This study evaluated whether scout-derived angular tube current modulation (ATCM) in cone beam computed tomography (CBCT) can reduce radiation dose while preserving image quality. ATCM is routine in helical CT but remains absent from clinical CBCT and is largely unexplored in the literature. A projection-based intensity-dependent noise model was developed using polymethyl methacrylate slab measurements at 80–120 kV. Posterior-anterior and lateral scout views were used to derive angle-dependent elliptical attenuation estimates and to derive per-angle current-time product profiles. ATCM was emulated using projection data acquired with a clinical CBCT system and tested using anthropomorphic phantoms and a uniform water phantom. Dose-area product (DAP), effective dose based on Monte Carlo calculations, image noise magnitude, and noise texture were compared between modulated and unmodulated reconstructions. Image quality was subjectively evaluated by five board-certified radiologists using blinded paired preference scoring (60 paired reads; 12 pairs × 5 radiologists). ATCM reduced DAP by up to 17
This study aims to analyze the balance responses and foot placement strategies of above-knee amputees following controlled mediolateral perturbations and to compare these responses with those of non-amputee individuals. We recruited eight above-knee amputees and eight non-amputee participants to walk on an instrumented treadmill while a cable-driven system delivered mediolateral perturbations to the waist. We applied perturbations at three gait timings (heel strike, late swing, and late single support), two directions (prosthesis and sound side, left and right for non-amputees), and two force magnitudes (7.5 and 15
An emerging concept for plate fixation of fractured bones in humans changed from absolute stability through relative stability to a new concept of a predefined range of elastic movement with mechanical amplitude limitation (“Biphasic Plate,” 41 medical, Bettlach, Switzerland). As this represents a novel implant design, only a few studies have investigated it so far, particularly with respect to human cadaveric specimens and in vivo patient-specific cases. Therefore, this study aims to investigate the biphasic behavior using in vitro, in silico, and in vivo methodologies concerning stress distributions under realistic boundary conditions. To test this, an implantable in vivo strain measurement system (“Fracture Monitor,” AO Foundation, Davos Switzerland) and a camera system are employed for experimental data acquisition, while patient monitoring captures actual boundary conditions during daily activities. Finite element simulations are performed on patient-specific 3D models generated from computed tomography imaging, allowing for detailed stress analysis. These simulations are validated through experiments on human cadaveric specimens in a custom-designed test rig by investigating the specific behavior of the biphasic plate with its two different modes (rigid and flexible). Under axial loading, the stresses are distributed centrally; while in patient-specific cases, the stress pattern changes markedly, with maximum stresses concentrating along the lateral side of the plate. By the chosen biomechanical testing setup, the in silico simulations could be validated, demonstrating the ability to accurately replicate the postulated motion sequence of the biphasic plate. This iteration emphasizes both validation and reproducibility as the key outcomes.
To develop a framework that benchmarks penalized statistical inference against explainable artificial intelligence (XAI) methods for predicting cranial and spinal surgical need in pediatric achondroplasia under small-cohort, class-imbalanced, and phenotypically heterogeneous conditions. We analyzed 34 clinical, demographic, and imaging variables in a multicenter cohort (four U.S. hospitals) of 150 pediatric patients with achondroplasia. We benchmarked penalized statistical inference (ridge regression and generalized additive models [GAMs]) against nine cost-sensitive classifiers, and applied post hoc SHapley Additive exPlanations (SHAP) to interpret the best-performing classifier. The stacked ensemble achieved superior test-set performance (accuracy and macro-F1 = 0.77) with stable generalization and significantly higher discrimination than the GAM baseline (cranial AUROC 0.78 vs. 0.68; spinal AUROC 0.76 vs. 0.66). Calibration was acceptable, and decision curve analysis showed positive net benefit across relevant thresholds for both outcomes. SHAP highlighted class-specific drivers: foramen magnum (FM) stenosis, hydrocephalus, family history, frontal bossing, sleep disturbance, and age for cranial surgery, and spinal stenosis, Chiari malformation, family history, back pain, FM stenosis, and age for spinal surgery. SHAP dependence patterns suggested context-dependent age attributions in relation to FM or spinal stenosis, rather than a consistent standalone age effect. Several SHAP-highlighted predictors (e.g., sleep disturbance, back pain, and age) were nonsignificant in the inferential baselines. Kaplan–Meier curves indicated earlier intervention in high-risk phenotypes. In a reproducible dual-stream benchmark, XAI improved discrimination over conventional inference by capturing clinically contextualized, complex nonlinear predictive dependencies and interactions associated with cranial and spinal surgical risk in achondroplasia using preoperative variables.
Implant-associated infections remain a major cause of failure in titanium-based orthopedic devices, often compromising long-term clinical outcomes and necessitating revision surgeries. Surface modification strategies that provide antibacterial protection while preserving osseointegration are therefore highly desirable. The aim of this study was to develop and evaluate a multifunctional titanium surface capable of preventing bacterial colonization without impairing biological integration. A titanium surface was engineered using sol–gel technology incorporating silver nanoparticles (Solgel_Ti). Antibacterial performance was assessed through microbiological analyses against Staphylococcus aureus, a clinically relevant pathogen in orthopedic implant infections. In vitro biological safety was evaluated by cytocompatibility assays and Ames’s mutagenicity test. The in vivo performance of the coating was investigated using a rat femoral implantation model, with implants analyzed after 90 days through histological and histomorphometric assessments to quantify bone-to-implant contact and new bone formation. Solgel_Ti demonstrated a significant short-term antibacterial effect against S. aureus when compared with uncoated titanium controls. In vitro analyses confirmed a high biological safety profile, showing excellent cytocompatibility and no evidence of mutagenic activity. In vivo evaluations revealed that Solgel_Ti-coated implants supported effective osseointegration, with bone-to-implant contact and new bone formation comparable to those observed for unmodified titanium. No adverse tissue reactions or signs of impaired bone healing were detected. These findings indicate that Solgel_Ti represents a promising dual-functional surface modification strategy that successfully combines antibacterial activity with preserved osseointegration. This approach offers a viable pathway for the development of safe, infection-resistant orthopedic implants with long-term clinical potential.
Approximately half of triple-negative breast cancer (TNBC) patients attain a complete response to neoadjuvant therapy (NAT). Thus, methods to predict and optimize NAT response are essential to improving patient outcomes. Previously, mathematical models with a reaction term describing cell growth and death and a diffusion term describing cell invasion have accurately forecasted NAT response but have had limited flexibly in capturing cell movement. We investigate the relative contributions of reaction, diffusion, and advection terms in predicting TNBC response to NAT. We compare a reaction-only model to three models capturing cell movement using advection and diffusion terms that can be coupled to tissue mechanics. When compared to a reaction-only model, the reaction–diffusion model did not improve calibration or prediction accuracy for tumor volume or cell count. For example, the median absolute difference between the predicted and measured percent change in tumor volume across the cohort was 9.1
Here, we build on a letter published in 2023 in this journal ( https://doi.org/10.1007/s10439-023-03272-4 ) on prompt engineering for academic writing. Drawing on ethics literature, both new and old, we reflect on the prompting of large language models (LLMs) from three different moral perspectives: consequentialism (the ethics of outcomes), deontology (the ethics of duties), and virtue ethics (the ethics of character). While we recognize that human authors who prompt LLMs to derive text can simplify knowledge creation, the question becomes: are prompters equivalent to authors, and should they be rewarded as such? We push back against over-reliance on prompt engineering by arguing in favor of the craft of human draft as a task worth preserving in academia to encourage authors to retain control over published content that bears their name.
Accurate evaluation of dentin–composite bond integrity is essential for predicting the longevity of resin-based restorations. However, conventional mechanical testing methods are destructive and do not provide a non-destructive means of assessing early interfacial changes. Owing to the complex microstructure and heterogeneous composition of dentin, reliable non-destructive assessment remains challenging. This study proposes a novel optical-machine learning framework based on laser speckle imaging for non-destructive assessment of dentin-composite interfacial bond integrity. Different bonding protocols involving an Au NP-modified adhesive system and laser irradiation at controlled stages of the bonding procedure were investigated to produce variations in the dentin–composite interfacial conditions. Speckle patterns were collected from the bonded interface using a 632 nm He-Ne laser and morphologically processed to extract statistical features characterizing interfacial heterogeneity. Permutation-based feature importance ranking was applied using a random forest (RF) model. The top seven ranked features were subsequently used to train support vector regression (SVR) and RF models using specimen-level measurements of shear bond strength as the reference outcome. SVR achieved a calibration coefficient of determination (R2) of 0.914 and a test RMSE of 0.234 MPa, while RF achieved a calibration R2 of 0.919 and a test RMSE of 0.228 MPa. Both models showed strong agreement between predicted and mechanically measured shear bond strength values. Morphological laser speckle descriptors provide a promising non-destructive approach for characterizing dentin–composite interface heterogeneity under controlled in vitro conditions, supporting the potential feasibility of laser speckle imaging as an optical predictive tool for interfacial bond integrity assessment. These encourgaing results support further investigation of the proposed approach under clinically relevant conditions.
While magnetocardiography (MCG) is a non-invasive tool for assessing cardiac electrophysiology, its conventional planar sensor array often struggles to accurately estimate sources distant from the sensor. This limitation hinders the precise localization of deep cardiac activity. This study aims to overcome this challenge by investigating a novel cylindrical sensor array designed to improve source current estimation accuracy. We combined numerical simulations, phantom measurements, and in vivo rat experiments to compare a cylindrical sensor array with a conventional planar array. Single dipole localization was used in the simulation and phantom experiments under controlled focal source conditions, whereas both single-dipole fitting and minimum norm estimation (MNE)-based distributed source analyses were applied to the animal MCG data. The simulations and phantom measurements showed that the cylindrical array achieved more accurate single-dipole localization than the planar array under comparable sensor–source distance conditions. In the animal experiments, this advantage was not clearly observed using single-dipole fitting. The MNE-based distributed source analysis showed geometry-dependent reconstruction patterns, with the cylindrical array producing more posteriorly distributed activity and higher source map repeatability. The cylindrical sensor array provided complementary spatial information to anterior planar sampling, particularly for deeper or posterior cardiac activity. These findings suggest that surrounding sensor coverage may provide complementary information for MCG-based reconstruction, although further validation with realistic activation models remains necessary.
Exposure to head acceleration events in rugby can lead to both physical and mental health problems, even without a diagnosed traumatic brain injury. This study aimed to explore relationships between gameplay head impact conditions and head acceleration metrics experienced by rugby players. Using head impact kinematic data from two high school rugby union teams in New Zealand (one male and one female), linear mixed-effect models identified significant trends between gameplay techniques, head impact mechanisms, and the resulting linear and rotational peak accelerations and injury criteria. Head impacts to bony body regions, crushing head impacts, and head impacts to the ground produced the highest linear and rotational acceleration metrics. Impacts between the head and the ball also resulted in higher rotational accelerations than head impacts with the hip and soft body regions. Indirect head acceleration events and head-to-tackle pad impacts resulted in the lowest acceleration metrics. Males showed lower acceleration metrics during training drills, mauls, and scrums than females. Impacts to the chin resulted in the highest peak accelerations, followed by impacts to the rear of the head, with impacts to the front, top, or side of the head resulting in the lowest acceleration metrics. Indirect head acceleration events resulting in flexion/extension or lateral rotational motion induced the highest acceleration metrics. The tackling technique had limited effect on the acceleration metrics of the tackler, with tackles made to the torso/hip region and upper leg region, both with the head beside the tackle location, resulting in the highest peak accelerations. This analysis could indicate gameplay events that require further investigation to improve player safety without employing protective equipment.
Aortic Stenosis (AS) involves progressive fibrocalcific degeneration of the aortic valve. Current imaging methods for valve tissue quantification remain semi-automated and lack detailed spatial tissue characterization. We present a CT-based pipeline that fully automates the quantification of fibrotic and calcific components and enables semi-automated regional tissue analysis. We developed a standardized pipeline that performs automatic multiplanar reconstruction of pre-TAVI cardiac CT scans, automatic fibrocalcific volume quantification, and semi-automated regional volume assessment requiring only a single user input. We evaluated the accuracy of the pipeline by comparing each module with manual analyses performed independently by two radiologists with different levels of expertise. Agreement between automated and manual results was assessed using Bland–Altman analysis and standard similarity metrics. We analyzed 25 patients with severe AS (13 females/12 males; median age 82 years, IQR 78–86). The pipeline processed for each patient in ≤ 2.5 min. Multiplanar reconstruction showed an angular error of 5.38° (IQR 3.29–8.33) between the expert user and the pipeline. Bland–Altman analysis revealed minimal bias for fibrocalcific volume and density, comparable to inter-user variability. Expert-pipeline Dice coefficients for regional divisions were high (0.86–0.89), and entropy of fibrocalcific ratio demonstrated substantial tissue heterogeneity among patients. The proposed pipeline enables accurate assessment of aortic valve disease severity by integrating fibrotic and calcific quantification with spatial distribution, offering potential to inform TAVI planning and clinical decision-making with minimal user input and analysis time. Trial Registration: NCT06029400, registered on 1 September 2023
Theia3D, a markerless motion capture system, offers a practical alternative to marker-based systems for movement assessment. However, its concurrent validity in patients with knee osteoarthritis remains unevaluated. This study aimed to evaluate the concurrent validity of Theia3D against a marker-based system in measuring joint kinematics and spatiotemporal gait parameters in patients with knee osteoarthritis and to compare error metrics between patients and healthy controls. A total of 162 patients with advanced knee osteoarthritis and 50 healthy controls performed self-selected speed walking and sit-to-stand tasks. Data were simultaneously collected using markerless and marker-based systems. Measurement agreement and errors for spatiotemporal and kinematic parameters were assessed using Bland–Altman analysis, mean difference, Pearson correlation, intraclass correlation coefficient (ICC), and root mean square error (RMSE). Group differences in ICC and RMSE were evaluated using independent-samples t tests or Wilcoxon rank-sum tests. Theia3D showed excellent agreement and very strong correlation with the marker-based system for spatiotemporal gait parameters. Sagittal hip and knee angles demonstrated good to excellent agreement, while agreement in most frontal and transverse joint angles was poor. Compared to controls, knee osteoarthritis patients showed significantly lower ICCs and higher RMSEs for certain joint angles. These findings suggest that Theia3D has potential as an alternative to marker-based systems for assessing spatiotemporal gait parameters and sagittal knee angles during walking. However, since most joint angle errors exceed the 5° clinical acceptability threshold, future work should expand training datasets to include diverse clinical populations and refine algorithms to enable broader clinical implementation.
During off-nominal landings, spacecraft reentry capsules can experience impact accelerations far exceeding the 8–12 g typical of nominal returns, with peak values surpassing 50 g and rise times under 100 ms. Such loading conditions pose multi-organ injury risks to astronauts. This study systematically investigates how seatback inclination modulates these risks. Drop tower tests were performed at three severity levels (Low: 18 g, Mid: 32 g, High: 45 g) using a Hybrid III 50th percentile male anthropomorphic test device. The measured velocity pulses then drove THUMS AM50 v7.0 simulations across five seatback inclinations (0°, 10°, 20°, 30°, 40°). Injury risk was evaluated using dynamic injury criteria (HIC15, Nij, Thoracic Dmax, and Lumbar Fmax) supplemented by tissue-level mechanical parameters (Brain CSDM0.2, Lung MPS and CSDM0.343, Myocardial VMS, Aorta MPS). Segmented linear regression with logit-transformed probabilities quantified the inclination–injury probability relationship. A sensitivity analysis varying the occupant–seat contact condition (friction coefficient ± 50
The mobility and life quality of a user with an above-knee amputation depend on the performance of their prosthetic leg. Prosthetic leg performance is a result of the quality of the prosthetic components, the interaction between them, and the availability of continuous care. Access to these factors is limited in low- and middle-income countries where the majority of lower limb amputees live. This study presents a method to design prosthetic legs based on the operation of all components in tandem to minimize the time for adjustment and the cost of a prosthetic leg. The full leg optimization (FLO) framework provides a method to quantitatively design prosthetic legs by predicting the interaction of the prosthetic foot and prosthetic knee at the initiation of knee flexion during the late stance phase. A prosthetic leg was designed for a person with an above-knee amputation, consisting of a Hip Trajectory Error foot and a passive prosthetic knee. The leg was designed so that the prosthetic knee would unlock at 72% of stance based on the predicted interaction of the foot and knee. The preliminary results demonstrate consistent knee unlocking at 73 ± 3.8% of stance after minimal acclimation time. This suggests that the FLO framework can accurately predict the interaction between the prosthetic foot and the prosthetic knee. Hence, the FLO framework is a design method for prosthetic legs for users with an above-knee amputation, providing a predictable performance, which would be beneficial for users living in low-resource settings.
Quantitative assessment of myocardial deformation is increasingly important in clinical cardiology, yet conventional two-dimensional (2D) echocardiography and standard three-dimensional (3D) approaches remain limited by out-of-plane motion and incomplete characterization of transmural mechanics. To address these limitations, we introduce a physics-informed framework for 3D echocardiography that reconstructs the full finite strain tensor across the entire myocardial wall. As an initial methodological study, we demonstrate the framework and validate it against cardiac magnetic resonance in a small cohort. Endocardial and epicardial surfaces were segmented from 3D echocardiographic datasets and tracked throughout the cardiac cycle using speckle-tracking techniques. An optimization framework with a soft volumetric penalty was implemented, permitting volume change at finite cost while maintaining tracking fidelity and geometric smoothness. The resulting deformation field enabled reconstruction of the complete 3D strain tensor. Global strain measurements derived from the method were validated against cardiac magnetic resonance (CMR) measurements obtained in two subjects. Global longitudinal and circumferential strain values obtained from the proposed framework showed strong agreement with CMR measurements. The optimization procedure also demonstrated robustness to segmentation variability and reduced errors associated with epicardial tracking. Beyond conventional strain indices, the method enabled reconstruction of spatially resolved principal strain fields throughout the ventricular wall, revealing physiologically consistent transmural gradients and contraction patterns. Physics-informed integration of speckle tracking with biomechanical constraints enables robust reconstruction of 3D myocardial deformation from echocardiography. This framework provides a comprehensive and physically consistent characterization of myocardial mechanics from widely available 3D echocardiographic data. These initial results support the feasibility of the proposed framework and motivate future evaluation in larger, more diverse patient cohorts to establish its clinical reliability.
To investigate the ability of multiscale entropy (MSE) methods to characterize alterations in airflow dynamics in chronic obstructive pulmonary disease (COPD) and to evaluate their associations with lung function and diagnostic performance. Spontaneous airflow signals were recorded during 150 s from 29 controls and 25 patients with COPD. Signal complexity was quantified using MSE, refined composite MSE (RCMSE), refined composite multiscale fuzzy entropy (RCMFE), and multiscale permutation entropy (MPE). The area under the entropy curve was calculated over predefined scale ranges (S0-50, S50-100, and S0-100), and entropy at τ = 1 was also analyzed. Associations with spirometry and respiratory resistance (R6) were evaluated, and diagnostic performance was assessed using ROC analysis. COPD patients showed a consistent reduction in airflow complexity across all entropy measures. Entropy indices were positively associated with spirometric parameters and inversely associated with R6, indicating progressive loss of complexity with increasing airway obstruction. The strongest correlations were observed at τ = 1, particularly with R6. Among multiscale indices, S0-50 showed the strongest associations with lung function, suggesting predominant involvement of short-term airflow dynamics, consistent with small airway abnormalities in COPD. ROC analysis demonstrated high diagnostic accuracy for MSE and RCMSE (AUC > 0.90 for S0-50 and SampEn1), with comparable performance for MPE and lower accuracy for RCMFE. Multiscale entropy analysis of spontaneous airflow provides a sensitive, noninvasive approach for detecting and characterizing respiratory dysfunction in COPD. These methods capture clinically relevant alterations in airflow dynamics and offer complementary information to conventional lung function tests. Further studies are warranted to validate these findings and enhance their clinical applicability.
To develop and evaluate an automated workflow for the setup of single-vertebra finite element (FE) simulations from clinical CT data. Specifically, we quantified how automated endplate identification, vertebra-specific coordinate system definition, and load-application-point assignment influence the simulated fracture-load estimates. We analyzed 113 vertebrae from 70 patients that had previously undergone manual FE setup. The automated pipeline identified vertebral endplates, assigned a vertebra-specific coordinate system, and assigned the load application point for axial compression simulations. Automated setups were visually graded as good, acceptable, or bad using predefined criteria. Agreement with manual reference models was evaluated, and the effects of each setup component on fracture-load estimates were quantified separately. Of 113 vertebrae, 53 (47
For over half a century, our method of annuloplasty ring sizing for mitral (tricuspid) repair has remained unchanged. Commercially available sizers provide the surgeon estimates but no objective measurement. This manuscript defines the engineering design process for a novel, objective sizing system and illustrates its clinical application. An early cadaver animal heart model was developed to validate the sizer's ability to both manipulate and measure coaptation length. An overview of the design process is presented to establish why the paired-ring sizing method is able to deliver precise mitral leaflet coaptation length measurements. The method of clinical application of a novel surgical technique is illustrated in detail. Preliminary animal heart validation results demonstrated the sizer's ability to effectively manipulate mitral annular structure and correctly predict changes in coaptation length. Because the paired-ring sizing technique allows coaptation length to be accurately measured and controlled, it has the potential—pending further study—to serve as the basis for a scientific annuloplasty sizing approach.
Inertial microfluidics is well established for separating particles and cells by size and mechanical properties; however, its application for isolating cell clusters from single blood cells in meandering channels remains largely unexplored. This study demonstrates the potential of inertial focusing in S-shaped meandering channels to separate two-cell clusters from single cells. We first analyzed the cumulative secondary transport using passive-tracer trajectories recorded at phase-matched sections of repeated S-shaped periods. This Lagrangian analysis revealed four recurrent transport regions that help explain the observed size- and shape-dependent focusing behavior. We further developed a computational model, validated against experimental data in a nine-section meandering microchannel, which accurately predicts the distinct focusing positions of 8μm and 15μm particles–characteristic sizes of human red blood cells (RBCs) and leukocytes. The model was further validated against experiments with red blood cells and the MCF7 cell line, reproducing the observed focusing modes and transverse focusing positions. Subsequently, we evaluated in silico the separation of two-cell clusters (formed by two adhered 8μm or 15μm cells, modeling circulating tumor cell clusters or leukocyte aggregates) from single blood cells across six microchannel designs of three varying widths and two heights of the channel. For different cell mixture compositions, we provide specific guidelines for optimizing channel geometry and flow velocity to achieve high-purity separation. This work establishes a systematic framework for designing inertial microfluidic devices to isolate cell clusters using meandering microchannels.
