Laboratory automation enhances repeatability and throughput in scientific and industrial applications. An important aspect of automation is transportation of standardized components, such as microplates (MPs). Previously, we developed the compliant-mechanism-based gripper ("CrocoGrip"), enabling secure and contamination-free MP transport. However, we have yet to optimize the gripping jaws to increase the gripper's maximum load capacity (LC). In this paper, we optimize the design and surface of CrocoGrip's jaws to optimize its LC. CrocoGrip works like a torsion spring, so its opening width, which defines the CrocoGrip's deformation, its jaw arm length, and the jaw's surface material affect its LC. We tested the LC of six different jaw materials at different opening widths. The best-performing jaw surface was further evaluated at a second jaw arm length. Jaws equipped with silicone inserts performed the best (LC=2.93 N). Aluminum with different surface smoothness performed the worst (0.68 N≤LC≤1.56 N). Decreasing the jaw arm length from 81 mm to 66 mm increased the LC of the silicone-insert jaws to 3.71 N. The improved LC allows the safe handling of MP (weight ≤127 g), enabling the use of the CrocoGrip for a broader range of tasks.
The atrial repolarization (Ta wave) characteristics remains largely unexplored, given its inherently low amplitude and obscured by the QRS complex. Hence, this study aims to witness Ta wave within QRS complex. 10 s ECGs of 50 Sinus Rhythm (SR), 50 Sinus Tachycardia (SiT) and 20 Atrial Tachycardia (AT) were recorded using standard 12-lead. The datapoints were extracted from pre-processed Lead-II and three spline model interpolated hidden Ta wave post fiducial point detection. Further, validation analysis was performed with and without QRS complex to select the optimal spline model with the Ta wave of SiT Modified Limb Lead (MLL) & Atrio-Ventricular block (AVB) ECG. It was noted that the cubic spline interpolation model gave the best SSIM score of 0.85 and lowest power spectrum % difference of 0.1 % for Ta wave interpolation without QRS complex. Further, ECG-based Ta temporal and voltage features were crafted. Statistically significant features were used for five ML models multi-class classification. Extra-Trees model gave the best output with 99 % using P-Ta feature combined. Overall, the proposed method demonstrated that along with the existing P wave features, Ta wave features have potential in better classification of atrial arrhythmia, while interpolation model offers ease of implementation and adaptability to diverse clinical applications.
Tendon and ligament injuries significantly affect patients' quality of life, with current treatments, such as surgery and tissue grafting, often resulting in recurrence and additional complications. Tissue engineering has emerged as a promising alternative by integrating cells, biomaterials, and bioactive cues. This review summarizes recent studies (2015 to present) on the application of polysaccharides in developing tissue-engineered tendons and ligaments, emphasizing the critical role of cell-material interactions in various stages of tissue engineering. Polysaccharide-based biomaterials have gained attention for their structural and functional versatility, biocompatibility, and abundance from renewable sources. Their resemblance to the native extracellular matrix of tendons makes them excellent candidates for scaffolding biomaterials in tendon and ligament tissue engineering. This review highlights advancements in using polysaccharides, demonstrating their potential to enhance regenerative outcomes by closely mimicking the native tissue environment.
The antimicrobial efficacy and operating parameters of a cold plasma prototype were investigated. The device concept is intended for use on laparoscopic trocar incisions. The device (50 mm length, 10 mm diameter, glass-ceramic body with two wire-wound electrodes) was operated with two gas compositions and in different environments. In vitro decontamination was tested on wet agar plates inoculated with Staphylococcus aureus (Gram-positive) and Escherichia coli (Gram-negative) as well as on inoculated stainless steel and polypropylene strips. A 5-min plasma exposure was applied. Both microorganisms were effectively inactivated on wet agar surfaces (inhibition zone assay: up to 36 mm) using different process gas compositions. Additionally, antimicrobial action was confirmed for S. aureus on stainless steel and polypropylene substrates. The prototype thus shows consistent decontamination performance across tested modes. The plasma source offers a promising, minimally invasive adjunct for preventing surgical site infections during laparoscopic procedures. Further development, advanced biological models, and compliance with regulatory standards (e.g., DIN SPEC 91315) are required before clinical implementation.
Conventional cuff-based blood pressure (BP) measurement provides only intermittent readings, whereas photoplethysmography (PPG)-based methods enable continuous and noninvasive monitoring. This study aims to develop a deep learning framework for accurate, cuffless BP estimation using a single PPG signal. A hybrid deep neural network, termed ResNet-BiGRU, was developed by integrating residual convolutional blocks with bidirectional gated recurrent units to jointly capture morphological and temporal features. The UCI Cuff-Less Blood Pressure Estimation Dataset (a subset of MIMIC-II), which contains synchronized PPG and arterial blood pressure (ABP) signals from 942 subjects, was used for model training and validation. After applying a 0.5-8 Hz bandpass filter and segmenting into 5 s windows, the data were split 80/20 for training and validation. External evaluation was conducted using the VitalDB dataset, which provides synchronized PPG and ABP recordings from surgical patients under diverse physiological conditions. The model achieved mean absolute errors (MAE) of 4.78 mmHg for systolic BP (SBP) and 2.98 mmHg for diastolic BP (DBP) on UCI, and 8.15 mmHg for SBP and 4.59 mmHg for DBP on VitalDB. The ResNet-BiGRU model demonstrates accurate, robust, and generalizable cuffless BP estimation, showing strong potential for wearable health monitoring applications.
To introduce ICEEMDAN-PC, a novel quantitative ultrasound (QUS) approach for accurate and noise-robust estimation of mean scatterer spacing (MSS), enabling refined characterization of liver and breast tissue microstructures in health, disease, and post-treatment states. ICEEMDAN-PC integrates the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and power cepstrum (PC). The intrinsic mode function with the highest energy is extracted, and its cepstrum is analyzed to determine MSS. The method was validated using 30,100 simulated noisy RF signals (semi-periodic/diffuse scatterers) and applied to: (1) 301 ex vivo porcine liver signals, (2) 31,488 paired RF signals pre/post microwave ablation (MWA), and (3) RF data from 100 clinical breast lesions (52 malignant, 48 benign). Simulations recovered the theoretical MSS (1.25 mm) with low variance despite high noise. In a healthy liver, MSS was 1.02 mm, with significant shifts post-MWA indicating microstructural disruption. Breast lesion MSS values (0.8736 mm benign, 0.9068 mm malignant) matched literature trends. ICEEMDAN-PC consistently achieved high accuracy and sensitivity across simulated, experimental, and clinical datasets, demonstrating strong potential for non-invasive QUS-based tissue characterization and therapeutic monitoring.
Diabetic Retinopathy (DR) causes major vision loss, requiring precise segmentation of retinal vessels and the Foveal Avascular Zone (FAZ). Accurate structural masks enable quantitative biomarkers that support early diagnosis and long-term monitoring. We propose a Retinal Graph Neural Network (RGNNNet) for OCTA segmentation. It combines multi-scale feature extraction with a graph representation, where node relations derive from an affinity matrix of feature maps. A symmetric normalization strategy stabilizes graph propagation and integrates local-global vascular context. A hybrid Dice-Focal loss refines fine-structure segmentation. On OCTA-500, RGNNNet achieved superior Dice and IoU to existing methods. For FAZ, it attained Dice values of 96.78 % (6 mm) and 98.02 % (3 mm), and maintained 0.915 on ROSE-0 without retraining. It outperformed baselines by 1-3 % Dice for other classes and remained lightweight (0.83 M params, 11.25 ms per 400 × 400 image). By coupling residual feature learning with graph-based relational reasoning, RGNNNet provides accurate structure-specific masks that can serve as a foundation for downstream biomarker extraction. Its compact design and stable generalization highlight its potential for large-scale ophthalmic screening and integration into clinical workflows.
Micro-computed tomography (Micro-CT) is renowned for its high resolution, holding a pivotal role in advancing medical science research. However, compared to CT medical imaging datasets, there are fewer publicly available Micro-CT datasets, especially those annotated for multiple objects, leading to segmentation models with limited generalization abilities. In order to improve the accuracy of multi-organ segmentation in Micro-CT, we developed a novel segmentation model called MOSnet which can utilize annotations from different datasets to enhance the whole segmentation performance. The proposed MOSnet includes a control module coupled with a reconstruction block that forms a multi-task structure, effectively addressing the absence of complete annotations. Experiments on 85 contrast-enhanced micro-CTscans and 140 native micro-CTscans for mice demonstrate that MOSnet is superior to the most of advanced segmentation networks. Compared to the best results of ResUnet, Unet3+, DAVnet3+ and AIMOS, our method improved dice similarity coefficient by 4.1 and 2.4 %, increased jaccard similarity coefficient by 4.1 and 3.1 %, and reduced HD95 by 16.3 and 19.3 % on the two datasets respectively at least. Our proposed model proves to be a robust and effective method for multi-organ segmentation in micro-CT, especially in situations where comprehensive annotations are lacking within a dataset.
Early hematoma expansion is a major determinant of poor outcome in hypertensive basal ganglia hemorrhage. This study evaluated whether CT-based radiomic texture analysis could improve early prediction of hematoma expansion. A retrospective cohort of 104 patients with hypertensive basal ganglia hemorrhage who underwent baselinef CT within 6 h of symptom onset and follow-up CT within 48 h was analyzed. Hematoma regions of interest were manually segmented, and 256 texture features were extracted using MaZda. Fisher's score, probability of error and average correlation coefficient, and mutual information were used for dimensionality reduction. Classification performance was assessed using raw data analysis, principal component analysis, linear discriminant analysis, and nonlinear discriminant analysis, followed by ROC analysis. Hematoma expansion occurred in 40 of 104 patients (38.4 %). Nonlinear discriminant analysis showed the lowest misclassification rate overall, including 0 under the POE-ACC feature set. ROC analysis demonstrated good diagnostic performance for several texture features, with S(3, -3)Difvarnc (AUC 0.944), S(4, -4)Difvarnc (AUC 0.942), and GrVariance (AUC 0.917) showing the strongest predictive value. CT-based texture analysis provides quantitative imaging biomarkers that may support early risk stratification of hematoma expansion in basal ganglia hemorrhage.
Existing diaphragm pacing (DP) system use an open-loop control method with a fixed stimulation mode to control breathing. It requires doctors to manually adjust stimulation parameters to meet the patient's ventilation needs. A neural network adaptive controller was tested to control breathing in DP system. For the diaphragm motion model, the respiratory airflow of healthy adults and rabbits was collected and compared with the simulation calculation results to verify the accuracy of the model. The performance of the adaptive controller was evaluated comparatively with that of the PID controller. Adaptive controller consists of a neural network and adaptively adjusts stimulation parameters to produce the desired respiratory volume waveform. Superiority of the output performance of the adaptive controller was further studied by setting various diaphragm model parameters. We further verified the feasibility of the adaptive controller through animal testing. The adaptive controller is better than the PID controller in maintaining the stability of the desired breath volume. Application of the adaptive controller can reduce the root mean square (RMS) error between the desired breath volume and the actual value to less than 6 %. This study demonstrates the potential application of adaptive controllers in closed-loop DP systems.
The study aims on evaluation criteria applicable for the development process of new mobile 3D X-ray imaging devices, based on existent knowledge from literature and user expertise. For this purpose, a structured literature review was performed to outline limitations and opportunities of established systems on the market. Additionally, medical professionals from various medical fields were interviewed, following a structured interview protocol, to identify critical evaluation aspects from a user perspective. The study identified a total of 52 evaluation criteria, which were categorized into the following domains: clinical application and manageability, image quality, radiation dose and risk of infection. The analysis of user needs, subsequent summarization, and rating of these needs will serve as a new benchmark for the evaluation of both existing and new devices. This will provide guidance to procurement, medical experts, and manufacturers. While the structured evaluation of mobile X-ray devices is possible, consultation of the users regarding their weighting of varying needs will remain an important aspect of the process. The evaluation items can serve as a starting point for this analysis.
the aim of this study was to clarify the biocompatibility of different tooth repositioning splints (aligners). This included the characterization of the volatile fraction by headspace solid phase microextraction coupled to gas chromatography mass spectrometry, simulation of leachable organic compounds using artificial saliva and the quantitation of bisphenol A and bisphenol S after extraction. Four different aligners were characterized in this study, these included two splints with novel shape memory properties: a printable aligner made of the resin Tera Harz TC-85 DAC (Graphy Inc., South Korea) and a self-manufactured SMP-Aligner consisting of the components polypropylene carbonate and thermoplastic polyurethane. The other two aligners were conventional, thermoformable aligners: CA® Pro Clear Aligner (Scheu Dental GmbH, Germany) and Erkodur-al (Erkodent Erich Kopp GmbH, Germany). no BPA was found in all four samples after 72 h. BPS was found in one Aligner. The most leachable compounds were found in the samples of the SMP- and the direct printed Graphy-Aligner. Most of the compound release occurred during the first 24 h. the cumulative exposure effect from aligner wear should be carefully considered, although the current research indicates that the levels of leachable compounds are generally low.
We developed a multimodal fusion model combining clinical data and deep transfer learning for early progressive cerebral contusion (PCC) prediction, providing precise clinical support for treatment decisions. Using a single-center retrospective cohort design, we analyzed 196 cerebral contusion patients between January 2022 and June 2024. PCC was characterized by a contusion volume increase of at least 30 % on CT scans within 24 h. Patients were categorized into a progression group (n=98) and a non-progression group (n=98). The dataset was split into a training coh59 participants, maintaining a 7:3 ratort of 137 participants and a validation cohort of io. A nomogram was developed by combining ResNet-50-based deep transfer learning features with clinical variables. Model performance was assessed through ROC curves, calibration plots, and decision curve analysis, while Grad-CAM was used to evaluate interpretability. The integrated nomogram demonstrated superior performance with AUC values of 0.999 (95 % CI: 0.998-1.000) in the training cohort and 0.972 (95 % CI: 0.939-1.000) in the validation cohort, surpassing the standalone DTL and clinical models. Grad-CAM demonstrated accurate lesion localization. The multimodal fusion model integrating DTL and clinical features shows excellent predictive performance and significant clinical value in early PCC prediction.
Accurate identification of Parkinson's disease (PD), particularly during its prodromal stage, remains a major clinical challenge due to heterogeneous symptom presentation and overlapping neurological patterns. This study proposes an LLM-Guided Multimodal Attention Network (LLM-MAN) to improve PD staging by jointly modeling structural MRI and clinical/cognitive metadata. We develop a unified multimodal framework that encodes structural MRI using a ResNet-18 backbone enhanced with Convolutional Block Attention Modules (CBAM) for discriminative neuroimaging feature extraction, and represents clinical/cognitive metadata using an LLM-based text encoder (pre-trained BERT) for contextualized semantic modeling. A Meta-Guided Cross-Attention (MGCA) module is introduced to align clinical semantic knowledge with imaging features, enabling robust cross-modal fusion for multiclass classification (Normal Control, prodromal PD, and diagnosed PD). The model is evaluated on the Parkinson's Progression Markers Initiative (PPMI) dataset and further validated on an independent external cohort. On the PPMI dataset, LLM-MAN achieved an accuracy of 95.68 % for distinguishing Normal Control, prodromal PD, and diagnosed PD. External validation on an independent cohort yielded 94.10 % accuracy, indicating strong generalization performance across datasets. LLM-guided multimodal fusion via MGCA provides reliable and interpretable approach for PD staging, substantially improving prodromal PD identification by integrating semantic clinical knowledge with neuroimaging representations.
To develop and validate a nnU-Net-based clinical radiomics model for predicting poor outcome in patients with sudden sensorineural hearing loss (SSNHL). A retrospective cohort of 124 SSNHL patients undergoing temporal bone high-resolution computed tomography (HRCT) was analyzed (54 good prognosis; 70 poor prognosis). Patients were randomly divided into training (n=87) and test (n=37) sets. The cochlea, vestibule, and internal auditory canal were manually segmented and used to train a nnU-Net 3D full-resolution model. Segmentation performance was evaluated using the Dice similarity coefficient (DSC). Radiomics features were extracted and reduced through variance thresholding, correlation analysis, univariate Cox regression, and random survival forest modeling to construct a radiomics score (Radscore). Independent prognostic factors were identified using multivariate Cox regression. A combined clinical-radiomics nomogram was developed and compared with clinical-only and Radscore-only models using C-index, calibration, and decision curve analysis (DCA). The nnU-Net achieved DSCs of 0.91 ± 0.07 (training) and 0.73 ± 0.14 (test). Twelve radiomics features were selected. High-risk Radscore and four clinical factors were independent predictors. The combined model showed superior discrimination (C-index: 0.812 training; 0.783 test) and the highest clinical net benefit. The nnU-Net-based clinical radiomics model provides accurate prognostic stratification for SSNHL.
To predict abnormal pulmonary artery hemodynamics caused by ventricular septal defect (VSD) using Physics-Informed Neural Networks (PINN) and address the challenges of high computational cost in traditional Computational Fluid Dynamics (CFD) and difficulty in obtaining measurement data. The PINN model was trained using boundary conditions and scattered clinical CFD data, with dynamic weighting factors incorporated to enhance training efficiency and optimize predictions. Model outputs for blood flow velocity and pressure were subsequently evaluated against CFD simulation results. The PINN accurately reproduced velocity and pressure fields across pulmonary artery models using only boundary conditions and sparse internal measurements. For velocity prediction, the average RMSE, MAE, and MRE for components u, v, and w ranged from 0.274 to 0.832 %, 0.448-1.096 %, and 0.833-1.341 %, respectively. For pressure prediction, the average RMSE, MAE, and MRE ranged from 2.953 to 5.145 %, 3.264-5.679 %, and 0.376-0.565 %, respectively. These findings demonstrate that the framework generalizes well and provides reliable hemodynamic estimation with limited input data. The PINN model compensates for incomplete measurement data through physical constraints, enabling rapid and accurate prediction of pulmonary artery hemodynamics and offering a promising non-invasive alternative for pulmonary artery pressure measurement.
To develop and validate a deep learning framework for classifying postoperative time-points as a proxy task for monitoring longitudinal fracture healing progression from serial radiographs. This retrospective study included 150 patients with paired pre-treatment and follow-up X-ray images. We built a detection-guided pipeline comprising (1) fracture-region localization using an enhanced YOLOv11 detector integrating attention mechanism, Focal-SIoU loss, and data augmentation, and (2) healing-status prediction from detected regions of interest by quantifying callus formation and fracture-line changes over time. Data were split at the patient level into training/validation/test cohorts. Performance was evaluated using accuracy, F1 score, ROC/AUC, and calibration, and compared with clinician readings. The YOLOv11-guided framework achieved reliable fracture localization and consistent healing assessment on serial radiographs. On the independent test set, it showed stable discriminative ability across follow-up stages and improved robustness over manual interpretation, particularly at early postoperative time points when radiographic changes are subtle. This single-center study demonstrates a technical framework for objective and scalable radiograph-based longitudinal fracture-healing monitoring. External, multi-center validation is required before broader clinical deployment. The proposed detection-enhanced YOLOv11 framework may support clinical follow-up and decision-making after fracture surgery.
This paper presents a systematic literature research and review of ventilator systems developed during the COVID-19 pandemic. Peer-reviewed journal and conference articles published through January 16th, 2025 were screened, and eligible systems were classified by actuation principle. Performance criteria were derived from Emergency Use Authorization requirements and used to generate a score-based ranking for each class. Performance was analyzed within and across classes to identify the situations in which each actuation principle is most advantageous. As an indicator of study quality, we evaluated the testing modalities reported for each device. Valve-based ventilators emerged as the most mature class in terms of ventilation functionality and testing. An emerging class of bag-based systems performed remarkably well compared with established valve- and blower-based designs. Across all three classes, the most frequent shortcomings concerned oxygen dosage of the inspired gas and the implementation of monitoring and alarm functions. Finally, we provide recommendations on development processes, testing procedures, and mitigation of supply-chain vulnerabilities that may support ventilator development in future pandemics.
The objective of this study was to develop and characterize a novel low-cost, flexible sensor system for ground reaction force (GRF) measurements for biomedical applications. The system aims to provide GRF measurements across customizable areas up to 2 m2, suitable for integration into various medical and rehabilitation devices. The sensor system was constructed using multiple discrete resistive sensor modules. Each module had a quadratic shape and an edge length of 7.5 cm. The system utilized ESD packing-foam as resistive sensing material and conductive textile as electrodes. Measurements were conducted using an Arduino Nano microcontroller, a Wheatstone bridge circuit and analogue multiplexers. A demonstrator, integrating the sensor modules in a sports mat was built to show the functionality. The proposed system was capable of measuring forces up to 330 N. The sensor modules have an exponential force-resistance characteristic curve and showed inter-module and inter-day variability in the range of commercially available sensor systems' accuracy. The demonstrator enabled to visualize changes in weight distribution on its surface. The developed sensor system offers a reliable, flexible, and low-cost solution for GRF analysis in biomedical applications, providing data e.g. for rehabilitation feedback.
To evaluate a multimodal deep learning model integrating preoperative transvaginal ultrasound (TVUS)-based radiomics features and clinical indicators for predicting 1-year postoperative recurrence of endometrial polyps (EP) after hysteroscopic polypectomy. A total of 116 patients with pathologically confirmed EP were assigned to a training cohort (n=81) and validation cohort (n=35). Radiomics features were extracted from TVUS images, and deep learning features were obtained using ResNet-based networks. These features, with clinical variables, were combined to build a multimodal model. Feature selection in the training cohort used reproducibility filtering (intraclass correlation coefficient [ICC] >0.80), univariate analysis, Pearson correlation (|r|>0.90), and least absolute shrinkage and selection operator (LASSO) regression. Model performance was evaluated by receiver operating characteristic (ROC) curves, area under the curve (AUC), calibration, and decision curve analysis (DCA). The multimodal model achieved AUCs of 0.941 (95 % CI: 0.897-0.985) and 0.922 (95 % CI: 0.852-0.992) in training and validation cohorts, outperforming clinical (0.812, 0.791) and radiomics-only models (0.871, 0.843). DeLong tests were significant (p<0.05). DCA showed higher clinical net benefit. This multimodal model effectively predicts 1-year recurrence after hysteroscopic EP resection, supporting individualized postoperative management.