Medical & Biological Engineering & Computing最新文献

Cardiac segmentation and quantification of cardiac function indicators play a crucial role in the clinical diagnosis and treatment of cardiovascular diseases. To address the issue of blurred cardiac chamber boundaries and adjacent tissue interference resulting from similar intensity in computed tomograph (CT) images, this paper proposes a 3D cardiac multi-structure segmentation network utilizing Multi-scale Channel Enhancement Attention (MCEA) and Spatial Decomposition with Channel Fusion Attention (SD-CA). The MCEA module integrates channel information from feature maps of various scales within the coding layer, thereby enhancing contextual linkage, strengthening the network's multi-scale feature representation capability, and improving decoding and segmentation performance. The SD-CA module generates spatial and channel attention weights in parallel and combines the three directional features of height, width, and depth. This enables the network to effectively concentrate on the region of interest and mitigate the interference of irrelevant structures. Experimental evaluations were conducted using a dataset of 192 cases provided by the People's Hospital of Liaoning Province and the MM-WHS dataset. Segmentation was achieved for the left ventricle, myocardium, left atrium, right ventricle, and right atrium, with average Dice coefficients of 94.21% and 93.9%, and average 95% Hausdorff distances of 6.5483 and 4.36, respectively. Furthermore, quantitative predictions of the left ventricular ejection fraction (LVEF) and substructure volumes were derived from the segmentation results. The correlation coefficients between the predicted and true values exceeded 0.9587, and all fell within the maximum error range of the Bland-Altman test for over 94.8% of the data, indicating a strong correlation and agreement between the predicted and true values.

Accurate muscle volume measurement is crucial for evaluating muscle impairment in healthcare and sports medicine. Compared to traditional methods, 3D ultrasound imaging offers noninvasive, flexible, cost-effectiveness advantages. This study aims to develop a precise volume assessment method for skeletal muscle, specifically gastrocnemius muscle, based on 3D ultrasound imaging. A feasible practice integrating 3D freehand ultrasound imaging based on optical tracking, slice extraction and alpha-shape-based surface reconstruction was proposed for precise volume assessment. 2D ultrasound images with spatial positions were acquired. Target slices were extracted for segmentation, and the alpha‑shape algorithm reconstructed the 3D muscle mesh for volume calculation. Phantom experiment using a pork tenderloin validated our method with a relative deviation of 0.47% compared to water displacement method. Clinical validation against MRI yielded relative deviations of 0.66% to 5.06% for manual segmentation and 0.28% to 2.58% for automated segmentation (using TransUNet). The method achieved smooth, detailed surfaces and outperformed Marching Cubes and Poisson reconstruction in accuracy and morphological fidelity. The proposed 3D freehand ultrasound workflow enables precise, detailed muscle volume assessment, showing strong agreement with MRI. Its accessibility and accuracy suggest significant potential for clinical and sports medicine applications in monitoring muscle health.

Coronary artery disease (CAD), one of the leading causes of mortality worldwide, necessitates effective risk assessment strategies, with coronary artery calcium (CAC) scoring via computed tomography (CT) being a key method for prevention. Traditional methods, primarily based on UNET architectures implemented on pre-built models, face challenges like the scarcity of annotated CT scans containing CAC and imbalanced datasets, leading to reduced performance in segmentation and scoring tasks. In this study, we address these limitations by introducing DINO-LG, a novel label-guided extension of DINO (self-distillation with no labels) that incorporates targeted augmentation on annotated calcified regions during self-supervised pre-training. Our three-stage pipeline integrates Vision Transformer (ViT-Base/8) feature extraction via DINO-LG trained on 914 CT scans comprising 700 gated and 214 non-gated acquisitions, linear classification to identify calcified slices, and U-NET segmentation for CAC quantification and Agatston scoring. DINO-LG achieved 89% sensitivity and 90% specificity for detecting CAC-containing CT slices, compared to standard DINO's 79% sensitivity and 77% specificity, reducing false-negative and false-positive rates by 49% and 57% respectively. The integrated system achieves 90% accuracy in CAC risk classification on 45 test patients, outperforming standalone U-NET segmentation (76% accuracy) while processing only the relevant subset of CT slices. This targeted approach enhances CAC scoring accuracy by feeding the UNET model with relevant slices, improving diagnostic precision while lowering healthcare costs by minimizing unnecessary tests and treatments.

Nowadays, the research of image fusion methods focuses on two-dimensional medical images, and almost no three-dimensional medical image fusion methods based on deep learning have been proposed. However, 3D image fusion is significant in clinical diagnosis. Therefore, this paper proposed a 3D medical image fusion enhancement network (3DFE-Net) for the gap in deep learning. 3DFE-Net included a feature extraction module, a multi-attention fusion module, and a feature reconstruction module. Firstly, multi-receptive field convolution blocks (MRFC) and multi-receptive field bottleneck blocks (MRFB) were devised instead of the traditional convolutional blocks to extract features of multiple receptive fields. Then, the multi-attention fusion module was designed using channel attention, self-attention, and spatial attention to make the network focus on the critical information in source images. Finally, the 3D fused image was obtained by the feature reconstruction module. In addition, a multivariate loss function was proposed for network training so that the fused image retains more edge structural information and texture details. MR-T1ce/MR-T2 fusion experiments show that, compared with the traditional method, 3DFE-Net improved the evaluation metrics EN (Information Entropy), MI (Mutual Information), SD (Standard Deviation), Qabf (Quality assessment of binary), and VIF (Visual Information Fidelity) by 0.0501, 0.1003, 5.2682, 0.1874, and 0.2129, respectively. 3DFE-Net can focus on the glioma lesion region in glioma slice fusion to achieve outstanding results and keep the structural information in MR-T1ce and the brightness information in MR-T2 well in normal slices. In qualitative and quantitative evaluations, 3DFE-Net performs better than conventional methods.

This paper introduces a novel densely connected neural network architecture designed for the pervasive visualisation of musculoskeletal system simulations. These simulations are built upon continuum-mechanical frameworks, which effectively integrate the diverse structural and physiological properties of the musculoskeletal system. A significant drawback of continuum-mechanical musculoskeletal models is their substantial computational resource requirement, making them difficult to transfer to/visualise the results on resource-poor systems like augmented reality or mobile devices. Such technologies, however, will be crucial for future advancements in human-machine interaction, surgical support tools, or physiotherapy. We use an activation-driven five-muscle continuum-mechanical upper limb model to obtain the activation-induced deformations of the respective muscles. Exemplified on the m. biceps brachii, we fit a sparse grid surrogate to capture the surface deformation and train a deep learning model that is subsequently used in our real-time visualisation. Based on the activation levels of the five muscles, the result of our trained neural network leads to an average positional error of 0.97±0.16 mm, or 0.57±0.10% for the 2809 mesh nodes of the m. biceps brachii's surface. With the novel deep neural network model, we achieved evaluation times for the m. biceps brachii's surface deformation of 9.88 ms on CPU-only architectures and 3.48 ms on architectures with GPU support. This leads to theoretical frame rates of 101 fps and 287 fps, respectively. The combination of surrogates and deep neural networks presented here succeeds as a proof-of-concept for real-time visualisation of a complex musculoskeletal system model, and does not rely on the inherent characteristics of the musculoskeletal system, and, hence, is also applicable to other real-time visualisations of complex meshed models in other applications.

In this paper, we extend the SonoNet architecture to capture spatio-temporal information from ultra-sound (US) sequences. More specifically, we propose 3D-SonoNet32 - which lifts 2D convolutions to 3D - and to an efficient (2+1)D variant - to keep the computational cost under control while preserving the benefits of the spatio-temporal model. We investigate the potential of these architectures on a scan-plane detection problem and discuss how these methodologies can be beneficial for AI-driven online "scan assistants", to enhance the quality and reproducibility of the evaluation and ultimately support the clinicians in the US examination. Our main contributions are (i) the design of novel Space-Time SonoNet architectures for analysing US video sequences, (ii) an in depth experimental analysis to show the benefit of using space-time models with respect to purely spatial ones, and to discuss the potential improvements gained by exploiting domain-specific properties like temporal coherence and prior knowledge of the ongoing scan. Overall, we show that the proposed models are specifically designed to be computationally lightweight, but also competitive in performance, making them suitable for real-time deployment on portable US devices.

Vaginal deliveries are frequently associated with perineal trauma, including severe tearing in some cases. Understanding of pelvic floor muscle damage and perineal tearing during childbirth is of great clinical relevance. However, the knowledge of these complex phenomena is incomplete. The objective of the present study is to explore the multifactorial view of pelvic floor muscle damage and perineal tearing during childbirth. Using nonlinear finite element modeling coupled to statistical surrogate modeling, we modeled fetal descent with imposed displacement and used active maternal for muscle contraction to estimate the pelvic floor muscle damage and perineal tearing indicators under different influencing factors such as fetal head deformability and biometry, as well as constitutive behaviors. The obtained results show that fetal head deformability reduces stress and strain concentrations in the pelvic floor muscles (PFM) and perineal region, while increasing fetal head size leads to heightened internal tissue responses. Linear regression analysis demonstrated strong model performance (R² = 0.782-0.981) and statistically predictive relationships between fetal biometric parameters, soft tissue constitutive behaviors, and associated mechanical responses. By integrating advanced finite element modeling with statistical modeling and regression, this work provides new quantitative insights into the biomechanical factors, highlighting tissue deformation patterns and indicating potential risk of tissue damage in highly strained areas due to localized mechanical stress. This approach offers a predictive and non-invasive strategy for assessing maternal tissue vulnerability during childbirth.