Search Results (3,206)

Plasma Oncostatin M (OSM) has been proposed as a biomarker of disease activity in ulcerative colitis (UC). We evaluated OSM’s relationship with clinical, endoscopic, and histological activity across two follow-up visits, compared with fecal calprotectin (FC). One hundred UC patients were assessed at Visit 1 (baseline) and Visit 2 (12 months) for clinical (partial Mayo score), endoscopic (Mayo Endoscopic Score), and histological (Nancy Histological Index) activity, alongside plasma OSM, FC, CRP, and fibrinogen; 30 healthy controls were included for comparison. Given the marked right-skewness of OSM and FC, associations were assessed using both Pearson and Spearman correlations, and logistic regression effect sizes are reported as odds ratios per doubling of biomarker concentration (log2 scale). OSM correlated significantly with all activity measures at both visits, most strongly with histological activity (r = 0.521–0.602), but weakly or not with CRP/fibrinogen. OSM was significantly elevated versus controls at both visits (p < 0.001) and declined significantly with treatment response (Δ = −89.06 vs. +0.50 in controls, p = 0.006). FC was the dominant predictor of clinical (AUC = 0.788) and endoscopic (AUC = 0.926) remission at Visit 2, with OSM contributing little independent value; however, for histological remission, both baseline OSM and ΔOSM were independent, complementary correlates (OR = 8.43 and 9.64 per doubling, respectively; p < 0.001 for both), yielding the strongest model overall (AUC = 0.949). OSM’s discriminative accuracy improved markedly from Visit 1 to Visit 2 for endoscopic (AUC 0.695 → 0.897) and histological (AUC 0.738 → 0.927) activity. Across two follow-up assessments, OSM consistently reflected UC activity, with its strongest and most independent signal for histological inflammation, both as a concurrent correlate and—more modestly—as a genuine prospective marker independent of baseline treatment exposure. OSM’s incremental value over FC was most apparent for histological status and for stringent (MES 0) endoscopic remission, but not for conventional clinical or endoscopic remission thresholds. Full article

Tea bud detection is a key technology for enabling automated and intelligent tea harvesting. However, accurate detection remains challenging because tea buds are generally small, densely distributed, partially occluded, and visually similar to the surrounding background. In addition, resource-constrained devices impose strict requirements on model size and computational complexity. To address these challenges, a lightweight tea bud detection model, termed CLBD-YOLO, is proposed for resource-constrained tea bud detection applications. Based on YOLOv11n, a C3K2_GD module is designed for the backbone and neck to combine ghost-style lightweight feature generation with input-conditioned dynamic feature modeling. A three-level weighted BiFPN is employed for bidirectional multi-scale feature fusion, while LAE, DAT, and MPDIoU are incorporated for downsampling feature extraction, adaptive spatial representation, and bounding box regression, respectively. Experiments on the predefined TeaBud test subset containing 625 images show that CLBD-YOLO achieves 74.31% precision, 77.07% recall, 85.67% mAP${}_{50}$, and 62.69% mAP${}_{50--95}$. Compared with YOLOv11n, the parameter count is reduced from 2.582M to 1.426M and the computational cost from 6.3G to 5.6G FLOPs, corresponding to reductions of approximately 44.8% and 11.1%, respectively. For the pre-declared seed-0 run, mAP${}_{50}$ and mAP${}_{50--95}$ decrease by 0.22 and 1.40 percentage points, respectively; when averaged over the three matched random seeds, the corresponding decreases are 0.77 and 1.89 percentage points. Because near-duplicate screening revealed cross-subset augmented siblings in the predefined TeaBud split, these results are interpreted as within-distribution reference values rather than as strictly leakage-free generalization estimates. Direct evaluation on the 591-image external Xinyang Maojian tea bud validation subset using the fixed TeaBud-trained weights without additional fine-tuning or threshold tuning yields 81.2% mAP${}_{50}$ and 62.5% mAP${}_{50--95}$, respectively, for CLBD-YOLO, compared with 80.4% and 61.2% for YOLOv11n. These results show that the compact model remains competitive under this external data distribution. Overall, CLBD-YOLO substantially reduces model complexity while maintaining competitive tea bud detection performance, demonstrating a favorable accuracy–complexity trade-off for resource-constrained applications. Full article

Investigation of the molecular mechanisms of postmortem processes to identify postmortem interval (PMI) markers is highly relevant in forensic medicine, given the insufficient accuracy and substantial limitations of current routine PMI estimation methods. This exploratory study aimed to describe changes in NADH fluorescence decay parameters in rat skeletal muscle during the early postmortem period and to assess their potential for PMI estimation. Using time-resolved fluorescence spectroscopy, the short- and long-lifetime components (τ₁ and τ₂), the amplitude-weighted mean fluorescence lifetime (τ_m), the relative contributions of the short- and long-lifetime components (α₁ and α₂, respectively) and fluorescence intensity were measured in 10 Wistar rats before death, immediately after death, every 30 min during the first 6 h, and at 24 h postmortem. Immediately after death, τ₁ decreased compared with the antemortem value (p = 0.02). τ₁, τ_m, α₁, and α₂ were dependent on the postmortem interval. For PMI prediction within the first 6 h after death, a partial least squares regression model combining α₁, τ_m, and the difference between rectal and room temperatures was constructed. The model demonstrated promising predictive performance, with a root mean square error (RMSE) of 0.89 h (95% CI: 0.63–0.99), a mean absolute error (MAE) of 0.68 h (95% CI: 0.49–0.82) and an R² of 0.762 (95% CI: 0.689–0.886); the prediction error did not exceed 1 h in 73.2% of cases. These data are preliminary and require confirmation in studies with larger samples under different ambient temperature conditions and subsequent validation on human cadaveric material. Full article

Background/Objectives: Breast cancer is the most common malignancy among women worldwide and poses a significant burden on healthcare systems. Despite evidence that treatment type is associated with hospital length of stay (LOS), current Germany-wide data on this topic are lacking. The present study aims to identify clinical and procedural factors associated with prolonged hospitalisation in women hospitalised for breast cancer in Germany. Methods: This multicentre cross-sectional study analysed anonymised inpatient data from 38 German hospitals (IQVIA database), including 16,062 female breast cancer hospitalisation episodes (inpatient cases; patient-level linkage was not possible due to anonymisation, so these may not represent 16,062 distinct women) treated between January 2019 and December 2024. The primary outcome was LOS; prolonged hospitalisation was defined as LOS > 7 days (75th percentile). Comorbidities were assessed using individual comorbidity categories defined according to the Elixhauser classification (retained at ≥1% prevalence). The primary multivariable analysis was restricted to characteristics documented independently of the in-hospital course (age, tumour site, metastatic status, chronic comorbidities). In-hospital complications, procedures, and diagnoses of ambiguous timing were analysed separately as secondary, descriptive associations. LOS was modelled using negative binomial regression (Poisson regression as a sensitivity analysis) and prolonged LOS using logistic regression, both with hospital-clustered robust (sandwich) standard errors. Results: The overall median LOS was 4 days (IQR 3–7); 19.3% of hospitalisation episodes were prolonged. In the primary analysis, longer LOS was associated with age > 70 years (RR 1.14; 95% CI 1.08–1.20), distant metastases (RR 1.76; 95% CI 1.51–2.04), lymph node metastases (RR 1.09; 95% CI 1.05–1.12), congestive heart failure (RR 1.44; 95% CI 1.31–1.59), anaemia (RR 1.41; 95% CI 1.14–1.75) and depression (RR 1.14; 95% CI 1.03–1.26). In the secondary descriptive analysis, in-hospital complications showed the strongest associations with longer LOS, particularly postoperative infection (RR 2.28), wound disruption (RR 1.92) and pneumonia (RR 1.45). Breast-conserving surgery and partial breast resection were associated with shorter LOS, while mastectomy, blood transfusions, and complex intensive care were associated with prolonged stays; the small inpatient radiotherapy subgroup (1.9%) most likely reflects a highly selected, predominantly palliative population rather than an association attributable to radiotherapy itself. Conclusions: Prolonged hospitalisation was associated with older age, documented lymph node and distant metastases, and chronic somatic comorbidities including congestive heart failure, anaemia, chronic kidney disease and depression, and, in secondary descriptive analyses, with perioperative complications and intensive procedures. The strongest correlates of prolonged stay—in-hospital complications and intensive procedures—are documented during rather than before the hospitalisation and therefore cannot themselves support early risk identification; however, the baseline characteristics documented independently of the in-hospital course (age, metastatic disease, chronic comorbidity burden) may usefully inform structured perioperative management and discharge planning in Germany. Full article

Background/Objectives: Vitamin D deficiency may increase the risk of IBD and be associated with disease activity. The aim of the study was to assess vitamin D levels, their association with disease activity in patients with IBD. Methods: A total of 231 subjects took part in the cross-sectional comparative study, including 129 patients with IBD and 102 healthy individuals Disease Activity Index and the Montreal classification were used to assess disease activity in patients with CD. For patients with UC, the Partial Mayo Score and the Montreal classification were applied. To determine total 25(OH)D chemiluminescent immunoassay (CLIA) technology was used. Results: The concentration of 25(OH)D was significantly lower in the group of patients with IBD compared to the control group (25.5 ng/mL vs. 28.8 ng/mL, p = 0.0027). Differences in 25(OH)D concentrations depended on IBD activity, with significantly higher vitamin D concentrations found in patients in remission compared to those with active IBD (28.9 ± 7.2 ng/mL vs. 22.3 ± 6.1 ng/mL, p = 0.0022). This association was confirmed for both UC and CD patients. A multivariate adaptive regression model using spline curves revealed a relationship between serum 25(OH)D concentrations in patients with IBD and total dietary vitamin D intake, including vitamin D supplementation, consumption of one serving of fish per week, and disease remission. Conclusions: Serum 25(OH)D levels in IBD patients may serve as an additional, useful, and non-invasive marker of disease activity. Full article

Background/Objectives: Diet quality is an important determinant of health during young adulthood; however, associations between dietary behaviours and anthropometric outcomes in university students remain insufficiently understood. This study examined the associations between dietary behaviours and anthropometric indicators in Chilean university students. Methods: A cross-sectional study was conducted among university students from southern Chile between June and November 2025. Dietary habits were assessed using a validated Chilean dietary habits questionnaire, from which Healthy Eating, Unhealthy Eating, and Overall Diet scores were derived. Anthropometric measurements included body mass index (BMI), waist circumference, and waist-to-height ratio (WHtR). Age- and sex-adjusted multivariable linear regression models were fitted. Exploratory analyses included psychometric evaluation, principal component analysis, false discovery rate (FDR) correction, and dietary behaviour network analysis. Results: A total of 374 students were included, of whom 40.3% were women. Psychometric analyses provided only partial support for the predefined two-domain structure, with stronger properties for the healthy domain. None of the predefined dietary scores were associated with BMI. The Unhealthy Eating Score was nominally positively associated with waist circumference (β = 0.285 cm per score point, 95% CI: 0.011–0.560), although this association did not remain statistically significant after FDR correction. Exploratory item-level analyses identified positive associations between sugar-sweetened beverage consumption and waist circumference and inverse associations between breakfast consumption and BMI after FDR correction. Conclusions: Predefined dietary behaviour scores showed limited associations with anthropometric outcomes. Psychometric analyses provided only partial support for the predefined scoring structure, particularly for the unhealthy dietary domain. Exploratory item-level analyses suggested that specific dietary behaviours may provide complementary information beyond that obtained by composite dietary scores, although these findings require confirmation in prospective longitudinal studies and further psychometric validation. Full article

Background: This study aims to investigate the relationship between malocclusion severity and periodontal clinical parameters and periodontal disease-related biochemical markers in adolescents. Materials and Methods: The study’s registration number is NCT07127744. It included 88 individuals aged 12–17 years. Participants were divided into 4 different groups (n = 22) with equal distribution in each group according to the Dental Aesthetic Index (DAI). Clinical (PI, GI, BOP, PPD) and biochemical (RANKL, OPG, RANKL/OPG, TOS, TAS, OSI in gingival crevicular fluid) measurements were performed. Sociodemographic data (age, gender, education, BMI) were recorded. The relationship of all parameters with DAI and with each other was statistically analyzed. Multiple comparison tests, correlation analysis, and multiple linear regression analysis were applied. Statistical significance was accepted at p < 0.05. Results: Significant differences were found between minor and very severe malocclusion in all clinical periodontal and biochemical parameters (except TAS) (p < 0.05). With increasing education level, a decrease in BOP (p = 0.041) and a significant increase in OPG (p = 0.003) were observed. A strong positive correlation was found between DAI and TOS and RANKL; and a moderate correlation was found with periodontal parameters except PI. Multivariate general linear model analysis revealed that DAI score was significantly associated with all clinical and biochemical parameters (Pillai’s Trace = 0.652, p < 0.001). The largest effect sizes were observed for TOS (partial η² = 0.455), RANKL (partial η² = 0.402), and PI (partial η² = 0.300). Conclusions: There is a significant association between increased malocclusion severity and periodontal disease-related parameters. TOS and RANKL may be sensitive biomarkers of malocclusion-related periodontal tissue destruction. Full article

Background and Objectives: Femoral arterial calcification (FAC) on routine hip radiographs may provide opportunistic vascular information within orthogeriatric assessment. We evaluated the associations of FAC and the Atherogenic Index of Plasma (AIP) with 1-year mortality after hip-fracture surgery. Materials and Methods: This retrospective single-centre cohort included 229 patients aged 65 years or older. FAC was graded from 0 to 3 on preoperative radiographs using an exploratory semiquantitative classification. AIP was calculated as log₁₀(TG/HDL-C). Logistic regression and complementary survival analyses were partially adjusted for age, sex, FAC, and AIP. Incremental discrimination was assessed using nested age + sex, age + sex + FAC, and age + sex + FAC + AIP models. Results: Thirty-eight patients (16.6%) died within 1 year. Interobserver agreement for FAC grading was almost perfect (κ = 0.87). Each one-grade increase in FAC was associated with mortality in the partially adjusted logistic (OR = 1.44, 95% CI 1.09–1.91) and Cox models (HR = 1.36, 95% CI 1.07–1.73). AIP was not significantly associated with mortality. AUCs were 0.726 for age + sex, 0.757 for age + sex + FAC, and 0.761 for age + sex + FAC + AIP; neither incremental increase was significant. Conclusions: Radiographic FAC was associated with 1-year mortality but did not demonstrate significant incremental discrimination beyond age and sex. FAC may serve as an opportunistic marker within multidisciplinary orthogeriatric assessment, not as a stand-alone prognostic tool. AIP findings remain inconclusive because sampling occurred during acute admission without standardised fasting. Prospective multicentre validation with comprehensive geriatric covariates is required. Full article

Background/Objectives: The purpose of this study was to evaluate associations between quantitative ellipsoid zone (EZ) integrity metrics derived from spectral-domain OCT (SD-OCT) and visual outcomes, including clinically meaningful visual acuity (VA) loss, in dry age-related macular degeneration (AMD). Methods: This single-center retrospective image-analysis cohort included 351 eyes from 351 patients with intermediate AMD or atrophic changes. Macular SD-OCT scans were analyzed using a certified-reader-validated machine learning-based segmentation platform to delineate the EZ, retinal pigment epithelium (RPE), and Bruch’s membrane (BM). EZ metrics included EZ-RPE thickness (i.e., outer segment thickness) and EZ attenuation, defined as partial attenuation (EZ-RPE thickness of ≤20 µm, abnormal outer segment thinning)) or total attenuation/loss (0 µm). Drusen volume was quantified from the RPE-BM compartment. VA was converted to ETDRS letters. Associations between baseline OCT metrics and baseline VA, Year 2 VA, and ≥15-letter VA loss were assessed using nonparametric testing, linear regression, and ANCOVA. Results: Greater EZ-RPE thickness and lower EZ attenuation were associated with better baseline VA (all p < 0.001), and EZ metrics remained significant in cross-sectional regression models. Among 188 eyes with Year 2 follow-up, 38 (20%) experienced ≥15-letter VA loss. These eyes had significantly lower baseline EZ-RPE thickness (i.e., abnormal outer segment thinning) and greater EZ attenuation than eyes without such loss (all p < 0.001). In ANCOVA models, EZ metrics remained associated with Year 2 VA (all p < 0.001), whereas drusen volume was not. Conclusions: Quantitative EZ integrity metrics were associated with cross-sectional and longitudinal visual outcomes in dry AMD. EZ-RPE thickness and EZ attenuation may serve as clinically relevant OCT biomarkers reflecting photoreceptor integrity and risk of meaningful VA loss. Full article

Red mud, an alkaline industrial by-product of alumina refining generated in enormous volumes worldwide, poses a persistent environmental disposal challenge; using it as a partial cement replacement offers a promising route toward more sustainable concrete, but the resulting compressive strength is governed by complex, nonlinear interactions among the mix constituents that conventional empirical and regression-based models struggle to capture accurately. To address this challenge, the present study develops and compares four machine learning and deep learning models, namely the Deep Gradient Boosting Machine (DGBM), the Differentiable Neural Decision Tree (DNDT), Long Short-Term Memory (LSTM), and the Monte Carlo Dropout Neural Network (MCDNN), for the accurate and uncertainty-aware prediction of the compressive strength of red mud concrete. A dataset of 183 data points, compiled from the literature and supplemented with experimental results, was used to capture the influence of red mud content, curing period, and other mix parameters, including cement dosage, water content, and admixture proportions. The data were pre-processed prior to model training, and predictive performance was evaluated using R², RMSE, MAE, and WMAPE, among other indicators. The results show that the deep learning models outperformed the tree-based models: LSTM achieved the highest accuracy (R² = 0.942 on the testing dataset), while MCDNN additionally provided reliable uncertainty estimates alongside comparable prediction accuracy; DNDT and DGBM were comparatively less effective. Global sensitivity analysis identified fly ash and water content as the most influential contributors to strength development. By combining rigorous data-driven modeling with sensitivity and uncertainty analysis, this study contributes to the literature a validated, uncertainty-aware deep learning framework for sustainable concrete strength prediction, and offers practical value to the construction industry by providing engineers with a reliable, data-driven tool for optimizing red mud content in concrete mix design, thereby supporting the safe and wider industrial utilization of this problematic waste stream. Full article

Soil ammonia (NH₃) emissions are a major pathway for reactive nitrogen loss, but their spatial variability is difficult to explain using equilibrium-based approaches alone. This study evaluated whether machine learning can improve the interpretation of NH₃ flux by identifying informative soil predictors. NH₃ flux, soil temperature and volumetric water content (approximately 0–6 cm) were measured at 20 sampling points on bare agricultural soil, while NH₄⁺, NO₃⁻, total nitrogen, humus content and pH were determined at three soil depths. Five regression models were evaluated using leave-one-out cross-validation. Pairwise correlations showed that no single measured variable exhibited a dominant linear relationship with NH₃ flux. Nevertheless, Random Forest achieved the best cross-validated performance (R² = 0.418), suggesting that the observed spatial variability was better represented by multivariate and potentially nonlinear relationships among soil properties than by individual linear associations. Partial Least Squares Regression provided complementary information on the multivariate soil gradients. Within the Random Forest model, NO₃⁻ concentration at 10–20 cm and near-surface soil moisture had the highest permutation importance. These findings indicate that small-scale spatial variation in NH₃ flux was associated with the combined statistical information contained in several soil properties rather than with a single dominant linear predictor. The importance of NO₃⁻ represents a dataset-specific statistical association and should not be interpreted as evidence of direct causal control or nitrification intensity. Full article

Background/Objectives: Dentists are uniquely positioned to detect child abuse and neglect (CAN), including dental neglect, yet reporting rates remain low and Bulgarian data are scarce. This study assessed the knowledge, attitudes, and self-reported intention to report CAN, including dental neglect, among clinical dental students in Plovdiv, Bulgaria, and their views on further education. We additionally tested whether recognition and attitudes differed by sex, stage of training, and prior CAN education, and identified independent correlates of the intention to report. Methods: A single-centre, analytical cross-sectional survey (Microsoft Forms) was conducted among 4th-, 5th-, and 6th-year (intern) dental students at the Medical University of Plovdiv (March–May 2025). The 17-item questionnaire covered demographics, knowledge, attitudes, the perceived role of the dentist, and interest in further training. An indicator-recognition score (0–24) was derived from four “select-all-that-apply” items. Group differences were examined with Mann–Whitney U and χ²/Fisher tests; independent correlates were identified with a general linear model (recognition score) and binary logistic regression (reporting intention; dentist’s role), with effect sizes reported. Results: Of 200 invited students, 115 responded (57.5%; 74 female, 41 male). Physical abuse was recognised by all participants (100%) but neglect by only ~75%. Female students scored higher than males (adjusted means 19.97 vs. 18.08; p = 0.005; partial η² = 0.07), with no difference by year of study. Overall recognition scores were similar in students with and without prior CAN education (p = 0.713), but trained students more often endorsed the dentist’s safeguarding role (88.3% vs. 57.9%; p = 0.001) and intended to report suspected cases (66.2% vs. 39.5%; p = 0.011). In multivariable models, prior education (adjusted OR 3.31) and the recognition score (adjusted OR 1.32 per point) were each independently associated with the intention to report (both p ≤ 0.012). Over half (51%) considered their training insufficient and 94.8% wanted further education, preferring case-based seminars (71.3%). Conclusions: A clear gap exists between awareness of CAN and readiness to act; prior education related to readiness to act rather than to broader recognition. Structured, practice-oriented training with clear local reporting pathways and interprofessional learning is needed. Full article

Metabolic syndrome (MetS), characterized by a cluster of interrelated metabolic abnormalities including central obesity, elevated blood pressure, dysglycemia, hypertriglyceridemia, and reduced high-density lipoprotein cholesterol (HDL-C), substantially increases type 2 diabetes and cardiovascular disease risk. Conventional genome-wide association studies (GWASs) analyze individual traits or a dichotomized MetS status, only partially capturing its heritability and potentially overlooking variants with shared (pleiotropic) effects across correlated traits. Here, we performed a multiple-trait GWAS using data from the Korean Genome and Epidemiology Study. Using the Korea Biobank Array (K-Chip), we analyzed a discovery cohort from the Ansan and Ansung studies (1490 cases and 3856 controls) and validated the findings in CAVAS (3620 cases and 4461 controls) and HEXA (15,257 cases and 41,807 controls) cohorts. We jointly modeled six MetS-related traits using the complementary multivariate frameworks Multiple Phenotype Association Tests and Genome-wide Efficient Mixed Model Association, alongside a baseline logistic regression. Whereas logistic regression detected 2 APOA5 variants (rs662799 and rs2075291), the multivariate analyses identified 79 significant single-nucleotide polymorphisms; all were replicated in the HEXA cohort. Functional annotation prioritized 27 high-risk variants mapping to 12 genes, whereas pathway analysis (DAVID) implicated eight Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathways related to lipid metabolism. Our findings demonstrate that multivariate analysis substantially improves the identification of pleiotropic susceptibility loci for MetS in the Korean population and indicates candidate genes for early detection and management. Full article

Background/Objectives: Alpha-Klotho is a circulating protein with anti-aging and vasculoprotective properties that may play a role in vascular calcification. This study aimed to investigate the association between serum alpha-Klotho levels and coronary atherosclerotic burden assessed by the Agatston score and CAD-RADS grading, as well as cardiometabolic risk indices. Methods: This single-center, prospective, cross-sectional study included 86 patients undergoing coronary computed tomography angiography. Serum alpha-Klotho levels were measured by ELISA. Associations with imaging and metabolic variables were assessed using Spearman correlation and age-adjusted partial correlation analyses. Independent associations were evaluated using multivariable linear regression with HC3-robust standard errors. Binary and ordinal logistic regression models with CAD-RADS as the dependent variable and receiver operating characteristic (ROC) analysis were additionally performed. Results: Serum Klotho levels were negatively correlated with age (ρ = −0.598), Agatston score (ρ = −0.540), and CAD-RADS (ρ = −0.549) (all p < 0.001). Associations with the Agatston score and CAD-RADS remained significant after age adjustment (ρ = −0.348 and −0.352, respectively). AIP also remained independently associated with Klotho after age adjustment (ρ = −0.289, p = 0.007), whereas associations with TyG and METS-IR were attenuated. Multivariable models explained 50–52% of the variance in Klotho. In ordinal logistic regression, Klotho remained independently associated with CAD-RADS severity after adjustment for age, sex, BMI, and AIP (OR = 0.525, p = 0.033). ROC analysis demonstrated moderate discrimination for the presence of coronary atherosclerosis, defined as Agatston score > 0 (equivalently, CAD-RADS > 0 in this cohort) (AUC = 0.754, 95% CI 0.654–0.855); an exploratory, internally derived cutoff of ≤823 pg/mL—not a clinically validated threshold—yielded 72.0% sensitivity and 75.0% specificity (bootstrap-validated optimism-corrected: 69.3% and 72.5%). Conclusions: Lower serum alpha-Klotho levels were independently associated with greater coronary atherosclerotic burden and showed moderate discriminative ability for its presence. These findings are preliminary and support further investigation of alpha-Klotho as a candidate circulating biomarker of coronary atherosclerosis, pending validation in larger, prospective cohorts. Full article

Lipid metabolites in maize kernels determine grain quality by influencing nutritional value, oxidative stability, and post-harvest deterioration, making their profiling essential for quality improvement and breeding. This study applies hyperspectral imaging (HSI) with a spectral range of 900–1700 nm to detect maize lipid metabolites. Six lipid metabolites—9-Octadecynoic acid (stearolic acid), pinolenic acid (Δ5,9,12 18:3), N-Acylethanolamine (16:0), N-Acylethanolamine (18:0), N-Acylethanolamine (18:1), and propionic acid—were selected due to their strong relevance to maize kernel quality and favorable spectral response. First, two-trace two-dimensional (2T2D) correlation spectroscopy with heterogeneous preprocessing is employed to capture both synchronous and asynchronous correlations across different preprocessing spectra. A convolutional autoencoder (CAE) was subsequently used to extract low-dimensional latent features from heterogeneous 2T2D-COS representations, followed by regression modeling using random forest (RF), support vector regression (SVR), gradient boosting (GB), and partial least squares regression (PLSR). A total of 82 maize seed varieties were employed for experimental validation. Compared with one-dimensional spectral, homogeneous preprocessing, and PCA-based feature extraction, the proposed approach provided improved predictive performance across the six lipid metabolites, with the optimal CAE-based models achieving R_P² values of 0.629–0.887, RMSEP values of 0.241–0.565, and RPD values of 1.656–2.069. Overall, this approach provides a rough screening solution for metabolite prediction in maize crop. Full article