Search

IntroductionThis study is a retrospective evaluation of the performance of deep learning models that were developed for the detection of COVID-19 from chest x-rays, undertaken with the goal of assessing the suitability of such systems as clinical decision support tools.MethodsModels were trained on the National COVID-19 Chest Imaging Database (NCCID), a UK-wide multi-centre dataset from 26 different NHS hospitals and evaluated on independent multi-national clinical datasets. The evaluation considers clinical and technical contributors to model error and potential model bias. Model predictions are examined for spurious feature correlations using techniques for explainable prediction.ResultsModels performed adequately on NHS populations, with performance comparable to radiologists, but generalised poorly to international populations. Models performed better in males than females, and performance varied across age groups. Alarmingly, models routinely failed when applied to complex clinical cases with confounding pathologies and when applied to radiologist defined “mild” cases.DiscussionThis comprehensive benchmarking study examines the pitfalls in current practices that have led to impractical model development. Key findings highlight the need for clinician involvement at all stages of model development, from data curation and label definition, to model evaluation, to ensure that all clinical factors and disease features are appropriately considered during model design. This is imperative to ensure automated approaches developed for disease detection are fit-for-purpose in a clinical setting.

Image-to-text radiology report generation aims to automatically produce radiology reports that describe the findings in medical images. Most existing methods focus solely on the image data, disregarding the other patient information accessible to radiologists. In this paper, we present a novel multi-modal deep neural network framework for generating chest x-rays reports by integrating structured patient data, such as vital signs and symptoms, alongside unstructured clinical notes. We introduce a conditioned cross-multi-head attention module to fuse these heterogeneous data modalities, bridging the semantic gap between visual and textual data. Experiments demonstrate substantial improvements from using additional modalities compared to relying on images alone. Notably, our model achieves the highest reported performance on the ROUGE-L metric compared to relevant state-of-the-art models in the literature. Furthermore, we employed both human evaluation and clinical semantic similarity measurement alongside word-overlap metrics to improve the depth of quantitative analysis. A human evaluation, conducted by a board-certified radiologist, confirms the model’s accuracy in identifying high-level findings, however, it also highlights that more improvement is needed to capture nuanced details and clinical context.

Musculoskeletal models have been widely used for detailed biomechanical analysis to characterise various functional impairments given their ability to estimate movement variables (i.e., muscle forces and joint moments) which cannot be readily measured in vivo. Physics-based computational neuromusculoskeletal models can interpret the dynamic interaction between neural drive to muscles, muscle dynamics, body and joint kinematics and kinetics. Still, such set of solutions suffers from slowness, especially for the complex models, hindering the utility in real-time applications. In recent years, data-driven methods have emerged as a promising alternative due to the benefits in speedy and simple implementation, but they cannot reflect the underlying neuromechanical processes. This paper proposes a physics-informed deep learning framework for musculoskeletal modelling, where physics-based domain knowledge is brought into the data-driven model as soft constraints to penalise/regularise the data-driven model. We use the synchronous muscle forces and joint kinematics prediction from surface electromyogram (sEMG) as the exemplar to illustrate the proposed framework. Convolutional neural network (CNN) is employed as the deep neural network to implement the proposed framework. Simultaneously, the physics law between muscle forces and joint kinematics is used the soft constraint. Experimental validations on two groups of data, including one benchmark dataset and one self-collected dataset from six healthy subjects, are performed. The experimental results demonstrate the effectiveness and robustness of the proposed framework.

How prevalent is spontaneous thrombosis in a population containing all sizes of intracranial aneurysms? How can we calibrate computational models of thrombosis based on published data? How does spontaneous thrombosis differ in normo- and hypertensive subjects? We address the first question through a thorough analysis of published datasets that provide spontaneous thrombosis rates across different aneurysm characteristics. This analysis provides data for a subgroup of the general population of aneurysms, namely, those of large and giant size (>10 mm). Based on these observed spontaneous thrombosis rates, our computational modeling platform enables the first in silico observational study of spontaneous thrombosis prevalence across a broader set of aneurysm phenotypes. We generate 109 virtual patients and use a novel approach to calibrate two trigger thresholds: residence time and shear rate, thus addressing the second question. We then address the third question by utilizing this calibrated model to provide new insight into the effects of hypertension on spontaneous thrombosis. We demonstrate how a mechanistic thrombosis model calibrated on an intracranial aneurysm cohort can help estimate spontaneous thrombosis prevalence in a broader aneurysm population. This study is enabled through a fully automatic multi-scale modeling pipeline. We use the clinical spontaneous thrombosis data as an indirect population-level validation of a complex computational modeling framework. Furthermore, our framework allows exploration of the influence of hypertension in spontaneous thrombosis. This lays the foundation for in silico clinical trials of cerebrovascular devices in high-risk populations, e.g., assessing the performance of flow diverters in aneurysms for hypertensive patients.

In-silico trials rely on virtual populations and interventions simulated using patient-specific models and may offer a solution to lower costs. Here, the authors present the flow diverter performance assessment in-silico trial, which models the treatment of intracranial aneurysms with a flow-diverting stent.

Abstract Invasive coronary angiography remains the gold standard for diagnosing coronary artery disease, which may be complicated by both, patient-specific anatomy and image quality. Deep learning techniques aimed at detecting coronary artery stenoses may facilitate the diagnosis. However, previous studies have failed to achieve superior accuracy and performance for real-time labeling. Our study is aimed at confirming the feasibility of real-time coronary artery stenosis detection using deep learning methods. To reach this goal we trained and tested eight promising detectors based on different neural network architectures (MobileNet, ResNet-50, ResNet-101, Inception ResNet, NASNet) using clinical angiography data of 100 patients. Three neural networks have demonstrated superior results. The network based on Faster-RCNN Inception ResNet V2 is the most accurate and it achieved the mean Average Precision of 0.95, F1-score 0.96 and the slowest prediction rate of 3 fps on the validation subset. The relatively lightweight SSD MobileNet V2 network proved itself as the fastest one with a low mAP of 0.83, F1-score of 0.80 and a mean prediction rate of 38 fps. The model based on RFCN ResNet-101 V2 has demonstrated an optimal accuracy-to-speed ratio. Its mAP makes up 0.94, F1-score 0.96 while the prediction speed is 10 fps. The resultant performance-accuracy balance of the modern neural networks has confirmed the feasibility of real-time coronary artery stenosis detection supporting the decision-making process of the Heart Team interpreting coronary angiography findings.

Abstract By 2040, ~100 million people will have glaucoma. To date, there are a lack of high-efficiency glaucoma diagnostic tools based on visual fields (VFs). Herein, we develop and evaluate the performance of ‘iGlaucoma’, a smartphone application-based deep learning system (DLS) in detecting glaucomatous VF changes. A total of 1,614,808 data points of 10,784 VFs (5542 patients) from seven centers in China were included in this study, divided over two phases. In Phase I, 1,581,060 data points from 10,135 VFs of 5105 patients were included to train (8424 VFs), validate (598 VFs) and test (3 independent test sets—200, 406, 507 samples) the diagnostic performance of the DLS. In Phase II, using the same DLS, iGlaucoma cloud-based application further tested on 33,748 data points from 649 VFs of 437 patients from three glaucoma clinics. With reference to three experienced expert glaucomatologists, the diagnostic performance (area under curve [AUC], sensitivity and specificity) of the DLS and six ophthalmologists were evaluated in detecting glaucoma. In Phase I, the DLS outperformed all six ophthalmologists in the three test sets (AUC of 0.834–0.877, with a sensitivity of 0.831–0.922 and a specificity of 0.676–0.709). In Phase II, iGlaucoma had 0.99 accuracy in recognizing different patterns in pattern deviation probability plots region, with corresponding AUC, sensitivity and specificity of 0.966 (0.953–0.979), 0.954 (0.930–0.977), and 0.873 (0.838–0.908), respectively. The ‘iGlaucoma’ is a clinically effective glaucoma diagnostic tool to detect glaucoma from humphrey VFs, although the target population will need to be carefully identified with glaucoma expertise input.

Between 2014 and 2023, 100,000 UK Biobank participants are undergoing brain, heart and abdominal MRI, as well as DXA and carotid ultrasound scans. In this review, authors provide a detailed overview of the rationale for the collection of these imaging data, the procedures of data collection and management, and the future directions of the UK biobank imaging enhancement.

Accurate segmentation of pediatric echocardiography images is essential for a wide range of diagnostic and pre-interventional planning, but remains challenging (e.g., low signal to noise ratio and internal variability in heart appearance). To address these problems, in this paper, we propose a novel Cardiac Attention-guided Dual-path Network (i.e., AIDAN). AIDAN comprises a convolutional block attention module (CBAM) attached to a spatial (i.e., SPA) and context paths (i.e., CPA), which can guide the network and learn the most discriminative features. The spatial path captures low-level spatial features, and the context path is designed to exploit high-level context. Finally, features learned from the two paths are fused efficiently using a specially designed feature fusion module (FFM), and these are used to predict the final segmentation map. We experiment on a self-collected dataset of 127 pediatric echocardiography cases which are videos containing at least a complete cardiac cycle, and obtain a Dice coefficient of 0.951 and 0.914, in the left ventricle and atrium segments, respectively. AIDAN outperforms other state-of-the-art methods and has great potential for pediatric echocardiography images analysis.

Built on the foundation of the randomized controlled trial (RCT), Evidence Based Medicine (EBM) is at its best when optimizing outcomes for homogeneous cohorts of patients like those participating in an RCT. Its weakness is a failure to resolve a clinical quandary: patients appear for care individually, each may differ in important ways from an RCT cohort, and the physician will wonder each time if following EBM will provide best guidance for this unique patient. In an effort to overcome this weakness, and promote higher quality care through a more personalized approach, a new framework has been proposed: Medicine-Based Evidence (MBE). In this approach, big data and deep learning techniques are embraced to interrogate treatment responses among patients in real-world clinical practice. Such statistical models are then integrated with mechanistic disease models to construct a “digital twin,” which serves as the real-time digital counterpart of a patient. MBE is thereby capable of dynamically modeling the effects of various treatment decisions in the context of an individual's specific characteristics. In this article, we discuss how MBE could benefit patients with congenital heart disease, a field where RCTs are difficult to conduct and often fail to provide definitive solutions because of a small number of subjects, their clinical complexity, and heterogeneity. We will also highlight the challenges that must be addressed before MBE can be embraced in clinical practice and its full potential can be realized.

A common feature of morphogenesis is the formation of three-dimensional structures from the folding of two-dimensional epithelial sheets, aided by cell shape changes at the cellular-level. Changes in cell shape must be studied in the context of cell-polarised biomechanical processes within the epithelial sheet. In epithelia with highly curved surfaces, finding single-cell alignment along a biological axis can be difficult to automate in silico. We present ‘Origami’, a MATLAB-based image analysis pipeline to compute direction-variant cell shape features along the epithelial apico-basal axis. Our automated method accurately computed direction vectors denoting the apico-basal axis in regions with opposing curvature in synthetic epithelia and fluorescence images of zebrafish embryos. As proof of concept, we identified different cell shape signatures in the developing zebrafish inner ear, where the epithelium deforms in opposite orientations to form different structures. Origami is designed to be user-friendly and is generally applicable to fluorescence images of curved epithelia. Author summary During embryonic development, two-dimensional epithelial sheets bend and fold into complex three-dimensional structures–like paper in the origami art form. The genetic and biomechanical processes driving epithelial folding can be polarised in the epithelium, leading to asymmetric shape changes at the single cell level. Defects in such epithelial shaping have been linked to many developmental anomalies and diseases. It is, therefore, important not only to quantify shape change at the single cell level, but also to orientate these asymmetrical changes along an epithelial axis of polarity when studying morphogenetic processes. Origami is a MATLAB-based software that has been developed to automatically extract such single-cell asymmetrical shape features along the epithelial apico-basal axis from fluorescence microscopy images of folding epithelia. Origami provides a solution to computing directional vectors along the epithelial apico-basal axis followed by extracting direction-variant shape features of each segmented cell. It is generally applicable to epithelial structures regardless of complexity or direction of folding and is robust to imaging conditions. As proof of concept, Origami successfully differentiated between different cell shape signatures in highly curved structures at different developmental timepoints in the zebrafish inner ear.

Biomedical image analysis challenges have increased in the last ten years, but common practices have not been established yet. Here the authors analyze 150 recent challenges and demonstrate that outcome varies based on the metrics used and that limited information reporting hampers reproducibility.

Background: Approximately 30% of patients with diffuse large B-cell lymphoma (DLBCL) will have recurrence. The aim of this study was to develop a radiomic based model derived from baseline PET/CT to predict 2-year event free survival (2-EFS). Methods: Patients with DLBCL treated with R-CHOP chemotherapy undergoing pre-treatment PET/CT between January 2008 and January 2018 were included. The dataset was split into training and internal unseen test sets (ratio 80:20). A logistic regression model using metabolic tumour volume (MTV) and six different machine learning classifiers created from clinical and radiomic features derived from the baseline PET/CT were trained and tuned using four-fold cross validation. The model with the highest mean validation receiver operator characteristic (ROC) curve area under the curve (AUC) was tested on the unseen test set. Results: 229 DLBCL patients met the inclusion criteria with 62 (27%) having 2-EFS events. The training cohort had 183 patients with 46 patients in the unseen test cohort. The model with the highest mean validation AUC combined clinical and radiomic features in a ridge regression model with a mean validation AUC of 0.75 ± 0.06 and a test AUC of 0.73. Conclusions: Radiomics based models demonstrate promise in predicting outcomes in DLBCL patients.

Microbubbles are used as contrast agents in clinical ultrasound for Left Ventricular Opacification (LVO) and perfusion imaging. They are also the subject of promising research in therapeutics as a drug delivery mechanism or for sonoporation and co-administration. For maximum efficacy in these applications, it is important to understand the acoustic characteristics of the administered microbubbles. Despite this, there is significant variation in the experimental procedures and equipment used to measure the acoustic properties of microbubble populations. A chamber was designed to facilitate acoustic characterisation experiments and was manufactured using additive manufacturing techniques. The design has been released to allow wider uptake in the research community. The efficacy of the chamber for acoustic characterisation has been explored with an experiment to measure the scattering of SonoVue® microbubbles at the fundamental frequency and second harmonic under interrogation from emissions in the frequency range of 1.6 to 6.4 MHz. The highest overall scattering values were measured at 1.6 MHz and decreased as the frequency increased, a result which is in agreement with previously published measurements. Statistical analysis of the acoustic scattering measurements have been performed and a significant difference, at the 5% significance level, was found between the samples containing contrast agent and the control sample containing only deionised water. These findings validate the proposed design for measuring the acoustic scattering characteristics of ultrasound contrast agents.

This immunohistochemistry dataset contains the main structures in deep subcortical white matter (axons, astrocytes, and myelinated axons) in a representative cohort of an ageing population. A set of samples from 90 subjects of the Cognitive Function and Ageing Study (CFAS) were analysed, stratified into three groups of 30 subjects each, in relation to the presence of age-associated deep subcortical lesions. High-resolution microscopy data enables the extraction of valuable information, such as volume fractions, for the construction and validation of diffusion MRI (dMRI) models. The dataset provided here was used in Coelho et al. [1].

White matter lesions represent a major risk factor for dementia in elderly people. Magnetic Resonance Imaging (MRI) studies have demonstrated cerebral blood flow reduction in age-related white matter lesions, indicating that vascular alterations are involved in developing white matter lesions. Hypoperfusion and changes in capillary morphology are generally linked to dementia. However, a quantitative study describing these microvascular alterations in white matter lesions is missing in the literature; most previous microvascular studies being on the cortex. The aim of this work is to identify and quantify capillary microstructural changes involved in the appearance of deep subcortical lesions (DSCL). We characterize the distribution of capillary diameter, thickness, and density in the deep white matter in a population of 75 elderly subjects, stratified into three equal groups according to DSCL: Control (subject without DSCL), Lesion (sample presenting DSCL), and Normal Appearing White Matter (NAWM, the subject presented DSCL but not at the sampled tissue location). Tissue samples were selected from the Cognitive Function and Aging Study (CFAS), a cohort representative of an aging population, from which immunohistochemically-labeled histological images were acquired. To accurately estimate capillary diameters and thicknesses from the 2D histological images, we also introduce a novel semi-automatic method robust to non-perpendicular incidence angle of capillaries into the imaging plane, and to non-circular deformations of capillary cross sections. Subjects with DSCL presented a significant increase in capillary wall thickness, a decrease in the diameter intra-subject variability (but not in the mean), and a decrease in capillary density. No significant difference was observed between controls and NAWM. Both capillary wall thickening and reduction in capillary density contribute to the reduction of cerebral blood flow previously reported for white matter lesions. The obtained distributions provide reliable statistics of capillary microstructure useful to inform the modeling of human cerebral blood flow, for instance to define microcirculation models for their estimation from MRI or to perform realistic cerebral blood flow simulations. Keywords: Vascular dementia, Deep white matter, Deep subcortical lesion, Capillary microstructure

Late-onset dementia is a major health concern in the ageing population. Alzheimer's disease (AD) accounts for the largest proportion (65–70%) of dementia cases in the older population.Despite considerable research effort, the pathogenesis of late-onset AD remains unclear. Substantial evidence suggests that the neurodegenerative process is initiated by chronic cerebral hypoperfusion (CCH) caused by ageing and cardiovascular conditions. CCH causes reduced oxygen, glucose and other nutrient supply to the brain, with direct damage not only to the parenchymal cells, but also to the blood–brain barrier (BBB), a key mediator of cerebral homeostasis. BBB dysfunction mediates the indirect neurotoxic effects of CCH by promoting oxidative stress, inflammation, paracellular permeability, and dysregulation of nitric oxide, a key regulator of regional blood flow. As such, BBB dysfunction mediates a vicious circle in which cerebral perfusion is reduced further and the neurodegenerative process is accelerated. Endothelial interaction with pericytes and astrocytes could also play a role in the process. Reciprocal interactions between vascular dysfunction and neurodegeneration could further contribute to the development of the disease.A comprehensive overview of the complex scenario of interacting endothelium-mediated processes is currently lacking, and could prospectively contribute to the identification of adequate therapeutic interventions.This study reviews the current literature of in vitro and ex vivo studies on endothelium-mediated mechanisms underlying vascular dysfunction in AD pathogenesis, with the aim of presenting a comprehensive overview of the complex network of causative relationships. Particular emphasis is given to vicious circles which can accelerate the process of neurovascular degeneration.

Thrombosis is a condition closely related to cerebral aneurysms and controlled thrombosis is the main purpose of endovascular embolization treatment. The mechanisms governing thrombus initiation and evolution in cerebral aneurysms have not been fully elucidated and this presents challenges for interventional planning. Significant effort has been directed towards developing computational methods aimed at streamlining the interventional planning process for unruptured cerebral aneurysm treatment. Included in these methods are computational models of thrombus development following endovascular device placement. The main challenge with developing computational models for thrombosis in disease cases is that there exists a wide body of literature that addresses various aspects of the clotting process, but it may not be obvious what information is of direct consequence for what modeling purpose (e.g., for understanding the effect of endovascular therapies). The aim of this review is to present the information so it will be of benefit to the community attempting to model cerebral aneurysm thrombosis for interventional planning purposes, in a simplified yet appropriate manner. The paper begins by explaining current understanding of physiological coagulation and highlights the documented distinctions between the physiological process and cerebral aneurysm thrombosis. Clinical observations of thrombosis following endovascular device placement are then presented. This is followed by a section detailing the demands placed on computational models developed for interventional planning. Finally, existing computational models of thrombosis are presented. This last section begins with description and discussion of physiological computational clotting models, as they are of immense value in understanding how to construct a general computational model of clotting. This is then followed by a review of computational models of clotting in cerebral aneurysms, specifically. Even though some progress has been made towards computational predictions of thrombosis following device placement in cerebral aneurysms, many gaps still remain. Answering the key questions will require the combined efforts of the clinical, experimental and computational communities.

In the original version of this Article the values in the rightmost column of Table 1 were inadvertently shifted relative to the other columns. This has now been corrected in the PDF and HTML versions of the Article.

Merging multimodal information about myocardial scar tissue can help electrophysiologists to find the most appropriate target during catheter ablation of ventricular arrhythmias. A framework is presented to analyze and combine information from delayed enhancement magnetic resonance imaging (DE-MRI) and electro-anatomical mapping data. Using this information, electrical, mechanical, and image-based characterization of the myocardium are performed. The presented framework allows the left ventricle to be segmented by DE-MRI and the scar to be characterized prior to the intervention based on image information. It allows the electro-anatomical maps obtained during the intervention from a navigation system to be merged together with the anatomy and scar information extracted from DE-MRI. It also allows for the estimation of endocardial motion and deformation to assess cardiac mechanics. Therefore, electrical, mechanical, and image-based characterization of the myocardium can be performed. The feasibility of this approach was demonstrated on three patients with ventricular tachycardia associated to ischemic cardiomyopathy by integrating images from DE-MRI and electro-anatomical maps data in a common framework for intraoperative myocardial tissue characterization. The proposed framework has the potential to guide and monitor delivery of radio frequency ablation of ventricular tachycardia. It is also helpful for research purposes, facilitating the study of the relationship between electrical and mechanical properties of the tissue, as well as with tissue viability from DE-MRI.