A DXA-based mathematical model predicts midthigh muscle mass from magnetic resonance imaging in typically developing children but not in those with quadriplegic cerebral palsy.

Valid methods for assessing regional muscle mass in children are needed. The aim of this study was to determine whether dual-energy X-ray absorptiometry (DXA) can accurately estimate midthigh muscle mass from MRI (muscle(MRI)) in typically developing children and children with quadriplegic cerebral palsy (CP). A mathematical model predicting muscle(MRI) from midthigh, fat-free soft tissue mass from DXA (FFST(DXA)) was developed using 48 typically developing children (6-13 y) and was validated using the leave-one-out method. The model was also tested in children with quadriplegic CP (n = 10). The model produced valid estimates of midthigh muscle mass (muscle(DXA)) in typically developing children, as indicated by a very strong relationship between muscle(DXA) and muscle(MRI) (r(2) = 0.95; SEE = 68 g; P < 0.001), no difference in muscle(DXA) and muscle(MRI) (P = 0.951), and visual examination using a Bland-Altman plot. Muscle(DXA) was very strongly related to muscle(MRI) in children with CP (r(2) = 0.96; SEE = 54 g; P < 0.001); however, muscle(DXA) overestimated muscle(MRI) by 15% (P = 0.006). The overestimation of muscle(MRI) by muscle(DXA) was strongly related to the lower ratio of muscle(MRI) to FFST(DXA) (muscle(MRI)/FFST(DXA)) in children with CP (r(2) = 0.75; P = 0.001). The findings suggest that the DXA-based mathematical model developed in the current study can accurately estimate midthigh muscle mass in typically developing children. However, a population-specific model that takes into account the lower muscle(MRI)/FFST(DXA) is needed to estimate midthigh muscle mass in children with quadriplegic CP.