Search Results (2,506)

Background: Achieving a complete response after therapy is an important predictor of long-term survival in lymphoma patients. However, previous predictive models have primarily focused on overall survival (OS) and progression-free survival (PFS), often overlooking treatment response. Predicting the likelihood of complete response before initiating therapy can provide more immediate and actionable insights. Thus, this study aims to develop and validate predictive models for treatment response to first-line therapy in patients with B-cell lymphomas. Methods: The study used 2763 patients from the Lymphoma and Related Diseases Registry (LaRDR). The data were randomly divided into training (n = 2221, 80%) and validation (n = 553, 20%) cohorts. Seven algorithms: logistic regression, K-nearest neighbor, support vector machine, random forest, Naïve Bayes, gradient boosting machine, and extreme gradient boosting were evaluated. Model performance was assessed using discrimination and classification metrics. Additionally, model calibration and clinical utility were evaluated using the Brier score and decision curve analysis, respectively. Results: All models demonstrated comparable performance in the validation cohort, with area under the curve (AUC) values ranging from 0.69 to 0.70. A nomogram incorporating the six variables, including stage, lactate dehydrogenase, performance status, BCL2 expression, anemia, and systemic immune-inflammation index, achieved an AUC of 0.70 (95% CI: 0.65–0.75), outperforming the international prognostic index (IPI: AUC = 0.65), revised IPI (AUC = 0.61), and NCCN-IPI (AUC = 0.63). Decision curve analysis confirmed the nomogram’s superior net benefit over IPI-based systems. Conclusions: While our nomogram demonstrated improved discriminative performance and clinical utility compared to IPI-based systems, further external validation is needed before clinical integration. Full article

Background/Objectives: Cardiovascular disease remains a leading cause of death worldwide, with chest pain often serving as an initial reason for emergency visits. However, the severity of chest pain does not necessarily correlate with the severity of myocardial infarction. Facial expressions are an essential medium to convey the intensity of pain, particularly in patients experiencing speech difficulties. Automating the recognition of facial pain expression may therefore provide an auxiliary tool for monitoring chest pain without replacing clinical diagnosis. Methods: Using streaming technology, the system captures real-time facial expressions and classifies pain levels using a deep learning framework. The PSPI scores were incorporated with the YOLO models to ensure precise classification. Through extensive fine-tuning, we compare the performance of YOLO-series models, evaluating both computational efficiency and diagnostic accuracy rather than focusing solely on accuracy or processing time. Results: The custom YOLOv4 model demonstrated superior performance in pain level recognition, achieving a precision of 97% and the fastest training time. The system integrates a web-based interface with color-coded pain indicators, which can be deployed on smartphones and laptops for flexible use in healthcare settings. Conclusions: This study demonstrates the potential of automating pain assessment based on facial expressions to assist healthcare professionals in observing patient discomfort. Importantly, the approach does not infer the underlying cause of myocardial infarction. Future work will incorporate clinical metadata and a lightweight edge computing model to enable real-time pain monitoring in diverse care environments, which may support patient monitoring and assist in clinical observation. Full article

High-grade serous carcinoma (HGSC) of the ovary is characterized by two major histological patterns: a classic papillary/micropapillary architecture and a solid pseudo-endometrioid transitional (SET) variant. We investigated whether the distinct morphologic subtypes are underpinned by transcriptomic differences in the tumor microenvironment (TME). We profiled 21 HGSC tumors (7 SET, 14 classic) using a 770-gene NanoString PanCancer Progression panel. Differential expression analysis revealed ~20 genes with significantly different expression (>4-fold, adjusted p < 0.01) between SET and classic tumors. Unsupervised clustering partially separated SET and classic tumors, suggesting that global gene expression patterns correlate with histologic subtype. SET tumors exhibited upregulation of cell-cycle and epithelial genes (e.g., PTTG1, TRAIL, HER3) and downregulation of genes involved in epithelial–mesenchymal transition (EMT), extracellular matrix (ECM) organization, and angiogenesis (e.g., TWIST2, FGF2, decorin) relative to classic tumors. Notably, PTTG1 and TRAIL were upregulated ~6–9-fold in SET tumors, whereas TWIST2 was ~7-fold downregulated, consistent with reduced EMT in SET tumors. Pathway analysis indicated that SET tumors appear to have an immune-active, stroma-poor microenvironment, in line with an “immunoreactive” phenotype, whereas classic tumors showed a mesenchymal, stroma-rich profile. These molecular distinctions could have diagnostic utility and may inform therapeutic stratification, with key dysregulated genes (e.g., HER3, TRAIL, FGF2) representing potential prognostic or predictive biomarkers. For example, high HER3 expression in SET tumors might predict sensitivity to ERBB3/PI3K inhibitors, whereas stromal factors (e.g., FGF2) enriched in classic HGSC could be targeted with microenvironment-modulating therapies. These preliminary findings require validation before translation into pathology practice via immunohistochemical (IHC) assays (e.g., for HER3 or TRAIL), potentially enabling improved classification and personalized treatment of HGSC. We report effect sizes as log₂ fold change with 95% confidence intervals and emphasize FDR-adjusted q-values. Given the small sample size and the absence of outcome data (OS/PFS/PFI), results are preliminary and hypothesis-generating. Orthogonal protein-level validation and replication in larger, independent cohorts are required before any translational inference. Full article

Background/Objective: Temporal eating patterns and ultra-processed food (UPF) consumption have independently been associated with obesity and non-communicable diseases. Little is known about the temporal patterns of UPF consumption, as data is challenging to collect. Temporal data can be extracted from mobile food records (mFRs). The aim of this study was to identify the temporal eating patterns of those consuming UPFs using an mFR. Methods: A combined sample of 243 young (18–30 years) and 148 older (>30 years) adults completed a 4-day mFR. The time of eating was extracted from the mFR image metadata. UPFs were identified using the NOVA food classification system. The proportion of total energy intake (EI) from UPFs was calculated hourly. Using chi-square tests, a day-of-the-week analysis compared weekends (Friday–Sunday) with weekdays (Monday–Thursday). A multivariate logistic regression of UPF EI terciles was conducted, expressed as odds ratios and 95% confidence intervals. Results: The proportion of total EI from UPFs was significantly different between younger adults (mean ± SD = 48.8 ± 15.6%) and older adults (36.1 ± 15.1%) (p < 0.001). Age-differentiated 24 h temporal eating pattern analysis found that younger adults had two distinct UPF EI peaks, with the highest at 8 pm, followed by 1 pm. Older adults followed a more conventional three-meal pattern with an additional peak at 7 am. Weekend UPF EI was higher than on weekdays for older adults (~560 kJ, p = 0.003), with no difference for younger adults. Multivariable logistic regression found no significant associations between UPF intake terciles and demographic variables (sex, BMI, education). Conclusions: The peak UPF EI occurred at conventional mealtimes, and UPFs accounted for a substantial proportion of energy intake, especially for younger adults. The timing of UPF EI provides important information for developing public health nutrition interventions. Full article

Background/Objectives: Ultra-processed food consumption has been shown to be linked with clinical cardiovascular disease. This study aims to examine the associations of ultra-processed food consumption with biomarkers for subclinical-level myocardial damage [high-sensitivity cardiac troponin I and T (hs-cTnI and hs-cTnT)] and myocardial stretch (NT-proBNP) in U.S. adults. Methods: We used data from 6615 U.S. adults aged ≥20 years without prevalent cardiovascular disease from the National Health and Nutrition Examination Survey 2001–2004. We identified ultra-processed food by applying the Nova classification to dietary recall data, and we divided participants into quartiles based on their consumption, expressed as a proportion of total daily energy (%kcal) and gram intakes (%grams). We defined elevated cardiac biomarkers as hs-cTnI > 12 ng/L in men and >10 ng/L in women, hs-cTnT ≥ 14 ng/L for all participants, and NT-proBNP ≥ 125 pg/mL for age < 75 y and ≥450 pg/mL for age ≥ 75 y. We used multivariable logistic regression with adjustment for socio-demographic, total energy intake, behavioral, and clinical characteristics. Results: Higher ultra-processed food intake in %grams was associated with elevated NT-proBNP [odds ratio (OR) for quartile 4 vs. 1: 1.27, 95% CI: 1.00–1.61] when socio-demographic characteristics and total energy intake were adjusted for, but this was not the case with hs-cTnI or hs-cTnT. Further adjusting for clinical characteristics attenuated the association with NT-proBNP (OR: 1.26, 95% CI: 0.98, 1.61). There was no consistent association between ultra-processed food in %kcal and elevated NT-proBNP, hs-cTnT, or hs-cTnI. Conclusions: Ultra-processed food consumption is associated with subclinical myocardial stretch, a precursor to early heart failure. This supports the potential risks of subclinical cardiovascular disease associated with consuming ultra-processed food. Full article

Elucidating drug response mechanisms in human melanoma is crucial for improving treatment outcomes. Moreover, the existing tools intended to provide deeper insight into melanoma drug resistance have notable limitations. Therefore, we propose a deep learning (DL)-based approach that works as follows. First, we processed two single-cell datasets related to human melanoma from the GEO (GSE108383_A375 and GSE108383_451Lu) database and trained a fully connected neural network with five adapted methods (L1-Regularization, DeepLIFT, SHAP, IG, and LRP). We then identified 100 genes by ranking all genes from the highest to the lowest based on the sum of absolute values for corresponding weights across all neurons in the first hidden layer. From a biological perspective, compared to existing bioinformatics tools, the presented DL-based methods identified a higher number of expressed genes in four well-established melanoma cell lines: MALME-3M, MDA-MB435, SK-MEL-28, and SK-MEL-5. Furthermore, we identified FDA-approved melanoma drugs (e.g., Vemurafenib and Dabrafenib), critical genes such as ARAF, SOX10, DCT, and AXL, and key TFs including MITF and TFAP2A. From a classification perspective, we utilized five-fold cross-validation and provided gene sets using all the abovementioned methods to three randomly selected machine learning algorithms, namely, support vector machines, random forests, and neural networks with different hyperparameters. The results demonstrate that the integrated gradients (IG) method adapted in our DL approach contributed to 2.2% and 0.5% overall performance improvements over the best-performing baselines when using A375 and 451Lu cell line datasets. Additional comparison against no gene selection demonstrated that IG is the only method to generate statistically significant results, with 14.4% and 11.7% overall performance improvements. Full article

Background/Objectives: The epidemiology of dermatomyositis (DM) and other idiopathic inflammatory myopathies (IIMs) remains not well established, especially in the Mediterranean region. We aimed to estimate the prevalence and incidence of IIM in a well-defined population of South Europe using standardized classification criteria. Methods: This population-based study included all IIM patients diagnosed from January 2000 to December 2022 in Cantabria, Northern Spain. IIM diagnosis was confirmed by fulfillment of the 2017 EULAR/ACR classification criteria or, alternatively, by European Neuro Muscular Center criteria for immune-mediated necrotizing myopathy (IMNM) and Connors’ criteria for antisynthetase syndrome (ASyS). Prevalence and incidence were expressed in cases per 100,000. A literature review was also performed. Results: A total of 60 patients (41 women, 19 men; mean age 52.6 ± 18.8 years) were included. The prevalence of IIM was 20 cases per 100,000 population [95% CI 14.5–25.1], and the annual incidence rate was 0.9 cases per 100,000 person-years [95% CI 0.6–1.14]. A significant upward trend in IIM incidence was observed with an estimated annual percentage change of 5.74% (95% CI: 2.16%–9.44%, p = 0.0015). The most common subtype was DM (n = 31, 51.7%), followed by ASyS (n = 17, 24%), IMNM (n = 9, 14.6%), and polymyositis (PM) (n = 3, 4.7%). No inclusion body myositis (IBM) cases were identified. Conclusions: Incidence and prevalence of IIM align with prior reports. We observed an increase in IIM incidence and a shift in subtype distribution, with ASyS and IMNM becoming more frequent. These findings have clinical relevance, as each IIM subtype carries distinct prognostic and therapeutic implications. Full article

A core challenge in multimodal emotion recognition lies in the precise capture of the inherent multimodal interactive nature of human emotions. Addressing the limitation of existing methods, which often process visual signals (facial expressions) and physiological signals (EEG, ECG, EOG, and GSR) in isolation and thus fail to exploit their complementary strengths effectively, this paper presents a new multimodal emotion recognition framework called the Gated Biological Visual Network (GBV-Net). This framework enhances emotion recognition accuracy through deep synergistic fusion of facial expressions and physiological signals. GBV-Net integrates three core modules: (1) a facial feature extractor based on a modified ConvNeXt V2 architecture incorporating lightweight Transformers, specifically designed to capture subtle spatio-temporal dynamics in facial expressions; (2) a hybrid physiological feature extractor combining 1D convolutions, Temporal Convolutional Networks (TCNs), and convolutional self-attention mechanisms, adept at modeling local patterns and long-range temporal dependencies in physiological signals; and (3) an enhanced gated attention fusion module capable of adaptively learning inter-modal weights to achieve dynamic, synergistic integration at the feature level. A thorough investigation of the publicly accessible DEAP and MAHNOB-HCI datasets reveals that GBV-Net surpasses contemporary methods. Specifically, on the DEAP dataset, the model attained classification accuracies of 95.10% for Valence and 95.65% for Arousal, with F1-scores of 95.52% and 96.35%, respectively. On MAHNOB-HCI, the accuracies achieved were 97.28% for Valence and 97.73% for Arousal, with F1-scores of 97.50% and 97.74%, respectively. These experimental findings substantiate that GBV-Net effectively captures deep-level interactive information between multimodal signals, thereby improving emotion recognition accuracy. Full article

The zygomaticus major and minor (ZMa/ZMi) are key determinants of smile dynamics and midface contour, yet they exhibit substantial morphological variability—including bifid or multibellied bellies, accessory slips, and atypical insertions. Such variants can alter force vectors, fat-compartment boundaries, and SMAS planes, increasing the risk of asymmetry, contour irregularities, or “joker smile” following facelifts, fillers, thread lifts, and smile reconstruction. To our knowledge, this is the first review to integrate the Landfald classification of ZMa/ZMi variants with a standardized dynamic imaging-based workflow for aesthetic and reconstructive midface procedures. We conducted a narrative literature synthesis of anatomical and imaging studies. Bifid or multibellied variants have been reported in up to 35% of cadaveric specimens. We synthesize anatomical, biomechanical, and imaging evidence (MRI, dynamic US, 3D analysis) to propose a practical protocol: (1) focused history and dynamic examination, (2) US/EMG mapping of contraction vectors, (3) optional high-resolution MRI for complex cases, and (4) individualized adjustment of surgical vectors, injection planes, and dosing. Procedure-specific adaptations are outlined for deep-plane releases, thread-lift trajectories, filler depth selection, and muscle-transfer orientation. We emphasize that standardizing preoperative dynamic mapping and adopting a “patient-specific mimetic profile” can enhance safety, predictability, and preservation of authentic expression, ultimately improving patient satisfaction across diverse midface interventions. Full article

IgA nephropathy (IgAN) is the most prevalent type of primary glomerulonephritis, with heterogeneous clinical outcomes. Conventional prognostic factors, such as proteinuria, eGFR, and Oxford histologic classification, have poor sensitivity and specificity. Recently, transcriptomic profiling has been employed to provide insights into the molecular definition of IgAN and facilitate patient stratification in those at risk of disease progression. In this review, we summarize our current understanding of IgAN derived from bulk RNA sequencing, single-cell transcriptomics, spatial transcriptomics, and gene expression profiling to elucidate the molecular characteristics of IgAN. Bulk transcriptomics of glomerular and tubulointerstitial compartments highlighted consistently upregulated genes (e.g., CCL2, CXCL10, LCN2, HAVCR1, COL1A1) and altered pathways (e.g., NF-κB, TGF-β, JAK/STAT, and complement) that are associated with clinical decline. Single-cell and single-nucleus RNA-sequencing has also identified the value of pathogenic cell types and regulatory networks in mesangial cells, tubular epithelium, and immune infiltrates. Furthermore, noninvasive transcriptomic signatures developed from urine and blood may represent useful real-time surrogates of tissue activity. With the advent of integrated analyses and machine learning approaches, personalized risk models that outperform traditional metrics are now available. While challenges remain, particularly related to standardization, cohort size, and clinical deployment, transcriptomics is likely to revolutionize IgAN by providing early risk predictions and precision therapeutics. Unlike prior reviews, our work provides an integrative synthesis across bulk, single-cell, spatial, and noninvasive transcriptomics, linking molecular signatures directly to clinical translation in risk stratification and precision therapeutics. Full article

Understanding spatial variability is central to precision agriculture, yet terrain features are often overlooked in remote sensing workflows that inform agronomic decision-making. This work introduces a terrain-aware visual analytics approach that integrates landform classification with crop performance analysis to better support field-level decisions. Terrain features are an important contributor to yield variability, alongside environmental conditions, soil properties, and management practices. However, they are rarely integrated systematically into performance analysis and decision-making workflows—limiting the potential for terrain-aware insights in precision agriculture. Addressing this gap requires approaches that incorporate terrain attributes and landform classifications into agricultural performance analysis and management zone (MZ) delineation—ideally through visual analytics that offer interpretable insights beyond the constraints of purely data-driven methods. We introduce an interactive focus+context visualization tool that integrates multiple data layers—including terrain features, vegetation index–based performance metric, and management zones—into a unified, expressive view. The system leverages freely available remote sensing imagery and terrain data derived from Digital Elevation Models (DEMs) to evaluate crop performance and landform characteristics in support of agronomic analysis. The tool was applied to eleven agricultural fields across the Canadian Prairies under diverse environmental conditions. Fields were segmented into depressions, hilltops, and baseline areas, and crop performance was evaluated across these landform groups using the system’s interactive visualization and analytics. Depressions and hilltops consistently showed lower mean performance and higher variability (measured by coefficient of variation) compared to baseline regions, which covered an average of 82% of each field. We also subdivided baseline areas using slope and the Sediment Transport Index (STI) to investigate soil erosion effects, but field-level patterns were inconsistent and no systematic differences emerged across all sites. Expert evaluation confirmed the tool’s usability and its value for field-level decision support. Overall, the method enhances terrain-aware interpretation of remotely sensed data and contributes meaningfully to refining management zone delineation in precision agriculture. Full article

Classifying brain tumour transcriptomic data is crucial for precision medicine but remains challenging due to high dimensionality and limited interpretability of conventional models. This study benchmarks three image-based deep learning approaches, DeepInsight, Fotomics, and a novel saliency-guided convolutional neural network (CNN), for transcriptomic classification. DeepInsight utilises dimensionality reduction to spatially arrange gene features, while Fotomics applies Fourier transforms to encode expression patterns into structured images. The proposed method transforms each single-cell gene expression profile into an RGB image using PCA, UMAP, or t-SNE, enabling CNNs such as ResNet to learn spatially organised molecular features. Gradient-based saliency maps are employed to highlight gene regions most influential in model predictions. Evaluation is conducted on two biologically and technologically different datasets: single-cell RNA-seq from glioblastoma GSM3828672 and bulk microarray data from medulloblastoma GSE85217. Outcomes demonstrate that image-based deep learning methods, particularly those incorporating saliency guidance, provide a robust and interpretable framework for uncovering biologically meaningful patterns in complex high-dimensional omics data. For instance, ResNet-18 achieved the highest accuracy of 97.25% on the GSE85217 dataset and 91.02% on GSM3828672, respectively, outperforming other baseline models across multiple metrics. Full article

Rosa hybrida (R. hybrida), a widely cultivated ornamental species, is increasingly threatened by climate-induced abiotic stresses, including heat and drought. Heat shock transcription factors (HSFs) are critical for plant stress responses, yet their roles in R. hybrida remain understudied. In this research, 71 HSF genes were identified from the haplotype-resolved genome of the tetraploid variety ‘Samantha’. These genes were classified into three major classes (HSFA, HSFB, HSFC) and 15 subgroups based on phylogenetic and motif analysis. Gene structure and conserved motifs revealed subgroup-specific functional divergence. Promoter analysis identified abundant hormone- and stress-responsive cis-elements, particularly for abscisic acid (ABA) and jasmonic acid. Synteny analysis suggested that segmental duplication contributed to the RhHSF family’s expansion. Tissue-specific expression profiling revealed distinct roles for HSFs, with HSFB genes predominantly expressed in reproductive tissues and HSFA genes in vegetative organs. Expression under heat and drought stress showed dynamic, subgroup-dependent responses, with HSFC members playing significant roles. Functional assays demonstrated that RhHSF17, induced by both stresses and ABA, localized to the nucleus, and its overexpression in Arabidopsis enhanced drought tolerance. This study provides a comprehensive characterization of the RhHSF gene family, offering insights into their roles in stress tolerance and laying the foundation for future functional research. Full article

Background/Objectives: CDH1 gene is widely studied, as pathogenic variants are involved in diffuse gastric cancers and lobular breast cancers. CDH1 genotype contributes to the management of clinical practice recommendations for cancer prevention. We proposed a qualitative and quantitative description of CDH1 alternative splicing profile on lymphoblastoid cell lines (LCLs). The aim of this description was to allow a comprehensive interpretation of the effect of variants of uncertain significance (VUS) on CDH1 splicing. Methods: We studied, using RNAseq, the splicing profile of 22 LCLs (untreated and treated with puromycin) with no pathogenic variant on CDH1 and evaluated the effect on CDH1 splicing of four VUS. Results: We highlighted a total of eleven alternative splicing events including four junctions starting from intron 2, defining novel isoforms of CDH1. We also identified an isoform causing the skip of exon 11 and leading to a disruption of the reading frame with high levels of expression on negative CDH1 control LCLs, confirmed by ddPCR. Splicing RNAseq results for CDH1 VUS: c.1008+1G>A and c.1936+5G>A showed complex splicing patterns but allowed their classification as pathogenic. We studied CDH1 VUS exon 4 to exon 11 duplication with RNA analysis combined with Bionano optical genome mapping. Depending on alternative splicing of proximal and distal exons 11 within the duplication, we identified four distinct transcripts, leading to truncated proteins, classifying the duplication as pathogenic. Conclusions:CDH1 has a complex alternative splicing profile characterized by a dynamic splicing of intron 2 making CDH1 a good candidate for a study using long-read RNAseq. Full article

B-cell acute lymphoblastic leukemia (B-ALL) is characterized by the abnormal proliferation of B-lineage lymphocytes in the bone marrow (BM). The roles of immune cells within the BM microenvironment remain incompletely understood. Single-cell RNA sequencing (scRNA-seq) provides the potential for groundbreaking insights into the pathogenesis of B-ALL. In this study, scRNA-seq was conducted on BM samples from 17 B-ALL patients (B-ALL cohorts) and 13 healthy controls (HCs). Bioinformatics analyses, including clustering, differential expression, pathway analysis, and gene set variation analysis, systematically identified immune cell types and assessed T-cell prognostic and metabolic heterogeneity. A metabolic-feature-based machine learning model was developed for B-ALL subtyping. Furthermore, T-cell–monocyte interactions, transcription factor (TF) activity, and drug enrichment analyses were performed to identify therapeutic targets. The results indicated significant increases in Pro-B cells, alongside decreases in B cells, NK cells, monocytes, and plasmacytoid dendritic cells (pDCs) among B-ALL patients, suggesting immune dysfunction. Clinical prognosis correlated significantly with the distribution of T-cell subsets. Metabolic heterogeneity categorized patients into four distinct groups (A–D), all exhibiting enhanced major histocompatibility class I (MHC-I)-mediated intercellular communication. The metabolic-based machine learning model achieved precise classification of B-ALL groups. Analysis of TF activity underscored the critical roles of MYC, STAT3, and TCF7 within the B-ALL immunometabolic network. Drug targeting studies revealed that dorlimomab aritox and palbociclib specifically target dysregulation in ribosomal and CDK4/6 pathways, offering novel therapeutic avenues. This study elucidates immunometabolic dysregulation in B-ALL, characterized by altered cellular composition, metabolic disturbances, and abnormal cellular interactions. Key TFs were identified, and targeted drug profiles were established, demonstrating the significant clinical potential of integrating immunological mechanisms with metabolic regulation for the treatment of B-ALL. Full article