A predictive model for estimating regional skeletal muscle size by two-point dixon magnetic resonance imaging in healthy Koreans

This study was undertaken to develop and cross-validate reference and individual predictive models for estimating functional thigh muscle cross-sectional area (TCSA) by 2-point Dixon magnetic resonance imaging (MRI). TCSAs of dominant sides at the mid-thigh level were measured by 2-point Dixon MRI (MRITCSA). Functional MRITCSA were compared with the predictive models in a sample of 92 younger (20-40 years; 28.55 ± 4.87; n = 50) and older (>65 years; 71.22 ± 4.82; n = 42) Koreans. Lean body masses were measured by dual energy X-ray absorptiometry (DXALBM), and thigh isokinetic muscle strengths, extension peak torque at 60°/sec, were measured using a Biodex® dynamometer (BiodexEPT). Multiple regression analysis generated the reference model (R2 = 0.75 and SEE = 1472.63 mm2 (8%)) as follows: The reference model: functional TCSA(mm2) = −1230.49 + 62.81*height + 3061.78*gender −2692.57*age + 58.91*weight. The individual model (R2 = 0.80, SEE = 1158.34 mm2 (7%)) was as follows: The individual model: functional TCSA(mm2) = 1631.62 + 1.76* DXALBM + 9.51*BiodexEPT where height is in centimeters; weight is in kilograms; for gender, female = 0 and male = 1; and for age, age under 40 = 1 and age over 65 = 2. PRESS statistics of R2 and SEE were 0.78 and 1382.98 mm2 for the reference model, and 0.88 and 979.02 mm2 for the individual model. The 2-point Dixon MRI appears to be valid for measuring functional muscle size. Our results suggest that the reference and individual models provide acceptable estimates of functional thigh muscle CSA in healthy Korean adults. Therefore, the models developed in the study could be useful as a research tool to establish indexes for functional muscle composition in healthy Koreans.