Search Results (838)

To address the low efficiency and high subjectivity of manual interpretation in fluorescence in situ hybridization (FISH) tissue and cell images, this study proposes an intelligent FISH image classification model based on an improved ResNet50 architecture. By analyzing the characteristics of multi-channel fluorescence signals and the bottlenecks of clinical interpretation, a Convolutional Block Attention Module (CBAM) is introduced to enhance the representation of salient fluorescence features through dual channel–spatial attention mechanisms. A Pyramid Pooling Module (PPM) is integrated to fuse multi-scale contextual information, improving the detection accuracy of small targets such as microdeletions. Furthermore, the shortcut connections in residual blocks are optimized to reduce feature loss. To mitigate the limitation of insufficient annotated samples, transfer learning is employed, combined with a focal loss function to enhance classification performance under class-imbalanced conditions. Experiments conducted on a clinical dataset of 12,000 FISH images demonstrate that the proposed model achieves an overall classification accuracy of 92.4%, representing a 9.9% improvement over the original ResNet50. The recall rate for complex categories (e.g., translocation and fusion) exceeds 90.7%, with an inference time of 22.3 ms per sample, meeting the real-time requirements of clinical diagnosis. These results provide an efficient and practical solution for the automated intelligent interpretation of FISH images, offering significant potential for precision-assisted diagnosis of tumors and genetic disorders. Full article

Sleep spindles, together with K-complexes, are the distinctive patterns of neuronal activity in EEG recordings during stage 2 sleep. When the mechanisms of sleep spindle generation are impaired, e.g., in epilepsy, their quantitative parameters change. The analysis of these changes can provide valuable insights into the formation of epileptiform activity patterns and help to develop an additional tool for more accurate medical diagnosis. Despite the central role of EEG in the diagnosis of epilepsy, disorders of consciousness, and neurological research, resources specifically dedicated to large-scale EEG data analysis are under-represented. In our study, we collect a specialized database of clinical EEG recordings from epilepsy patients and controls during N2 sleep, characterized by rhythmic spindle activity in frontocentral and vertex regions, and manually annotate them. We then quantify four key sleep spindle characteristics using a comparison of manual annotation by a clinician and artificial intelligence technologies. A thorough evaluation of state-of-the-art deep learning architectures for detecting and characterizing sleep spindles in EEG recordings from epilepsy patients is conducted. The results show that the 1D U-Net and SEED architectures achieve competitive overall performance, but their precision-to-recall ratios differ markedly in clinical settings. This suggests that different approaches may be appropriate for each clinical situation. Furthermore, our results demonstrate that epilepsy is associated with significant and quantifiable changes in sleep spindle morphology and frequency. Automated analysis of these characteristics using artificial intelligence provides a reliable biomarker that provides a detailed picture of thalamocortical dysfunction in epilepsy. This approach has great potential for accelerated diagnosis and the development of targeted therapeutic strategies for epilepsy. Full article

Lung cancer remains the leading cause of cancer-related deaths worldwide, with over 2.4 million new diagnoses in 2022. Early diagnosis remains challenging due to the non-specificity of symptoms, often resulting in late-stage detection. Although 2-D and 3-D medical imaging, particularly computed tomography (CT), is widely used for detecting lung cancer, it is associated with manual segmentation, which remains time-consuming and user-dependent. This study proposes LuCa as an innovative 2.5-D deep learning model for lung cancer delineation, which combines the benefits of 2-D segmentation with 3-D volume delineation. The main novelty of LuCa is focused on its pipeline, specifically designed to be of clinical use, in order to guarantee the usability of the method. LuCa employs a U-Net architecture for segmentation, followed by a post-image-processing step for 3-D tumor volume delineation and false-positive correction. The method was trained and evaluated using the “NSCLC-Radiomics” database, comprising CT images of 422 non-small cell lung cancer patients, with clinical manual tumor annotations as ground truth. The model achieved strong performance, with high dice coefficients (87 ± 12%), intersection over union (81 ± 17%), sensitivity (84 ± 16%), and positive predictive value (94 ± 10%) on the test set. Performance was particularly high for larger tumors, reflecting the ability of the model to delineate more visible lesions accurately. Statistical analysis confirmed the high correlation and minimal error between predicted and ground truth tumor volumes. The results highlight the potential of the 2.5-D approach to improve clinical efficiency by enabling accurate tumor segmentation with reduced computational cost, compared to traditional 3-D methods. Future research will focus on assessing the use of LuCa as real-time clinical decision support, particularly for assessing tumors during treatment. Full article

Background/Objectives: Endometriosis is a chronic inflammatory condition that often requires laparoscopic examination for definitive diagnosis. Automated analysis of laparoscopic images using Deep Learning (DL) may support clinicians by improving diagnostic consistency and efficiency. This study aimed to develop and evaluate explainable DL models for the binary classification of endometriosis using laparoscopic images from the publicly available GLENDA (Gynecologic Laparoscopic ENdometriosis DAtaset). Methods: Four representative architectures—ResNet50, EfficientNet-B2, EdgeNeXt_Small, and Vision Transformer (ViT-Small/16)—were systematically compared under class-imbalanced conditions using five-fold cross-validation. To enhance interpretability, Gradient-weighted Class Activation Mapping (Grad-CAM) and SHapley Additive exPlanations (SHAP) were applied for visual explanation, and their quantitative alignment with expert-annotated lesion masks was assessed using Intersection over Union (IoU), Dice coefficient, and Recall. Results: Among the evaluated models, EdgeNeXt_Small achieved the best trade-off between classification performance and computational efficiency. Grad-CAM produced spatially coherent visualizations that corresponded well with clinically relevant lesion regions. Conclusions: The study shows that lightweight convolutional neural network (CNN)–Transformer architectures, combined with quantitative explainability assessment, can identify endometriosis in laparoscopic images with reasonable accuracy and interpretability. These findings indicate that explainable AI methods may help improve diagnostic consistency by offering transparent visual cues that align with clinically relevant regions. Further validation in broader clinical settings is warranted to confirm their practical utility. Full article

Hereditary breast and ovarian cancer (HBOC) syndrome accounts for 5–10% of all breast and ovarian cancers, with BRCA1 and BRCA2 pathogenic variants being the most common genetic alterations. However, additional genes such as BARD1, whose protein product interacts with BRCA1 via its N-terminal RING domain, have been implicated as low-penetrance contributors to cancer risk. This study aimed to investigate the frequency and distribution of the BARD1 variant c.1518_1519delinsCA (p.Val507Met) in a cohort of 920 patients undergoing genetic testing for hereditary cancer predisposition. Next Generation Sequencing (NGS) was performed using a 28-gene panel, and allelic frequencies of BARD1 were analyzed. Among 920 patients, 159 (17.28%) were pure heterozygous for the c.1518_1519delinsCA variant. Notably, c.1519G>A was never observed without c.1518T>C, suggesting a strong linkage between the two variants. The allele frequencies observed (34.51% for A at c.1519 and 77.88% for C at c.1518) challenge current reference genome expectations. Data from the ALFA database confirmed that these frequencies are consistent with population-level variation, not sample bias. Our findings raise the hypothesis that the reference allele at position c.1518 may not reflect the true wild-type sequence. While both c.1518T>C and c.1519G>A are individually classified as benign, their combined occurrence as a dinucleotide substitution (c.1518_1519delinsCA) warrants further investigation. These results underscore the importance of accurate variant annotation and population-specific frequency data for clinical interpretation of NGS findings. Although BARD1 remains a low-frequency contributor to HBOC compared to BRCA1/2, its inclusion in multigene panels is supported by the potential relevance of such complex variants. Full article

Background/Objectives: Bacteroides species are key members of the human gut microbiota but can act as opportunistic pathogens. This study investigated the genomic features of clinical Bacteroides isolates from southern Thailand. Methods: Sixteen isolates were collected from body fluids, tissues, and pus at Songklanagarind Hospital (2022–2024). Whole-genome sequencing was performed on the BGI platform, followed by genome assembly, annotation, average nucleotide identity (ANI), pairwise single-nucleotide polymorphism (SNP) analysis, antimicrobial resistance (AMR) gene profiling, plasmid prediction, virulence screening, and phylogenetic analysis. Results: ANI and SNP analysis revealed two clusters: one comprising B. ovatus, B. intestinigallinarum, and B. thetaiotaomicron, and another mainly B. fragilis with one B. hominis isolate. All isolates were resistant to ampicillin, cephalothin, and penicillin; six B. fragilis strains were resistant to all tested antibiotics. The β-lactamase gene cepA was detected in all B. fragilis isolates, and plasmids were predicted in two genomes. Three virulence types (capsule formation, lipopolysaccharide modification, and stress response) were identified. Phylogenomic analysis confirmed species-level assignments and revealed underrecognized lineages, emphasizing the value of genome-based approaches for accurate classification. Conclusions: Clinical Bacteroides isolates display diverse resistance and virulence profiles, highlighting the importance of strain-level genomic surveillance. Full article

Background: Gait event detection from inertial sensors offers scalable insights into locomotor health, with applications in clinical monitoring and mobile health. However, supervised methods are limited by scarce annotations, device variability, and sensor placement shifts. This in silico study evaluates self-supervised learning (SSL) as a resource-efficient strategy to improve robustness and generalizability. Methods: Six public smartphone and wearable inertial measurements unit (IMU) datasets (WISDM, PAMAP2, KU-HAR, mHealth, OPPORTUNITY, RWHAR) were harmonized within a unified deep learning pipeline. Models were pretrained on unlabeled windows using contrastive SSL with sensor-aware augmentations, then fine-tuned with varying label fractions. Experiments systematically assessed included (1) pretraining scale, (2) label efficiency, (3) augmentation contributions, (4) device/placement shifts, (5) sampling-rate sensitivity, and (6) backbone comparisons (CNN, TCN, BiLSTM, Transformer). Results: SSL consistently outperformed supervised baselines. Pretraining yielded accuracy gains of ΔF1 +0.08–0.15 and reduced stride-time error by −8 to −12 ms. SSL cut label needs by up to 95%, achieving competitive performance with only 5–10% labeled data. Sensor-aware augmentations, particularly axis-swap and drift, drove the strongest transfer gains. Robustness was maintained across sampling rates (25–100 Hz) and device/placement shifts. CNNs and TCNs offered the best efficiency–accuracy trade-offs, while Transformers delivered the highest accuracy at greater cost. Conclusions: This computational analysis across six datasets shows SSL enhances gait event detection with improved accuracy, efficiency, and robustness under minimal supervision, establishing a scalable framework for human performance and sports medicine in clinical and mobile health applications. Full article

Age-related macular degeneration (AMD) is a leading cause of irreversible blindness, yet current tools for forecasting treatment outcomes remain limited by either the opacity of deep learning or the rigidity of rule-based systems. To address this gap, we propose a hybrid neuro-symbolic and large language model (LLM) framework that combines mechanistic disease knowledge with multimodal ophthalmic data for explainable AMD treatment prognosis. In a pilot cohort of ten surgically managed AMD patients (six men, four women; mean age 67.8 ± 6.3 years), we collected 30 structured clinical documents and 100 paired imaging series (optical coherence tomography, fundus fluorescein angiography, scanning laser ophthalmoscopy, and ocular/superficial B-scan ultrasonography). Texts were semantically annotated and mapped to standardized ontologies, while images underwent rigorous DICOM-based quality control, lesion segmentation, and quantitative biomarker extraction. A domain-specific ophthalmic knowledge graph encoded causal disease and treatment relationships, enabling neuro-symbolic reasoning to constrain and guide neural feature learning. An LLM fine-tuned on ophthalmology literature and electronic health records ingested structured biomarkers and longitudinal clinical narratives through multimodal clinical-profile prompts, producing natural-language risk explanations with explicit evidence citations. On an independent test set, the hybrid model achieved AUROC 0.94 ± 0.03, AUPRC 0.92 ± 0.04, and a Brier score of 0.07, significantly outperforming purely neural and classical Cox regression baselines (p ≤ 0.01). Explainability metrics showed that >85% of predictions were supported by high-confidence knowledge-graph rules, and >90% of generated narratives accurately cited key biomarkers. A detailed case study demonstrated real-time, individualized risk stratification—for example, predicting an >70% probability of requiring three or more anti-VEGF injections within 12 months and a ~45% risk of chronic macular edema if therapy lapsed—with predictions matching the observed clinical course. These results highlight the framework’s ability to integrate multimodal evidence, provide transparent causal reasoning, and support personalized treatment planning. While limited by single-center scope and short-term follow-up, this work establishes a scalable, privacy-aware, and regulator-ready template for explainable, next-generation decision support in AMD management, with potential for expansion to larger, device-diverse cohorts and other complex retinal diseases. Full article

This study examines the genomic composition and resistance potential of eight putative plasmid-derived contig assemblies reconstructed from marine Enterobacterales isolated in the central Adriatic Sea. Using a combination of Illumina-based whole genome sequencing, de novo assembly, and a multi-tool bioinformatics pipeline, we annotated antimicrobial resistance genes (ARGs), insertion sequences (ISs), and plasmid replicon types. Clinically significant resistance markers such as bla_KPC, bla_TEM, aacA4, tetA, and folP were identified, frequently co-localised with mobile genetic elements including IS110, IS4, and IS1182. The plasmid-associated contigs were assigned to MOBP and MOBQ types and contained replicon markers (IncP6, IncA/C2) characteristic of broad-host-range plasmids. Our findings provide valuable insight into the plasmidome of environmental Enterobacterales, emphasising the role of coastal pollution in shaping the distribution and potential mobility of antimicrobial resistance genes. This supports the One Health framework by linking environmental reservoirs to clinically relevant resistance mechanisms. Full article

Background: Recurrent cervical cancer is one of the most defining threats to patient longevity, underscoring the need for prognostic models to identify high-risk patients. Objectives: The aim of the study is to integrate clinical data with the GSE44001 Dataset to identify key risk factors associated with the recurrence of cervical cancer. Patients are stratified into high-, moderate-, and low-risk groups using selected clinical and molecular features. Identifying a long non-coding RNA (lncRNA) gene signature associated with recurrent cervical cancer. Methods: From the total data collected, 138 recurrent cervical cancer patients were identified. GSE44001 Dataset is downloaded from the NCBI GEO Database. When using the GENCODE Annotation tool, the long non-coding RNA is filtered. The dataset is then linked with filtered long non-coding RNA. The Least Absolute Shrinkage Selection Operator (LASSO) is employed to find attributes in gene expression analysis. Risk factors of recurrent cervical cancer are identified. Risk value is assigned to each individual based on the selected lncRNAs and the corresponding overfitting coefficients. Result: The RNN Long Short-Term Memory model demonstrates a prognostic value, where high-risk patients experience a shorter duration of recurrence-free survival (p < 0.05). Individuals with a recurrence of cervical carcinoma, a progressive disease, were associated with the ATXN8OS marker, the C5orf60 indicator, and the INE1 index gene. In contrast, patients diagnosed at earlier stages are aligned with the KCNQ1DN marker, LOH12CR2 gauge, RFPL1S value, and KCNQ1OT1 indicator. Patients in moderate stages were primarily associated with the EMX2OS score. Conclusions: The research findings demonstrate that the nine-lncRNA signature, when combined with deep learning, offers a powerful approach for recurrence risk stratification in cervical cancer. Full article

Cancer progression is primarily driven by disruptions in critical biological pathways, including ErbB signaling, p53-mediated apoptosis, and GSK3 signaling. However, experimental and clinical studies typically identify only limited disease-associated genes, challenging traditional pathway analysis methods that require larger gene sets. To overcome this limitation, reliably expanded gene sets are required to align with cancer-related pathways. Although various propagation methods are available, the key challenge is to select techniques that can effectively propagate signals from limited seed gene sets through protein interaction networks, thereby generating robust, expanded sets capable of revealing pathway disruptions in cancer. In this study, the number of seed genes was systematically varied to evaluate the alignment of pathways obtained from different propagation methods with known pathways using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations. Among the evaluated propagation methods, normalized Laplacian diffusion (NLD) demonstrated the strongest alignment with reference pathways, with an average area under the ROC curve (AUC) of 95.11% and an area under precision–recall (AUPR) of 71.20%. Focusing specifically on well-established cancer pathways, we summarized the enriched pathways and discussed their biological relevance with limited gene input. Results from multiple runs were aggregated to identify genes consistently prioritized but absent from core pathway annotations, representing potential pathway extensions. Notable examples include RAC2 (ErbB pathway), FOXO3 and ESR1 (GSK3 signaling), and XIAP and BRD4 (p53 pathway), which were significantly associated with patient survival. Literature validation confirmed their biological relevance, underscoring their potential as prognostic markers and therapeutic targets. In summary, NLD-based diffusion proves effective for pathway discovery from limited input, extending beyond annotated members to reveal clinically relevant genes with therapeutic and biomarker potential. Full article

Background/Aim: Furcation lesions in primary molars are critical in pediatric dentistry, often guiding treatment decisions between root canal therapy and extraction. This study introduces a deep learning-based clinical decision-support system that directly maps radiographic lesion characteristics to corresponding treatment recommendations—a novel contribution in the context of pediatric dental imaging, also represents the first integration of panoramic radiographic classification of primary molar furcation lesions with treatment planning in pediatric dentistry. Materials and Methods: A total of 387 anonymized panoramic radiographs from children aged 3–13 was labeled into five distinct bone lesion categories. Three object detection models (YOLOv12x, RT-DETR-L, and RT-DETR-X) were trained and evaluated using stratified train-validation-test splits. Diagnostic performance was assessed using precision, recall, mAP@0.5, and mAP@0.5–0.95. Additionally, qualitative accuracy was evaluated with expert-annotated samples. Results: Among the models, RT-DETR-X achieved the highest performance (mAP@0.5 = 0.434), representing modest but clinically promising diagnostic capability, despite the limitations of a relatively small, single-center dataset. Specifically, RT-DETR-X achieved the highest diagnostic accuracy (mAP@0.5 = 0.434, Recall = 0.483, Precision = 0.440), followed by YOLOv12x (mAP@0.5 = 0.397, Precision = 0.442) and RT-DETR-L (mAP@0.5 = 0.326). All models successfully identified lesion types and supported corresponding clinical decisions. The system reduced diagnostic ambiguity and showed promise in supporting clinicians with varying levels of experience. Conclusions: The proposed models have potential for standardizing diagnostic outcomes, especially in resource-limited settings and mobile clinical environments. Full article

Background: Large Language Models (LLMs), such as ChatGPT, are increasingly applied in medicine for summarization, clinical decision support, and diagnostic assistance, including recent work in dermatology. Previous AI and NLP models in dermatology have mainly focused on lesion classification, diagnostic support, and patient education, while extracting structured treatment information from unstructured dermatology records remains underexplored. We evaluated ChatGPT-4o’s ability to identify psoriasis treatments from free-text documentation, compared with expert annotations. Methods: In total, 94 electronic medical records (EMRs) of patients diagnosed with psoriasis were analyzed. ChatGPT-4o extracted treatments used for psoriasis from each unstructured clinical note. Its output was compared to manually curated reference annotations by expert dermatologists. A total of 83 treatments, including topical agents, systemic medications, biologics, phototherapy, and procedural interventions, were evaluated. Performance metrics included recall, precision, F1-score, specificity, accuracy, Cohen’s Kappa, and Area Under the Curve (AUC). Analyses were conducted at the individual-treatment level and grouped into pharmacologic categories. Results: ChatGPT-4o demonstrated strong performance, with recall of 0.91, precision of 0.96, F1-score of 0.94, specificity of 0.99, and accuracy of 0.99. Agreement with expert annotations was high (Cohen’s Kappa = 0.93; AUC = 0.98). Group-level analysis confirmed these results, with the highest performance in biologics and methotrexate (F1 = 1.00) and lower recall in categories with vague documentation, such as systemic corticosteroids and antihistamines. Conclusions: Our study highlights the potential of LLMs to extract psoriasis treatment information from unstructured clinical documentation and structure it for research and decision support. The model performed best with well-defined, commonly used treatments. Full article

Background/Objectives: Hallux valgus (HV), a common foot deformity, is difficult to quantify beyond traditional angular measurements. This study introduces a novel dynamic distance mapping technique to visualize HV progression and identify spatial features linked to severity. Methods: A retrospective analysis of 335 feet from 178 patients undergoing HV surgery at Hasharon Hospital, Israel (2014–2024), utilized custom Python software to annotate 24 landmarks on preoperative standing anteroposterior radiographs. This generated 276 normalized Euclidean distances, analyzed via Pearson correlation against HV angles (HVA, IMA, DMAA, HIA). Results: Seven distances correlated negatively (r > 0.4, p < 0.05) and seven positively with HVA, involving the distal phalanx, sesamoids, and second metatarsal. Eleven distances showed strong positive correlation (r > 0.4, p < 0.05) with IMA, reflecting displacement patterns. Moderate correlations were observed with DMAA (six negative, r −0.3 to −0.4; two positive, r 0.3 to 0.4, p < 0.05) and HIA (two negative, r −0.3 to −0.4, p < 0.05). Visualizations highlighted progressive spatial changes. Conclusions: Dynamic distance mapping provides valuable insights into hallux valgus (HV) progression, as evidenced by significant correlations with HVA and IMA, supporting its potential role in surgical planning. However, its ability to capture 3D deformities requires validation against weightbearing computed tomography (WBCT). Future research should explore correlations with specific indications for corrective osteotomies to enhance clinical applicability. Full article

Antimicrobial resistance (AMR) is a major public health threat that urgently requires alternative strategies to address this challenge. Klebsiella spp. are among the most important clinical pathogens and a leading cause of opportunistic nosocomial infections, with high morbidity and mortality associated with strains resistant to last-line antimicrobials such as carbapenems. Bacteriophages are considered a promising therapeutic option for treating infections caused by Klebsiella strains. Hence, the aim of this work was to isolate and characterize a phage capable of infecting carbapenem-resistant Klebsiella strains. The phage KpCCP1 was isolated using the double layer agar method (DLA), from the influent of a wastewater treatment plant, which was characterized through phenotypic and genomic analyses. Morphological characteristics were determined using TEM, and its host range was evaluated against a collection of 133 Klebsiella strains. Its whole genome was sequenced using the Illumina NovaSeq X Plus platform and then assembled and annotated. VICTOR was used for phylogenetic analysis of the isolated phage, and VIRIDIC to compare its genome with those of its closest relatives. KpCCP1 is a tailed dsDNA lytic phage with a genome size of 177,276 bp and a GC content of 41.82%. It encodes 292 ORFs, including two tRNA genes. Phage KpCCP1 is a member of the Slopekvirus genus in the Straboviridae family. It is capable of infecting 22 carbapenem-resistant Klebsiella strains, including K. pneumoniae and K. michiganensis. Notably, it does not contain virulence or antibiotic resistance genes and harbors putative anti-CRISPR genes, therefore representing a promising candidate for phage therapy against clinically critical Klebsiella strains. Full article