Journal of Medical and Biological Engineering最新文献

Purpose

Cardiac amyloidosis (CA) is a highly underdiagnosed disease characterized by the accumulation of misfolded amyloid protein fragments in the heart, resulting in reduced heart functionality and myocardial stiffness. Artificial intelligence (AI) has garnered considerable interest as a potential tool for diagnosing cardiovascular diseases, including CA. This systematic review concentrates on the application of AI in the diagnosis of CA.

Methods

A comprehensive systematic search was performed on the databases of PubMed, Embase, and Medline, to identify relevant studies. The screening process was conducted in two stages, using predetermined inclusion and exclusion criteria, and was carried out in a blinded manner. In cases where discrepancies arose, the reviewers discussed and resolved the issue through consensus.

Results

Following the screening process, a total of 10 studies were deemed eligible for inclusion in this review. These investigations evaluated the potential utility of AI models that analyzed routine laboratory data, medical records, ECG, transthoracic echocardiography, CMR, and WBS in the diagnosis of CA.

Conclusion

AI models have demonstrated utility as a diagnostic tool for CA, with comparable or in one case superior efficacy to that of expert cardiologists.

Purpose

Surface-enhanced Raman scattering (SERS) is a technique for trace analysis detection based on the interaction of light with matter and between materials. In the past development of SERS, precious metals were primarily chosen as substrates due to their high electromagnetic effect, which leads to significantly enhanced SERS signals. However, the effect of using only precious metals is limited. Therefore, this study utilizes the characteristic micro-nano V-shaped pits that appear on the surface of c-plane GaN after wet etching. By depositing a gold film of various thicknesses, we aim to increase the contact area with the target molecule Rhodamine 6G (R6G), thereby further enhancing the sensitivity of SERS detection.

Methods

After fabricating pitted c-plane GaN using chemical etching techniques, we analyzed the sample surface with a scanning electron microscope and assessed the impact of different gold film thicknesses on the SERS intensity of R6G using Raman spectroscopy. The comprehensive biomedical detection effectiveness was also evaluated using contact angle measurement, and fluorescence microscopy.

Results

For the target molecule R6G, after depositing a 25 nm gold film, the enhancement factor of the substrate for detection reached 2.21×10⁸, and the limit of detection was achieved at a concentration of 10^− 10 M.

Conclusion

This study confirms the feasibility of using wet etching techniques on hexagonal materials like GaN for SERS applications. The GaN substrate with V-shaped pits provides an increased surface area, effectively enhancing SERS signal strength. This offers different choices and perspectives for SERS substrate selection in the detection of various target molecules.

Purpose

Objective motor and gait assessment is crucial for assessing fall risk and predicting intervention efficacy. This study compares the postural control and gait complexity in adults with and without vertebral compression fracture (VCF) during the timed up-and-go (TUG) test.

Methods

The groups of eligible older adults were divided into VCF (n = 21) and Control (without VCF; n = 43). The Biodex Balance System was used for postural stability and motor control tests. A TUG test was conducted, during which an inertial motion system was used to record joint kinematics and center of mass (CoM) trajectories. The gait complexity was assessed using multiscale entropy (MSE) analysis of pelvic acceleration.

Results

The VCF group had poor postural stability and longer times in the motor control test than the control group. During the sit-to-stand phase of the TUG test, the VCF group exhibited more significant mediolateral CoM displacement and less anteroposterior displacement than the control group. The VCF group had more significant vertical CoM displacement, lower acceleration, and lower ranges of motion in the cervical, thoracic, lumbar, and hip joints and longer stance phases than the control group while performing the TUG test. Furthermore, the VCF group had a higher complexity index of gait, indicating lower adaptability than the control group during walking.

Conclusion

Patients with VCF exhibited lower postural stability, potentially increasing their risk of falls. The patients adopted various less stable body configurations during the TUG test. Gait quality characteristics measured through MSE analysis may help evaluate the walking adaptability of individuals at risk of falls.

Purpose

Mild cognitive impairment (MCI) and Alzheimer’s disease (AD) are known to cause geometrical changes in the integrity of the fornix, which plays a crucial role in memory formation and retrieval. The objective of this study is to analyse structural variations in the fornix region using structural magnetic resonance (sMR) images and geometrical features.

Methods

Initially, the fornix region of the brain is segmented from the sMR images of normal cognitive (NC), MCI and AD using the FreeSurfer software package. Further, geometrical features such as volume, equivalent diameter, extent, major axis length, and solidity are extracted to investigate the changes in the structure of the fornix in MCI and AD conditions.

Results

The segmentation results show that FreeSurfer software is able to delineate the irregular boundaries of the fornix region accurately. The extent, major axis length, and solidity features are found to be statistically significant (p < 0.001) in discriminating NC, MCI and AD. It indicates that the considered features can capture the geometrical variation in the fornix structure.

Conclusion

The reported approach can facilitate the early diagnosis of the disease, as the distinction of AD in the preclinical stage is complex and clinically significant.

Purpose

Image registration is a critical component in medical image analysis applications. Optimization algorithms for energy functions play a crucial role in registration. Most registration methods improve the performance by modifying the energy function and optimizing it directly, neglecting the impact of the optimization algorithm. This paper is to investigate how to efficiently design an attention allocation strategy and improve the convergence of the optimization algorithm.

Methods

This paper introduces a novel image registration method that leverages the distributed alternating direction method of multipliers to perform optimization, named DADMMreg. Compared to the optimization algorithm using the alternating direction method of multipliers (ADMM), the optimization algorithm used in DADMMreg achieves better convergence by altering the optimization order of the similarity and regularization terms within the energy function. To overcome the limitations of intensity-based or structural-based similarity metrics, a modified structural similarity measure (SSIM) is proposed that takes into account both intensity and structural information. Considering that homogeneous smoothing prior at the sliding surface leads to inaccurate registration, a novel vector-modulus-based regularization metric is proposed to avoid physically implausible displacement fields.

Results

Experimental results on 4D-CT image dataset and COPD image dataset demonstrate the satisfactory registration performance of DADMMreg, with an average target registration error (TRE) of 0.9105 mm and 0.9201 mm, respectively. Meanwhile, the experimental results show that the DADMMreg method exhibits better convergence performance than other registration methods.

Conclusion

Compared to classical methods, the attention allocation strategy of DADMMreg enables faster convergence with comparable registration accuracy.

Purpose

This systematic review and meta-analysis was conducted to evaluate the usefulness of deep learning (DL) models for aorta segmentation in computed tomography (CT) images.

Methods

Adhering to 2020 PRISMA guidelines, we systematically searched PubMed, Embase, and Web of Science for studies published up to March 13, 2024, that used DL models for aorta segmentation in adults’ chest CT images. We excluded studies that did not use DL models, involved nonhuman subjects or aortic diseases (aneurysms and dissections), or lacked essential data for meta-analysis. Segmentation performance was evaluated primarily in terms of Dice scores. Subgroup analyses were performed to identify variations related to geographical location and methodology.

Results

Our review of 16 studies indicated that DL models achieve high segmentation accuracy, with a pooled Dice score of 96%. We further noted geographical variations in model performance but no significant publication bias, according to the Egger test.

Conclusion

DL models facilitate aorta segmentation in CT images, and they can therefore guide accurate, efficient, and standardized diagnosis and treatment planning for cardiovascular diseases. Future studies should address the current challenges to enhance model generalizability and evaluate clinical benefits and thus expand the application of DL models in clinical practice.

Purpose

To adopt the merits of the Bayesian Penalized Likelihood (BPL) reconstruction algorithm (incl. improved contrast recovery), a deep learning ResNet model was trained to estimate BPL-like images using the non-attenuation, non-scatter corrected PET images (PET-nonAC) as inputs.

Methods

Images of 112 patients were used for model training (79 patients), validation (13 patients) and testing (20 patients). The ResNet model used PET-nonAC images as input and predicted corresponding BPL-like images. The model performance regarding image quality was evaluated using metrics such as contrast-to-noise ratio (CNR).

Results

The CNR of the reference BPL images was 2.40, while estimated BPL-like images using the deep learning model have a CNR value of 2.42 indicative of comparable performance.

Conclusion

The estimated BPL-like images of the deep learning model offer comparable quality to the reference BPL images especially regarding the CNR metric. This deep learning model can be used to improve the image quality PET-nonAC by adopting the characteristics of the BPL images.

Purpose

From the myofibrils to the whole muscle scale, muscle micro-constituents exhibit passive and active mechanical properties, potentially coupled to electrical, chemical, and thermal properties. Experimental characterization of some of these properties is currently not available for all muscle constituents. Multiscale multiphysics models have recently gained interest as a numerical alternative to investigate the healthy and diseased physiological behavior of the skeletal muscle.

Methods

This paper refers to the multiscale mechanical models proposed in the literature to investigate the mechanical properties and behavior of skeletal muscles. More specifically, we focus on the scale transition methods, constitutive laws and experimental data implemented in these models.

Results

Using scale transition methods such as homogenization, coupled to appropriate constitutive behavior of the constituents, these models explore the mechanisms of ageing, myopathies, sportive injuries, and muscle contraction.

Conclusion

Emerging trends include the development of multiphysics simulations and the coupling of modeling with the acquisition of experimental data at different scales, with increasing focus to little known constituents such as the extracellular matrix and the protein titin.

Purpose

The objective of this study was to evaluate the effect of thread shape outline, thread pitch and preload on the central screw loosening or fracturing by using 3D finite element (FE) method.

Methods

By using a commercial implant system as prototypes, nine central screw models and matching implant components were created by combinations of two parameters with different values: three screw shapes (triangle, buttress and reverse buttress) and three screw pitches (0.2 mm, 0.3 mm and 0.4 mm). These models were inserted into a bone block. After applying pre-tightening torque to each central screw, a 100 N load tilted at 45 degrees was applied to the abutment to simulate an occlusal load. The stability performance of the central screws was evaluated through FE software.

Results

For the central screw shape, the highest stress was always located at the level of the first thread, while the highest stress of the reverse buttress group was always located at the level of the second thread. Triangle (average 650 MPa) and reverse buttress (average 729 MPa) threads were more conducive to reducing central screw stress than buttress thread (average 920 MPa). Moreover, it is necessary to avoid thread parameters design with buttress shape/0.3 mm pitch and buttress shape/0.4 mm pitch.

Conclusion

Thread shape outline and thread pitch significantly influenced central screw stability performance. The central screw with a triangle thread shape and a pitch of 0.2 mm presented the best mechanical properties.

Purpose

This work focuses on automated epileptic seizure diagnosis (ESD) and prediction (ESP) to clarify the expanding role of machine learning (ML) in epileptic analysis. It outlines the current approaches and challenges in the diagnosis and prognosis of epilepsy and examines the convergence of magnetic resonance imaging (MRI), electroencephalogram (EEG), and ML.

Methods

This paper lists current methods for segmentation, localization, feature extraction, diagnosis, and prognosis after providing a brief medical review to distinguish between different forms of epilepsy. A particular focus is on using ML to EEG and MRI data, describing classification techniques to differentiate normal and epileptic activity.

Results

We highlight the potential of ML-driven methods for computer-aided epilepsy diagnosis and prognosis. We discuss achievements, challenges, and future directions, including devising novel techniques for automated alerts and seizure frequency estimation with minimal computational burden.

Conclusion

ML interfaces offer new possibilities for real-time seizure diagnosis in refractory epilepsy patients through wearables and implants. This discovery opens the door for improved diagnostic precision and individualized treatment plans in this field by using ML’s capabilities.

Graphical Abstract

The graphical abstract presents the machine Learning (ML) workflow for epileptic seizure diagnosis (ES) in detail. It begins with collecting data, such as magnetic resonance imaging (MRI) and electroencephalogram (EEG) data. Subsequently, features were extracted from the MRI and EEG data and used to train and evaluate machine learning models. The trained models were then applied to ES classification. Finally, ML algorithms proved to have the potential to revolutionize the diagnosis and treatment of epilepsy. By enabling early detection and personalized treatment, ML algorithms can help improve patient outcomes and quality of life.