent polypropylene fibers of size 0.4 mm. Moreover, MP contamination did not vary between species or throughout the developmental stages, showing similar levels and profiles of MPs contamination. Conclusions: These findings provide new evidence that MP contamination begins at the earliest developmental stages of the fish, from hatching onwards. The results further suggest that MP uptake in fish larvae is primarily driven by environmental availability rather than fish larvae’s preferences or ecological guild, physical characteristics, or even the ontogenetic developmental stage. Full article

Saudi Arabia’s audit market faces three governance challenges that existing frameworks may not fully address. These challenges concern a potential regulatory gap around autonomous AI accountability, a trust dimension that standard technology-adoption models may not fully capture, and limited mechanisms for independently verified ESG assurance under Vision 2030. This study adopts a conceptual design approach within the design science research tradition and proposes the CMA Agentic AI Platform as a practical response to these challenges. The platform comprises two segments. Segment 1 deploys autonomous drone swarms to verify corporate assets across four audit tasks—asset valuation, ESG compliance, anomaly detection and construction progress—using deep learning, thermal imaging and social-media cross-referencing. Segment 2 continuously monitors discretionary accruals and uses objective earnings-management data to inform auditor assignment and rotation decisions. This approach replaces subjective reputational assessments with transparent, quantifiable governance criteria. The platform is governed through the Triadic Agentic Framework, which extends classical agency theory by distributing authority across the Principal, the Human Agent and the AI Agent. The framework also operationalises Trust Expectancy as the primary adoption condition. The evidence base draws on two complementary streams: a PRISMA-guided systematic review and bibliometric analysis of thirty-nine peer-reviewed studies, and a documentary analysis of four national agentic-AI regulatory frameworks (SDAIA, MDDI/IMDA, NIST and ICO). The study contributes the concept of Algorithmic Accountability as a distinct governance domain, the Triadic Agentic Framework as an operational architecture for autonomous regulatory monitoring, and a reframing of the UTAUT trust construct for agentic-AI adoption in mature professional contexts. The platform converts theoretical governance into a regulatory architecture with direct implications for concentrated capital market regulators. Full article

Background/Objectives: Artificial intelligence (AI) systems for detecting lymph node metastases in histopathological images achieve near-expert classification performance but remain opaque to clinicians, limiting their clinical adoption and regulatory acceptance. This study presents the first rigorous quantitative framework for evaluating and comparing explainable AI (XAI) methods in digital pathology, providing actionable evidence-based guidance for clinical deployment. Methods: Four XAI techniques—LIME, GradCAM, GradCAM++, and SHapley Additive exPlanations (SHAP) via DeepExplainer—were applied to three convolutional neural networks (VGG19, ResNet50, and EfficientNetB3) trained on the PatchCamelyon (PCam) benchmark (220,026 patches). Quantitative evaluation employed two complementary frameworks: spatial agreement with expert pathologist annotations (Intersection over Union and Sørensen–Dice coefficient on 2847 annotated patches) and faithfulness metrics (Area Over the Perturbation Curve and insertion/deletion Area Under the Curve) independent of external annotations. Threshold sensitivity analysis was also conducted at fixed binarisation thresholds (τ = 0.3 and τ = 0.7) in addition to Otsu automatic thresholding. Results: GradCAM++ achieved the highest spatial agreement with pathologist annotations (mean IoU = 0.52 ± 0.14 for EfficientNetB3), while SHAP yielded the highest faithfulness scores (AOPC = 0.61 ± 0.08). The parameter-free squaregrid LIME variant offered a favourable trade-off (IoU = 0.44 ± 0.17) at 3.8× lower computational cost than LIME AVG. Relative method rankings were preserved across all binarisation thresholds, confirming the robustness of the evaluation framework. A Spearman correlation of ρ = 0.81 was found between model classification AUC and spatial agreement, indicating that superior classification performance systematically produces more spatially coherent explanations. Conclusions: GradCAM++ is recommended for high-throughput clinical workflows; SHAP for research contexts requiring maximal faithfulness; and squaregrid LIME as a transparent, parameter-free baseline for clinical communication and audit, preferred over LIME AVG on account of its parameter-free operation and 3.8× lower computational cost. A tiered deployment strategy integrating GradCAM++, SHAP, and squaregrid LIME is proposed. These findings provide quantitative, technical evidence of a type relevant to regulatory frameworks such as the FDA SaMD Action Plan and EU IVDR 2017/746; formal regulatory acceptance would additionally require prospective, multi-site external validation and a pathologist reader study, which lie beyond the scope of this single-benchmark study. Full article

Background/Objectives: Despite well-documented chemical and physical hazards in healthcare settings, existing reviews of occupational reproductive risks have largely focused on single-agent risk estimation and have rarely translated occupational hygiene evidence into clinical decision-making frameworks for reproductive counseling and surveillance. This narrative review synthesizes evidence across multiple occupational exposure categories—antineoplastic agents, high-level disinfectants (HLDs), sterilants, and work-organization factors—and proposes an integrated, clinically operational framework for preconception counseling, pregnancy-sensitive risk stratification, exposure-control verification, and reproductive health surveillance among women healthcare workers. Methods: A structured narrative literature search was conducted across PubMed/MEDLINE, Scopus, Web of Science, and Embase from database inception through January 2025 and updated in March 2026. The review was guided by a Population–Exposure–Comparison–Outcome (PECO) framework and structured using Search–Appraisal–Synthesis–Analysis (SALSA) principles and the Scale for the Assessment of Narrative Review Articles (SANRA). Evidence quality was summarized using a modified hierarchy-of-evidence classification provided as a reader aid. This narrative review employed structured transparency tools but does not claim the methodological status of a systematic review. Quantitative meta-analytic pooling was not performed owing to substantial heterogeneity across study designs, exposure assessment methods, and outcome definitions; findings were synthesized narratively by exposure category. Results: The strongest and most consistent evidence was identified for occupational exposure to antineoplastic agents, which has been associated with spontaneous abortion, stillbirth, congenital abnormalities, impaired fecundability, and selected cancer-related concerns. HLDs and sterilants represent exposure categories warranting precautionary attention, with some evidence suggesting possible adverse effects on fecundability and early pregnancy maintenance; however, findings are considerably more heterogeneous, context-dependent, and reliant on self-reported exposure assessment than those for antineoplastic agents. Broader workplace factors, including shift work, prolonged working hours, physical workload, and mixed exposures, may further contribute to reproductive risk. The synthesis supports task-specific occupational history taking, exposure-control verification, and pregnancy-sensitive risk stratification. Conclusions: This review provides a multi-exposure, clinically operational framework that bridges occupational hygiene evidence with reproductive healthcare delivery, offering practical decision-support tools for clinicians managing women healthcare workers during preconception, pregnancy, and lactation. The framework includes structured occupational history-taking questions, a clinical decision pathway with evidence-tier classification, and a prevention matrix linking exposure sources to workplace controls and clinical actions. Integrating task-specific occupational history taking into routine reproductive care may improve detection of preventable workplace risks and support timely accommodation, while clinicians should calibrate recommendation strength to the underlying evidence quality for each exposure category. Full article

Background and Objectives: Breast cancer remains one of the most common malignancies worldwide, and early detection plays a crucial role in improving treatment outcomes and reducing mortality. Several experimental studies using animal models of breast cancer have explored the potential of terahertz-based technologies in this field. However, their preclinical evidence base in breast cancer remains heterogeneous and has not been systematically synthesized with a focus on experimental models, imaging protocols, and barriers to translation. Methods: We conducted a descriptive systematic review, according to PRISMA guidelines, of 10 articles selected from a total of 372 identified across four databases—PubMed, Embase, Web of Science, and Cochrane—regarding the diagnostic performance of terahertz (THz) imaging in breast cancer animal models. We included studies that used rodent models diagnosed with breast cancer, subsequently confirmed through histological examination, and extracted relevant data. Results: The results were synthesized using a narrative approach. Most studies used C57BL/6J mice with E0771 cell line-induced breast tumors, with histopathology as the reference standard. In the reflection mode, at frequencies between 0.1 and 4 THz, the identification of tumoral, fibrous, fat, and muscle tissues was possible. Conclusions: Overall, the available preclinical evidence supports THz imaging as a promising proof-of-concept approach for breast tissue characterization, but not yet as a standardized or clinically translatable diagnostic platform. Future studies should use harmonized animal models, standardized acquisition and specimen-handling protocols, transparent reporting of classification workflows, and consistent outcome metrics to enable comparison across studies and to clarify the biological and biophysical determinants of THz contrast in breast cancer. Full article