Frontiers in medical technology最新文献

Background: Gastric reactance (XL) is a bioelectrical property of the stomach lining that responds to changes in gut perfusion. It is measured through bioimpedance spectroscopy, a technology that assesses the tissue's electrical resistance and capacity to store electrical charge, providing insight into the physiological state of the gastric mucosa.

Objective: This prospective observational study explored the relationship between XL and hemodynamic variables in high-risk adult patients undergoing elective cardiac surgery with cardiopulmonary bypass (CPB) at the National Institute of Cardiology, Mexico City.

Methods: A binary composite endpoint was constructed to aggregate major perioperative complications into a single outcome measure. The sample size was calculated based on anticipated event rates. Associations among variables were examined using nonparametric statistical tests. Predictive performance, including confidence intervals, was estimated using bootstrapped receiver operating characteristic (ROC) curves.

Results: Thirty-seven patients were enrolled and categorized according to the development of major perioperative complications (MPOC; n = 23) or absence thereof (Non-MPOC; n = 14). Baseline demographic and intraoperative variables did not differ significantly between groups. However, the MPOC group exhibited higher postoperative severity scores (APACHE II: 21.5 vs. 18.5, p = 0.231; SOFA: 12.5 vs. 12.0, p = 0.249) and greater postoperative bleeding (1.0 L vs. 0.4 L, p < 0.001). XL minimum values (XL_Min) were consistently elevated in the MPOC group throughout all perioperative events, with a significant shift of 6.14 -jΩ (95% CI [1.06, 11.34], p = 0.022) in Post-CPB.

Conclusion: These findings suggest that gastric impedance spectroscopy is a safe and feasible technique for intraoperative and postoperative monitoring, and that elevated XL_Min values may aid in the early identification of patients at risk for MPOC by detecting gastric mucosal hypoperfusion during high-risk cardiac surgery.

Introduction: Functional Near-Infrared Spectroscopy (fNIRS) is widely used to monitor cerebral hemodynamics, however, it is limited by shallow penetration depth and susceptibility to hemodynamic noise from the scalp. A novel intracranial fNIRS (ifNIRS) system, featuring depth optrodes (optode-electrodes) and optical anchor bolts (OABs), has been proposed to address these limitations. This study investigates the feasibility of ifNIRS in a swine model under controlled interventions.

Methods: Three animals were implanted with ifNIRS. Each animal with three OABs, with depth optrodes (DO) inserted into two of the OABs. Hemodynamic changes were recorded using OAB-to-OAB (OAB-OAB) and DO-to-OAB (DO-OAB) channels. Two interventions were performed to generate hemodynamic changes: rapid infusion of hypotonic saline to induce cerebral edema and blood withdrawal. Postmortem assessment for tissue damage and hemorrhage was performed. Hemoglobin concentration changes were analyzed using the Beer-Lambert equation.

Results: A decrease in total hemoglobin (tHb) levels during blood withdrawal was observed in all channel configurations that displayed relevant signals. During hypotonic saline infusion, variable patterns of tHb were observed. Postmortem findings showed minor extra-axial hemorrhages near OABs, but no intracerebral or heat-related injuries.

Discussion: This study demonstrates the feasibility of the ifNIRS system in detecting hemodynamic changes in vivo. While technical refinements are needed, ifNIRS shows promise for improving cerebral hemodynamic monitoring and enhancing diagnostic accuracy in invasive monitoring of patients with epilepsy.

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide. Due to its high invasiveness and poor prognosis, it ranks among the top three causes of cancer-related deaths globally. Accurate segmentation of the liver and lesion areas is crucial. It provides key support for diagnosis, surgical planning, and rehabilitation therapy. Deep learning technologies have been applied to the automatic segmentation of the liver and tumors. However, several issues remain, such as insufficient utilization of inter-pixel relationships, lack of refined processing after fusing high-level and low-level features, and high computational costs. To address insufficient inter-pixel modeling and high parameter costs, we propose DGA-Net (Dual-branch Group Aggregation Network for Liver Tumor Segmentation in Medical Images), a dual-branch architecture that includes two main components, i.e., a dual-branch encoder and a decoder with a specific module. The dual-branch encoder consists of the Fourier Spectral Learning Multi-Scale Fusion (FSMF) branch and the Multi-Axis Aggregation Hadamard Attention (MAHA) branch. The decoder is equipped with a Group Multi-Head Cross-Attention Aggregation (GMCA) module. The FSMF branch uses a Fourier network to learn amplitude and phase information. This helps capture richer features and details. The MAHA branch combines spatial information to enhance discriminative features. At the same time, it effectively reduces computational costs. The GMCA module merges features from different branches. This not only improves localization capabilities but also establishes long-range inter-pixel dependencies. We conducted experiments on the public LiTS2017 liver tumor dataset. Experiments on the public LiTS2017 liver tumor dataset show that the proposed method outperforms existing state-of-the-art approaches, achieving Dice-per-case (DPC) scores of 94.84% for liver and 69.51% for tumors, outperforming competing methods such as PVTFormer by 0.72% (liver) and 1.68% (tumor), and AGCAF-Net by 0.97% (liver) and 2.59% (tumor). We also carried out experiments on the 3DIRCADb dataset. The method still delivers excellent results, which highlights its strong generalization ability.

Objectives: This article is a narrative review that synthesizes current evidence on orotracheal intubation-related oral mucosal membrane pressure injuries in intensive care unit (ICU) patients, focusing on mechanisms, risk factors, and prevention strategies. The review is intended to inform clinicians and researchers by integrating insights from intensive care, biomechanics, biomaterials, and oral microbiology.

Methods: A comprehensive literature search was conducted in PubMed, Web of Science, Embase, and CNKI using the terms "orotracheal intubation", "oral mucosal injury", "device-related pressure injury", "biomechanics", "biomaterials" and "oral microbiome". Studies published between 2000 and 2025, including both clinical and experimental research, were considered without language restrictions.

Results: Evidence indicates that vertical pressure, shear force, and friction from endotracheal tubes are key contributors to oral mucosal injury. Reported risk factors include advanced age, prolonged intubation, malnutrition, and inflammation. Preventive strategies have been explored in four domains: biomechanical modeling using finite element analysis, biomaterial optimization such as hydrogel and nanocoatings, regulation of the oral microecosystem through probiotics, and intelligent monitoring systems incorporating artificial intelligence and Internet of Things technologies.

Conclusions: Orotracheal intubation-related oral mucosal pressure injuries are multifactorial and preventable. This narrative review integrates biomechanical insights, optimized biomaterials, microbiome regulation, and intelligent monitoring into a multidimensional prevention framework. Such strategies may enhance early identification, reduce complications, and improve clinical outcomes in ICU patients.

Nanoparticle-based drug delivery systems, such as liposomes, polymeric micelles, dendrimers, and other nanosized carriers, have emerged as promising strategies to improve the targeted delivery of therapeutic agents to the brain. These nanoparticles can be engineered to encapsulate drugs, facilitating their passage across the BBB, enabling localized treatment of the regions affected by neurodegeneration. Nanoparticles are characterized by their small size, large surface area, and possibility of functionalization, which allows them to be useful in many areas, including improved bioavailability, decreased systemic side effects, and improved therapeutic efficacy. Additionally, nanoparticles may also be surface-modified with appropriate ligands like antibodies, peptides, or small molecules, which exhibit specific interactions with receptors or cellular targets associated with the disease process. Such targeting has the potential to make targeted drug delivery possible, allowing therapeutic factors that can damage the healthy part of the brain to be delivered only to the diseased region. Furthermore, probable treatments for neurodegenerative diseases are also reviewed with the potential for complexation of different therapeutic agents, including small molecules, proteins, RNA, lipid nanoparticles and gene therapies with nanoparticle-based systems.

Introduction: Plant-derived exosome-like nanoparticles (PELNs) are currently a hot research topic, which have been confirmed to have similar structures and functions to mammalian-derived exosomes. PELNs are lipid bilayer membrane nanovesicles containing bioactive constituents such as miRNA, mRNA, protein, and lipids obtained from plant cells, that can participate in intercellular communication and mediate transboundary communication, have high bioavailability and low immunogenicity, are relatively safe, and have been shown to play an important role in maintaining cell homeostasis and preventing, and treating a variety of diseases.

Methods: The author has read recent articles on PELNs and summarized them.

Results: We summarized the importance and challenges of PELNs and provided a theoretical basis for the future research and clinical application of PELNs.

Discussion: In this review, we describe the biogenesis, isolation and purification methods, structural composition, stability and function of PELNs, mainly introducing the role of PELN in anti-inflammatory, anti-tumor, and drug delivery.

The rising global rates of metabolic disorders, such as obesity, type 2 diabetes, non-alcoholic fatty liver disease, and metabolic syndrome, call for new treatment methods beyond traditional drugs. The human gut microbiota, made up of trillions of microorganisms that plays a crucial role in maintaining metabolic balance through complex biochemical processes and interactions between hosts and microbes. Dysbiosis, which involves changes in microbial composition and a decrease in diversity, has become a major factor in metabolic problems. This disruption impacts the production of short-chain fatty acid, increase in permeability of intestine, and causes enduring low-grade inflammation. This review features into the potential of treatments based on microbiome for metabolic syndromes, focusing on probiotics, prebiotics, synbiotics, and postbiotics. It also encompasses innovative methods such as engineered microbial consortium, fecal microbiota transplantation (FMT), and vaginal microbiota transplantation (VMT). Probiotics show significant promise in improving blood sugar control and enhancing lipid levels. Prebiotics help bring about positive changes in microbial composition and the production of beneficial metabolites. Synbiotic combinations provide added benefits by helping good microbes thrive while supplying nutrients they can ferment. Postbiotics have recent research focus because they are safer, more stable, easier to store, and less likely to contribute to antibiotic resistance comparative to live probiotics. Even now there are substantial complications in translating microbiome research into standardized therapeutics despite of promising pre-clinical outcomes and some initial clinical data. These comprises individual variances, strain-specificity, dosage problems, regulation issues, and the necessity for personalised treatment strategies. Future success will depend upon personalized medicine, technological developments, and the incorporation of multi-omics strategy to generate metabolic health therapeutics depending on targeted microbiomes.

Digital Health Technologies (DHTs) hold immense potential for transforming drug development. Although innovation in the DHT space has been rapid, the approval process for these technologies remains slow due to fragmented efforts from industry and researchers, as well as regulatory challenges. In this position paper, we propose a hybrid methodology and approach for developing fit-for-purpose DHTs for assessment by integrating both patient-centric and data-centric elements. By emphasizing patient relevance while considering device and data feasibility, we can advance the development of patient-centric digital measures efficiently without compromising measurement precision. Ultimately, this hybrid approach aims to streamline the approval process, foster collaboration among stakeholders, and accelerate the integration of DHTs into clinical practice, thereby enhancing the overall efficiency and effectiveness of drug development.

Background: Magnetic Resonance Imaging (MRI) and ultrasound are central to tumour diagnosis and treatment planning. Although Deep learning (DL) models achieve strong prediction performance, high computational demand and limited explainability can hinder clinical adoption. Common post hoc Explainable Artificial Intelligence (XAI) methods namely Grad-CAM, LIME, and SHAP often yield fragmented or anatomically misaligned saliency maps.

Methods: We propose SpikeNet, a hybrid framework that combines Convolutional Neural Networks (CNNs) for spatial feature encoding with Spiking Neural Networks (SNNs)for efficient, event driven processing. SpikeNet includes a native saliency module that produces explanations during inference. We also introduce XAlign, a metric that quantifies alignment between explanations and expert tumour annotations by integrating regional concentration, boundary adherence, and dispersion penalties. Evaluation follows patient level cross validation on TCGA-LGG (MRI, 22 folds) and BUSI (ultrasound, 5 folds), with slice level predictions aggregated to patient level decisions and BUSI treated as a three class task. We report per image latency and throughput alongside accuracy, precision, recall, F1, AUROC, and AUPRC.

Results: SpikeNet achieved high prediction performance with tight variability across folds. On TCGA-LGG it reached $97.12\pm 0.63\%$ accuracy and $97.43\pm 0.60\%$ F1; on BUSI it reached $98.23\pm 0.58\%$ accuracy and $98.32\pm 0.50\%$ F1. Patient level AUROC and AUPRC with 95% confidence intervals further support these findings. On a single NVIDIA RTX 3090 with batch size 16 and FP32 precision, per image latency was about 31 ms and throughput about 32 images per second, with the same settings applied to all baselines. Using XAlign, SpikeNet produced explanations with higher alignment than Grad-CAM, LIME, and SHAP on both datasets. Dataset level statistics, paired tests, and sensitivity analyses over XAlign weights and explanation parameters confirmed robustness.

Conclusion: SpikeNet delivers accurate, low latency, and explainable analysis for MRI and ultrasound by unifying CNN based spatial encoding, sparse spiking computation, and native explanations. The XAlign metric provides a clinically oriented assessment of explanation fidelity and supports consistent comparison across methods. These results indicate the potential of SpikeNet and XAlign for trustworthy and efficient clinical decision support.