AUTHOR=Zhao Debbie , Ferdian Edward , Maso Talou Gonzalo D. , Quill Gina M. , Gilbert Kathleen , Wang Vicky Y. , Babarenda Gamage Thiranja P. , Pedrosa João , D’hooge Jan , Sutton Timothy M. , Lowe Boris S. , Legget Malcolm E. , Ruygrok Peter N. , Doughty Robert N. , Camara Oscar , Young Alistair A. , Nash Martyn P. 

TITLE=MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging

JOURNAL=Frontiers in Cardiovascular Medicine

VOLUME=Volume 9 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/cardiovascular-medicine/articles/10.3389/fcvm.2022.1016703

DOI=10.3389/fcvm.2022.1016703

ISSN=2297-055X

ABSTRACT=Segmentation of the left ventricle (LV) in echocardiography is an important task for the quantification of volume and mass in heart disease. Continuing advances in echocardiography have extended imaging capabilities into the 3D domain, subsequently overcoming the geometric assumptions associated with conventional 2D acquisitions. Nevertheless, analysis of 3D echocardiography (3DE) poses several challenges associated with limited spatial resolution, poor contrast-to-noise ratio, complex noise characteristics, and image anisotropy.

To develop automated methods for 3DE analysis, a sufficiently large, labelled dataset is typically required. However, ground truth analyses have historically been difficult to obtain due to the high inter-observer variability associated with manual analysis. We address this lack of expert consensus by registering labels derived from higher-resolution cardiac magnetic resonance (CMR) imaging from the same subject, to produce 536 annotated 3DE images from a mixed cohort of 134 human subjects consisting of healthy controls and patients with cardiac disease across a range of demographics. To demonstrate the utility of such a dataset, a state-of-the-art, self-configuring deep learning network for semantic segmentation is employed for automated 3DE analysis. Using the proposed dataset for training, the network produced measurement biases of -9 ± 16 ml, -1 ± 10 ml, -2 ± 5 %, and 5 ± 23 g, for end-diastolic volume, end-systolic volume, ejection fraction, and mass, respectively, outperforming an expert human observer in terms of accuracy as well as scan-rescan reproducibility.

As part of the ongoing efforts to improve frameworks for 3DE, we present here a large, publicly available dataset with ground truth labels that leverage the higher resolution and contrast of CMR, to provide a new benchmark for automated 3DE analysis. Such an approach not only significantly reduces the effect of observer-specific bias present in manual 3DE annotations, but also improves the agreement between 3DE and CMR, enabling more accurate diagnostic and prognostic information to be obtained from echocardiography.