Obesity is a major risk factor for cardiometabolic disease, but the effect of body composition on vascular aging and arterial stiffness remains uncertain. We investigated relationships among body ...composition, blood pressure, age, and aortic pulse wave velocity in healthy individuals. Pulse wave velocity in the thoracic aorta, an indicator of central arterial stiffness, was measured in 221 volunteers (range, 18–72 years; mean, 40.3±13 years) who had no history of cardiovascular disease using cardiovascular MRI. In univariate analyses, age (r=0.78; P<0.001) and blood pressure (r=0.41; P<0.001) showed a strong positive association with pulse wave velocity. In multivariate analysis, after adjustment for age, sex, and mean arterial blood pressure, elevated body fat% was associated with reduced aortic stiffness until the age of 50 years, thereafter adiposity had an increasingly positive association with aortic stiffness (β=0.16; P<0.001). Body fat% was positively associated with cardiac output when age, sex, height, and absolute lean mass were adjusted for (β=0.23; P=0.002). These findings suggest that the cardiovascular system of young adults may be capable of adapting to the state of obesity and that an adverse association between body fat and aortic stiffness is only apparent in later life.
To characterize the genetic architecture of left ventricular noncompaction (LVNC) and investigate the extent to which it may represent a distinct pathology or a secondary phenotype associated with ...other cardiac diseases.
We performed rare variant association analysis with 840 LVNC cases and 125,748 gnomAD population controls, and compared results to similar analyses on dilated cardiomyopathy (DCM) and hypertrophic cardiomyopathy (HCM).
We observed substantial genetic overlap indicating that LVNC often represents a phenotypic variation of DCM or HCM. In contrast, truncating variants in MYH7, ACTN2, and PRDM16 were uniquely associated with LVNC and may reflect a distinct LVNC etiology. In particular, MYH7 truncating variants (MYH7tv), generally considered nonpathogenic for cardiomyopathies, were 20-fold enriched in LVNC cases over controls. MYH7tv heterozygotes identified in the UK Biobank and healthy volunteer cohorts also displayed significantly greater noncompaction compared with matched controls. RYR2 exon deletions and HCN4 transmembrane variants were also enriched in LVNC, supporting prior reports of association with arrhythmogenic LVNC phenotypes.
LVNC is characterized by substantial genetic overlap with DCM/HCM but is also associated with distinct noncompaction and arrhythmia etiologies. These results will enable enhanced application of LVNC genetic testing and help to distinguish pathological from physiological noncompaction.
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 (4D
), 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 = .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.
Incorporation of prior knowledge about organ shape and location is key to improve performance of image analysis approaches. In particular, priors can be useful in cases where images are corrupted and ...contain artefacts due to limitations in image acquisition. The highly constrained nature of anatomical objects can be well captured with learning-based techniques. However, in most recent and promising techniques such as CNN-based segmentation it is not obvious how to incorporate such prior knowledge. State-of-the-art methods operate as pixel-wise classifiers where the training objectives do not incorporate the structure and inter-dependencies of the output. To overcome this limitation, we propose a generic training strategy that incorporates anatomical prior knowledge into CNNs through a new regularisation model, which is trained end-to-end. The new framework encourages models to follow the global anatomical properties of the underlying anatomy (e.g. shape, label structure) via learnt non-linear representations of the shape. We show that the proposed approach can be easily adapted to different analysis tasks (e.g. image enhancement, segmentation) and improve the prediction accuracy of the state-of-the-art models. The applicability of our approach is shown on multi-modal cardiac data sets and public benchmarks. In addition, we demonstrate how the learnt deep models of 3-D shapes can be interpreted and used as biomarkers for classification of cardiac pathologies.
Accurate discrimination of benign and pathogenic rare variation remains a priority for clinical genome interpretation. State-of-the-art machine learning variant prioritization tools are imprecise and ...ignore important parameters defining gene–disease relationships, e.g., distinct consequences of gain-of-function versus loss-of-function variants. We hypothesized that incorporating disease-specific information would improve tool performance.
We developed a disease-specific variant classifier, CardioBoost, that estimates the probability of pathogenicity for rare missense variants in inherited cardiomyopathies and arrhythmias. We assessed CardioBoost’s ability to discriminate known pathogenic from benign variants, prioritize disease-associated variants, and stratify patient outcomes.
CardioBoost has high global discrimination accuracy (precision recall area under the curve AUC 0.91 for cardiomyopathies; 0.96 for arrhythmias), outperforming existing tools (4–24% improvement). CardioBoost obtains excellent accuracy (cardiomyopathies 90.2%; arrhythmias 91.9%) for variants classified with >90% confidence, and increases the proportion of variants classified with high confidence more than twofold compared with existing tools. Variants classified as disease-causing are associated with both disease status and clinical severity, including a 21% increased risk (95% confidence interval CI 11–29%) of severe adverse outcomes by age 60 in patients with hypertrophic cardiomyopathy.
A disease-specific variant classifier outperforms state-of-the-art genome-wide tools for rare missense variants in inherited cardiac conditions (https://www.cardiodb.org/cardioboost/), highlighting broad opportunities for improved pathogenicity prediction through disease specificity.
IntroductionImprovements in cardiac magnetic resonance (CMR) have enabled better phenotyping of the left atrium (LA). However, little is known of the incremental prognostic value of the novel LA ...measurements (phasic LA strain, LA ejection fraction LAEF, and LA minimum volume LAVImin) compared to LA maximum volume LAVImax in dilated cardiomyopathy (DCM). Thus, we decided to evaluate the prognostic value of each LA measure in DCM.Materials and MethodsCMR studies of 580 DCM patients, in sinus rhythm, prospectively enrolled into a biobank between 2009 and 2017 were used. The primary endpoint was a composite of cardiovascular (CV) mortality and non-fatal major heart failure (HF) events, which included HF hospitalisations, heart transplantation or Left Ventricular (LV) assist device implantation. Event rates were compared between patients in sinus rhythm and those with persistent atrial fibrillation (AF).ResultsOver a median follow-up duration of 7.4 years (IQR 4.7–9.3), 103 patients (18%) met the primary endpoint. On univariable Cox regression analysis, all LA metrics were significantly associated with the primary endpoint (all, p<0.05). All indices, apart from LA conduit strain, remained associated with the endpoint on multivariate analyses adjusted for age, sex, NYHA, LV ejection fraction and the presence of fibrosis (all, p<0.05). The addition of the LA metrics to a baseline model containing conventional risk predictors improved model discrimination, with LAVImin providing the greatest improvement (C-statistic 0.702 to 0.738), similar to that of LAVImax (C-Statistic: 0.702 to 0.732) and LAEF (C-Statistic: 0.702 to 0.734). LA strain variables did not improve baseline model discrimination over LA volumes. Patients in the highest tercile of LAVImin had similar event rates to those with persistent atrial fibrillation.DiscussionIn line with previous studies, LA structure and function was independently associated with CV death and HF events. LA volumes and LAEF provided better prognostication than LA strain. Amongst the volumes, LAVImin improved baseline model discrimination better than LAVImax, perhaps because it reflects both LA structure and function. This is important as LAVImin can easily be added to CMR reporting protocols.ConclusionLA metrics provide incremental prognostic information in DCM patients. LA strain did not provide any additional prognostic information over LA volumes. AcknowledgementsWe would like to thank the Royal Brompton and Harefield Cardiovascular Research Centre nurses and support staff.
IntroductionWith expansion of family screening, genetic testing and advanced cardiovascular imaging, more patients are being diagnosed with early-stage non-ischaemic cardiomyopathy (early-NICM), ...often before symptom-onset. Observational studies of such patients are lacking. We sought to characterise early-NICM phenotype, evaluate risk of adverse outcomes and assess the rate of disease progression.Materials and MethodsWe conducted a prospective observational cohort study of patients with early-NICM assessed by LGE-CMR. Cases were classified based on presence/absence of LV dilatation and reduced LVEF into subgroups: isolated LV dilation (early-NICM H-/D+) in cases with LV dilatation without hypokinesia; non-dilated LV cardiomyopathy (early-NICM H+/D-) in cases with hypokinesia without dilatation; early dilated cardiomyopathy (early-NICM H+/D+) in cases with LV dilatation and mild hypokinesia. Follow-up for major adverse cardiovascular events (MACE) included life-threatening arrhythmia, unplanned cardiovascular hospitalisation or cardiovascular death. A subset of patients (n=119) underwent a second CMR to assess changes in cardiac structure and function.ResultsOf 254 patients with early-NICM (median age 46 years IQR 36–58, 94 37% women, median LVEF 55% 52–59), myocardial fibrosis was present in 65 (26%). There was no difference in the prevalence of fibrosis among subgroups (p=0.90), however fibrosis mass was lowest in early-NICM H-/D+, higher in early-NICM H+/D- and highest in early-NICM H+/D+ (p=0.03). Over a median follow-up of 7.9 (5.5–10.0) years, 28 patients (11%) experienced MACE. Myocardial fibrosis (HR 3.77, 95%CI 1.73–8.20, p<0.001), non-sustained VT (HR 5.10, 95%CI 2.36–11.00, p<0.001) and diabetes mellitus (HR 5.12, 95%CI 1.73–15.18, p=0.003) were associated with MACE in a multivariable model. Only 8% of patients progressed from early-NICM to DCM with LVEF<50% over a median of 16 (11–34) months.DiscussionMyocardial fibrosis deposition occurs early in the phenotypic course of NICM, often preceding obvious adverse cardiac remodelling or symptom-onset. The observation of a gradient in fibrosis mass across early-NICM subgroups suggests it may be integral to disease progression. The finding that 11% of patients with early-NICM experienced MACE over long-term follow-up highlights the importance of risk stratification and surveillance, despite the low rate of early phenotypic progression.Abstract 26 Figure 1ConclusionEarly-NICM is not benign. Fibrosis is an early feature of disease. In-depth characterisation enhances risk stratification and might aid clinical management.AcknowledgementsWe thank the Royal Brompton and Harefield Cardiovascular Research Centre nurses and support staff.
IntroductionGreater precision is required for arrhythmic risk stratification of patients with non-ischaemic cardiomyopathy (NICM). We sought to evaluate whether fibrosis entropy,a measure of scar ...texture heterogeneity derived from late gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR), had incremental utility to fibrosis presence for arrhythmic risk prediction in NICM.Materials and MethodsProspective observational cohort study of consecutive patients with NICM and myocardial fibrosis detected by LGE-CMR. Entropy (computed as standard Shannon Entropy) was calculated for core fibrosis, gray zone (GZ), and combined core and GZ fibrosis. Core fibrosis was classified using full width half maximum method and GZ fibrosis as regions with ≥35% but <50% maximum signal intensity. Patients were followed up for life-threatening arrhythmia LTA (sudden cardiac death, aborted sudden cardiac death or sustained ventricular tachycardia).ResultsOf 292 patientswith NICM and mid-wall/subepicardial fibrosis, 38 patients (13.0%) experienced LTA over median follow-up of 6.3 years (IQR 4.6–9.1years). Core fibrosis entropy (HR 1.63, 95%CI 1.15–2.30, p=0.006), GZ fibrosis entropy (HR 1.68, 95%CI 1.16–2.44, p=0.006) and combined fibrosis entropy (HR 1.65, 95%CI 1.11–2.44, p=0.013) were independently associated with LTA after adjustment for variables used to guide ICD implantation (LVEF<35% and NYHA class >1) and remained associated in multivariable models accounting separately for core and GZ fibrosis mass. The addition of core fibrosis entropy, GZ fibrosis entropy and combined fibrosis entropy to a baseline clinical model improved the C-statistic from 0.49 to 0.63, 0.62 and 0.63 respectively. LVEF<35% was not associated with LTA (HR 1.45, 95%CI 0.77–2.74, p=0.25).DiscussionFibrosis entropy was associated with LTA in patients with NICM and mid-wall/subepicardial fibrosis. This association was independent of benchmark variables used in clinical practice and the addition of fibrosis entropy to arrhythmic risk models enhanced their predictive power. The comparable increment in C-statistic for core and GZ fibrosis entropy indicates equivalent effect on arrhythmic risk from each. The lack of association between LVEF and LTA highlights the inadequacy of the current paradigm for determining primary prevention ICD candidacy and provides further impetus for transition towards methods directly evaluating underlying arrhythmic substrate.ConclusionFibrosis entropy has incremental utility to fibrosis presence in arrhythmic risk prediction of patients with NICM.AcknowledgementsWe thank the Royal Brompton and Harefield Cardiovascular Research Centre nurses and support staff, led by Ms Geraldine Sloane.