Search Results (123)

Background: Accurate prognosis in diffuse large B-cell lymphoma (DLBCL) is limited by biological heterogeneity and the absence of formal per-patient uncertainty quantification for treatment-response prediction. This study introduces a multi-layer evidence fusion framework combining gene expression profiling and clinical features with distribution-free uncertainty quantification. Methods: The proposed framework integrates four evidence layers—WGCNA co-expression eigengenes, ssGSEA pathway scores, bootstrap-stable prognostic genes, and the International Prognostic Index—through AUC-proportional reliability discounting and sequential Dempster–Shafer fusion. The primary endpoint was three-year overall survival (OS₃yr) as a surrogate for R-CHOP treatment response. Inductive conformal prediction (ICP, ε = 0.10) was applied to provide per-patient uncertainty sets with a distribution-free coverage guarantee. Training used GSE10846 (n = 223, Affymetrix); external validation used GSE181063 (n = 479, Illumina). Results: The proposed framework achieved internal AUC = 0.808 (95% CI [0.750, 0.863]), significantly outperforming logistic stacking (AUC = 0.786, p = 0.0009) and unweighted DS fusion (AUC = 0.767, p = 0.037). External AUC = 0.791 was statistically comparable to logistic stacking (DeLong p = 0.21). AUC-proportional discounting reduced inter-source conflict K^- by 75% (0.093→0.023). ICP achieved 90.1% internal and 94.6% external coverage; 43.5% of training patients received uncertain predictions ({S,R}). Conclusions: The proposed framework provides an uncertainty-aware approach for multi-layer genomic–clinical evidence fusion in DLBCL, with cross-platform discrimination validated on an independent Illumina cohort. Full article

Background/Objectives: CT-based trabecular attenuation in Hounsfield units (HU) is used as a surrogate marker of bone quality, but most approaches rely on a single vertebral level, usually L1, where local abnormalities may limit reliability and availability. This study evaluated whether multilevel lumbar, reference-adjusted HU assessment beyond isolated L1 improves the identification of QCT-defined osteoporosis and the discrimination of prevalent fragility fracture status. Methods: Between 2021 and March 2024, 800 patients referred for evaluation of bone mineral density underwent QCT of the lumbar spine with a calibration phantom. Cancellous attenuation was measured in HU at L1–L3 using manually positioned ellipsoid regions of interest, with predefined adjacent levels (T12, L4) substituted when a target vertebra was unsuitable. Single-level HU, the mean multilevel HU value and the lowest valid HU value were compared with QCT-defined osteoporosis and prevalent fragility fracture status (vertebral or sacral) using ROC analyses and pairwise DeLong testing. Results: Mean multilevel HU correlated closely with mean vBMD (Spearman ρ = 0.988; p < 0.001); as both derive from the same acquisition, the same vertebral bodies and the same phantom, this reflects a strong association between two related measurements rather than validation against an independent standard. Osteoporosis was present in 483 patients (60.4%). Mean multilevel HU was associated with osteoporosis with an AUC of 0.996 (95% CI, 0.993–0.998) at an internally derived cut-off of ≤99.6 HU. In the 699 patients with a valid L1 measurement, in whom a paired comparison was possible, mean multilevel HU exceeded isolated L1-HU (0.992; p = 0.0028). For fracture status, the AUC was 0.975 (0.966–0.983) at ≤80.5 HU, without advantage over L1-HU (p = 0.816). A valid value was obtainable in all 800 patients versus 699 (87.4%) with L1 alone. All patients with a fracture had a vBMD below the osteoporosis threshold, so this endpoint is structurally dependent on the densitometric one. Conclusions: Multilevel lumbar HU assessment provides a practical marker of trabecular bone quality beyond isolated L1. Its principal advantage is measurement availability rather than a clinically meaningful gain in discrimination. The reported thresholds are internally derived, cohort- and protocol-specific, and require external validation. Full article

Urban transmission of dengue and other Aedes aegypti-borne diseases is shaped by the interaction of vector ecology, human mobility, climate, and spatial heterogeneity. Capturing these interactions in city-specific settings requires models that are detailed enough to represent local transmission processes, while remaining computationally feasible for calibration, validation, and sensitivity analysis. This study presents a hybrid agent-based modeling and simulation (HABMS) approach, supported by high-performance computing (HPC), for simulating urban vector-borne disease transmission. Human residents are represented as mobile agents with stochastic infection dynamics, while mosquito populations are represented at the patch level through discrete-time equations. The model incorporates geospatial structure, temperature, land-use-based human movement, and local contextual information to represent transmission within urban environments. The framework was applied to Santa Marta, Colombia, as a city-specific case study. Transmission parameters were calibrated using surrogate-based Bayesian optimization, and their influence was assessed through sensitivity analysis. High-performance computing made the large number of stochastic simulations required for calibration and sensitivity analysis feasible. The calibrated model reproduced the magnitude and main seasonal shape of the observed dengue epidemic, including the peak and early decline. However, the model did not fully reproduce the late low-incidence tail of the season, indicating the need to incorporate external introductions of infection and rainfall-driven seasonal forcing of vector recruitment in future versions. A control scenario run on the calibrated baseline, a 30% reduction in larval carrying capacity, lowered the simulated seasonal attack rate by about three quarters and moved the system below the threshold at which local transmission is self-sustaining, illustrating the relative comparisons the calibrated model supports. This study contributes an adaptable hybrid model architecture and a high-performance implementation that make city-specific calibration and sensitivity analysis computationally feasible, demonstrated through a case study in Santa Marta. Full article

Rapid cooperative detection of stationary targets by multiple unmanned aerial vehicles (UAVs) is important in time-critical missions such as search and rescue. However, the online coordination of probabilistic inference, distributed communication, and detection–motion decisions under local information remain challenging when targets exhibit spatial correlations that existing methods typically neglect. To address this challenge, we develop a distributed event-driven Bayesian search framework for stationary, spatially correlated targets at unknown locations. The framework couples three components. A pairwise spatial model and a distance-dependent Neyman–Pearson detector yield a Bayesian belief update whose unclipped product form is order-invariant to event-processing sequence. A distributed selective flooding algorithm propagates only positive detection events, achieving finite-time event-set consensus over connected graphs while avoiding full-map exchange. A decoupled detection–motion planner exhausts high-belief cells within each UAV’s field of view before selecting a waypoint that balances surrogate detection probability against travel cost, with responsibility regions dynamically renegotiated among neighbors when local high-value cells are depleted. In numerical experiments, the proposed method achieved zero uncoordinated repeat detection in all simulations and significantly reduced first-discovery coverage relative to static-partition and no-communication baselines, while adapted external baselines required 90-fold and 6-fold larger communication payloads and had nonzero repeat-detection rates. The framework thus occupies a specific tradeoff point of zero revisit, sparse communication, and early discovery gain in scenes where targets span multiple UAV search regions. Full article

Background: Patients with coronary artery disease (CAD) receiving dual antiplatelet therapy (DAPT) after percutaneous coronary intervention (PCI) remain at risk of early adverse outcomes, including in-hospital mortality. Simple risk stratification based on routinely available variables may help identify higher-risk patients, but a limited number of outcome events constrains robust prediction-model development and validation. Aim: This exploratory study aimed to derive a preliminary, interpretable clinical score based on routinely available variables for risk stratification of all-cause in-hospital mortality in CAD patients receiving DAPT after PCI. In the clopidogrel-dominant practice setting of the participating centers, the score was conceived as a hypothesis-generating risk-enrichment framework rather than a validated treatment-selection tool or a surrogate measure of platelet reactivity. Methods: We analyzed a retrospective cohort of 1600 adults with CAD admitted between 2022 and 2024; 36 in-hospital deaths occurred. Twenty demographic, clinical, laboratory, and instrumental variables were evaluated. The primary outcome was all-cause in-hospital mortality during the index hospitalization. For exploratory score derivation, the dataset was randomly divided into a derivation subset (75%; n = 1200) and a hold-out assessment subset (25%; n = 400). Predictors were explored using univariable and multivariable logistic regression with stepwise selection. Continuous variables were categorized using Weight of Evidence binning, and an integer point score was derived. Performance was summarized using ROC analysis, AUC, sensitivity, specificity, and accuracy. Given the small number of deaths and the data-driven modelling workflow, all performance estimates were considered preliminary rather than definitive internal validation. Results: The exploratory six-variable score included age ≥ 57 years, estimated glomerular filtration rate < 45 mL/min/1.73 m², body mass index ≥ 25 kg/m², troponin I ≥ 100, prior myocardial infarction, and current smoking. In the derivation subset, each additional point was associated with higher odds of mortality (OR 1.39; 95% CI 1.29–1.51; p < 0.001), and the AUC was 0.654. A Youden-index threshold of approximately 6 points yielded sensitivity of 0.41, specificity of 0.80, and accuracy of 0.72. In the hold-out assessment subset, sensitivity was 0.53, specificity was 0.70, accuracy was 0.70, and AUC was 0.61. These estimates indicate modest discrimination and should be interpreted cautiously because only 36 outcome events were available. Conclusions: This exploratory clinical score showed modest discrimination for all-cause in-hospital mortality and should be regarded as a preliminary, hypothesis-generating risk-stratification approach. It is not sufficiently validated for routine prognostic classification, platelet-reactivity triage, or antiplatelet treatment selection. Model redevelopment using event-efficient methods, resampling-based internal validation, and subsequent external validation are required before clinical implementation. Full article

Advances in machine learning, predictive analytics, and clinical prediction modeling have accelerated the development of algorithmic tools for estimating reproductive outcomes after cancer treatment. In female oncofertility counseling, these models may support individualized assessment of treatment-related amenorrhea, premature ovarian insufficiency, and fertility risk, thereby improving risk communication and timely fertility-preservation referral. However, their use raises ethical concerns beyond predictive accuracy. This narrative review examines algorithmic prognostication in female oncofertility counseling, focusing on predictive uncertainty, surrogate reproductive endpoints, missing data, heterogeneous datasets, limited external validation, algorithmic bias, reproductive inequity, and the influence of algorithmic authority on patient autonomy and shared decision-making. We argue that predictive algorithms should be understood as decision-support tools rather than determinants of reproductive futures. Responsible implementation requires transparency, explainability, fairness assessment, ongoing validation, and meaningful human oversight. Algorithmic risk estimates should be communicated as conditional and contextual probabilities within patient-centered counseling, ensuring that predictive tools support informed, transparent, and value-concordant fertility-preservation decisions for women facing cancer treatment. Full article

Topological indices provide numerical descriptions of graph structure and support graph analysis in cheminformatics, network design, and graph mining. This study presents a reproducible computational framework that combines controlled cyclic-graph generation, structure-preserving transformations, exact computation of six classical topological indices, and multi-output graph neural network regression. The framework evaluates the Wiener, Merrifield–Simmons, Hosoya, first Zagreb, second Zagreb, and Randi’c indices for unicyclic and bicyclic graphs. It represents each graph using sparse connectivity and node-level features that encode degree, cycle membership, pendant connectivity, leaf status, and normalized eccentricity. A graph isomorphism network (GIN) jointly predicts the six indices and is compared with graph convolutional networks (GCNs), graph attention networks (GATs), and descriptor-based regression baselines. The controlled benchmark shows that nonlinear descriptor-based models achieve the lowest aggregate errors because the supplied graph-level descriptors contain strong prior information about graph size, degree structure, branching, and cycle complexity. Although GIN does not achieve the highest overall accuracy, it provides the strongest graph-native performance by learning directly from sparse connectivity and node-level features without requiring a fixed handcrafted graph-level descriptor vector. The proposed surrogate does not replace exact evaluation for isolated small graphs, where exact computation remains more appropriate. Instead, its practical value emerges through repeated evaluations of larger, more complex graph instances. To examine this setting, a computational stress experiment evaluates sparse multicyclic graphs under increasing cyclomatic complexity and measures exact computation time, timeout frequency, prediction accuracy, and the amortized break-even point. The results indicate that surrogate prediction becomes beneficial when combinatorial index computation becomes sufficiently expensive, and the trained model is reused across many structurally related graph queries. An external experiment on circulant graphs also demonstrates that the framework can extend beyond the original graph generators by modifying only the graph-construction stage. Full article

Broadband antenna, frequency and channel selection requires efficient prioritisation of finite candidate configurations, yet strong surrogate performance within a simplified analytical model does not ensure that the learned ordering will transfer to another electromagnetic representation. This study developed a validation-aware surrogate-shortlisting framework using a controlled breast-mimetic benchmark comprising 120 scenarios and 80 antenna–frequency–channel candidates per scenario. Surrogate models were developed using grouped scenario-level validation, and the model and five-candidate shortlist policy were frozen before external evaluation. Random forest produced the lowest-regret analytical-domain shortlist and remained stable across model-initialisation seeds. The frozen ranking was then challenged on held-out scenarios using a separately implemented restricted two-dimensional transverse-magnetic finite-difference time-domain model. Although numerical-reference checks supported shortlist-level use of the operational grid, candidate ordering did not transfer reliably: optimum inclusion, shortlist agreement and rank association were weak, although a qualified candidate within 2 mm of the finite-library FDTD optimum was retained in 58.3% of cases. Transfer failure varied by frequency band and channel family, while incomplete alignment between the analytical and FDTD candidate libraries prevented attribution of the discrepancy to electromagnetic-model shift alone. The framework therefore positions analytical surrogates as auditable shortlisting tools that reduce downstream candidate-level assessment while retaining independent electromagnetic evaluation before final design selection. Full article

Background: Retroperitoneoscopic laparoscopic adrenalectomy (RLA) is widely adopted for benign adrenal lesions due to its direct access and minimal bowel interference. However, expanding indications to larger tumors and complex anatomies increase technical difficulty because of the confined retroperitoneal space and obscured hilar landmarks. Current predictive models mainly focus on transperitoneal approaches and lack specificity for RLA. This study aimed to develop and internally validate a preoperative nomogram for predicting RLA-specific surgical difficulty. Methods: All patients undergoing RLA from April 2024 to February 2025 at a single center were included. Operative time and postoperative hemoglobin (Hb) reduction were used as surrogate markers of difficulty. Patients were classified into high- and standard-risk groups. Univariate and multivariate analyses were performed to identify predictors of difficult RLA and develop a composite difficulty score. Results: Tumor position (between the upper renal pole and renal pedicle: OR 8.819, 95% CI 3.981–21.462, p < 0.001) and pheochromocytoma (PHEO) pathology (OR 34.881, 95% CI 3.841–4719.428, p < 0.001) were the strongest independent predictors of high surgical difficulty, whereas tumor size (≥40 mm: OR 3.926, p = 0.071) showed limited predictive value. The composite score demonstrated excellent discrimination, with an optimism-corrected area under the curve (AUC) of 0.828 (95% CI 0.796–0.835). The nomogram showed robust calibration, and a cutoff score ≥ 2 yielded a negative predictive value of 0.940. Conclusions: This study developed an RLA-specific nomogram incorporating tumor size, position, and pathology. The tumor–renal pedicle relationship was more predictive of surgical difficulty than tumor size alone, providing risk stratification to guide surgical planning, optimize patient selection, and enhance safety as RLA indications expand. External validation is required to confirm its generalizability. Full article

Internal and external stress exerted by forces such as oxidative, xenobiotic, radiant, ischemic, mechanical, metabolic, neurohormonal, infectious, immunoinflammatory, emotional and psychological increases the risk of cardiovascular morbidity and mortality. Initiation and propagation of the explosive self-amplifying chain reaction of lipid peroxidation chain reaction (LPO) generates atherogenic and inflammatory toxic reactive oxygen species that cause cardiovascular tissue lesion that are the ultimate determinants of cardiovascular disease risk (CVD-R). Because the LPO toxins, cellular stress sensors and defenses, inter- and intracellular signaling, and pathogenic lesions are closely similar among these stressors, simultaneous exposure to multiple stresses can amplify LPO to accelerate accumulation of pathogenic lesions synergistically. Numerous blood tests measure LPO-derived toxins, stress-responsive metabolism and cytokines, which are used clinically as surrogate biomarkers of CVD-R. They can predict population risks quite well, but prediction of individual patient CVD-R using multiplex biomarker panels is hampered by lack of true independence between biomarkers, lack of understanding of their relative hierarchy in disease etiology or progression, and interference from comorbid diseases or acute-phase reactions. We present this historical review of landmark studies that led to the current clinical paradigms of CVD-R prediction to provide mechanistic and clinical context that will aid the development of integrated etiological biomarkers that reflect the multiple types of stress that promote CVD-R. Full article

To address low seeding accuracy and poor seed-fertilization coordination caused by wheel slip and vibration in undulating terrains, a model 2BJD-4 precision corn planter featuring an independent electric-drive transmission was developed. The planter integrates furrow opening, fertilization, single-seed precision metering, soil covering, and compaction into a coordinated one-pass operation. Key mechanical assemblies include a servo-motor-driven finger-clamp seed meter, a parallel four-bar terrain-following mechanism, and an external fluted-roller fertilization meter. To capture complex non-linear soil-tool interactions, a predictive surrogate model was established using Support Vector Regression (SVR) and coupled with the Dung Beetle Optimizer (DBO) for global parameter optimization. Comprehensive field trials validated that the SVR-DBO framework outperformed traditional Response Surface Methodology, securing an optimal qualified spacing index of 92.8% and a planting depth qualification rate of 93.0% under experimental conditions. These findings demonstrate the technical feasibility of the proposed design in maintaining seed spacing and depth uniformity under tested topographies, offering a practical reference for the development of precision planters in hilly and plain regions. Full article

Surrogate endpoints are increasingly used in chronic lymphocytic leukemia (CLL) to accelerate treatment evaluation, but their validity is context-dependent. This review examines key endpoints—progression-free survival (PFS), time to next treatment (TTNT), measurable residual disease (MRD), and quality of life (QoL)/patient-reported outcomes (PROs). PFS, while standard, is limited by competing risks and poor reflection of toxicity and patient experience. TTNT captures both efficacy and tolerability but is influenced by external factors. MRD is a strong predictor of outcomes in fixed-duration venetoclax-based regimens but less reliable in continuous therapies. QoL and PROs provide essential patient-centered insight often missed by traditional endpoints. We propose a practical framework in which endpoint selection depends on treatment type, patient characteristics, and intended use. MRD is most informative after fixed-duration therapy, TTNT in continuous treatment, and PROs in vulnerable populations. Overall, surrogate endpoints in CLL require setting-specific validation to ensure they reflect meaningful clinical benefit. Full article

Near-fault pulse-like ground motions produce coupled intensity, duration, and period-matching effects, making nonlinear seismic assessment of multi-ribbed composite slab structures (MCSS) computationally expensive and difficult to generalize. To address this problem, a 4000-case OpenSees nonlinear time-history analysis (NLTHA) database is generated from Wenchuan ground motions through Latin hypercube sampling, and a mechanically informed feature-enhanced deep neural network (MIFE-DNN, previously denoted as PE-DNN in the first submission) is trained using equivalent stiffness, equivalent yield strength, mass proxy, demand-capacity ratios, period-matching ratio, normalized duration, and energy-capacity proxy; a validation-weighted stacked surrogate is further constructed from multi-seed MIFE-DNN and residual learners. On the independent test set, the mean $R^2$ increases from 0.9645 for the ordinary deep neural network (DNN) and 0.9686 for the single MIFE-DNN to 0.9782 for the stacked mechanically informed surrogate, while the maximum inter-story drift-ratio $R^2$ reaches 0.9541. Additional checks include 16 active-learning OpenSees enrichment cases, 12 analyses under two external near-fault records, 3 out-of-domain parameter cases, 100 cross-story tests, SHAP-based interpretation, and multi-EDP fragility post-processing. These checks show that the surrogate is reliable for interpolation and screening within the calibrated equivalent-model domain, but direct OpenSees recalculation is required for boundary, out-of-domain, and cross-configuration use. Parameter-importance, SHAP, and fragility analyses identify peak ground acceleration (PGA), pulse index, period matching, rib height, rib spacing, and damping ratio as dominant factors, indicating that mechanically informed feature-enhanced surrogate modeling provides an interpretable and efficient tool for MCSS response prediction and conditional fragility assessment within the sampled structural and ground-motion domain. Full article

Adherence to right-of-way (RoW) rules at uncontrolled T-intersections helps avoid accidents and alleviate congestion. In non-uniform traffic, right-turning behaviour can be characterised by distinct driving traits, such as non-compliance (failure to yield), a nonchalant attitude, and competitive behaviour. This paper presents a cost-effective computer vision framework using UAV videos to analyse right-turning behaviour and assess safety and operational performance (congestion) at uncontrolled T-intersections. A conflict cone of a vehicle is defined to automatically detect a right-of-way violation (RoWV) and yield. The impact of driving-related parameters and external traffic on non-compliant behaviour is analysed using the Tweedie generalised linear model. This paper proposes an aggregated surrogate safety measure, condPET, and a novel parameter, congValue, to identify critical conflicts and congestion due to non-compliant behaviour. Lateral evasive action is used to detect a constrained path because of nonchalant and competitive behaviours. The results indicate that only 7.50% of vehicles yielded, 6.25% of conflicts were critical (compared to 38.94% using PET alone and 11.05% using CS), and localised congestion occurred for 44.00% of the total video time. Overall, 45.34% of vehicles created a constrained path, and 26.00% committed RoW violations, causing congestion and increasing the average travel time on major roads by 2.0 and 3.5 times, respectively. Our methodology enables computer vision-based automated assessments of both road traffic safety and operational performance at uncontrolled T-intersections under non-uniform traffic conditions, providing a valuable tool for road-traffic-monitoring systems. Full article

Background/Objectives: Triple-negative breast cancer (TNBC) is an aggressive subtype in which neoadjuvant chemotherapy plays an important role in initial treatment. Pathological complete response (pCR) following neoadjuvant therapy can serve as a surrogate marker for long-term survival. Our study aimed to evaluate the value of the Pan-Immune-Inflammation Value (PIV), Neutrophil-to-Lymphocyte Ratio (NLR), and Platelet-to-Lymphocyte Ratio (PLR), measured at multiple time points, in predicting pCR. Methods: We retrospectively included 89 patients with non-metastatic TNBC treated with neoadjuvant chemotherapy with or without immunotherapy, followed by surgery. Neutrophil-to-Lymphocyte Ratio (NLR), Platelet-to-Lymphocyte Ratio (PLR), and Pan-Immune-Inflammation Value (PIV) were calculated at baseline, before the second treatment cycle, and at the pragmatic pre-transition assessment before a subsequent treatment phase, when applicable. The primary endpoint was pCR, defined as ypT0N0. Multivariable logistic regression models included age, Ki-67, clinical T stage, and tumor grade. Biomarker-extended models were compared with the clinical model on identical complete-case samples. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), likelihood ratio testing, calibration measures, Brier scores, and bootstrap internal validation with 2000 resamples. Results: pCR was achieved in 25 of 89 patients (28.1%). Baseline platelet counts were lower in patients with pCR than in those without pCR (median 246 × 10³/µL vs. 272 × 10³/µL; p = 0.021). At the pre-transition assessment, patients with pCR had lower monocyte counts (0.20 × 10³/µL vs. 0.57 × 10³/µL; p = 0.048) and lower PIVs (378.9 vs. 746.0; p = 0.011). Baseline PIV did not improve the clinical model. On the same 83-patient sample, adding baseline platelet count increased the apparent AUC from 0.751 to 0.807; however, the bootstrap 95% confidence interval (CI) for the AUC difference included zero (−0.003 to 0.128). On the same 75-patient sample, adding pre-transition PIV increased the apparent AUC from 0.735 to 0.798, with a bootstrap 95% confidence interval for the AUC difference of 0.005 to 0.142. The optimism-corrected AUC for the overall clinical model was 0.716, indicating lower internally validated performance than suggested by the apparent AUC. A smaller absolute increase in PIV from baseline to the pre-transition assessment was associated with pCR, but this finding was definition-dependent and remained exploratory. Conclusions: Standalone baseline markers have limited utility in predicting pCR in non-metastatic TNBC. Pre-transition PIV showed the most consistent incremental association with pCR beyond conventional clinical variables, whereas the added value of baseline platelet count was uncertain and baseline PIV provided no incremental benefit. Because of the retrospective design, limited sample size, treatment heterogeneity, and evidence of model optimism, these findings should be regarded as hypothesis-generating and require external validation. Full article