Annals of Biomedical Engineering最新文献

Purpose: Hemolysis remains a concern in mechanical circulatory support devices (MCSDs). Capturing flow-induced red blood cell (RBC) deformation is important to improve these technologies. Deformation models that are feasible for macroscale MCSD flows have not been calibrated with human RBC deformation data across multiple conditions. The purpose of this study is to modify and test a droplet deformation model that is applicable for MCSD flows for predicting human RBC deformation in silico.

Methods: In vitro human RBC deformation is studied in microfluidic flows in two suspension viscosities (2.05 and 4.17 cP) at MCSD relevant strain rates (5,000 - 200,000 s^-1 in shear flow; 330 - 13,160 s^-1 in extensional flow). Modifications are made to the deformation model's constitutive parameters to represent the observed RBC deformation in silico.

Results: The calibrated model reproduces the unique RBC deformation behaviors observed in shear and extensional flows across a range of conditions. In silico shear deformation index data have mean absolute error (MAE) ≤ 0.15 compared to in vitro results for both viscosity conditions from 5,000 to 200,000 s^-1. Peak in silico extensional deformation data demonstrate MAE ≤ 0.11 compared to our in vitro results for both viscosity conditions from 670 to 1,300 s^-1, while MAE is higher (up to 0.17) for conditions at 330 s^-1.

Conclusion: The model adaptations successfully produce accurate RBC deformation results at MCSD relevant strain rates for two flow types and two suspension viscosities. The strengths of the model are in relatively high velocity gradient magnitudes and/or suspension viscosities where RBCs emulate liquid droplets.

Purpose: A diagnostic challenge in the management of chronic kidney disease (CKD) is distinguishing diabetic kidney disease (DKD) from hypertensive kidney disease (HKD) in patients with coexisting diabetes mellitus (DM) and hypertension (HTN), because accurate diagnosis often depends on renal biopsy as a reference standard. This study proposes a modeling approach to identify cardiovascular biomarkers for differentiating DKD from HKD.

Methods: An existing whole-body circulation model of the vascular tree was extended with a detailed renal circulation network to predict biomarkers measured at different locations. The model parameterized sex, age, and disease factors and was used to conduct virtual clinical trials that identified individual and combined biomarkers for DKD-HKD differentiation. Biomarkers were identified with univariate and multivariate analysis and characterized with the area under the receiver operating characteristic curve (AUC).

Results: Results show that the strongest individual biomarker that is commonly used in clinical practice is pulsatility index (PI) measured in the main renal artery, with an AUC of 0.87. Among all evaluated two-biomarker combinations, PI and resistive index (RI) measured in the same artery achieved the highest classification performance (AUC 0.94). In comparison, the highest performance among three-biomarker combinations (AUC 0.96) is achieved by mean blood flow rate, systolic blood flow rate, and diastolic flow rate.

Conclusion: This modeling work suggests that cardiovascular biomarkers can assist in differentiating DKD and HKD, and proposes specific hypotheses that form a strong rationale for targeted clinical trials. If confirmed, these methods could enable non-invasive assessment of renal vascular alterations associated with DKD and HKD, reducing reliance on kidney biopsies for diagnostic evaluation.

Discrete subaortic stenosis (DSS) is a congenital heart disease in which a fibrotic membrane forms below the aortic valve; the underlying cellular mechanisms are currently unknown. Since an elevated pressure gradient in the left ventricular outflow tract (LVOT) is a distinguishing feature of DSS, it is hypothesized that the membrane formation is caused by elevated wall shear stress applied to the endocardial endothelial cells (EECs) that line the LVOT, triggering fibrosis. To correlate shear stress to an EEC fibrotic phenotype, we applied fluid shear stress to EECs at physiological and pathological shear rates using a cone-and-plate device, designed to recapitulate physiological wall shear stress in a controlled in vitro environment. Controlled shear stress regimes were applied to EECs to replicate the conditions observed in DSS patients. We found that elevated shear stress triggered EEC alignment as well as endothelial-to-mesenchymal transformation (EndMT) signaling pathways driven by upregulation of SNAI1 gene expression. The EECs were then treated with a small molecule inhibitor of Snail1 protein, CYD19, to attempt to attenuate EndMT signaling, and subsequently subjected to pathological shear stress. The Snail1 inhibitor did downregulate selected markers of EndMT signaling, although only transiently. Interestingly, the application of shear stress had a greater effect on the EEC gene and protein expression than did the Snail1 inhibition. This investigation of EEC response to shear stress reveals the pronounced and complex effect of this mechanical stimulation on the EEC phenotype. Further study should reveal the mechanisms that drive fibrosis and the formation of the DSS membrane.

Purpose: Prediction of Torsades de Pointes (TdP) risk using hiPSC-CM assays remains challenging, as many models fail to capture nonlinear patterns and exhibit unstable performance across different drugs. We examined whether a stacking ensemble can improve robustness when only two simple $\Delta FPDc$ -derived predictors are available.

Methods: Two electrophysiological predictors were derived from MEA-based FPDc measurements: the maximum change ( $\Delta FPDc$ ) and the interpolated $\Delta FPDc$ at Cmax. These features were used to train a stacking model comprising random forest (RF), XGB, and a shallow artificial neural network (ANN). Performance was assessed on 16 unseen CiPA reference compounds using AUC, likelihood ratios, pairwise accuracy, and classification error. Class weighting addressed class imbalance, and likelihood-based metrics assessed diagnostic consistency.

Results: The stacking ensemble consistently outperformed all single classifiers. The XGB-based meta-classifier achieved perfect discrimination for both AUC1 and AUC2 (1.000), with pairwise accuracy reaching 1.000 and classification error remaining below 0.125 across repeated evaluations.

Conclusions: Combining a simple MEA-derived feature set with a stacking architecture provides a more reliable framework for early TdP risk assessment than single-model approaches. External validation using additional MEA datasets and the integration of interpretable modeling strategies will be important for future translational use.

Purpose: To develop and validate a subject-specific finite element (FE) modeling framework that estimates breast hyperelastic parameters from in vivo, gravity-driven deformation, and to investigate whether post-augmentation breasts with silicone implants exhibit altered mechanical characteristics compared with natural breasts.

Methods: Left breasts of twelve participants (six normal, six post-augmentation) were modeled from prone MRI, and upright morphology was simulated using a two-step gravity procedure in Abaqus using a first-order isotropic Ogden hyperelastic law, chosen as a simple, widely used model for large soft-tissue deformations. For each breast, the Ogden exponent (α) and shear modulus (μ) were identified as apparent organ-level stiffness parameters by minimizing the Hausdorff distance between the simulated shape and the standing 3D surface scan.

Results: Natural breasts achieved HDF 0.94-3.09 mm (mean 2.14 mm). Post-augmentation breasts achieved HDF 1.38-2.54 mm (mean 1.90 mm). Across participants, α showed marked inter-individual variability. Post-augmentation cases required higher effective shear modulus than natural breasts.

Conclusion: A subject-specific FE workflow driven by in vivo deformation accurately reproduces standing breast morphology and reveals increased apparent stiffness after implant augmentation. This framework enables individualized parameterization for surgical planning, implant selection, and prediction of post-operative outcomes in breast biomechanics.

Purpose: The aim of this study was to compare and assess T₂ values of compartmental femorotibial cartilage and subregional menisci in patients with diabetes mellitus (DM) at 3 T.

Methods: Forty-two subjects were enrolled in the study and subdivided into two subgroups: 20 healthy controls (11 females, mean age = 51.6 ± 5.9 years, range 44-67 years) and 22 DM patients (8 females, mean age = 60.1 ± 8.9 years, range 40-75 years). All subjects were evaluated on a 3 T MR scanner applying a 15-channel phased-array knee coil. Wilcoxon rank sum test and analysis of covariance were exploited for detecting group differences, with false discovery rate utilized to control for false positives due to multiple comparisons.

Results: Lateral tibial cartilage in healthy controls had significantly lower (p < 0.05) T₂ values (38.8 ± 5.4 ms) than lateral and medial femoral cartilage, respectively, in DM patients. T₂ values of medial body segment (Mc) of meniscus (33.8 ± 8.5 ms) in DM patients were significantly higher (p < 0.05) than those of all subregional menisci except Mc (p = 0.7496) in controls. Mc in controls had significantly higher (p < 0.05) T₂ values (31.4 ± 6.3 ms) than lateral and medial anterior horn, and lateral posterior horn of meniscus, respectively, in DM patients.

Conclusion: T₂ values in certain cartilage subcompartments in DM patients are abnormally elevated. T₂ mapping may be a potential imaging biomarker for detecting early subtle damage to femorotibial cartilage resulted from DM.

Purpose: Instrumented mouthguards (iMGs) are commonly used to quantify head acceleration event (HAE) exposure, but accurate interpretation requires rigorous data cleaning methods. This study compared six data cleaning methods for determining HAE rates and magnitudes, as well as cleaning method validity compared to the 5^th method video verification in youth tackle football.

Methods: Fifty athletes (ages 8-12) wore Impact Monitoring Mouthguards during games across one season. Six data cleaning methods were applied to HAEs, including uncleaned data, time-windowing, proprietary classification algorithms, video verification, and combinations thereof. Impact rate, peak linear acceleration (PLA), and peak rotational velocity (PRV) were compared across methods using rate ratios, and intra-class correlation coefficients (ICCs), and non-parametric analyses.

Results: Data cleaning methods significantly influenced HAE rate but had minimal effect on magnitude. The uncleaned dataset produced the highest HAE rate (67.75 per athlete exposure), while the most stringent method (i.e., time-windowed, proprietary algorithm-classified, video-verified data) yielded the lowest (0.70 per athlete exposure). Although the time-windowed, proprietary algorithm-classified data demonstrated high specificity (0.96), it demonstrated low sensitivity (0.37) and positive predictive value (0.39) when compared to video-verified data. Differences in PLA across methods were not significant; only one significant difference in PRV was observed.

Conclusions: These findings highlight the impact of data cleaning on HAE quantification in youth tackle football. Although video verification remains best practice, it is resource intensive. Time-windowed, algorithm-classified data may serve as an efficient proxy in similar cohorts, though researchers should recognize its limitations. Findings support the need for standardized data cleaning methods and transparent reporting to ensure accurate and comparable HAE exposure estimates.

The conductive response in thigh muscle compartments has been imaged by electrical impedance tomography (EIT) under bicycle-operation conditions and muscle-stimulation conditions to determine an efficient training strategy. EIT is already applied to evaluate the effectiveness of electrical muscle stimulation (EMS) on human muscle, and hybrid of EMS which combines EMS on biceps and voluntary resistance training simultaneously. In this study, we newly applied EIT to the thigh muscle compartments and imaged the conductive response under bicycle-operation conditions and muscle-stimulation conditions. The bicycle-operation conditions are high pedal rate in low torque (HPLT) and low pedal rate in high torque (LPHT). The muscle-stimulation conditions are bicycle condition and EMS-combined bicycle condition. As a result, EIT successfully imaged the conductive response in three muscle compartments which are T1 compartment (quadriceps), T2 compartment (hamstrings), and T3 compartment (other than the quadriceps and hamstrings). The spatial-mean conductivity changes $\Delta ⟨{\sigma }_P⟩$ (the power of bicycle training $P=103,152,or194\left[W\right]$ ) increased with increasing power $P$ of the bicycle training across all four conditions and in all muscle compartments ( $\beta =0.531,p<0.001$ ). The $\Delta ⟨{\sigma }_P⟩$ of HPLT tended to be larger than $\Delta ⟨{\sigma }_P⟩$ of LPHT ( $\beta =-35.52,p<0.001$ ). The $\Delta ⟨{\sigma }_P⟩$ of EMS-combined bicycle condition tended to be larger than $\Delta ⟨{\sigma }_P⟩$ of bicycle condition ( $\beta =12.85,p<0.05$ ). The key findings are that HPLT and EMS-combined bicycle condition are a more efficient bicycle training than LPHT and bicycle condition for increasing $\Delta ⟨{\sigma }_P⟩$ .

Purpose: This study aimed to quantify the effect of seated weight-bearing on the rotational motion characteristics of lumbar facet joints (LFJs) using the dual fluoroscopic image system (DFIS), providing kinematic evidence for the prevention and clinical management of lumbar degenerative disorders.

Methods: Sixteen healthy volunteers (8 males, 8 females; age: 25-39 years, mean: 32 ± 4.29 years) were enrolled. Thin-layer computed tomography (CT) scans of the L3-S1 segment (64 vertebrae total) were acquired to reconstruct 3D lumbar spine models via Mimics 21.0. DFIS was used to simulate seated rotational motion (left rotation, neutral position, right rotation) under 0 kg (non-weight-bearing, NWB) and 10 kg weight-bearing (WB) conditions. Translational displacements (X: coronal axis, Y: sagittal axis, Z: vertical axis) of L3/4, L4/5, and L5/S1 LFJs were measured and compared using Rhinoceros 5.0. Statistical analysis was performed via SPSS 26.0 with paired t-tests; P < 0.05 was considered statistically significant.

Results: At L4/5, WB significantly reduced translational displacements: left side (X-axis:0.01 ± 0.80 mm [NWB] vs1.00 ± 1.01 mm [WB], P = 0.014; Y-axis: 0.90 ± 1.76 vs 0.55 ± 0.55 mm, P = 0.014) and right side (X-axis: 0.04 ± 0.79 vs 1.05 ± 0.94 mm, P = 0.023; Z-axis: 1.25 ± 0.98 vs 0.92 ± 1.22 mm, P = 0.001). At L5/S1, WB induced significant displacement reductions: left side (Y-axis: 2.42 ± 1.16 vs 0.15 ± 1.82 mm, P < 0.001; Z-axis: 3.43 ± 3.30 vs 3.03 ± 0.76 mm, P = 0.002) and right side (Z-axis: 1.57 ± 2.01 vs 0.64 ± 1.53 mm, P = 0.002).

Conclusion: Seated weight-bearing alters the rotational kinematics of LFJs, particularly at L5/S1, with significant reductions in translational displacements. This finding provides a biomechanical basis that may inform our understanding of load-related adaptations in the lumbar spine.

Irreversible electroporation (IRE) is a minimally invasive, non-thermal, and cell-selective technique. When combined with the noninvasive nature of contact electrodes, they hold great promise for the treatment of cardiac conditions, gastrointestinal tumors, and superficial lesions. However, its broader clinical application is hindered by its reliance on a kilovolt-level, high-voltage pulse characteristics power supply and the lack of real-time postoperative assessment methods for evaluating ablation efficacy. To address these challenges, a contact electrode system with integrated IRE and impedance monitoring functions was developed. Numerical simulations were performed to optimize the anode, gap, and cathode widths in the concentric electrode design. This ensured efficient electric-field focusing under low-voltage conditions. The ablation performance was verified using a potato model. A four-electrode impedance measurement technique was used to capture the spectral characteristics of biological tissues. The impedance changes were analyzed using a double-shell equivalent circuit model. The system achieved a 2 mm ablation depth at 125 V, which is suitable for the treatment of superficial lesions. This reduces the required voltage from the kilovolt level to the hundred-volt level. The four-electrode method reduced contact resistance interference, and the Nyquist plots showed a unique double-arc pattern. Changes in cell wall resistance correlated with ablation depth ( ${\text{R}}^2$ = 0.86) with a prediction error of <10%. This study presents an innovative approach for IRE therapy that combines low-voltage operation with real-time feedback through impedance spectroscopy, thereby offering improved safety and treatment monitoring.