Search Results (57,178)

Background/Objectives: Pancreatic ductal adenocarcinoma (PDAC) has exceptionally high mortality, largely because of late diagnosis, while established blood biomarkers such as CA19-9 have limited sensitivity for early detection. We evaluated whether modelling serum proteomic measurements as supervised pairwise statistical graphs using synolitic graph neural networks (SGNNs) could provide a useful framework for pre-diagnostic PDAC classification. Methods: We analysed 217 pre-diagnostic serum samples from UKCTOCS, comprising 100 PDAC samples from 75 women and 117 control samples, profiled for 97 proteins. Samples were divided into a development set collected 0–1 year before diagnosis (n = 110) and a temporal holdout set collected 1–2 years before diagnosis (n = 107); 25 PDAC participants contributed samples to both windows. The final SGNN used fold-internal mutual-information selection of 30 proteins, pairwise RBF-SVM-derived edge information, minimum-connected sparsification, and a 10-member GATv2 ensemble per cross-validation fold. Conventional machine-learning comparators were tuned using development data only. Results: The SGNN achieved a mean 5-fold development ROC-AUC of 75.33 ± 10.43%, lower than the tuned Random Forest (81.67%); the other tuned comparators achieved 77.67% (XGBoost), 77.17% (logistic regression), and 73.17% (SVM). Averaging predictions across the five-fold-specific SGNN ensembles yielded a temporal-holdout ROC-AUC of 64.35% (95% CI 52.6–74.6%). In the participant-independent subset of 25 PDAC cases not represented in development and 57 controls, ROC-AUC was 66.04% (95% CI 52.56–78.88%). Conclusions: SGNNs provide a feasible graph-based representation of pre-diagnostic proteomic data but did not outperform optimised conventional machine-learning methods within development cross-validation. The temporal and participant-independent performance estimates warrant further investigation in larger fully independent cohorts. A graph-specific predictive advantage was not directly tested—no such advantage was demonstrated in this dataset—and the clinical utility has not been established. Full article

We examine how distinct equity sectors on the Warsaw Stock Exchange respond to inflationary shocks, extending the Fisher hedge hypothesis beyond aggregate market indices. Drawing on monthly data from October 2009 to September 2025, our analysis captures the effects of headline inflation, long-term inflation expectations, and cyclical inflation components on nominal and real returns. To address volatility clustering and asymmetric market conditions, we employ GARCH specifications alongside quantile regressions, while also controlling for Economic Policy Uncertainty (EPU) and Geopolitical Risk (GPR). The empirical evidence points to pronounced cross-industry heterogeneity. Banking stocks stand out by offering the strongest inflation protection, with the chemicals sector providing more modest benefits. Under traditional mean-based models, real returns across most industries appear disconnected from price level changes, which aligns with the Generalized Fisher hypothesis. Yet, quantile estimates reveal a highly state-dependent reality: equities successfully protect purchasing power during market expansions, but this capacity deteriorates under market stress. Interestingly, although elevated EPU and GPR directly depress real returns, the core inflation-hedging features of resilient sectors survive these macro-institutional shocks. Ultimately, inflation protection depends heavily on both the specific industry and the prevailing market regime. Full article

Background. Body mass index (BMI) remains the most widely used tool for obesity classification and cardiometabolic risk stratification. However, increasing evidence suggests that normal BMI does not necessarily reflect metabolic health, particularly in individuals with hidden visceral adiposity and residual atherogenic dyslipidemia. The present study aimed to evaluate the prevalence of residual atherogenic dyslipidemia among normal-weight adults and to analyze its association with visceral adiposity, metabolic dysfunction, and lifestyle-related factors. Methods. A cross-sectional observational study was conducted in 166,008 Spanish workers aged 18–69 years with normal BMI (18.5–24.9 kg/m²) who underwent standardized occupational health examinations between 2022 and 2024. Anthropometric, biochemical, vascular, and lifestyle-related variables were evaluated. Visceral adiposity was estimated using waist-to-height ratio (WtHR) and the Metabolic Score for Visceral Fat (METS-VF). Residual atherogenic dyslipidemia was primarily assessed using the triglyceride-to-HDL cholesterol ratio (TG/HDL-c). Multivariable logistic regression analyses were performed to identify independent associations with elevated TG/HDL-c. Results. Despite normal BMI values, 8.0% of participants presented with an elevated TG/HDL-c ratio, 11.8% had central adiposity, 11.4% showed low HDL cholesterol, and 17.9% exhibited elevated blood pressure. Individuals with an elevated TG/HDL-c ratio demonstrated significantly higher waist circumference, WtHR, METS-VF, triglycerides, non-HDL cholesterol, fasting glucose, blood pressure, and TyG index values (all p < 0.001). A progressive worsening of lipid and metabolic parameters was observed across increasing METS-VF quartiles. Elevated TG/HDL-c ratio prevalence increased from 2.7% in the lowest METS-VF quartile to 15.8% in the highest quartile. Men showed a substantially higher prevalence of residual atherogenic dyslipidemia than women (12.2% vs. 3.9%, p < 0.001). Physical inactivity, smoking, poor Mediterranean diet adherence, and regular alcohol consumption were strongly associated with adverse lipid-metabolic profiles. In multivariable analyses using separate adiposity measures, BMI and METS-VF remained independently associated with elevated TG/HDL-c, whereas WtHR ≥ 0.50 was not independently associated after adjustment. Male sex, physical inactivity, smoking, impaired fasting glucose, elevated blood pressure, and low Mediterranean diet adherence showed consistent associations across models. Conclusions. Residual atherogenic dyslipidemia is highly prevalent among adults with normal BMI and is strongly associated with visceral adiposity, metabolic dysfunction, and unfavorable lifestyle-related behaviors. Conventional BMI classification alone substantially underestimates hidden cardiometabolic risk. These findings highlight the metabolic heterogeneity present among apparently healthy normal-weight individuals and the limitations of BMI as a sole marker of cardiometabolic health. Full article

Low Earth orbit (LEO) maneuver analysis remains challenging because maneuver events are sparse and short in duration, while accurate analysis requires simultaneous maneuver detection, onset-time localization, and continuous three-dimensional velocity-increment (Δv) estimation from long multivariate tracking sequences. To address these challenges, an improved multi-task iSpikformer framework is proposed for the joint detection, onset-time localization, and three-dimensional velocity-increment estimation of low Earth orbit maneuvers from multivariate tracking time series. The model combines a local convolutional encoder, stacked spiking Transformer blocks, and task-specific output heads within a unified multi-task framework. Event-centered supervision, hard-negative optimization, and robust Δv regression are employed to improve sparse maneuver detection and continuous parameter estimation. For full-scene inference, predictions from overlapping windows are fused and converted into discrete maneuver events through boundary-aware event extraction and validation-based calibration. Unlike approaches that separately handle maneuver detection and parameter estimation or rely on conventional dense sequence modeling, the proposed framework jointly learns maneuver occurrence, onset location, and three-dimensional Δv from a shared temporal representation. Evaluation on a synthetic dataset of Starlink-like LEO trajectories showed that the proposed method achieved an F1-score of 0.9018, an onset-time MAE of 14.54 s, a component-wise Δv MAE of 0.0307 m/s, and a vector RMSE of 0.1302 m/s. Compared with representative baseline methods, the proposed method showed improved maneuver-detection performance and more accurate Δv estimation while maintaining comparable onset-localization accuracy. The inference-stride analysis further showed that inference time could be substantially reduced over a moderate stride range with limited changes in detection and estimation performance, whereas an excessively large stride reduced detection sensitivity. The proposed framework therefore provides a unified data-driven approach to event-level LEO maneuver analysis and demonstrates the applicability of spiking temporal modeling to joint maneuver detection and continuous orbital-parameter estimation. Full article

Background/Objectives: Single-cell transcriptomics provides detailed maps of the tumour microenvironment, but translating these maps to large clinical cohorts requires methods that can recover biologically meaningful cell states from routinely available bulk transcriptomes and establish whether these states provide information beyond standard clinical factors. The objective of this study was to develop ReCAST, a donor-aware method that estimates tumour microenvironment cell states from bulk tumour transcriptomes, and to test whether these estimates improved survival prediction in esophageal cancer beyond routine clinical variables. Methods: ReCAST constructed outcome-independent cell-state anchors from single-cell transcriptomes by first aggregating expression within donors and then across donors, followed by robust non-negative projection of bulk tumour profiles. The framework was evaluated using synthetic mixtures, 720 genuine-cell pseudo-bulk mixtures from 60 held-out donors, cross-study mixtures from an independent single-cell cohort, an independent adenocarcinoma single-cell cohort, and independently measured tumour compositions in 182 TCGA esophageal cancers. Biological validity was examined against Human Protein Atlas cell-type annotations. ReCAST-derived tumour microenvironment features were subsequently tested for incremental prognostic value beyond age, sex, histology, and stage using repeated cross-validation, and the clinical prediction models were evaluated by locked external validation in an independent cohort of 60 patients with squamous-cell carcinoma. Results: ReCAST recovered 13 biologically coherent cell states whose markers showed strong enrichment for corresponding independent cell-type annotations (fold enrichment 5.1–29.8; FDR < 10⁻¹⁵). Cell-state profiles were reproducible across held-out donors and independent studies and donor-aware reference construction improved cross-study transfer. ReCAST outperformed BisqueRNA and unbalanced optimal transport on held-out pseudo-bulk mixtures and, because it worked on within-sample ranks, extended cell-state profiling to data without raw counts, such as variance-stabilised, microarray or summarised reference data; where raw counts were available, count-based methods (MuSiC, DWLS, BayesPrism and a CIBERSORT-type ν-support-vector regression) achieved lower errors and were the preferred complement. In clinical tumours, inferred immune composition tracked independently measured lymphocyte infiltration and leukocyte fraction. Routine clinical variables remained the strongest predictors of survival and were transferred to the independent squamous-cell carcinoma cohort (Uno C-index 0.61); adding ReCAST-derived features produced no incremental improvement (ΔUno C-index −0.033, 95% CI −0.089 to 0.023), a result that was consistent within each histology and with an adenocarcinoma-matched reference. Conclusions: ReCAST provides reproducible cell-state profiles across donors, studies and histologies and extends tumour microenvironment profiling to expression data without raw counts. In esophageal cancer, routine clinical variables already capture the prognostic information carried by cell-state composition, which positions cell-state profiling as a tool for biological characterisation rather than risk prediction. Establishing measurement validity and clinical utility separately provides a robust route for evaluating molecular biomarkers. Full article

Because of the increasing negative effects of drought on sectors such as water resources, the economy and agriculture, there is a strong need to study drought and its projections. Given the importance of studying different scenarios regarding drought and climate change, this study presents an integrated modelling framework to generate future drought projections for the Western Black Sea Basin, Türkiye, under four Shared Socioeconomic Pathway (SSP) scenarios derived from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Monthly precipitation data from seven General Circulation Models—ACCESS-CM2, CanESM5, CNRM-CM6-1, IPSL-CM6A-LR, MIROC6, MPI-ESM1-2-LR and MRI-ESM2-0—were statistically downscaled to 31 meteorological observation stations using a Multivariate Adaptive Regression Splines (MARS) approach, trained over the 1979–2014 period. Precipitation records from all 32 available meteorological stations were quality-controlled, and the 31 stations with continuous temperature records were retained for the full analysis, so that both indices could be computed at the same locations. Systematic biases in the downscaled outputs were subsequently corrected using the Quantile Delta Mapping (QDM) method, an essential step that simultaneously reduces distributional errors and preserves the future climate change signal. Both the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI) were computed at 3-, 6- and 12-month scales, with distribution parameters estimated once over the 1979–2014 window of the bias-corrected model series and held fixed for 2015–2100 so that baseline and future statistics remain comparable. The two indices yield markedly different projections. Although the number of drought events falls, the share of months below the moderate-drought threshold at the 12-month scale rises from 16.5% in the baseline to 47.0% by 2061–2100 under SSP5-8.5 for the SPI and to 88.4% for the SPEI. All 217 station–GCM combinations show the SPEI as the drier index in every scenario and period. The reason is the temperature term: under SSP5-8.5, basin-mean precipitation changes by −20 mm yr⁻¹ by the late century while potential evapotranspiration rises by +192 mm yr⁻¹, so that about 91% of the change in the climatic water balance is attributable to evaporative demand rather than to rainfall. What changes is therefore not how often drought occurs but how long it lasts and how much deficit it accumulates. Projections under SSP1-2.6 indicate substantially less severe drought conditions than those under the higher-emission pathways at every station, although drought exposure still increases relative to the baseline. Full article

This study examined associations between self-perceived caffeine addiction severity, mental health, and lifestyle choices in a clinic-based sample of young adults, with consideration of their broader public health implications. A cross-sectional survey was conducted through an internal medicine clinic in Weimar, California, among 1438 individuals aged 18–24 years from diverse cultural backgrounds across 37 countries, 77% of whom were from the United States. Self-perceived caffeine addiction was categorized as none, moderate, or severe. Chi-square analyses showed significant differences across self-perceived caffeine addiction severity groups in all examined lifestyle variables, including dietary factors, physical activity, sleep patterns, sleep and meal regularity, sunlight exposure, entertainment-related screen use, conscience-related behaviors, and substance-use behaviors. One-way analysis of variance (ANOVA) indicated that greater caffeine addiction severity was associated with higher depression and anxiety scores and lower emotional intelligence (EI) scores. Among these mental health indicators, only EI showed a significant association with severe self-perceived caffeine addiction after adjustment. In the adjusted multinomial logistic regression model, sleeping 7–9 h per night; obtaining 1–2 h of sleep before midnight; consuming 3–4 (but not ≥5; a category with only one severe case) servings of fruits, vegetables, and whole grains daily; not reporting tobacco addiction; and having higher EI scores were significantly associated with lower odds of severe self-perceived caffeine addiction. For moderate caffeine addiction, consuming cheese no more than twice per week, viewing entertainment media for two hours or less per week, and not reporting tobacco addiction were associated with lower odds. These findings indicate that self-perceived caffeine addiction severity is associated with poorer mental health and unhealthy lifestyle behaviors, supporting the need for a holistic approach to prevention and intervention. Full article

Soil salinization restricts crop production and complicates field management in arid farming systems. We assessed low-cost red-green-blue and near-infrared sensing across three contiguous management zones assigned to deep vertical rotary tillage depths of 40, 55, and 70 cm in Bachu County, Xinjiang, China. The dataset contained 90 paired sampling points, with 30 spatial subsamples in each zone. Because each depth was represented by one zone, the design had no independent plot-level biological replication and zone contrasts were interpreted descriptively. We compared support vector regression and extreme gradient boosting after feature screening by the maximal information coefficient, Pearson correlation, or grey relational analysis. Hyperparameters were tuned by conventional random search, the Sparrow Search Algorithm, or the Grey Wolf Optimizer within repeated nested cross-validation. Mean soil total salt contents were 12.12, 14.63, and 7.24 g kg⁻¹ in the 40, 55, and 70 cm zones, respectively. Welch’s analysis of variance and Games–Howell comparisons indicated that the 70 cm zone differed from the other two zones, whereas the 40 and 55 cm zones did not differ. These point-level tests describe spatial contrasts and do not establish a causal tillage-depth effect. Using field-derived RGB-NIR predictors, the best screened and conventionally tuned extreme-gradient-boosting model achieved a pooled out-of-fold coefficient of determination of 0.694 and a root mean square error of 3.832 g kg⁻¹. Its paired outer-fold error was lower than that of the all-feature default baseline after multiplicity correction, while the two swarm optimizers did not significantly improve on conventional tuning under the matched evaluation budget. The results support field-derived RGB-NIR sensing as a practical aid for within-field monitoring of soil total salt content, while non-spatial cross-validation may overestimate performance at spatially independent locations and external validation remains necessary. Full article

Background/Objectives: Fetal aortic isthmus reference data have characterized its size as a function of gestational age; whether aortic and pulmonary valve diameters are associated with isthmic caliber in addition to gestational age has not been systematically examined. We examined the correlates of aortic isthmus diameter in fetuses considered structurally normal on second-trimester assessment. Methods: This retrospective single-center cohort study included 472 fetuses considered structurally normal on second-trimester assessment at 19 + 0–24 + 6 weeks. Aortic isthmus (AoI), aortic valve (AV), pulmonary valve (PV) and thymic transverse diameters were measured by a single operator. Three-step hierarchical linear regression modeled the AoI diameter (Model 1, gestational age; Model 2, +AV and PV; Model 3, +thymus), with variance inflation factors, regression diagnostics and internal validation conducted via repeated five-fold cross-validation (100 repeats). For sensitivity analyses, the gestational-age equivalents recorded in the clinical reports for femur length, biparietal diameter and head circumference were converted into total days and analyzed as continuous variables in place of the clinical gestational age. Results: AoI diameter was correlated with gestational age (r = 0.512), AV (r = 0.507), PV (r = 0.516) and thymic diameter (r = 0.397) (all p < 0.001). Gestational age alone accounted for 26.2% of the variance (R² = 0.262). Adding AV and PV increased the explained variance (ΔR² = 0.060; total R² = 0.322; p < 0.001); both AV (β = 0.186; p = 0.005) and PV (β = 0.187; p = 0.006) were independently associated with isthmic caliber, and the highest VIF in Model 2 was 3.22. Adding thymic diameter produced no meaningful increment (ΔR² = 0.0003; p = 0.636). Across the four size-index specifications, the valve-term increment ranged from 0.037 to 0.060. Internal validation via repeated five-fold cross-validation gave a mean cross-validated R² of 0.299 (SD of repeat-specific means, 0.011). Conclusions: In fetuses considered structurally normal on second-trimester assessment, aortic isthmus diameter was associated with gestational age, and, in addition, with aortic and pulmonary valve diameters, which accounted for a modest further share of the variance. No independent association with thymic diameter was demonstrated. Full article

Background/Objective: Lung cancer remains a leading cause of cancer-related mortality, and access to low-dose CT screening is limited in resource-constrained settings. This study aimed to develop and evaluate an interpretable machine learning model for preliminary, symptom-based lung cancer risk discrimination using clinical indicators and lifestyle factors, without requiring imaging. Methods: A cross-sectional dataset of 561 participants (38 with a prior lung cancer diagnosis), following exclusion of a minors-eligible age bracket for which independent age verification was not possible, was analyzed. Five machine learning algorithms (Logistic Regression, Random Forest, XGBoost, LightGBM, CatBoost) were compared using a leakage-free pipeline, with ADASYN class balancing and hyperparameter tuning performed exclusively within cross-validation folds; feature selection was verified to be stable across folds (98% average overlap with the final feature set). The final model was selected solely on the basis of repeated cross-validated performance, without reference to the test set. Model interpretability was assessed using LIME, and a four-tier risk stratification system was developed and calibrated. Results: Random Forest achieved the best cross-validated F1-score (0.850 ± 0.092) among the five candidates and, given a clinically motivated preference for higher sensitivity at comparable F1, was selected as the final model. On the independent test set (n = 113), it achieved Accuracy = 0.956, Recall = 0.875, F1-score = 0.737, and AUC-ROC = 0.912. Sensitivity analysis indicated moderate reliance on a potential diagnostic proxy feature, and balancing-method comparisons (ADASYN, random oversampling, class weighting) confirmed that conclusions were not contingent on the specific resampling technique. Conclusions: The proposed model demonstrates solid discriminative performance and interpretability; however, given the cross-sectional design and limited number of positive cases, results should be interpreted as preliminary risk discrimination rather than validated early-detection prediction, warranting further prospective validation. Full article

Background: Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are widely used in obesity and type 2 diabetes mellitus (T2DM), but comparative real-world evidence from the Middle East remains limited. We compared the effectiveness, documented adverse events, adherence and persistence of once-weekly semaglutide and once-daily liraglutide in Omani adults. Methods: Retrospective cohort study of adults treated with semaglutide or liraglutide at Sultan Qaboos University Hospital between January 2024 and December 2025. The primary outcome was change in body mass index (BMI) at 3 and 6 months; change in body weight, percentage total body weight loss and the proportions achieving at least 5% and at least 10% weight loss. Because treatment indication and exposure differed markedly between groups, the primary comparison used propensity-score overlap weighting with additional covariate adjustment, implemented in a design-based general linear model. A mixed model for repeated measures supported the primary analysis and handled missing visits under a missing-at-random assumption, with a delta-adjusted pattern-mixture analysis used to test departures from that assumption. Persistence was analysed with Kaplan–Meier estimation and Cox regression adjusted for baseline covariates only. Secondary outcomes were corrected for multiplicity using the Benjamini–Hochberg procedure. Analyses are labelled throughout as primary, prespecified secondary, or post hoc exploratory. No patient received semaglutide 2.4 mg weekly, the dose licenced for chronic weight management, due to unavailability during the study period. Results: Of 287 treatment records supplied, 257 unique patients were analysed (liraglutide n = 135; semaglutide n = 122). Semaglutide was associated with a greater reduction in BMI than liraglutide at 3 months (adjusted difference −0.60 kg/m², 95% CI −1.12 to −0.07; p = 0.026) and at 6 months (−1.42 kg/m², 95% CI −2.60 to −0.24; p = 0.019). In clinical units, the adjusted difference in body weight was −1.70 kg at 3 months (95% CI −3.10 to −0.29; p = 0.018) and −3.66 kg at 6 months (95% CI −6.72 to −0.60; p = 0.020). At 6 months, 56.5% of semaglutide-treated and 24.7% of liraglutide-treated patients had lost at least 5% of body weight (p < 0.001), and 30.4% versus 4.9% had lost at least 10% (p < 0.001). The mixed model gave the same conclusion (group-by-time interaction p < 0.001). No statistically detectable difference in glycated hemoglobin HbA1c change was observed at either time point (p = 0.132 and p = 0.158; interaction p = 0.360). Of eight secondary cardiometabolic outcomes, only total cholesterol at 3 months remained significant after correction for multiplicity (adjusted p = 0.024). Documented adverse events were infrequent and did not differ between groups (12.6% versus 8.2%, p = 0.251). Hypoglycaemia occurred in seven liraglutide-treated and one semaglutide-treated patient (p = 0.069 by exact test) and was not significant after weighting and adjustment for background insulin or sulfonylurea use (p = 0.345). Crude discontinuation was 48.9% with liraglutide versus 9.0% with semaglutide, but median exposure was 17 versus 6 months. After accounting for time at risk, no difference in persistence was statistically apparent (log-rank p = 0.328; adjusted hazard ratio for semaglutide 1.60, 95% CI 0.61 to 4.19). Documented adherence was higher with semaglutide (95.9% versus 83.7%, p = 0.001). Conclusions: In this cohort, semaglutide was associated with a modestly greater reduction in BMI and body weight than liraglutide, and with a higher proportion of patients achieving clinically meaningful weight loss. Glycaemic and most cardiometabolic comparisons were inconclusive once confounding, multiplicity and limited precision were accounted for, and the large apparent difference in discontinuation was no longer statistically apparent after allowing for unequal time at risk. These findings are associative. They describe the drugs as prescribed in one health-care setting, at doses that are not pharmacologically equivalent, and should not be read as evidence of comparative causal superiority. Full article

Preliminary selection of bracing systems in high-rise steel buildings requires comparing many configurations, each demanding computationally intensive minimum-weight design. This study uses machine learning to estimate the best-observed feasible weight directly from building-level configuration parameters. A full factorial dataset of 45 configurations covered three storey numbers (17, 24, and 30), three bracing types (V, X, and K), and five brace-placement layouts. Each configuration was optimized in a coupled SAP2000–MATLAB procedure with Particle Swarm, Grey Wolf, Honey Badger, and Aquila optimization (180 runs in total); no single algorithm was best for every configuration. Mean weight increased from 2537 kN at 17 storeys to 7103 kN at 30 storeys. X-bracing gave the highest mean weight at all heights and K-bracing the lowest at 17 and 24 storeys, whereas V- and K-bracing differed by only 2.4% at 30 storeys. Bracing type changed the mean weight by up to about 30%, and brace placement by about 10–15%. Among six regression models, Ridge achieved the highest mean five-fold cross-validated R² (0.914; 0.904 over 1000 repeated partitions), while Random Forest gave the lowest RMSE (511.7 kN); in leave-one-out prediction, Ridge ordered 90.3% of configuration pairs correctly. The resulting Ridge equation offers an interpretable screening tool within the investigated design domain. Full article

Stepwise fixed-effect selection in complex linear mixed models is expensive because each candidate can require the re-estimation of variance components. Permanent pruning lowers this cost but can discard predictors that become relevant after other terms enter. We propose adaptive re-screened stepwise selection (ARSS), which uses a fixed-covariance generalized least-squares score to rank all remaining candidates, fully refits only a short list, and re-screens after every accepted update. We further introduce pair-screened joint rescue (ARSS-JPS) for correlated candidates that are jointly informative, although neither enters alone. A worked suppression example explains the motivation for this method. With the full covariance fixed, the scalar screening score equals the one-degree-of-freedom generalized least-squares likelihood-ratio improvement; this identity does not establish exact forward path equivalence after covariance re-estimation. In 340 known-covariance simulations, ARSS-JPS achieved a masked-scenario mean true-positive rate of 0.999 using 35.5 logical full-model fit requests versus 414.4 for exact forward selection. For an 80-dataset full-refit benchmark, it increased the masked-scenario true-positive rate from 0.768 to 0.925 and exact recovery from 0.429 to 0.600 while reducing mean full-model fit requests from 69.0 to 37.4. In the corrected reconstruction of 95,890 Polish commuting flows using orthogonal–triangular (QR) factorization, ARSS reproduced exact forward selection through the fixed-effect Stages 1–5, with 2587 rather than 62,651 logical candidate evaluations. The archived random-effect extension retained 87 fixed terms and 12 random-effect variance components. Full article

This study examines the relationship between digitalisation, digital marketing innovation, and sustainable consumption using macro-level longitudinal data for 2012–2025 (T = 14 annual observations). The core empirical analysis is correlational and descriptive in character: ordinary least squares (OLS) regression is used to estimate associations between digital marketing indicators and sustainable-market outcomes. The Theory of Planned Behaviour (TPB) and the Technology Acceptance Model (TAM) are used as interpretive theoretical lenses for discussing possible behavioural mechanisms; mediation and moderation mechanisms motivated by these theories are discussed as exploratory propositions but are not formally tested as statistical models. The study conducts an exploratory assessment of these interrelationships at the macro level. The findings show that digitalisation, reflected in the growth of digital advertising expenditure and internet penetration, is associated with the expansion of sustainable-product markets, particularly in the organic and eco-friendly sectors. A 1% increase in global digital advertising expenditure is associated with an approximately 0.607% increase in the organic market in a log-log specification. However, this level association attenuates substantially once common temporal trends are removed via first-differencing; this contrast between level and first-differenced results is treated as a central empirical finding of the study. Complementary survey evidence indicates that Generation Z and Millennials account for an estimated 75% of sustainability-related purchasing decisions. The findings are also consistent with the persistence of a structural “Green Gap” between pro-environmental attitudes and actual purchasing behaviour. Descriptive and survey-based evidence further suggests that this gap co-varies with price sensitivity, informational frictions, habitual routines, and consumer trust. As descriptive contextual evidence only, the difference between survey-based attitudinal prevalence and aggregate sustainable market share (a group-level comparison, not an individual-level measurement of the intention–behaviour gap) is approximately 53–58 percentage points. The evidence is also consistent with the possibility that digital marketing’s contribution to sustainable consumption may be constrained by the risk of greenwashing, which can erode consumer trust, and by supply-side and economic barriers that lie outside the scope of digital marketing interventions. Overall, the study provides descriptive and correlational evidence that digitalisation and digital marketing innovation may provide an enabling context for responsible consumption, while their contribution likely depends on complementary supply-side, regulatory, and trust-related conditions. Full article

Electronic noses provide a pattern-based approach for investigating volatile profiles associated with colorectal cancer; however, the combined analysis of participant-matched sensor profiles from different biological matrices remains largely unexplored. This study evaluated breath, urine, and axillary sweat profiles acquired from the same participants using a single microelectromechanical system (MEMS)-based electronic nose system for colorectal cancer (CRC) classification. Matched records from 60 participants were analyzed, including 27 with CRC and 33 controls (CO). Signals from each biological matrix were processed independently using fractional-difference feature extraction, min–max scaling, Orthogonal Signal Correction (OSC), and Principal Component Analysis (PCA). Because OSC is a supervised preprocessing method that incorporates class information, the PCA score plots were used only for descriptive visualization of the transformed feature space, whereas predictive performance was assessed exclusively within the cross-validation framework. The selected PCA scores were subsequently integrated into a six-feature participant-level representation. Five supervised classifiers were evaluated using stratified five-fold cross-validation, with all data-dependent transformations fitted exclusively on the corresponding training sets. In this exploratory setting, Logistic Regression showed the highest observed mean accuracy of 98.3 ± 3.3% among the evaluated models, though this represents an internal estimate rather than a claim of statistical superiority. Aggregated out-of-fold predictions yielded 98.3% accuracy, 100.0% sensitivity, 97.0% specificity, a 98.2% F1-score, and an AUC of 1.00 (95% CI: 0.99–1.00). Our comparative analysis revealed that axillary sweat alone demonstrated a high discriminative power and that the pairwise combination of urine and sweat achieved 100% accuracy. While the complete three-matrix integration maintained high robustness, it did not provide incremental value over sweat alone. These findings highlight the remarkable potential of the axillary volatilome for CRC classification and suggest that selectively combining highly discriminative matrices may be more effective than integrating all available biological sources. Full article