This project aims to develop and validate automated whole-body segmentation tools for UK Biobank imaging data, focusing on fat, organ, and muscle compartments. By extracting volumetric and shape-based biomarkers from MRI scans, we will create robust imaging phenotypes that can be linked to health outcomes.
The primary research questions are:
– Can advanced segmentation tools improve the accuracy and reproducibility of whole-body fat, muscle, and organ measurements in large-scale imaging datasets?
– What is the relationship between imaging-derived biomarkers (e.g., adipose tissue distribution, muscle quality, organ size/shape) and clinical outcomes, including osteoarthritis (OA), metabolic disorders, and other age-related conditions?
– Can imaging biomarkers enhance risk stratification and prediction of disease progression beyond conventional clinical and biochemical measures?
The objectives are to (i) design and test novel segmentation algorithms tailored to UK Biobank imaging protocols, (ii) generate large-scale, standardized imaging-derived phenotypes, and (iii) integrate these biomarkers with clinical and outcome data to identify associations with musculoskeletal and metabolic diseases.
The scientific rationale is that quantitative imaging biomarkers provide a unique window into whole-body structure and composition, enabling early detection of disease processes and better understanding of pathophysiology. By leveraging the scale and diversity of UK Biobank data, this project will advance precision imaging and contribute to improved prevention, diagnosis, and treatment strategies