Meeting Banner
Abstract #3867

Bone Marrow Adiposity Evaluation using Dixon MRI Fat Fraction with Machine Leaning Segmentation in the UK Biobank – Initial Validation 

David M Morris1, Chengjia Wang1, Giorgos Papanastasiou2, Calum Gray3, Scott IK Semple1, Tom MacGillivray1, and William Cawthorn1
1Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, United Kingdom, 2University of Essex, Colchester, United Kingdom, 3Edinburgh Imaging, University of Edinburgh, Edinburgh, United Kingdom

The UK Biobank is a research tissue data bank that contains genetic, health and MRI data. This includes 2-point Dixon sequences allowing for the assessment of fat fraction to evaluate bone marrow adiposity. Bone marrow adipose tissue is a potential biomarker of diseases including osteoporosis and endocrine disorders. Manual processing of this big data is not feasible and automated processes must be developed and validated. Machine Learning has been developed for the automated segmentation of anatomically appropriate bone marrow regions of interest and the extracted Fat Fraction values validated against expected values across different subject groups.

This abstract and the presentation materials are available to members only; a login is required.

Join Here