Search Results (699)

Background: Stroke remains a leading cause of mortality, disability, and health-system burden in Korea’s rapidly aging population. We aimed to describe national stroke incidence trends from 2011 to 2023, characterize ecological associations between stroke incidence and health-screening indicators, and generate model-based projections through 2028 to support health-system planning. Methods: This retrospective ecological longitudinal analysis used three publicly available aggregate national data sources: (1) NHIS annual aggregate statistics on crude and age-standardized stroke incidence, stroke case counts, first-onset vs. recurrent stroke, and case-fatality rates (2011–2023); (2) regional standardized health-awareness survey rates for stroke symptoms, myocardial infarction symptoms, blood pressure, and blood glucose (2017–2025); and (3) national cancer-screening outcome tallies for breast and cervical cancer (2010–2024). All analyses used pre-aggregated annual summary data; individual-level NHIS records were not used. Annual trends were modeled with ordinary least-squares linear regression (n = 13 annual observations). Pearson correlations were computed only for overlapping observation windows. Model-based projections are presented with 95% prediction intervals and are explicitly distinguished from observed NHIS values. This study is purely descriptive and ecological; no causal inference is made. Results: Crude stroke incidence increased from 199.2 to 221.1 per 100,000 (2011–2023; slope +2.32/year, R² = 0.83), whereas age-standardized incidence declined from 158.3 to 113.2 per 100,000 (slope −3.41/year, R² = 0.96), a pattern consistent with demographic aging as a contributing factor to growing absolute burden, though formal age-decomposition analysis would be required to confirm this inference. Total cases increased from 99,837 to 113,098; the 30-day case-fatality rate declined from 8.5% to 7.5%. Ecological correlations showed that blood glucose awareness was strongly negatively correlated with age-standardized incidence (r = −0.944, p = 0.001, n = 7), though these are ecological associations and must not be interpreted as individual-level causal relationships. Model-based projections estimate crude incidence near 230.7 (95%PI 219.2–242.2) and age-standardized incidence near 103.2 (95%PI 95.7–110.8) per 100,000 by 2026. Conclusions: Concurrent increase in crude burden and decline in age-standardized incidence reflects demographic aging as the primary driver of Korea’s stroke burden. Projections support integrated cardiovascular prevention, public health education, and age-sensitive service planning. All projections are short-horizon statistical extrapolations intended for policy scenario planning only and must not be interpreted as observed future NHIS outcomes. Full article

Background: Dupilumab, a monoclonal antibody blocking IL-4Rα, has recently demonstrated efficacy in patients with type 2 (T2)-inflamed chronic obstructive pulmonary disease (COPD) in the BOREAS and NOTUS trials. Real-world experience in older patients with predominant chronic bronchitis phenotype remains limited. Methods: We report a single-centre case series of 12 consecutive patients with T2-inflamed COPD treated with dupilumab 300 mg every two weeks under a compassionate-use programme at San Carlo di Nancy Hospital, Rome (first administration: April 2025). Eligibility required ≥2 moderate or ≥1 severe exacerbation in the prior 12 months despite triple inhaled therapy and a blood eosinophil count ≥300 cells/µL. Follow-up ranged from 3 to 12 months, with 6 months pre-specified as the primary analysis timepoint; data at 9 and 12 months are reported as descriptive observations. Endpoints included paired changes in annualised exacerbation rate (AER), CAT score and item-level CAT, and FEV₁, with exploratory univariate Spearman analyses of candidate baseline predictors of response. Results: The cohort was elderly (mean age 73.6 ± 5.2 years, range 65–82), predominantly female (8/12, 67%) and characterised by a chronic bronchitis phenotype with high symptom burden (mean baseline CAT 22.8 ± 7.5; CAT item 2 [phlegm] median 3, IQR 3–4). Severe exacerbations decreased significantly (Wilcoxon p = 0.0156; mean AER 0.75 → 0.19 events/patient-year; 6/12 improved, 0/12 worsened). The mean cumulative function showed a standardised incidence ratio of 0.46 (95% CI 0.19–0.95; p = 0.033) versus the pre-dupilumab rate. Mean FEV₁ increased by +66 mL at 1 month (n = 11, paired Wilcoxon p = 0.025), +78 mL at 3 months (n = 10, p = 0.082) and +120 mL at 6 months (n = 10, p = 0.007). Total CAT decreased from 22.9 to 12.5 at 6 months (Friedman p = 0.0007), with the largest absolute reductions in item 2 (phlegm; Δ = −2.6 at 6 months, p < 0.001) and item 3 (chest tightness; Δ = −2.5 at 6 months, p = 0.002). Higher baseline CAT was associated with greater reduction in severe AER (Spearman ρ = −0.79, p = 0.002). Conclusions: In this elderly real-world cohort with phlegm-driven T2 COPD, dupilumab was associated with a significant decrease in severe exacerbations, a clinically meaningful gain in lung function and a marked improvement in mucus-related symptoms. Further studies are warranted to confirm these findings and to clarify whether the reduction in severe exacerbations translates into a measurable mortality benefit. Full article

Manually counting sunflower seeds on capitula is labor-intensive, requiring approximately one person-hour per head, and can be inconsistent for densely packed heads. Existing phenotyping approaches often depend on laboratory-based equipment, limiting their accessibility. In this study, we developed a benchtop image-based pipeline for rapid, non-destructive estimation of developed and aborted seeds on intact dried sunflower heads. A dataset of 1093 sunflower capitula was imaged under fixed indoor lighting, and individual seeds were annotated as developed or aborted. A YOLOv8m one-stage object detector was trained and evaluated using a counting-focused protocol, in which a single confidence threshold was selected on the validation set and then applied unchanged to an independent test set of 109 images. The baseline model was compared with recent YOLO variants and different augmentation strategies. On the test set, the model achieved a mean absolute count error of 61.3 seeds per image, a mean relative error of 12.0%, and an mAP50 of 0.18 at the locked confidence threshold of 0.15. Only 13.8% of test images had relative errors below 2%. Larger YOLO models and augmentation variants did not improve performance. These findings show that the proposed system provides approximate, non-destructive seed-count estimation under controlled imaging conditions, while highlighting the need for improved localization in dense regions and domain adaptation for fresh heads or field conditions. The annotated dataset and trained model weights are made available to support reproducible research. Full article

Background/Objective: During the last two decades, several cases presenting with circulating clonal B-cells with features consistent with possible marginal zone (MZ) derivation have been described under different terminologies. The present study aims to shed more light onto the main disease characteristics and determine prognostic factors for outcome in 98 consecutive cases with clonal B-cell lymphocytosis of marginal zone origin (CBL-MZ). Methods: This is a multicenter retrospective analysis including 98 consecutive CD5(−) CBL cases referred to our Departments between 1999 and 2017. These cases were selected based on the presence of circulating CD5(−) clonal B-cells, irrespectively of their absolute number, without B-symptoms, organomegaly, lymphadenopathy or cytopenias or any other features consistent with a known lymphoproliferative disorder. Clinical, morphologic, biochemical, immunophenotypic, histologic and molecular features of CBL-MZ cases were analyzed. Results: The median absolute lymphocyte counts (ALCs) and circulating CBLs were 6.7 × 10⁹/L and 3.447 × 10⁹/L, respectively. Paraproteinemia was found in 38%. Bone marrow (ΒΜ) was involved in all but one case. MYD-88L265P mutation was positive in 11%. Two subcategories of CBL-MZ were identified: One was characterized by lower ALC/CBL, paraproteinemia, CD38 expression, BM lymphoplasmacytic morphology and more frequent MYD-88L265P mutation. The second category displayed a leukemic picture, higher frequency of CD11c expression, hypogammaglobulinemia and elevated LDH. Treatment-free survival (TFS) was 91%, and median freedom from progression (FFP) was 95.6 months. For TFS, two factors proved significant using multivariate analysis: BM infiltration ≥ 50% and elevated LDH (RR 5.6 and 5.4, respectively). Evolution to splenic marginal zone lymphoma was a rare event (5%). A novel pattern of progression emerged, namely development of cytopenias due to extensive BM infiltration without any other disease localization. Conclusions: CBL-MZ is an indolent lymphoproliferative disorder with excellent outcome and low probability of progression. Based on our findings, CBL-MZ represents a heterogeneous entity where the vast majority of cases remain stable or develop increasing lymphocytosis. Clinically, our data also highlight that the extent of BM infiltration and elevated LDH levels appear to be the most notable predictors for introducing therapy. Full article

Shear-wave velocity (V_s) estimation from the Standard Penetration Test (SPT) can support preliminary site characterization when direct geophysical data are limited, but empirical correlations require validation schemes that reflect transferability between sites. This study evaluates V_s prediction using an interval-paired dataset derived from geotechnical investigations of school foundations in Piura, Peru. Its novelty lies in comparing the raw SPT blow count (NSPT) and the overburden- and energy-corrected SPT blow count ((N₁)₆₀) on the same strict common sample, using grouped cross-validation by school, thereby emphasizing transferability across sites rather than only internal fit. Five predictive scenarios were tested, from penetration-only formulations to geotechnically enriched specifications. The lowest grouped out-of-fold error among the evaluated models was obtained by a generalized power baseline using (N₁)₆₀ and the integral geotechnical predictor set, yielding root mean square error (RMSE) = 80.48 m/s, mean absolute error (MAE) = 60.15 m/s, and coefficient of determination (R²) = 0.338. This moderate R² indicates limited standalone predictive capacity under transfer to unseen schools; therefore, the model is interpreted as a preliminary transfer-oriented correlation rather than as a substitute for direct V_s measurements or as an independent design equation. In the complementary full-data analysis, the strongest descriptive fit was obtained with Hist Gradient Boosting, whereas the strongest explicit equation corresponded to the log-semi baseline. Overall, the findings show that externally validated transferability, descriptive full-data fit, and equation-based interpretability represent different analytical roles in V_s-SPT modeling. Full article

Flow cytometry is an essential diagnostic method in hematology, and one of its main applications is the assessment of the clonality of mature B cells. We present a case report of a patient referred for the investigation of absolute lymphocytosis. The flow cytometry study revealed an increased percentage of B cells, but it could not establish B-cell clonality, based on the study of surface light chains in combination with the pattern of expression of mature B-cell markers. The diagnosis of Persistent Polyclonal B-cell Lymphocytosis (PPBL) was considered in the differential diagnosis as the mature B cells were found to be immunophenotypically memory B cells. However, due to the markedly elevated count of B cells, molecular testing with Polymerase Chain Reaction (PCR) for B-cell clonality based on IGH (Immunoglobulin Heavy Chain) gene rearrangements was performed, and it revealed the presence of two clones of B cells. Approximately one year later, the same work-up was repeated in the patient’s bone marrow aspirate. By flow cytometry, a distinct clonal B-cell population was isolated, while the molecular testing with PCR for B cell clonality based on IGH heavy-chain gene rearrangements revealed the presence of three clones of B cells. In addition, evaluation of the sample with high-dimensional mass cytometry showed the presence of four major immunophenotypically abnormal B-cell subsets. Full article

Background: Obesity is a chronic endocrine–metabolic disorder. The risk of comorbidities increases with a higher body mass index (BMI), particularly when BMI ≥ 35.0 kg/m². Common complications include insulin resistance, type 2 diabetes, dyslipidemia, and chronic low-grade inflammation, which collectively impair DNA stability by promoting the formation of genotoxic species. Methods: This non-randomised, quasi-experimental clinical intervention study included 53 participants (both sexes) with a BMI ≥ 35.0 kg/m², who were assigned to parallel experimental or control streams based on clinical needs and institutional eligibility. During a three-week intervention, the experimental group received a hospital-supervised very-low-calorie diet (VLCD; ~600 kcal/day) under continuous medical monitoring. Conversely, the control group followed a standard reduced-calorie diet (SRD) of 1500 kcal/day in a free-living home environment. Before and after the intervention, primary, oxidative, and permanent DNA damage were measured using alkaline, FPG-modified comet (peripheral blood mononuclear cells), and cytokinesis-block micronucleus cytome assays (phytohaemagglutinin-stimulated binucleated lymphocytes), alongside anthropometric and biochemical tracking. Results: Within-group evaluations revealed that both dietary regimens improved several metabolic health indicators, notably modulating insulin resistance, lipid profiles, and leukocyte counts. However, participants in the VLCD stream experienced significantly greater downward changes in body weight, BMI, and absolute lipid values. Crucially, the VLCD intervention was associated with a highly significant within-group reduction in parameters of permanent chromosomal damage, effectively halving the frequencies of micronuclei and nuclear buds, independent of baseline variations, in adjusted multivariate regression models. Conversely, the home-based SRD regimen demonstrated no measurable impact on permanent genomic damage. Neither diet induced a significant change in repairable primary or oxidative DNA lesions over this short timeframe. Conclusions: These exploratory findings suggest that strict calorie restriction can rapidly stabilise genome stability in advanced clinical settings, warranting future randomised controlled trials with long-term longitudinal follow-up to assess permanent risk reductions. Due to structural baseline variations in age, chronic comorbidities, and compliance environments between the cohorts, direct comparative superiority cannot be definitively established. Full article

Urban planners need scalable ways to monitor pedestrian conditions across heterogeneous cities, but conventional Level-of-Service (LOS) methods are often locally calibrated and difficult to transfer. This study proposes a city-adaptive framework for pedestrian LOS mapping using spatio-temporal graph models and few-shot transfer learning. Pedestrian count data from Melbourne, Dublin, and Zurich were converted into six ordinal LOS classes using city-specific percentile thresholds computed from the training data, yielding a relative congestion measure rather than an absolute cross-city standard. We developed a spatio-temporal graph transformer with an ordinal prediction head and evaluated it under in-domain, zero-shot, few-shot, and domain-adaptive settings. The results show strong in-domain performance in Melbourne (accuracy 79.7%; Acc ± 1 99.1%) and effective adaptation to the city-adaptive ordinal classification task. Few-shot fine-tuning with only 5% labeled target city data recovered 95–99% of in-domain performance, suggesting that small amounts of local supervision can substantially reduce calibration requirements in data-scarce environments. KernelSHAP analysis indicates that short-term temporal lag features dominate predictions across cities, whereas spatial and contextual features vary more strongly with local urban structure. The findings suggest that few-shot transfer learning can support pedestrian LOS estimation in cities with limited labeled data; however, the proposed LOS formulation should be interpreted as a city-specific relative indicator rather than an absolute measure of pedestrian comfort, crowding, or service quality. While the framework was evaluated across three cities, additional validation in diverse urban contexts and against perceptual measures of pedestrian experience remains necessary. Overall, the study contributes a city-adaptive framework for transferable relative LOS prediction rather than a universal cross-city LOS standard. Full article

Background: Baseline lymphopenia is common among advanced solid tumors and may influence the efficacy/safety of immune checkpoint inhibitors (ICIs), but large real-world evidence is limited. We evaluated the association between baseline absolute lymphocyte count (ALC) and clinical outcomes in adults with solid tumors treated with ICIs in routine practice. Methods: We performed a retrospective cohort study using TriNetX. Adults with solid tumors who received pembrolizumab, nivolumab, or atezolizumab (ICI-Naïve) between January 2015 and June 2026 were included. Baseline ALC was measured within 30 days before first treatment and was classified as lymphopenic (ALC < 1.5 × 10⁹/L) or non-lymphopenic (ALC ≥ 1.5 × 10⁹/L). Propensity score matching (1:1) yielded 5249 patients per group. The index date was the first immunotherapy date, and outcomes were assessed at 6, 12, 24, 36 months, and 5 years. The primary outcome was 24-month overall survival (OS); secondary outcomes were OS at 6 and 12 months and 6-month risks of healthcare utilization, immune-related adverse events (irAEs), and serious infections; and exploratory outcomes included OS at 36 months and 5 years. All outcomes were analyzed using Kaplan–Meier analysis, Cox proportional hazards models, and risk ratios. Subgroup analysis included OS stratified by solid tumor subtypes and prior lines of therapy. Results: After matching, patients with baseline lymphopenia had consistently worse OS. Compared with patients without lymphopenia, the lymphopenia cohort had lower OS at 6 months (HR 1.29, 95% CI 1.22–1.37), 12 months (HR 1.28, 95% CI 1.21–1.35), 24 months (HR 1.26, 95% CI 1.2–1.33), and, in exploratory analyses with substantial right censoring and limited observed follow-up, 36 months (HR 1.26, 95% CI 1.2–1.33) and 5 years (HR 1.26, 95% CI 1.2–1.33), though these estimates should be considered hypothesis-generating only. At 6 months, baseline lymphopenia was associated with a greater healthcare utilization (RR 1.05, 95% CI 1.02–1.09), a higher infection risk (RR 1.08, 95% CI 1.01–1.15), and similar rates of clinically coded irAEs (RR 1.0, 95% CI 0.93–1.09), an observation subject to competing risk from early mortality in the lymphopenic cohort. Subgroup analysis, stratified by tumor subtypes and prior lines of therapy, showed consistently lower OS in the lymphopenia group, consistent with the primary outcome results. Conclusions: In this large propensity-matched real-world analysis of 10,498 patients with diverse solid tumors, baseline lymphopenia at ICI initiation was associated with persistently inferior OS at 6, 12, and 24 months (primary and secondary endpoints), greater early healthcare utilization, and a higher serious infection risk. Critically, lymphopenic patients developed irAEs at an identical rate to non-lymphopenic patients despite worse survival, a dissociation suggesting that baseline ALC stratifies patients along mortality risk and immune activation capacity as partially independent axes. These findings could support the use of baseline ALC as a simple, universally available biomarker that informs not only survival prognosis but also the anticipated toxicity profile of ICI therapy and highlight the need for competing-risk analyses and prospective immune phenotyping to characterize this relationship fully. Full article

Urban Unmanned Aerial Vehicle (UAV) logistics operations are frequently constrained by the intersection of limited battery endurance and dynamic ground traffic. When mobile depots are delayed by congestion, onboard drone fleets experience extended idling periods, leading to constrained sortie generation and reduced asset utilization. To address this bottleneck, this paper introduces a traffic-predictive multi-UAV dispatch framework for deterministic day-ahead planning under modeled urban operating conditions. By coupling a count-derived macroscopic speed surrogate learned using XGBoost with a Particle Swarm Optimization (PSO)–Mixed-Integer Linear Programming (MILP) optimization architecture, the framework synchronizes mobile depot trajectories with forecasted low-congestion windows and pre-allocates endurance-feasible parallel aerial sorties. Controlled computational experiments across 30 synthetic routing instances demonstrate the potential value of this approach within the stated modeling assumptions. Compared to baseline clustered deployments, the traffic-aware framework raises mean fleet utilization from 0.43 to 0.63—a 46.2% relative improvement driven by temporal compression of the mission window rather than an absolute increase in flight hours. Furthermore, the proposed framework reduces total mission completion time by 69.87% relative to the conventional truck-only baseline, while achieving a 29.58% incremental gain over static speed drone deployments. These findings suggest that incorporating predictive ground traffic information into day-ahead UAV scheduling can improve modeled fleet efficiency; however, field validation with measured route-level speeds, real delivery demand, and operational constraints remains necessary before deployment-level claims can be made. Full article

Fourier Neural Operators (FNOs) exhibit mode saturation on high-contrast inhomogeneous media, and recent multi-scale extensions (MscaleFNO) further worsen out-of-distribution (OOD) generalization. We introduce the Helmholtz Neural Operator (HNO), a physics-informed, FFT-free branch–trunk operator in the DeepONet family, with a hybrid SIREN+learnable-Fourier trunk and a dual-path rank-32 hypernetwork branch, with bounded multiplicative gating on per-mode coefficients. At a matched parameter count (∼1.05 M, five seeds), HNO achieves a 2.6× lower OOD generalization gap than FNO (19.6% vs. 50.6%, $p=1.7\times {10}^{-3}$, Cohen’s $d=5.1$), 5.1× lower than vanilla DeepONet (19.6% vs. 99.9%, $p=8.2\times {10}^{-3}$), and 6.0× lower than MscaleFNO (19.6% vs. 117.4%, $p=2.4\times {10}^{-6}$); MscaleFNO’s deficit grows at 4.2× more parameters, ruling out capacity starvation. HNO is 4.6×/16.4× faster than FNO/MscaleFNO and 64×–245× faster than multi-threaded FD-PML (MKL PARDISO, 12 cores; 183×–698× vs. single-thread scipy.spsolve), making it suitable as a forward surrogate inside many-query workflows. Absolute accuracy on extreme-contrast (15:1) OOD samples is limited (relative $L^2\approx 1$), so HNO is positioned as a many-query surrogate or warm start for refinement loops, not a stand-alone replacement for direct solvers. A scope limitation is that HNO underperforms FNO on elliptic Darcy Flow, confirming specialization for hyperbolic/wave equations rather than universal operator learning. Full article

For microbiological diagnostics, accurately counting and segmenting microbial colonies is extremely important. However, manual methods are labor-intensive and yield inconsistent results. We develop a hybrid model using swarm intelligence, combining a convolutional transformer with nested skip connections and global context with channel and spatial attention. Parameter tuning is supported by a variety of swarm optimization algorithms (e.g., Particle Swarm Optimization, Quantum-behaved Particle Swarm Optimization, and Differential Evolution Particle Swarm Optimization). Morphological refinement, including a further watershed transform, an attention graph, and post-processing, enhances colony boundaries by separating them. Grad-CAM++, Integrated Gradients, and temperature scaling provide a transparent and trustworthy model through explainability and post hoc calibration. The proposed model was extensively tested on the Microbial Colony Recognition and Circular Bacterial Colony Datasets, achieving a Dice score of 94.2%, an Intersection over the Union of 88.6%, and a mean absolute counting error of 2.7 colonies. These results significantly outperform several baseline models, including U-Net (88.1%), U-Net++ (89.7%), Attention U-Net (90.6%), and Swin-Unet (91.4%). Statistically significant improvements were confirmed (p < 0.01). A cross-dataset analysis demonstrates the framework’s robustness and cross-domain applicability, and positions it as a trustworthy, explainable automated model for assessing microbial colonies in laboratory and clinical settings. Full article

Background and Objectives: Brucellosis remains a significant zoonotic disease in endemic regions such as Saudi Arabia, with children being particularly vulnerable. Pediatric brucellosis often presents with nonspecific symptoms, and hematological abnormalities can serve as important yet underrecognized diagnostic clues. This study aimed to evaluate the demographic, clinical, microbiological, treatment, and hematological characteristics of patients with pediatric brucellosis in the Hail region of Saudi Arabia and to assess the diagnostic value of hematological parameters. Methods: This retrospective observational study included children aged ≤15 years who were diagnosed with brucellosis at a tertiary care hospital in Hail between 2014 and 2025. Demographic, clinical, laboratory, microbiological, and treatment data were analyzed. Hematological parameters were compared between culture-confirmed and non-culture-confirmed brucellosis cases using multivariate and receiver operating characteristic (ROC) analyses. Results: A total of 38 pediatric patients were included (mean age 8.6 years; 57.9% male). Positive culture results were observed in 42.1% of the cases, with Brucella melitensis being the predominant species (68.75%). Fever (89.5%) and bone/joint pain (71.1%) were the most frequent symptoms. Culture-confirmed brucellosis patients had significantly lower hemoglobin levels (10.8 vs. 12.1 g/dL; p = 0.020), white blood cell counts (p = 0.046), and absolute neutrophil counts (p = 0.037). ROC analysis revealed a fair diagnostic performance for hemoglobin (AUROC = 0.695), WBC (0.699), and ANC (0.680). Leukopenia demonstrated high specificity (95.5%) and positive predictive value. Conclusions: Pediatric brucellosis is commonly associated with anemia, leukopenia, and neutropenia. Although no single hematological parameter independently predicts infection, the combination of these abnormalities may support early clinical suspicion, particularly in resource-limited endemic settings. Full article

Background/Objectives: Angiogenic T (Tang) cells support endothelial repair and vascular homeostasis. This cross-sectional study compared circulating Tang cell levels in young adults with T1DM vs. healthy controls, and assessed associations between Tang cells and continuous glucose monitoring (CGM) metrics. Methods: Sixty-five young adults with T1DM and 55 healthy controls were enrolled at the University of Campania “Luigi Vanvitelli,” Naples, Italy. Clinical and biochemical data were collected. Tang cells (CD3⁺CD31⁺CD184⁺) were quantified by flow cytometry as absolute counts and percentage of CD3⁺ T cells. In T1DM, CGM metrics from the preceding 14 days were analyzed, including time in range (TIR), time above range (TAR), and time below range (TBR). Results: Individuals with T1DM had higher fasting glucose and HbA1c than controls. Total CD3⁺ T cell counts were lower in T1DM. Tang cells were significantly reduced in T1DM both as absolute number and percentage (21% [10–31] vs. 48% [39–62]; p < 0.001). In multivariable analyses, Tang cell percentage was positively associated with TIR and inversely associated with HbA1c and TAR. Conclusions: Young adults with T1DM exhibit significantly reduced circulating Tang cells. Associations with CGM metrics support a link between real-world glucose control and endothelial vascular health. Full article

Accurate bee counting is important for colony monitoring, pollination assessment, and precision beekeeping, but manual counting and dense point annotation are labor-intensive. This study proposes M3DANet, a lightweight semi-supervised density regression network with a handheld edge deployment system for bee colony counting. A dataset containing 586 valid high-resolution images and 34,869 point annotations was constructed for training and evaluation. M3DANet uses the first seven stages of MobileNetV3-Large as the lightweight backbone and combines multi-scale context encoding, attention-guided low-level feature fusion, and teacher–student consistency learning with confidence masking and warm-up training. The 10%, 30%, and 50% labeled data settings refer to the proportions of labeled images in the training set, and the remaining training images are used as unlabeled data. Mean absolute error (MAE) and root mean square error (RMSE) are used as evaluation metrics. On the main dataset, M3DANet achieved MAE values of 9.937, 7.003, and 5.570 and RMSE values of 13.093, 9.387, and 7.620 under the 10%, 30%, and 50% settings, respectively, outperforming representative semi-supervised baselines. Under the fully supervised setting, it achieved an MAE of 5.201 and an RMSE of 6.989 with only 2.095 M parameters and 416.64 FPS, using 87.1% fewer parameters and running 17.7 times faster than CSRNet. Cross-species experiments confirmed its low-label generalization ability. Jetson Orin NX deployment achieved 65.75 ms/image inference latency and 10.44 FPS complete-pipeline throughput. These results show that M3DANet balances counting accuracy, annotation efficiency, generalization, and edge deployment practicality. Full article