Search Results (133,723)

Background: Rising survival rates have created a growing population of childhood cancer survivors (CCSs) who have an increased risk of late effects and require long-term follow-up (LTFU) care. Existing services are often fragmented and may not reflect survivors’ priorities; optimizing such care requires understanding not only the self-reported burden of late effects and their impact on daily lives but also survivors’ information status and their preferences for future care. Methods: Survivor Deutschland e.V. conducted a nationwide cross-sectional online survey assessing the current late effects, daily-life impairments, subjective information statuses, current follow-up structures, and preferences for future LTFU care. A total of 339 CCSs were included, covering all childhood cancer entities, most frequently leukemia (30.7%), central nervous system tumors (18.3%), and lymphoma (16.5%). Data were analyzed descriptively and supplemented by paired non-parametric analyses. Results: Overall, 74.3% (252/339) reported at least one late effect. Self-rated limitations in daily life had a median of five (Q1–Q3 3–7) on a 1–10 scale, which increased with the number of reported late effects. The most frequently affected domains were endocrine (36.3%, n = 123), fertility (33.6%), psychological (30.4%, n = 103), neurocognitive (26.0%, n = 88), and orthopedic (24.8%, n = 84) problems. A majority (69.3%, n = 235) knew that late effects existed yet felt insufficiently informed, and 4.4% only learned of these through the survey. Among 172 survivors in adult follow-up care, only 26.8% (46/172) reported access to structured, specialized LTFU care, whereas 56.4% (97/172) preferred this model. Survivors rated the importance of LTFU care highly (median 9/10) but rated satisfaction with their current care as much lower (median 3/10). The most valued components were the coverage of follow-up costs, sufficient consultation time, a dedicated contact person, and clear communication of results. Psychological support was a notable gap (5.8%, 10/172 current access vs. 17.9% 31/172 preferred), and 84.1% were willing to travel up to two hours or more for high-quality care. Conclusions: German CCSs report a high late-effect burden, meaningful impairment of their daily lives, a pronounced information gap, and substantial unmet care needs. These patient-centered findings support structured, risk-adapted LTFU with proactive information, integrated psychological support, and sustainable financing at specialized LTFU centers. Full article

Background/Objectives: Molecular docking and molecular dynamics are accurate but computationally expensive, so the compounds entering them must be chosen well. The present study proposes CADT, a confidence-gated affinity–ADME-T docking-triage cascade that decides which compounds are worth docking. Methods: The gate combines an ensemble estimate of drug–target affinity with its epistemic uncertainty and an applicability-domain check. Predicted absorption, distribution, metabolism, excretion, and toxicity (ADME-T) developability is added as a soft flag. All components were trained on openly licensed Therapeutics Data Commons data. Ranking was assessed on the DAVIS and KIBA kinase panels and on BindingDB $K_d$, under three split protocols over five seeds. The routing decision was then examined against molecular docking, in which 407 compound–target pairs were docked into six withheld kinases. Results: A Morgan-fingerprint gradient-boosting model reached a concordance index of $0.866\pm 0.006$, with $0.813$ for unseen targets and $0.720$ for unseen drugs. Across eight ADME-T endpoints, the area under the ROC curve ranged from $0.65$ to $0.91$. On the cold-target split the cascade reduced the compounds sent to docking by $86\%$ while retaining $61\%$ of the true strong binders. Docking measured that reduction at $85\%$, and at an equal budget, the gate enriched true binders more than the docking score itself. Conclusions: A transparent pre-screen can prioritise compounds ahead of structure-based calculation at a fraction of its cost. However, the uncertainty and applicability-domain terms act as an abstention mechanism rather than an accuracy gain, and that abstention is not free. Full article

Access to reliable medical images is essential for clinical training. To address this need, this paper presents an extended version of MIRAGE, a multimodal retrieval and generation system that utilizes a shared latent space to process medical queries by retrieving real images from the ROCO dataset, generating synthetic scans, and providing LLM-based clinical descriptions alongside dual-concept visual comparisons. To overcome previous computational limits and the lack of clinical validation, we introduce three core enhancements: first, an Auto-$\alpha$ module to dynamically weight visual and textual similarities; second, the integration of LCM-LoRA to accelerate synthetic image generation; and third, an automated clinical auditor based on Gemini 2.5 Flash. Experimental results demonstrate that Auto-$\alpha$ improves retrieval accuracy for heterogeneous queries, reaching $38.83$% Top-1 Recall over a 65,419-image gallery and outperforming nine fusion baselines evaluated under a unified configuration, with a controlled ablation attributing most of this gain to learning the weight rather than merely making it query-adaptive, while the LCM-LoRA module reduces computational costs by a factor of 12.5× in CPU environments, with a blinded radiologist evaluation confirming only a small drop in clinical quality. Furthermore, the clinical auditor achieves a 0.805 Pearson correlation against an expert radiologist, effectively correcting the systematic overestimation of traditional CLIP scores. Finally, the optimized platform is publicly deployed on Hugging Face. Full article

Convolution is spatially fixed and channel-specific, whereas involution is location-specific and channel-agnostic, capturing spatially varying patterns efficiently. Existing involutional networks, however, generate point-estimate kernels and expose no native measure of where the operator is uncertain, while prevailing uncertainty and calibration methods act on the network output rather than on the aggregation operator itself. We propose Evidential Involution–Attention (EvIA) networks, which recast involution as a distributional operator whose per-location neighborhood aggregation is a Dirichlet distribution. This yields, in a single forward pass and at the same parameter cost as involution, a closed-form epistemic-uncertainty (vacuity) map. The vacuity drives a parameter-free, precision-weighted gate that fuses the local involution branch with a global branch, instantiated as windowed self-attention (EvIA-W) or lightweight channel attention (EvIA-C). A lemma and three propositions establish that normalized involution is the infinite-evidence limit of the operator, that convex aggregation makes it non-expansive, that the Dirichlet strength is the precision of the aggregated feature, and that the gate is the minimum-variance unbiased fusion of the two branches. An evidential head trained with a differentiable calibration objective produces reliable confidences. Over five seeds with paired tests on brain magnetic resonance imaging, chest radiography, and dermatoscopy, the windowed variant EvIA-W is significantly stronger on brain MRI, attaining 0.803 ± 0.026 accuracy against 0.714 ± 0.019 (p = 0.006) by EvIA-C, and reducing the area under the risk–coverage curve from 0.187 to 0.092 (p = 0.001). Architecture-matched controls show that the discrimination gain on brain MRI comes from the global branch, while substituting evidential for standard involution leaves accuracy, calibration, and selective risk statistically unchanged, so the operator supplies its uncertainty machinery at no measurable cost. We further report that the spatial vacuity map does not localize input corruption, because the aggregation weights are invariant to the evidence scale and no objective term supervises it, an analysis that motivates the operator-level regularizers we define for future work. Full article

This study focuses on a three-step iterative scheme, referred to as the NIP iteration, for the approximation of fixed points associated with nonexpansive mappings in uniformly convex Banach spaces. Weak convergence is established using Fejér monotonicity, asymptotic regularity and the demiclosedness principle. Strong convergence is proved under uniform convexity, compactness, and Condition (I) of Senter and Dotson. A numerical convergence and computational-cost comparison is developed numerically, showing that the NIP iteration performs better than the Ishikawa, S, Noor, Abbas–Nazir and SP schemes. Numerical experiments for nonlinear nonexpansive mappings validate the theoretical findings. An application to convex optimization via fixed point reformulation is also presented, illustrating the effectiveness of the method. Full article

Background/Objectives: Schatzker type is assigned early, usually from the anteroposterior (AP) radiograph. A single benchmark accuracy cannot say whether a model read the fracture, the anatomy around it, the annotation, or how the archive was assembled. We ran four inexpensive controls to separate those contributions. Methods: We benchmarked a ResNet-50 on PlaTiF, a 2026 public release built for artificial-intelligence research that pairs 421 AP knee radiographs from 186 patients with expert Schatzker labels and per-image tibial segmentations. Evaluation used stratified group five-fold cross-validation grouped by patient, five seeds and balanced accuracy. Inputs were cropped to the expert tibial segmentation shipped with the dataset, an oracle localisation unavailable at deployment. Four controls ran on identical folds: a regression given no pixel content; ablation of the tibial pixels with its complement; a regression on the expert mask alone; and an augmentation audit for label-erasing invariances. Results: Among the 128 fracture patients the network reached 0.345 ± 0.030 six-class balanced accuracy, +0.168 over a non-anatomical baseline fitted on the same folds and the same labels (95% CI +0.106 to +0.230, p = 0.002). Recall was graded: 0.72 for Schatzker VI, 0.11 for V and 0.04 for IV, the last two below chance (0.167). Erasing the tibial pixels left 0.257 ± 0.025, read on its own as the target bone being unused; its complement, the tibia with everything else removed, reached 0.367 ± 0.012, and the whole radiograph, which carries both, only 0.297 ± 0.032 (+0.071 for the tibia alone, 95% CI +0.031 to +0.111, p = 0.008). A regression on the expert mask alone reached 0.213 ± 0.034 and was not distinguishable from the erased model. On fracture versus no classifiable fracture the network reached 0.833 ± 0.028 against 0.814 ± 0.016 for a model given no pixels (p = 0.264), and a coronal computed tomography section accompanied 126 of 128 fracture patients but 24 of 58 others (p = 2.9 × 10⁻¹⁹). Conclusions: Each headline number admitted an explanation other than the fracture in the target bone. An ablation reported without its complement misstated where the signal lay, and an augmentation audit overturned our own explanation for the failure of type IV. Controls of this kind cost minutes, and this study illustrates why they can be informative when a benchmark is built on a retrospective clinical archive. Full article

Coal-fired power generation remains a major source of carbon dioxide (CO₂) emissions, and metal–organic framework-74 (MOF-74) materials offer high adsorption capacity but rely on costly 2,5-dihydroxyterephthalic acid linkers that limit scalability. This study synthesized bimetallic Zn_0.75Mg_0.25-MOF-74 using terephthalic acid (TPA) as a cheaper alternative linker and evaluated the effect of synthesis reaction temperature (89–160 °C) and time (5–55.5 h) on its physicochemical properties for carbon capture. Samples were prepared solvothermally and characterized by FTIR, XRD, SEM-EDS, and N₂ (77 K) and CO₂ (293 K) adsorption analysis. FTIR confirmed metal–ligand coordination; XRD verified crystalline MOF-74 formation, and SEM showed well-defined rod-like morphology at 100 °C, 12 h and 125 °C, 30 h. Direct CO₂ adsorption on the 125 °C, 30 h sample yielded a Type I isotherm characteristic of micropore filling, with an uptake of 0.31 mmol/g at ~1 bar and 0.072 mmol/g at flue-gas-relevant conditions (~0.135 bar). Its CO₂-derived BET surface area (60.10 m²/g) and Dubinin–Astakhov micropore area (131.69 m²/g) far exceeded N₂-derived values, confirming ultra-micropores accessible to CO₂ but not to N₂ at 77 K. TPA therefore yields a stable, microporous CO₂-adsorbing framework, trading some capacity for lower cost and scalability. Future investigations should systematically evaluate long-term cycling stability and adsorption performance under mixed-gas operating conditions. Full article

High-dimensional gene expression data pose computational challenges for feature selection due to redundancy, noise, and limited sample sizes. Recursive Cluster Elimination (RCE) addresses these issues through iterative feature clustering and elimination. Conventional implementations of RCE require repeated cross-validation steps within each cluster that increase the computational cost. The RCE with Center Weights (RCE-CW) framework addresses this by evaluating all clusters simultaneously through a single linear model. This study extends RCE-CW by replacing synthetic cluster centroids with medoid representatives, introducing Medoid-SVM-RCE and Medoid-RF-RCE. Experiments on 17 gene expression datasets show that Medoid-based variants select approximately 46–51% fewer genes than RCE-CW and SVM-RCE, while maintaining statistically comparable classification performance. Across 17 datasets, Medoid-based variants completed all experiments within 9 h, whereas SVM-RCE did not complete within 24 h on 11 of the 17 datasets. These results indicate that medoid-based cluster representation reduces gene subset size and computational cost within the RCE framework while preserving predictive performance. Full article

Background/Objectives: Obesity care requires nutritional prescriptions that are individualized, nutritionally coherent, understandable, and feasible within high-volume clinical services. We describe the NUTRI-OBEDIGITHUM pathway, which integrates multidimensional digital phenotyping, clinician-supervised exchange-based nutritional prescription, patient-oriented materials, and structured education, and we report preliminary descriptive outcomes from a retrospective real-world cohort. Methods: This descriptive methodological and operational study included a retrospective observational analysis of anonymized routine-care data. NUTRI-OBEDIGITHUM integrates OBEDIGITHUM, a digitally enabled obesity-care pathway, with Dietcamb^®, a software-supported method that translates individualized energy and macronutrient targets into food exchanges, meal distributions, household measures, food lists, and example menus. The source database comprised 3928 unique patients. Among patients with a plausible baseline weight and an observable 6-month window, the eligible outcome was the measurement closest to day 183 within ±60 days, after prespecified plausibility checks. Baseline characteristics were compared descriptively according to outcome availability. Baseline medication exposure was assessed using free-text mentions only. Results: By the data cutoff, the 6-month window had opened for 3215 patients; 929 (28.9%) had an eligible 6-month weight and 2286 (71.1%) did not. In the analyzed cohort, mean age was 53.7 ± 14.0 years, 70.5% were women, and mean body mass index was 35.9 ± 6.3 kg/m². Mean weight changed from 98.5 ± 20.7 kg at baseline to 94.2 ± 20.3 kg at follow-up, corresponding to a descriptive paired change of −4.3 kg (95% CI −4.7 to −3.9); median total body-weight loss was 3.4%. Losses of at least 5%, 10%, and 15% were observed in 39.6%, 17.8%, and 6.4% of patients, respectively. A baseline GLP-1 receptor-agonist mention was present in 151/929 patients (16.3%); medication timing, dose, duration, dispensing, persistence, and adherence were unavailable. Conclusions: NUTRI-OBEDIGITHUM provides an operational framework for converting multidimensional obesity phenotyping into a structured, flexible, clinician-supervised, and patient-oriented nutritional prescription. The observed weight trajectories are descriptive complete-case findings from a selected, uncontrolled cohort and cannot establish the independent effectiveness of the pathway, Dietcamb^®, or pharmacotherapy. Prospective studies should evaluate reach, adoption, fidelity, usability, adherence, nutritional adequacy, safety, NUTRI-specific workload and cost, and metabolic outcomes. Full article

The initiation of detonation is a critical yet challenging task for pulse and rotating detonation engines. Conventional approaches often rely on nanosecond repetitively pulsed discharges to generate nonequilibrium plasma for ignition assistance, but the complexity and high cost of the required power supplies limit practical applications. In this work, we experimentally investigate a simplified method using ordinary electrode breakdown discharge to produce an arc plasma that promotes jet-induced detonation initiation. Two ignition strategies are compared under the same total energy: dual-spark-plug ignition (energy concentrated at the jet tube head) and single-spark-plug coupled with electrode discharge (energy split between the jet tube head and an electrode pair placed near the detonation chamber inlet). High-speed schlieren measurements are performed to capture the dynamic flame evolution and shock wave structures. The results show that the electrode-discharge approach dramatically increases the detonation success rate from 13.33% to 66.67% over 30 repeated runs. The electrode discharge is found to occur after the emerging flame has already covered the electrodes. Therefore, the promoting mechanism is attributed not to the high temperature or free radicals generated in the already-burned products, but rather to the discharge-induced shock wave. This shock wave interacts with the corner expansion waves generated by the sudden area expansion, thereby delaying the unsteady decay of the leading shock and promoting re-initiation. This study provides the first experimental evidence that ordinary electrode breakdown discharge promotes jet-induced detonation via a shock-wave reinforcement mechanism. The findings enable a low-cost, compact plasma-assisted initiation strategy for practical detonation engines. Full article

Distributed Denial of Service (DDoS) attacks are one of the most important cybersecurity problems today since they are one of the biggest threats against network service continuity and accessibility. Increasing network traffic volume combined with heterogeneous data structures and variations in attack types creates greater complexity in detection and mitigation. Traditional methods based on signatures and static rule sets are not adequate, especially for unknown and evolving attack types. In recent years, deep learning techniques have become a powerful alternative for DDoS detection systems with their ability to automatically extract features from high-volume network traffic data. This study tries to give a detailed comparison of deep learning-based DDoS detection approaches by examining their datasets, model architecture, evaluation metrics, and computational limitations. The review shows that many recent studies have only focused on accuracy; however, critical system parameters such as latency, computational cost, and energy consumption are insufficiently evaluated. Furthermore, generalization problems and dependencies on specific datasets create important vulnerabilities. In this context, this work tries to define future discussions for resource-aware, generalizable, and real-time DDoS detection systems. Full article

Early-exit deep neural networks (DNNs) can reduce edge-inference latency, but abrupt variations in wireless and computing resources can destabilize split-inference policies. This paper proposes a meta-learning-driven adaptive control framework for joint backbone splitting and early-exit routing in MobileViT. The framework formulates multi-exit splitting as a constrained Markov decision process (CMDP) and introduces splitting-aware multi-dimensional adaptive proximal policy optimization (SMAPPO). SMAPPO combines nonlinear quality-of-service (QoS) penalties with topology-aware action masking, while cross-environment meta-initialization supports edge-local adaptation after resource disturbances. Under the stated simulation assumptions, SMAPPO reached the highest performance-index plateau among six methods in a representative 500-episode stationary trace and achieved the lowest normalized total cost across three latency–energy preference settings. Across ten seeds and nine stationary or disturbed scenarios, online SMAPPO achieved a 77.20% measured accuracy and 22.40 mJ of system energy. With an adaptation horizon of $K=14$, SMAPPO yielded a post-disturbance mean latency of 37.68 ms, a QoS-violation rate of 2.24%, and an on-time completion rate of 98.69%. These results indicate that combining meta-initialization, nonlinear constraint shaping, and topology-aware action masking improves stationary optimization and disturbance recovery within the controlled simulator. Full article

Electroencephalography (EEG) provides a non-invasive and cost-effective tool for supporting the assessment of attention-deficit/hyperactivity disorder (ADHD). However, the nonstationary nature of EEG produces substantial variability among signal windows recorded from the same participant, which can obscure diagnostic structure and lead to inconsistent predictions. To address this problem, we propose the Contextualized Transfer Entropy Network (CTE-Net), an end-to-end deep-learning architecture that combines global content-based contextualization with nonlinear and directed EEG connectivity estimation. CTE-Net first employs a Transformer encoder to contextualize the multichannel representations within each EEG window. The resulting signals are processed using channel-wise nonlinear temporal filters and Takens delay-coordinate embeddings. A differentiable matrix-based Transfer Entropy module, formulated using Rényi’s $\alpha$-entropy and a rational quadratic kernel, then estimates directed predictive information dependencies between all ordered electrode pairs. The resulting connectivity coefficients are used for ADHD-versus-control classification. The model was evaluated on a publicly available pediatric EEG dataset comprising 120 participants, equally divided between ADHD and control groups, using five fixed subject-wise folds and ten random training repetitions. At the window level, CTE-Net achieved an accuracy of $80.9\pm 1.7\%$, precision of $82.7\pm 2.1\%$, and sensitivity of $84.2\pm 2.3\%$. At the participant level, it achieved an accuracy of $83.4\%$ (95% CI: $78.2$–$88.2$) and an ROC-AUC of $90.2\%$ (95% CI: $85.1$–$94.6$), demonstrating competitive and comparatively balanced classification performance. Beyond classification performance, the directed Transfer Entropy representation exhibited the lowest within-subject dispersion among the analyzed representation stages, with a median reduction of $38.35\%$ relative to raw EEG. This reduction remained consistent across different PCA dimensionalities and distance definitions. These single-dataset findings support CTE-Net as a compact and interpretable methodological framework for representing directed EEG interactions while attenuating window-specific variability within individual participants. Full article

Cyber incidents can disrupt many virtual power plant (VPP) assets through shared software and communication services. This study links preventive finance, cyber defense, dispatch, and restoration in one multi-timescale model. An attacker, a VPP operator, a bond vehicle, and a regulator interact in a partially observable stochastic game. The operator controls hardening, dispatch, isolation, and recovery. The bond provides restricted pre-event capital and releases collateral through an auditable index of service loss, control availability, network stress, and recovery delay. A risk-constrained multi-agent policy enforces power-system feasibility, investor impairment, sponsor affordability, and trigger–loss limits. Tests use transparent synthetic VPP-39 and VPP-118 portfolios and 20 out-of-sample seeds. The proposed design lowers normalized social cost to 0.691 and 0.704 and weighted basis risk to 0.065 and 0.071. It also improves critical-load continuity and restoration relative to self-insurance and three bond baselines. The VPP-118 case recovers in 11.8 h, compared with 14.3 h for the closest rule-based benchmark. Ablations separate the effects of finance and control. Removing the coupon–control link reduces verified hardening from 0.672 to 0.519. Removing the safety projection raises unsafe proposals from 0.4% to 5.9%. These results show that stochastic multi-timescale control can support adaptable and resilient VPP operation while keeping the financial mechanism within explicit risk limits. Full article