WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
We propose a method based on deep learning to perform cardiac segmentation on short axis MRI image stacks iteratively from the top slice (around the base) to the bottom slice (around the apex). At each iteration, a novel variant of U-net is applied to propagate the segmentation of a slice to the adjacent slice below it. In other words, the prediction of a segmentation of a slice is dependent upon the already existing segmentation of an adjacent slice. 3D-consistency is hence explicitly enforced. The method is trained on a large database of 3078 cases from UK Biobank. It is then tested on 756 different cases from UK Biobank and three other state-of-the-art cohorts (ACDC with 100 cases, Sunnybrook with 30 cases, RVSC with 16 cases). Results comparable or even better than the state-of-the-art in terms of distance measures are achieved. They also emphasize the assets of our method, namely enhanced spatial consistency (currently neither considered nor achieved by the state-of-the-art), and the generalization ability to unseen cases even from other databases.
Description of cardiovascular phenotype in the UK Biobank population based on cardiovascular magnetic resonance and carotid ultrasound
Imaging of the heart and blood vessels is performed in a large subset of the UK Biobank cohort. Many measures defining the state of the heart and blood vessels can be derived from the images acquired. These measures are influenced by various health conditions and modifiable and non-modifiable factors, such as age, gender and ethnicity. The aim of this proposal is to describe the measures of the heart and blood vessel in the UK Biobank population and investigate how much modifiable and non-modifiable factors influence them. All new data will be made available for future research. Knowing the reference ranges for common imaging measures of the heart and circulation and how they are influenced by factors, such as age, gender, ethnicity, risk factors for heart attacks and strokes, is key for improving making diagnoses and predicting health outcomes. Descriptive statistics will be performed for all image derived phenotypes (IDPs) from the cardiovascular magnetic resonance (CMR) and carotid ultrasound images. We will perform subgroup analysis for important clinical factors, such as age, gender, cardiovascular risk, chronic conditions (e.g. Diabetes). We will apply descriptive statistics to a subpopulation considered `healthy without cardiovascular disease or presence of modifiable risk factors`. Univariate and multivariate regression analysis will be used to assess relationships between IDPs and relevant co-variates. We will also assess intra- and inter-observer variability for IDP measurement when repeat analysis is available. Initial 5000 subjects from the imaging enhancement study.