Research Corner - Machine Learning DXA for Prediction of Hepatic Steatosis
Following on the heels of our first paper, this is our second paper.
We created a machine learning algorithm that can predict accurately hepatic steatosis using DXA parameters.
Our study demonstrates value in body composition information captured by dxa (dual-energy x-ray absorptiometry) which can be used in a prediction model to predict hepatic steatosis using machine learning .
Not surprisingly visceral fat was very predictive, but several other dxa parameters were also useful in the prediction model.
There are also differences between genders, with android and trunk fat mass being more important in females.
Body composition imaging using dxa is underutilised. We use dxa commonly for bone density assessment, but not so much for body composition such as extracting muscle mass and visceral fat. It is relatively cheap, with very low-level radiation, and has a ton of value particularly if you want to optimise health metrics relating to muscle and fat composition.
For more details:
https://www.linkedin.com/posts/dr-vincev_bodycomposition-dxa-machinelearning-activity-7150470403260317696-lMU4?utm_source=share&utm_medium=member_desktop