Bioengineering

15 pages, 3064 KB

Open AccessArticle

Effect of Axial and Coronal Malalignment on Polyethylene Stresses in Lumbosacral Total Joint Replacement: A Finite Element Study

by Steven A. Rundell, Steven M. Kurtz, Hannah Spece, Scott D. Hodges and Ron V. Yarbrough

Bioengineering 2026, 13(9), 987; https://doi.org/10.3390/bioengineering13090987 (registering DOI) - 27 Aug 2026

Abstract

The purpose of this study was to analyze the sensitivity of polyethylene stresses and strains, as proxies for wear and surface damage, to coronal and axial misalignment for a novel lumbar total joint replacement (LTJR) implanted at L5-S1. We hypothesized that these stresses [...] Read more.

The purpose of this study was to analyze the sensitivity of polyethylene stresses and strains, as proxies for wear and surface damage, to coronal and axial misalignment for a novel lumbar total joint replacement (LTJR) implanted at L5-S1. We hypothesized that these stresses and strains would remain below the levels associated with worst-case impingement in a spine wear simulator. A finite element model (FEM) of the L4-S1 spine was developed using CT-based anatomy from a representative Investigational Device Exemption study patient. An appropriately sized LTJR was virtually implanted at L5-S1 in a standardized baseline orientation, and component positioning was altered to achieve eight independent misalignment scenarios, spanning axial convergence angle, anterior–posterior offset, and coronal tilt. Physiologic forward-bending loads were applied, and polyethylene contact pressure, von Mises stress, and effective strain were recorded. Peak values for the baseline scenario were 35.0 MPa, 19.9 MPa, and 4.8%, respectively. Convergence angle and anterior–posterior offset had minimal effect, whereas coronal tilt produced the largest changes. Contact pressure and von Mises stress varied by <20% due to misalignment, whereas effective strain was more sensitive, increasing from 4.8% at baseline to 9.9% with 20 degrees of coronal tilt. All values were comparable to prior L4-L5 findings and remained well below impingement levels. Within the bounds of reasonable misalignment, the device maintained bearing congruency and did not approach conditions associated with accelerated wear. Full article

(This article belongs to the Special Issue Bioengineering Technologies for Spine Research)

► Show Figures

Figure 1

20 pages, 1154 KB

Open AccessArticle

AI-Driven Three-Dimensional Reconstruction of Fetal Cardiovascular Pathways from Ultrasound Sweep Videos for Congenital Heart Disease Screening

by Naoaki Harada, Masaaki Komatsu, Naoki Teraya, Shohei Wakisaka, Katsuji Takeda, Takashi Natsume, Tomonori Taniguchi, Nobuji Kouno, Satoshi Takahashi, Ken Asada, Syuzo Kaneko, Mayumi Kaneko, Reina Komatsu, Kazuki Iwamoto, Ryu Matsuoka, Akihiko Sekizawa and Ryuji Hamamoto

Bioengineering 2026, 13(9), 986; https://doi.org/10.3390/bioengineering13090986 - 26 Aug 2026

Abstract

Fetal cardiac ultrasound screening is vital for the prenatal diagnosis of congenital heart disease (CHD). However, understanding complex cardiovascular structures and continuity from two-dimensional (2D) ultrasound images remains challenging. Therefore, this study developed and evaluated an artificial intelligence (AI)-based three-dimensional (3D) visualization approach [...] Read more.

Fetal cardiac ultrasound screening is vital for the prenatal diagnosis of congenital heart disease (CHD). However, understanding complex cardiovascular structures and continuity from two-dimensional (2D) ultrasound images remains challenging. Therefore, this study developed and evaluated an artificial intelligence (AI)-based three-dimensional (3D) visualization approach that reconstructs fetal cardiovascular structures from 2D ultrasound sweep video. To this end, information from the adjacent frames was used. Instead of interpreting each ultrasound frame independently, this method incorporates neighboring frames, enabling the continuous visualization of the right- and left-sided functional circulatory pathways. Our method achieved a mean Dice coefficient of 0.785 on an independent test dataset, demonstrating that the adjacent frame information can provide stable 3D segmentation. Diffusion-based anatomical plausibility analysis trained exclusively on normal anatomies showed exploratory discrimination between normal fetuses and fetuses with CHD (AUC, 0.861; 95% CI, 0.695–0.980). In a clinical comparison study involving 12 obstetricians, the introduction of this AI-driven 3D cardiovascular visualization slightly improved the screening accuracy from 0.779 to 0.808 and significantly improved the confidence-weighted accuracy score (p = 0.002). Our approach can potentially support examiners and facilitate fetal cardiac ultrasound screenings. Full article

(This article belongs to the Special Issue Artificial Intelligence in Biomedical Engineering Applications Towards Clinical Translation)

19 pages, 2529 KB

Open AccessArticle

An Explainable Machine Learning Framework for Predicting Hearing Aid Satisfaction: Integrating the HATASS Instrument and Clinical Insights

by Seyma Arslanbas, Tahir Cetin Akinci, Ümit Can Çetinkaya and Sengul Terlemez

Bioengineering 2026, 13(9), 985; https://doi.org/10.3390/bioengineering13090985 - 26 Aug 2026

Abstract

Hearing aid technology adaptation and user satisfaction are influenced by multiple interacting demographic, clinical, and behavioral factors, making reliable prediction of outcomes challenging with conventional statistical approaches alone. This study proposes an explainable machine learning framework to investigate the multidimensional determinants of hearing [...] Read more.

Hearing aid technology adaptation and user satisfaction are influenced by multiple interacting demographic, clinical, and behavioral factors, making reliable prediction of outcomes challenging with conventional statistical approaches alone. This study proposes an explainable machine learning framework to investigate the multidimensional determinants of hearing aid satisfaction by integrating demographic characteristics, hearing aid-related variables, and patient-reported outcomes obtained from the Hearing Aid Technology Adaptation and Satisfaction Scale (HATASS). Five regression algorithms—Linear Regression, Decision Tree Regression (DTR), Random Forest Regression (RFR), Support Vector Regression, and Gradient Boosting Regression (GBR)—were comparatively evaluated using a five-fold cross-validation strategy. Predictive performance was assessed using the root mean square error (RMSE), mean absolute error (MAE), and the coefficient of determination (

R^{2}

), while model interpretability was investigated through cross-validated out-of-bag permutation feature importance analysis. Among the evaluated algorithms, Random Forest Regression achieved the most consistent predictive performance, yielding the lowest average RMSE (14.225) and the highest average

R^{2}

(0.146) under the adopted validation framework. Although the overall predictive performance remained modest, the explainability analysis consistently identified age as the most influential predictor, followed by onset year, education level, hearing aid usage duration, and daily hearing aid use. In contrast, gender, battery type, tinnitus, and vertigo contributed comparatively less to model predictions. These findings indicate that hearing aid adaptation and satisfaction arise from complex nonlinear interactions among demographic, behavioral, clinical, and device-related characteristics rather than isolated linear associations. The proposed framework provides an interpretable, internally validated analytical approach for investigating hearing aid technology adaptation and satisfaction and establishes a foundation for future studies that integrate comprehensive audiological measurements, longitudinal follow-up data, and independent external validation to support the development of more transparent and personalized hearing healthcare systems. Full article

(This article belongs to the Section Biosignal Processing)

► Show Figures

Graphical abstract

31 pages, 1530 KB

Open AccessArticle

Edge-Deployable Lightweight Deep Learning for Hypertensive Retinopathy Grading from En-Face OCT: A Patient-Level Feasibility Study

by Süleyman Burçin Şüyun, Mustafa Yurdakul, Şakir Taşdemir and Serkan Biliş

Bioengineering 2026, 13(9), 984; https://doi.org/10.3390/bioengineering13090984 - 26 Aug 2026

Abstract

Background: Hypertensive retinopathy (HR) is an early marker of hypertension-related end-organ damage, and multi-grade classification is limited by overlap between the early stages. No prior study has reported edge-deployable HR grading from optical coherence tomography (OCT). We assess patient-level–validated four-grade HR classification from [...] Read more.

Background: Hypertensive retinopathy (HR) is an early marker of hypertension-related end-organ damage, and multi-grade classification is limited by overlap between the early stages. No prior study has reported edge-deployable HR grading from optical coherence tomography (OCT). We assess patient-level–validated four-grade HR classification from en-face OCT with on-device inference. Methods: MobileViT-XXS and EfficientNetV2-B0 were evaluated on 478 en-face OCT images from 221 patients. To avoid leakage, all partitions were patient-level: stratified group 5-fold cross-validation with a patient-disjoint test set. Each model was evaluated as a 5-fold soft-voting ensemble, with inference profiled on an NVIDIA Jetson Orin Nano. Accuracy, weighted F1, quadratic-weighted kappa, and ROC–AUC were reported; the models were compared by McNemar’s test, and clinical utility by referable-HR (Grade ≥ 2) triage. Results: MobileViT-XXS achieved 84.4% accuracy, ROC–AUC 0.944, and quadratic-weighted kappa 0.916, significantly outperforming EfficientNetV2-B0 (McNemar p = 0.013). Mild-grade overlap limited four-class accuracy, but referable-HR triage reached 95.6% accuracy, 95.2% sensitivity, and 95.8% negative predictive value. MobileViT-XXS required only 3.83 MB versus 23.70 MB, and on-device TensorRT FP16 five-fold ensemble inference achieved 10.4 ms (96 FPS) at 8.9 W. Conclusions: Lightweight models enable feasible, power-efficient point-of-care HR triage from en-face OCT; broader multi-center validation is warranted. Full article

(This article belongs to the Special Issue Application of Deep Learning in Medical Image Classification and Segmentation for Precise Disease Diagnosis)

27 pages, 15343 KB

Open AccessArticle

AMFF-Net: An Adaptive Multi-Layer Feature Fusion Network Based on ConvNeXtV2-B for Medical Image Classification

by Min-Seo Kim and Hyoung-Gook Kim

Bioengineering 2026, 13(9), 983; https://doi.org/10.3390/bioengineering13090983 - 26 Aug 2026

Abstract

Although ConvNeXtV2 has shown promising performance in medical image classification, approaches relying primarily on final-stage features may underutilize low-level structural and complementary hierarchical information. To address this limitation, we propose an Adaptive Multi-Layer Feature Fusion Network (AMFF-Net) based on ConvNeXtV2-B for medical image [...] Read more.

Although ConvNeXtV2 has shown promising performance in medical image classification, approaches relying primarily on final-stage features may underutilize low-level structural and complementary hierarchical information. To address this limitation, we propose an Adaptive Multi-Layer Feature Fusion Network (AMFF-Net) based on ConvNeXtV2-B for medical image classification. The proposed framework employs a Feature Alignment (FA) module to project multi-stage features into a unified representation space and an Adaptive Multi-Layer Feature Fusion (AMFF) module to compute a single input-dependent scalar weight for each stage and dynamically adjust the relative contributions of hierarchical features. An Efficient Channel Attention (ECA) module is subsequently incorporated to enhance the fused representation through lightweight channel-wise recalibration. AMFF-Net was evaluated on three medical image classification datasets: Kvasir-v2, HAM10000, and ChestXray14. Experimental results demonstrate that AMFF-Net consistently improves classification performance over the baseline ConvNeXtV2-B and achieves competitive performance compared with representative convolutional neural network (CNN)- and Transformer-based architectures, while incurring relatively modest additional computational overhead. Ablation results further support the contribution of FA, AMFF, and ECA to the overall performance of the proposed framework. Full article

(This article belongs to the Special Issue AI and Machine Learning in Medical Image Processing: Innovations and Applications—2nd Edition)

► Show Figures

Figure 1

41 pages, 4581 KB

Open AccessArticle

Multimuscle Surface-EMG Characterization of Upper-Limb Fatigue During Repetitive Haptic Interaction for Health 5.0 Applications

by Mohammad Alja’afreh, Nasser Mustafa and Ali Karime

Bioengineering 2026, 13(9), 982; https://doi.org/10.3390/bioengineering13090982 - 26 Aug 2026

Abstract

Health 5.0 increasingly involves medical robots and haptic systems that sustain physical interaction with patients and clinicians. This proof-of-concept study examined fatigue-related surface electromyography (sEMG) spectral changes during a 400 s repetitive haptic-writing task in 20 adults. Five upper-limb muscles were monitored at [...] Read more.

Health 5.0 increasingly involves medical robots and haptic systems that sustain physical interaction with patients and clinicians. This proof-of-concept study examined fatigue-related surface electromyography (sEMG) spectral changes during a 400 s repetitive haptic-writing task in 20 adults. Five upper-limb muscles were monitored at 1000 Hz, and mean frequency (MNF) and median frequency (MDF) trajectories were summarized by fitted start-to-end spectral decline and combined into arm-level indices. MDF had the higher association with the archived participant-level fatigue-analysis score (

r = 0.954

versus

r = 0.783

;

Δ r = 0.171

; Holm-adjusted

p = 0.0035

). Small-sample influence analysis supported the stability of this within-sample ordering: after omitting each participant in turn,

Δ r

remained positive in 20/20 analyses (range 0.125–0.218), although the exact paired label-swap sensitivity test remained inconclusive (

p = 0.082

). Exploratory leave-one-participant-out calibration produced lower held-out error for MDF (MAE 4.38, RMSE 5.87 percentage points) than for MNF (MAE 9.24, RMSE 12.02 percentage points). The raw seven-category Q2 responses and the archived 0–100 analysis score are reported as separate data products because they are not numerically equivalent under direct linear rescaling. Accordingly, the results identify MDF as the more promising spectral summary for prospective validation in this task, rather than establishing universal superiority or numerical replacement of subjective fatigue. The experiment was a laboratory haptic-writing study and did not test a surgical robot, rehabilitation robot, patient population, or clinical controller; Health 5.0 is therefore presented as a translational motivation rather than a demonstrated application. Full article

(This article belongs to the Special Issue Progress in Biosensors for Biomedical Engineering Applications)

► Show Figures

Figure 1

21 pages, 7587 KB

Open AccessPerspective

Translational Initiative to Minimize Childhood Hemiparesis After Asymmetrical Perinatal Brain Injury

by Susan V. Duff, Emily Garavatti and Terrie Inder

Bioengineering 2026, 13(9), 981; https://doi.org/10.3390/bioengineering13090981 - 26 Aug 2026

Abstract

The purpose of this perspective paper is to present a translational approach aimed at identifying the critical periods in brain development during which to provide neuroprotective and neurorestorative strategies, followed by prehensile training, to minimize hemiparesis and foster skill acquisition after asymmetrical perinatal [...] Read more.

The purpose of this perspective paper is to present a translational approach aimed at identifying the critical periods in brain development during which to provide neuroprotective and neurorestorative strategies, followed by prehensile training, to minimize hemiparesis and foster skill acquisition after asymmetrical perinatal brain injury (APBI). We contend that hemiparesis could be minimized after APBI through a multimodal sequential approach based on timely implementation of evidence-based intervention strategies. If APBI is identified early, neuroprotective techniques can be used in the first 10 days of life to minimize cell death. This can be followed in the next 3 months by the provision of timely neurorestorative strategies to enhance neuron survival and promote brain growth. Ideally these strategies will prime the system for targeted prehensile training from 3 months and beyond. Building on animal work, scientifically sound pilot work in human infants is needed to determine the best timing and therapy dose verified with sensitive outcome measures available before widespread implementation. Strategic planning among a multidisciplinary group working in basic science, medicine, and rehabilitation will inform efforts to bring this translational approach to clinical fruition. Full article

(This article belongs to the Special Issue Technological Advances in Neurorehabilitation)

► Show Figures

Figure 1

21 pages, 4726 KB

Open AccessArticle

Optimization of a 3D Skin Equivalent Incorporating Melanocytes or Melanoma Cells for In Vitro Melanoma Research

by Sylwia Hasterok, Skaidre Jankovskaja, Zdenka Prgomet, Lars Ohlsson and Anna Gustafsson

Bioengineering 2026, 13(9), 980; https://doi.org/10.3390/bioengineering13090980 - 26 Aug 2026

Abstract

Replicating the complex physiology of human skin in vitro remains a challenge for cutaneous oncology. Although several three-dimensional (3D) skin and melanoma models have been developed, reproducible skin equivalent systems that enable direct and standardized comparisons between healthy and melanoma-containing skin remain valuable [...] Read more.

Replicating the complex physiology of human skin in vitro remains a challenge for cutaneous oncology. Although several three-dimensional (3D) skin and melanoma models have been developed, reproducible skin equivalent systems that enable direct and standardized comparisons between healthy and melanoma-containing skin remain valuable experimental tools. This study aimed to optimize and characterize skin equivalents containing either melanocytes or melanoma cells and to evaluate their utility as platforms for investigating melanoma-associated skin biology and UVB-induced responses in vitro. To achieve this, 3D skin equivalents containing fibroblasts, keratinocytes, and either melanocytes or melanoma cells were reconstructed on a polystyrene scaffold to generate healthy (mc) and melanoma (mm) models. Morphological characteristics were compared with clinically verified human tissue sections, and functional responses to UVB irradiation were assessed with a focus on the kynurenine pathway. The models demonstrated a distinct dermal–epidermal architecture and distinguishable melanoma-associated features, including epidermis-confined melanoma cell clusters, in mm constructs. UVB exposure induced differential responses between models, including significant differences in kynurenine pathway regulation. These findings suggest that the developed skin equivalents provide a reproducible in vitro platform for studying melanoma-associated skin biology and treatment-related metabolic responses and may support future mechanistic studies of skin cancer progression. Full article

(This article belongs to the Special Issue Advances in Three-Dimensional Culture Systems: Bridging Biology and Medicine)

► Show Figures

Graphical abstract

20 pages, 11228 KB

Open AccessArticle

Shared-Scale Predictive Benchmarking of an Acoustic-Radiation-Force Model for Frequency-Dependent Retinal Ganglion Cell Responses

by Bingao Zhang and Shengyong Xu

Bioengineering 2026, 13(9), 979; https://doi.org/10.3390/bioengineering13090979 - 26 Aug 2026

Abstract

Retinal responses to ultrasound depend on carrier frequency, but whether a constrained tissue-to-neuron model can predict this dependence on a common published response scale remains unclear. We developed an acoustic-radiation-force (ARF) modelling workflow that couples an effective tissue-scale stress proxy to a stochastic [...] Read more.

Retinal responses to ultrasound depend on carrier frequency, but whether a constrained tissue-to-neuron model can predict this dependence on a common published response scale remains unclear. We developed an acoustic-radiation-force (ARF) modelling workflow that couples an effective tissue-scale stress proxy to a stochastic Hodgkin–Huxley retinal ganglion cell population. One non-negative affine observation model, shared across frequencies, linked raw simulated spike counts to the published three-frequency ex vivo RGC response scale without frequency-specific rescaling. In a nested leave-one-frequency-out model-selection assessment, the workflow achieved an overall

Q^{2}

of 0.915 and an RMSE of 0.0752 across 25 held-out observations. Mean-response transfer remained strong across frequencies, although probabilistic calibration was weakest at 1.9 MHz. An in-sample fixed-map sensitivity analysis using moment-matched log-normal, Gamma, and Weibull thresholds yielded overall RMSEs of 0.0434–0.0461 and preserved positive overall K-ablation NLPD differences of 0.219–0.238, with the effect concentrated at 43 MHz. These results show that shared-scale modelling captures the main cross-frequency response structure and that a high-threshold inhibitory functional component improves the description of high-intensity rolloff. Within the published three-frequency dataset, the workflow provides a reproducible benchmark for targeted comparisons of retinal ultrasound mechanisms. Full article

(This article belongs to the Special Issue Mathematical and Computational Models in Biomedical Science and Engineering (3rd Edition))

► Show Figures

Figure 1

20 pages, 1041 KB

Open AccessReview

Machine Learning on the Pediatric Intensive Care (PIC) Database: PIC-Powered Prediction

by Hammad Ashraf Ganatra, Daniah Shamim, Shawn B. Sood and Amr Mohamed Ali

Bioengineering 2026, 13(9), 978; https://doi.org/10.3390/bioengineering13090978 - 26 Aug 2026

Abstract

Machine learning (ML) applied to pediatric intensive care data could enable earlier risk stratification and more precise decision support, but limited shareable pediatric datasets have constrained progress. The Pediatric Intensive Care database (PIC) on PhysioNet provides a de-identified, bilingual electronic health record resource [...] Read more.

Machine learning (ML) applied to pediatric intensive care data could enable earlier risk stratification and more precise decision support, but limited shareable pediatric datasets have constrained progress. The Pediatric Intensive Care database (PIC) on PhysioNet provides a de-identified, bilingual electronic health record resource spanning 2010–2018. In this narrative review we synthesize ten PIC-based ML studies identified through forward citation tracking on the original PIC publication and PhysioNet dataset record in PubMed, Web of Science, and Google Scholar. For each study we abstracted the clinical question, cohort, label construction, feature engineering, model class, validation design, calibration reporting, and release of reproducibility artifacts. The reviewed studies addressed catheter-associated thrombosis, sepsis, in-hospital mortality, and organ-dysfunction phenotyping. We organize the synthesis around four substrate properties of PIC: irregular time series, constructed labels, single-center temporal drift, and bilingual identifier semantics. The review advances three claims. First, upstream design choices appear to drive performance at least as much as classifier selection across the studies reviewed. Second, within-site discrimination metrics are insufficient without calibration, decision-curve analysis, and deployment-realistic validation. Third, PIC should be viewed as the seed for a collaborative pediatric ICU data ecosystem. Full article

(This article belongs to the Special Issue Machine Learning and Artificial Intelligence in Pediatric Healthcare)

► Show Figures

Figure 1

24 pages, 1841 KB

Open AccessReview

From Reactive to Proactive Healthcare: Synergizing Wearable Biomarkers and Machine Learning in Digital Therapeutics

by Kwanjoon Park, Eunice Kwan Chae Park, Woo Hyun Park and Eun-Young Jeon

Bioengineering 2026, 13(9), 977; https://doi.org/10.3390/bioengineering13090977 - 25 Aug 2026

Abstract

The integration of digital therapeutics (DTx), wearable electronic devices, and artificial intelligence (AI) represents a significant advancement in personalized healthcare. The primary purpose of this structured narrative review is to evaluate the convergence of these technologies, providing a consolidated framework that bridges the [...] Read more.

The integration of digital therapeutics (DTx), wearable electronic devices, and artificial intelligence (AI) represents a significant advancement in personalized healthcare. The primary purpose of this structured narrative review is to evaluate the convergence of these technologies, providing a consolidated framework that bridges the gap between raw biometric data acquisition and actionable, AI-driven clinical insights. This paper synthesizes the latest literature on the intersection of mobile health (mHealth), machine learning (ML), and physiological tracking, with a primary focus on heart rate variability (HRV) and associated biochemical markers, such as cortisol, salivary alpha-amylase, and interleukins. Instead of viewing wearable outputs simply as raw data, we critically evaluate the technical verification and clinical validation required to define them as true “digital biomarkers.” By evaluating multimodal sensor technologies and advanced predictive algorithms, this paper outlines the clinical utility of digital biomarkers in diagnosing and proactively managing cardiovascular, neurological, metabolic, and psychiatric conditions, noting classification accuracies frequently exceeding 85% in controlled settings. However, we strongly caution that internally validated performance in controlled settings does not inherently demonstrate external clinical utility. The clinical relevance of this study lies in its holistic approach to identifying how continuous monitoring can broaden healthcare accessibility while improving precision medicine. Furthermore, it deeply addresses the technical challenges of highly variable ambulatory data quality, the necessity for robust artifact reduction (e.g., via LSTM and GAN architectures), and the limitations of small, homogeneous training datasets. We highlight the essential need for demographic-aware algorithmic models, external validation, and decentralized privacy-preserving models (e.g., federated learning) in diverse populations to ensure the safe, equitable clinical translation of DTx, mHealth, ML, and AI technologies. Full article

(This article belongs to the Section Biomedical Engineering and Biomaterials)

► Show Figures

Graphical abstract

43 pages, 797 KB

Open AccessReview

Artificial Intelligence-Driven Reproductive Bioengineering: Integrating Fertility Diagnostics, Organ-on-Chip Systems, Cryobiology and Epigenetic Safety for Precision Reproductive Medicine

by Mohamad Warda, Ali Doğan Ömür, Hae-Jin Park, Jaehoon Bae and A. M. Abd El-Aty

Bioengineering 2026, 13(9), 976; https://doi.org/10.3390/bioengineering13090976 - 25 Aug 2026

Abstract

Traditional assisted reproductive technologies (ART) remain constrained by subjective, descriptive diagnostics and empirical, one-size-fits-all preservation strategies that expose gametes to nonphysiological stressors, risking disruptions to cellular homeostasis and epigenetic programming. This review explores the technological convergence of artificial intelligence (AI), reproductive organ-on-chip bioengineering, [...] Read more.

Traditional assisted reproductive technologies (ART) remain constrained by subjective, descriptive diagnostics and empirical, one-size-fits-all preservation strategies that expose gametes to nonphysiological stressors, risking disruptions to cellular homeostasis and epigenetic programming. This review explores the technological convergence of artificial intelligence (AI), reproductive organ-on-chip bioengineering, multiomics, and translational cryobiology and proposes an integrated, systems-level paradigm for next-generation precision reproductive medicine. By evaluating the clinical readiness, mechanistic insights, and translational trajectories of these emerging platforms, we show how AI architectures transition fertility diagnostics from descriptive metrics to predictive computational phenotyping by integrating high-dimensional imaging, multiomics, and sperm functional datasets. Concurrently, microphysiological platforms—such as testis-, ovary-, and endometrium-on-a-chip systems—recapitulate complex multicellular architecture and endocrine dynamics. When embedded with miniaturized biosensors and machine learning loops, these “smart” closed-loop microfluidic devices enable real-time biological monitoring and adaptive culture regulation. Furthermore, integrating AI analytics into cryobiology optimizes nonlinear thermodynamic variables, shifting the field from basic postthaw morphologic survival toward safeguarding macromolecular fidelity, mitochondrial competence, and long-term epigenetic safety across lifespans and generations. Ultimately, this computational–bioengineering roadmap transitions reproductive healthcare from a reactive discipline into a predictive, personalized, and adaptive framework. Overcoming persistent challenges in biological complexity, data interoperability, and multicenter clinical validation will lead to the establishment of safe, scalable, and ethically governed healthcare infrastructures capable of protecting developmental integrity. Full article

(This article belongs to the Special Issue Artificial Intelligence in Reproductive Medicine: Challenges and New Perspectives)

► Show Figures

Graphical abstract

22 pages, 2595 KB

Open AccessArticle

A Contrastive Domain Adaptation Framework for Knee Osteoarthritis Severity Grading

by Weiqiang Liu, Minghui Wu, Keming Liu, Mingyao Wu and Yunfeng Wu

Bioengineering 2026, 13(9), 975; https://doi.org/10.3390/bioengineering13090975 - 25 Aug 2026

Abstract

Kneeosteoarthritis (KOA) is a common degenerative joint disease that causes pain, stiffness, and impaired mobility. Automated Kellgren–Lawrence (KL) grading from knee X-ray images facilitates efficient screening and follow-up assessment. However, models trained on a single-source dataset frequently suffer performance degradation when applied to [...] Read more.

Kneeosteoarthritis (KOA) is a common degenerative joint disease that causes pain, stiffness, and impaired mobility. Automated Kellgren–Lawrence (KL) grading from knee X-ray images facilitates efficient screening and follow-up assessment. However, models trained on a single-source dataset frequently suffer performance degradation when applied to external cohorts, due to heterogeneities in image quality, acquisition protocols, class distributions, and annotation patterns. Furthermore, conventional domain adaptation approaches typically treat all source samples uniformly, making them vulnerable to negative transfer induced by ambiguous or distributionally divergent instances. To overcome these limitations, the present study develops a supervised contrastive domain adaptation framework designed for robust KOA severity grading under domain shift. The framework incorporates two task-specific modules: (1) a source-domain sample screening module that dynamically allocates class-wise quotas based on transferability and identifies high-value source samples by evaluating target intra-class affinity, inter-class separability, and source-class compactness; and (2) a target-balanced ordinal contrastive learning module that aligns the screened source samples with target features and imposes stronger constraints on negative pairs with larger KL-grade distances. The framework was evaluated bidirectionally on KneeKL (8260 images) and MedicalExpert-I (1650 images), two public knee radiograph datasets for KOA grading. With ResNet-18, it achieved a Quadratic Weighted Kappa (QWK) of 0.8557 for KneeKL-to-MedicalExpert-I transfer, exceeding source-only training and direct source–target merging by 0.2652 and 0.0468, respectively. Comparisons with representative existing methods and multiple experimental analyses further validate the competitiveness of the proposed framework. Full article

(This article belongs to the Special Issue Advanced Computer Methods and Programs in Biomedicine)

► Show Figures

Figure 1

42 pages, 7721 KB

Open AccessArticle

Conditional Diffusion–Augmented Cascaded Multi-View Attention BiLSTM for Non-Invasive Blood Glucose Estimation from Photoplethysmography

by Chaofan Mo and Jianfeng He

Bioengineering 2026, 13(9), 974; https://doi.org/10.3390/bioengineering13090974 - 25 Aug 2026

Abstract

Photoplethysmography (PPG)-based blood glucose estimation is attractive for non-invasive monitoring, but its development is constrained by limited paired PPG–glucose data and the weak representation of glucose-related waveform variations. This study proposes a framework combining conditional diffusion-based data augmentation with a cascaded multi-view attention [...] Read more.

Photoplethysmography (PPG)-based blood glucose estimation is attractive for non-invasive monitoring, but its development is constrained by limited paired PPG–glucose data and the weak representation of glucose-related waveform variations. This study proposes a framework combining conditional diffusion-based data augmentation with a cascaded multi-view attention bidirectional long short-term memory (BiLSTM) network. Blood glucose level, heart rate, and body mass index were used as physiological conditions to generate synthetic PPG segments, while channel and temporal attention together with cascaded BiLSTM layers were used to extract discriminative spatiotemporal features. A moth–flame optimization algorithm was employed to tune key hyperparameters. Experiments were conducted on a public dataset containing 67 PPG recordings from 23 participants. The framework was evaluated using subject-wise five-fold cross-validation, ensuring that all recordings from the same participant remained within a single fold. In the primary seed-42 analysis, the proposed method achieved a recording-level RMSE of 0.59 mmol/L, an MAE of 0.38 mmol/L, a MARD of 5.48%, and Clarke Zone A and A + B proportions of 97.0% and 100%, respectively; repeating the complete evaluation with seeds 2026 and 3407 produced closely similar recording-level results. These results suggest that physiologically conditioned generative augmentation may improve PPG-based blood glucose estimation in small-sample settings. Full article

(This article belongs to the Special Issue Artificial Intelligence in Biomedical Imaging and Biomedical Signal Processing: Second Edition)

► Show Figures

Figure 1

14 pages, 11379 KB

Open AccessArticle

Record Extent and Occlusal Support in Scan-Assisted Interocclusal Registration: A Three-Dimensional In Vitro Study

by Ragai Edward Matta, Paula Töpel, Fabian Lotz, Manfred Wichmann and Johannes Matthias Ries

Bioengineering 2026, 13(9), 973; https://doi.org/10.3390/bioengineering13090973 - 25 Aug 2026

Abstract

Scan-assisted interocclusal registration may vary with record extent and occlusal support. Using separately fabricated high-hardness vinyl polysiloxane (VPS) records, a fixed bilateral buccal intraoral scanner (IOS) acquisition protocol, optical reference articulations, and position-level three-dimensional analysis, this in vitro study compared prepared-tooth (Hard-Prep) and [...] Read more.

Scan-assisted interocclusal registration may vary with record extent and occlusal support. Using separately fabricated high-hardness vinyl polysiloxane (VPS) records, a fixed bilateral buccal intraoral scanner (IOS) acquisition protocol, optical reference articulations, and position-level three-dimensional analysis, this in vitro study compared prepared-tooth (Hard-Prep) and complete-arch (Hard-Arch) records across three support configurations and assessed a no-record IOS workflow. Ninety VPS-based and 45 no-record IOS acquisitions were analyzed. VPS records were fabricated under a 3.0 kg load; measurements used a 1.0 kg load. Overall three-dimensional deviation (dXYZ) was averaged over five mandibular positions. Log-transformed dXYZ was analyzed by ordinary least squares with Holm-adjusted contrasts. The Hard-Arch/Hard-Prep contrast was not statistically significant under bilateral extended support. Hard-Arch showed higher dXYZ under unilateral extended support (ratio, 1.75; 95% confidence interval (CI), 1.54–1.99; p < 0.001) and unilateral limited support (ratio, 1.58; 95% CI, 1.39–1.80; p < 0.001). In the no-record workflow, unilateral extended support exceeded bilateral extended support (ratio, 1.50; 95% CI, 1.31–1.72; p < 0.001), whereas unilateral limited support was lower than unilateral extended support (ratio, 0.68; 95% CI, 0.59–0.77; p < 0.001). Prepared-tooth records yielded lower deviation in both unilateral configurations, while the no-record workflow varied with support configuration. Full article

(This article belongs to the Special Issue Advanced Restorative Dental Materials and Implant Technologies)

► Show Figures

Graphical abstract

4 pages, 2960 KB

Open AccessCorrection

Correction: Lago et al. Generation of Gellan Gum-Based Adipose-Like Microtissues. Bioengineering 2018, 5, 52

by Manuela E. L. Lago, Lucília P. da Silva, Catarina Henriques, Andreia F. Carvalho, Rui L. Reis and Alexandra P. Marques

Bioengineering 2026, 13(9), 972; https://doi.org/10.3390/bioengineering13090972 - 25 Aug 2026

Abstract

In the original publication [...] Full article

(This article belongs to the Section Biomedical Engineering and Biomaterials)

► Show Figures

Figure 4

17 pages, 5180 KB

Open AccessArticle

CNN Sample-Size Effects Across Biomedical Datasets: A Reliability Pattern in Overfitting, Ranking, and Monotonicity

by Giacinto Angelo Sgarro, Melle Mendikowski, Domenico Santoro and Luca Grilli

Bioengineering 2026, 13(9), 971; https://doi.org/10.3390/bioengineering13090971 - 25 Aug 2026

Abstract

Convolutional neural networks (CNNs) are widely used for biomedical image classification, yet it remains unclear under which conditions training on reduced subsets of available data can provide reliable guidance during model development, how much training data is required to achieve stable and comparable [...] Read more.

Convolutional neural networks (CNNs) are widely used for biomedical image classification, yet it remains unclear under which conditions training on reduced subsets of available data can provide reliable guidance during model development, how much training data is required to achieve stable and comparable performance across CNN architectures, and whether increasing the training set size always leads to improved generalization or can sometimes result in degraded performance. We study this question across four biomedical datasets (breast mammography, pediatric chest X-ray, brain tumor MRI, and skin lesion dermoscopy) using the full grid of 39 CNN architectures (1–3 convolutional layers, 16/32/64 filters) from our companion architectural study, training each configuration from scratch on seven proportions of the training data (5%, 10%, 20%, 40%, 60%, 80%, and 100%) over 5 independent runs per configuration, with the test set held at a fixed size across all sample-size conditions to ensure a like-for-like comparison of generalization performance. The analysis investigates three complementary aspects of sample-size sensitivity: the stabilization of the training–test generalization gap as training-set size increases, the reliability of architecture rankings obtained from reduced training fractions as a proxy for the full-dataset ranking, and the monotonicity of test performance with respect to training-set size. Taken together, the results point to a rough four-band pattern of reliability across the sampled fractions—unstable below

20 %

of the training set, of uncertain overfitting status between

20 %

and

60 %

, comparatively stable between

60 %

and

80 %

, and potentially counterproductive beyond

80 %

—while showing that this pattern is itself dataset-dependent and offers no guarantee on architecture ranking, arguing against reduced-fraction screening as a reliable shortcut for CNN architecture selection in biomedical imaging. All code and datasets are publicly released for reproducibility. Full article

(This article belongs to the Special Issue Machine Learning and Artificial Intelligence for Biomedical Applications, 4th Edition)

► Show Figures

Figure 1

34 pages, 16146 KB

Open AccessArticle

Hybrid CNN–Transformer Framework for Automated Detection of Developmental Coordination Disorder from Motion Imaging Sequences

by Khaled Mahmoud Heba, Abbas Hassan Abbas Atya, Noor Hazim Saleh Alrawashdeh, Sana Shahab and Mohd Anjum

Bioengineering 2026, 13(9), 970; https://doi.org/10.3390/bioengineering13090970 - 25 Aug 2026

Abstract

Hybrid CNN–Transformer (HCT) synthesis for automated neurodevelopmental diagnostics is an effective approach to constructing intelligent detection systems that are not merely oriented toward feature classification but primarily toward solving spatiotemporal pattern recognition problems in motor disorder assessment. In neurodevelopmental diagnostics, existing automated methods [...] Read more.

Hybrid CNN–Transformer (HCT) synthesis for automated neurodevelopmental diagnostics is an effective approach to constructing intelligent detection systems that are not merely oriented toward feature classification but primarily toward solving spatiotemporal pattern recognition problems in motor disorder assessment. In neurodevelopmental diagnostics, existing automated methods rely on fixed, single-model architectures that process spatial or temporal motion features independently, failing to adapt to the heterogeneous motor irregularities characteristic of developmental coordination disorder and degrading detection sensitivity and generalization across diverse patient populations. There is therefore a pressing need for models capable of simultaneously capturing intra-frame spatial coordination patterns and inter-frame temporal movement dependencies against interrelated diagnostic criteria including accuracy, sensitivity, and motor irregularity specificity. To address this challenge, this paper proposes HCT, a novel framework that integrates ResNet-based spatial feature extraction from optical flow maps and pose estimation skeletons with multi-head self-attention Transformer encoding for modeling long-range temporal dependencies across multi-frame motion sequences. Unlike conventional single-stream approaches, where spatial and temporal processing remain confined to independent architectures, HCT decouples spatiotemporal feature learning through a cross-modal fusion pipeline, constructing a unified discriminative architecture that captures motor coordination dependencies between motion imaging inputs and multiple diagnostic criteria simultaneously. The convolutional encoder generates diverse joint displacement features, which are consolidated through cross-modal attention fusion into a robust, unified embedding with enhanced generalization and resilience to inter-individual motor variability. Integration within neurodevelopmental assessment frameworks facilitates reliable developmental coordination disorder classification, motor irregularity prediction, and interpretable diagnostic decision support, advancing the accuracy, flexibility, and clinical validity of intelligent motor disorder diagnostic systems. Full article

(This article belongs to the Special Issue Biomedical Application of Big Data and Artificial Intelligence—Second Edition)

► Show Figures

Figure 1

11 pages, 2094 KB

Open AccessArticle

3D-Printed PLA/PETG Sterilizable Static/Dynamic Modular External Finger Fracture Fixator

by Xavier Soong, Alyssa Rothman, Eric Tolo, Maximillian Soong and N. George Kasparyan

Bioengineering 2026, 13(9), 969; https://doi.org/10.3390/bioengineering13090969 - 24 Aug 2026

Abstract

Finger fractures are common and potentially disabling. Certain complex injuries, particularly with small fragments and/or open wounds, are not manageable with conventional orthopedic hardware. We created 3D-printed finger fracture fixators using polylactic acid (PLA) and polyethylene terephthalate glycol (PETG). We tested flexion stiffness [...] Read more.

Finger fractures are common and potentially disabling. Certain complex injuries, particularly with small fragments and/or open wounds, are not manageable with conventional orthopedic hardware. We created 3D-printed finger fracture fixators using polylactic acid (PLA) and polyethylene terephthalate glycol (PETG). We tested flexion stiffness and lateral stiffness, before and after sterilization for surgical use, and compared them against standard 316L surgical stainless-steel (SSS) wires. Because PLA and PETG are compromised at high temperatures, these were sterilized using hydrogen peroxide gas plasma, while the SSS wires were autoclaved. The PLA fixators demonstrated greater flexion stiffness, and equivalent lateral stiffness, to the SSS wires, both before and after sterilization. The PETG fixators demonstrated inferior flexion stiffness and lateral stiffness to the SSS wires, and these properties worsened with sterilization. The PLA fixator was then applied to a cadaveric hand in static and dynamic configurations, and demonstrated excellent radiolucency for visualization of bone alignment and healing. This novel PLA device secures numerous surgical stainless-steel wires, maintains greater flexion stiffness and equivalent lateral stiffness compared to the wires even after sterilization, allows for static and dynamic fixation with compact, lightweight, and modular radiolucent components, and is both low-cost and customizable on demand, which may benefit under-resourced communities. Full article

(This article belongs to the Section Biomedical Engineering and Biomaterials)

► Show Figures

Graphical abstract

Journal Description

Bioengineering

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI