Search

Additional file 3. Supplementary Materials C.

Computational fluid dynamics (CFD) can be used to simulate vascular haemodynamics and analyse potential treatment options. CFD has shown to be beneficial in improving patient outcomes. However, the implementation of CFD for routine clinical use is yet to be realised. Barriers for CFD include high computational resources, specialist experience needed for designing simulation set-ups, and long processing times. The aim of this study was to explore the use of machine learning (ML) to replicate conventional aortic CFD with automatic and fast regression models. Data used to train/test the model consisted of 3,000 CFD simulations performed on synthetically generated 3D aortic shapes. These subjects were generated from a statistical shape model (SSM) built on real patient-specific aortas (N=67). Inference performed on 200 test shapes resulted in average errors of 6.01% +/-3.12 SD and 3.99% +/-0.93 SD for pressure and velocity, respectively. Our ML-based models performed CFD in +/-0.075 seconds (4,000x faster than the solver). This proof-of-concept study shows that results from conventional vascular CFD can be reproduced using ML at a much faster rate, in an automatic process, and with high accuracy., Comment: 22 pages, 19 figures

Additional file 1. Supplementary Materials A.

Additional file 2. Supplementary Materials B.

Purpose: We implemented the Machine Learning (ML) aided k-t SENSE reconstruction to enable high resolution quantitative real-time phase contrast MR (PCMR). Methods: A residual U-net and our U-net M were used to generate the high resolution x-f space estimate for k-t SENSE regularisation prior. The networks were judged on their ability to generalise to real undersampled data. The in-vivo validation was done on 20 real-time 18x prospectively undersmapled GASperturbed PCMR data. The ML aided k-t SENSE reconstruction results were compared against the free-breathing Cartesian retrospectively gated sequence and the compressed sensing (CS) reconstruction of the same data. Results: In general, the ML aided k-t SENSE generated flow curves that were visually sharper than those produced using CS. In two exceptional cases, U-net M predictions exhibited blurring which propagated to the extracted velocity curves. However, there were no statistical differences in the measured peak velocities and stroke volumes between the tested methods. The ML aided k-t SENSE was estimated to be ~3.6x faster in processing than CS. Conclusion: The ML aided k-t SENSE reconstruction enables artefact suppression on a par with CS with no significant differences in quantitative measures. The timing results suggest the on-line implementation could deliver a substantial increase in clinical throughput., Comment: 26 pages, 6 figures, 6 videos

Additional file 1. Table S1. Subject characteristics.

Magnetic Resonance Imaging (MRI) plays a vital role in diagnosis, management and monitoring of many diseases. However, it is an inherently slow imaging technique. Over the last 20 years, parallel imaging, temporal encoding and compressed sensing have enabled substantial speed-ups in the acquisition of MRI data, by accurately recovering missing lines of k-space data. However, clinical uptake of vastly accelerated acquisitions has been limited, in particular in compressed sensing, due to the time-consuming nature of the reconstructions and unnatural looking images. Following the success of machine learning in a wide range of imaging tasks, there has been a recent explosion in the use of machine learning in the field of MRI image reconstruction. A wide range of approaches have been proposed, which can be applied in k-space and/or image-space. Promising results have been demonstrated from a range of methods, enabling natural looking images and rapid computation. In this review article we summarize the current machine learning approaches used in MRI reconstruction, discuss their drawbacks, clinical applications, and current trends., Comment: 34 pages, 3 figures, 1 table. review article

Additional file 5. Bland-Altman plots of agreement with high-resolution WH-bSSFP for the individual vessels.

Additional file 3. Synthetic test results from different network structures.

Additional file 2. Flow diagram showing the steps taken to convert the high-resolution WH-bSSFP data, to synthetic low-resolution WH-bSSFP data used to train/test the residual U-Net.

Additional file 7. Application of the super-resolution network applied to abdominal WH-bSSFP data.

Additional file 6. Diagnostic Accuracy and Confidence scoring for all lesions.

Supplemental Demographics. (DOCX 12 kb)

Multi-dimensional non-Cartesian MRI encoding using the precomputed interpolator can encounter the curse of dimensionality, in which the interpolator size exceeds the available memory on the parallel accelerators. Here we reformulate the multi-dimensional non-uniform fast Fourier transform (NUFFT) to a tensor form. The exponentially growing size of the fully precomputed interpolator can be reduced by tensor analysis. We propose a tree-like, mixed-radix tensor method which flexibly reduces the storage of the NUFFT. A parallel tensor product algorithm is proposed and tested with in vivo cardiac MRI data. Cross-architecture comparisons show that up to 88.1% and 62.4% memory savings are seen in 3D and 2D CINE MRI, respectively, subject only to a negligible loss of accuracy compared to the double-precision CPU version., Comment: The submitted paper in its current form is inappropriate

This table outlines the clinically relevant blood and urine tests in the mild, moderate and severe CKD groups. (DOCX 17 kb)

Background The right ventricle is a complex structure that is challenging to quantify by two-dimensional (2D) echocardiography. Unlike disk summation three-dimensional (3D) echocardiography (3DE), single-beat 3DE can acquire large volumes at high volume rates in one cardiac cycle, avoiding stitching artifacts or long breath-holds. The aim of this study was to assess the accuracy and test-retest reproducibility of single-beat 3DE for quantifying right ventricular (RV) volumes in adult populations of acquired RV pressure or volume overload, namely, pulmonary hypertension (PH) and carcinoid heart disease, respectively. Three-dimensional and 2D echocardiographic indices were also compared for identifying RV dysfunction in PH. Methods A prospective cross-sectional study was performed in 100 individuals who underwent 2D echocardiography, 3DE, and cardiac magnetic resonance imaging: 49 patients with PH, 20 with carcinoid heart disease, 11 with metastatic carcinoid tumors without cardiac involvement, and 20 healthy volunteers. Two operators performed test-retest acquisition and postprocessing for inter- and intraobserver reproducibility in 20 subjects. Results: RV single-beat 3DE was attainable in 96% of cases, with mean volume rates of 32 to 45 volumes/sec. Bland-Altman analysis of all subjects (presented as mean bias ± 95% limits of agreement) revealed good agreement for end-diastolic volume (−2.3 ± 27.4 mL) and end-systolic volume (5.2 ± 19.0 mL) measured by 3DE and cardiac magnetic resonance imaging, with a tendency to underestimate stroke volume (−7.5 ± 23.6 mL) and ejection fraction (−4.6 ± 13.8%) by 3DE. Subgroup analysis demonstrated a greater bias for volumetric underestimation, particularly in healthy volunteers (end-diastolic volume, −11.9 ± 18.0 mL; stroke volume, −11.2 ± 20.2 mL). Receiver operating characteristic curve analysis showed that 3DE-derived ejection fraction was significantly superior to 2D echocardiographic parameters for identifying RV dysfunction in PH (sensitivity, 94%; specificity, 88%; area under the curve, 0.95; P = .031). There was significant interobserver test-retest bias for RV volume underestimation (end-diastolic volume, −12.5 ± 28.1 mL; stroke volume, −10.6 ± 23.2 mL). Conclusions Single-beat 3DE is feasible and clinically applicable for volumetric quantification in acquired RV pressure or volume overload. It has improved limits of agreement compared with previous disk summation 3D echocardiographic studies and has incremental value over standard 2D echocardiographic measures for identifying RV dysfunction. Despite the ability to obtain and postprocess a full-volume 3D echocardiographic RV data set, the quality of the raw data did influence the accuracy of the data obtained. The technique performs better with dilated rather than nondilated RV cavities, with a learning curve that might affect the test-retest reproducibility for serial RV studies.

Background Undesirable outcomes in health care are associated with patient harm and substantial excess costs. Coarctation of the aorta (CoA), one of the most common congenital heart diseases, can be repaired with stenting but requires monitoring and subsequent interventions to detect and treat disease recurrence and aortic wall injuries. Avoidable costs associated with stenting in patients with CoA are unknown. Methods We developed an economic model to calculate potentially avoidable costs in stenting treatment of CoA in the United Kingdom over 5 years. We calculated baseline costs for the intervention and potentially avoidable complications and follow-up interventions and compared these to the costs in hypothetical scenarios with improved treatment effectiveness and complication rates. Results Baseline costs were £16 688 ($25 182) per patient. Avoidable costs ranged from £137 ($207) per patient in a scenario assuming a 10% reduction in aortic wall injuries and reinterventions at follow-up, to £1627 ($2455) in a Best-case scenario with 100% treatment success and no complications. Overall costs in the Best-case scenario were 90.2% of overall costs at Baseline. Reintervention rate at follow-up was identified as most influential lever for overall costs. Probabilistic sensitivity analysis showed a considerable degree of uncertainty for avoidable costs with widely overlapping 95% confidence intervals. Conclusions Significant improvements in the treatment effectiveness and reductions in complication rates are required to realize discernible cost savings. Up to 10% of total baseline costs could be avoided in the best-case scenario. This highlights the need to pursue patient-specific treatment approaches which promise optimal outcomes. Electronic supplementary material The online version of this article (doi:10.1186/s12913-017-2215-2) contains supplementary material, which is available to authorized users.

The supplemental material provides additional information on model inputs and results of the extensive sensitivity analyses conducted. The supplemental material is provided in one document under the following section headings: A1. Results of probabilistic sensitivity analysis. A2. Calculation of cost inputs and standard deviations. A3. Univariate sensitivity analyses of all input parameters. (PDF 1.14 mb)

The human stress response activates the autonomic nervous system and endocrine systems to increase performance during environmental challenges. This response is usually beneficial, improving the chance of overcoming environmental challenges, but costs resources such as energy. Humans and other animals are known to adapt their responses to acute stress when they are stimulated chronically, presumably to optimise resource utilisation. Characterisation of these adaptations has been limited. Using advanced imaging techniques, we show that cardiovascular and endocrine physiology, reflective of energy utilisation during acute stress, and energy storage (fat) differ between the sexes when they are exposed to chronic stress. We examine possible evolutionary explanations for these differences, related to energy use, and point out how these physiological differences could underpin known disparities between the sexes in their risk of important cardiometabolic disorders such as obesity and cardiovascular disease.Obesity and associated diseases, such as cardiovascular disease, are the dominant human health problems in the modern era. Humans develop these conditions partly because they consume excess energy and exercise too little. Stress might be one of the factors contributing to these disease-promoting behaviours. We postulate that sex-specific primordial energy optimisation strategies exist, which developed to help cope with chronic stress but have become maladaptive in modern societies, worsening health. To demonstrate the existence of these energy optimisation strategies, we recruited 88 healthy adults with varying adiposity and chronic stress exposure. Cardiovascular physiology at rest and during acute stress (Montreal Imaging Stress Task), and body fat distribution were measured using advanced magnetic resonance imaging methods, together with endocrine function, cardiovascular energy use and cognitive performance. Potential confounders such as lifestyle, social class and employment were accounted for. We found that women exposed to chronic stress had lower adiposity, greater acute stress cardiovascular responses and better cognitive performance. Conversely, chronic stress-exposed men had greater adiposity and lower cardiovascular responses to acute stress. These results provide initial support for our hypothesis that differing sex-specific energy conservation strategies exist. We propose that these strategies have initially evolved to benefit humans but are now maladaptive and increase the risk of disorders such as obesity, especially in men exposed to chronic stress.