Biomedical Physics & Engineering Express最新文献

Hemodialysis is a crucial procedure for removing toxins and waste from the body when kidneys fail to perform this function effectively. This study addresses the need to improve the efficiency and biocompatibility of membranes used in dialyzers. We simulate fluid flow through two types of membranes, Cuprophan (cellulosic) and AN69ST (synthetic), to understand the complex mechanisms involved and quantify key variables such as pressure, concentration, and flow. This study presents a detailed model that applies mass conservation equations and Navier-Stokes principles adapted for porous media, along with heat and mass transfer considerations. The results revealed significant differences in the flow behavior and filtration efficiency between the two membranes, highlighting the superiority of the AN69ST membrane in terms of flow rate and toxin removal. This model serves as a valuable tool for characterizing new porous membranes in dialysis applications, enabling the prediction of the temperature, pressure, and concentration profiles. By providing this information without requiring extensive experimentation, the model complements the design and evaluation of new membranes and, optimizes their development. The ability to predict these profiles is crucial because they directly influence the parameters that determine treatment effectiveness. Moreover, this study underscores the importance of continued innovation in membrane materials and designs, contributing to improved clinical outcomes and treatment efficiency, representing a significant advancement in healthcare.

Capacitive-based radiofrequency (Rf) radiation at 27 MHz offers a non-invasive approach for inducing hyperthermia, making it a promising technique for thermal cancer therapy applications. To achieve focused and site-specific hyperthermia, Rf-responsive materials is required to convert Rf radiation into localized heat efficiently. Nanoparticles capable of absorbing Rf energy and convert into heat for targeted ablation are of critical importance. In this study, we developed and evaluated an Intra-tumoral injectable magnetic hydrogel (IT-MG) composed of Superparamagnetic Iron Oxide Nanoparticles (SPIONs) impregnated in low molecular weight Hyaluronic Acid (HA) forming HA-SPIONs. Our systematic investigation revealed that HA-SPIONs exposed to Rf radiation significantly increased temperature, reaching up to 50 °C. Further testing in tissue-mimicking phantom models also showed consistent heating, with temperatures stabilizing at 43 °C, ideal for localized hyperthermia. The ability of HA-SPIONs to act as an effective localized heating agent when exposed to 27 MHz Rf radiation, reaching apoptosis-inducing temperature, has not been previously reported. In conclusion, synergistic effects of IT-MG in bothin-vitroand tumor-mimicking phantom models demonstrate improved and localized hyperthermia, facilitating adjuvant cancer treatment.

Purpose. PDD and profile curves play a crucial role in analyzing the beam quality and energy stability of accelerators. The aim of this study was to assess the efficacy of GPR in machine QA and compare it with traditional methods for analyzing dose outputs.Methods. GPRs were employed to assess the quality of radiation beams by comparing 1D and 2D Profile metrics and PDD data against commissioning data. The data used were obtained from the ASCII data files derived from the water tank. GPRs were calculated for all plots with a lower percentage dose cutoff of 10%. The local GPRs and dose influence for the 2D PDD metrics and dose influence were calculated for an open field 10 × 10 cm²photon beam at SSD = 100 cm. In both 1D and 2D GPRs analyses, criterion of 1%/1 mm was adopted, as this approach allows for the capture of more subtle variations in the data. To substantiate the viability of the study, a comparative analysis was conducted by comparing the outcomes of the gamma analysis with those derived from traditional methods, such as manual machine quality assurance checks.Results. GPRs demonstrated a superior capability for comprehensive data analysis compared to traditional methods. For the 1D curves, the passing rates (γ≤ 1) are 96.19%, 100%, and 93.46%, respectively. With respect to the 2D dose influence, the PDD image passing rate was 99.57%, and significant dose differences were observed at the four corners of the open field, indicating areas that require further investigation.Conclusions. Compared to traditional methods, GPRs are more sensitive to subtle changes in the data, providing valuable insights into the accelerator beam status.

Passive brain-computer interfaces (passive BCIs, pBCIs) enable computers to unobtrusively decipher aspects of a user's mental state in real time from recordings of brain activity, e.g. electroencephalography (EEG). When used during human-computer interaction (HCI), this allows a computer to dynamically adapt for enhancing the subjective user experience. For transitioning from controlled laboratory environments to practical applications, understanding BCI performance in real contexts is of utmost importance. Here, Virtual Reality (VR) can play a unique role: both as a fully controllable simulation of a realistic environment and as an independent, increasingly popular real application. Given the potential of VR as a dynamic and controllable environment, and the capability of pBCIs to enable novel modes of interaction, it is tempting to envision a future where pBCI and VR are seamlessly integrated. However, the simultaneous use of these two technologies - both of which are head-mounted - presents new challenges. Due to their immediate proximity, electromagnetic artifacts can arise, contaminating the EEG. Furthermore, the active movements promoted by VR can induce mechanical and muscular artifacts in the EEG. The varying body postures and display preferences of users further complicate the practical application of pBCIs. To address these challenges, the current study investigates the influence of body posture (sitting vs. standing) and display media (computer screen vs. VR) on the performance of a pBCI in assessing cognitive load. Our results show that these conditions indeed led to some changes in the EEG data; nevertheless, the ability of pBCIs to detect cognitive load remained largely unaffected. However, when a classifier trained in one context (body posture or modality) was applied to another (e.g., cross-task application), reductions in classification accuracy were observed. As HCI moves towards increasingly adaptive and more interactive designs, these findings support the expansive potential of pBCIs in VR contexts.

3D-printed boluses in radiation therapy are of increasing interest for enhancing treatment precision and patient comfort. A comprehensive clinical validation of these boluses remains to be established. This study aims to confirm the effectiveness of a 3D-printed bolus through a proof-of-concept comparative validation, by implementing in a clinical setting a bolus made of PLA and designed to ensure uniform dose coverage for a case in the eye region. In this study the 3D-printed bolus was compared to two commercially available boluses (one thermoplastic and one skin type) by using a refecence where no bolus was present (with the optimal dose distribution scenario). All boluses were placed on an anthropomorphic head phantom and BeOSL detectors were used to measure dose values to determine the level of their effectiveness on delivery. During the scanning process, a thermoplastic mask was used to prevent bolus movement and to accurately reproduce clinical scenarios. Differences in dose values at D_maxand D_50%revealed the performance of each bolus. The treatment planning system (TPS) and BeOSL readings for the 3D printed bolus were within 2% (the clinical tolerance), with 0.66% dose difference for the customized 3D-printed bolus. Although the thermoplastic bolus had the closest value to the detector reading, with a score of 0.30%, this result was influenced by improper shaping of the bolus on the phantom and the presence of a wide air gap, which caused lack of eye covering. Whereas, the skin bolus, due to higher volume of air between phantom surface and bolus, showed a 1.29% dose difference between the TPS values and the OSL detector readings. We provide a comparative validation for the use of 3D printed boluses and highlight that proper bolus fitting is essential in clinical settings to avoid air gaps and to maintain dose distribution over multiple treatment sessions.

For brain-dedicated positron emission tomography (PET) scanners, depth-of-interaction (DOI) information is essential to achieve uniform spatial resolution across the field-of-view (FOV) by minimizing parallax error. Time-of-flight (TOF) information can enhance the image quality. In this study, we proposed a novel monolithic U-shaped crystal design that had a tapered geometry to achieve good coincidence timing resolution (CTR) and DOI resolution simultaneously. We compared a novel tapered U-shaped crystal design with a conventional flat-top geometry for PET detectors. Each crystal had outer dimensions of 5.85 × 2.75 × 15 mm³, with a 0.2 mm central gap forming physically isolated bottom surfaces (2.85 × 2.75 mm²). The novel U-shape crystal design with a tapered top roof resulted in the best CTR of 201 ± 3 ps, and DOI resolution of 3.1 ± 0.6 mm, which were better than flat top geometry. In the next study, we plan to optimize the crystal surface treatment and reflector to further improve the CTR and DOI resolution.

This study presents the design and validation of a neonatal head phantom using innovative heterogeneous composite materials customized to replicate the x-ray attenuation properties of neonatal cranial structures. Analysis of Hounsfield Unit (HU) data from 338 neonatal head CT scans informed the design of epoxy resin-based composites with additives such as sodium bicarbonate, fumed silica, and acetone to simulate bone, brain matter, cerebrospinal fluid (CSF) and hyperdense abnormalities. The cranial bone substitute (60% epoxy resin, 40% sodium bicarbonate) achieved a density of 1.60 g cm^-3, with HU values (574.67-608.04) closely matching clinical ranges. Brain matter (95% epoxy resin, 5% acetone) achieved HU values (35.27-43.61), aligning with clinical means, while the CSF-equivalent material (80% epoxy resin, 15% fumed silica, 5% acetone) matched neonatal CSF HU values (14.53-17.02). A mass substitute for hyperdense abnormalities exhibited HU values (56.16-61.07), enabling differentiation from normal brain. Validation included Monte Carlo simulations and experimental CT imaging, showing close agreement in linear attenuation coefficients, with deviations below 11% across energy levels. Mass attenuation coefficients from simulations and XCOM software were consistent, with deviations under 0.7%, confirming the materials dosimetric reliability. The phantom, with a cylindrical geometry (9 cm diameter, 10 cm length), provides accurate attenuation properties across 80-120 kVp energy levels, with deviations below 5% between experimental CT numbers and simulation data. This phantom offers a robust platform for neonatal imaging research, enabling impactful dose optimization and imaging protocol adjustment and supports improved diagnostic accuracy in pediatric imaging.

Purpose. To investigate the relationship between spatial parotid dose and the risk of xerostomia in patients undergoing head-and-neck cancer radiotherapy, using machine learning (ML) methods.Methods. Prior to conducting voxel-based ML analysis of the spatial dose, two steps were taken: (1) The parotid dose was standardized through deformable image registration to a reference patient; (2) Bilateral parotid doses were regrouped into contralateral and ipsilateral portions depending on their proximity to the gross tumor target. Individual dose voxels were input into six commonly used ML models, which were tuned with ten-fold cross validation: random forest (RF), ridge regression (RR), support vector machine (SVM), extra trees (ET), k-nearest neighbor (kNN), and naïve Bayes (NB). Binary endpoints from 240 patients were used for model training and validation: 0 (N = 119) for xerostomia grades 0 or 1, and 1 (N = 121) for grades 2 or higher. Model performance was evaluated using multiple metrics, including accuracy, F₁score, areas under the receiver operating characteristics curves (auROC), and area under the precision-recall curves (auPRC). Dose voxel importance was assessed to identify local dose patterns associated with xerostomia risk.Results. Four models, including RF, SVM, ET, and NB, yielded average auROCs and auPRCs greater than 0.60 from ten-fold cross-validation on the training data, except for a lower auROC from NB. The first three models, along with kNN, demonstrated higher accuracy and F₁scores. A bootstrapping analysis confirmed test uncertainty. Voxel importance analysis from kNN indicated that the posterior portion of the ipsilateral gland was more predictive of xerostomia, but no clear patterns were identified from the other models.Conclusion. Voxel doses as predictors of xerostomia were confirmed with some ML classifiers, but no clear regional patterns could be established among these classifiers, except kNN. Further research with a larger patient dataset is needed to identify conclusive patterns.

Brain-computer interfaces (BCIs) offer disabled individuals the means to interact with devices by decoding the electroencephalogram (EEG). However, decoding intent in fine motor tasks can be challenging, especially in stroke survivors with cortical lesions. Here, we attempt to decode graded finger extension from the EEG in stroke patients with left-hand paresis and healthy controls. Participants extended their fingers to one of four levels: low, medium, high, or 'no-go' (none), while hand, muscle (electromyography: EMG), and brain (EEG) activity were monitored. Event-related desynchronization (ERD) was measured as the change in 8-30 Hz EEG power during movement. Classifiers were trained on EEG features, EMG power, or both (EEG+EMG) to decode finger extension, and accuracy assessed via four-fold cross-validation for each hand of each participant. Mean accuracy exceeded chance (25%) for controls (n = 11) at 62% for EMG, 60% for EEG, and 71% for EEG+EMG on the left hand; and 67%, 60%, and 74%, respectively, on the right hand. Accuracies were similar on the unimpaired right hand for the stroke group (n = 3): 61%, 68%, and 78%, respectively. But on the paretic left hand, EMG only discriminated no-go from movement above chance (41%); in contrast, EEG gave 65% accuracy (68% for EEG+EMG), comparable to the non-paretic hand. The median ERD was significant (p < 0.01) over the cortical hand area in both groups and increased with each level of finger extension. But while the ERD favored the hemisphere contralateral to the active hand as expected, it was ipsilateral for the left hand of stroke due to the lesion in the right hemisphere, which may explain its discriminative ability. Hence, the ERD captures effort in finger extension regardless of success or failure at the task; and harnessing residual EMG improves the correlation. This marker could be leveraged in rehabilitative protocols that focus on fine motor control.

Photoacoustic tomography (PAT) is a non-destructive, non-ionizing, and rapidly expanding hybrid biomedical imaging technique, yet it faces challenges in obtaining clear images due to limited data from detectors or angles. As a result, the methodology suffers from significant streak artifacts and low-quality images. The integration of deep learning (DL), specifically convolutional neural networks (CNNs), has recently demonstrated powerful performance in various fields of PAT. This work introduces a post-processing-based CNN architecture named residual-dense UNet (RDUNet) to address the stride artifacts in reconstructed PA images. The framework adopts the benefits of residual and dense blocks to form high-resolution reconstructed images. The network is trained with two different types of datasets to learn the relationship between the reconstructed images and their corresponding ground truths (GTs). In the first protocol, RDUNet (identified as RDUNet I) underwent training on heterogeneous simulated images featuring three distinct phantom types. Subsequently, in the second protocol, RDUNet (referred to as RDUNet II) was trained on a heterogeneous composition of 81% simulated data and 19% experimental data. The motivation behind this is to allow the network to adapt to diverse experimental challenges. The RDUNet algorithm was validated by performing numerical and experimental studies involving single-disk, T-shape, and vasculature phantoms. The performance of this protocol was compared with the famous backprojection (BP) and the traditional UNet algorithms. This study shows that RDUNet can substantially reduce the number of detectors from 100 to 25 for simulated testing images and 30 for experimental scenarios.