Predicting VAT using Machine Learning

What are we doing? Visceral adipose tissue (VAT) is independently associated with increased cardiometabolic disease risk, but its measurement requires radiation exposure and expensive equipment. Despite longstanding interest in predicting VAT, existing prediction equations rely on small, targeted samples and are not well-validated. The aim of this project is to leverage existing data to better predict VAT using common anthropometric and clinical variables.

What am I learning? With this study, I’m combining my training as a physiologist with my interest in statistics and machine learning. I’m learning how to collaborate with researchers at other universities to access data, how to rigorously design and test machine learning models, and how to best communicate my findings in a clinically relevant manner.