Annals of Biomedical Engineering最新文献

Tooth cracks, one of the most common dental diseases, can result in the tooth falling apart without prompt treatment; dentists also have difficulty locating cracks, even with X-ray imaging. Indocyanine green (ICG) assisted near-infrared fluorescence (NIRF) dental imaging technique can solve this problem due to the deep penetration of NIR light and the excellent fluorescence characteristics of ICG. This study extracted 593 human cracked tooth images and 601 non-cracked tooth images from NIR imaging videos. Multiple imaging analysis methods such as classification, object detection, and super-resolution were applied to the dataset for cracked image analysis. Our results showed that machine learning methods could help analyze tooth crack efficiently: the tooth images with cracks and without cracks could be well classified with the pre-trained residual network and squeezenet1_1 models, with a classification accuracy of 88.2% and 94.25%, respectively; the single shot multi-box detector (SSD) was able to recognize cracks, even if the input image was at a different size from the original cracked image; the super-resolution (SR) model, SR-generative adversarial network demonstrated enhanced resolution of crack images using high-resolution concrete crack images as the training dataset. Overall, deep learning model-assisted human crack analysis improves crack identification; the combination of our NIR dental imaging system and deep learning models has the potential to assist dentists in crack diagnosis.

An estimated 6.8 million people in the United States have an unruptured intracranial aneurysms, with approximately 30,000 people suffering from intracranial aneurysms rupture each year. Despite the development of population-based scores to evaluate the risk of rupture, retrospective analyses have suggested the limited usage of these scores in guiding clinical decision-making. With recent advancements in imaging technologies, artery wall motion has emerged as a promising biomarker for the general study of neurovascular mechanics and in assessing the risk of intracranial aneurysms. However, measuring arterial wall deformations in vivo itself poses several challenges, including how to image local wall motion and deriving the anisotropic wall strains over the cardiac cycle. To overcome these difficulties, we first developed a novel in vivo MRI-based imaging method to acquire cardiac gated images of the human basilar artery (BA) over the cardiac cycle. Next, complete BA endoluminal surfaces from each frame were segmented, producing high-resolution point clouds of the endoluminal surfaces. From these point clouds we developed a novel B-spline-based surface representation, then exploited the local support nature of B-splines to determine the local endoluminal surface strains. Results indicated distinct regional and temporal variations in BA wall deformation, highlighting the heterogeneous nature BA function. These included large circumferential strains (up to $\sim$ 20 $\%$ ), and small longitudinal strains, which were often contractile and out of phase with the circumferential strains patterns. Of particular interest was the temporal phase lag in the maximum circumferential perimeter length, which indicated that the BA deforms asynchronously over the cardiac cycle. In summary, the proposed method enabled local deformation analysis, allowing for the successful reproduction of local features of the BA, such as regional principal stretches, areal changes, and pulsatile motion. Integrating the proposed method into existing population-based scores has the potential to improve our understanding of mechanical properties of human BA and enhance clinical decision-making.

Purpose: Ultrasound imaging is key in the management of patients with an abdominal aortic aneurysm (AAA). It was recently shown that the cyclic diameter variations between diastole and systole, which can be quantified with US imaging, increase significantly with the strength of the applied probe pressure on the patient's abdomen. The goal of this study is to investigate this effect more thoroughly.

Methods: With finite-element modeling, pulsatile blood pressure and probe pressure are simulated in three patient-specific geometries. Two distinct models for the aortic wall were simulated: a nonlinear hyperelastic and a linear elastic model. In addition, varying stiffness was considered for the surrounding tissues. The effect of light, moderate, and firm probe pressure was quantified on the stresses and strains in the aortic wall, and on two in vivo stiffness measures. In addition, the Elasticity Loss Index was proposed to quantify the change in stiffness due to probe pressure.

Results: Firm probe pressure decreased the measured aortic stiffness, and material stiffness was affected only when the wall was modeled as nonlinear, suggesting a shift in the stress-strain curve. In addition, stiffer surrounding tissues and a more elongated aneurysm sac decreased the responsiveness to the probe pressure.

Conclusion: The effect of probe pressure on the AAA wall stiffness was clarified. In particular, the AAA wall nonlinear behavior was found to be of primary importance in determining the probe pressure response. Thus, further work will intend to make use of this novel finding in a clinical context.

Purpose: To obtain high-resolution velocity fields of cerebrospinal fluid (CSF) and cerebral blood flow by applying a physics-guided neural network (div-mDCSRN-Flow) to 4D flow MRI.

Methods: The div-mDCSRN-Flow network was developed to improve spatial resolution and denoise 4D flow MRI. The network was trained with patches of paired high-resolution and low-resolution synthetic 4D flow MRI data derived from computational fluid dynamic simulations of CSF flow within the cerebral ventricles of five healthy cases and five Alzheimer's disease cases. The loss function combined mean squared error with a binary cross-entropy term for segmentation and a divergence-based regularization term for the conservation of mass. Performance was assessed using synthetic 4D flow MRI in one healthy and one Alzheimer' disease cases, an in vitro study of healthy cerebral ventricles, and in vivo 4D flow imaging of CSF as well as flow in arterial and venous blood vessels. Comparison was performed to trilinear interpolation, divergence-free radial basis functions, divergence-free wavelets, 4DFlowNet, and our network without divergence constraints.

Results: The proposed network div-mDCSRN-Flow outperformed other methods in reconstructing high-resolution velocity fields from synthetic 4D flow MRI in healthy and AD cases. The div-mDCSRN-Flow network reduced error by 22.5% relative to linear interpolation for in vitro core voxels and by 49.5% in edge voxels.

Conclusion: The results demonstrate generalizability of our 4D flow MRI super-resolution and denoising approach due to network training using flow patches and physics-based constraints. The mDCSRN-Flow network can facilitate MRI studies involving CSF flow measurements in cerebral ventricles and association of MRI-based flow metrics with cerebrovascular health.

Purpose: Craniofacial osteotomies involving pterygomaxillary disjunction are common procedures in maxillofacial surgery. Surgeons still rely on their proprioception to determine when to stop impacting the osteotome, which is important to avoid complications such as dental damage and bleeding. Our group has developed a technique consisting in using an instrumented hammer that can provide information on the mechanical properties of the tissue located around the osteotome tip. The aim of this study is to determine whether a mallet instrumented with a force sensor can be used to predict the crossing of the osteotome through the pterygoid plates.

Methods: 31 osteotomies were carried out in 16 lamb skulls. For each impact, the force signal obtained was analysed using a dedicated signal processing technique. A prediction algorithm based on an SVM classifier and a cost matrix was applied to the database.

Results: We showed that the device could always detect the crossing of the osteotome, sometimes before its occurrence. The prediction accuracy of the device was 94.7%. The method seemed to be sensitive to the thickness of the plate and to crack apparition and propagation.

Conclusion: These results pave the way for the development of a per-operative decision support system in maxillofacial surgery.

Purpose: This study aimed to assess the feasibility of early detection of fatigued gait patterns for older adults through the development of a smart portable device.

Methods: The smart device incorporated seven force sensors and a single inertial measurement unit (IMU) to measure regional plantar forces and foot kinematics. Data were collected from 18 older adults walking briskly on a treadmill for 60 min. The optimal feature set for each recognition model was determined using forward sequential feature selection in a wrapper fashion through fivefold cross-validation. The recognition model was selected from four machine learning models through leave-one-subject-out cross-validation.

Results: Five selected characteristics that best represented the state of fatigue included impulse at the medial and lateral arches (increased, p = 0.002 and p < 0.001), contact angle and rotation range of angle in the sagittal plane (increased, p < 0.001), and the variability of the resultant swing angular acceleration (decreased, p < 0.001). The detection accuracy based on the dual signal source of IMU and plantar force was 99%, higher than the 95% accuracy based on the single source. The intelligent portable device demonstrated excellent generalization (ranging from 93 to 100%), real-time performance (2.79 ms), and portability (32 g).

Conclusion: The proposed smart device can detect fatigue patterns with high precision and in real time.

Significance: The application of this device possesses the potential to reduce the injury risk for older adults related to fatigue during gait.