Bioengineering doi: 10.3390/bioengineering13050532

Authors: Misael Zambrano-de la Zambrano-de la Torre Sebastian Guzman-Alfaro Andrea Acuña-Correa Manuel A. Soto-Murillo Maximiliano Guzmán-Fernández Ricardo Robles-Ortiz Karen E. Villagrana-Bañuelos Jose G. Arceo-Olague Carlos H. Espino-Salinas Ana G. Sánchez-Reyna Erik O. Cuevas-Rodriguez

Recent advances in edge computing and Tiny Machine Learning (TinyML) have enabled the deployment of artificial intelligence models directly on microcontrollers with extremely limited computational and memory resources. In this context, this work presents the design, implementation, and validation of a real-time cardiac arrhythmia classification system based on a quantized one-dimensional convolutional neural network (1D-CNN), deployed on an 8-bit Arduino UNO microcontroller. The proposed system integrates end-to-end processing, including ECG signal acquisition using a low-cost AD8232 analog front-end, signal preprocessing, heartbeat segmentation, classification, and real-time visualization on an OLED display. The model was trained and evaluated using the MIT-BIH Arrhythmia Database, considering a reduced three-class problem (Normal, Ventricular, and Supraventricular) to meet the constraints of ultra-low-cost hardware deployment. Under benchmark conditions, the quantized model achieved an accuracy of 97.6%, with a memory footprint below 24 KB and an average inference time of 200 ms per heartbeat, enabling real-time operation on a resource-constrained microcontroller. Real-time experiments were conducted using signals acquired from healthy volunteers to validate system functionality, although no annotated ground truth was available for these recordings, and therefore no diagnostic performance was derived from them. The results demonstrate the feasibility of deploying lightweight deep learning models on ultra-constrained embedded systems using the TinyML paradigm, implemented using TensorFlow 2.15 and TensorFlow Lite. This work should be interpreted as a proof-of-concept platform that highlights the trade-off between classification performance and hardware limitations, providing a foundation for future development of low-cost cardiac monitoring technologies in resource-limited environments.

Bioengineering doi: 10.3390/bioengineering13050531

Authors: Haochen Tian Jiaxin Wang Shijie Guo Feng Cao Lei Liu

Wearable lower-limb exoskeletons can enhance mobility, reduce metabolic cost, and aid rehabilitation. Effective human-exo cooperation requires personalized assistance profiles that match biomechanical principles. Existing methods often rely on fixed curves, involve complex tuning, and lack biomechanical interpretability. To address this, we propose a “Physics-guided perception and physiology-driven optimization” approach. First, a Physics-guided Dynamic Fusion Model (PDFM) is proposed, which integrates Newton–Euler dynamics, LSTM, and NTM to estimate multi-plane hip joint moments without ground reaction forces, employing biomechanical models as complementary fusion factors rather than the embedded hard constraints used in conventional physics-informed neural networks (PINNs). The model achieved correlation coefficients of 0.938, 0.924, and 0.929, and relative root mean square error (rRMSE) values of 5.29%, 9.79%, and 5.61%, in the sagittal, coronal, and transverse planes, respectively. These results outperformed all single-network baselines across all three anatomical planes. Second, an assistance profile derived from estimated moments is individually optimized using Bayesian optimization based on multi-muscle sEMG. Compared to no-exo walking, the optimized system reduced target muscle loading by 49.31% and metabolic cost by 14.75%; relative to the pre-optimized profile, the reductions were 23.64% and 5.74%, respectively. This work provides a laboratory-validated framework for personalized hip exoskeleton assistance in healthy adults, establishing a foundation for future clinical translation.

Bioengineering doi: 10.3390/bioengineering13050530

Authors: LaBreesha Batey Enrique Jackson Changchun Zeng Selvum Pillay

Research into auxetic foams—materials with a negative Poisson’s ratio— is expanding, yet their integration into orthotics for diverse neuropathic conditions remains largely unexplored. This pilot study evaluates the feasibility of fabricating custom auxetic foam insoles and characterizing vertical ground reaction force (vGRF) trends across a heterogeneous cohort. In collaboration with the NASA/Marshall Space Flight Center, six participants, including five representing varied neuropathic etiologies and one healthy control, performed randomized walking trials under three conditions: barefoot, over-the-counter (OTC) insoles, and custom auxetic prototypes. The healthy control was retained in the cohort-level analysis to preserve methodological symmetry across experimental conditions. To maintain physical rigor, vGRF data were mass-normalized (N/kg). A Friedman test (n = 6) evaluated global differences, supplemented by a dual-bootstrap analysis (1000 resamples) to quantify effect magnitudes (r) and numerical uncertainty. Although the Friedman test revealed no statistically significant global differences (Q = 0.333, df = 2, p = 0.846), a descriptively large effect size (r = 0.58) was observed for the auxetic material versus barefoot walking. However, wide 95% bootstrap confidence intervals prevent population-level inference, reinforcing the exploratory nature of these findings. Subject-specific observations showed descriptive differences in vGRF in three participants (0.17 to 1.18 N/kg), while increases in others occurred alongside confounding factors such as self-selected walking velocity. This work demonstrates the mechanical application of auxetic insole prototypes, providing a foundational rationale for future trials utilizing standardized walking velocity to isolate material performance.

Bioengineering doi: 10.3390/bioengineering13050529

Authors: Şükrü Demir Mehmet Vural Buğra Can Fatih Demir Abdulkadir Sengur

Background/Objectives: Osteonecrosis of the femoral head (ONFH) is difficult to diagnose, particularly in the early stages, because radiological findings may be subtle. Delayed or inaccurate staging may increase the risk of femoral head collapse and functional loss. Although magnetic resonance imaging is highly sensitive for early-stage lesion detection, interpretation may vary depending on observer experience. Therefore, reliable and explainable automated decision support approaches are needed. Methods: In this study, a deep learning-based approach was proposed to classify ONFH into early and late stages according to the Ficat–Arlet staging system. Stage I–II cases were defined as early-stage, whereas Stage III–IV cases were defined as late-stage. Axial and coronal MR images were evaluated separately to investigate plane-dependent classification performance. The images were converted into a three-channel format, resized to a common spatial resolution, normalized, and augmented during training. Feature extraction was performed using transfer learning with modern convolutional neural network architectures. ConvNeXt Tiny was used as the main classification backbone. Weighted loss was applied to reduce the effect of class imbalance, and the decision threshold was optimized on validation data to reduce missed clinically critical late-stage cases. Results: A dataset collected from the Orthopedics and Traumatology Department of Firat University Hospital was used in the experimental evaluation. The dataset was divided into training and test sets using an 80:20 split, and 10-fold cross-validation was additionally performed to assess model stability. In the hold-out test, the axial plane model achieved 94.51% accuracy, 96.80% sensitivity, 93.49% specificity, 0.9162 F1-score, and 0.981 AUC. In the coronal plane model, 92.84% accuracy, 96.13% sensitivity, 90.96% specificity, 0.9072 F1-score, and 0.988 AUC were obtained. The 10-fold cross-validation results provided a more conservative estimate of generalization performance. Conclusions: The findings indicate that deep learning-based plane-wise analysis of MR images can distinguish early- and late-stage ONFH with high performance. Grad-CAM-based visual explanations showed that the model focused mainly on clinically relevant subchondral and weight-bearing regions of the femoral head. The proposed approach may serve as an explainable decision support tool for reducing observer-dependent variability in clinical staging. Future studies should validate the method using external, multicenter datasets and paired patient-level axial–coronal images.

Bioengineering doi: 10.3390/bioengineering13050528

Authors: Jiameng Xu Huiquan Wang Shaofan Yang Xiang Liao Kuan Zhang Guang Zhang

Two-photon microscopy enables in vivo imaging of astrocytic Ca2+ activity, yet detecting weak, transient, and background-coupled signals remains challenging due to low signal-to-noise ratios and heterogeneous noise. Here, we propose a low-parameter, adaptive framework for detecting weak astrocytic Ca2+ signals in two-photon imaging. After short-window frame accumulation, static background suppression, and Gaussian smoothing to stabilize intensity statistics, signal candidates are identified via segment-wise Gaussian mixture modeling, temporal persistence masking, and adaptive threshold updates. In simulated videos, the proposed method improved the Dice coefficient from 0.06 to 0.77 and increased the reference SNR from −9.82 to 3.40 dB. In in vivo recordings, the local SNR increased from 5.58 to 7.28 dB. Compared with fixed thresholding, AQuA, and AQuA2, our method was more robust under high-noise conditions while requiring only three user-defined parameters (minimum area, minimum duration, and an initialization coefficient). This framework provides an interpretable and computationally practical front-end module for the robust extraction of astrocytic Ca2+ signal in low-SNR two-photon imaging.

Bioengineering doi: 10.3390/bioengineering13050527

Authors: Qiuping Ding Qiqi Tong Hongjian He Bin Gao Ling Xia

Background: Microstructural abnormalities of the corpus callosum (CC) are a consistent finding in schizophrenia, yet conventional diffusion tensor imaging (DTI) metrics provide limited biological specificity. Neurite orientation dispersion and density imaging (NODDI) can disentangle the neurite density index (NDI) and the orientation dispersion index (ODI), providing indirect, model-based markers of white matter microstructure in vivo. Methods: We applied NODDI to diffusion-weighted MRI data in patients with first-episode schizophrenia (FES) and matched healthy controls (HCs). The CC was used as a mask and subdivided into the genu (GCC), body (BCC), and splenium (SCC). Group differences in z-scores of the NDI and ODI were assessed using voxel-wise statistics within the CC and region of interest (ROI) analyses in the GCC, BCC, and SCC, controlling for age and sex. Associations between NODDI metrics and clinical symptoms were examined using the Positive and Negative Syndrome Scale (PANSS). Results: FES patients showed a significantly increased ODI in portions of the GCC, BCC, and SCC, as well as region-specific NDI alterations, with decreased NDI in parts of the SCC and increased NDI in sub-regions of the GCC/BCC (voxel-wise p < 0.05, FWE-corrected). ROI analyses confirmed a significant reduction in NDI z-scores in the SCC in FES patients compared with HCs (p = 0.009), whereas the ODI z-scores in the SCC did not differ significantly between groups (p = 0.124). Despite the absence of group-level ODI differences in the SCC, the SCC ODI was positively correlated with PANSS negative symptom scores in FES patients (r = 0.554, p = 0.002) and was also positively correlated with PANSS total scores in FES (r = 0.457, p = 0.014). This association remained significant in the region of the SCC after regressing out NDI from ODI (residual z_ODI), which was correlated with PANSS negative scores (r = 0.503, p = 0.006) and PANSS total scores (r = 0.474, p = 0.011), and the ODI/NDI ratio in the SCC was also correlated with negative symptom severity (r = 0.457, p = 0.014). Conclusions: Our findings suggest that, in the SCC, negative symptoms in schizophrenia are linked to altered neurite orientation dispersion under conditions of reduced neurite density. The dissociation between group-level NDI and ODI effects and their distinct relationship with psychopathology highlights the value of composite microstructural indices (e.g., residual z_ODI, ODI/NDI) for capturing clinically relevant white matter abnormalities.

Bioengineering doi: 10.3390/bioengineering13050526

Authors: Kotomi Inada Yuka Nozaki Takahiko Yamamoto

(1) Background: Breast cancer is the most prevalent cancer among women. Conventional screening method have drawbacks, including pain and radiation exposure. Microwave mammography has emerged as a promising diagnostic modality, and its development involves assessing equipment performance; however, ethical concerns limit its use on actual animals or humans. Therefore, an electromagnetic phantom mimicking the relative permittivity and conductivity of the human body has become crucial. (2) Methods: In this study, the electrical properties of a phantom were adjusted by modifying the material composition and additives based on a previous study. We used a network analyzer and dielectric probe to measure the electrical properties using the coaxial probe method. (3) Results: One issue with the existing phantom was the large average error rate in conductivity. Therefore, we increased the conductivity by adding sodium chloride (NaCl). Additionally, we investigated the effects of the amounts of cooking oil, TX-151, and detergent on the electrical properties to ensure a stronger correlation with target values. (4) Conclusions: The average error rates for the relative permittivity and conductivity were 8.26% and 16.9%, respectively, demonstrating an improvement in the agreement with the target values compared to the previous formulations.

Bioengineering doi: 10.3390/bioengineering13050525

Authors: Kahina Amara Oussama Kerdjidj Mohamed Amine Guerroudji Shadi Atalla Naeem Ramzan

Cardiovascular diseases are among the most prevalent global health conditions, making the accurate diagnosis and classification of cardiac abnormalities crucial for effective treatment and patient management. While the electrocardiogram (ECG) is the primary tool for assessing cardiac electrical activity, its manual analysis is often time-consuming and susceptible to interpretive error. To address these limitations, this work proposes a comprehensive deep learning pipeline for the automated classification of arrhythmias, incorporating specific strategies to mitigate the challenge of imbalanced datasets. Furthermore, we introduce a novel three-dimensional (3D) visualisation framework that provides interactive, anatomically precise renderings of the heart regions implicated by the ECG classification, thereby delivering enhanced diagnostic insight. Our evaluation demonstrates that the proposed data balancing techniques yield significant performance gains, and under our current experimental setup, the results are competitive with or exceed several previously reported methods. We acknowledge that a more rigorous inter-patient cross-validation is needed to fully establish generalisation. The resulting 3D visualisations not only enable precise anatomical localisation of arrhythmia substrates but also serve as a powerful interactive tool for clinical practice and medical education.

Bioengineering doi: 10.3390/bioengineering13050524

Authors: Jiarong Wu Jun Wu Qiuxia Zhang Wanli Zang

Accurate estimation of knee joint moment is important for biomechanical monitoring and injury-risk assessment, yet model generalizability under altered sensory environments remains unclear. This study evaluated a support vector regression model for predicting sagittal knee moment during the landing–takeoff cycle of the drop vertical jump (DVJ) under visuo-proprioceptive conflict and examined whether adding hip and ankle kinematics improved performance. Fourteen healthy men performed DVJs under one real and four virtual perturbation conditions with a fixed physical drop height and virtual heights of 0, 10, 30, and 50 cm. Bilateral surface electromyography and three-dimensional lower-limb kinematics were used as inputs, and the inverse-dynamics-derived sagittal knee moment served as the target. Basic and extended feature sets were compared under leave-one-subject-out (LOSO) and leave-one-condition-out (LOCO) frameworks. Within the present experimental design, prediction performance was generally higher under LOCO than under LOSO. Adding hip and ankle kinematics improved prediction mainly under LOCO, whereas gains under LOSO were limited. Waveform similarity showed a non-monotonic decrease-then-recovery pattern across perturbation levels. Residual analysis showed no directional bias, and errors were greater during landing absorption and push-off than during flight. These findings suggest that under the present study design and in this sample, lower performance was observed under LOSO than under LOCO, and that multijoint kinematics may improve prediction robustness under cross-condition settings.

Bioengineering doi: 10.3390/bioengineering13050523

Authors: Edgar Rafael Ponce de León-Sánchez Jorge Domingo Mendiola-Santibañez Omar Arturo Domínguez-Ramírez Ana Marcela Herrera-Navarro Alberto Vázquez-Cervantes Hugo Jiménez-Hernández José Alfredo Acuña-García Rafael Duarte-Pérez José Manuel Álvarez-Alvarado

Multiple sclerosis (MS) is an inflammatory disease of the central nervous system (CNS) that impacts nearly 3 million people worldwide. While the etiology and pathogenesis of MS are not yet fully understood, current evidence suggests that it results from complex interactions between genetic and environmental conditions. Clarifying the autoimmune mechanisms underlying MS remains a central objective in the development of effective therapeutic strategies. Interferon-beta (IFN-β) is one of the most frequently prescribed disease-modifying treatments for individuals with MS. However, despite its established efficacy, recent studies report that approximately 30–50% of patients exhibit inadequate response to IFN-β, largely due to genetic variability. Machine learning (ML), a branch of artificial intelligence (AI), employs data-driven computational models to enhance predictive accuracy and classification. In recent MS research, unsupervised learning techniques such as hierarchical clustering and K-means have been applied for classification purposes. However, these methods often fail to yield optimal solutions because they require numerous arbitrary decisions and perform adequately only when datasets contain clusters of similar sizes and lack significant outliers. Fuzzy systems (FSs) are designed to model complex, ambiguous real-world phenomena. In this study, an AI algorithm incorporating a fuzzy system, informed by expert neurologist input, is proposed to enhance the assignment of unknown class labels related to IFN-β response in MS patients. Additionally, a genetic algorithm (GA) is introduced to identify optimal solutions within the search space, facilitating hyperparameter optimization of a deep learning (DL) model trained with genetic biomarkers to identify patients likely to benefit from this therapy. Experimental results demonstrate that the fuzzy system achieved 80% classification efficiency, in contrast to 64% with conventional hierarchical clustering. Furthermore, an artificial neural network (ANN) model, with hyperparameters optimized by the GA, achieved an accuracy of 0.8–1.0, surpassing the multi-layer perceptron (MLP), which achieved 0.6–0.8 accuracy using conventional tuning methods.

Bioengineering doi: 10.3390/bioengineering13050522

Authors: Gabriela Bernadeta Orzeł-Łomozik Łukasz Łomozik Maciej Podolski Martyna Rożek Kalina Światlak Weronika Radwan Zuzanna Przybylska Paulina Michalska Maciej Pruski Katarzyna Mizia-Stec

Background: In recent years, artificial intelligence (AI) methods, including deep convolutional neural networks (CNNs), have gained increasing importance in supporting the automated analysis of echocardiograms. The aim of this study was to evaluate the impact of selected image artifacts—motion blur, acoustic shadowing, and speckle noise—on the performance of automatic classification of standard transthoracic echocardiographic (TTE) views using deep learning models. Methods: The analysis included 217 TTE video clips (2170 frames) covering apical views: two-chamber (A2C), three-chamber (A3C), four-chamber (A4C), and five-chamber (A5C). Two convolutional neural network architectures—ResNet-18 and ResNet-34—were applied, initialized with weights pretrained on the ImageNet dataset (transfer learning). In a limited comparative scope, EfficientNet-B0, a ViT model used as a frozen feature extractor combined with Logistic Regression, and a classical HOG + SVM model, were also included as reference methods. Classification performance was evaluated under conditions of controlled image degradation caused by motion blur, acoustic shadowing, and speckle noise. Results: All analyzed artifacts reduced classification performance, although the magnitude of this effect depended on artifact type. Speckle noise proved to be the most destructive, causing performance collapse across all evaluated methods at high severity. Motion blur and acoustic shadowing produced more differentiated degradation profiles. The ResNet models achieved the highest performance under reference conditions; however, after degradation, the ranking of models was no longer stable. In the comparative analysis, HOG + SVM showed the smallest relative performance loss under motion blur and the highest balanced accuracy under severe acoustic shadowing, whereas severe speckle remained critical for all models. Conclusions: Image quality degradation significantly impairs TTE view classification performance, and evaluation based solely on reference-quality images does not fully reflect model robustness to artifacts. These findings indicate the need to complement standard model evaluation with a structured robustness analysis under degraded imaging conditions and highlight the importance of training and validation settings that better reflect real clinical practice.

Bioengineering doi: 10.3390/bioengineering13050521

Authors: Zijun Huang Yu Li Jia Kou Shanqi Bao Ying Sun Li Lin

Habitat analysis enables spatial characterization of intratumoral heterogeneity; however, its application in nasopharyngeal carcinoma (NPC), particularly regarding metastatic lymph node (MLN), remains limited. This study aims to systematically compare the prognostic performance of various models using different sequences and spatial region combinations for predicting overall survival in NPC. The study retrospectively included 725 NPC patients (543 training, 182 testing). Habitat analysis was conducted based on T1, T1C, and T2 sequences in three regional strategies: primary gross tumor volume (GTVp), metastatic lymph nodes (MLNs), and the combined region of GTVp-MLN. The tumor area was divided into six subregions, and a multi-region spatial interaction (MSI) matrix was constructed to extract MSI features. On this basis, a radiomics model (R Model) and a clinical–radiomics model (CR Model) were established, and the model performance was evaluated using C-index and Kaplan–Meier survival analysis. The results show that the combined GTVp-MLN model based on the T1 sequence achieved the best overall predictive performance (R Model: C-index = 0.693; CR Model: C-index = 0.722). Significant survival differences were observed between the high- and low-risk groups. These findings suggest that habitat analysis incorporating the combined GTVp–MLN region may improve prognostic prediction and risk stratification in patients with NPC.

Bioengineering doi: 10.3390/bioengineering13050520

Authors: Lixia Zhu Xinyi Chen Xiaoling Li Thomas W. Leung Lawrence Ka Sing Wong Jack Jiaqi Zhang Yiao Liu Bin Luo Jianhang Du Yiliang Li Li Xiong

Background and purpose: Short-term external counterpulsation (ECP) noninvasively augments cerebral blood flow by elevating blood pressure in ischemic stroke. The current retrospective case–control study examined the effect of long-term ECP treatment on blood pressure and beat-to-beat blood pressure variability (BPV) in patients with recent ischemic stroke. Method: The ECP group included data from 20 recent ischemic stroke patients who received five daily 1 h sessions each week for seven weeks, for a total of 35 sessions of ECP treatment from our ECP registry. An equivalent comparative control group without ECP treatment was composed from the same pool of patients and matched with cases by sex and age. Beat-to-beat heart rate and blood pressure were monitored before and after the long-term intervention. Power spectral analysis calculated the beat-to-beat BPV oscillations at very low frequency (VLF; <0.04 Hz), low frequency (LF; 0.04–0.15 Hz), high frequency (HF; 0.15–0.40 Hz), and the total power spectral density (TP; <0.40 Hz) and LF/HF ratio. Result: There was a significant reduction in systolic blood pressure (SBP) after the intervention compared with that before intervention in both groups (p < 0.05), but only the ECP group displayed a statistically significant reduction in diastolic blood pressure (DBP) (p = 0.023). The changes in SBP and DBP (delta SBP and delta DBP) from pre-intervention to completion showed no significant differences between the two groups (all p > 0.05). The ECP group exhibited a more pronounced and significant decrease in each spectral component of BPV after the intervention than at pre-intervention, with a substantial decrease in systolic BPV at TP (p = 0.048) and in the LF/HF ratios (p = 0.021 in diastolic BPV and p = 0.004 in systolic BPV, respectively) compared to the control group. Conclusions: A standard 35-session ECP treatment decreases beat-to-beat BPV but does not change SBP and DBP in patients with recent ischemic stroke. This implies that long-term ECP treatment may enhance autonomic regulation to benefit post-stroke clinical outcomes.

Bioengineering doi: 10.3390/bioengineering13050519

Authors: Nahid Nuhash Zhang

Aortic arch aneurysms are uncommon but clinically significant due to their rapid growth and increasing rupture risk. Analyzing flow changes associated with aneurysm enlargement is essential for understanding mechanisms of disease progression. However, computational studies focusing on the aortic arch aneurysm remain limited. In this study, computational fluid dynamics (CFD) simulations were conducted under pulsatile flow conditions to investigate flow characteristics across different aneurysm sizes. A patient-specific aortic geometry was reconstructed and modified to generate three idealized aneurysm models with diameters of 45, 55, and 65 mm, along with a healthy reference model. Key hemodynamic parameters, including velocity distribution, flow recirculation, wall shear stress (WSS), oscillatory shear index (OSI) and helicity, were analyzed. The results demonstrated that increasing aneurysm size significantly disrupts normal flow patterns, leading to reduced flow velocities and progressively enhanced recirculation zones, particularly during the deceleration phase of the cardiac cycle. Enlarged aneurysms also exhibited consistently low WSS, elevated OSI, and disrupted helical flow patterns along the vessel walls. These adverse hemodynamic conditions are associated with intraluminal thrombus (ILT) formation, localized wall thinning, and increased risk of dissection or rupture. Overall, this study highlights the critical role of aneurysm size in shaping aortic arch hemodynamics and provides a computational framework for assessing disease progression and rupture potential.

Bioengineering doi: 10.3390/bioengineering13050518

Authors: Timothy A. Blenkinsop Anne Z. Eriksen

Background: Corneal organoids derived from pluripotent stem cells have emerged as powerful tools for studying corneal development, disease modeling, and regenerative medicine applications. While previous protocols have successfully generated corneal tissue structures, there remains a need for three-dimensional models that recapitulate the complex cellular architecture and diversity of native human cornea. Methods: We developed a modified spontaneous three-dimensional corneal organoid model using human embryonic stem cells (hESCs) through an adapted Self-formed Ectoderm Autonomous Multi-zone (SEAM) protocol. hESCs were cultured as spheroids in ultra-low-binding plates under normoxic conditions and differentiated over 7–8 weeks. Organoids were characterized using immunofluorescence staining for corneal-specific markers and single-cell RNA sequencing to assess cellular composition and gene expression patterns. Results: Approximately 20% of organoids developed transparent regions characteristic of corneal tissue by day 30 of differentiation. Immunofluorescence analysis revealed spatially organized expression of corneal markers, including ZO-1 and E-cadherin in the outermost epithelial layers, P63α-positive putative limbal stem cells at the epithelial–stromal interface, vimentin-positive stromal cells in the interior, and laminin-1 deposition that suggests Bowman’s membrane formation. The organoids expressed cornea-specific keratins (K3, K12, and K15) and the master regulator PAX6 in appropriate cellular compartments. Single-cell RNA sequencing identified 18 distinct cell clusters, including three corneal epithelium subclusters with differential expression of MUC16, KRT12, and ΔNp63α, two stromal populations with distinct inflammatory profiles, and a corneal endothelium cluster. Transcriptomic analysis confirmed expression of key corneal genes, including AQP3, CDH1, multiple keratins, mucins, and extracellular matrix components (HAS2, CD34, CD44, COL8A1, and KERA). Conclusions: This three-dimensional spheroid-based putative corneal organoid model successfully recapitulates the multilayered architecture and cellular diversity of human cornea, including stratified epithelium, putative limbal stem cells, stroma, and endothelium in spatially appropriate arrangements. The model demonstrates molecular signatures consistent with native corneal tissue and provides a valuable platform for studying corneal development, disease mechanisms, and potential therapeutic applications. Future optimization to improve organoid formation efficiency and functional maturation will enhance the utility of this system for both basic research and translational medicine.

Bioengineering doi: 10.3390/bioengineering13050517

Authors: Antonio Esquivel Herrera Liangju Kuang Mark Krauthammer Michael Bednar Eleftherios I. Paschalis Thomas H. Dohlman

This study evaluates how titanium and polymethyl methacrylate (PMMA) Boston Keratoprosthesis backplate substrates influence human corneal fibroblast proliferation, cytotoxicity, morphology, activation phenotype, and mechanotransductive signaling. Human corneal fibroblasts were cultured on titanium and PMMA, with tissue culture plastic or glass as controls. Proliferation was assessed over 7 days using metabolic assays, and cytotoxicity was measured by lactate dehydrogenase release. Cell morphology and surface coverage were examined by scanning electron microscopy, while immunofluorescence quantified fibroblast-specific protein 1 (FSP-1) and α-smooth muscle actin (α-SMA). Gene expression of α-SMA, collagen I, FSP-1, and focal adhesion kinase (FAK) was analyzed by quantitative PCR. Cells cultured on both substrates maintained stable viability with modest increases in estimated cell numbers and comparable proliferation curves, indicating preserved metabolic activity without growth suppression. Cytotoxicity remained low and similar between groups. SEM demonstrated broader and more continuous cell spreading on titanium, whereas cells on PMMA were more sparsely distributed. Immunofluorescence showed higher FSP-1 expression on titanium and increased α-SMA on PMMA. Gene expression analysis revealed higher FAK transcripts on PMMA, with no significant differences in α-SMA, FSP-1, or collagen I. These results confirm the cytocompatibility of both titanium and PMMA backplates with human corneal fibroblasts and support their use with the Boston Keratoprosthesis.

Bioengineering doi: 10.3390/bioengineering13050516

Authors: Ali Ari Pelin Atalan Efkere Ecem Yıldız Çangur Kamile Uzun Akkaya Berna Gurler Ari Bülent Elbasan Abdulkadir Sengur Yan Tian

Background/Objectives: Assessment of infant General Movements (GMs) is essential for early detection of neurological disorders such as cerebral palsy, but current methods depend on expert interpretation. This study proposes an automated and interpretable framework for infant movement classification using pose-based representations from RGB videos. Methods: A pose-driven pipeline was developed to extract 2D skeletal key points using a two-stage tracking strategy. Joint coordinates were normalized using the shoulder center and inter-shoulder distance. Videos were segmented into overlapping temporal windows, and each segment was represented using Pose-LBP histograms and motion ratio features. Classification was performed with a cost-sensitive subspace k-nearest neighbor ensemble (CSS-kNN-E). Performance was evaluated using stratified 10-fold cross-validation on a five-class infant movement dataset. Results: The proposed method achieved 99.16% (±0.48%) accuracy, 99.19% (±0.50%) sensitivity, 99.76% (±0.13%) specificity, and 99.23% (±0.48%) F1-score. The model demonstrated strong discrimination across classes and robustness to class imbalance. Conclusions: The framework provides an accurate and scalable solution for automated infant movement analysis. It reduces dependency on expert evaluation and has strong potential for early clinical screening and decision support.

Bioengineering doi: 10.3390/bioengineering13050515

Authors: Kandasamy Nagarajan ArulJothi Ramya Lakshmi Rajendran Byeong-Cheol Ahn Prakash Gangadaran

The study of regenerative medicine is an ongoing revolution in modern healthcare, with the aim of restoring the function of a diseased organ or a system through repair, replacing or regenerative modes, as opposed to conventional treatments that merely treat symptoms [...]

Bioengineering doi: 10.3390/bioengineering13050514

Authors: Naznin Sultana

Bone is a hierarchically organized composite material with unique mechanical properties and an intrinsic regenerative capacity that conventional repair strategies, including autografts, allografts, xenografts, and metallic or ceramic implants, fail to fully replicate due to donor scarcity, immunogenicity, mechanical mismatch, and poor long-term integration. Bone tissue engineering (TE) offers a biologically informed alternative by integrating osteoconductive scaffolds, osteogenic progenitor cells, and osteoinductive signaling molecules into a unified regenerative framework. Unlike existing reviews that evaluate these components in isolation, this review provides a mechanistically integrated analysis that repositions scaffold design as a biologically instructive platform whose topography, stiffness, porosity, and surface chemistry collectively govern cell adhesion, mechanotransduction, osteogenic differentiation, and extracellular matrix remodeling. Critically, it moves beyond cataloging materials and fabrication approaches to evaluate how specific scaffold features drive biological outcomes and to identify frequently understated limitations, including polymer-ceramic degradation kinetics and the inadequacy of small-animal models for clinical translation. By synthesizing advances in biomaterials, additive manufacturing, and smart scaffold technologies within this integrative framework, this review provides researchers and clinicians with a structured framework for evaluating emerging strategies and prioritizing future directions in functional bone regeneration.

Bioengineering doi: 10.3390/bioengineering13050513

Authors: Bernardo G. Collaco Srinivasagam Prabha Cesar A. Gomez-Cabello Syed Ali Haider Ariana Genovese Nadia G. Wood Narayanan Gopala Raghunath Raman Erik O. Hester Antonio Jorge Forte

Background: Agentic artificial intelligence (AI) systems, characterized by autonomous goal-directed behavior, multi-step reasoning, task decomposition, and tool use, are increasingly proposed for healthcare applications. However, their autonomy raises concerns regarding transparency, accountability, and human oversight. While explainable AI (XAI) has been widely studied in traditional predictive models, less is known about how explainability is implemented within agentic architectures. Objective: To map the emerging literature on explainable agentic AI (XAAI) in healthcare and characterize the types, scope, and forms of explainability used in these systems. Methods: A scoping review was conducted following PRISMA-ScR guidelines. PubMed, Embase, IEEE Xplore, and ACM Digital Library were searched through November 2025. Eligible studies described healthcare-related agentic AI systems incorporating explicit explainability mechanisms. Data were extracted on system architecture, explainability type (intrinsic, post hoc, hybrid), explanation scope (local, global), explanation form, and reported clinical outcomes. Results: Nine studies met the inclusion criteria. All systems demonstrated core agentic features, including autonomy, task decomposition, and tool integration, often within multi-agent frameworks. Explainability was predominantly intrinsic and workflow-native, typically delivered through textual reasoning traces and example-based grounding in retrieved clinical evidence. Feature-based and global explanations were comparatively rare and largely confined to hybrid architectures. Across domains including radiology, neurology, psychiatry, and biomedical research, XAAI systems were reported to improve performance and interpretability relative to baseline models in the included studies. However, these findings were derived from heterogeneous, predominantly experimental or retrospective studies, and structured human-in-the-loop oversight was infrequently described. Conclusions: Current XAAI systems appear to emphasize process transparency and evidence grounding rather than mechanistic model-level attribution. The available evidence remains limited and heterogeneous, and findings should be interpreted as early trends rather than established characteristics. Further progress will require standardized evaluation frameworks, clearer reporting of oversight mechanisms, and validation in real-world clinical settings to support safe and trustworthy integration of agentic AI into healthcare practice.

Bioengineering doi: 10.3390/bioengineering13050512

Authors: Tao Huang Yihang Jiang Xiaomei Wu

Background: Atrial electrical activity in patients with persistent atrial fibrillation (PeAF) is extremely complex, and identifying optimal ablation targets during ablation procedures remains a significant challenge. This manuscript aims to identify spatiotemporally stable highest dominant frequency (HDF) in the left atrium to provide a reliable basis for Electrophysiologists to locate ablation targets beyond pulmonary veins. Methods: Filtering and spectral estimation were performed on the left atrial intracardiac electrogram (LA-EGM) of PeAF patients to recognize the dominant frequency (DF). Spatiotemporally stable DF features within the left atrium were extracted using spatial aggregation and others to construct a 3D DF distribution model. HDF areas were automatically identified based on personalized thresholds derived from the patient’s DF distribution. Results: Data analysis of 43 PeAF patients demonstrated that spatial aggregation with 2 mm voxel size accurately constructs spatiotemporally stable DF distribution models. The proposed DF model enables the automatic identification of stable HDF areas in PeAF patients. In retrospective clinical cases, 72.1% of patients underwent ablation at these identified sites with effective therapeutic outcomes. Conclusion: Recurring HDF areas during PeAF serve as potential ablation targets. The results of this study provide a reliable basis for determining personalized ablation targets for PeAF patients.

Bioengineering doi: 10.3390/bioengineering13050511

Authors: Eric Macias-Fassio Aythami Morales Cristina Pruenza Julian Fierrez Carlos Espósito

Background: Machine learning-based Artificial Intelligence (AI) models have shown significant potential in the biomedical field, offering promising advances in diagnostics, personalized medicine, and patient care. However, to build these models, we have to deal with important challenges, including (1) the scarcity and low quality of available datasets in many important applications and (2) privacy concerns associated with sensitive patient data. Synthetic data (SD) generation has emerged as a promising strategy to address these challenges, yet many existing approaches struggle to simultaneously preserve privacy and accurately model tabular data, the predominant format in healthcare. Methods: We propose Kernel Density Estimation–K-Nearest Neighbors (KDE-KNN), a privacy-aware tabular data generation method, and evaluate its performance against state-of-the-art techniques. Using sepsis detection as a real-world case study, we assess both data utility and privacy protection. Results: Models trained on KDE-KNN-generated SD outperformed those trained on real data across both internal testing and external validation. In particular, a support vector machine achieved superior performance when trained on SD relative to real data. This gain is likely driven by the balanced class distribution of the synthetic dataset, underscoring KDE-KNN’s utility as an effective data balancing strategy. Consistent performance in external validation further supports the robustness and generalizability of the proposed approach. Privacy evaluation indicated a lower re-identification risk, with a mean distance to closest record of 4.971 between synthetic and real samples, compared with 2.715 among real samples. Conclusions: KDE-KNN effectively captures underlying population distributions while generating high-quality SD that preserve statistical fidelity and protect sensitive information. By balancing the trade-off between utility and privacy, the method produces representative datasets without exposing individual records. These findings position KDE-KNN as a valuable tool for data-scarce and privacy-sensitive applications, with broad potential across healthcare and other data-driven domains.

Bioengineering doi: 10.3390/bioengineering13050510

Authors: Emily Vivianne Freitas da Silva Carolina Lucena e Ortiz Marina Silveira Gomes Wendy Julliet Alvarado Baldeon Condor Karina Felix Santos Savio José Cardoso Bezerra Paulo Francisco Cesar Natalia Almeida Bastos-Bitencourt Sandro Basso Bitencourt Blanca Liliana Torres Léon

This study evaluated the physical and mechanical properties of acrylic resins used for interim fixed prostheses, with and without metal reinforcement, before and after aging. A total of 138 samples were divided into three groups: VIPI + Wire (control), VIPI, and Diamond D. Samples were assessed for microhardness, porosity, roughness, and flexural strength. Aging was simulated using 500 thermocycling cycles at 5 and 55 ± 1 °C. Data were analyzed using ANOVA and Tukey’s test. Group Diamond D did not fracture during flexural testing, but it exhibited significantly lower microhardness at both baseline and after aging. Before aging, Group Diamond D had higher roughness than Group VIPI, which exhibited greater porosity. Aging increased the microhardness of Group VIPI and the roughness of Group Diamond D. The percentage of porosity decreased significantly for Groups VIPI + Wire and VIPI, and pore size was reduced in all groups. Based on the results obtained from Diamond D material, this resin does not meet the required properties for the proposed indication for temporary fixed prostheses, whereas VIPI with reinforcement showed superior properties and greater stability after aging.

Bioengineering doi: 10.3390/bioengineering13050509

Authors: Masuma Firoz Neloy Shome Noah Wong Prisha Jonnalagadda Hari Tunga Amirmohammad Shafiee Amirmahdi Shafiee Sohan Bobba Karanjit S. Kooner

Retinal ganglion cell (RGC) degeneration underlies glaucomatous optic neuropathy and remains a leading cause of irreversible vision loss worldwide. Although elevated intraocular pressure (IOP) is the primary modifiable risk factor, RGC death reflects converging mechanisms including mechanical stress, vascular insufficiency, metabolic dysfunction, and neuroinflammation. We conducted a PRISMA-guided systematic review with PICOS-defined eligibility criteria, searching PubMed, Cochrane Library, ScienceDirect, Scopus, Google Scholar, and ProQuest for studies through January 2026 on RGC degeneration and neuroprotective or regenerative therapies in glaucoma. Included studies supported OCT-based structural assessment and imaging biomarkers as essential tools for early detection, risk stratification, and monitoring of progression and treatment response. Continued RGC loss despite IOP control in many patients highlights the need for mechanism-based interventions; neuroprotective strategies targeting excitotoxicity, oxidative stress, mitochondrial dysfunction, and neurotrophic insufficiency are emerging, while stem cell and gene-based regenerative therapies remain under active investigation. Integrating molecular insights with advanced imaging and biomarker-guided endpoints may enable earlier, more individualized intervention and help explain progression despite adequate pressure control.

Bioengineering doi: 10.3390/bioengineering13050508

Authors: Chengjie Huang Changqing Li Zhihai Su Qiwei Guo Quan Wang Tao Chen Yuhan Wang Zhen Yuan Hai Lu

Background: Lumbar disc herniation (LDH) is the most common etiological cause of low back pain (LBP). Objective and precise pain evaluation is of significant clinical value. Functional near-infrared spectroscopy (fNIRS) as a noninvasive neuroimaging modality, has been increasingly validated to reflect subjective pain perception through hemodynamic correlates. This study aimed to analyze the fNIRS changes in patients with LDH about to receive Unilateral Biportal Endoscopy and to further explore the feasibility of fNIRS as an objective biomarkers for clinical assessment of LDH. Methods: Resting-state fNIRS data were acquired from 67 preoperative LDH patients and 20 healthy controls (HC). Brain functional maps—including z-standardized fractional amplitude of low-frequency fluctuations (zfALFF) and seed-based functional connectivity (FC)—were extracted and quantified. Group-level comparisons were performed between LDH and HC groups across four predefined regions of interest; additionally, correlation analyses were conducted between fNIRS metrics and clinical assessment scores within the LDH cohort. Results: Compared with HC, LDH patients exhibited significantly altered zfALFF in the medial prefrontal cortex (mPFC): decreased amplitude at channel CH12 (t = −2.031, p = 0.045) and increased amplitude at CH21 (t = 2.462, p = 0.016). Whole-brain FC analysis further revealed widespread changes—particularly between the parietal somatosensory cortex and prefrontal regions. Among all tested FC–clinical indicator associations, 56 reached statistical significance after FDR correction (q < 0.05). VAS_ lumbar and SF-36_SF exhibited the highest number of significant connections. Conclusions: LDH patients with LBP exhibit notable alterations in prefrontal resting-state ALFF and FC between the parietal somatosensory cortex and prefrontal cortex relative to HC. Importantly, these neural alterations exhibit significant associations with both pain severity (VAS) and long-term health-related quality of life (SF-36), thereby strengthening their candidacy as neural correlates meriting prospective validation as objective, mechanism-informed biomarkers for clinical evaluation of lumbar disc herniation (LDH). Moreover, these findings highlight candidate neural targets for future longitudinal studies investigating early prognostic prediction and treatment response monitoring in LDH.

Bioengineering doi: 10.3390/bioengineering13050507

Authors: Taro Mukaibo Takafumi Watanabe Ayako Miura Kanna Saimoto Misaki Matsuo Hiromichi Ogusu Chihiro Masaki Hiroshi Ikeda

Poly (ether ether ketone) (PEEK) has attracted increasing attention for dental applications because of its favorable mechanical properties, physicochemical stability, and biocompatibility. However, its inherently poor bonding characteristics remain a major limitation in clinical practice. This study investigated the effect of sequential alumina airborne-particle abrasion (sandblasting) followed by plasma treatment on the bonding performance of methyl methacrylate (MMA)-based luting systems to dental CAD-CAM PEEK. PEEK specimens were prepared as plates and divided into four surface-treatment groups: untreated, airborne-particle abraded, plasma-treated, and airborne-particle abraded followed by plasma treatment. Surface characteristics were evaluated using SEM–EDX analysis and surface roughness measurements, and surface wettability was assessed by contact angle measurements using primers from two MMA-based luting systems (Beautylink [BL] and Super-Bond [SB]). Shear bond strength (SBS) between treated PEEK and each luting system was determined after 24 h of water storage (initial) and after 20,000 thermocycles (aged). Airborne-particle abrasion significantly increased surface roughness, whereas plasma treatment enhanced surface wettability without altering roughness. The combined treatment resulted in the highest surface roughness and the lowest contact angles and demonstrated superior or comparable SBS compared with the single treatments. After aging, the combined treatment significantly improved bonding durability. These findings indicate that airborne-particle abrasion followed by plasma treatment enhances the bonding performance and durability of MMA-based luting systems to PEEK.

Bioengineering doi: 10.3390/bioengineering13050506

Authors: Abdulrahman Alabduljabbar Tallha Akram Youssef N. Altherwy Muhammad Adeel Akram Imran Ashraf

Explainable Artificial Intelligence (XAI) has become a critical requirement in medical image analysis, where transparency and interpretability are essential for clinical trust and decision support. Melanoma is recognized as one of the most deadly types of skin cancer, with its occurrence exhibiting an increasing pattern in recent times. However, detecting this cancer in its initial stages greatly increases patients’ chances of long-term survival. Various computer-based techniques have recently been proposed to diagnose skin lesions at their early stages. Even though the machine learning community has achieved a certain degree of success, there is still an unresolved research challenge regarding high error margins and the limited interpretability of automated systems. This study focuses on addressing both segmentation and classification tasks, with particular emphasis on two key concepts: (1) improving image quality to maximize distinguishability between foreground and background regions, thereby enhancing visual interpretability and segmentation accuracy and (2) eliminating redundant and cluttered feature information to generate the most discriminative and compact feature representations. The input images are initially processed using a novel metaheuristic contrast-stretching method to estimate image-specific key parameters, thereby enhancing lesion boundary clarity in a clinically interpretable manner. Following this, the improved images are fed into selected pre-trained deep models, including DenseNet-201, Inception-ResNet v2, and NASNet-Mobile. The extracted features from all pre-trained models are fused to produce resultant vectors, which are then refined using a bio-inspired feature selection method, termed entropy-controlled whale optimization, to retain only the most informative attributes. The selected discriminative feature set is subsequently classified using multiple classifiers. The results indicate that the proposed framework achieves superior performance compared to existing methods in terms of accuracy, sensitivity, specificity, and F1-score. Additionally, it facilitates a more explainable, transparent, and structured diagnostic pipeline appropriate for medical applications.

Bioengineering doi: 10.3390/bioengineering13050505

Authors: Maximilian Lackner Arabi Sivanesapillai Dirk Holtmann

C1 biotechnology—bioprocesses that valorize one-carbon feedstocks such as CO2, CO-rich gases (blast furnace gas or synthesis gas), CH4 and CH3OH—has evolved from laboratory curiosity to industrial reality. In the quest to de-fossilize the chemical industry, the circular bioeconomy is widely seen as a solution. However, today it is still mostly based on primary agricultural feedstocks. Compared to thermochemical and catalytic processes, bioprocesses (fermentations) are carried out at ambient conditions, achieve high selectivities and good productivities. By decoupling fermentation from sugar-based substrates, gas fermentation of C1 substrates offers a scalable technology platform for producing biofuels, bioplastics, bio-based building blocks and alternative proteins, to name a few large-volume products. C1 platforms enable a circular, resource-efficient and virtually feedstock-independent bioeconomy that directly supports multiple United Nations Sustainable Development Goals (SDGs). In this article, we analyze the current technological landscape and discuss the (potential) impact of C1 routes on key SDGs using recent research advances and commercial case studies.

Bioengineering doi: 10.3390/bioengineering13050503

Authors: Woshington Valdeci de Sousa Rodrigues Armando Luz José Denes Lima Araújo João Diniz Antonio Oseas

The automatic classification of ISUP (International Society of Urological Pathology) grade groups in prostate histopathological images remains challenging due to the high similarity between adjacent classes, class imbalance, and label noise. In this work, we propose a deep learning pipeline based on EfficientNet convolutional neural networks combined with a hybrid loss function that integrates ordinal regression and Focal Loss to better capture the ordered nature of ISUP grades. A noise-filtering strategy based on the entropy of predictions from multiple EfficientNet models was first applied to identify and remove high-uncertainty samples from the training set. The problem was then reformulated as an ordinal regression task to explicitly model the hierarchical relationship among grades. Experiments conducted on the PANDA dataset demonstrate that removing noisy samples improved performance from κ=0.826 to κ=0.833. Incorporating ordinal loss further increased performance to κ=0.851. The best configuration, combining ordinal regression and Focal Loss, achieved κ=0.857 and an accuracy of 0.669, while reducing severe misclassifications and concentrating errors among adjacent classes. These results indicate that explicitly modeling ordinal structure and mitigating label noise are effective strategies for improving prostate cancer grading systems.

Bioengineering doi: 10.3390/bioengineering13050504

Authors: Angelo Michele Inchingolo Alessio Danilo Inchingolo Filippo Cardarelli Francesco Inchingolo Daniela Di Venere Elisabetta de Ruvo Laura Ferrante Grazia Marinelli Andrea Palermo Gianna Dipalma

Background: This retrospective case series evaluated the descriptive clinical observations of the bio-activator AMCOP® TC in the treatment of growing patients with Class III dento-skeletal malocclusion. In recent years, elastodontic appliances have been introduced as an evolution of conventional functional appliances. Elastodontic therapy could be an excellent therapeutic alternative in the early treatment of patients with Class III dento-skeletal malocclusion. Aim: This retrospective experimental study evaluated the descriptive clinical observations of the bio-activator AMCOP® TC in the treatment of patients with Class III dento-skeletal malocclusion and described four clinical cases. Materials and methods: The study included 11 subjects (5 males and 6 females, aged between 3 and 12 years) treated with the AMCOP® TC bio-activator for Class III dento-skeletal malocclusion. Patients used the AMCOP® TC device for two hours in the afternoon and all night for 6–8 months and then only at night. For each patient, cephalometric analyses were performed on latero-lateral teleradiographs both at the beginning of treatment (T0) and at the end of treatment (T1). Analyses were performed using DeltaDent® software. Conclusions: Cephalometric observations between T0 and T1 showed changes in sagittal relationship parameters, including ANB values; however, these findings should be interpreted cautiously. Elastodontic therapy with an AMCOP® TC appliance improved the correction of a Class III dento-skeletal malocclusion and postural restoration of the first cervical vertebrae. Although further studies are needed, AMCOP® TC bio-activators may be considered a possible interceptive treatment approach in selected growing patients; however, the present findings should be interpreted with caution. Findings should be considered preliminary and interpreted with caution.

Bioengineering doi: 10.3390/bioengineering13050502

Authors: Vinicius de Moura Silva Lima Eduarda Faria Arthur Rafaela Rodrigues Dousseau Gonzaga Luan Faria Diniz Rodrigo Cunha de Mello Pedreiro Osmar Pinto Neto

Transcranial direct current stimulation (tDCS) over motor–premotor regions may modulate motor performance, though underlying mechanisms remain unclear. Twenty-four athletes (9 females, 15 males) were randomly assigned to receive anodal tDCS (2 mA, 20 min) over the left sensorimotor cortex (n = 12) or sham stimulation (n = 12). Participants performed a bimanual isometric force-matching task at 30% maximal voluntary contraction, with visual feedback initially provided and then removed. Force undershoot, root mean square error (RMSE), spectral power (1–3 Hz), and inter-hand coherence were analyzed. A computational model was developed to test whether enhanced proprioceptive feedback processing could account for observed effects. Following tDCS, force undershoot decreased significantly (p = 0.002, d = −1.15) and RMSE improved (p = 0.010, d = −0.91). Spectral power in the 1–3 Hz band increased (p = 0.012, d = 0.87), suggesting enhanced corrective oscillations. These within-group changes were absent in the sham group (all p > 0.20), although Group × Epoch interactions did not reach significance (all p > 0.05), likely due to limited statistical power. Inter-hand coherence remained unchanged. The computational model demonstrated that enhanced proprioceptive feedback gain qualitatively reproduces the observed behavioral pattern. Anodal tDCS over the left sensorimotor/premotor region may enhance bimanual force control under conditions requiring proprioceptive feedback. Replication with larger samples is needed to confirm between-group specificity.

Bioengineering doi: 10.3390/bioengineering13050501

Authors: Ahmad Z. Rao Michael Houston Hao Meng Shengai Li Yingchun Zhang Sheng Li

Breathing-controlled electrical stimulation (BreEStim) is an innovative neuromodulation intervention that synchronizes deep voluntary breathing with peripheral electrical stimulation. Prior studies have shown its analgesic effects in healthy adults and spinal cord injury patients with neuropathic pain. The present study used EEG to examine BreEStim’s neural effects on sensory, affective, and cognitive components of pain. Fourteen healthy participants (7 M, 7 F) completed 30 min of BreEStim and conventional electrical stimulation (EStim) interventions in a randomized, crossover within-subject design. Electrical pain thresholds (EPT) and EEG were recorded pre- and post-intervention. Event-related potentials (ERPs) at pre-EPT-level stimuli before and immediately after each intervention were analyzed for early sensory (P30) and affective (P250) processing, while resting-state EEG assessed spectral power across delta, theta, alpha, and beta bands for cognitive processing. Both BreEStim and EStim increased EPT, indicating short-term habituation. There was no change in early ERP responses (P30) after each intervention, suggesting preserved sensory perception. BreEStim selectively reduced P250, reflective of the affective component of pain. BreEStim significantly increased delta and theta band power and reduced alpha band power on resting-state EEG analyses, whereas no significant changes after EStim were observed. Collectively, BreEStim preserves sensory encoding while selectively modulating affective and cognitive dimensions of pain, supporting its potential as a targeted, non-pharmacological neuromodulation strategy.

Bioengineering doi: 10.3390/bioengineering13050500

Authors: Fan Zhang Zhaoyan Wang Dou Ye Jialan Liang Hui Yang Suqing Qu Qian Wang Zuo Luan

Although transplantation of induced oligodendrocyte progenitor cells (iOPCs) is a promising strategy for white matter injury, the therapeutic efficacy of in vitro-generated iOPCs remains limited due to insufficient differentiation potential. Here, we aimed to identify key transcription factors and small-molecule drugs to optimize iOPC quality. Through transcriptome sequencing and bioinformatics analysis, we identified the transcription factor SOX10, which is differentially expressed between endogenous fetal OPCs and exogenous iOPCs. We established lentivirus-mediated SOX10 overexpression in neural stem cells (NSCs) before iOPC induction and performed cellular assays and multi-omics analysis. Early SOX10 overexpression reduced cell migration but promoted maturation into oligodendrocytes and suppressed astrocyte differentiation. Multi-omics analyses revealed that SOX10 overexpression is associated with the extensive redistribution of SOX10 chromatin binding and enrichment of regulatory programs linked to oligodendroglial differentiation, including the activation of the key signaling downstream transcription factors JUN/FOS. Moreover, TSA, Dabrafenib, and Fedratinib effectively upregulated SOX10 and improved iOPC differentiation. This study identifies SOX10 as a core upstream regulator governing the fate of iOPCs, providing a potential strategy for optimizing iOPC induction for future investigation of white matter injury therapy.

Bioengineering doi: 10.3390/bioengineering13050499

Authors: Christian Said Christopher Hayward Michael Stevens Gabriel Matus Vazquez Laurence Boss Ricardo Deveza Sumita Barua Kavitha Muthiah Pankaj Jain

Venoarterial extracorporeal membrane oxygenation (VA-ECMO) has failed to demonstrate mortality benefit in randomised controlled trials of cardiogenic shock. We aimed to determine whether a novel ‘Hemi-ECMO’ configuration, involving aortic occlusion to isolate the left ventricle from the VA-ECMO circuit, improves cardiac haemodynamics. We utilised a pulsatile biventricular mock circulatory loop with variable contractility to compare standard VA-ECMO with Hemi-ECMO support under left ventricular or biventricular failure conditions. When averaged across all pump speeds, mean left atrial pressure was significantly reduced with Hemi-ECMO compared to VA-ECMO (21.11 ± 1.32 mmHg vs. 26.53 ± 0.87 mmHg, p < 0.001), with more pronounced benefit at higher pump speeds. Aortic ejection increased with Hemi-ECMO at higher pump speeds: 0.14 ± 0.03 vs. 0.00 ± 0.00 L/min (p = 0.002) at 3000 revolutions per minute (RPM). Aortic ejection was greater with Hemi-ECMO in the descending aorta compared to the ascending aorta position (0.27 ± 0.03 L/min vs. 0.17 ± 0.05 L/min, p = 0.015). In conclusion, Hemi-ECMO demonstrates significant haemodynamic advantages in severe cardiogenic shock, including reductions in mean left atrial pressure and increases in aortic ejection, with greater benefits when positioned in the descending aorta. Further in vivo studies are warranted to assess clinical viability.

Bioengineering doi: 10.3390/bioengineering13050498

Authors: Sebastian Hetzler Stefan Rues Andreas Zenthöfer Peter Rammelsberg Reinald Kühle Christopher J. Lux Ralf Erber Christoph J. Roser

Additive manufacturing enables the fabrication of patient-specific zirconia devices with integrated surface features; however, the biological effects of meso-scale topographies remain insufficiently understood. This in vitro study evaluated the influence of defined meso-scale surface modifications on osteoblast behavior using Digital Light Processing (DLP)-fabricated 3Y tetragonal zirconia polycrystal (3Y-TZP) and 5Y partially stabilized zirconia (5Y-PSZ). Planar control specimens and surfaces incorporating regularly distributed columnar structures (height: 100 µm; width: 40 µm; center-to-center spacing: 80, 120, and 160 µm; Mod-80, Mod-120, Mod-160) were fabricated and characterized after sintering. Cytotoxicity was assessed by elution testing and showed cell viability >98% for all groups. Osteoblast adhesion and proliferation (hFOB 1.19) were quantified using metabolic assays. Meso-scale modifications significantly increased early cell adhesion compared to planar controls (p < 0.05), with the strongest effect observed for Mod-160. No significant differences in proliferation rates were detected between groups (p > 0.05). Osteogenic differentiation was evaluated by RT-qPCR (RUNX2, ALPL, COL1A1, BGLAP), revealing material- and geometry-dependent responses. On 3Y-TZP, meso-scale structures, particularly Mod-160, were associated with sustained upregulation of BGLAP, whereas 5Y-PSZ exhibited less pronounced effects. Within the limitations of this in vitro study, meso-scale surface structuring of additively manufactured zirconia enhances early osteoblast adhesion without affecting proliferation and may influence osteogenic differentiation in a material-dependent manner.

Bioengineering doi: 10.3390/bioengineering13050497

Authors: Runze Li Zhuyi Shen Chengkai Wu Jingsong Li Yu Tian

Clinical prediction from electronic health records (EHRs) is complicated by pervasive missingness and label scarcity, which make performance sensitive to the match between data conditions and pipeline choice. Choosing the best pipeline for a new incomplete dataset still requires costly trial-and-error. We cast this as an algorithm selection problem and address two bottlenecks—instance scarcity and distance quality—that have so far prevented meta-learning from reaching clinical settings. Graph neural networks offer diverse strategies (patient similarity networks, bipartite imputation graphs, attention-driven feature interaction), yet no single architecture dominates across missingness patterns, and selecting the best pipeline for a new dataset remains a trial-and-error approach. Formal algorithm selection could automate this choice but requires many characterized meta-instances—more than clinical settings typically provide. We propose two solutions: (1) constructive instance augmentation, applying controlled quality perturbations (MCAR and MNAR missingness injection, label trimming) to 20 base EHR datasets to expand the meta-knowledge base to 83 characterized meta-instances, each described by a 10-dimensional missingness fingerprint, without additional model training; and (2) dynamic-supervised metric learning, using differential evolution to optimize fingerprint feature weights so that static distances preserve method-performance similarity captured by dynamic fingerprints, which require model sweeps and are unavailable at deployment. Under base-dataset-level leave-one-dataset-out cross-validation over 21 pipelines, the resulting metric-learned kNN recommender attains the highest win rate (20.5%) among non-oracle strategies on the augmented store, selecting the correct pipeline more often than any fixed default. At deployment, the recommender needs only the 10-dimensional static fingerprint with pre-learned weights; no sweep data is required for new datasets. Cross-domain evaluation on 25 external subsets (colorectal cancer, kidney disease, MIMIC-IV) demonstrates framework modularity: when the fingerprint module is adapted (standard meta-features in place of the missingness-specific set), the recommender achieves regret of 0.025 (55% below random selection).

Bioengineering doi: 10.3390/bioengineering13050496

Authors: Ryunosuke Watanabe Shota Yasunaga Fumi Hirose Koshiro Shimasaki Tomohiro Yoshizawa Yasuhiro Homma Tomofumi Nishino Hajime Mishima Yoshihisa Harada

Proximal femoral fractures associated with osteoporosis are an important clinical problem, yet how bone quality independently influences stress distribution remains insufficiently understood. This study aimed to quantitatively compare surface stress distribution between normal and osteoporotic proximal femoral models using thermoelastic stress analysis (TSA). Fourth-generation composite femurs with identical external geometries were subjected to cyclic compressive loading at a 9° adduction angle, with different maximum loads applied to avoid structural failure (normal: 1900 N; osteoporotic: 1000 N). TSA was performed using an infrared lock-in system to obtain surface stress maps, and stress values were evaluated across key proximal regions and along the medial and lateral cortices. The osteoporotic group showed higher maximum stress values in the medial neck (−37.79 vs. −11.52 MPa), lateral neck (24.70 vs. 8.75 MPa), and intertrochanteric crest (−17.98 vs. −6.05 MPa), corresponding to approximately 1.8–3.5-fold increases compared with the normal model values normalized to 1000 N. Mean stress values were also higher by approximately 1.9–2.4-fold across regions. These results suggest that reduced bone quality is associated with increased proximal stress concentration. They may also help guide implant and fixation strategies, including stem selection and fixation configuration, by identifying regions susceptible to stress concentration under different bone quality conditions.

Bioengineering doi: 10.3390/bioengineering13050495

Authors: Luka Mejić Olja Šovljanski Milada Pezo Lato Pezo Tiana Milović Ana Tomić

In recent years, bacteria-based self-healing has emerged as a promising bioengineering strategy to address the self-repair of cracks in cement-based materials, which represent one of the persistent durability challenges. This approach relies on microbiologically induced calcium carbonate (CaCO3) precipitation (MICP), in which metabolically active bacteria promote CaCO3 formation of crystals that can heal cracks and restore material integrity. This study compares the self-healing potential of a natural (N-) alkaline soil Bacillus licheniformis strain with a UV-strain (phenotypic mutant) generated through controlled UV exposure followed by adaptive evolution. Both strains were evaluated under conditions relevant to cementitious environments. The UV-strain exhibited enhanced ureolytic performance, reaching urease activity of 0.32 U/mg compared to 0.24 U/mg in the N-strain. This translated into improved biomineralization, with CaCO3 precipitation reaching 2.37 mg versus 2.23 mg/100 mL in the N-strain. Additionally, the UV-strain showed increased cell hydrophobicity and aggregation, indicating improved nucleation potential and surface-mediated mineral deposition. Multivariate analysis confirmed strong correlations between ureolytic metabolism, alkalization, and mineral formation, while artificial neural network (ANN) modeling (MLP 6-10-14) successfully predicted biomineralization-related parameters with high accuracy (R2 > 0.90 for urease activity, NH4+, ΔpH, and CaCO3). The results demonstrate that UV-induced phenotypic adaptation can enhance biomineralization efficiency with minor trade-offs in physiological robustness. For the first time, that controlled UV-induced phenotypic adaptation can be used as a targeted strategy to enhance biomineralization efficiency in B. licheniformis, while maintaining functional stability under cement-relevant conditions. These findings provide a novel framework for tailoring bacterial performance in self-healing systems for construction biotechnology.

Bioengineering doi: 10.3390/bioengineering13050494

Authors: Andreas Heinrich

Identification of unknown individuals is challenging, and radiological imaging databases provide rich anatomical information for automated recognition. This study evaluated whether a single routine magnetic resonance imaging (MRI) slice contains sufficient person-specific features to identify individuals in large databases. It analyzed 11,078 head MRI examinations from 5770 individuals (age 52 ± 18 years, 2714 men) acquired between 2002 and 2025. For identification, 112 individuals were randomly selected across eight 10-year age groups, and one slice from four anatomical regions was extracted. The remaining 10,966 MRI examinations with 247,804 slices formed the reference database. Distinctive anatomical features were automatically extracted using computer vision (CV), and the identification rate was evaluated by rank. Using a single MRI slice, the identification rate at rank 1 reached 96% (107/112) for the best-performing region, the maxillary sinus, among 5770 potential identities. Across all regions, the rank 1 identification rate ranged from 91% to 96%; combining them increased rank 1 and 10 identification rates to 98% (110/112) and 99% (111/112). Identification rate remained stable over several years, with only two cases showing reduced rank 1 performance, likely due to age-related morphological changes. A single MRI slice contains stable, individualized features sufficient for reliable identification in large databases, supporting automated CV-based personal identification across years.

Bioengineering doi: 10.3390/bioengineering13050493

Authors: Md. Zihad Bin Jahangir Muhammad Ashad Kabir Sumaiya Akter Israt Jahan Minh Chau

Background: The goal of automated radiology report generation is to help radiologists in their task of creating descriptive reports from chest radiographs. However, the process of creating coherent and contextually accurate reports has been challenging, mainly due to the intricacies of medical language and the need to correlate visual data with textual descriptions. Methods: This study presents LLaMA-XR, a novel framework that integrates Meta LLaMA 3.1 Large Language Model with DenseNet-121-based image embeddings and Quantized Low-Rank Adaptation (QLoRA) fine-tuning. Results: The experiment conducted on the IU X-ray dataset demonstrates that LLaMA-XR outperforms a range of state-of-the-art methods. It achieves an ROUGE-L score of 0.433 and a METEOR score of 0.336, establishing new performance benchmarks in the domain. Conclusions: These results underscore LLaMA-XR’s potential as an effective artificial intelligence system for automated radiology reporting, offering enhanced performance.

Bioengineering doi: 10.3390/bioengineering13050492

Authors: Jie Wang Zongjie Wang

Translating preclinical findings into effective clinical cancer immunotherapies remains a major challenge, mainly because conventional in vitro and animal models often fail to capture the complexity, dynamics, and species-specific features of human immune responses. Organ-on-a-chip (OoC) technologies that combine engineered tissue architectures with precisely controlled microfluidic transport provide human-relevant microphysiological platforms for mechanistic studies of immune–tumor interactions and evaluation of therapeutic efficacy and immunotoxicity under defined microenvironmental conditions. However, immune responses involve time-dependent and interconnected processes, including immune cell trafficking, cytokine programs, metabolic shifts, and cytolysis, that are not adequately resolved by static or endpoint assays. Engineering immune-competent OoC systems therefore requires coordinated design of platform architectures, immune cell incorporation strategies, and integrated measurement workflows capable of capturing dynamic and state-dependent responses. In this review, we summarize engineering strategies for building immune-competent OoC platforms for cancer immunotherapy, focusing on platform architectures, immune cell incorporation methods, and fit-for-purpose assay workflows. Emphasis is placed on embedded sensing modalities (e.g., cytokine, oxygen, and impedance readouts) that provide valuable kinetic and state-variable data. Finally, we discuss key translational challenges, including reproducibility, standardization, and benchmarking, and outline near-term priorities to accelerate the adoption of immune-competent OoC systems in immunotherapy research and development.

Bioengineering doi: 10.3390/bioengineering13050491

Authors: Mirko Zisi Vincenzo Ricci Alessandro Rocchi Vincenzo Canali

Scoliosis is a three-dimensional spinal deformity that may affect musculoskeletal alignment, respiratory mechanics, and neuromotor control. Rigid thoraco-lumbo-sacral orthoses (TLSOs) remain the primary conservative treatment during skeletal growth. Most brace systems rely on three-point pressure mechanisms that primarily generate lateral compression forces, while the contribution of shear components to corrective biomechanics has been insufficiently quantified. This study presents the experimental and analytical validation of the Canali Front-Push Orthopedic Brace, a rigid orthotic system designed to generate controlled compressive and shear forces through a frontal thrust mechanism and anterior rib cage engagement. By applying anterior force, the device reduces the frontal-plane lever arm, thereby limiting the mechanical moment that contributes to transverse plane rotation. An instrumented four-segment torso model derived from the internal CAD geometry of the brace was developed to independently measure upper compression, lower compression, and intersegmental shear forces. Controlled misalignment conditions (0 mm, 2 mm, and 4 mm) were introduced to simulate asymmetric engagement of the orthosis. Three load cell configurations (200 N and 500 N capacity) were tested. Mechanical endurance of the rack–latch fastening system was also evaluated. A predictive shear–misalignment relationship was derived and experimentally validated. Peak compressive forces reached approximately 370 N, while shear forces increased from less than 40 N under symmetric alignment (D0) to approximately 170 N under maximal misalignment (D4). Shear activation demonstrated near-linear proportionality to imposed geometric asymmetry (R2 > 0.94). Following cyclic loading, the fastening system stabilized mechanically around 300 N. Measurement repeatability showed a coefficient of variation below 5%. These findings demonstrate that the brace produces predictable and controllable shear activation while maintaining high mechanical repeatability. The results provide a quantitative biomechanical framework for understanding shear-induced corrective mechanics in scoliosis bracing and support future studies integrating computational modeling and clinical validation. The proposed mechanical framework may contribute to the development of next-generation orthotic strategies aimed at controlling spinal rotation through vector modulation rather than purely compressive correction.

Bioengineering doi: 10.3390/bioengineering13050490

Authors: Philip Chung Yin Mak Luoyi Kong Lawrence Wing Chi Chan

The high complexity of radiation therapy for lung cancer necessitates effective planning of advanced treatments such as Volumetric Modulated Arc Therapy (VMAT) by radiation oncologists. The current VMAT treatment planning process typically involves extensive manual interaction and a time-consuming, trial-and-error, iterative approach that requires planners’ experience. This can lead to varying levels of plan quality. To improve the quality of radiotherapy treatment plans quickly and accurately, this research presents a new architecture, Enhanced UNet3D, to generate three-dimensional (3-D) dose distributions for lung cancer patients. Enhanced UNet3D utilises a symmetric encoder–decoder architecture with residual connections and a target region-attention module to achieve high accuracy in dose shaping within the PTV. A new composite objective function, Enhanced Combined Loss (ECLoss), that includes both SharpLoss, a structure-aware DVH-guided loss, and 3D gradient regularisation, has been developed to address voxel-level class imbalance and achieve realistic spatial dose falloff. This research utilised a retrospective dataset of 170 VMAT plans to train and validate the proposed model. On the test set (n = 14), the model demonstrated exceptional overall accuracy, with a Mean Absolute Error (MAE) of 0.238 ± 0.075 Gy and a structural similarity index measure (SSIM) of 0.970 ± 0.005. Moreover, the model can perform near-real-time inference at approximately 0.5 s per patient, representing a significant reduction in computational resources compared to other architectures. Therefore, these results demonstrate that the Enhanced UNet3D model with ECLoss is a clinically feasible tool for the rapid evaluation and quality assurance of radiotherapy treatment plans and may reduce the need for manual trial-and-error in VMAT workflows.

Bioengineering doi: 10.3390/bioengineering13050488

Authors: Xinyi Wei Xiaolong Wang Xin Yang Mingjing Qiao Yannan Chen Andre Hoerning Xianhu Liu Chenchen Ren

Pelvic organ prolapse (POP) is a common benign gynecological disorder that substantially affects quality of life, particularly in aging female populations. Current management strategies, including standardized vaginal pessaries and synthetic surgical meshes, are often limited by poor anatomical adaptability, mechanical mismatch with native pelvic tissues, and long-term safety concerns. These limitations have driven increasing interest in personalized and biomechanically compatible therapeutic solutions. Three-dimensional (3D) printing, also known as additive manufacturing, has emerged as a promising bioengineering technology to address these unmet clinical needs. By enabling layer-by-layer fabrication directly from digital models, 3D printing allows for precise control over device geometry, mechanical properties, and material composition, facilitating patient-specific design. This narrative review summarizes recent progress in 3D printing for POP management across three major application domains: (i) next-generation meshes based on biodegradable polymers, elastomeric materials, natural biomaterials, and hydrogel systems; (ii) customized vaginal pessaries tailored to individual pelvic anatomy using imaging-assisted workflows; and (iii) imaging-based pelvic models and prototype devices for surgical planning, education, and exploratory assessment. Overall, existing studies demonstrate that 3D printing enables improved biomechanical compatibility, enhanced tissue integration, and multifunctional device design, including drug delivery capability. Although current evidence is largely pre-clinical or based on pilot studies, additive manufacturing holds strong potential to advance POP management toward safer, personalized, and functionally optimized clinical solutions.

Bioengineering doi: 10.3390/bioengineering13050489

Authors: Xinbei Dong Qinglong Wang Zhipeng Huang Yucai Wang

Background: Postoperative anemia associated with intertrochanteric hip fracture is a detrimental complication that detrimentally impairs patients’ outcomes. This study is designed to develop an online predictive tool to assist physicians in developing surgical blood preparation strategies to prevent the occurrence of postoperative anemia. Methods: This study included data collected from June 2017 to June 2025 on intertrochanteric hip fracture patients at Tangdu Hospital, including demographic information, comorbidities, vital signs, and laboratory results. LASSO regression was used to select predictive variables, and seven machine learning techniques: Logistic Regression, Support Vector Machine, Decision Tree, LightGBM, XGBoost, Neural Networks, and Random Forest, were compared to identify the best tool for predicting postoperative anemia risk. We created a patient-specific risk prediction tool with SHAP-driven interpretability for clinical decision support. Results: A total of 815 patients were included in the analysis, of whom 208 (25.5%) presented with postoperative anemia. Eight variables were selected to build seven machine learning models. Among these, the SVM model exhibited the best predictive performance in terms of discrimination, calibration, and clinical applicability, with an AUC range of 0.827–0.831. In test sets encompassing diverse population characteristics, SVM achieved the highest sensitivity (72.73%), accuracy (77.78%), and F1 score (57.14%). Conclusions: We established an online prediction platform for clinical practice, enabling clinicians to assess anemia risk in intertrochanteric hip fracture patients and support early prevention of postoperative anemia.

Bioengineering doi: 10.3390/bioengineering13050487

Authors: Elvira Maranesi Roberta Bevilacqua Elisa Casoni Ilaria Barboni Federico Barbarossa Arianna Margaritini Chiara Polverigiani Arianna Sgolastra Emanuela Bertino Renato Baldoni Marco Benadduci Giulio Amabili Pietro Scendoni Giuseppe Pelliccioni Valentina Di Donna Giovanni R. Riccardi

(1) Background: Parkinson’s disease (PD) is a neurodegenerative disorder characterized by gait and postural impairments. Recently, physical activity has emerged as a key strategy in PD management. This study aimed to evaluate the effectiveness of an innovative technology-assisted rehabilitation program in improving gait and reducing fall risk in older adults with PD. (2) Methods: Fifty-eight patients were randomly assigned to three groups: conventional rehabilitation (CG), or conventional therapy combined with technology-assisted rehabilitation using Tymo (TG) or Walker View (WG). The intervention consisted of 10 sessions over 5 weeks. Assessments were conducted at baseline (T0), post-intervention (T1), and at 6-month follow-up (FW). Outcomes included gait and balance performance, fear of falling, quality of life, activities of daily living, and physical function. (3) Results: The CG showed no significant improvements, with a decline in Barthel Index from T1 to FW. The WG demonstrated significant improvement in POMA Gait scores, while the TG improved both POMA Total and Balance scores at T1. Post-treatment, TG and WG outperformed CG in POMA outcomes; however, these differences were lost at follow-up. (4) Conclusions: Technology-assisted rehabilitation can improve gait and balance in older adults with PD, although sustained or repeated interventions may be necessary to maintain long-term benefits.

Bioengineering doi: 10.3390/bioengineering13050486

Authors: Carlo Cosimo Quattrocchi Claudia Piervincenzi Raffaella Di Giacopo Donatella Ottaviani Maria Chiara Malaguti Chiara Longo Francesca Cattoi Nikolaos Petsas Loredana Verdone Micaela Caserta Sabrina Venditti Bruno Giometto Rossana Franciosi Federica Vaccarino Marco Parillo Tal Dotan Ben-Soussan

Parkinson’s disease (PD) may benefit from non-pharmacological motor–cognitive rehabilitation, but sensitive neuroimaging markers of training-related brain changes remain limited. This study investigated whether 4 weeks of daily Quadrato Motor Training (QMT) modulate resting-state functional connectivity (FC) in PD and secondarily explored whether whole-brain radiomic features derived from T1-weighted and fractional anisotropy (FA) images could detect pre–post differences over this short intervention interval. Fifty patients with idiopathic PD were randomized to QMT or a SHAM repetitive stepping condition, and 48 completed the protocol (25 SHAM, 23 QMT). MRI was acquired at baseline and after 4 weeks and included resting-state fMRI, 3D T1-weighted imaging, and diffusion-derived FA maps. Resting-state fMRI was analyzed using independent component analysis and dual regression, whereas an IBSI-compliant radiomics workflow and machine-learning models were used for exploratory scan-level classification. Compared with baseline, the SHAM group showed reduced synchronization across several resting-state networks, whereas the QMT group showed increased synchronization in the right sensorimotor and frontoparietal networks and no significant reductions. Between-group analyses showed lower delta-FC in SHAM than QMT in the cerebellar and sensorimotor networks. In contrast, radiomics showed limited discrimination between pre- and post-QMT scans; the best model achieved a ROC-AUC of 0.65 with near-chance accuracy, and no selected predictor remained significant after multiple-comparison correction. These findings suggest that QMT may support short-term functional network stability or task-relevant reorganization in PD relative to the SHAM condition, whereas whole-brain structural radiomics appears less sensitive for detecting early training-related effects in this setting.

Bioengineering doi: 10.3390/bioengineering13050485

Authors: Karol Nowak Anna Szymczak-Graczyk Aram Cornaggia Tomasz Garbowski

The squat is a fundamental multi-joint movement widely studied in strength training and biomechanics. While numerous experimental and computational studies have examined squat kinematics and joint loading, the mechanisms governing how squat technique adapts to increasing external load remain insufficiently understood. In particular, inverse-dynamics-based approaches often overlook explicit constraints imposed by limited joint moment capacity. This study presents a computational framework for predicting load-dependent adaptations of squat posture. The human body was represented as a multi-segment rigid-body system, with joints modeled as nonlinear rotational elements with bounded moment capacity. A reference squat trajectory was first generated kinematically, and a constrained optimization procedure was then applied at each motion frame to determine a mechanically admissible posture under increasing barbell load. The results show that higher loads lead to systematic posture adaptations, including increased torso inclination and redistribution of rotational demand from the knee toward the hip joint. For the highest load, peak torso pitch increased from 30° to over 40°, while joint utilization exceeded unity, indicating the onset of yielding. These findings identify joint moment capacity as a key constraint governing squat technique and demonstrate the potential of the proposed framework for predictive biomechanical analysis.

Bioengineering doi: 10.3390/bioengineering13050484

Authors: Leandro Norberto da Silva-Júnior Maria Angelica Miglino Bianca de Oliveira Horvath-Pereira João Victor Barbosa Tenório Fireman Giovanna Macedo da Siqueira Maria Laura dos Reis Ferre Pereira Letícia dos Santos Bezerra Luís Vicente Franco de Oliveira Samuel de Sousa Morais Márcia Zilioli Bellini Carlos Henrique Bertoni Reis Rogerio Leone Buchaim Daniela Vieira Buchaim

Pulmonary bioengineering holds significant promise for the development of functional lungs suitable for transplantation in patients with terminal lung diseases; however, it encounters considerable challenges. The inherent structural complexity, diverse cellular composition, and the intricate process of re-endothelialization the pulmonary vasculature complicate efforts to reconstruct viable lungs for transplantation. This study aimed to establish an innovative re-endothelialization technique utilizing decellularized scaffolds, integrating canine yolk sac-derived endothelial precursor cells with mechanical respiratory stimuli within a bioreactor framework. Wistar rat lungs were subjected to a decellularization protocol employing SDS + Triton X-100 0.5% and subsequently assessed for cytocompatibility with murine fibroblasts (3T3) and yolk sac (YS) cells in fragments. Following this, the recellularization of the whole-lung scaffold was evaluated under constant mechanical respiratory stimulation with YS cells. Each stage of the process was rigorously analyzed using histological staining, DAPI, scanning electron microscopy (SEM), and genomic DNA quantification. The findings reveal that the implemented alternating decellularization protocol resulted in a structured scaffold conducive to the culture of various cell types in fragments. When subjected to the complete scaffold recellularization model, the results indicated that YS cells are advantageous for the re-endothelialization process. Moreover, when employed in conjunction with the bioreactor model incorporating respiratory stimulation, these cells demonstrated enhanced cellular diffusion capacity and facilitated more homogeneous recellularization of the entire organ. These results signify a notable advancement in the reconstruction of new tissues for pulmonary transplantation.

Bioengineering doi: 10.3390/bioengineering13040483

Authors: AmirHossein MajidiRad Iram Azam Japp Adhikari Mehrnoosh Damircheli

Rotator cuff injuries are among the most common musculoskeletal conditions that affect shoulder function and can ultimately impact quality of life. While physical therapy is essential in the care of rotator cuff injuries, the ideal dose of therapeutic exercises continues to be a significant clinical dilemma because of the generalized nature of rehabilitation protocols. This pilot study proposes a machine learning approach to personalize rehabilitation using surface electromyography (sEMG) data collected from eight healthy individuals by testing four key shoulder movements: scaption, internal rotation, external rotation, and external rotation at 90° abduction. In this research, the XGBoost algorithm was used to model muscle activation patterns by achieving a high predictive accuracy (R2 = 0.5325; MSE = 0.0084 μV2). Because sEMG reliably measures superficial muscle activity, a linear programming model was used to divide a 60 min therapy session in a way that increases activation of superficial muscles (such as deltoid and trapezius) while reducing strain on deep muscles (such as supraspinatus and infraspinatus). Three optimization scenarios were tested by reflecting a different clinical goal: prioritizing superficial muscles, minimizing deep muscle strain, or balancing both. Optimized time allocations assigned more time to external rotation at 90° abduction and scaption. This research demonstrates the potential for data-driven methods to transform rotator cuff rehabilitation through personalized and evidence-based treatment plans. The results enhance clinical practice by enabling adaptive rehabilitation planning and show that machine learning can support decision-making in complex muscle activation analysis with strong performance and low latency.

Bioengineering doi: 10.3390/bioengineering13040482

Authors: Christian Liebsch

As the central musculoskeletal element of the human body, the spine simultaneously enables trunk movement, upright posture, and load transfer from the upper to the lower body [...]

Bioengineering doi: 10.3390/bioengineering13040481

Authors: Stavros Patsouris Panagiotis Mallis Efstathios Michalopoulos Nefeli Papadopoulou Michalis Katsimpoulas Nikolaos Nikiteas

Revisions to Authorship and Affiliation [...]

Bioengineering doi: 10.3390/bioengineering13040480

Authors: Diyar Qader Zeebaree Merdin Shamal Salih Danial William Odeesho Dilovan Asaad Zebari Nechirvan Asaad Zebari Omar I. Dallal Bashi Reving Masoud Abdulhakeem Yahya Ahmed Yahya

Diabetes mellitus is a health issue that is rapidly increasing worldwide, and it affects more than 347 million people globally. It is important to note that the disease can be successfully detected in its early stages, enabling physicians to avoid complications and improve patient outcomes. Despite the fact that machine learning (ML) has been extensively used in diabetes classification, the available solutions tend to place little or no emphasis on feature selection and ensembles, which limits prediction accuracy and generalizability. In this study, we introduce a hybrid framework that is based on three feature-selection algorithms, specifically, genetic algorithm (GA), correlation-based feature selection (CFS) and recursive feature elimination (RFE), in single and hybrid forms, and three classifiers, namely, multi-layer perceptron (MLP), support vector machine (SVM) and random forest (RF), to achieve a greater predictive robustness with the aid of soft voting. Experimental findings obtained from a benchmark diabetes dataset indicate that the RFE + CFS + SVM combination achieves the best performance, with an accuracy of 98.0%, sensitivity of 97.43%, specificity of 99.03%, precision of 99.51% and F1-score of 98.72%. These results indicate that the suggested hybrid feature-selection and ensemble learning model can offer a robust and highly effective approach for early-stage diabetes diagnosis, one which clinicians may use to make timely and accurate decisions.

Bioengineering doi: 10.3390/bioengineering13040479

Authors: Darryl Narcisse Robert Benkowski Matthew Dwyer Samarendra Mohanty

Optogenetic modulation employs light-sensitive proteins known as opsins to regulate cellular activity. A unique therapeutic application of this technique involves modulating pain perception by selectively targeting neural pathways within the spinal cord. Multi-Characteristic Opsin (MCO) represents an innovative optogenetic actuator capable of activation across a broad spectrum of light wavelengths, exhibiting a slow depolarizing phase that resembles natural photoreceptors. This study examines the current advancements in spinal optogenetic modulation utilizing MCO for pain management. Due to its high sensitivity, MCO facilitates minimally invasive, remotely controlled optogenetic modulation of spinal neurons. This approach enables the regulation of extensive spatial regions, provided the MCO channel receives sufficient light intensity to surpass the activation threshold. Nano-enhanced optical delivery (NOD) successfully transfected spinal neurons with the GAD67-MCO2-mCherry construct, as confirmed by membrane-localized mCherry fluorescence with DAPI-labeled nuclei. Using this platform, 5 Hz spinal optogenetic stimulation produced a significant reduction in formalin-evoked pain behaviors, demonstrating frequency-specific modulation of spinal pain circuits. Neither 2 Hz nor 10 Hz stimulation yielded comparable analgesic effects, underscoring the importance of precise stimulation parameters. The therapeutic impact also depended on transfection efficiency: reducing the fGNR–plasmid concentration diminished MCO expression and weakened the analgesic response. Together, these results show that effective spinal optogenetic pain modulation requires both optimal stimulation frequency and robust gene delivery.

Bioengineering doi: 10.3390/bioengineering13040478

Authors: Abdullah Jabri Mohamed Alsharif Bader Taftafa Tasnim Abbad Dania Sibai Abdulaziz Mhannayeh Abdulrahman Elsalti Islam M. Saadeldin Jahan Salma Tanveer Ahmad Mir Ahmed Yaqinuddin

Diabetes mellitus is primarily caused by the loss or malfunction of insulin-producing β-cells, and although current therapies improve glycemic control, they do not restore physiologic insulin secretion. Advances in stem cell biology and organoid engineering have led to the development of pancreatic organoids and induced pluripotent stem cell (iPSC)-derived β-cells as promising platforms for disease modeling, drug testing, and regenerative medicine. Pancreatic organoids generated from ductal, acinar, or progenitor populations can recapitulate key anatomical and functional features of native pancreatic tissue, enabling studies of development, injury, and regeneration. In parallel, improvements in iPSC differentiation protocols have produced β-like cells capable of insulin secretion in response to glucose, although achieving full functional maturity remains a challenge. Bioengineering strategies, including biomaterial scaffolds, microfluidic platforms, endothelial co-culture systems, three-dimensional bioprinting, and CRISPR-based genome editing, have enhanced the stability, vascular compatibility, and functional performance of both organoid and iPSC-derived systems. Despite these advances, variability in differentiation efficiency, limited β-cell maturity, and poor long-term survival continue to hinder clinical translation. Together, pancreatic organoids and iPSC-derived β-cells represent complementary platforms that advance fundamental research and support the development of β-cell replacement therapies, with ongoing integration of bioengineering approaches expected to accelerate progress toward reproducible, scalable, and clinically relevant β-cell regeneration.

Bioengineering doi: 10.3390/bioengineering13040477

Authors: Chih-Hao Chang Mei-Ling Chan Yu-Hung Fang Po-Lin Huang Tsung-Yi Chen Tsun-Kuang Chi I Elizabeth Cha Tzong-Rong Ger Kuo-Chen Li Shih-Lun Chen Liang-Hung Wang Jia-Ching Wang Patricia Angela R. Abu

Concurrent with the rising consumption of ultra-processed, high-calorie diets and the decline in physical activity, obesity and related cardiovascular conditions among young adults have continued to increase, becoming an important global public health concern. This study integrates non-invasive Internet of Things (IoT) sensing devices, including the TERUMO ES-P2000 blood pressure monitor (Terumo Corp., Tokyo, Japan) and the PhysioFlow PF07 Enduro cardiac hemodynamic analyzer (Manatec Biomedical, Poissy, France), with an artificial neural network (ANN) for cardiac index (CI) prediction. Through appropriate data preprocessing and model training strategies, the generalization ability and stability of the proposed CI prediction model were significantly enhanced. Experimental results demonstrate that, when using three physiological parameters as input, the ANN achieved a classification accuracy of 97.78%, substantially outperforming traditional approaches. Even under two-parameter input conditions, the model maintained strong predictive performance. These findings confirm the effectiveness and practical potential of the proposed framework for real-time, non-invasive CI assessment. Moreover, this research has received rigorous assessment and approval from the Institutional Review Board (IRB) under application number 202501987B0.

Bioengineering doi: 10.3390/bioengineering13040476

Authors: Panangattukara Prabhakaran Praveen Kumar Dong-Kwon Lim Taeho Kim

Photoacoustic imaging (PAI) is an emerging hybrid biomedical imaging modality that combines the high molecular contrast of optical excitation with the deep tissue penetration of ultrasound detection. This review presents recent advances in PAI-based techniques for the detection and characterization of gynecological diseases in women, with particular focus on endometriosis and uterine-related disorders. We summarize the application of PAI across preclinical and translational studies, highlighting progress in photoacoustic microscopy, spectroscopic photoacoustic imaging, and endoscopic and probe-based implementations for non-invasive, high-resolution tissue evaluation. The role of functional and contrast-enhanced PAI approaches is discussed, emphasizing their ability to enhance diagnostic sensitivity, enable longitudinal monitoring, and provide detailed information on vascular, biochemical, and structural tissue characteristics. Furthermore, the expanding applications of PAI in assessing uterine, cervical, and ovarian pathologies, including tumor detection and tissue remodeling, are reviewed. Finally, current challenges, limitations, and future directions toward clinical translation are addressed. Collectively, this review underscores the potential of photoacoustic imaging as a powerful, non-invasive platform for early diagnosis, disease monitoring, and improved management of women’s health conditions.

Bioengineering doi: 10.3390/bioengineering13040475

Authors: Roberta Fusco Vincenza Granata Paolo Vallone Teresa Petrosino Maria Daniela Iasevoli Roberta Galdiero Mauro Mattace Raso Davide Pupo Filippo Tovecci Annamaria Porto Gerardo Ferrara Modesta Longobucco Giulia Capuano Roberto Morcavallo Caterina Todisco Fabiana Antenucci Mario Sansone Mimma Castaldo Daniele La Forgia Antonella Petrillo

Background: This study investigates the impact of anatomically constrained preprocessing and deep learning architecture selection on benign versus malignant breast lesion classification in contrast-enhanced mammography (CEM), with the goal of improving robustness and clinical reliability across heterogeneous data sources. Methods: In this retrospective multicenter study, CEM images from 300 patients (314 lesions) were combined with 1003 publicly available CEM images, yielding a total of 1120 breast cases. Automatic breast segmentation was performed using the LIBRA framework to generate breast-mask images. Eleven deep learning models, including classical convolutional neural networks, attention-based networks, hybrid convolutional neural networks (CNNs), Transformer architectures, and mammography-specific models, were trained and evaluated using both original DICOM images and breast-mask inputs. Performance was assessed using accuracy, balanced accuracy, sensitivity, specificity, AUROC, and AUPRC on cross-validation and independent test sets. Hyperparameter optimization was conducted for the best-performing architecture. Results: Models trained on breast-mask images consistently outperformed those trained on original DICOM images across all architectures and metrics, with AUROC improvements ranging from +0.06 to +0.21. Among all models, ResNet50 trained on breast-mask images achieved the best performance (AUROC = 0.931; AUPRC = 0.933; balanced accuracy = 0.834), further improved after optimization (balanced accuracy = 0.886; sensitivity = 0.842; specificity = 0.930). Classical CNN architectures demonstrated performance comparable to or exceeding that of more complex hybrid CNN–Transformer models when anatomically focused preprocessing and rigorous optimization were applied. Conclusions: Anatomically constrained preprocessing through breast-mask segmentation substantially enhances deep learning performance and stability in CEM-based breast lesion classification. These findings indicate that input representation quality and training optimization are critical determinants of clinically relevant performance, often outweighing architectural complexity, and may support more reliable AI-assisted decision support in CEM workflows.

Bioengineering doi: 10.3390/bioengineering13040474

Authors: Shu Xu Zheng Chang Zenghui Ding Xianjun Yang Tao Wang Dezhang Xu

Wearable neurotechnology depends critically on continuous movement monitoring to characterize motor impairment and recovery in real-world settings. While joint torque serves as a clinically essential kinetic marker, estimating it directly on-device from inertial signals remains challenging due to stringent computational, memory, and energy constraints. Lightweight pipelines typically omit computationally expensive time–frequency processing; however, this omission degrades the observability of dynamics encoded in 1D IMU signals and diminishes the effectiveness of standard knowledge distillation strategies. To enable reliable on-device torque inference, we propose a Physically Guided Dual-Consistency Knowledge Distillation (PDC-KD) framework that explicitly integrates biomechanical priors into the learning process through two collaborative pathways: parameter-manifold alignment and physics-guided compensation. The student network receives guidance through Fisher-information-weighted parameter transfer, ensuring robust knowledge distillation despite significant model capacity mismatch. Furthermore, the framework incorporates a physics-guided regularization term that enforces dynamically consistent torque trajectories via a numerically stable Cholesky-parameterized constraint. Experiments demonstrate that the student model preserves teacher-level predictive accuracy while operating within the stringent resource constraints of edge devices (achieving a 98% parameter reduction, ∼2× faster inference, and ∼1 ms latency). Moreover, the proposed method yields torque estimates with enhanced dynamical consistency, providing an efficient biosignal-processing solution for wearable neurotechnology platforms demanding real-time movement analytics.

Bioengineering doi: 10.3390/bioengineering13040473

Authors: Catalin Ionite Lucian Indrei Andrei Gheorghiță Bogdan Caba Marius Turnea Irina Duduca Cezar Mucileanu Iustina Condurache Mariana Rotariu

Sports-related muscle injuries represent a major challenge in both recreational and professional sports, accounting for a substantial proportion of time-loss injuries and frequently leading to recurrent episodes. The aim of this narrative review was to analyze the biomechanical and neuromuscular mechanisms involved in the occurrence of muscle injuries and to synthesize evidence-based prevention strategies reported in the scientific literature. The literature search was conducted in the Web of Science database using the keyword “muscle injury prevention”, focusing on studies published between 2010 and 2025. The analyzed literature indicates that muscle injuries are strongly associated with eccentric contractions at long muscle lengths, neuromuscular fatigue, strength imbalances, impaired lumbopelvic stability, and inadequate load management. Preventive strategies based on biomechanical principles, particularly eccentric strength training, neuromuscular training programs, and core stability exercises, have demonstrated consistent effectiveness in reducing injury incidence and recurrence rates across multiple sports disciplines. In addition, emerging technological approaches, including wearable sensors and machine learning models, show promising potential for injury risk prediction and individualized prevention strategies.

Bioengineering doi: 10.3390/bioengineering13040472

Authors: Shray Mehra Hana Shah Sara E. Munkwitz Nicholas J. Iglesias Tina Joshua Kashyap K. Tadisina Natalia Fullerton Vasudev Vivekanand Nayak Lukasz Witek Paulo G. Coelho

Current evidence suggests that achieving the desired level of osseointegration necessitates a hierarchical approach to implant design. This is particularly relevant for osseointegration around implant systems such as those presenting vertical decompression chambers and acid-etched surfaces which could further be augmented by non-thermal plasma (NTP) treatment. Three implant systems were compared in this study: (i) ND (GM Helix Acqua Implant; Neodent®, Curitiba, PR, Brazil—hybrid, acid-etched thread design treated with isotonic sodium chloride solution), (ii) Sin (Epikut Plus; S.I.N. Implant System, São Paulo, Brazil—V-shaped, acid-etched thread design treated with nano-hydroxyapatite), and (iii) Mp (Maestro; Implacil De Bortoli, São Paulo, Brazil—buttress, acid-etched thread design with decompressing vertical chambers). The ND and Sin implants were used directly as supplied by the manufacturer. For the Mp implants, the manufacturer-supplied surface was subjected to supplemental acid etching with 37% hydrochloric acid followed by Argon-based NTP treatment administered with a pulsed plasma generator prior to implantation into the iliac crest of n = 12 adult female sheep. Histomorphometric analysis was conducted at 3- and 12-week post-implantation (n = 6 sheep per time point) to assess bone-to-implant contact (BIC) and bone area fraction occupancy (BAFO). After 3 weeks in vivo, the healing chambers of all implant groups consisted predominantly of newly forming woven bone. By 12 weeks, bone maturation was observed, with the presence of remodeling sites and some areas of well-organized lamellar structures occupying the healing chambers. At both 3 and 12 weeks, the Mp implants demonstrated significantly higher BAFO values relative to ND (p = 0.015 and p = 0.008, respectively). The combination of vertical healing chambers, acid etching, and NTP treatment promoted early vascular infiltration and sustained bone deposition.

Bioengineering doi: 10.3390/bioengineering13040471

Authors: Jia-Sheng Hong Yun-Hsuan Tzeng Kuan-Ting Wu Shih-Yu Huang Ting-Wei Wang Guan-Yu Li Chun-Yi Lin Ho-Ren Liu Hai-Neng Fu Yung-Tsai Lee Wei-Hsian Yin Yu-Te Wu

Early detection of aortic dilatation is clinically important for preventing progression to serious aortic disease and enabling timely intervention. We aimed to develop an AI method for quantifying the aorta in both contrast-enhanced and non-contrast CT scans, assisting early detection of aortic dilation. A total of 190 patient cases were analyzed, each having paired contrast-enhanced and non-contrast CT scans acquired in the same session, resulting in 380 scans. Our approach, based on open-source tools, demonstrated strong agreement with manual annotations, particularly in the ascending aorta. For contrast-enhanced CT, the AI achieved a correlation coefficient of 0.987 and intraclass correlation coefficient (ICC) of 0.986; for non-contrast CT, both were 0.945. Compared with clinical records, the sensitivity of AI detection was 97% for contrast-enhanced CT and 94% for non-contrast CT. This AI-based workflow enables highly sensitive automated aortic quantification in both contrast-enhanced and non-contrast CT scans, supporting broader clinical applicability across different imaging conditions.

Bioengineering doi: 10.3390/bioengineering13040470

Authors: Marcin Majos Artur Klepaczko Ilona Kurnatowska

Chronic kidney disease (CKD) represents a growing medical, diagnostic and social challenge, and it is estimated to effect 8.5–9.8% of the global population and requires expensive modes of treatment, such as hemodialysis or renal transplants. Currently, a diagnosis of CKD is set based on the level of creatinine in the blood, which is the gold standard of renal function diagnostics. Unfortunately, decrease in GFR is secondary to damage of the kidney parenchyma and indicates that the best time to start more aggressive treatment has already passed. Therefore, several non-invasive methods have been proposed for predicting increased risk of CKD progression; however, in most of the cases kidney biopsy is essential. Currently, the greatest hopes for a method that can confirm CKD are associated with the development of MRI, the most tissue-specific imaging method, and it is already proven to be capable to detect inflammatory and edematous changes, fibrosis, as well as perfusion and oxygenation disturbances. Therefore, in our manuscript we decided to present up-to-date knowledge about kidney MRI from a clinical point of view.

Bioengineering doi: 10.3390/bioengineering13040469

Authors: Xiaoyu Gu Wei Zhong Zhihao Mao Yan Shi Yixuan Wang

Objectives: This study aimed to demonstrate the feasibility of a novel wireless energy-transmitting implantable diaphragm pacemaker for restoring respiratory ventilation. Methods: The diaphragm pacing (DP) system was designed based on the principle of electromagnetic resonance coupling. The safety of device implantation was analyzed through finite-element simulations of multi-field coupling between electromagnetic heating and biological tissue. In vitro testing with coils embedded in pork demonstrated the system output characteristics. This device was used in miniature Bama pigs that underwent deep anesthesia and respiratory arrest (N = 8). Respiratory airflow, diaphragmatic displacement, and blood gases were used to evaluate the effectiveness of the designed DP system. Results: Thermal effect simulation results show that the temperature rise of the surrounding tissue does not exceed 2 °C during 1 h of transmission power (0.5–1.3 W) operation of the receiver. In vitro tests with two receivers embedded in pork showed that the DP system can effectively output stimulation waveforms over a certain transmission distance (5–35 mm). The stimulation waveform output by the receiver is consistent with the parameters set by the external controller. In phrenic nerve electrical stimulation experiments, the peak respiratory airflow and tidal volume remained stable over 50 consecutive respiratory cycles. The tidal volume (108.63 mL) and diaphragmatic displacement (0.883–2.15 cm) in a pig induced by DP demonstrate the effectiveness of respiratory ventilation. The arterial blood gas analysis results and temperature rise experiment during implantation further confirmed the effectiveness and safety of the ventilation. Conclusions: The implantable diaphragmatic pacemaker developed in this study exhibits good thermal safety, stable output, and effective respiratory ventilation. A control group with commercial diaphragmatic pacemakers and data from chronic implantation experiments are needed to further evaluate its effectiveness.

Bioengineering doi: 10.3390/bioengineering13040468

Authors: Ruixin Dai Liping Cui Kun Hu Jiye An Ning Deng

Providing dietary feedback is important for promoting healthy behaviors in weight management, but the rapid development of obesity and the shortage of medical nutrition human resources have limited this health service. The rise of large language models (LLMs) offers the possibility of using artificial intelligence (AI) to simulate the behavior of human dietitians. However, existing studies have only explored LLM performance when generating answers to common nutrition-related questions; the use of LLMs to generate situation-adapted dietary feedback in practical weight management scenarios still needs further research. In this study, we collected dietary records and dietary feedback from primary dietitians through an mHealth weight management application. We conducted topic modeling to generalize how dietitians deliver nutrition guidance in real-world dietary feedback scenarios. Combining the in-context learning capability of LLMs with real-world data, we proposed a synthetic data generation approach (HDI-SDG) and trained an LLM for dietary feedback with the synthetic data (LLMDF-EXP). Experiments on automatic and manual evaluation of LLMDF-EXP and an LLM trained directly with the real-world data as well as generalized LLMs illustrated that LLMDF-EXP performed most similarly to human experts. Notably, there were no significant differences from human experts in terms of professionalism (p-value = 0.510) and usefulness (p-value = 0.498). The study highlights that integrating LLMs with real-world data in health management processes can enhance the situational adaptability of LLMs in practical health management environment applications.

Bioengineering doi: 10.3390/bioengineering13040467

Authors: Chaoxin Li Chenghong Lu Jiuqiang Li Kai Guo

Precise monitoring of forearm muscle groups is crucial for decoding motor intentions in human–machine interfaces (HMIs) and rehabilitation. However, traditional surface electromyography (sEMG) electrodes face significant challenges in densely packed muscle regions with large skin deformations, leading to severe signal crosstalk and unstable contact. Here, we report a flexible, low-cost 16-channel copper electrode array system designed for the high-density monitoring of multiple forearm muscle activities. Through a facile fabrication process, rigid copper is transformed into a conformable sensing interface. The optimized serpentine interconnects endow the array with excellent stretchability and effectively isolate motion-induced stress, ensuring high-quality signal acquisition under complex deformations. The high-density 2 × 8 array enables the spatiotemporal mapping of distributed flexor and extensor muscle groups. Integrated with a customized wireless data acquisition system, the array successfully demonstrates real-time, multi-channel sEMG monitoring of various hand movements (e.g., fist clenching, wrist flexion/extension), clearly revealing specific muscle activation patterns. This low-cost, high-performance flexible sensor array provides a highly promising tool for complex gesture decoding, electromyographic imaging, and next-generation wearable HMIs.

Bioengineering doi: 10.3390/bioengineering13040466

Authors: Musa Adamu Wakili Aminu Bashir Suleiman Kaloma Usman Majikumna Harisu Abdullahi Shehu Huseyin Kusetogullari Md. Haidar Sharif

The demand for advanced detection methods and accurate staging remains a global challenge in cancer diagnosis. Even though traditional deep learning models in medical imaging achieve high precision, they suffer from limited explainability and multimodal reasoning due to their black-box nature, thereby limiting their clinical applicability. To address this gap, recent research has increasingly explored multimodal approaches that integrate visual and textual clinical data to enhance diagnostic accuracy and interpretability. This study presents a bibliometric analysis of 408 publications from 2021 to 2025, collected from Web of Science and Scopus, using VOSviewer and R-Bibliometrix to map citation networks, co-authorship, and keyword co-occurrences. The results reveal a rapid growth from 1 publication in 2021 to 269 in 2025, with significant contributions from leading countries and institutions. Thematic analysis indicates a shift from conventional convolutional approaches toward transformer-based and self-supervised methods, alongside increasing attention to multimodal learning in cancer imaging tasks such as breast, lung, and brain cancer analysis. Overall, this study provides a structured overview of the evolving research landscape, highlighting key trends, emerging themes, and research gaps to inform future developments in multimodal artificial intelligence for cancer diagnosis.

Bioengineering doi: 10.3390/bioengineering13040465

Authors: Marija Stanišić Milica Crnoglavac Popović Nikola Knežević Marko Radenković Branimir Bajac Olivera Prodanović Radivoje Prodanović

Biomineralization has recently emerged as a highly effective strategy for enzyme immobilization. Zeolitic imidazolate frameworks (ZIFs), a subclass of metal–organic frameworks (MOFs), are particularly attractive carriers due to their structural tunability and chemical stability. While ZIF-8 has been extensively studied, its denser and thermodynamically more stable analog ZIF-zni has received far less attention. In this work, we report the biomineralization of glucose oxidase (GOx) from Aspergillus niger within the ZIF-zni framework and systematically investigate the influence of zinc and imidazole (Im) concentration on immobilization performance. The optimized biocomposite, obtained at 10 mM Zn2+ and a Zn:Im ratio of 1:10, exhibited a specific activity of 2051 IU g−1, which is more than twice the activity obtained for GOx@ZIF-8 in our previous study (874 IU g−1). Furthermore, the GOx@ZIF-zni biocomposite demonstrated remarkable resistance to sodium dodecyl sulfate (SDS) and retained up to 50% of its activity after incubation at 65 °C for one hour. These results demonstrate that ZIF-zni is a highly promising carrier for enzyme immobilization and suggest that framework topology and synthesis conditions play a crucial role in determining the catalytic performance and stability of enzyme@MOF biocomposites.

Bioengineering doi: 10.3390/bioengineering13040464

Authors: Sandra P. Hager Vahid Ziaei-Rad Jenny S. Choy Mengjun Wang Ghassan S. Kassab Lik Chuan Lee

Cardiac and hemodynamic conditions such as myocardial infarct, cardiomyopathy, hypertension, and aortic valve disease can impair conduction within the Purkinje fiber network and compromise left ventricular (LV) pump function. We developed a computational framework that couples electrical propagation in a structurally organized Purkinje fiber network with LV electromechanics to analyze the impact of conduction abnormalities on cardiac performance. A baseline simulation reproduced physiological activation patterns and pump indices consistent with healthy human data. Conduction block was then introduced at different locations within the Purkinje fiber network. LV pump function was strongly dependent on block location: left bundle branch block (LBBB) produced the largest reduction in ejection fraction (EF) (59% to 46%) and peak pressure (119 to 97 mmHg), whereas left anterior fascicle block caused smaller functional changes. Across simulations, myocardial activation delay and systolic dyssynchrony index (SDI) exhibited a nonlinear relationship with EF and myocardial strain. A threshold behavior was identified at a simulated LV activation duration of approximately 240 ms and an SDI of 8.4%, beyond which EF and strain decreased by about 5% relative to baseline. These findings provide a mechanistic framework to investigate how Purkinje fiber network conduction abnormalities influence LV pump dysfunction.

Bioengineering doi: 10.3390/bioengineering13040463

Authors: Anagha Shinde Virendra Shete Ninad Mehendale

Electromyography (EMG) signals are critical for prosthetic control, rehabilitation, and human–machine interaction, yet their classification remains challenging due to noise, non-stationarity, and inter-subject variability. This study presents a comprehensive comparative analysis of machine learning (ML), deep learning (DL), and reinforcement learning (RL) approaches for 1D EMG signal classification, with a systematic evaluation of signal acquisition parameters. Using both synthetic and real-world EMG datasets, we demonstrate that 8–10 bit quantization and a 2000 Hz sampling rate provide optimal signal fidelity while maintaining data efficiency. Among the evaluated models, ensemble methods (Gradient Boosting, Voting Ensemble) and advanced DL architectures (LSTM, Transformer) achieved superior performance on real EMG data, with accuracies reaching 100% and 96.3%, respectively. Notably, reinforcement learning agents (Deep Q-Networks) demonstrated 100% accuracy on multiclass synthetic data, revealing their potential for learning complex bio-signal representations. Our findings establish that meticulous optimization of preprocessing pipelines, combined with robust AI models, significantly enhances EMG classification accuracy. This work provides empirical guidance for selecting optimal acquisition parameters and AI architectures for practical EMG analysis systems, with direct implications for prosthetic control and rehabilitation technologies.

Bioengineering doi: 10.3390/bioengineering13040462

Authors: Ragai-Edward Matta Renan Belli Katrin Hurle Arulraj Sangarapillai Oleksandr Sednyev Manfred Wichmann Lara Berger

Multilayer zirconias have recently been introduced as dental biomaterials to combine improved translucency with sufficient mechanical reliability by implementing yttria-driven gradients in phase composition. Such materials can be considered functionally graded ceramics, where local phase stabilization influences strength and crack resistance. However, manufacturer-specific gradient profiles and their structure–property relationships remain insufficiently characterized. This study investigated two commercially available multilayer zirconias with distinct gradient concepts: IPS e.max® ZirCAD Prime (continuous gradient) and KATANA™ Zirconia YML (stepwise gradient). Ten equidistant sections along the blank height were analyzed using quantitative X-ray diffraction and Rietveld refinement to quantify zirconia phase fractions and estimate local Y2O3 content. Mechanical behavior was evaluated by biaxial flexural strength testing (ball-on-three-balls method) and fracture toughness testing using the chevron-notched beam technique. Both materials exhibited pronounced yttria- and phase-dependent gradients consistent with their reported layer designs. Regions with increased yttria content showed higher t″ fractions and reduced fracture toughness and strength, whereas deeper regions displayed increased mechanical performance associated with higher fractions of transformable tetragonal phase. These findings emphasize that multilayer zirconias exhibit spatially dependent mechanical properties, which should be considered in biomaterial selection and restoration design, particularly when balancing aesthetic demands and fracture resistance.

Bioengineering doi: 10.3390/bioengineering13040461

Authors: Ahmad Tijjani Garba Aminu Bashir Suleiman Wenze Du Ahmed Ibrahim Mahmud Harisu Abdullahi Shehu Huseyin Kusetogullari Md. Haidar Sharif

This study presents the first dedicated bibliometric analysis of artificial intelligence (AI) and deep learning applications in pediatric radiology and medical imaging, mapping the intellectual structure of a rapidly evolving field. A total of 2688 articles and conference proceedings published between 2005 and 2025 were retrieved from the Web of Science Core Collection and analyzed using Bibliometrix R and VOSviewer. The findings reveal exponential growth in publications, from 7 papers in 2005 to 559 in 2025, with journal articles dominating the corpus (85.9%). The most-cited contributions, led by Kermany et al. (2018) with 2886 citations, are predominantly technical feasibility studies rather than clinical outcome trials, indicating a field that has advanced methodologically but remains in early stages of clinical translation. Thematic mapping identifies convolutional neural networks, pneumonia, and transfer learning as Motor Themes representing methodological maturity in chest imaging, while neuroimaging and image segmentation clusters occupy Niche Themes, reflecting insular development with limited cross-field connectivity. Geographic analysis reveals concentrated co-authorship along US–China and US–Europe corridors, with African, Latin American, and Southeast Asian institutions largely absent from knowledge production networks. Eight of the ten most productive affiliations are North American, highlighting structural inequities that risk producing AI tools optimized for high-resource settings rather than the global pediatric population. This analysis provides an empirical foundation for reorienting the field toward clinical validation, geographic inclusion, and methodological integration across isolated research communities.

Bioengineering doi: 10.3390/bioengineering13040460

Authors: António Ramos Olga Noronha Orlando Simões José Noronha José Simões

Surgical fixation techniques for bone fracture healing are well established and effective; however, opportunities remain to improve both functional outcomes and the patient experience. The Biofiligree® concept integrates medicine, engineering, and design by reimagining conventional osteosynthesis plates as both therapeutic and aesthetic devices. Inspired by traditional Portuguese filigree, these plates allow patient participation through personalized geometries, patterns, or engravings and may later be transformed into wearable jewellery after removal, preserving them as symbolic artefacts of recovery. This study introduces and biomechanically evaluates a novel calcaneal fixation plate incorporating the biofiligree geometry concept. A biofiligree plate was designed for calcaneus fracture fixation and manufactured in stainless steel 306L. Experimental testing was conducted on synthetic composite calcaneus bone models to simulate anatomical conditions and compare the new design with a standard commercial plate. The biofiligree plate, 2 mm thick, was fixed using five screws and two percutaneous screws positioned at 45° to compress the fracture line. Results demonstrated comparable biomechanical performance between both systems, with similar strain distributions and fracture stabilization. The biofiligree plate showed stresses around 430 MPa and fracture displacement below 0.7 mm. Fixation stiffness values were 1445 N/mm for intact calcaneus, 1065 N/mm for the commercial plate, and 725 N/mm for the biofiligree plate, indicating adequate support for bone healing.

Bioengineering doi: 10.3390/bioengineering13040459

Authors: Liang Li Chen-Yi Jiang Xing-Jie Wang Yuan-Jun Wang Jian Zhuo

Objective: To enhance the information modeling capacity of large-kernel convolutional neural networks and to build a ureteral orifice detection framework for ureteroscopic imaging. Methods: A retrospective dataset of ureteroscopic images from 222 patients was collected. The patients were randomly divided into training and testing sets at a ratio of 7:3. Initially, video files were converted into image frames, and feature-relevant images were manually labeled by physicians. Subsequently, a ConvNeXt-based backbone augmented with squeeze-and-excitation (SE) modules was employed to extract diverse deep features. SCConv modules were incorporated across stages to strengthen the network’s feature extraction performance. Lastly, enhanced spatial excitation attention mechanisms were cascaded to achieve superior feature fusion and detection accuracy. Comparative experiments were conducted against baseline models, including ConvNeXt, assessing accuracy, computational overhead, and inference latency. Results: On a test set of 491 ureteroscopic images, all models achieved mAP@50 values above 0.75, whereas the proposed network achieved 0.890, markedly exceeding baseline performance. The model operated at 20 ms per frame, achieving a frame rate of 50 FPS. Conclusions: We developed an improved deep learning framework based on large-kernel convolutional networks for real-time ureteral orifice detection in endoscopic scenarios. This system achieves a favorable balance between detection accuracy and real-time efficiency. The method demonstrates significant potential as a training and feedback tool for residents and junior urologists in clinical environments.

Bioengineering doi: 10.3390/bioengineering13040458

Authors: Taieba Tuba Rahman Zhijian Pei Hongmin Qin Hamid R. Parsaei

Understanding human’s responses to extreme environments holds significant importance for space exploration, deep-sea research, and environmental adaptation. Traditionally, human subjects were used to study humans’ responses to extreme environments. The main limitations of this approach include the inability to independently investigate specific cellular mechanisms, ethical and safety constraints, limited experimental controllability, and inter-individual variability that complicates mechanistic interpretation. Another approach is to study humans’ responses at the cellular level using 2D culture. This approach often exhibits limited reproducibility due to its inability to recapitulate physiologically relevant microenvironments. Bioprinting can enable studies on human’s responses at the cellular level and within 3D environments. One way is to study human cells’ responses to localized and transient extreme environments created during printing. Another way is to expose 3D printed samples (embedded with human cells) to extreme environments. However, the literature does not contain comprehensive review papers to discuss the prospects and limitations of bioprinting for investigating human cells’ responses to extreme environments. This review paper aims to fill this gap in the literature. It begins with a brief description of the effects of extreme environments on human health and summarizes reported studies on cells’ responses to extreme environments. Afterward, it discusses the prospects and limitations of the two ways of using bioprinting to investigate cells’ responses to extreme environments. Finally, it concludes with identifying knowledge gaps and proposing research directions in the application of bioprinting to study human cells’ responses to extreme environments.

Bioengineering doi: 10.3390/bioengineering13040457

Authors: Akeel Qadir Saad Arif Prajoona Valsalan Osama Khan

This study introduces a physics-guided deep learning architecture designed for the simulation, reconstruction, and pattern recognition of biomedical images. By explicitly integrating physical priors into the learning model, the framework addresses the black-box nature of traditional artificial intelligence (AI). It provides an explainable AI pathway that enhances diagnostic accuracy, robustness, and clinical interpretation. The proposed framework was evaluated through systematic simulation studies. It involved complex geometric configurations, multimodal physical fields, and noise-corrupted synthetic three-dimensional brain volumes. Quantitative analysis demonstrates consistent improvements in reconstruction fidelity, with the peak signal-to-noise ratio (PSNR) reaching 47 dB and the structural similarity index exceeding 0.90 across all scenarios. Notably, at moderate noise levels (0.05), the framework maintains a PSNR greater than 32 dB, ensuring structural integrity essential for computer-aided diagnosis. Volumetric brain experiments further reveal a 38–44% reduction in activation localization errors, highlighting the framework’s utility in functional imaging and disease prognosis. By grounding deep learning in physical constraints, this study provides a transparent and robust solution for automated disease classification and advanced biomedical imaging tasks within clinical decision support systems.

Bioengineering doi: 10.3390/bioengineering13040455

Authors: Ahmed Abdirahman Ibrahim Michael Opoku Abakar Mahamat Abdramane Mingqing Fang Xu Liu Abdulraheem Mustapha Yusheng Li Wenfeng Xiao Kai Zhang Shuguang Liu

Purpose: To critically evaluate the role or effect of platelet-rich plasma (PRP) in anterior cruciate ligament (ACL) reconstruction in terms of clinical and radiological outcomes. Method: We conducted a systematic search of PubMed, Embase, the Cochrane Library, and Web of Science to identify relevant studies. Clinical outcomes included the Visual Analogue Scale (VAS), International Knee Documentation Committee (IKDC) subjective and objective evaluations, Lysholm score, Tegner score, anterior knee laxity, Knee Injury and Osteoarthritis Outcome Score (KOOS), Kujala score, Victorian Institute of Sport Assessment (VISA) scale, proprioception, isokinetic strength, and physical examination tests (anterior drawer, Lachman, and pivot-shift tests). Radiological outcomes encompassed measures obtained via magnetic resonance imaging (MRI), computed tomography (CT), X-ray, and ultrasound. Statistical significance was defined as a p value < 0.05, and all analyses were performed using R software (version 4.1.3). Results: A total of 23 studies, including 19 randomized controlled trials, met the inclusion criteria, encompassing 1072 patients overall. The meta-analysis showed significant differences between PRP group and non-PRP group with regard to VAS score at 6- and 12-month follow-up, Lysholm score at 6-month follow-up, and Tegner score at 6-month follow-up. Meta-regression showed that the two group differences in VAS score changed significantly with follow-up time (p < 0.01). In terms of radiological findings, about half of the assessments favored PRP to facilitate the graft maturation and integration at 6-month follow-up. Conclusions: PRP application in ACL reconstruction compared with non-PRP, may produce short-term but not long-term clinical outcomes such as VAS score, Lysholm score and Tegner score. While some short-term statistical differences exist, their magnitude and durability do not yet justify routine clinical adoption of PRP in ACL reconstruction. Larger samples and higher-quality studies are needed to support our results and further explore the advantages of PRP in other aspects. Level of evidence: Level II.

Bioengineering doi: 10.3390/bioengineering13040456

Authors: Anchal Kumari Punam Rattan Anand Kumar Shukla Sita Rani Aman Kataria Hong Min Taeho Kim

Skin cancer is caused by aberrant cells that proliferate uncontrollably after unrepaired DNA damage results in mutations in the epidermis. The majority of skin cancer is caused by high UV exposure from the sun, tanning beds, or sunlamps. Due to sociocultural hurdles, limited access to specialized dermatological care, and low public knowledge, many nations, including India, have higher mortality rates and late-stage presentations. The unequal distribution of specialized dermatological treatments, particularly in rural and underdeveloped areas, makes detection and treatment more difficult. For skin cancer, one of the most prevalent malignancies with a high death rate, early detection is crucial. This study gathered 1200 dermoscopic images from two clinics in Himachal Pradesh in order to solve these problems. In order to automatically classify dermoscopic clinical images into melanoma and non-melanoma skin cancer categories, this study compares VGG16 with ResNet-50. Preprocessing, lesion segmentation, and classification are all part of the suggested approach. A collection of 1200 dermoscopic images with clinical annotations was used to improve the models. ResNet-50 outperformed VGG16 in tests, with 93% accuracy and 96% AUC-ROC as opposed to 89% and 94%, respectively. These results emphasize how crucial model selection and preprocessing are to diagnostic performance. Ensemble methods, multi-class classification, explainability integration, and clinical validation will be investigated in order to facilitate the implementation of AI-assisted dermatological diagnostic tools.

Bioengineering doi: 10.3390/bioengineering13040454

Authors: Michael D. Lopez Jonathan Marrero Bermudez David Berard Lawrence Holland Austin J. Ruiz Jose M. Gonzalez Sofia I. Hernandez Torres Eric J. Snider

Hemorrhagic shock remains one of the leading causes of preventable death for both civilian and military trauma. Fluid resuscitation is the primary treatment but requires constant monitoring, particularly for volume non-responsive patients susceptible to fluid overload, pulmonary edema, and other life-threatening conditions. To overcome fluid non-responsiveness, vasoactive drugs or vasopressors can be necessary adjuvants to fluid therapy but require tedious titrations that can be difficult to manage during mass-casualty situations. This study developed and evaluated automated closed-loop vasopressor controllers for hemorrhage scenarios. Ten physiological closed-loop controller (PCLC) configurations with different underlying functionalities were tuned to be either more aggressive or conservative to reach the target mean arterial pressure. A hardware-in-loop test platform with fluid-pressure responsiveness, derived from animal data, tested each controller across three different starting pressure scenarios. The platform successfully differentiated controller designs based on performance metrics. While some configurations overshot the target and others could not reach the target pressure, strong-performing PCLCs consistently reached and maintained the target quickly. Three candidate PCLCs outperformed the rest and will be evaluated across wider scenarios to develop a robust controller design. This work accelerates PCLC-driven vasopressor administration development, providing a necessary fluid resuscitation adjuvant for precise hemodynamic management in hemorrhagic trauma.

Bioengineering doi: 10.3390/bioengineering13040453

Authors: Mirabel Ewura Esi Acquah Zengguang Wang Wei Chen Dongyun Gu

Turning impairment is a major contributor to falls in Parkinson’s disease (PD), yet the mechanisms linking cortical dysfunction to altered motor behavior remain unclear. In particular, it is unknown whether disrupted cortical communication impairs turning by altering muscle coordination. This study investigates a novel mechanistic pathway: whether muscle synergy complexity mediates the relationship between cortical network connectivity and turning performance in PD. Specifically, electroencephalography (EEG) and electromyography (EMG) were recorded from 12 individuals with PD and 12 age-matched healthy controls during a 180° turning task. Directed cortical connectivity, muscle synergy complexity, and spatiotemporal turning performance were quantified. Mediation analysis was used to determine whether cortical influences on behavior operate indirectly through neuromuscular coordination. Compared to controls, individuals with PD performed slower turns with shorter stride lengths and reduced synergy complexity (p < 0.05), alongside altered frontal cortical connectivity (p < 0.05). Across participants, higher synergy complexity was associated with faster, longer strides (p < 0.04). Cortical connectivity strength strongly predicted synergy complexity (R2 = 0.66, p < 0.001) and exerted a significant indirect effect on turning performance (β = 0.312; 95% CI [0.072, 0.605]; p = 0.008). In PD, reliance on this indirect pathway increased with disease severity and poorer turning ability (r > 0.57, p < 0.03). This work establishes how muscle synergy complexity significantly mediates the relationship between cortical connectivity and turning performance in PD. Our findings provide evidence of a cortical–neuromuscular–behavioral pathway underlying turning deficits, highlighting coordination as a key target for neurorehabilitation.

Bioengineering doi: 10.3390/bioengineering13040452

Authors: Shima Shojaei Catherine N. Mulligan

Biosurfactants offer a green, sustainable approach to many environmental bioremediations, especially for oil contamination. In this study, the aim is to evaluate the effectiveness of biosurfactants in accelerating hydrocarbon removal from mature fine tailings under anaerobic conditions. The bacteria were isolated from mature fine tailings and tested for biosurfactant production using different biosurfactant screening methods (i.e., blood agar, cetyltrimethylammonium bromide (CTAB) blue agar, oil displacement, and drop collapse). The most efficient strain showed high similarity to Stutzerimonas stutzeri by 16S rRNA gene sequencing. Results showed that this strain produces rhamnolipids with a critical micelle concentration (CMC) of 600 mg/L and a minimum surface tension of 38.70 ± 0.08 mN/m. Moreover, when supplemented with whey, the strain showed a high emulsification index of 24 toward toluene (66%) and hexane (60%). The bioremediation of mature fine tailings (MFTs) was conducted under anaerobic conditions by adding a consortium of the four strains that were positive in biosurfactant screening tests. The results showed 53% removal of n-alkane C9-C30 and a reduction in surface tension from 69 ± 0.5 mN/m to a minimum of 54.33 ± 0.5 mN/m. The results suggest the potential successful application of bioaugmentation for in situ biological treatment in the oil sands industry.

Bioengineering doi: 10.3390/bioengineering13040451

Authors: Yu Tian Congliang Yang Lingfeng Sang Cicao Ping Lili Feng Weixiong Chen Hongbo Wang Wenxian Li Yuan Han

Accurate and rapid detection of anatomical structures, such as the glottis, is critical during tracheal intubation (TI) to ensure patient safety and procedural success. However, it remains a challenge due to the limited field of view and computational resources of video laryngoscopy, especially for difficult airway situations. Existing deep learning (DL) models struggle to balance high accuracy and real-time clinical deployment. To address these issues, we propose TI-YOLO (TI-You Only Look Once), a lightweight and efficient object detection model built upon the YOLOv11 architecture. TI-YOLO introduces the Bidirectional Feature Pyramid Network (BiFPN) module for multi-scale feature fusion, effectively enhancing the ability to detect anatomical structures of different sizes. TI-YOLO integrates the Deformable Attention Transformer (DAT) module to enhance the perception of crucial regions, improving detection accuracy and robustness. To further reduce the consumption of computational resources while maintaining efficiency, TI-YOLO is optimized by reconstructing the backbone based on MobileNetV4. Furthermore, TI-YOLO employs the Slide Weight Function (SWF) as a loss function during model training to mitigate the class imbalance within the dataset. One self-built dataset is used to validate the effectiveness of TI-YOLO. Compared to the original YOLOv11, TI-YOLO achieves mean Average Precision at IoU 0.50 (mAP50) scores of 0.902, with improvements of 3.8%. Meanwhile, TI-YOLO balances detection accuracy and computational efficiency with a 10.5% reduction in floating-point operations (FLOPs) and a 28.9% reduction in parameters, and the model weight is only 4.6 MB. Additionally, to evaluate TI-YOLO real-time inference capability, we quantize and deploy it on a low-cost embedded OrangePi 5 platform. The inference speed reaches over 50 frames per second (FPS), meeting real-time clinical requirements.

Bioengineering doi: 10.3390/bioengineering13040450

Authors: Shrouk Mohamed Osman Ahmed Alksas Hossam Magdy Balaha Ali Mahmoud Ahmed Gamal Mohamed El-Said Abdel-Hady Mohamed Moawad Abdelsalam Abeer Twakol Khalil Ashraf Sewelam Ayman El-Baz

Diabetic retinopathy (DR) remains a major cause of vision loss in patients with diabetes, and earlier recognition of retinal vascular abnormalities may improve risk stratification and clinical follow-up. Optical coherence tomography angiography (OCTA) provides a noninvasive way to visualize the retinal microvasculature and may detect DR-related changes before they are evident on routine clinical assessment. In this work, we investigated whether dividing OCTA images into anatomically defined retinal regions could improve DR classification and clarify which regions carry the greatest discriminative information. The study included 188 OCTA images: 67 from normal eyes, 57 from eyes with mild DR, and 64 from eyes with moderate DR. Each image was divided into seven concentric regions centered on the fovea, and vessel-density features were extracted from each region. Ten machine learning classifiers were trained and compared at the regional level. For each region, the best-performing classifier was retained, and the final prediction was obtained with a majority-voting ensemble. To examine model behavior, Local Interpretable Model-Agnostic Explanations (LIME) were applied. Performance was also compared with that of a transfer-learning MobileNet model trained on whole OCTA images. On the held-out patient-level test set, the ensemble model achieved 97% accuracy, 98% precision, 97% recall, and a 97% F1-score for three-class classification. These results were higher than those obtained with the tested whole-image transfer-learning baselines. The interpretability analysis consistently identified the parafoveal regions as the most informative for classification. Among the seven regions, Region 3 showed the highest overall contribution, followed by Regions 2 and 5, whereas Region 5 became more influential in moderate DR. These results suggest that regional analysis of OCTA-derived vessel density can improve both classification performance and interpretability in DR assessment. The findings also indicate that parafoveal vascular alterations carry substantial discriminative value in distinguishing normal, mild DR, and moderate DR cases. Validation in larger, independent cohorts from multiple centers will be necessary to confirm the generalizability of these findings.

Bioengineering doi: 10.3390/bioengineering13040449

Authors: Rana Gunoz Comert Ravza Yilmaz Eda Cingoz Zuhal Bayramoglu Aysel Bayram Baran Mollavelioglu Mahmut Muslumanoglu Ulas Bagci

Purpose: To compare the efficacy of Superb Microvascular Imaging (SMI) with grayscale ultrasound (US) and dynamic contrast-enhanced MRI in predicting pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) in invasive breast cancer. Methods: A total of 115 patients included in the study were evaluated based on their pre-treatment imaging findings (US, mammography, and MRI). Following completion of NAC, all patients underwent grayscale US and SMI examinations. In patients with available post-NAC MRI, treatment response was additionally assessed by comparing MRI findings. Imaging results were correlated with postoperative pathological outcomes, which served as the reference standard. pCR was defined as the absence of residual invasive carcinoma, regardless of ductal carcinoma in situ. Molecular subtype, Ki-67, and axillary status were recorded. Statistical analyses included chi-square tests and stepwise multiple logistic regression. Significance was set at p < 0.05 (95% CI). Results: The median age was 51 years (range: 30–75). Most tumors were high-grade (55%) and invasive ductal carcinoma (95%). Breast-pCR was achieved in 43% of patients. Significant predictors of pCR included hormone receptor negativity, HER-2 positivity, high Ki-67 expression (≥40%), non-luminal subtype, and complete radiologic response on US and MRI (p < 0.05). Lower SMI index values were strongly associated with pCR (p < 0.001), with an optimal cut-off of 1.8 demonstrating good diagnostic performance (AUC = 0.804, 95% CI: 0.721–0.887). In multivariate analysis, the combined model including US, SMI, HER-2 status, and MRI showed the highest predictive performance (AUC = 0.890, 95% CI: 0.829–0.950), explaining 55.1% of the variance in pCR. Conclusions: An SMI index < 1.8, HER-2 positivity, and complete response on US and MRI are independent predictors of pCR after NAC. Combining SMI with multimodal imaging significantly improves predictive accuracy.

Bioengineering doi: 10.3390/bioengineering13040448

Authors: Lawrence Wing Chi Chan

Artificial intelligence (AI) plays an important role in bioengineering that has promoted a paradigm shift in how disease diagnosis, treatment planning, and patient monitoring are performed [...]

Bioengineering doi: 10.3390/bioengineering13040447

Authors: Anja Heselich Ramin Najafi Sami Alammawi Joanna Śmieszek-Wilczewska Shahram Ghanaati

Alveolar ridge preservation using biomaterials is a well-established approach to counteract post-extraction bone resorption and optimize conditions for implant placement. However, most studies rely only on a single evaluation method and thereby risk overlooking essential aspects of alveolar regeneration. This preliminary analysis aimed to assess alveolar ridge preservation outcome using a multimodal approach combining histomorphometric, radiological, and image-based visualization methods. Twenty out of a planned 60 patients per group from an ongoing randomized controlled clinical trial were included. Patients were allocated to alveolar ridge preservation with a bone substitute material (BSM), a collagen-based material, a combination of both, or natural healing as control. Outcomes included CBCT-based volumetric analysis, histomorphometry, and primary implant stability via ISQ. Mineralized bone volume was significantly better preserved in bone substitute material groups compared to other groups, with BSM combined with collagen yielding the highest values. Histomorphometrically determined hard tissue fractions and implant stability were comparable across groups. Notably, CBCT-based visualization revealed non-ossified hypodense regions, so-called cavitations or covered socket residuum (CSR) within the former extraction sockets across all groups, independent of the biomaterial applied. BSM-based alveolar ridge preservation, particularly combined with a collagen membrane, most effectively maintained mineralized bone volume after extraction. Beside volumetric benefits, this preliminary in-dept analysis of the first part of the trial highlights cavitations/CSRs as a potentially underrecognized feature of post-extraction healing. Integrating quantitative with qualitative visualization-based assessments provides a more complete understanding of alveolar bone regeneration.

Bioengineering doi: 10.3390/bioengineering13040446

Authors: Luis Roberto Mercado-Diaz Javier O. Pinzon-Arenas Paul A. Constable Irene O. Lee Lynne Loh Dorothy A. Thompson Hugo F. Posada-Quintero

Objective biomarkers for neurodevelopmental disorders remain an unmet clinical need. The electroretinogram (ERG), a non-invasive recording of the retinal response to light, has shown promise as a physiological marker for autism spectrum disorder (ASD) and attention deficit/hyperactivity disorder (ADHD), yet existing classification approaches based on time-domain and time–frequency features achieve limited accuracy in clinically relevant multi-group scenarios. This study introduces ERG-Graph, a novel graph signal processing (GSP) framework that transforms each ERG waveform into a weighted, undirected graph through amplitude quantization and temporal-adjacency connectivity. Nine topological and spectral features, including total load centrality, clique number, algebraic connectivity, and clustering coefficient, were extracted from each graph to characterize the structural dynamics of the signal. Using light-adapted ERG recordings from 278 participants (ASD = 77, ADHD = 43, ASD + ADHD = 21, Control = 137), we evaluated these features across binary, three-group, and four-group classification scenarios using seven machine learning classifiers with 10-fold subject-wise cross-validation. The proposed ERG-Graph features achieved balanced accuracies of 0.91 (ASD vs. control, males) and 0.88 (ADHD vs. control, females). Critically, fusing ERG-Graph with time-domain features yielded a balanced accuracy of 0.81 for three-group classification (ASD vs. ADHD vs. control), representing an 11-percentage-point improvement over the previous benchmark of 0.70. Statistical analysis confirmed significant topological differences between groups (Kruskal–Wallis, p < 0.001; Cliff’s delta: large effect sizes), and SHAP analysis revealed that graph-theoretic features dominated the top-ranked predictors. These results demonstrate that graph-based topological features capture discriminative information in the ERG waveform that is inaccessible to conventional signal analysis methods, advancing the development of objective biomarkers for neurodevelopmental disorder screening.

Bioengineering doi: 10.3390/bioengineering13040445

Authors: Wissem Dhahbi

Aim: Conventional free-fall kinematic models applied to plyometric push-up assessment treat the upper body as a vertically translating point mass, ignoring the curvilinear trajectory imposed by the ankle pivot and systematically biasing flight-time and height estimates. Methods: A planar rigid-body pendulum pivoting about the ankle axis was formulated via two independent derivation pathways (static moment equilibrium and a gravitational-torque coordinate approach), yielding effective pendulum length L = (MW/M) × LOS. Closed-form expressions for flight time, arc displacement, maximum height, and mean mechanical power were derived analytically from energy conservation and compared against free-fall predictions across seven pendulum arm lengths (LOW = 0.50–2.00 m) and 500 initial hand velocities per length, using adaptive Gauss–Kronrod quadrature (relative tolerance 10−10) with ODE cross-validation (maximum discrepancy < 2.5 × 10−7 s). Results: Flight time equivalence (tH = tG) was formally established. The free-fall model overestimated flight time by up to 18.82% (Δt = 0.096 s; LOW = 0.50 m, VH,0 = 2.50 m/s) and maximum height by up to 28.43% (Δh = 0.087 m; LOW = 0.50 m, tflight = 0.50 s), with both errors growing nonlinearly with initial velocity. Overestimation in height was proportionally larger at shorter pendulum arm lengths (18.18% at tflight = 0.30 s for LOW = 0.50 m vs. 10.91% for LOW = 1.00 m). Conclusions: The pendulum model provides a physically consistent, analytically tractable framework for geometry-adjusted upper-body power assessment from four field-obtainable anthropometric inputs. These results reflect computational self-consistency; prospective experimental validation against force-plate kinematics is required before applied deployment. Prospective empirical validation against dual force-plate and motion-capture reference data is required to establish the model’s accuracy boundaries under real push-up kinematics.

Bioengineering doi: 10.3390/bioengineering13040444

Authors: Abdulrahman Aloraynan Eunice Chu Jishen Wang Dawood Alsaedi Dayan Ban

According to the International Diabetes Federation, 589 million adults worldwide live with diabetes in 2025 (approximately 1 in 9 adults). The development of convenient noninvasive blood glucose monitoring systems has been a central focus in diabetes management. Optical spectroscopy has advanced significantly among all noninvasive glucose detection techniques. A photoacoustic system has been developed using a single-wavelength near-infrared laser, operating at 1625 nm, where glucose exhibits an overtone absorption band with relatively low water interference. The noninvasive system has been evaluated using artificial skin phantoms, with different glucose concentrations, covering both normoglycemic and hyperglycemic blood glucose levels. The detection sensitivity of the developed system has been enhanced to ±15 mg/dL across the entire clinically relevant glucose range. K-nearest neighbours and wide neural network machine learning models were developed for noninvasive glucose classification. The models achieved prediction accuracies of 80.0% and 81.5%, respectively, with 100% of the predicted data located within zones A and B of Clarke’s error grid analysis. These findings satisfy the regulatory requirements for glucose monitors established by Health Canada and the U.S. Food and Drug Administration.

Bioengineering doi: 10.3390/bioengineering13040442

Authors: Franck Simon Christophe Caneparo Jadson Moreira-Pereira Stéphane Chabaud

The rising popularity of sports practiced without adequate preparation has increased the incidence of anterior cruciate ligament (ACL) injuries, particularly among young individuals. Because the ACL has a very limited intrinsic healing capacity, surgical reconstruction—most often using autologous grafts—remains the standard of care. However, current techniques frequently lead to donor-site morbidity and do not consistently restore long-term joint stability, contributing to early post-traumatic osteoarthritis in active patients. Over the past decades, tissue engineering (TE) has opened promising avenues for developing biological substitutes capable of overcoming these limitations. Despite substantial progress, no strategy has yet demonstrated reliable and clinically validated functional regeneration of the human ACL. Meanwhile, artificial intelligence is emerging as a complementary tool for diagnosis, surgical planning, biomechanical assessment, and personalized reconstruction strategies. This review aims to provide a comprehensive overview of current TE-based approaches for ACL repair and reconstruction, analyzes their biological and biomechanical limitations, and discusses emerging concepts that may enhance future clinical outcomes. We first summarize the fundamental principles of tissue engineering, then examine the major strategies proposed for ACL regeneration—highlighting their respective strengths and shortcomings—and finally outline perspectives for a novel approach currently under development.

Bioengineering doi: 10.3390/bioengineering13040443

Authors: Marcelo A. F. Ribeiro Sharon Henry

Necrotizing soft tissue infections (NSTIs) remain one of the most unforgiving diseases encountered in modern surgical practice [...]

Bioengineering doi: 10.3390/bioengineering13040441

Authors: Asha Parmar Shantanu Chauhan Sora Alghziwatalkhawaldh Kanwarpal Singh

This paper introduces a laser interferometric elastography (LIE) system that uses a narrow linewidth laser and a single photodetector to measure mechanical displacements induced by surface acoustic waves (SAWs) generated by an electrically driven piezoelectric transducer. The method relies on phase delay analysis of the resulting interference signal to determine displacement within the medium, thereby eliminating the need for complex interferometers and broadband light sources. By substantially reducing optical hardware requirements, the system provides a compact and cost-effective platform for elasticity mapping in biological samples. Quantitative assessment of mechanical properties is achieved through controlled mechanical excitation and phase-resolved signal collection, demonstrating the practicality of simplified LIE for real-world applications.

Bioengineering doi: 10.3390/bioengineering13040440

Authors: Hong-Yun Ou Takahiro Hasegawa Osamu Fukayama Eizo Miyashita

Brain–machine interfaces (BMIs) aim to decode motor intentions from neural activity to enable direct control of external devices. However, most existing decoders rely on monolithic architectures that fail to capture the distinct neural representations of different joint movement directions, limiting their generalizability. In this work, we propose a Single-Direction CNN-LSTM decoder inspired by motor cortex encoding mechanisms, which separately models extension and flexion dynamics through parallel CNN-LSTM branches. Each branch extracts spatial–temporal features from neural spike data and predicts directional joint variables, which are then combined by subtraction to yield the net angular velocity and torque of upper-limb joints. Using invasive recordings from a macaque during a 2D center-out reaching task, we demonstrate that our decoder achieves comparable performance to a conventional CNN-LSTM when trained on all tasks, while significantly outperforming both CNN-LSTM and linear regression baselines in cross-target generalization scenarios. Moreover, the model can capture physiologically meaningful co-contraction patterns, providing richer insights into motor control. These results suggest that incorporating neuroscience-inspired modular decoding into deep neural architectures enhances robustness and adaptability across tasks, offering a promising pathway for BMI applications in prosthetics and rehabilitation.

Bioengineering doi: 10.3390/bioengineering13040439

Authors: Gaia Roccaforte Arianna Sinardi Sofia Ruello Carmela Lipari Flavio Corpina Antonio Epifanio Anna Isgrò Francesco Davide Russo Alfio Puglisi Giovanni Pioggia Flavia Marino

Telerehabilitation—the remote delivery of rehabilitation services—is undergoing a paradigm shift with the convergence of immersive virtual reality (VR) and wearable biosensor technologies. This perspective article outlines a vision for home-based motor and cognitive rehabilitation that is engaging, personalized, and data-driven. We describe how immersive VR environments (for example, simulations of home settings or supermarkets) coupled with wearable sensors can address current challenges in rehabilitation by increasing patient motivation, enabling real-time biofeedback, and supporting remote clinician supervision. Gamification mechanisms and rich sensory feedback in VR are highlighted as key strategies to enhance user engagement and adherence to therapy. We discuss conceptual innovations such as multi-sensor data integration, dynamic difficulty adaptation, and AI-driven personalization of exercises, derived from recent research and our development experience, and consider their potential benefits for patients with neuro-cognitive-motor impairments (e.g., stroke, Parkinson’s disease, and multiple sclerosis). Implementation scenarios for home-based therapy are presented, emphasizing scalability, standardized digital metrics for monitoring progress, and seamless involvement of clinicians via telehealth platforms. We also critically examine the current limitations of VR and telehealth rehabilitation and how an integrative model could overcome these barriers. More specifically, this perspective defines the engineering requirements of a closed-loop VR-based telerehabilitation framework, including multimodal data synchronization, calibration, signal-quality management, interpretable adaptive control, digital biomarker validation, and practical strategies to improve accessibility, privacy, and scalability in home-based neurological rehabilitation.

Bioengineering doi: 10.3390/bioengineering13040438

Authors: Alessandra Franco Francesca Angelone Danilo Calderone Alfonso Maria Ponsiglione Maria Romano Carlo Ricciardi Francesco Amato

This systematic review analyzes how the introduction and progressive deployment of 5G networks have influenced the evolution of telemedicine between 2014 and 2024, focusing on their impact on performance, accessibility, and the feasibility of advanced clinical applications across the pre-COVID-19, COVID-19, and post-COVID-19 periods. The review was conducted in accordance with PRISMA guidelines and included publications retrieved from SCOPUS, PubMed, and Web of Science using a PICO-based search strategy. Studies were selected based on predefined inclusion and exclusion criteria, and extracted data included clinical parameters, network characteristics such as bandwidth and latency, geographic setting, and type of telemedicine service. A total of 45 studies met the inclusion criteria, with most published between 2020 and 2024. The most frequently reported applications were telediagnosis, particularly robotic ultrasound, followed by telesurgery and teleconsultation. The low latency enabled by 5G networks supported complex telesurgical procedures over distances exceeding 5000 km, while in ultra-remote areas, hybrid solutions combining 5G and fiber-optic networks were often adopted to ensure stable connections. The integration of robotic platforms and AI-based tools further enhanced the precision and reliability of remote procedures. Overall, 5G technology has significantly advanced telemedicine by enabling real-time, high-quality care over long distances, improving access to specialist services and supporting more equitable and efficient digital healthcare delivery, particularly in underserved regions.

Bioengineering doi: 10.3390/bioengineering13040437

Authors: Zachary R. Davis P. Cameron Gossett Robert L. Wilson Woong Kim Yue Mei Kent D. Butz Nancy C. Emery Eric A. Nauman Stéphane Avril Corey P. Neu Deva D. Chan

Intervertebral disc degeneration is the most recognized cause of low back pain, characterized by the decline in tissue structure and mechanics. Image-based mechanical parameters (e.g., strain, stiffness) may provide an ideal assessment of disc function that is lost with degeneration, but unfortunately, these remain underdeveloped. Moreover, it is unknown whether strain or stiffness of the disc may be predicted by MRI relaxometry (e.g., T1 or T2), an increasingly accepted quantitative measure of disc structure. In this study, we quantified T1 and T2 relaxation times and compared to in-plane strains measured with displacement-encoded MRI within human cadaveric discs under physiological levels of compression and bending. Using a novel inverse approach, we then estimated shear modulus in orthogonal image planes and regionally compared these values to relaxation times and 2D strains. Intratissue strain depended on the loading mode, and shear modulus in the nucleus pulposus was typically an order of magnitude lower than the annulus fibrosus. Relative shear moduli estimated from strain data derived under compression generally did not correspond with those from bending experiments. Only one anatomical region showed a significant correlation between relative shear modulus and relaxometry (T1 vs. µrel, coronal plane under bending). Together, these results suggest that future inverse analyses may be improved by incorporating multiple loading conditions into the same model and that image-based elastography and relaxometry should be viewed as complementary measures of disc structure and function to assess degeneration in future studies.

Bioengineering doi: 10.3390/bioengineering13040436

Authors: Andrea Sambri Alessandro Bruschi Cristina Scollo Massimiliano De Paolis

Local antibiotic delivery has gained a central role as an adjunct to radical debridement in chronic osteomyelitis, allowing high antimicrobial concentrations at the infection site while reducing systemic toxicity. This narrative review summarizes the current clinical evidence on commercially available antibiotic-loadable bone substitutes, with particular focus on calcium sulfate (CaSO4)-based systems and biphasic calcium sulfate/hydroxyapatite (CaS/HA) composites. Nineteen studies were included. Differences in formulation, resorption kinetics, antibiotic elution profile and osteoconductive behavior are discussed, alongside clinical outcomes including recurrence of infection, reoperation rates and complication patterns. Finally, based on the currently available evidence and expert recommendations, practical guidance is proposed to support carrier selection in different clinical scenarios (cavitary vs. corticomedullary defects; high-risk soft tissue; polymicrobial or resistant infections). Across published series, although heterogeneous, infection eradication rates are generally high when local carriers are integrated into structured surgical protocols. Calcium sulfate carriers provide rapid resorption and robust early antibiotic release but are associated with higher rates of sterile wound drainage. In contrast, CaS/HA biocomposites demonstrate more gradual remodeling and radiographic integration, potentially improving defect consolidation and reducing wound-related morbidity, although leakage and cost considerations remain relevant.

Bioengineering doi: 10.3390/bioengineering13040435

Authors: Ilaria Neri Francesca Gallivanone Elena Venturini Carla Canevari Chiara Caleri Nicole Rotmensz Samuele Ghezzo Carolina Bezzi Paola Mapelli Pietro Panizza Maria Picchio Rosa Di Micco Arturo Chiti Oreste Davide Gentilini Paola Scifo

Breast cancer (BC) requires the evaluation of tumor aggressiveness features to guide treatment decisions. Biopsy-derived prognostic information may differ from surgical histopathology due to tumor heterogeneity. Hybrid PET/MRI can provide additional information for tumor characterization, supporting initial therapy planning and prognosis. In this work, we acquired 157 BC patients using a hybrid PET/MRI scanner. The PET data were combined with ADC and semi-quantitative DCE-MRI metrics to derive “hybrid PET/MRI parameters.” Pathological data such as tumor grade, hormone receptors, proliferation index (Ki67), and surrogate molecular subtype were collected, and we evaluated their associations with hybrid imaging, also comparing with the PET and MRI data analyzed separately. Ki67 showed moderate correlations with PET, ADCmin, and most hybrid parameters. The PET and hybrid data differentiate histopathological factors, while ADCmin differentiates G1 vs. G2 and luminal A vs. luminal B. In the ROC analysis, hybrid SUVmax/ADCmin shows better performance to predict luminal B from luminal A (AUC 0.720, sensitivity 73.1%, specificity 63.2%, PPV 54.3%, NPV 79.7%) than SUVmean alone. Our findings suggest that these novel hybrid PET/MRI parameters may help the characterization of tumor tissue in IDC. However, a multivariate analysis is needed to confirm our preliminary results.

Bioengineering doi: 10.3390/bioengineering13040434

Authors: Yoshiko Bamba Michio Itabashi Hirotoshi Kobayashi Kenjiro Kotake Masayasu Kawasaki Yukihide Kanemitsu Yusuke Kinugasa Hideki Ueno Kotaro Maeda Takeshi Suto Kimihiko Funahashi Heita Ozawa Fumikazu Koyama Shingo Noura Hideyuki Ishida Masayuki Ohue Tomomichi Kiyomatsu Soichiro Ishihara Keiji Koda Hideo Baba Kenji Kawada Yojiro Hashiguchi Takanori Goi Yuji Toiyama Naohiro Tomita Eiji Sunami Yoshito Akagi Jun Watanabe Kenichi Hakamada Goro Nakayama Kenichi Sugihara Yoichi Ajioka

Background: Owing to significant clinical heterogeneity, the achievement of accurate survival forecasting for individuals with colorectal cancer and peritoneal metastasis continues to be a complex undertaking. We aimed to transcend traditional prognostic limitations by evaluating machine learning boosting models against standard regression-based methods in terms of estimating overall survival (OS). Methods: Utilizing a multi-institutional registry of 150 patients diagnosed with synchronous peritoneal metastasis of colorectal cancer, we integrated 124 clinicopathological variables to refine our predictive models. Beyond standard preprocessing—including standardization and median imputation—we rigorously compared XGBoost and LightGBM against Ridge, Lasso, and linear regression via five-fold cross-validation. To specifically address right-censoring, an XGBoost Cox model was implemented and validated using Harrell’s C-index, with SHAP and LIME providing essential model interpretability. Results: Boosting models consistently outperformed linear alternatives, which struggled with high error rates and negative R2 values. Specifically, XGBoost achieved an MAE of 475 ± 60 and an RMSE of 585 ± 88. The XGBoost Cox model reached a C-index of 0.64 ± 0.06. SHAP analysis highlighted inflammatory markers and peritoneal disease extent as the most influential prognostic drivers. Conclusions: While boosting models offer a clear accuracy advantage over linear methods, their prognostic power remains moderate. These findings underscore the potential of ensemble learning in oncology, yet mandate external validation before these tools can be integrated into clinical decision-making.

Bioengineering doi: 10.3390/bioengineering13040433

Authors: Anastasia Kourtesa Konstantinos Skarentzos Georgios S. Dimtsas Periklis G. Foukas Marilita Moschos

Recent evidence links lens epithelial cell (LEC) dysfunction and cellular senescence—an irreversible cell cycle arrest with a pro-inflammatory secretory phenotype—to age-related cataract (ARC) progression. This systematic review synthesizes current knowledge on LEC senescence, its molecular features, and laboratory methods for senescence assessment in the ARC. Following PRISMA guidelines, a comprehensive search of PubMed, Scopus and Cochrane databases retrieved 3417 records from inception to 9 February 2025, with 14 studies ultimately included (821 patients and multiple in vitro LEC models). The following multiple senescence expression pathways were identified: SA-β-gal activity, p53/p21 and p16INK4A pathway activation, mitochondrial dysfunction, oxidative stress, and secretion of senescence-associated secretory phenotype (SASP) factors. Notably, cortical cataract demonstrated direct association with local senescent cell accumulation, while nuclear cataract reflected cumulative oxidative damage from impaired LEC-mediated antioxidant defense. Senescence markers correlated positively with cataract severity across multiple studies. Several potential therapeutic targets emerged, including metformin (AMPK activation/autophagic restoration), circMRE11A silencing, NLRP3 inflammasome inhibition, and modulation of FYCO1/PAK1 and MMP2 pathways. This review establishes LEC senescence as a central process in ARC pathogenesis and highlights promising senotherapeutic approaches. Future research should prioritize human surgical samples, develop standardized senescence detection panels (SA-β-gal + p21/p16 + SASP factors), and conduct longitudinal studies to establish causal relationships between senescence accumulation and cataract progression.