Search

Abstract Background Type 2 diabetes mellitus (T2DM) is a major risk factor for heart failure with preserved ejection fraction and cardiac arrhythmias. Precursors of these complications, such as diabetic cardiomyopathy, remain incompletely understood and underdiagnosed. Detection of early signs of cardiac deterioration in T2DM patients is critical for prevention. Our goal is to quantify T2DM-driven abnormalities in ECG and cardiac imaging biomarkers leading to cardiovascular disease. Methods We quantified ECG and cardiac magnetic resonance imaging biomarkers in two matched cohorts of 1781 UK Biobank participants, with and without T2DM, and no diagnosed cardiovascular disease at the time of assessment. We performed a pair-matched cross-sectional study to compare cardiac biomarkers in both cohorts, and examined the association between T2DM and these biomarkers. We built multivariate multiple linear regression models sequentially adjusted for socio-demographic, lifestyle, and clinical covariates. Results Participants with T2DM had a higher resting heart rate (66 vs. 61 beats per minute, p

Electronic skin (e-skin) systems are devices that mimic the different sensing modalities of skin. While the modalities sensed can vary from temperature to tactile, the main inspiration for much of the research being conducted is mimicking the sensing modalities of human skin. Much research has been conducted on tactile sensing through e-skin, as interest grows in the use of e-skin with smart prosthetics and Virtual and Augmented Reality (VR/AR) applications. Being able to mimic the sophisticated ability for human hands to complete complex tasks of dexterity, using an e-skin as essential input system, is the goal of many research groups. In this systematic review, we provide a full overview on e-skin systems that focus on developing tactile sensing. The objective is to assess how accurately these systems mimic human skin tactile sensing modalities and how accurately the detected stimuli are sensed. The outcomes of the review show where the focus of the community is with regards to which modalities are being developed. Furthermore, information extracted from the papers that detail their quantitative accuracy in sensing these modalities is provided. With a total of 205 systems included insights on trends and a quantitative comparison of current e-skin systems are discussed. The limitations of the current methods applied and how they could be overcome are explored, in addition to highlighting the need for standardized experimental protocols to ensure e-skin systems can be assessed and compared more easily.

Abstract Aim Acute respiratory distress syndrome or ARDS is an acute, severe form of respiratory failure characterised by poor oxygenation and bilateral pulmonary infiltrates. Advancements in signal processing and machine learning have led to promising solutions for classification, event detection and predictive models in the management of ARDS. Method In this review, we provide systematic description of different studies in the application of Machine Learning (ML) and artificial intelligence for management, prediction, and classification of ARDS. We searched the following databases: Google Scholar, PubMed, and EBSCO from 2009 to 2023. A total of 243 studies was screened, in which, 52 studies were included for review and analysis. We integrated knowledge of previous work providing the state of art and overview of explainable decision models in machine learning and have identified areas for future research. Results Gradient boosting is the most common and successful method utilised in 12 (23.1%) of the studies. Due to limitation of data size available, neural network and its variation is used by only 8 (15.4%) studies. Whilst all studies used cross validating technique or separated database for validation, only 1 study validated the model with clinician input. Explainability methods were presented in 15 (28.8%) of studies with the most common method is feature importance which used 14 times. Conclusion For databases of 5000 or fewer samples, extreme gradient boosting has the highest probability of success. A large, multi-region, multi centre database is required to reduce bias and take advantage of neural network method. A framework for validating with and explaining ML model to clinicians involved in the management of ARDS would be very helpful for development and deployment of the ML model.

Abstract This study aims at identifying risk-related patterns of left ventricular contraction dynamics via novel volume transient characterization. A multicenter cohort of AMI survivors (n = 1021) who underwent Cardiac Magnetic Resonance (CMR) after infarction was considered for the study. The clinical endpoint was the 12-month rate of major adverse cardiac events (MACE, n = 73), consisting of all-cause death, reinfarction, and new congestive heart failure. Cardiac function was characterized from CMR in 3 potential directions: by (1) volume temporal transients (i.e. contraction dynamics); (2) feature tracking strain analysis (i.e. bulk tissue peak contraction); and (3) 3D shape analysis (i.e. 3D contraction morphology). A fully automated pipeline was developed to extract conventional and novel artificial-intelligence-derived metrics of cardiac contraction, and their relationship with MACE was investigated. Any of the 3 proposed directions demonstrated its additional prognostic value on top of established CMR indexes, myocardial injury markers, basic characteristics, and cardiovascular risk factors (P

Cardiovascular diseases (CVDs) are the most common health threats worldwide. 2D X-ray invasive coronary angiography (ICA) remains the most widely adopted imaging modality for CVD assessment during real-time cardiac interventions. However, it is often difficult for the cardiologists to interpret the 3D geometry of coronary vessels based on 2D planes. Moreover, due to the radiation limit, often only two angiographic projections are acquired, providing limited information of the vessel geometry and necessitating 3D coronary tree reconstruction based only on two ICA projections. In this paper, we propose a self-supervised deep learning method called NeCA, which is based on neural implicit representation using the multiresolution hash encoder and differentiable cone-beam forward projector layer, in order to achieve 3D coronary artery tree reconstruction from two 2D projections. We validate our method using six different metrics on a dataset generated from coronary computed tomography angiography of right coronary artery and left anterior descending artery. The evaluation results demonstrate that our NeCA method, without requiring 3D ground truth for supervision or large datasets for training, achieves promising performance in both vessel topology and branch-connectivity preservation compared to the supervised deep learning model.

Abstract Objectives Positron emission tomography (PET) imaging is a costly tracer-based imaging modality used to visualise abnormal metabolic activity for the management of malignancies. The objective of this study is to demonstrate that non-contrast CTs alone can be used to differentiate regions with different Fluorodeoxyglucose (FDG) uptake and simulate PET images to guide clinical management. Methods Paired FDG-PET and CT images (n = 298 patients) with diagnosed head and neck squamous cell carcinoma (HNSCC) were obtained from The cancer imaging archive. Random forest (RF) classification of CT-derived radiomic features was used to differentiate metabolically active (tumour) and inactive tissues (ex. thyroid tissue). Subsequently, a deep learning generative adversarial network (GAN) was trained for this CT to PET transformation task without tracer injection. The simulated PET images were evaluated for technical accuracy (PERCIST v.1 criteria) and their ability to predict clinical outcome [(1) locoregional recurrence, (2) distant metastasis and (3) patient survival]. Results From 298 patients, 683 hot spots of elevated FDG uptake (elevated SUV, 6.03 ± 1.71) were identified. RF models of intensity-based CT-derived radiomic features were able to differentiate regions of negligible, low and elevated FDG uptake within and surrounding the tumour. Using the GAN-simulated PET image alone, we were able to predict clinical outcome to the same accuracy as that achieved using FDG-PET images. Conclusion This pipeline demonstrates a deep learning methodology to simulate PET images from CT images in HNSCC without the use of radioactive tracer. The same pipeline can be applied to other pathologies that require PET imaging.

Cardiac anatomy and function vary considerably across the human population with important implications for clinical diagnosis and treatment planning. Consequently, many computer-based approaches have been developed to capture this variability for a wide range of applications, including explainable cardiac disease detection and prediction, dimensionality reduction, cardiac shape analysis, and the generation of virtual heart populations. In this work, we propose a variational mesh autoencoder (mesh VAE) as a novel geometric deep learning approach to model such population-wide variations in cardiac shapes. It embeds multi-scale graph convolutions and mesh pooling layers in a hierarchical VAE framework to enable direct processing of surface mesh representations of the cardiac anatomy in an efficient manner. The proposed mesh VAE achieves low reconstruction errors on a dataset of 3D cardiac meshes from over 1,000 patients with acute myocardial infarction, with mean surface distances between input and reconstructed meshes below the underlying image resolution. We also find that it outperforms a voxelgrid-based deep learning benchmark in terms of both mean surface distance and Hausdorff distance while requiring considerably less memory. Furthermore, we explore the quality and interpretability of the mesh VAE's latent space and showcase its ability to improve the prediction of major adverse cardiac events over a clinical benchmark. Finally, we investigate the method's ability to generate realistic virtual populations of cardiac anatomies and find good alignment between the synthesized and gold standard mesh populations in terms of multiple clinical metrics.

Human cardiac function is characterized by a complex interplay of mechanical deformation and electrophysiological conduction. Similar to the underlying cardiac anatomy, these interconnected physiological patterns vary considerably across the human population with important implications for the effectiveness of clinical decision-making and the accuracy of computerized heart models. While many previous works have investigated this variability separately for either cardiac anatomy or physiology, this work aims to combine both aspects in a single data-driven approach and capture their intricate interdependencies in a multi-domain setting. To this end, we propose a novel multi-domain Variational Autoencoder (VAE) network to capture combined Electrocardiogram (ECG) and Magnetic Resonance Imaging (MRI)-based 3D anatomy information in a single model. Each VAE branch is specifically designed to address the particular challenges of the respective input domain, enabling efficient encoding, reconstruction, and synthesis of multi-domain cardiac signals. Our method achieves high reconstruction accuracy on a United Kingdom Biobank dataset, with Chamfer Distances between reconstructed and input anatomies below the underlying image resolution and ECG reconstructions outperforming multiple single-domain benchmarks by a considerable margin. The proposed VAE is capable of generating realistic virtual populations of arbitrary size with good alignment in clinical metrics between the synthesized and gold standard anatomies and Maximum Mean Discrepancy (MMD) scores of generated ECGs below those of comparable single-domain approaches. Furthermore, we observe the latent space of our VAE to be highly interpretable with separate components encoding different aspects of anatomical and ECG variability. Finally, we demonstrate that the combined anatomy and ECG representation improves the performance in a cardiac disease classification task by 3.9% in terms of Area Under the Receiver Operating Characteristic (AUROC) curve over the best corresponding single-domain modeling approach.

AimsPatient-to-patient anatomical differences are an important source of variability in the electrocardiogram, and they may compromise the identification of pathological electrophysiological abnormalities. This study aims at quantifying the contribution of variability in ventricular and torso anatomies to differences in QRS complexes of the 12-lead ECG using computer simulations.MethodsA computational pipeline is presented that enables computer simulations using human torso/biventricular anatomically based electrophysiological models from clinically standard magnetic resonance imaging (MRI). The ventricular model includes membrane kinetics represented by the biophysically detailed O’Hara Rudy model modified for tissue heterogeneity and includes fiber orientation based on the Streeter rule. A population of 265 torso/biventricular models was generated by combining ventricular and torso anatomies obtained from clinically standard MRIs, augmented with a statistical shape model of the body. 12-lead ECGs were simulated on the 265 human torso/biventricular electrophysiology models, and QRS morphology, duration and amplitude were quantified in each ECG lead for each of the human torso-biventricular models.ResultsQRS morphologies in limb leads are mainly determined by ventricular anatomy, while in the precordial leads, and especially V1 to V4, they are determined by heart position within the torso. Differences in ventricular orientation within the torso can explain morphological variability from monophasic to biphasic QRS complexes. QRS duration is mainly influenced by myocardial volume, while it is hardly affected by the torso anatomy or position. An average increase of 0.12 ± 0.05 ms in QRS duration is obtained for each cm3 of myocardial volume across all the leads while it hardly changed due to changes in torso volume.ConclusionComputer simulations using populations of human torso/biventricular models based on clinical MRI enable quantification of anatomical causes of variability in the QRS complex of the 12-lead ECG. The human models presented also pave the way toward their use as testbeds in silico clinical trials.

IntroductionAortic distensibility can be calculated using semi-automated methods to segment the aortic lumen on cine CMR (Cardiovascular Magnetic Resonance) images. However, these methods require visual quality control and manual localization of the region of interest (ROI) of ascending (AA) and proximal descending (PDA) aorta, which limit the analysis in large-scale population-based studies. Using 5100 scans from UK Biobank, this study sought to develop and validate a fully automated method to 1) detect and locate the ROIs of AA and PDA, and 2) provide a quality control mechanism.MethodsThe automated AA and PDA detection-localization algorithm followed these steps: 1) foreground segmentation; 2) detection of candidate ROIs by Circular Hough Transform (CHT); 3) spatial, histogram and shape feature extraction for candidate ROIs; 4) AA and PDA detection using Random Forest (RF); 5) quality control based on RF detection probability. To provide the ground truth, overall image quality (IQ = 0-3 from poor to good) and aortic locations were visually assessed by 13 observers. The automated algorithm was trained on 1200 scans and Dice Similarity Coefficient (DSC) was used to calculate the agreement between ground truth and automatically detected ROIs.ResultsThe automated algorithm was tested on 3900 scans. Detection accuracy was 99.4% for AA and 99.8% for PDA. Aorta localization showed excellent agreement with the ground truth, with DSC ≥ 0.9 in 94.8% of AA (DSC = 0.97 ± 0.04) and 99.5% of PDA cases (DSC = 0.98 ± 0.03). AA×PDA detection probabilities could discriminate scans with IQ ≥ 1 from those severely corrupted by artefacts (AUC = 90.6%). If scans with detection probability < 0.75 were excluded (350 scans), the algorithm was able to correctly detect and localize AA and PDA in all the remaining 3550 scans (100% accuracy).ConclusionThe proposed method for automated AA and PDA localization was extremely accurate and the automatically derived detection probabilities provided a robust mechanism to detect low quality scans for further human review. Applying the proposed localization and quality control techniques promises at least a ten-fold reduction in human involvement without sacrificing any accuracy.

Assessment of coronary stenosis severity is crucial in clinical practice. This study proposes a novel method to generate 3D models of stenotic coronary arteries, directly from 2D coronary images, and suitable for immediate assessment of the stenosis severity.From multiple 2D X-ray coronary arteriogram projections, 2D vessels were extracted. A 3D centreline was reconstructed as intersection of surfaces from corresponding branches. Next, 3D luminal contours were generated in a two-step process: first, a Non-Uniform Rational B-Spline (NURBS) circular contour was designed and, second, its control points were adjusted to interpolate computed 3D boundary points. Finally, a 3D surface was generated as an interpolation across the control points of the contours and used in the analysis of the severity of a lesion. To evaluate the method, we compared 3D reconstructed lesions with Optical Coherence Tomography (OCT), an invasive imaging modality that enables high-resolution endoluminal visualization of lesion anatomy.Validation was performed on routine clinical data. Analysis of paired cross-sectional area discrepancies indicated that the proposed method more closely represented OCT contours than conventional approaches in luminal surface reconstruction, with overall root-mean-square errors ranging from 0.213mm2 to 1.013mm2, and maximum error of 1.837mm2. Comparison of volume reduction due to a lesion with corresponding FFR measurement suggests that the method may help in estimating the physiological significance of a lesion.The algorithm accurately reconstructed 3D models of lesioned arteries and enabled quantitative assessment of stenoses. The proposed method has the potential to allow immediate analysis of the stenoses in clinical practice, thereby providing incremental diagnostic and prognostic information to guide treatments in real time and without the need for invasive techniques.

We introduce a tool to build a surface mesh able to deal with sparse, heterogeneous, non-parallel, cross-sectional, non-coincidental contours and show its application to reconstruct surfaces of the heart. In recent years, much research has looked at creating personalised 3D anatomical models of the heart. These models usually incorporate a geometrical reconstruction of the anatomy in order to better understand cardiovascular functions as well as predict different cardiac processes. As MRIs are becoming the standard for cardiac medical imaging, we tested our methodology on cardiac MRI data from standard acquisitions. However, the ability to accurately reconstruct heart anatomy in three dimensions commonly comes with fundamental challenges—notably, the trade-off between data fitting and expected visual appearance. Most current techniques can either require contours from parallel slices or, if multiple slice orientations are used, require an exact match between these contours. In addition, some methods introduce a bias by the use of prior shape models or by trade-offs between the data matching terms and the smoothing terms. Our approach uses a composition of smooth approximations towards the maximization of the data fitting, ensuring a good matching to the input data as well as pleasant interpolation characteristics. To assess our method in the task of cardiac mesh generations, we evaluated its performance on synthetic data obtained from a cardiac statistical shape model as well as on real data. Using a statistical shape model, we simulated standard cardiac MRI acquisitions planes and contour data. We performed a multi-parameter evaluation study using plausible cardiac shapes generated from the model. We also show that long axes contours as well as the most extremal slices (basal and apical) contain the most amount of structural information, and thus should be taken into account when generating anatomically relevant geometrical cardiovascular surfaces. Our method is both used on epicardial and endocardial left ventricle surfaces as well as on the right ventricle.

Tissue microstructure, in particular the alignment of myocytes (fibre direction) and their lateral organisation into sheets, is fundamental to cardiac function. We studied the effect of microstructure on contraction in a computational model of rat left ventricular electromechanics. Different fibre models, globally rule-based or locally optimised to DT-MRI data, were compared, in order to understand whether a subject-specific fibre model would enhance the predictive power of our model with respect to the global ones. We also studied the impact of sheets on ventricular deformation by comparing: (a) a transversely isotropic versus an orthotropic material law and (b) a linear model with a bimodal model of sheet transmural variation. We estimated ejection fraction, wall thickening and base-to-apex shortening and compared them with measures from cine-MRI. We also evaluated Lagrangian strains as local metrics of cardiac deformation. Our results show that the subject-specific fibre model provides little improvement in the metric predictions with respect to global fibre models while material orthotropy allows closer agreement with measures than transverse isotropy. Nonetheless, the impact of sheets in our model is smaller than that of fibres. We conclude that further investigation of the modelling of sheet dynamics is necessary to fully understand the impact of tissue structure on cardiac deformation.

Images are ubiquitous in biomedical applications from basic research to clinical practice. With the rapid increase in resolution, dimensionality of the images and the need for real-time performance in many applications, computational requirements demand proper exploitation of multicore architectures. Towards this, GPU-specific implementations of image analysis algorithms are particularly promising. In this paper, we investigate the mapping of an enhanced motion estimation algorithm to novel GPU-specific architectures, the resulting challenges and benefits therein. Using a database of three-dimensional image sequences, we show that the mapping leads to substantial performance gains, up to a factor of 60, and can provide near-real-time experience. We also show how architectural peculiarities of these devices can be best exploited in the benefit of algorithms, most specifically for addressing the challenges related to their access patterns and different memory configurations. Finally, we evaluate the performance of the algorithm on three different GPU architectures and perform a comprehensive analysis of the results.