Approved Research

Automated and robust cardiac function quantification from cardiac MRI

Lay summary

Cardiovascular disease is the leading cause of death globally, according to the World Health Organization. Cardiovascular magnetic resonance imaging (MRI) is considered the gold standard for evaluating heart function. Estimating the ventricular end-systolic (ESV) and end-diastolic (EDV) volumes, stroke volume (SV) and ejection fraction (EF) from cardiac MRI is a prerequisite for assessing cardiovascular diseases, and typically requires careful and precise contouring of the ventricles. Although having been the subject of intense research over the years, cardiac function quantification from MRI is still not a fully automatic process in clinical practice. This is partly due to the shortage of training data covering all relevant cardiovascular disease phenotypes. We propose to develop a deep learning convolutional neural network to predict the end-diastolic volume, end-systolic volume, and implicitly the ejection fraction from cardiac MRI without explicit segmentation. The regression network will take as input short axis images and directly predict ESV, EDV, SV and EF. Depending on the results we will likely resort to the generation of synthetic to expand the available data sets that consist of predominantly healthy subjects to include more cases with reduced ejection fraction.