Segmentation of heart substructures in 2D echocardiography images is an important step in diagnosis and management of cardiovascular disease. Given the ubiquity of echocardiography in routine cardiology practice, the time-consuming nature of manual segmentation, and the high degree of inter-observer variability, fully automatic segmentation is a goal common to both clinicians and researchers. The recent publication of the annotated CA- MUS dataset will help catalyze these efforts. In this work we develop and validate against this dataset a deep fully convolutional neural network architecture for the multi-structure segmentation of echocardiography, in- cluding the left ventricular endocardium and epicardium, and the left atrium. In ten-fold cross validation with data augmentation, we obtain mean Dice overlaps of 0.93, 0.95, and 0.89 on the three structures respectively, representing state of the art on this dataset. We further report small biases and narrow limits of agreement between the automatic and manual segmentations in derived clinical indices, including median absolute errors for left ventricular diastolic (7.4mL) and systolic (4.8mL) volumes, and ejection fraction (4.1%), within previously reported inter-observer variability. These encouraging results must still be validated against large-scale independent clinical data.
Segmentation of heart substructures in cardiac magnetic resonance (CMR) is an important step in the quantitative assessment of the impact of cardiovascular disease. Manual delineation of these structures, over many patients and multiple time phases, is time consuming and prone to human error and fatigue. In this work we use a deep fully convolutional neural network architecture to automatically segment heart substructures in CMR, achieving state of the art results on a recent benchmark dataset. We further apply our process to a much larger study of CMR subjects, automatically segmenting both left and right ventricular endocardiums (LV, RV) with full thirty-phase time resolution, and LV epicardium (Epi) at end-diastole. We validate our automatically obtained results against manual delineations using Dice overlap and Hausdorff distance, as well as Bland-Altman limits of agreement on the derived blood volumes, ejection fraction, and LV mass. We obtain median Dice overlaps of 0.97, 0.94, and 0.97 on the three structures respectively, and further find small biases and narrow limits of agreement between the two assessments (manual, automatic) of volumes and mass. Our results show promise for the fully automated analysis of the CMR data stream in the near future.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.