Effects of filtering methods on muscle and fat cross-sectional area measurement by pQCT: a technical note

Peripheral quantitative computed tomography (pQCT) is most commonly used for bone density and morphology assessment of the limbs, but it can also be used for soft tissue area quantification by segmenting regions representing different tissues. Scanning and analyzing cross-sectional areas of larger thighs present a special challenge due to increased statistical noise created from fewer detected x-ray photons. The purpose of this technical note is to compare total, muscle and fat cross-sectional area (CSA) measurements of the midthigh with Stratec 3000 pQCT scans using no filter, a weak smoothing filter and a strong smoothing filter to CSA measurements of midthigh MRI scans analyzed by Image J, a public domain image processing program. Nine healthy men and women participated in this study. CSAs did not differ significantly between MRI and strongly filtered pQCT images with per cent differences ranging from -3.1% for muscle to +6.5% for fat. The per cent difference in muscle CSA values between MRI and pQCT with the weak filter (-24.0 +/- 38.0%) or no filter (-44.9 +/- 22.7%) was strongly related to total thigh CSA (r = 0.78-0.92, p < 0.05). We propose that the midthigh can be assessed for soft tissue area measurements with pQCT, provided that strong smoothing filter is utilized.