Large-scale medical image analysis for detailed studies of causes and consequences of body composition in relation to cardiovascular disease

We are facing a global epidemic of obesity and related cardiovascular complications. Improved understanding of underlying mechanisms is crucial to optimize intervention strategies.

Both total fat mass, its distribution in the body, and fat infiltration of, for example, liver and muscle are linked to cardiovascular risk.

Magnetic resonance imaging (MRI) and computed tomography (CT) can measure this physiological information (tissue volume and fat content) in millions of voxels throughout the whole body.

We have experience and new ideas for multiple advanced image analysis techniques based on image registration and machine learning that allow unique integrated analysis of medical image data and non-image (pheno- and genotype) data.The overall aim of this project is to develop and apply these image analysis techniques to improve our understanding of human body composition and its causes and consequences in relation to cardiovascular disease.Specific aims include:1) Development of multiple novel approaches including tailored deep regression, cohort saliency analysis, and causality imaging for both MRI and CT.2) Application to multiple large-scale cohort studies (n>86,000) including detailed studies of cardiovascular disease, and type 2 diabetes. 3) Studies of genetic data in relation to image data for detailed analysis of causes and consequences of body composition.We anticipate multiple important findings that may serve as a springboard for novel intervention strategies.