P.10.19 Net muscle volumetry using MRI and CT

Abstract

Muscle volumetry by MRI has been previously reported using T1 weighted images. We reported muscle volumetry using CT and compared the results obtained by CT and DIXON images. Five patients with muscular dystrophy, 3 patients with myopathy, 2 patients with myositis, and 3 patients with neurogenic amyotrophy were enrolled in this study. CT and MRI DIXON images of the 26 limbs (the thigh or lower leg) of these patients were investigated. The CT and MRI images of fourteen 1-cm and seventy 2-mm thickness slices of the mid-thigh, 7cm above and below the midpoint of the trochanter and patella, and ten 1-cm and fifty 2-mm thickness slices of the mid-lower leg, 5cm above and below the midpoint of the fibula head and lateral malleolus, were analyzed, respectively. A muscle density map was obtained using estimated linear-function. Obvious vessels and skin tissues were excluded manually from the density map and the net muscle volumes of the thighs and lower legs were calculated. Percentage water images were simultaneously calculated using DIXON images, and the sum of the %water images resulted in the net muscle volume from MRI. The results obtained by DIXON closely corresponded to those by CT, with a correlation coefficient of 0.993 and intraclass correlation coefficient of 0.993. There was a significant difference between the two results by the paired <i>t</i>-test (<i>p</i>=0.001), and the results obtained by DIXON were slightly larger than those by CT, with an SEM of 47.95cm³. The results obtained by DIXON were correlated to those by CT because these results were calculated from the division of water images by the addition of water images and fat images, and the division procedure reduced the effect of magnetic field inhomogeneity on these images. The results obtained by DIXON were larger than those by CT because the bone marrow was not able to be eliminated on %water images.