Meeting Banner
Abstract #3264

Automated Muscle Fat Segmentation in DTI Data of Post-Polio Patients Based on Parameter Distirbutions

Martijn Froeling1, Gustav J. Strijkers2, Frans Nollet3, Marianne de Visser4, Klaas Nicolay2, Aart J. Nederveen5

1Department of Radiology, Academical Medical Center , Amsterdam, Netherlands; 2Biomedical NMR, Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands; 3Department of Rehabilitation Medicine, Academical Medical Center, Amsterdam, Netherlands; 4Department of Neurology, Academical Medical Center, Amsterdam, Netherlands; 5Department of Radiology, Academical Medical Center, Amsterdam, Netherlands

In DTI of skeletal muscle in post-polio patients one would like to make a distinction between healthy and affected muscles and also investigate if the fat-infiltrated muscles show organized structures and thus differs from the subcutaneous fat. The aim of this work was therefore to develop a fully automated algorithm for segmentation of healthy and fat-infiltrated muscle based on the distribution of the diffusion tensor imaging parameters.

Keywords

academical acquisitions addition affected anisotropy anterior arch architecture assessing atrophy automated belonging biomedical blue bring channel chosen coil combined compartment compartments completely complication computed consisted cumulative damage dataset degrees density describe develop differ diffusion diffusivity disease distinction distribution distributions edges eigenvalues engineering entire expected fiber filter fitted five fraction fractional fully function functions gradient green guess healthy hinders histogram histograms homogenized identified impact individual infiltrated initial input insight integration interventions investigate known legs length like little made make male masks measured mechanism medical minimal much muscle muscles needed neurology next organized overlaid overlap partially pathogenesis patient patients polio post posterior precise preserves probability progression proposed radiology rehabilitation remnants repeated replaced represents residual revealed scanner segment segmentation segmented sense severe similarly skeletal skew slice spatially stacks step steps still stopped structures subcutaneous suffer suppression syndrome technology tensor therapeutic therapy threshold together tracking tracts upper user variance view volume volunteer weakness whole yellow