The emergence of deep learning has considerably advanced the state-of-the-art in cardiac magnetic resonance (CMR) segmentation. Many techniques have been proposed over the last few years, bringing the accuracy of automated segmentation close to human performance. However, these models have been all too often trained and validated using cardiac imaging samples from single clinical centres or homogeneous imaging protocols. This has prevented the development and validation of models that are generalizable across different clinical centres, imaging conditions or scanner vendors. To promote further research and scientific benchmarking in the field of generalizable deep learning for cardiac segmentation, this paper presents the results of the Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Segmentation (M&Ms) Challenge, which was recently organized as part of the MICCAI 2020 Conference. A total of 14 teams submitted different solutions to the problem, combining various baseline models, data augmentation strategies, and domain adaptation techniques. The obtained results indicate the importance of intensity-driven data augmentation, as well as the need for further research to improve generalizability towards unseen scanner vendors or new imaging protocols. Furthermore, we present a new resource of 375 heterogeneous CMR datasets acquired by using four different scanner vendors in six hospitals and three different countries (Spain, Canada and Germany), which we provide as open-access for the community to enable future research in the field.

The pathophysiology of cardiovascular complications in people with type 1 diabetes (T1DM) remains unclear. An increase in epicardial adipose tissue (EAT) and alterations in the composition of high-density lipoprotein (HDL) are associated with coronary artery disease, but information on its relationship in T1DM is very limited. Our aim was to determine the association between EAT volume, subclinical atherosclerosis, and HDL composition in type 1 diabetes. Seventy-two long-term patients with T1DM without clinical atherosclerosis were analyzed. EAT volume and subclinical atherosclerosis were measured using cardiac computed tomography angiography. EAT was adjusted according to body surface to obtain an EAT index (iEAT). HDL composition was determined. The mean iEAT was 40.47 ± 22.18 cc/m2. The bivariate analysis showed positive associations of the iEAT with gender, age, hypertension, dyslipidemia, smoking, body mass index, waist circumference, insulin dose, and triglyceride (P < 0.05). The iEAT correlated positively with small HDL, increased content of apolipoprotein (apo)A-II and apoC-III, and decreased content of apoE and free cholesterol. Multiple linear regression showed that age, apoA-II content in HDL, and waist circumference were independently associated with the iEAT. Fifty percent of the patients presented subclinical atherosclerotic lesions. These patients had a higher iEAT, and their HDL contained less cholesterol and more apoA-II and lipoprotein-associated phospholipase A2 than patients without subclinical atherosclerosis. Alterations in the composition of HDL in TIDM are associated with increased iEAT and the presence of subclinical atherosclerosis. We propose that these abnormalities of HDL composition could be useful to identify T1DM patients at highest cardiovascular risk.

Few studies have analyzed the potential of biophysical parameters as markers of cardiac remodeling post-myocardial infarction (MI), particularly in human hearts. Fourier transform infrared spectroscopy (FTIR) illustrates the overall changes in proteins, nucleic acids and lipids in a single signature. The aim of this work was to define the FTIR and lipidomic pattern for human left ventricular remodeling post-MI. A total of nine explanted hearts from ischemic cardiomyopathy patients were collected. Samples from the right ventricle (RV), left ventricle (LV) and infarcted left ventricle (LV INF) were subjected to biophysical (FTIR and differential scanning calorimetry, DSC) and lipidomic (liquid chromatography–high-resolution mass spectrometry, LC–HRMS) studies. FTIR evidenced deep alterations in the myofibers, extracellular matrix proteins, and the hydric response of the LV INF compared to the RV or LV from the same subject. The lipid and esterified lipid FTIR bands were enhanced in LV INF, and both lipid indicators were tightly and positively correlated with remodeling markers such as collagen, lactate, polysaccharides, and glycogen in these samples. Lipidomic analysis revealed an increase in several species of sphingomyelin (SM), hexosylceramide (HexCer), and cholesteryl esters combined with a decrease in glycerophospholipids in the infarcted tissue. Our results validate FTIR indicators and several species of lipids as useful markers of left ventricular remodeling post-MI in humans.

Epicardial adipose tissue (EAT) constitutes a novel parameter for cardiometabolic risk assessment and a target for therapy. Here, we evaluated for the first time the plasma microRNA (miRNA) profile as a source of biomarkers for epicardial fat volume (EFV). miRNAs were profiled in plasma samples from 180 patients whose EFV was quantified using multidetector computed tomography. In the screening study, 54 deregulated miRNAs were identified in patients with high EFV levels (highest tertile) compared with matched patients with low EFV levels (lowest tertile). After filtering, 12 miRNAs were selected for subsequent validation. In the validation study, miR-15b-3p, miR-22-3p, miR-148a-3p miR-148b-3p and miR-590-5p were directly associated with EFV, even after adjustment for confounding factors (p value < 0.05 for all models). The addition of miRNA combinations to a model based on clinical variables improved the discrimination (area under the receiver-operating-characteristic curve (AUC) from 0.721 to 0.787). miRNAs correctly reclassified a significant proportion of patients with an integrated discrimination improvement (IDI) index of 0.101 and a net reclassification improvement (NRI) index of 0.650. Decision tree models used miRNA combinations to improve their classification accuracy. These results were reproduced using two proposed clinical cutoffs for epicardial fat burden. Internal validation corroborated the robustness of the models. In conclusion, plasma miRNAs constitute novel biomarkers of epicardial fat burden.