Combined Automatic Segmentation of Fat and Muscle Compartments with T1 and T2* Measurements Using a Triple-Angle Multiple Gradient-Echo Acquisition Technique

Benjamin Leporq¹, Yann Le Fur², Patrick Cozzone², Olivier Beuf¹, David Bendahan²

Due to its sensitivity to key processes in the diseased muscle such as edema and fat infiltration, MRI is emerging as a suitable quantitative method which could provide reliable surrogate markers of disease severity and progression. The aim of this work was to evaluate the feasibility of 3D segmentation of fat and muscular tissue using a triple-angles multiple gradient-echo acquisition. Our approach includes a complex-based technique for fat volume fraction (FVF) quantification with independent measurements of fat and water T1 and T2*and corrections from T1-related bias and spectral complexity of fat.