Comparison of In Vivo Intramuscular Fat Quantification Techniques in MRI

Abstract: People with cerebral palsy (CP) may develop muscle contractures, which are defined as a loss of joint range due to an increase in passive muscle stiffness [1]. Computer models suggest that intramuscular fat can increase muscle stiffness [2]. There is evidence that children with spastic CP have elevated intramuscular fat fractions, but quantitative estimations of the intramuscular fat fraction are variable [1]. Therefore, a reliable method for in vivo quantification of intramuscular fat is required. The aim of this thesis was to compare two- and three-point Dixon fat-water separation methods, as well as the IDEAL algorithm for intramuscular fat quantification of the gastrocnemius medialis, for both children with CP and typically developed peers. As a reference standard, the water- and fat-only maps resulting from the Philips mDIXONXD turbo-spin-echo sequence were used (Ingenia CX, Philips Healthcare, The Netherlands). It was found that neither of the methods can compare to the reference standard, which is probably due to poor performance of the phase unwrapping algorithm applied on this data. Further studies need to be done in order to better quantify the phase error in multi-echo Dixon MRI.