3D Regional Shape Analysis of Left Ventricle Using MR Images: Abnormal Myocadium Detection and Classification

Linear Discriminant Analysis (LDA) and Multilinear Principal Component Analysis (MPCA) are leading subspace methods for achieving dimension reduction based on supervised learning. Both LDA and MPCA use class labels of data samples to calculate subspaces onto which these samples are projected. Furthermore, both methods have been successfully applied to face recognition. Although LDA and MPCA share common goals and methodologies, in previous research they have been applied separately and independently. In this paper, we propose an extension of LDA to multiple factor frameworks. Our proposed method, Multifactor Discriminant Analysis, aims to obtain multilinear projections that maximize the between-class scatter while minimizing the withinclass scatter, which is the same core fundamental objective of LDA. Moreover, Multifactor Discriminant Analysis (MDA), like MPCA, uses multifactor analysis and calculates subject parameters that represent the characteristics of subjects and are invariant to other changes, such as viewpoints or lighting conditions. In this way, our proposed MDA combines the best virtues of both LDA and MPCA for face recognition.

In the last few years, great efforts have been made to extend the linear projection technique (LPT) for multidimensional data (i.e., tensor), generally referred to as the multilinear projection technique (MPT). The vectorized nature of LPT requires high-dimensional data to be converted into vector, and hence may lose spatial neighborhood information of raw data. MPT well addresses this problem by encoding multidimensional data as general tensors of a second or even higher order. In this paper, we propose a novel multilinear projection technique, called multilinear spatial discriminant analysis (MSDA), to identify the underlying manifold of high-order tensor data. MSDA considers both the nonlocal structure and the local structure of data in the transform domain, seeking to learn the projection matrices from all directions of tensor data that simultaneously maximize the nonlocal structure and minimize the local structure. Different from multilinear principal component analysis (MPCA) that aims to preserve the global structure and tensor locality preserving projection (TLPP) that is in favor of preserving the local structure, MSDA seeks a tradeoff between the nonlocal (global) and local structures so as to drive its discriminant information from the range of the non-local structure and the range of the local structure. This spatial discriminant characteristic makes MSDA have more powerful manifold preserving ability than TLPP and MPCA. Theoretical analysis shows that traditional MPTs, such as multilinear linear discriminant analysis, TLPP, MPCA, and tensor maximum margin criterion, could be derived from the MSDA model by setting different graphs and constraints. Extensive experiments on face databases (ORL, CMU PIE, and the extended Yale-B) and the Weizmann action database demonstrate the effectiveness of the proposed MSDA method.

This paper introduces a pioneering Hybrid Parallel Multi-linear Face Recognition algorithm that capitalizes on multi-linear methodologies, such as Multi-linear Principal Component Analysis (MPCA), Linear Discriminant Analysis (LDA), and Histogram of Oriented Gradients (HOG), to attain exceptional recognition performance. The Hybrid Feature Selection (HFS) algorithm is meticulously crafted to augment the classification performance on the CK+ and FERET datasets by amalgamating the strengths of feature extraction techniques and feature selection methods. HFS seamlessly incorporates Principal Component Analysis (PCA), Local Discriminant Analysis (LDA), and HOG. The primary aim of this algorithm is to autonomously identify a subset of the most distinctive features from the extracted feature pool, thus elevating classification accuracy, precision, recall, and F1-Score. By amalgamating these methodologies, the algorithm adeptly diminishes dimensionality while conserving pivotal features. Experimental trials on facial image datasets, CK+ and FERET, underscore the algorithm's supremacy in terms of accuracy and computational efficiency when contrasted with conventional linear techniques and even certain deep learning approaches. The proposed algorithm proffers an encouraging solution for real-world face recognition applications where precision and operational efficiency are of paramount significance.

Highlights From the Circulation Family of Journals.

Objective: Psoriasis, which is a chronic inflammatory disease is associated with an increased risk of cardiovascular (CV) diseases (CVD). Currently, no method is available to evaluate subclinical CV dysfunction in patients with psoriasis. The aim of the study was to evaluate left ventricular (LV) function in patients with untreated psoriasis but no risk factors for CV disease (CVD). Material and Methods: Thirty patients with psoriasis without any CVD or risk factors for CVD and 20 age- and gender-matched controls were recruited. All patients underwent detailed transthoracic echocardiography, including speckle tracking-derived strain analysis. Results: Speckle tracking echocardiography showed significantly lower global circumferential strain in psoriatic patients compared to healthy controls (mean ±SD: -23.34±6.74% vs. -27.53±4.11%, respectively, p<0.001). No significant differences were found in terms of global longitudinal strain, aortic flow velocity, pulmonary arterial flow velocity, LV end-systolic volume, LV end-diastolic volume, LV ejection fraction, peak of early diastolic (E) and late diastolic (A) flow velocity, E⁄A ratio, and levels of biochemical markers, including C-reactive protein, between the patients and controls (p>0.05). Conclusion: Subclinical myocardial dysfunction was seen in the patients with psoriasis who had no conventional risk factors for CVD using two-dimensional strain imaging. We believe that this imaging technique may be useful for early detection of CVD development in patients with psoriasis.

LEFT VENTRICULAR MASS PREDICTS CARDIAC REVERSE REMODELING IN PATIENTS TREATED WITH EMPAGLIFLOZIN: AN EXPLORATORY SUB-ANALYSIS OF THE EMPA-HEART CARDIOLINK-6 RANDOMIZED CONTROLLED TRIAL

Background While left atrial (LA) dysfunction is often considered a reflection of diastolic dysfunction, typifying changes in the ageing heart, it may also serve as an indicator of early left ventricular (LV) systolic deterioration that remains undetectable through conventional LV evaluation in cardiovascular imaging. Despite the lack of gross LV dysfunction in population studies, the development of subsequent cardiovascular disease (CVD) over time among settings of LA dysfunction may suggest early perturbations in the LV. Left atrioventricular coupling (LACI), a marker of LA-to-LV physiological interactions, is believed to be a marker for characterizing early perturbations in the LV. We investigated the prognostic value of LACI and LA measures obtained by advanced cardiac magnetic resonance (CMR) imaging and incident outcomes in older adults. Methods This prospective, population-based cohort study included 255 adults (70.8±9.1years, 55% male) without CVD at baseline. LA and LV volumes were derived from CMR. LACI was calculated as the ratio of LA-to-LV end-diastolic volume indexed (LAEDVi/LVEDVi). Feature tracking on CMR assessed LA emptying fractions (LAEF), LA and LV strains. Clinical outcomes were defined as a composite of all-cause mortality or hospitalization. Results CMR measurements for the cohort showed a mean±SD of LACI (30.1±14.3%), with 60% having LACI&amp;gt;25%. Total LAEF was 52.0±9.5%, while LV parameters included an ejection fraction of 65.1±7.5%, LVEDVi of 64.3%±13.8 ml/m2 and LV mass-indexed (LVMi) of 46.3±11.6 g/m2. Over a follow-up of 5.21±1.60 years, 61 participants developed incident clinical outcomes. LACI&amp;gt;25% was most strongly associated with the risk of incident outcomes on univariable analysis (hazard ratio (HR) = 1.88, 95% confidence interval (CI) (1.07-3.29), p = 0.027) and multivariable analysis (HR = 2.18 95%CI (1.05-4.52), p = 0.035) adjusting for cardiovascular risk factors, LVMi and LV global longitudinal strain (Table 1). The log-rank test further supported this association (x2 = 5.27, p = 0.022) (Figure 1). LACI, when included in this model, demonstrated incremental discrimination and calibration in the prediction of incident outcomes (C-statistic = 0.656, 95%CI (0.564-0.747)); likelihood-ratio-test x2 = 4.26, p = 0.039; Gronnesby-Borgan’s score test (G-B) x2 = 14.07 p = 0.120). LACI&amp;gt;25% had the highest discriminatory value (C-statistic = 0.678, 95%CI (0.590-0.765)) and Total LAEF had the best model calibration (G-B x2 = 8.89 p = 0.448) in the prediction of incident outcomes, after adjusting for the multivariable model (Table 1). Conclusion Among a prospective cohort of older adults free of baseline CVD, an increased LACI and decreased LAEF was independently associated with and incrementally predictive of future clinical outcomes. Determination of LACI in asymptomatic older adults may reflect heightened risks of CVD development in ageing.

Introduction: Spherical remodeling of atria and ventricles is associated with cardiovascular disease (CVD) and mortality. We hypothesized that volumetric left atrial (LA) and left ventricular (LV) sphericity index (SI) are associated with adverse measures of LV structure and function in older adults free of CVD. Methods: Framingham Heart Study 2nd generation participants without prevalent CVD with echocardiography at exam 8 were studied (n=1701). Volumetric LA SI was calculated as the ratio of biplane LA volume to the volume of a sphere [4/3*pi*(r^3)] calculated using the radius of LA length (r=½ average length of the atria in 4- and 2-chamber views in diastole). Volumetric LV SI was calculated as the ratio of LV end-diastolic volume (LVEDV) to the volume of a sphere calculated using the radius of LV length (r= ½ length of LV in 4-chamber view in diastole). LVEDV was measured by biplane area-length method. In multivariable models adjusting for CV risk factors including age, sex, BMI, hypertension, dyslipidemia, diabetes, and smoking, we evaluated the relations of LA and LV SI with echocardiographic LV structure and function: wall thickness (septal and posterior wall average), mass, global longitudinal strain, ejection fraction (EF), and log-transformed mitral Doppler E/A. β coefficients are presented as standard deviation (SD) change in outcome per SD change in predictor variable. Results: Descriptives are presented in the Table . There was no association of LA or LV SI with sex (p=0.48, p=0.61, respectively). LA SI was associated with worse (positive) LV global longitudinal strain (std β±SE: 0.08±0.03, p=0.016). LV SI was associated with LV wall thickness (std β±SE: 0.24±0.03, p&lt;0.0001), mass (0.45±0.03, p&lt;0.0001) and log-Doppler E/A (0.07±0.03, p=0.015) and was inversely associated with LV EF (-0.12±0.03, p&lt;0.0001). Conclusions: In a large community sample without prevalent CVD, spherical remodeling of the heart was associated with unfavorable LV structure and both systolic and diastolic dysfunction.

Assessment of Left Ventricular Mass in Children and Adolescents: Current Status

Treatment of Diabetes in People with Heart Failure

In This Issue: March.

Introduction: Type 2 diabetes mellitus (T2DM) is associated with increased risk for left ventricular (LV) hypertrophy and heart failure (HF). Moreover, some drugs used to treat hyperglycaemia are associated with increased risk of HF. There are limited data on the effects of exogenous insulin on left ventricular mass (LVM) and function (LVF). Therefore, we evaluated the effects of insulin glargine on LVM and LVF in the echocardiographic substudy of the Outcome Reduction with Initial Glargine Intervention (ORIGIN) trial (funded by Sanofi; ORIGIN ClinicalTrials.gov number, NCT00069784). Methods: Patients aged ≥ 50 years with dysglycemia (early T2DM, impaired glucose tolerance, or impaired fasting glucose) and with cardiovascular (CV) disease or CV risk factors were randomized to receive insulin glargine (with a target fasting blood glucose level of ≤95 mg per deciliter [5.3 mmol per liter]) or standard glycemic care. Echocardiograms were performed at baseline and after 3 years. Changes in LVM, LV ejection fraction (LVEF), wall motion score (WMS), LV end-diastolic and end-systolic volume (LVEDV and LVESV), E/A, and E/E’ were compared between the study groups using mixed effects modelling. Results: 564 patients aged 64 ± 8 years were evaluated; 30% were women, 84% had a history of hypertension and 32% of prior myocardial infarction. This is the largest reported study on the effects of exogenous insulin on LVM and LVF. There were no significant differences in the changes over time in LVM and LV systolic and diastolic function between the study groups (Table). Conclusions: In people with CV disease and/or CV risk factors and dysglycemia three years of treatment with insulin glargine has a neutral effect on LVM and on LV systolic and diastolic function. These findings support the CV safety of insulin glargine.

Introduction: Myocardial contraction fraction (MCF) is the ratio of left ventricular (LV) stroke volume to myocardial volume, and thus a measure of LV pumping capacity per unit of myocardium. We sought to determine whether MCF measured using current steady-state free precession (SSFP) cardiac magnetic resonance (CMR) sequences was an independent predictor of incident “hard” cardiovascular disease (CVD) events, defined by myocardial infarction (MI), stroke, unstable angina (UA), hospitalized heart failure (HF) or CVD death in a community dwelling cohort initially free of these CVD events. Methods: 1794 members of the Framingham Heart Study Offspring cohort (aged 65±9 years) underwent CMR between 2002-2006 using a 1.5-Tesla system with contiguous multislice SSFP cine imaging to encompass the left ventricle. MCF was determined from the cine images by a single observer blinded to participant characteristics. We tracked incident hard CVD events over median 6.5-year follow up and used Cox proportional hazards models (adjusted for age, sex, body mass index, systolic blood pressure, diabetes, dyslipidemia, smoking, treatment for hypertension) to determine hazard of hard CVD events per increment (0.10) of MCF. Results: MCF was determined in 1776 (99%) Offspring (835 men). Overall, MCF was greater in women (0.92±0.14 vs. 0.78±0.15 for men), p&lt;0.0001. There were 60 incident hard CVD events during follow up. Incident hard events included 26 MI, 2 UA, 13 stroke, 14 hospitalized HF and 5 CVD deaths. Offspring experiencing an incident event had lower MCF (0.78±0.19 vs. 0.86±0.15 for those free of events), p=0.002. On MV-adjusted Cox proportional hazards analyses, a greater MCF was protective against hard CVD events, HR [95% confidence intervals] = 0.76 [0.63 - 0.93] per 0.10 increment of MCF. Conclusion: Over 6.5-year follow-up, greater MCF is protective against major adverse CVD events, even after adjustment for traditional CVD risk factors in a community dwelling cohort of middle-aged and older predominantly European-descended adults. Determination of MCF requires only knowledge of LV stroke volume and myocardial volume, both of which are routinely determined in a standard CMR examination of the left ventricle, and thus imposes no additional scan-time or analysis burden. While MCF may be clinically useful for prediction of risk for incident hard CVD events, its potential value in younger age groups and other ethnicities remains to be determined.

Clinical Summaries

Background The risk of cardiovascular (CV) disease in type 2 diabetes mellitus (T2DM) is highly heterogeneous and an adequate risk stratification is needed. CV risk mainly depends on concomitant risk factors, combined in the metabolic syndrome (MetS). Echocardiography is an useful tool for diagnosis of cardiac organ damage and CV risk stratification in T2DM. Purpose To investigate the effects of MetS on left ventricular (LV) structure and function in patients with T2DM with normal LV ejection fraction (LVEF) and without overt coronary artery disease (CAD) and heart failure symptoms/signs. Methods We prospectively recruited 384 consecutive, uncomplicated T2DM patients. All patients underwent clinical exam, blood sampling and complete echo-Doppler exam, including determination of 2D-echo derived global longitudinal strain (GLS). LV mass index ≥45 g/m^2.7 in women and ≥49 g/m^2.7 in men was used to characterize LV hypertrophy. LV longitudinal dysfunction was assumed for GLS &amp;lt; 20% in absolute values. LV diastolic dysfunction was identified according to 2016 ASE/EACVI recommendations. MetS was defined according to NCEP-ATP III criteria. Significant CAD including previous myocardial infarction, LV systolic dysfunction (= LVEF &amp;lt;50%), hemodinamically significant valvular heart disease, primary cardiomyopathies, permanent atrial fibrillation, glomerular filtration rate &amp;lt;30 ml/min and inadequate echo images were exclusion criteria. The study population was divided according to presence of MetS. Results 66% of the patients (254/384) met the criteria for MetS diagnosis. They had comparable age and heart rate with controls. Diabetic patients with MetS had higher glycated haemoglobin (HbA1c) (7.2 ± 1.3 vs. 6.9 ± 1.0%, p = 0.023) and uric acid (5.5 ± 1.4 vs. 4.9 ± 1.3 mg/dl, p = 0.001) than those without, and lower glomerular filtration rate (69.5 ± 15.0 vs 74.0 ± 12,1 ml/min, p = 0.004). MetS patients showed a higher LV mass index (p &amp;lt; 0.0001) and a greater prevalence of both LV hypertrophy (31.9 vs 12.5%, p &amp;lt; 0.0001) and diastolic dysfunction (52.6 vs. 32.8%, p = 0.007) (Figure). T2DM patients with MetS also had lower GLS (20.6 ± 2.1 vs. 21.9 ± 2.2%, p = 0.001), with a greater prevalence of LV longitudinal dysfunction (38.2 vs. 24.7%, p = 0.049) (Figure). After adjusting for age, T2DM duration, sex, HbA1c, uric acid, LV mass index and LV diastolic dysfunction by a multiple regression analysis in the pooled population, GLS reduction was independently associated with MetS (β coefficient = -0.184, p &amp;lt; 0.02) and LV mass index (β= -0.94, p = 0.04). Conclusion In patients with T2DM, the presence of MetS induces a greater prevalence not only of LV hypertrophy and diastolic dysfunction but also of LV longitudinal deformation impairment. GLS reduction in diabetic patients is associated with MetS independently of confounders including glycemic control and diabetic duration. Abstract 433 Figure. Rate of LV alterations according to MetS