Adiposity influences on myocardial deformation: a cardiovascular magnetic resonance feature tracking study in people with overweight to obesity without established cardiovascular disease.

The objective of this study was to assess whether dietary-induced weight loss improves myocardial deformation in people with overweight to obesity without established cardiovascular disease applying cardiovascular magnetic resonance (CMR) with feature tracking (FT) based strain analysis. Ninety people with overweight to obesity without established cardiovascular disease (age 44.6 ± 9.3 years, body mass index (BMI) 32.6 ± 4 kg/m2) underwent CMR. We retrospectively quantified FT based strain and LA size and function at baseline and after a 6-month hypocaloric diet, with either low-carbohydrate or low-fat intake. The study cohort was compared to thirty-four healthy normal-weight controls (age 40.8 ± 16.0 years, BMI 22.5 ± 1.4 kg/m2). At baseline, the study cohort with overweight to obesity without established cardiovascular disease displayed significantly increased global circumferential strain (GCS), global radial strain (GRS) and LA size (all p < 0.0001 versus controls) but normal global longitudinal strain (GLS) and normal LA ejection fraction (all p > 0.05 versus controls). Dietary-induced weight loss led to a significant reduction in GCS, GRS and LA size irrespective of macronutrient composition (all p < 0.01). In a population with overweight to obesity without established cardiovascular disease subclinical myocardial changes can be detected applying CMR. After dietary-induced weight loss improvement of myocardial deformation could be shown. A potential clinical impact needs further studies.

Predicting post-contrast information from contrast agent free cardiac MRI using machine learning: Challenges and methods.

Objectives: Currently, administering contrast agents is necessary for accurately visualizing and quantifying presence, location, and extent of myocardial infarction (MI) with cardiac magnetic resonance (CMR). In this study, our objective is to investigate and analyze pre- and post-contrast CMR images with the goal of predicting post-contrast information using pre-contrast information only. We propose methods and identify challenges. Methods: The study population consists of 272 retrospectively selected CMR studies with diagnoses of MI (n = 108) and healthy controls (n = 164). We describe a pipeline for pre-processing this dataset for analysis. After data feature engineering, 722 cine short-axis (SAX) images and segmentation mask pairs were used for experimentation. This constitutes 506, 108, and 108 pairs for the training, validation, and testing sets, respectively. We use deep learning (DL) segmentation (UNet) and classification (ResNet50) models to discover the extent and location of the scar and classify between the ischemic cases and healthy cases (i.e., cases with no regional myocardial scar) from the pre-contrast cine SAX image frames, respectively. We then capture complex data patterns that represent subtle signal and functional changes in the cine SAX images due to MI using optical flow, rate of change of myocardial area, and radiomics data. We apply this dataset to explore two supervised learning methods, namely, the support vector machines (SVM) and the decision tree (DT) methods, to develop predictive models for classifying pre-contrast cine SAX images as being a case of MI or healthy. Results: Overall, for the UNet segmentation model, the performance based on the mean Dice score for the test set (n = 108) is 0.75 (±0.20) for the endocardium, 0.51 (±0.21) for the epicardium and 0.20 (±0.17) for the scar. For the classification task, the accuracy, F1 and precision scores of 0.68, 0.69, and 0.64, respectively, were achieved with the SVM model, and of 0.62, 0.63, and 0.72, respectively, with the DT model. Conclusion: We have presented some promising approaches involving DL, SVM, and DT methods in an attempt to accurately predict contrast information from non-contrast images. While our initial results are modest for this challenging task, this area of research still poses several open problems.

Assessment of diastolic dysfunction: comparison of different cardiovascular magnetic resonance techniques.

AIMS: Heart failure with preserved ejection fraction is still a diagnostic and therapeutic challenge, and accurate non-invasive diagnosis of left ventricular (LV) diastolic dysfunction (DD) remains difficult. The current study aimed at identifying the most informative cardiovascular magnetic resonance (CMR) parameters for the assessment of LVDD. METHODS AND RESULTS: We prospectively included 50 patients and classified them into three groups: with DD (DD+, n = 15), without (DD-, n = 26), and uncertain (DD±, n = 9). Diagnosis of DD was based on echocardiographic E/E', invasive LV end-diastolic pressure, and N-terminal pro-brain natriuretic peptide. CMR was performed at 1.5 T to assess LV and left atrial (LA) morphology, LV diastolic strain rate (SR) by tissue tracking and tagging, myocardial peak velocities by tissue phase mapping, and transmitral inflow profile using phase contrast techniques. Statistics were performed only on definitive DD+ and DD- (total number 41). DD+ showed enlarged LA with LA end-diastolic volume/height performing best to identify DD+ with a cut-off value of ≥0.52 mL/cm (sensitivity = 0.71, specificity = 0.84, and area under the receiver operating characteristic curve = 0.75). DD+ showed significantly reduced radial (inferolateral E peak: DD-: -14.5 ± 6.5%/s vs. DD+: -10.9 ± 5.9%/s, P = 0.04; anterolateral A peak: DD-: -4.2 ± 1.6%/s vs. DD+: -3.1 ± 1.4%/s, P = 0.04) and circumferential (inferolateral A peak: DD-: 3.8 ± 1.2%/s vs. DD+: 2.8 ± 0.8%/s, P = 0.007; anterolateral A peak: DD-: 3.5 ± 1.2%/s vs. DD+: 2.5 ± 0.8%/s, P = 0.048) SR in the basal lateral wall assessed by tissue tracking. In the same segments, DD+ showed lower peak myocardial velocity by tissue phase mapping (inferolateral radial peak: DD-: -3.6 ± 0.7 ms vs. DD+: -2.8 ± 1.0 ms, P = 0.017; anterolateral longitudinal peak: DD-: -5.0 ± 1.8 ms vs. DD+: -3.4 ± 1.4 ms, P = 0.006). Tagging revealed reduced global longitudinal SR in DD+ (DD-: 45.8 ± 12.0%/s vs. DD+: 34.8 ± 9.2%/s, P = 0.022). Global circumferential and radial SR by tissue tracking and tagging, LV morphology, and transmitral flow did not differ between DD+ and DD-. CONCLUSIONS: Left atrial size and regional quantitative myocardial deformation applying CMR identified best patients with DD.

Auto-threshold quantification of late gadolinium enhancement in patients with acute heart disease.

PURPOSE: To assess the Otsu-Auto-Threshold (OAT) for accuracy and reproducibility for sizing irreversible injury in late gadolinium enhancement (LGE) images of patients with acute heart disease. The OAT method automatically identifies high signal intensity areas using a cutoff derived from the signal intensity histogram and therefore is user-independent. MATERIALS AND METHODS: LGE was performed in 28 patients with acute myocardial infarction (MI) and 30 patients with acute myocarditis. LGE mass was compared between OAT and thresholds using 2 standard deviations (SD), 3SD, and 5SD above remote myocardium, and full-width-at-half-maximum (FWHM). A separate, blinded visual assessment served as the standard of truth. RESULTS: In patients with acute MI, OAT and 5SD did not differ (26.1 ± 11.4 g vs. 25.4 ± 11.1 g, P = 0.088), but thresholds of 2SD and 3SD overestimated LGE mass by 37% and 20%, respectively, and FWHM underestimated by 15%. In acute myocarditis, OAT was not different from a visual quantification, but thresholds of 2SD and 3SD overestimated LGE mass by 46% and 19%, respectively, and thresholds of 5SD and FWHM underestimated LGE mass by 17% and 26%, respectively. OAT and FWHM showed the best intraobserver and interobserver reproducibility. CONCLUSION: Automatic thresholding using OAT may serve as an accurate and reproducible method to quantify irreversible myocardial injury in acute heart disease.