Frontiers in medical technology最新文献

{"title":"Dissecting and stimulating instruments in thyroid surgery: principles, applications and clinical role.","authors":"Francesco Brucchi, Carla Colombo, Giuseppe Mercante, Paolo Battaglia, Pietro Indelicato, Diego Barbieri, Davide Lombardi, Gabriele Materazzi, Gianlorenzo Dionigi","doi":"10.3389/fmedt.2026.1845958","DOIUrl":"10.3389/fmedt.2026.1845958","url":null,"abstract":"<p><p>Stimulating dissecting instruments (SDIs) combine fine dissection and calibrated nerve stimulation in a single tool, addressing the workflow limitations of conventional intermittent intraoperative neuromonitoring (IONM) in thyroid and parathyroid surgery. By eliminating repeated exchanges between a dissecting instrument and a separate stimulator, SDIs enable quasi-continuous functional assessment of the recurrent laryngeal nerve (RLN) and external branch of the superior laryngeal nerve (EBSLN) precisely during high-risk maneuvers, such as Berry's ligament dissection and superior pole management. Prospective clinical data in open thyroidectomy show that SDIs are feasible, safe, and effective in detecting early traction-related EMG changes, allowing timely release of stress and resulting in very low rates of postoperative RLN palsy, while experimental endoscopic and transoral models confirm reliable dose-response characteristics at clinically acceptable stimulation thresholds. In minimally invasive and narrow-corridor approaches, SDIs reduce instrument traffic and maintain neuromonitoring without disrupting operative flow, and their real-time feedback enhances training by directly linking dissection quality to neurophysiological signals. Within current guideline-based IONM frameworks, SDIs represent a practical evolution toward integrated, high-frequency monitoring, with future development expected to focus on their combination with energy-based devices, advanced imaging, and intelligent EMG analysis.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1845958"},"PeriodicalIF":4.6,"publicationDate":"2026-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13526971/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148868684","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of quantization on various CNN architectures for bone fracture detection.","authors":"Sushil Krishnan, Yokesh Babu Sundaresan","doi":"10.3389/fmedt.2026.1892595","DOIUrl":"10.3389/fmedt.2026.1892595","url":null,"abstract":"<p><p>The skeletal system, comprising bones, plays a crucial role in physical support and locomotion. Any deformation of bones, due to trauma or medical conditions like osteoporosis, can lead to a bone fracture. Bone fractures are painful conditions requiring immediate medical attention. Automating fracture detection can play a significant role in early diagnosis and intervention for bone fractures. Convolutional neural networks (CNN) have shown promising results in multiple medical imaging applications. There is a wide variety of convolutional neural network architectures. Due to their high computational and power requirements, they are not being adopted in resource-constrained environments. Quantization is a technique by which the size and power consumption a model are minimized with minimal loss of accuracy. It is performed by converting higher-precision data types to lower-precision data types. It enables the adoption of complex artificial intelligence models in resource-constrained environments such as mobile devices, edge devices, and IoT devices. The work focuses on finding the optimal convolutional neural network architecture and quantization method for bone fracture detection. For comparison, seven models were selected from different convolutional neural network architectures, namely, LeNet, AlexNet, VGG 19, Inception-V3, ResNet-152, MobileNetV2, and EfficientNet-B0. A baseline transformer, ViT-B/16, was also considered. Dynamic range quantization, float-16, and int-8 quantization were performed to analyze their impact on the diagnostic and deployment performances of the model. TOPSIS and Cohen's <i>d</i> analyses were performed. It was observed that lightweight models, such as LeNet and MobileNetV2, benefit greatly from quantization. This study serves as a practical guide for optimizing in a resource-constrained environment.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1892595"},"PeriodicalIF":4.6,"publicationDate":"2026-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13506755/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148835759","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intelligent data governance and quality control for chest Bi syndrome/coronary heart disease across the prevention-treatment-rehabilitation continuum: integrating a standardized framework, adversarially-optimized trigger engine, and domain-adaptive AI.","authors":"Huisi Hong, Guiyuan Yang, Weihan Shen, Shiqi Luo, Yiming Yuan, Yong Zhao, Hong Zhang, Shuangyan Li, Juhua Wu, Wenge Chen, Shaoyang Men, Kaixuan Lin","doi":"10.3389/fmedt.2026.1759773","DOIUrl":"10.3389/fmedt.2026.1759773","url":null,"abstract":"<p><strong>Background: </strong>Digital transformation in Traditional Chinese Medicine (TCM) is hindered by data silos, poor standardization, and inefficient quality control. This study developed and validated an intelligent data governance system to address these gaps.</p><p><strong>Methods: </strong>Using Chest Bi syndrome (coronary heart disease) as a model, we established a standardized data system (9 subsets, 352 elements; 42.6% TCM-specific) based on the \"Triple-Loop Coupling\" theory and HL7 FHIR (Fast Healthcare Interoperability Resources)/SNOMED CT standards. The system integrates a three-tier trigger engine (128 rules) for real-time alerts and a domain-adapted LLM, fine-tuned on TCM texts and guidelines, for unstructured data processing. Efficacy was evaluated via a prospective controlled study (<i>n</i> = 972).</p><p><strong>Conclusion: </strong>The intervention group showed significant improvements (<i>P</i> < 0.001): data completeness rose from 82.51% to 99.59% (<i>Δ</i>=17.08%), and TCM-specific data elements collection improved from 67.08% to 98.97% (<i>Δ</i>=31.89%). Logical consistency errors dropped from 15.84% to 1.65% (<i>Δ</i>=-14.20%). Unstructured data conversion increased from 45.06% to 91.77% (<i>Δ</i>=46.71%). Quality control time per record decreased by 58.95% (from 12.48 to 5.12 min). The AI model achieved a 91.4% F1-score and 91.8% alert accuracy.</p><p><strong>Conclusion: </strong>The \"standardization-trigger-AI\" model significantly enhances data quality and efficiency for TCM. This system provides a preliminary and verifiable exploratory path for the intelligent governance of TCM clinical data, offering a scalable solution for TCM modernization.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1759773"},"PeriodicalIF":4.6,"publicationDate":"2026-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13508055/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148835803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An automated deep learning-based pipeline for 3D characterisation of the murine upper airway from micro-CT.","authors":"Flore Belmans, Sergi Llambrich, Wim Vos, Frederik Maes, Neus Martínez-Abadías, Greetje Vande Velde","doi":"10.3389/fmedt.2026.1823048","DOIUrl":"10.3389/fmedt.2026.1823048","url":null,"abstract":"<p><strong>Introduction: </strong>Upper airway structure and function are affected in several respiratory (e.g., obstructive sleep apnoea, allergic rhinitis, and asthma) and congenital [e.g., Down syndrome (DS)] conditions. Although murine models are widely used to study airway pathology and to evaluate therapies, quantitative analysis of the upper airway on micro-computed tomography (micro-CT) remains limited as it is hampered by labour-intensive and observer-dependent manual segmentation of the images. Automated and standardized image analysis methods are therefore needed to support mechanistic and translational upper airway research.</p><p><strong>Materials and methods: </strong>We developed a deep learning-based pipeline for fully automated segmentation and characterization of the trachea, pharynx, and nasal cavity from micro-CT images. A full-resolution 2D CNN based on the nnU-Net framework was trained in the axial plane using 5-fold cross-validation on 106 scans from a DS study involving wild-type (WT) and trisomic (TS) mice. Generalisability was assessed on an external validation set of 74 scans from another DS study examining the effect of RSV infection in WT and TS mice. Model performance was evaluated using the Dice similarity coefficient (DSC), relative absolute volume difference (RAVD), Pearson correlation and Bland-Altman analysis between manual and automated measurements.</p><p><strong>Results: </strong>Strong segmentation performance was obtained on the internal validation set, with DSC of 0.988 ± 0.010, 0.985 ± 0.008, and 0.958 ± 0.004, and RAVD of 1.483% ± 1.800%, 1.506% ± 1.217% and 1.965% ± 1.563% for trachea, pharynx and nasal cavity, respectively. Performance remained robust on the external validation set, confirming generalisability across different experimental conditions and imaging protocols. Automatically extracted volumetric biomarkers showed strong agreement with manual measurements and small Bland-Altman biases. The pipeline reproduced biologically relevant group differences previously identified through manual analysis, including reduced pharynx and nasal cavity volumes in TS mice compared with WT controls.</p><p><strong>Conclusions: </strong>We present a deep learning-based pipeline for automated, standardized 3D characterization of the murine upper airway from micro-CT. By reducing manual workload and enhancing measurement consistency, it enables high-throughput phenotyping and mechanistic investigation of the upper airway in preclinical models. By facilitating direct, automated measurements from imaging, our pipeline strengthens the translational bridge between experimental and clinical research.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1823048"},"PeriodicalIF":4.6,"publicationDate":"2026-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13503376/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148820567","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A comparative performance analysis of ensemble learning and regularized neural networks in cardiovascular risk prediction.","authors":"Golam Md Mohiuddin, Md Shakil Hossain, Md Shohel Sayeed","doi":"10.3389/fmedt.2026.1902471","DOIUrl":"https://doi.org/10.3389/fmedt.2026.1902471","url":null,"abstract":"<p><strong>Introduction: </strong>Heart disease classification using small structured clinical datasets requires evaluation procedures that account for sampling variability, model-selection bias, probability calibration, and clinically relevant operating characteristics. This study presents an internally validated comparison of classical classifiers, modern tree ensembles, and regularized artificial neural networks using a cleaned Cleveland heart disease dataset containing 303 observations, 13 predictors, and a binary outcome.</p><p><strong>Methods: </strong>Continuous variables were standardized, nominal variables were one-hot encoded, and all preprocessing transformations were fitted exclusively within training folds. Model performance was evaluated using five-fold outer cross-validation repeated five times, with inner stratified cross-validation for hyperparameter optimisation. Twelve models were assessed using accuracy, sensitivity, specificity, positive and negative predictive values, F1-score, Matthew's correlation coefficient, ROC-AUC, PR-AUC, Brier score, and calibration measures.</p><p><strong>Results: </strong>CatBoost achieved the highest mean ROC-AUC of 0.909 ± 0.041 and F1-score of 0.862 ± 0.041, while Logistic Regression achieved a comparable ROC-AUC of 0.906 ± 0.041. Random Forest obtained the highest mean PR-AUC of 0.916 ± 0.040. CatBoost did not significantly outperform Logistic Regression in ROC-AUC, PR-AUC, F1-score, or paired classification errors. Held-out permutation analysis identified the number of major vessels, thalassemia status, and chest-pain category as the most influential predictors.</p><p><strong>Discussion: </strong>The findings indicate that no model family was universally superior and that a carefully regularized linear model remained competitive with substantially more complex approaches. These findings represent internal validation only and require confirmation using independent clinical cohorts before deployment.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1902471"},"PeriodicalIF":4.6,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13490151/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148803588","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ex vivo fluorescence confocal microscopy combined with rapid immunofluorescence for intraoperative intestinal mapping in near-total intestinal aganglionosis: a pediatric proof-of-concept case report.","authors":"Donatella Di Fabrizio, Alessandra La Contana, Edoardo Bindi, Alba Cruccetti, Gaia Goteri, Giovanni Cobellis","doi":"10.3389/fmedt.2026.1869954","DOIUrl":"10.3389/fmedt.2026.1869954","url":null,"abstract":"<p><strong>Introduction: </strong>In Hirschsprung disease (HD), intraoperative identification of ganglionated bowel remains central to surgical planning, particularly in extensive or near-total intestinal aganglionosis (nTIA), where bowel preservation is critical. Frozen-section histopathology remains the intraoperative standard. Ex-vivo fluorescence confocal microscopy (FCM) provides microscopic assessment of fresh tissue, while rapid immunofluorescence (rIF) has the potential to add targeted visualization of enteric neural structures. We describe a single pediatric proof-of-concept case in which FCM combined with selected rIF was integrated into intraoperative intestinal mapping.</p><p><strong>Case-report: </strong>A full-term newborn, diagnosed at birth with nTIA, underwent tube stoma formation at 40 cm from the Treitz ligament without bowel resection. At 4-years, surgery was scheduled to remove aganglionic segments and lengthen the remaining bowel. During surgery the small bowel segment proximal to the ostomy measured 70 cm. Twenty-four full-thickness biopsies were collected every 20 cm from sigmoid colon to presumed ganglionic areas at stoma level. All samples underwent FCM for morphological assessment of fresh tissue. rIF was applied to four selected samples, two from presumed ganglionated bowel and two from presumed aganglionic bowel, as an adjunct to enhance visualization of S100B-positive neural and glial structures. Fresh tissue was incubated for 30 min with a primary anti-S100B antibody followed by an Alexa Fluor 647-conjugated secondary antibody, then rinsed and counterstained with acridine orange. Under fluorescence laser visualization ganglion cells were identified in functional segments while hypertrophic nerve fibers were detected in aganglionic regions. Surgical planning relied on integrated intraoperative assessment, with final confirmation by conventional histopathology. The simultaneous application of a longitudinal intestinal lengthening (LILT) ensured maximization of functional bowel.</p><p><strong>Conclusion: </strong>This single-case experience demonstrates the technical feasibility of integrating <i>ex vivo</i> FCM and rIF into intraoperative intestinal mapping for complex Hirschsprung disease. The approach should be regarded as exploratory and complementary to established histopathological assessment.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1869954"},"PeriodicalIF":4.6,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13463412/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148727300","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of catheter navigation difficulty during mechanical thrombectomy using CTA-based carotid siphon morphometry.","authors":"Ferenc Dezső Bakó, Attila Tanács, József Tolnai, Pál Barzó, Tamás Németh, Attila Nagy","doi":"10.3389/fmedt.2026.1900255","DOIUrl":"10.3389/fmedt.2026.1900255","url":null,"abstract":"<p><strong>Introduction: </strong>Successful catheter navigation through the carotid siphon remains a major technical challenge during mechanical thrombectomy for acute ischemic stroke. Unfavorable vascular anatomy may require stepwise device escalation, prolonging procedural time and potentially delaying reperfusion. Because conventional two-dimensional measurements may not fully capture the three-dimensional vascular features influencing device behavior, this study evaluated whether CTA-derived carotid siphon morphometry can predict the level of device support required for aspiration catheter navigation.</p><p><strong>Methods: </strong>This retrospective study included patients who underwent endovascular thrombectomy at the Department of Neurosurgery, University of Szeged, between 15 April 2020 and 16 May 2022. From an initial cohort of 160 patients, 54 CTA datasets were suitable for three-dimensional segmentation, and 53 cases were included in the final morphometric analysis. Paired native CT and CTA scans were processed using a subtraction-based workflow in 3D Slicer, followed by vesselness filtering, segmentation, and centerline extraction. Local carotid siphon features, including cross-sectional area and curvature-related distance measurements, were extracted. Procedural difficulty was defined using a stepwise device-escalation scale reflecting the support required to advance the aspiration catheter through the carotid siphon. A logistic regression classifier with ridge regularization and correlation-based feature selection was trained using nine geometric attributes.</p><p><strong>Results: </strong>In the full 53-case dataset, the model achieved 58.5% accuracy, with substantial interclass overlap, particularly among higher-complexity categories. Applying a confidence-based decision criterion, retaining only predictions where the second-highest class probability was less than 60% of the highest class probability, yielded a high-confidence subset of 32 cases. In this subset, accuracy increased to 75.0%, the kappa statistic increased from 0.40 to 0.65, and the weighted F-measure reached 0.73. The model achieved 100% recall for M2 and 100% precision for M3, whereas M4 remained difficult to distinguish.</p><p><strong>Conclusion: </strong>CTA-derived three-dimensional morphometric analysis of the carotid siphon provides clinically relevant information on catheter navigation difficulty during mechanical thrombectomy. Confidence-gated classification identified anatomically well-defined cases in which device-support requirements could be predicted more reliably. Borderline cases remained influenced by factors beyond luminal geometry. Larger multicenter studies are warranted to validate this approach and assess its role in patient-specific preprocedural planning.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1900255"},"PeriodicalIF":4.6,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13461604/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148724104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Iron-oxide nanoparticle-integrated soft contact lenses: optical optimization, catalytic functionality, and biocompatibility.","authors":"Aswathi Ram P, Muhammed Shebeeb, Muhammed Hisham, Sanjana Chandran, Said El Turk, Haider Butt","doi":"10.3389/fmedt.2026.1782878","DOIUrl":"10.3389/fmedt.2026.1782878","url":null,"abstract":"<p><p>Millions of people suffer from corneal infections and inflammatory conditions, and a contact lens that not only corrects vision but also continuously protects against microbes and inflammation could transform ocular therapeutics. Compared with plasmonic metal nanoparticles, such as silver or gold, iron oxide nanoparticles are well known for their biocompatibility, low toxicity, and low cost. To our knowledge, this work is the first to demonstrate the integration of iron oxide nanoparticles into commercial contact lenses. Nanoparticles were characterized using X-ray diffraction, X-ray photoelectron spectroscopy, scanning electron microscopy and transmission electron microscopy. Optical properties were assessed using UV-Vis spectroscopy and COMSOL simulations. Catalytic activity was evaluated by methylene blue degradation in the presence of H₂O₂. Lens's biocompatibility, anti-inflammatory and anti-bacterial properties were investigated using fibroblast cell viability assays and antibacterial testing. X-ray diffraction verified the structure with a lattice parameter of 8.37 Å, while XPS shows that the obtained nanoparticles (NPs) consisted of a mixture of both Fe₃O₄/γ-Fe₂O₃ NPs. SEM and transmission electron microscope confirmed that the NPs are uniformly dispersed and have an average size of ∼8 nm. UV-vis spectroscopy results showed significant optical transmission after embedding the NPs. The COMSOL simulation data indicated that smaller nanoparticle radii maintain high transparency and larger radii reduce transmission. At higher nanoparticle loading, the lenses exhibited 96.7% methylene blue degradation in the presence of H₂O₂ over 2 weeks. Biocompatibility testing showed >70% fibroblast cell viability with normal morphology, along with anti-inflammatory and antibacterial activity against <i>Staphylococcus aureus</i>. Findings show the potential of iron-oxide nanoparticle-embedded contact lenses as transparent, catalytically active, and biologically safe platforms for future therapeutics and wearable optics applications.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1782878"},"PeriodicalIF":4.6,"publicationDate":"2026-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13447371/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148690632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Feature-reconstructed diabetic retinopathy classification using variational autoencoder with disentanglement factor.","authors":"Ashu Priya, Niranjana Banerjee, Manas Ranjan Prusty","doi":"10.3389/fmedt.2026.1826143","DOIUrl":"10.3389/fmedt.2026.1826143","url":null,"abstract":"<p><p>Diabetic Retinopathy or DR is one of the leading causes of blindness in the working age population across the world, making its early detection and accurate classification as one of the major challenges in healthcare and medical imaging. Over the years various approaches have been developed like conventional deep learning models to classify DR or grade DR according to its severity. However, it has been observed that some of the major challenges encountered by various approaches is the lack of a lightweight feature focused learning mechanism. In this study, the authors propose a Variational Autoencoder (VAE) with disentanglement factor (Beta) and Logistic Regression based framework to classify DR in both ways, binary and multilevel classification. The proposed architecture aims to extract the fine features and textures of the retinal images compress it into a latent vector via the encoder and then expand the image map into a reconstructed image which aids the Logistic Regression classifier to classify the images and distinguish the severity of the disease. The proposed model was evaluated on two well-known datasets, APTOS 2019 dataset and DDR dataset. In case of binary classification, the model achieved an accuracy of 98.64% on the APTOS 2019 dataset and a 97.83% accuracy when evaluated on the DDR dataset. On multilevel classification, the model recorded an accuracy of 97.80% and 97.23% on APTOS 2019 dataset and DDR dataset respectively. These findings highlight the potential of the proposed method as an accurate and effective tool for automated DR screening and severity grading.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1826143"},"PeriodicalIF":4.6,"publicationDate":"2026-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13429598/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148671755","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Nanosystems for delivery of indolicidin peptide.","authors":"José Gregorio Martín Bedoya, Katia Conceição","doi":"10.3389/fmedt.2026.1845120","DOIUrl":"10.3389/fmedt.2026.1845120","url":null,"abstract":"<p><p>The emergence of multidrug-resistant pathogens has necessitated the search for novel therapeutic alternatives such as antimicrobial peptides (AMPs). Indolicidin (Ind) represents a highly promising candidate owing to its broad-spectrum activity and multimodal mechanisms; however, its clinical translation is limited by rapid proteolytic degradation and intrinsic host cytotoxicity. To overcome these pharmacological bottlenecks, this review places nanocarrier based delivery of Ind at its analytical center, rigorously evaluating the integration of Ind into nanoscale delivery systems as a transformative therapeutic strategy. We critically analyze how diverse organic and inorganic platforms fundamentally reconfigure Ind's mechanism of action, including the bidirectional relationship between nanoparticle-generated ROS and peptide-mediated membrane disruption. Furthermore, we discuss emerging bio-inspired approaches utilizing stimuli-responsive hydrogels, lipid cubic phases, and liposomes for programmed intracellular release. A critical assessment of long-term genotoxicological and ecotoxicological safety profiles is subsequently provided to address mandatory regulatory prerequisites. This integrated analysis establishes a comprehensive mechanistic and translational framework for the rational development of next-generation Ind-based antimicrobial nanomedicines, balancing structural tethering with supramolecular assembly to effectively combat antibiotic-resistant infections.</p>","PeriodicalId":94015,"journal":{"name":"Frontiers in medical technology","volume":"8 ","pages":"1845120"},"PeriodicalIF":4.6,"publicationDate":"2026-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13429792/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148671872","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}