Exponential growth in data storage and computational power is rapidly narrowing the gap between translating findings from advanced clinical informatics into cardiovascular clinical practice. Specifically, cardiovascular imaging has the distinct advantage in providing a great quantity of data for potentially rich insights, but nuanced interpretation requires a high-level skillset that few individuals possess. A subset of machine learning, deep learning (DL), is a modality that has shown promise, particularly in the areas of image recognition, computer vision, and video classification. Due to a low signal-to-noise ratio, echocardiographic data tend to be challenging to classify; however, utilization of robust DL architectures may help clinicians and researchers automate conventional human tasks and catalyze the extraction of clinically useful data from the petabytes of collected imaging data. The promise is extending far and beyond towards a contactless echocardiographic exam—a dream that is much needed in this time of uncertainty and social distancing brought on by a stunning pandemic culture. In the current review, we discuss state-of-the-art DL techniques and architectures that can be used for image and video classification, and future directions in echocardiographic research in the current era.
1 November 2020 in Arteriosclerosis, Thrombosis, and Vascular Biology
Objective: Coronavirus disease 2019 (COVID-19) can infect patients in any age group including those with no comorbid conditions. Understanding the demographic, clinical, and laboratory characteristics of these patients is important toward developing successful treatment strategies. Approach and Results: In a retrospective study design, consecutive patients without baseline comorbidities hospitalized with confirmed COVID-19 were included. Patients were subdivided into ≤55 and >55 years of age. Predictors of in-hospital mortality or mechanical ventilation were analyzed in this patient population, as well as subgroups. Stable parameters in overall and subgroup models were used to construct a cluster model for phenotyping of patients. Of 1207 COVID-19–positive patients, 157 met the study criteria (80≤55 and 77>55 years of age). Most reliable predictors of outcomes overall and in subgroups were age, initial and follow-up d -dimer, and LDH (lactate dehydrogenase) levels. Their predictive cutoff values were used to construct a cluster model that produced 3 main clusters. Cluster 1 was a low-risk cluster and was characterized by younger patients who had low thrombotic and inflammatory features. Cluster 2 was intermediate risk that also consisted of younger population that had moderate level of thrombosis, higher inflammatory cells, and inflammatory markers. Cluster 3 was a high-risk cluster that had the most aggressive thrombotic and inflammatory feature. Conclusions: In healthy patient population, COVID-19 remains significantly associated with morbidity and mortality. While age remains the most important predictor of in-hospital outcomes, thromboinflammatory interactions are also associated with worse clinical outcomes regardless of age in healthy patients.
Background: Heart failure with preserved ejection fraction (HFpEF) is prevalent in women and is associated with atrial fibrillation (AF). However, sex associations in AF-related HFpEF are not well explored. Aim: We studied differences between men and women with and without AF-related HFpEF symptoms on left ventricular (LV) geometry and diastolic dysfunction (DD) and their effect on cardiovascular events. Methods: Retrospectively, HFpEF patients with and without a history of AF referred for echocardiography were studied. Echocardiographic assessments were focused on LV geometry and diastolic functions. Patients were followed for the occurrence of cardiac events defined as death and cardiac hospitalization. Results: We studied 556 patients [age: 66.7 ± 17 years, 320 (58%) women, 91 (16%) AF]. Compared to HFpEF without AF (HFpEF-AF), HFpEF with AF patients (HFpEF+AF) were older (76 ± 13.8 vs. 64.9 ± 17.3 years, p < 0.001), had more risk factors, comorbidities, left ventricular hypertrophy (32 vs. 13%, p < 0.001), higher relative wall thickness (0.50 ± 0.14 vs. 0.44 ± 0.15, p < 0.001), and DD (56 vs. 30%, all p < 0.001). HFpEF+AF women had the worst clinical, LV geometric, and diastolic functional profiles and highest rates of cardiovascular outcomes compared to HFpEF+AF men and were the only group to predict outcomes (HR: 2.7, 95%CI: 1.4–5.1), while HFpEF-AF women were a low-risk group; HFpEF+AF and HFpEF-AF men had intermediate cardiovascular outcomes which were confirmed after propensity score matching. Conclusions: Among patients with HFpEF, women with AF had more abnormal LV geometry and diastolic function and had an increased risk of adverse cardiovascular outcomes independent of traditional risk factors, comorbidities, and baseline diastolic function.