Figure 2 | Scientific Reports

Figure 2

From: AI driven analysis of MRI to measure health and disease progression in FSHD

Figure 2

(A) Example of the process to analyze a muscle’s (vastus medialis) cross-sectional area (CSA) muscle composition from a 3D segmentation. (B) The muscle boundary CSA, contractile muscle CSA, fat CSA, and fat fraction (%) were found at each axial slice as a function of axial distance slice-by-slice moving inferior (distal) to superior (proximal). As the muscle characteristics were expressed as a function of axial slice location, the first and last slice’s axial location in which the muscle was present were recorded. Muscle characteristics were then expressed as a function of percent along muscle from 0% (first, inferior location muscle is present) to 100% (last, superior location muscle is present) and were interpolated between points by 1% increments. Shown is the final CSA from 10 to 90%, the last 10% ends were excluded from visualization as the CSA is so small on the ends it increases errors. (C) The 3D model demonstrated again, this time with pixel-by-pixel fat fraction, in which each pixel is color coded based on fat fraction (%) from 0% (blue) to 100% (yellow). (D) Lastly, the frequency in which each pixel’s fat fraction is found is displayed as a histogram from 0% fat infiltration to 100% fat infiltration at baseline (solid line) and 1 year later (dotted line).

Back to article page