LU-Net: A Multistage Attention Network to Improve the Robustness of Segmentation of Left Ventricular Structures in 2-D Echocardiography.

Segmentation of cardiac structures is one of the fundamental steps to estimate volumetric indices of the heart. This step is still performed semiautomatically in clinical routine and is, thus, prone to interobserver and intraobserver variabilities. Recent studies have shown that deep learning has the potential to perform fully automatic segmentation. However, the current best solutions still suffer from a lack of robustness in terms of accuracy and number of outliers. The goal of this work is to introduce a novel network designed to improve the overall segmentation accuracy of left ventricular structures (endocardial and epicardial borders) while enhancing the estimation of the corresponding clinical indices and reducing the number of outliers. This network is based on a multistage framework where both the localization and segmentation steps are optimized jointly through an end-to-end scheme. Results obtained on a large open access data set show that our method outperforms the current best-performing deep learning solution with a lighter architecture and achieved an overall segmentation accuracy lower than the intraobserver variability for the epicardial border (i.e., on average a mean absolute error of 1.5 mm and a Hausdorff distance of 5.1mm) with 11% of outliers. Moreover, we demonstrate that our method can closely reproduce the expert analysis for the end-diastolic and end-systolic left ventricular volumes, with a mean correlation of 0.96 and a mean absolute error of 7.6 ml. Concerning the ejection fraction of the left ventricle, results are more contrasted with a mean correlation coefficient of 0.83 and an absolute mean error of 5.0%, producing scores that are slightly below the intraobserver margin. Based on this observation, areas for improvement are suggested.

Screening Rheumatic Heart Disease in 1530 New Caledonian Adolescents.

BACKGROUND In New Caledonia, a South Pacific archipelago whose inhabitants comprise Melanesians, Europeans/whites, Wallisians, Futunans, Polynesians, and Asians, the prevalence of rheumatic heart disease (RHD) is 0.9% to 1% at ages 9 and 10. It could be higher at the age of 16, but this remains to be verified. METHODS AND RESULTS A total of 1530 Melanesian, Métis, white, Wallisian, Futunan, Polynesian, and Asian adolescents benefited from a transthoracic echocardiogram. Definite or borderline RHD, nonrheumatic valve lesions, congenital heart defects, family and personal history of acute rheumatic fever, and socioeconomic factors were collected. The prevalence of cardiac abnormalities was 8.1%, made up of 4.1% RHD including 2.4% definite and 1.7% borderline RHD, 1.7% nonrheumatic valve lesions, and 2.3% congenital anomalies. In whites and Asians, there were no cases of RHD. RHD was higher in the Wallisian, Futunan, and Polynesian group (7.6%) when compared with Melanesians (5.3%) and Métis (2.9%). The number of nonrheumatic valve lesions was not statistically different in the different ethnicities. The prevalence of RHD was higher in adolescents with a personal history of acute rheumatic fever, in those living in overcrowded conditions, and in those whose parents were unemployed or had low-income occupations, such as the farmers or manual workers. CONCLUSIONS RHD was 4 times higher in adolescents at age 16 than at ages 9 and 10 (4.1% versus 0.9%-1%). No cases of RHD were observed in whites and Asians. The determining factors were history of acute rheumatic fever and socioeconomic factors.

Deep Learning for Segmentation Using an Open Large-Scale Dataset in 2D Echocardiography.

Delineation of the cardiac structures from 2D echocardiographic images is a common clinical task to establish a diagnosis. Over the past decades, the automation of this task has been the subject of intense research. In this paper, we evaluate how far the state-of-the-art encoder-decoder deep convolutional neural network methods can go at assessing 2D echocardiographic images, i.e., segmenting cardiac structures and estimating clinical indices, on a dataset, especially, designed to answer this objective. We, therefore, introduce the cardiac acquisitions for multi-structure ultrasound segmentation dataset, the largest publicly-available and fully-annotated dataset for the purpose of echocardiographic assessment. The dataset contains two and four-chamber acquisitions from 500 patients with reference measurements from one cardiologist on the full dataset and from three cardiologists on a fold of 50 patients. Results show that encoder-decoder-based architectures outperform state-of-the-art non-deep learning methods and faithfully reproduce the expert analysis for the end-diastolic and end-systolic left ventricular volumes, with a mean correlation of 0.95 and an absolute mean error of 9.5 ml. Concerning the ejection fraction of the left ventricle, results are more contrasted with a mean correlation coefficient of 0.80 and an absolute mean error of 5.6%. Although these results are below the inter-observer scores, they remain slightly worse than the intra-observer's ones. Based on this observation, areas for improvement are defined, which open the door for accurate and fully-automatic analysis of 2D echocardiographic images.

Pre-participation cardiovascular evaluation in Pacific Island athletes.

BACKGROUND: Pre-participation cardiovascular evaluation (PPE) aims to detect cardiac disease with sudden cardiac death (SCD) risk. No study has focused on Pacific Island athletes. METHODS: A total of 2281 Pacific Island athletes were studied with (i) a questionnaire on family, personal history and symptoms, (ii) a physical examination and (iii) a 12-lead ECG. RESULTS: 85% presented a normal history and examination. A positive family history was 1.4-1.9 fold higher in Melanesians, Polynesians and Métis than in Caucasians, while a positive personal history, abnormal symptoms and abnormal examination was 1.3 fold higher in Melanesians and Métis than in others. Neither gender nor training level had a bearing on these results. Melanesians had higher T wave inversions (TWIs) in V2-V4 leads but had no CV abnormalities. Lateral or infero-lateral TWIs were found in 6 male and in 5 highly trained athletes and cardiomyopathies were diagnosed in 3/6 athletes. Overall, 3.9% athletes were found to have a CV abnormality and 0.8% had a risk of SCD. Polynesians and males were more at risk than the others while the level of training made no difference. In athletes at risk of SCD, the main detected CV diseases were cardiomyopathies, Wolff-Parkinson-White (WPW) and severe valve lesions of rheumatoid origin. CONCLUSIONS: PPE revealed that 3.9% presented CV abnormalities. A risk of SCD was found in 0.8% with cardiomyopathies, WPW, and severe valve lesions of rheumatoid origin. Melanesians, Polynesians and male of high level of training were more at risk than others.