Explore the vast range of titles available.

Motion analysis is used in computer vision to understand the behaviour of moving objects in sequences of images. Optimizing the interpretation of dynamic biological systems requires accurate and precise motion tracking as well as efficient representations of high-dimensional motion trajectories so that these can be used for prediction tasks. Here we use image sequences of the heart, acquired using cardiac magnetic resonance imaging, to create time-resolved three-dimensional segmentations using a fully convolutional network trained on anatomical shape priors. This dense motion model formed the input to a supervised denoising autoencoder (4Dsurvival), which is a hybrid network consisting of an autoencoder that learns a task-specific latent code representation trained on observed outcome data, yielding a latent representation optimized for survival prediction. To handle right-censored survival outcomes, our network used a Cox partial likelihood loss function. In a study of 302 patients, the predictive accuracy (quantified by Harrell’s C -index) was significantly higher ( P  = 0.0012) for our model C  = 0.75 (95% CI: 0.70–0.79) than the human benchmark of C  = 0.59 (95% CI: 0.53–0.65). This work demonstrates how a complex computer vision task using high-dimensional medical image data can efficiently predict human survival. A fully convolutional neural network is used to create time-resolved three-dimensional dense segmentations of heart images. This dense motion model forms the input to a supervised system called 4Dsurvival that can efficiently predict human survival.

The inner surfaces of the human heart are covered by a complex network of muscular strands that is thought to be a remnant of embryonic development 1 , 2 . The function of these trabeculae in adults and their genetic architecture are unknown. Here we performed a genome-wide association study to investigate image-derived phenotypes of trabeculae using the fractal analysis of trabecular morphology in 18,096 participants of the UK Biobank. We identified 16 significant loci that contain genes associated with haemodynamic phenotypes and regulation of cytoskeletal arborization 3 , 4 . Using biomechanical simulations and observational data from human participants, we demonstrate that trabecular morphology is an important determinant of cardiac performance. Through genetic association studies with cardiac disease phenotypes and Mendelian randomization, we find a causal relationship between trabecular morphology and risk of cardiovascular disease. These findings suggest a previously unknown role for myocardial trabeculae in the function of the adult heart, identify conserved pathways that regulate structural complexity and reveal the influence of the myocardial trabeculae on susceptibility to cardiovascular disease. A genome-wide association study shows that myocardial trabeculae are an important determinant of cardiac performance in the adult heart, identifies conserved pathways that regulate structural complexity and reveals the influence of trabeculae on the susceptibility to cardiovascular disease.

ABSTRACT Artificial intelligence‐based cardiac motion mapping offers predictive insights into pulmonary hypertension (PH) disease progression and its impact on the heart. We proposed an automated deep learning pipeline for bi‐ventricular segmentation and 3D wall motion analysis in PH rodent models for bridging the clinical developments. A data set of 163 short‐axis cine cardiac magnetic resonance scans were collected longitudinally from monocrotaline (MCT) and Sugen‐hypoxia (SuHx) PH rats and used for training a fully convolutional network for automated segmentation. The model produced an accurate annotation in < 1 s for each scan (Dice metric > 0.92). High‐resolution atlas fitting was performed to produce 3D cardiac mesh models and calculate the regional wall motion between end‐diastole and end‐systole. Prominent right ventricular hypokinesia was observed in PH rats (−37.7% ± 12.2 MCT; −38.6% ± 6.9 SuHx) compared to healthy controls, attributed primarily to the loss in basal longitudinal and apical radial motion. This automated bi‐ventricular rat‐specific pipeline provided an efficient and novel translational tool for rodent studies in alignment with clinical cardiac imaging AI developments.

To the Editor: Myles et al. (Jan. 12 issue) 1 found that the risk of death was not significantly higher with tranexamic acid than with placebo among patients undergoing cardiac surgery. This drug has a class IA indication for bleeding prophylaxis, decreasing the use of blood products and the risk of reintervention. Doses that are less than 50 mg per kilogram of body weight are effective in preventing bleeding as well as in decreasing the inflammatory response that is associated with cardiopulmonary bypass. 2 Patients with the 5G/G genotype had a greater blood-sparing benefit with the use of tranexamic acid than those . . .

Background Cardiac phenotypes, such as left ventricular (LV) mass, demonstrate high heritability although most genes associated with these complex traits remain unidentified. Genome-wide association studies (GWAS) have relied on conventional 2D cardiovascular magnetic resonance (CMR) as the gold-standard for phenotyping. However this technique is insensitive to the regional variations in wall thickness which are often associated with left ventricular hypertrophy and require large cohorts to reach significance. Here we test whether automated cardiac phenotyping using high spatial resolution CMR atlases can achieve improved precision for mapping wall thickness in healthy populations and whether smaller sample sizes are required compared to conventional methods. Methods LV short-axis cine images were acquired in 138 healthy volunteers using standard 2D imaging and 3D high spatial resolution CMR. A multi-atlas technique was used to segment and co-register each image. The agreement between methods for end-diastolic volume and mass was made using Bland-Altman analysis in 20 subjects. The 3D and 2D segmentations of the LV were compared to manual labeling by the proportion of concordant voxels (Dice coefficient) and the distances separating corresponding points. Parametric and nonparametric data were analysed with paired t-tests and Wilcoxon signed-rank test respectively. Voxelwise power calculations used the interstudy variances of wall thickness. Results The 3D volumetric measurements showed no bias compared to 2D imaging. The segmented 3D images were more accurate than 2D images for defining the epicardium (Dice: 0.95 vs 0.93, P < 0.001; mean error 1.3 mm vs 2.2 mm, P < 0.001) and endocardium (Dice 0.95 vs 0.93, P < 0.001; mean error 1.1 mm vs 2.0 mm, P < 0.001). The 3D technique resulted in significant differences in wall thickness assessment at the base, septum and apex of the LV compared to 2D (P < 0.001). Fewer subjects were required for 3D imaging to detect a 1 mm difference in wall thickness (72 vs 56, P < 0.001). Conclusions High spatial resolution CMR with automated phenotyping provides greater power for mapping wall thickness than conventional 2D imaging and enables a reduction in the sample size required for studies of environmental and genetic determinants of LV wall thickness.

Cardiovascular imaging is perhaps the area in which ML methods have been most extensively tested and have demonstrated the greatest immediate potential.1 These algorithms have been used for more efficient image acquisition and reconstruction, automated quality control, image segmentation, myocardial motion and blood flow analysis, and computer-assisted diagnosis. Biases in the training data, model overfitting, inadequate statistical correction for multiple testing and limited transparency around the processes by which DL algorithms reach their output (‘black box’ systems) are only some of the pitfalls of AI that can have significant implications for patients and require careful evaluation by researchers, clinicians and regulatory entities. The British Heart Foundation has also recently announced a partnership with HDRUK to enable responsible research combining the power of advanced analytical methods with the UK’s large-scale and diverse cardiovascular data resources.

Stuart Cook and colleagues study the role of TTN (titin)-truncating variants using a combination of heart physiology experiments in rats and genomic analysis in humans. Their data show that TTN variants are associated with a range of cardiac phenotypes in healthy individuals and in patients with dilated cardiomyopathy. Titin-truncating variants (TTNtv) commonly cause dilated cardiomyopathy (DCM). TTNtv are also encountered in ∼1% of the general population, where they may be silent, perhaps reflecting allelic factors. To better understand TTNtv, we integrated TTN allelic series, cardiac imaging and genomic data in humans and studied rat models with disparate TTNtv. In patients with DCM, TTNtv throughout titin were significantly associated with DCM. Ribosomal profiling in rat showed the translational footprint of premature stop codons in Ttn , TTNtv-position-independent nonsense-mediated degradation of the mutant allele and a signature of perturbed cardiac metabolism. Heart physiology in rats with TTNtv was unremarkable at baseline but became impaired during cardiac stress. In healthy humans, machine-learning-based analysis of high-resolution cardiac imaging showed TTNtv to be associated with eccentric cardiac remodeling. These data show that TTNtv have molecular and physiological effects on the heart across species, with a continuum of expressivity in health and disease.

Motion analysis is used in computer vision to understand the behaviour of moving objects in sequences of images. Optimising the interpretation of dynamic biological systems requires accurate and precise motion tracking as well as efficient representations of high-dimensional motion trajectories so that these can be used for prediction tasks. Here we use image sequences of the heart, acquired using cardiac magnetic resonance imaging, to create time-resolved three-dimensional segmentations using a fully convolutional network trained on anatomical shape priors. This dense motion model formed the input to a supervised denoising autoencoder (4Dsurvival), which is a hybrid network consisting of an autoencoder that learns a task-specific latent code representation trained on observed outcome data, yielding a latent representation optimised for survival prediction. To handle right-censored survival outcomes, our network used a Cox partial likelihood loss function. In a study of 302 patients the predictive accuracy (quantified by Harrell's C-index) was significantly higher (p < .0001) for our model C=0.73 (95\\(\\%\\) CI: 0.68 - 0.78) than the human benchmark of C=0.59 (95\\(\\%\\) CI: 0.53 - 0.65). This work demonstrates how a complex computer vision task using high-dimensional medical image data can efficiently predict human survival.

MOTIVATION: Left ventricular (LV) hypertrophy is a strong predictor of cardiovascular outcomes, but its genetic regulation remains largely unexplained. Conventional phenotyping relies on manual calculation of LV mass and wall thickness, but advanced cardiac image analysis presents an opportunity for high-throughput mapping of genotype-phenotype associations in three dimensions (3D). RESULTS: High-resolution cardiac magnetic resonance images were automatically segmented in 1,124 healthy volunteers to create a 3D shape model of the heart. Mass univariate regression was used to plot a 3D effect-size map for the association between wall thickness and a set of predictors at each vertex in the mesh. The vertices where a significant effect exists were determined by applying threshold-free cluster enhancement to boost areas of signal with spatial contiguity. Experiments on simulated phenotypic signals and SNP replication show that this approach offers a substantial gain in statistical power for cardiac genotype-phenotype associations while providing good control of the false discovery rate. This framework models the effects of genetic variation throughout the heart and can be automatically applied to large population cohorts. AVAILABILITY: The proposed approach has been coded in an R package freely available at https://doi.org/10.5281/zenodo.834610 together with the clinical data used in this work.

IntroductionAs gene sequencing becomes more widespread, the clinical implications of incidental findings in patients' genomes are becoming more complex. We identified healthy volunteers with mutations known to cause penetrant, Mendelian aortic disease, and examined the association of these mutations with aortic pulse wave velocity (PWV); a key marker of cardiovascular risk and aortic elastic function.MethodsWe recruited 476 healthy volunteers with no known history of cardiovascular risk factors or disease for aortic phenotyping and gene sequencing. We measured aortic arch pulse wave velocity derived from Cardiovascular MRI (CMR) using ArtFun software and performed whole exome sequencing (Illumina HiSeq 2000). Sequence was alignedto hg19 reference using BWA v0.7.10 and variants were called using GATK and validated using IGV. Variants that are presumed causative for aortic disease were prioritised using HGMD and were further annotated by literature review. A binary variable, \"AOV status\" reflected the presence or absence of a variant linked with aortopathy by this approach. Statistical analysis was performed using linear regression modelling, Mann-Whitney U tests and bootstrapping in R. For Mann-Whitney U tests, we used age-corrected PWV [=PWV/log (Age)].Results17 of our healthy volunteers (3%) had previously reported pathogenic mutations in seven aortopathy genes (COL1A2, COL3A1, FBN1, MYH11, MYLK, TGFBR1 and TGFBR2; see Table 1).Abstract 175 Table 1Mutations in our cohort linked with aortic disease in HGMDGeneBase substitutionAmino acid substitutionNumber in cohortMinor Allele Frequency (ExAC,%)Disease associationEvidence of pathogenicity COL1A2 c.2123G >Ap. Arg708Gln10.06MarfanPatient + affected father; functional confirmation COL3A1 c.2002C >Ap. Pro668Thr30.17Ehlers DanlosSingle case FBN1 c.7379A >Gp. Lys2460Arg10.007MarfanSingle casec.6700G >Ap. Val2234Met10.079MarfanSingle casec.3422C >Tp. Pro1141Leu10.072MarfanSingle case MYH11 c.4604G >Ap. Arg1535Gln20.23FTAASingle case MYLK c.4195G >Ap. Glu1399Lys10.044FTAASingle casec.3637G >Ap. Val1213Met20.010FTAASingle case TGFBR1 c.1433A >Gp. Asn478Ser10.028Loeys DietzSingle case TGFBR2 c.1119G >Ap. Met373Ile10.139Loeys DietzSingle casec.1159G >Ap. Val387Met20.116Loeys Dietz2 family members; functional confirmationc.1657T >Ap. Ser553Thr10.14Loeys Dietz2 separate case series found variantThree mutations (in 4 individuals; 0.8% of our cohort) had evidence for pathogenicity (eg family linkage analysis) beyond just a single case report; two in TGFBR2 (3 individuals) and one in COL1A2. These four subjects had significantly higher aPWVs than control cases (Mann-Whitney U test; U=235, p = 0.01; see Figure 1), and than those cases where the evidence for variant pathogenicity was limited (p = 0.02).[Figure]Linear regression modelling of PWV was significantly improved by the addition of AOV status (ANOVA of nested linear models; p = 0.03; F=5.56(1), R2=0.07; multiple R2=0.57; p < 0.001), and this improvement was more evident with bootstrapped linear regression (p < 0.01). Stepwise model selection by AIC prioritised AOV status in the final linear model, which included log (age), gender, MAP, pulse rate, body surface area and fat mass.DiscussionIt is not unusual to find \"disease-causing\" variants in an apparently healthy population. Sometimes this is used to argue against the pathogenicity of a particular variant. However, our results imply that some of these \"healthy\" individuals may have penetrant aortic disease but of limited expressivity. These individuals may have increased risk of cardiovascular events, and so-called \"spontaneous\" aortic aneurysm and dissection. This finding therefore has implications for genetic counselling, as well as for the conduct of gene sequencing studies.