Reproducibility and Repeatability of Computer Tomography-based Measurement of Abdominal Subcutaneous and Visceral Adipose Tissues

Excessive accumulation of abdominal adipose tissue is a widely recognized as a major feature of obesity, and it can be quantified by dual-energy x-ray absorptiometry (DXA). However, in a phantom study, the inter- and intra-instrument reliability of DXA remains unpredictable. Thus, we attempted to determine the precision of estimates from computer tomography-based measurements and analysis with AZE Virtual Place software. To determine the inter-rater reproducibility and intra-rater repeatability of adipose tissue area estimates, we used the automatic boundary-tracing function of the AZE Virtual Place to generate cross-sectional areas of subcutaneous and visceral adipose tissues from the abdomen of reconstructed CT images. The variability of inter-rater and intra-rater estimates expressed as the coefficient of variation ranged from 0.47% to 1.43% for subcutaneous adipose tissue and 1.08% to 2.20% for visceral adipose tissue; the optimal coefficient of variation of the fat rate calculation ranged from 0.55% to 1.13%, respectively. There was high and significant correlation between adipose tissue areas as estimated in 40 obese subjects by two raters or repeatedly on 20 obese subjects by either rater. This indicates excellent reproducibility and repeatability via a computer tomography-based measurement of abdominal subcutaneous and visceral adipose tissues.