Curr Microbiol. 2025 Jul 11;82(9):379. doi: 10.1007/s00284-025-04346-3.

ABSTRACT

Resistance to antimicrobial agents has become a global threat, estimated to cause 10-million deaths annually by 2050. Antimicrobial peptides are emerging as an alternative and offer advantages over traditional antibiotics. Antimicrobial peptides generated by artificial intelligence (AI) strategies are potential alternatives that reduce costs and development time. This work optimized a set of peptides generated by two deep learning algorithms. The modifications made to the peptides were evaluated with bioinformatic and other AI tools as predictors of antimicrobial activity, hemolytic capacity, and toxicity. As a result, 26 synthetic peptides generated in silico were obtained with a high probability of being antimicrobial and biologically safe. Finally, 12 peptides were synthesized to perform in vitro tests against four bacterial species, Candida albicans, and cancer cells. Results indicate that 9 of the peptides have a MIC below 10 μM, and some have an inhibitory concentration at 2 μM, such as OrP1M for Escherichia coli, OrP9M for Pseudomonas aeruginosa, and VeP1 for Staphylococcus aureus. In addition, six peptides have activity against the breast cancer cell line (MCF-7), and peptide OrP1M had an IC50 of < 6.25 μM. It is concluded that the synthetic-generated peptides have high antimicrobial activity, but in most cases, their MICs were improved after the modifications were made.

PMID:40643674 | DOI:10.1007/s00284-025-04346-3

Int J Surg. 2025 Jul 9. doi: 10.1097/JS9.0000000000002840. Online ahead of print.

NO ABSTRACT

PMID:40643594 | DOI:10.1097/JS9.0000000000002840

ACS Appl Mater Interfaces. 2025 Jul 11. doi: 10.1021/acsami.5c08242. Online ahead of print.

ABSTRACT

Machine learning has emerged as a powerful tool for the inverse design of various device structures including metamaterials and multilayer coatings. This study presents an inverse design approach for transparent wave absorbers based on multiple ionic liquids, employing a mixture density network (MDN) architecture. The model focuses on the perfect absorption bandwidth as the design objective, treating both the ionic liquid type and the layer-specific structural parameters as design variables. It enables rapid prediction of design variables that meet specified conditions, offering multiple viable structural configurations, even as design goals change. Compared with other inverse design methods, this approach is highly practical and provides a broad range of solutions, facilitating the identification of the optimal configuration. Using this model, we designed a transparent broadband wave absorber with an absorption bandwidth of 4.18 to 34.9 GHz, an average transmittance of 76.5%, and a thickness of only 8.725 mm, achieving a high bandwidth while maintaining high transparency. This absorber is well-suited for applications in fighter cockpit shields and microwave anechoic chambers.

PMID:40643426 | DOI:10.1021/acsami.5c08242

FEMS Microbiol Lett. 2025 Jul 11:fnaf072. doi: 10.1093/femsle/fnaf072. Online ahead of print.

ABSTRACT

The GH11 xylanase XynCDBFV, derived from Neocallimastix patriciarum, is widely used in various industries. However, its relatively low thermostability limits its potential. In this study, two computational approaches-Rosetta Cartesian_ddG and the deep learning-based tool ​Pythia-were employed to identify key residues affecting XynCDBFV thermostability. Both methods highlighted residues D57 and G201 as promising targets. Site-saturation mutagenesis at these positions yielded eighteen variants with improved thermostability. Notably, three ​D57 variants (D57N/S/T) exhibited a 10°C increase in optimal temperature and retained 3.4%-21.7% higher residual activity than the wild-type after 1-hour incubation at 80°C. Five ​G201 variants (G201A/C/F/I/V) showed 5°C/10°C enhancements in optimal temperatures, with 10.1%-22.6% improved residual activity. These findings validate ​D57 and ​G201 as pivotal sites influencing thermostability. However, combining beneficial mutations from both sites led to reduced thermostability due to ​negative epistatic interactions. Comparative analysis revealed that while Rosetta Cartesian_ddG offers broader screening, it suffers from a high false discovery rate. In contrast, Pythia provides a balanced trade-off between precision and speed. This study offers a robust framework for enzyme thermostability enhancement and underscores the value of integrating computational predictions with experimental validation in protein engineering.

PMID:40643334 | DOI:10.1093/femsle/fnaf072

ACS Appl Mater Interfaces. 2025 Jul 11. doi: 10.1021/acsami.5c08568. Online ahead of print.

ABSTRACT

In recent years, smart textiles and flexible wearable products have garnered significant attention in fields such as human-computer interaction, medical rehabilitation training, and motion monitoring. Flexible pressure sensors have attracted significant attention due to their excellent flexibility, stability, and multifunctional integration. Herein, a multifunctional wearable MXene/polydopamine (PDA)@cotton fabric pressure sensor was developed by modifying weft-knitted cotton fabric based on a dual hydrogen bond self-assembly strategy. The MXene/PDA@cotton fabric pressure sensor demonstrates wide linear detection range (0-146 kPa), high sensitivity (0.95 kPa-1), fast response/recovery times (16.434 and 11.952 ms), and outstanding stability after over 5000 cyclic tests. This sensor can achieve the monitoring of physiological parameters for human state detection, such as facial expression signals, abdominal respiratory signals, and joint bending signals. Furthermore, by integrating the MXene/PDA sensors into the interphalangeal and metacarpophalangeal joints of a cotton glove, combined with intelligent algorithms and the human-computer interaction system, static gesture recognition and dynamic sign language translation were successfully realized based on the smart glove. This work demonstrates the potential application of flexible pressure sensors in intelligent human-computer interaction, providing new insights for developing next-generation sign language recognition systems.

PMID:40643219 | DOI:10.1021/acsami.5c08568

AIDS. 2025 Aug 1;39(10):1311-1321. doi: 10.1097/QAD.0000000000004220. Epub 2025 Jul 10.

ABSTRACT

Artificial intelligence (AI) holds significant potential to transform HIV prevention and treatment through the application of advanced technologies such as machine learning (ML), deep learning (DL), and generative AI (Gen AI). These technologies can enhance the monitoring, management, and analysis of vast and complex HIV-related datasets, enabling more timely predictions of potential risks and improving HIV care strategies. AI is poised to streamline HIV prevention interventions by increasing workforce efficiency, supporting expanded accessibility and sustainability of preexposure prophylaxis (PrEP) care in nontraditional settings, and supporting clinical decision-making. Additionally, when utilized within HIV care systems, AI can help close gaps in diagnosis, treatment, and continuous care engagement. However, to optimize AI's potential in HIV prevention, careful implementation is crucial. Challenges such as reducing bias, ensuring ethical standards (including health privacy standards) are maintained, and mitigating risks like AI hallucinations must be addressed. Thoughtful integration, community consultation, and continuous evaluation will be critical to ensuring that AI plays a beneficial role in HIV prevention and drives innovations that lead to more equitable health outcomes. This editorial review explores AI's transformative potential, focusing on the US CDC's key public health strategies for HIV prevention. When aligning with public health strategies - particularly in countries supported by initiatives like President's Emergency Plan for AIDS Relief (PEPFAR) - AI can contribute significantly to global efforts to end the HIV epidemic. It offers a vision for AI's future application in HIV prevention, emphasizing the need for a holistic and syndemic approach to improving HIV prevention worldwide.

PMID:40643081 | DOI:10.1097/QAD.0000000000004220

APMIS. 2025 Jul;133(7):e70051. doi: 10.1111/apm.70051.

ABSTRACT

miR-143 and miR-145 have been reported as downregulated in colorectal cancer (CRC) compared to normal mucosa, with regulatory effects on proteins involved in carcinogenesis. These findings primarily derive from tissue homogenate analyses and experimental models. The present study employs in situ methodology to reassess miR-143 and miR-145 expression in CRC and their associations with validated protein targets within the native tumor microenvironment. Expression patterns of miR-143, miR-145, and eight previously experimentally validated target proteins were analyzed in clinical samples from 100 CRC patients using in situ hybridization, immunohistochemistry, and deep learning-based epithelial segmentation. Expression levels of miR-143 and miR-145 showed no significant difference between CRC and normal mucosa, though considerable inter-patient variability was observed. Among 11 examined miRNA-protein relationships, only four showed significant correlations, exhibiting positive associations that contrast with previously reported inverse relationships. Subgroup analyses revealed no statistically significant association between miRNA expression variability and examined clinicopathological parameters. These findings highlight the importance of in situ validation for results obtained from tissue homogenates and in vitro experiments. Additional research is warranted to determine the prognostic significance of miR-143 and miR-145 in clinical outcomes.

PMID:40642870 | DOI:10.1111/apm.70051

Chin J Cancer Res. 2025 Jun 30;37(3):303-315. doi: 10.21147/j.issn.1000-9604.2025.03.02.

ABSTRACT

OBJECTIVE: This study aims to develop a deep multiscale image learning system (DMILS) to differentiate malignant from benign thyroid follicular neoplasms on multiscale whole-slide images (WSIs) of intraoperative frozen pathological images.

METHODS: A total of 1,213 patients were divided into training and validation sets, an internal test set, a pooled external test set, and a pooled prospective test set at three centers. DMILS was constructed using a deep learning-based weakly supervised method based on multiscale WSIs at 10×, 20×, and 40× magnifications. The performance of the DMILS was compared with that of a single magnification and validated in two pathologist-unidentified subsets.

RESULTS: The DMILS yielded good performance, with areas under the receiver operating characteristic curves (AUCs) of 0.848, 0.857, 0.810, and 0.787 in the training and validation sets, internal test set, pooled external test set, and pooled prospective test set, respectively. The AUC of the DMILS was higher than that of a single magnification, with 0.788 of 10×, 0.824 of 20×, and 0.775 of 40× in the internal test set. Moreover, DMILS yielded satisfactory performance on the two pathologist-unidentified subsets. Furthermore, the most indicative region predicted by DMILS is the follicular epithelium.

CONCLUSIONS: DMILS has good performance in differentiating thyroid follicular neoplasms on multiscale WSIs of intraoperative frozen pathological images.

PMID:40642490 | PMC:PMC12240243 | DOI:10.21147/j.issn.1000-9604.2025.03.02

VideoGIE. 2025 Mar 22;10(7):358-363. doi: 10.1016/j.vgie.2025.03.027. eCollection 2025 Jul.

ABSTRACT

BACKGROUND AND AIMS: The use of artificial intelligence (AI) has been introduced in several medical fields with promising results, including endoscopy. In the field of EUS, studies using AI are still limited and have mostly focused on the identification and characterization of pancreatic masses. Recently, AI systems based on deep learning have been developed to identify anatomical landmarks during diagnostic EUS.

METHODS: The Endoangel system (Wuhan ENDOANGEL Medical Technology, Wuhan, China), built using deep convolutional neural networks (DCNNs), is able to provide navigation hints and identify anatomical landmarks in real time during diagnostic EUS. The system was trained with more than 550 EUS procedures and uses a DCNN that processes images through multiple layers by extracting features, introducing nonlinearity, reducing complexity, and making predictions via fully connected layers.

RESULTS: The AI EUS system was tested in 3 patients undergoing diagnostic EUS. In each case, the correct recognition of anatomical landmarks by the AI EUS system was judged by a single expert performing the EUS examination. The system did not recognize pathologic alterations such as pancreatic masses or cystic lesions.

CONCLUSIONS: The AI EUS DCNN-based system is able to correctly identify EUS anatomical landmarks. In the near future, this system might play an important role in EUS training and quality control. In addition, many other features might progressively be added, with the next ideal step being the identification of pathologic alterations.

PMID:40642404 | PMC:PMC12237856 | DOI:10.1016/j.vgie.2025.03.027

Front Public Health. 2025 Jun 26;13:1416273. doi: 10.3389/fpubh.2025.1416273. eCollection 2025.

ABSTRACT

BACKGROUND: Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is a global pandemic affecting millions worldwide. This study aims to bridge the knowledge gap between acute and chronic symptoms, vaccination impact, and associated factors in patients across different low- and middle-income countries (LMICs).

MATERIALS AND METHODS: The study included 2,445 participants aged 18 years and older, testing positive for COVID-19. Data collection involved screening for medical histories, testing records, symptomatology, and persistent symptoms. Validated instruments, including the DePaul Symptom Questionnaire (DSQ-2) and the Patient Health Questionnaire-9 (PHQ-9), were used. We applied a self-supervised and unsupervised deep neural network to extract features from the questionnaire. Gradient boosted machines (GBM) model was used to build a risk calculator for chronic fatigue syndrome (CFS), depression, and prolonged COVID-19 symptoms.

RESULTS: Out of the study cohort, 68.1% of the patients had symptoms lasting longer than 2 weeks. The most frequent symptoms were loss of smell (46.8%), dry cough (40.1%), loss of taste (37.8%), headaches (37.2%), and sore throat (28.9%). The patients also reported high rates of depression (47.7%), chronic fatigue (6.5%), and infection after vaccination (23.7%). Factors associated with CFS included sex, age, and smoking. Vaccinated individuals demonstrated lower odds of experiencing prolonged COVID-19 symptoms, CFS, and depression. The predictive models achieved a high area under the curve (AUC) scores of 0.87, 0.82, and 0.74, respectively.

CONCLUSION: The findings underscore the significant burden of long-term symptoms such as chronic fatigue and depression, affecting a considerable proportion of individuals post-infection. Moreover, the study reveals promising insights into the potential benefits of vaccination in mitigating the risk of prolonged COVID-19 symptoms, CFS, and depression. Overall, this research contributes valuable knowledge towards comprehensive management and prevention efforts amidst the ongoing global pandemic.

CLINICAL TRIAL REGISTRATION: Clinical trials.gov, NCT05059184.

PMID:40642241 | PMC:PMC12240947 | DOI:10.3389/fpubh.2025.1416273

Front Robot AI. 2025 Jun 26;12:1609765. doi: 10.3389/frobt.2025.1609765. eCollection 2025.

ABSTRACT

This paper presents a general framework that integrates visual and acoustic sensor data to enhance localization and mapping in complex, highly dynamic underwater environments, with a particular focus on fish farming. The pipeline enables net-relative pose estimation for Unmanned Underwater Vehicles (UUVs) and depth prediction within net pens solely from visual data by combining deep learning-based monocular depth prediction with sparse depth priors derived from a classical Fast Fourier Transform (FFT)-based method. We further introduce a method to estimate a UUV's global pose by fusing these net-relative estimates with acoustic measurements, and demonstrate how the predicted depth images can be integrated into the wavemap mapping framework to generate detailed 3D maps in real-time. Extensive evaluations on datasets collected in industrial-scale fish farms confirm that the presented framework can be used to accurately estimate a UUV's net-relative and global position in real-time, and provide 3D maps suitable for autonomous navigation and inspection.

PMID:40642204 | PMC:PMC12240768 | DOI:10.3389/frobt.2025.1609765

Front Dent Med. 2025 Jun 26;6:1618246. doi: 10.3389/fdmed.2025.1618246. eCollection 2025.

ABSTRACT

OBJECT: In forensic dentistry, dental age estimation assists experts in determining the age of victims or suspects, which is vital for legal responsibility and sentencing. The traditional Demirjian method assesses the development of seven mandibular teeth in pediatric dentistry, but it is time-consuming and relies heavily on subjective judgment.

METHODS: This study constructed a largescale panoramic dental image dataset and applied various convolutional neural network (CNN) models for automated age estimation.

RESULTS: Model performance was evaluated using loss curves, residual histograms, and normal PP plots. Age prediction models were built separately for the total, female, and male samples. The best models yielded mean absolute errors of 1.24, 1.28, and 1.15 years, respectively.

DISCUSSION: These findings confirm the effectiveness of deep learning models in dental age estimation, particularly among northern Chinese adolescents.

PMID:40642202 | PMC:PMC12241049 | DOI:10.3389/fdmed.2025.1618246

Front Dent Med. 2025 Jun 26;6:1583455. doi: 10.3389/fdmed.2025.1583455. eCollection 2025.

ABSTRACT

INTRODUCTION: Accurate assessment of midpalatal suture (MPS) maturation is critical in orthodontics, particularly for planning treatment strategies in patients with maxillary transverse deficiency (MTD). Although cone-beam computed tomography (CBCT) provides detailed imaging suitable for MPS classification, manual interpretation is often subjective and time-consuming.

METHODS: This study aimed to develop and evaluate a lightweight two-dimensional convolutional neural network (2D CNN) for the automated classification of MPS maturation stages using axial CBCT slices. A retrospective dataset of CBCT images from 111 patients was annotated based on Angelieri's classification system and grouped into three clinically relevant categories: AB (Stages A and B), C, and DE (Stages D and E). A 9-layer CNN architecture was trained and evaluated using standard classification metrics and receiver operating characteristic (ROC) curve analysis.

RESULTS: The model achieved a test accuracy of 96.49%. Class-wise F1-scores were 0.95 for category AB, 1.00 for C, and 0.95 for DE. Area under the ROC curve (AUC) scores were 0.10 for AB, 0.62 for C, and 0.98 for DE. Lower AUC values in the early and transitional stages (AB and C) likely reflect known anatomical overlap and subjectivity in expert labeling.

DISCUSSION: These findings indicate that the proposed 2D CNN demonstrates high accuracy and robustness in classifying MPS maturation stages from CBCT images. Its compact architecture and strong performance suggest it is suitable for real-time clinical decision-making, particularly in identifying cases that may benefit from surgical intervention. Moreover, its lightweight design makes it adaptable for use in resource-limited settings. Future work will explore volumetric models to further enhance diagnostic reliability and confidence.

PMID:40642201 | PMC:PMC12241142 | DOI:10.3389/fdmed.2025.1583455

Phys Imaging Radiat Oncol. 2025 Jun 19;35:100793. doi: 10.1016/j.phro.2025.100793. eCollection 2025 Jul.

ABSTRACT

BACKGROUND AND PURPOSE: Radiotherapy targets DNA in cancer cell nuclei. Radiation dose, however, is prescribed to a macroscopic target volume assuming uniform distribution, failing to consider microscopic variations in dose absorbed by individual nuclei. This study investigated a potential link between pre-treatment tumour nuclear size distributions and post-radiotherapy outcomes in gynecological squamous cell carcinoma (SCC).

MATERIALS AND METHODS: Our multi-institutional cohort consisted of 191 non-metastatic gynecological SCC patients who had received radiotherapy with diagnostic whole slide images (WSIs) available. Tumour nuclear size distribution mean and standard deviation were extracted from WSIs using deep learning, and used to predict progression-free interval (PFI) and overall survival (OS) in multivariate Cox proportional hazards (CoxPH) analysis adjusted for age and clinical stage.

RESULTS: Multivariate CoxPH analysis revealed that a larger nuclear size distribution mean results in more favorable outcomes for PFI (HR = 0.45, 95% CI: 0.19 - 1.09, p = 0.084) and OS (HR = 0.55, 95% CI: 0.24 - 1.25, p = 0.16), and that a larger nuclear size standard deviation results in less favorable outcomes for PFI (HR = 7.52, 95% CI: 1.43 - 39.52, p = 0.023) and OS (HR = 4.67, 95% CI: 0.96 - 22.57, p = 0.063). The bootstrap-validated C-statistic was 0.56 for PFI and 0.57 for OS.

CONCLUSION: Despite low accuracy, tumour nuclear size heterogeneity aided prognostication over standard clinical variables and was associated with outcomes following radiotherapy in gynecological SCC. This highlights the potential importance of personalized multiscale dosimetry and warrants further large-scale pan-cancer studies.

PMID:40642183 | PMC:PMC12242011 | DOI:10.1016/j.phro.2025.100793

Afr J Lab Med. 2024 Nov 21;13(1):2516. doi: 10.4102/ajlm.v13i1.2516. eCollection 2024.

ABSTRACT

BACKGROUND: Artificial intelligence (AI) and digital technology, as advanced human-created tools, are influencing the healthcare sector.

AIM: This review provides a comprehensive and structured exploration of the opportunities presented by AI and digital technology to laboratory diagnostics and management of communicable and non-communicable diseases in Africa.

METHODS: The study employed the Preferred Reporting Items for Systematic Reviews, Meta-Analyses guidelines and Bibliometric analysis as its methodological approach. Peer-reviewed publications from 2000 to 2024 were retrieved from PubMed®, Web of Science™ and Google Scholar databases.

RESULTS: The study incorporated a total of 1563 peer-reviewed scientific documents and, after filtration, 37 were utilised for systematic review. The findings revealed that AI and digital technology play a key role in patient management, quality assurance and laboratory operations, including healthcare decision-making, disease monitoring and prognosis. Metadata reflected the disproportionate research outputs distribution across Africa. In relation to non-communicable diseases, Egypt, South Africa, and Morocco lead in cardiovascular, diabetes and cancer research. Representing communicable diseases research, Algeria, Egypt, and South Africa were prominent in HIV/AIDS research. South Africa, Nigeria, Ghana, and Egypt lead in malaria and tuberculosis research.

CONCLUSION: Facilitation of widespread adoption of AI and digital technology in laboratory diagnostics across Africa is critical for maximising patient benefits. It is recommended that governments in Africa allocate more funding for infrastructure and research on AI to serve as a catalyst for innovation.

WHAT THIS STUDY ADDS: This review provides a comprehensive and context-specific analysis of AI's application in African healthcare.

PMID:40642055 | PMC:PMC12242046 | DOI:10.4102/ajlm.v13i1.2516

Front Med (Lausanne). 2025 Jun 26;12:1589707. doi: 10.3389/fmed.2025.1589707. eCollection 2025.

ABSTRACT

Accurate segmentation of the prostate in T2-weighted MRI is critical for effective prostate diagnosis and treatment planning. Existing methods often struggle with the complex textures and subtle variations in the prostate. To address these challenges, we propose RaNet (Residual Attention Network), a novel framework based on ResNet50, incorporating three key modules: the DilatedContextNet (DCNet) encoder, the Multi-Scale Attention Fusion (MSAF), and the Feature Fusion Module (FFM). The encoder leverages residual connections to extract hierarchical features, capturing both fine-grained details and multi-scale patterns in the prostate. The MSAF enhances segmentation by dynamically focusing on key regions, refining feature selection and minimizing errors, while the FFM optimizes the handling of spatial hierarchies and varying object sizes, improving boundary delineation. The decoder mirrors the encoder's structure, using deconvolutional layers and skip connections to retain essential spatial details. We evaluated RaNet on a prostate MRI dataset PROMISE12 and ProstateX , achieving a DSC of 98.61 and 96.57 respectively. RaNet also demonstrated robustness to imaging artifacts and MRI protocol variability, confirming its applicability across diverse clinical scenarios. With a balance of segmentation accuracy and computational efficiency, RaNet is well suited for real-time clinical use, offering a powerful tool for precise delineation and enhanced prostate diagnostics.

PMID:40641983 | PMC:PMC12241084 | DOI:10.3389/fmed.2025.1589707

Sci Rep. 2025 Jul 10;15(1):24888. doi: 10.1038/s41598-025-08226-5.

ABSTRACT

Bug detection and resolution are pivotal to maintaining the quality, reliability, and performance of software systems. Manual debugging, along with traditional static rule-based methods, proves inefficient when applied to complex software structures in contemporary times. SynergyBug combines BERT and GPT-3 to autonomously detect and repair bugs across multiple sources. It resolves essential requirements by implementing an automated system that diagnoses and resolves software bugs automatically, thus minimising human involvement. The framework unites BERT as a contextual machinery with GPT-3 to produce bug fix generation capabilities. The semantic pattern within bug reports, together with error logs and documentation, feeds into BERT for contextual embedding generation. GPT-3 applies the generated embeddings to produce code fixes, code snippets, as well as detailed explanations that address detected problems. The system achieves continuous automatic debugging by enhancing both detection and resolution steps into one unified process. The experimental outcomes prove that it achieves superior performance than conventional bug detection methods by reaching 98.79% accuracy alongside 97.23% precision and 96.56% recall. The system demonstrated exceptional detection strength for functional and performance, and security bugs, where the detection rates reached 94% and 90% and 92%, respectively. SynergyBug showed its ability to expand as it processed bug reports exceeding 100,000 cases without noticeably impacting system performance. This proposed system provides faster debugging capabilities to improve the quality of the complete software development process. This paper discusses as a tool that can revolutionise bug management through proactive instead of just reactive strategies. The implementation of human monitoring within safety programs and managing training system biases represent essential organisational factors. The study terminates by recognising SynergyBug as a crucial development leading toward automated debugging tools that maintain operational safety within intricate software systems.

PMID:40640256 | DOI:10.1038/s41598-025-08226-5

Sci Rep. 2025 Jul 10;15(1):24840. doi: 10.1038/s41598-025-09120-w.

ABSTRACT

Super-resolution imaging has emerged as a rapidly advancing field in diagnostic ultrasound. Ultrasound Localization Microscopy (ULM) achieves sub-wavelength precision in microvasculature imaging by tracking gas microbubbles (MBs) flowing through blood vessels. However, MB localization faces challenges due to dynamic point spread functions (PSFs) caused by harmonic and sub-harmonic emissions, as well as depth-dependent PSF variations in ultrasound imaging. Additionally, deep learning models often struggle to generalize from simulated to in vivo data due to significant disparities between the two domains. To address these issues, we propose a novel approach using the DEformable DEtection TRansformer (DE-DETR). This object detection network tackles object deformations by utilizing multi-scale feature maps and incorporating a deformable attention module. We further refine the super-resolution map by employing a KDTree algorithm for efficient MB tracking across consecutive frames. We evaluated our method using both simulated and in vivo data, demonstrating improved precision and recall compared to current state-of-the-art methodologies. These results highlight the potential of our approach to enhance ULM performance in clinical applications.

PMID:40640235 | DOI:10.1038/s41598-025-09120-w

Sci Rep. 2025 Jul 10;15(1):24830. doi: 10.1038/s41598-025-05958-2.

ABSTRACT

Predicting time-series data is inherently complex, spurring the development of advanced neural network approaches. Monitoring and predicting PM2.5 levels is especially challenging due to the interplay of diverse natural and anthropogenic factors influencing its dispersion, making accurate predictions both costly and intricate. A key challenge in predicting PM2.5 concentrations lies in its variability, as the data distribution fluctuates significantly over time. Meanwhile, neural networks provide a cost-effective and highly accurate solution in managing such complexities. Deep learning models like Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM) have been widely applied to PM2.5 prediction tasks. However, prediction errors increase as the forecasting window expands from 1 to 72 hours, underscoring the rising uncertainty in longer-term predictions. Recurrent Neural Networks (RNNs) with continuous-time hidden states are well-suited for modeling irregularly sampled time series but struggle with long-term dependencies due to gradient vanishing or exploding, as revealed by the ordinary differential equation (ODE) based hidden state dynamics-regardless of the ODE solver used. Continuous-time neural processes, defined by differential equations, are limited by numerical solvers, restricting scalability and hindering the modeling of complex phenomena like neural dynamics-ideally addressed via closed-form solutions. In contrast to ODE-based continuous models, closed-form networks demonstrate superior scalability over traditional deep-learning approaches. As continuous-time neural networks, Neural ODEs excel in modeling the intricate dynamics of time-series data, presenting a robust alternative to traditional LSTM models. We propose two ODE-based models: a transformer-based ODE model and a closed-form ODE model. Empirical evaluations show these models significantly enhance prediction accuracy, with improvements ranging from 2.91 to 14.15% for 1-hour to 8-hour predictions when compared to LSTM-based models. Moreover, after conducting the paired t-test, the RMSE values of the proposed model (CCCFC) were found to be significantly different from those of BILSTM, LSTM, GRU, ODE-LSTM, and PCNN,CNN-LSSTM. This implies that CCCFC demonstrates a distinct performance advantage, reinforcing its effectiveness in hourly PM2.5 forecasting.

PMID:40640232 | DOI:10.1038/s41598-025-05958-2

Sci Rep. 2025 Jul 10;15(1):24940. doi: 10.1038/s41598-025-08249-y.

ABSTRACT

Gastric Tract Disease (GTD) constitutes a medical emergency, emphasizing the critical importance of early diagnosis and intervention to lessen its severity. Clinical practices often utilize endoscopy-supported examinations for GTD screening. The images obtained during this procedure are examined to identify the presence of the disease and investigate its severity. Autoimmune Gastritis (AIG) is a chronic inflammatory GTD and timely detection and treatment is crucial to reduce its harshness. This research aims to develop a deep-learning (DL) tool to detect the AIG from clinical-grade endoscopic images. Various stages in the DL tool comprise; (i) Image collection and resizing, (ii) image pre-processing using Entropy-function and Moth-Flame (MF) Algorithm, (iii) deep-features extraction using a chosen DL-model, (iv) feature optimization using MF algorithm and serial features concatenation, and (iv) classification and performance confirmation using five-fold cross-validation. This study aims to develop a DL tool to assist clinicians during the AIG examination and hence better detection accuracy is preferred. The merit of the DL model is demonstrated in the individual deep-features and serially concatenated-features and the experimental outcome of this study provides a detection accuracy of 99.33% when the detection is performed with fused-features and K-Nearest Neighbor classifier. This authenticates that this tool offers a clinically important outcome on the endoscopy database.

PMID:40640222 | DOI:10.1038/s41598-025-08249-y