Phenotyping cardiovascular disease using deep learning: Correlation of medical images, biomarkers and genetics to clinical outcome.
Principal Investigator: Johan Verjans
Approved Research ID: 33221
Approval date: June 25th 2018
A. To determine whether unsupervised approaches to quantify cardiac MR images can lead to novel quantifiable risk parameters in cardiac MR images in patients. The UK Biobank stated purpose is to improve the prevention, diagnosis and treatment of a wide range of illnesses, including heart diseases, stroke. Our proposal aims to examine the relation of novel quantifiable image parameters to phenotypic links within cardiovascular disease. If a novel (set of) imaging biomarker(s) were to be found by this research, this could lead to advances in the diagnosis, prediction, and treatment of cardiovascular disease. In addition, this strategy would be very low-cost since it makes use of existing imaging data. There is more than the eye of a radiologist can see and measure. We will use artificial intelligence to correlate cardiac MR images, biomarkers, and genetics to predict clinical outcome and find novel risk factors. We will use all the patients for which cardiac MRIs and parameters are available.