Annals of Biomedical Engineering最新文献

Phase-contrast magnetic resonance imaging (PC-MRI) is a widely used technique for measuring cerebral blood flow dynamics in humans. However, the application of this technique for velocity mapping of cerebral blood flow in rats has been very limited thus far. This study aimed to determine the accuracy of PC-MRI flow measurements and to establish reference flow values for the main intracerebral feeding and draining vessels in rats. PC-MRI velocimetry was conducted on a flow phantom and 12 adult Sprague Dawley rats. A semi-automatic segmentation was applied to the velocity images to enable the quantification of flows and the calculation of pulsatility index. PC-MRI flow measurements showed good agreement with the reference flow from the pump. The main findings of the in vivo measurements were as follows: the mean blood flow values in the internal carotid artery (ICA), external carotid artery (ECA), basilar artery (BA), and transverse sinuses were 9.6 ± 3.4, 5.7 ± 2.4, 2.0 ± 0.6, and 10.5 ± 3.6 ml/min, respectively. Mean ICA flow was higher than ECA flow, while the ICA pulsatility index was lower than the ECA pulsatility index. There was no significant difference in the mean cerebral inflow (combined ICA and BA flows) and cerebral outflow (left and right transverse sinus flows) measures. No differences in arterial or sinus flow were observed between the left and right sides of the brain. These reference flow values and pulsatility index may be useful for understanding flow dynamics in healthy and pathological conditions in rats.

Early detection of retinal lesions helps to avoid visual loss or blindness. The main lesions associated with eye diseases include soft exudates, hard exudates, microaneurysms, and hemorrhages. However, the segmentation of these four kinds of lesions is difficult and time-consuming due to their uncertainty in size, contrast, and high inter-class similarity. To address these issues, this study presents Deep Decoder-Focused U-Net (DDFU-Net), an asymmetric dense U-Net model for automatic and accurate multi-lesion segmentation using fundus images. Our approach simultaneously segments all four kinds of retinal lesions after proving that multi-task learning yields better results than single-task learning. DDFU-Net incorporates an asymmetric design with five dense blocks in the encoder and seven dense blocks in the decoder. This design enhances feature extraction while ensuring a more refined reconstruction of lesion boundaries, particularly for small and complex structures. By allocating more layers to the decoder, the model improves segmentation accuracy by gradually restoring spatial details lost during down-sampling, mitigating over-compression, and enhancing fine-grained feature preservation. Comprehensive experiments on IDRiD and DDR datasets well demonstrate the superiority of our approach, which outperforms state-of-the-art segmentation methods. Specifically, DDFU-Net achieved a mean Area Under the Precision-Recall Curve of 54.86%, a mean Intersection Over Union of 36.96%, and mean Dice scores of 52.24% on the DDR test set. On the IDRiD test set, it achieved 66.69%, 57.31%, and 69.93%, respectively. The asymmetric structure outperforms traditional symmetric U-Nets by capturing more detailed features during encoding while reducing complexity during decoding. The proposed method can be useful to aid in the diagnosis of eye diseases, reducing the workload of specialists and improving the attention to patients.

The increasing prevalence of chronic diseases, functional decline, and cognitive impairments, together with the limitations of episodic clinical assessments, has created a critical need for automated and continuous monitoring of Activities of Daily Living (ADLs) to support timely intervention, personalized care, and independent living. This review systematically examines how ambient sensing, wearable devices, and fusion-based systems have been applied to ADL monitoring in home environments. We analyze 109 peer-reviewed studies published between 2013 and 2024, focusing on reported performance trends, study scale, sensing configurations, and methodological characteristics. Given the substantial heterogeneity across experimental settings, participant populations, activity definitions, and evaluation protocols, reported metrics are interpreted descriptively rather than as directly comparable measures. Overall, wearable-based systems frequently report higher performance metrics in controlled settings, ambient systems offer advantages related to privacy and unobtrusiveness, and fusion-based approaches provide richer contextual information but face scalability challenges. Across all modalities, limited dataset diversity, small sample sizes, inconsistent evaluation practices, and insufficient reporting of real-world deployments remain persistent gaps. This review synthesizes recent advances in sensor-based activity recognition for monitoring activities of daily living, with a particular emphasis on AI-driven approaches.

Purpose: Simulating tissue deformation and flow alterations induced by external compression typically requires fluid-structure interaction (FSI) analysis, which is computationally demanding. This study presents a multi-fidelity FSI framework that efficiently captures tissue mechanics and hemodynamic responses to dynamic external pressure and demonstrates its applicability to compression therapy.

Methods: We developed an FSI model that couples a one-dimensional deformable blood flow formulation with the three-dimensional (3D) Cauchy equation of motion. Model performance was evaluated by comparing the multi-fidelity and full FSI solutions in simplified cylindrical and subject-specific geometries. As a practical demonstration, the framework was applied to simulate a full-cycle pulsatile intermittent pneumatic compression (IPC) operation.

Results: The model efficiently reproduced tissue deformation and hemodynamic changes under external compression, yielding <1% flow-rate error in both geometries and <2% pressure error in the simplified geometry for most of the cycle, with good agreement in the subject-specific geometry. Computational cost was reduced by a factor of 9 in the cylindrical geometry and 46 in the subject-specific geometry relative to full 3D FSI. In the IPC application, the model captured dynamic behavior over an extended temporal scale, completing a full cycle in 457 s for the simplified geometry and 42.2 min for the subject-specific geometry.

Conclusion: This multi-fidelity FSI framework enables efficient and accurate simulation of tissue deformation and hemodynamic responses under external pressure, providing a tractable platform for large-scale parametric and optimization studies. Its application to IPC highlights potential to enhance therapeutic device design and support broader applications in biomedical modeling and medical device development.

Purpose: The recurrence rate of ventral abdominal wall hernia repair remains high, partly due to incomplete understanding of biomechanical factors influencing fascial closure. This study aimed to develop and evaluate a patient-specific computational model to predict closure tension in the anterior rectus sheath (ARS) and posterior rectus sheath (PRS) during different steps of the transversus abdominis release (TAR) procedure.

Materials and methods: A preoperative CT scan of a patient with a 7 × 6 cm midline ventral hernia was segmented to reconstruct in 3D abdominal wall structures. Finite element meshes were generated for muscles, fasciae, aponeuroses, and the peritoneum. Tissue properties were assigned from experimental data and the abdominal cavity was modeled as a fluid-filled incompressible volume. The successive surgical steps of retrorectus dissection, incision of the posterior lamella of the internal oblique aponeurosis, and TAR, were simulated. Tension in ARS and PRS was computed and compared with in vivo measurements reported in the literature. Stress distributions in fascial components were also analyzed.

Results: The model successfully reproduced clinical trends of tension reduction reported in vivo. PRS tension decreased progressively from baseline to TAR (retrorectus dissection: - 16.7%, posterior lamella of the IO aponeurosis incision: - 60.0%, TAR: - 97.6%), while ARS tension dropped mainly after retrorectus dissection and remained stable thereafter (retrorectus dissection: - 45%, posterior lamella of IO aponeurosis incision: - 53.2%, TAR: - 54.0%). However, predicted baseline tensions appeared overestimated compared to clinical values. Stress analysis revealed redistribution between anterior and posterior fascial components during dissection.

Conclusion: This patient-specific computational model replicated in vivo tension trends, while overestimating absolute values. Model assumptions on material properties, fascial thickness, and abdominal cavity incompressibility may explain these discrepancies. A retrospective patient-specific validation study is warranted. Once validated, such models could guide surgical decision-making, help standardize TAR procedures, and enable patient-tailored hernia repair strategies.

Whole-body powered exoskeletons can augment human performance and reduce physical strain in occupational settings, but little is known about how users adapt to these complex devices during practical work scenarios. We compared novice and experienced users during simulated, occupationally relevant load-handling tasks. Six novice users completed exoskeleton familiarization and stationary load-handling tasks in three sessions while five experienced users performed the tasks once. Task performance, biomechanical demands, and perceived workload were compared in each novice session vs. the experienced group. Novice performance improved substantially across sessions, with task completion time reduced by nearly 50% and movement jerk by 30%. However, performance gaps still persisted in session three, compared to the experienced users. Novices also used consistently lower angular velocities (up to 52% lower) and adopted greater hip flexion throughout the sessions. In contrast, differences in shoulder flexion, muscle activity, perceived exertion, and workload diminished more rapidly, with novices approaching experienced levels by session three. Novice users adapted to using a powered exoskeleton over multiple sessions, especially in movement patterns and muscle activation, but differences in task completion time, jerk index, and angular velocities indicated that novices did not attain the skilled coordination and efficiency of experienced users after three sessions. Our results highlight the likely need for extended familiarization and training for the current powered exoskeleton design and provide baseline data for the novice learning curve in occupational settings.

Purpose: To analyze the development and application of microfluidic devices representative of the dentin-pulp complex.

Methods: A systematic literature search was conducted on electronic databases. The inclusion criteria were in vitro studies aimed at developing and/or using microfluidic devices that mimic the dentin-pulp interface. The aspects analyzed included device design, fabrication methods, interface and channel characteristics, culture methods, main outcomes and measurement techniques. The risk of bias was assessed via the QUIN tool.

Results: Five studies were included. Significant variability was observed in device design, construction, channels, and interface characteristics. Only one study incorporated dynamic fluid flow, whereas the others relied on static culture conditions. The main outcomes measured included cell viability, odontoblastic morphology and metabolic activity.

Conclusion: This systematic review identified limited research on microfluidic devices that replicate the dentin-pulp complex with high variability in terms of fabrication, design, and materials. Most studies have used static cultures despite the benefits of dynamic flow. Additionally, studies lack detailed information on critical device characteristics. Future research should integrate dynamic features, and comprehensive reporting to improve reproducibility and develop physiologically relevant models.

Purpose of review: Historically, bone-anchored prosthesis (BAP) treatment was developed for the treatment of individuals with a transfemoral amputation. However, increasing experience with osseointegration implant (OI) surgery has resulted in its use in individuals with a transtibial amputation. In this article, we describe the emergence and evolution of BAP treatment for individuals with a transtibial amputation, review the current literature, and report on 10 years of clinical experience at our center.

Recent findings: The first tibial OI cases experienced high failure rates due to infection and aseptic loosening. However, all the following published studies reported improvements to functional outcomes and low rates of major complications. In the past, there was a clear lack of fixed follow-up, longer term studies, and stratification of outcomes. At our center, we have treated a total of 124 patients with 133 tibial BAPs over the course of 10 years, starting in 2014, and followed up for a mean of 58 months, ranging from 9 to 131 months. Five different OI designs were used over this period, showing the evolution of design philosophies and adaptations to challenges over time. A total of 11 OIs failed/were removed for reasons such as implant breakage (2), chronic knee pain (2), early infection (3), late infection (2), failure of osseointegration (2). Tibial BAP treatment has proven increasingly safe with growing patient demand. Initial fears about failure of fixation/loosening seem unfounded due to optimization of the designs. Our current aim is to improve implant survival and to answer questions about: optimal patient selection, minimum area of osseointegration necessary, need for additional screw fixation, and optimal level of implant placement.

Purpose: The lymphatic system regulates fluid homeostasis, yet the relative roles of active vessel pumping, external mechanical forces, and lymphatic vessel/interstitium material properties on lymphatic uptake remain unclear. We sought to quantify how each factor affects lymphatic drainage.

Method: We developed a bond-graph-based computational framework for modelling initial lymphatic uptake. Two networks, a simple 7-branch and a complex 19-branch architecture, were constructed to evaluate active pumping, passive external forces, and their combination. We varied vessel compliance, interstitial permeability, anchoring filament stiffness, interstitial compliance, and interstitial thickness, and performed global sensitivity analysis to identify dominant and interacting parameters. Model outputs were compared with published experimental data to assess physiological plausibility.

Results: Simulated uptake rates were consistent with published ranges, supporting the model's physiological validity. Network topology exerted strong control: the 19-branch network consistently produced lower flow than the 7-branch network. Vessel compliance increased cycle-mean flow in the 7-branch network (+11% active; + 17.8% concurrent) but reduced flow in the 19-branch network (-4.3% active; - 9.3% concurrent), indicating that structural losses in larger networks can outweigh distensibility benefits. Under passive forcing, interstitial permeability dominated drainage; very low permeability nearly eliminated flow. Anchoring filament stiffness, interstitial compliance, and thickness had negligible effects. Sensitivity analyses showed compliance governed active and concurrent force cases, whereas permeability dominated passive flow, with additional permeability/compliance interactions emerging in the 19-branch network.

Conclusion: Branching topology and material properties fundamentally influence lymphatic drainage. Active pumping is particularly important in highly branched architectures, while interstitial properties mainly modulate passive flow. The framework enables quantitative exploration of regional physiological and pathological differences.

Purpose: The purpose of this article is to address the limitations of inconsistency between the impeller and motor when designing Percutaneous ventricular assist devices for different clinical scenarios, such as cardiogenic shock and high-risk percutaneous coronary intervention surgery. This inefficiency results in the waste of computational and experimental resources. This work presents a novel optimization method for systematically designing the pVAD.

Methods: The system-level optimization framework combines artificial neural networks (ANN), analytical target cascading (ATC), and NSGA-II. ATC ensures coordinated parameter matching between the motor and blade, NSGA-II refines specific design parameters, and ANN reduces computational overhead. The approach streamlines key design parameters.

Results: The integrated optimization method successfully achieves balanced performance between cardiac output and motor power demand. The prototype achieves a pressure head of > 80 mmHg at 5 L/min with a hemolysis index < 0.02. The total efficiency of the pVAD reaches 33.65%, compared to the baseline design's 7.59%; the algorithm significantly improves the design.

Conclusion: The proposed framework resolves the trade-off between blade performance and motor power in pVAD design, enabling more efficient and feasible device optimization. The results also revealed that the proper design of blades is vital. Proper inlet and outlet angles will yield optimal hydraulic performance, while a shorter blade chord may reduce the risk of hemolysis. The two aspects are not independent, and design points should be chosen comprehensively.