Quantifying Visceral Adipose Tissue by DXA

As previously described (13 (link)), full body DXA (Prodigy and iDXA models; GE/Lunar Radiation Corp, Madison, WI) scans were performed by a single certified technician using standard subject positioning and data acquisition protocols. All scan analyses were completed using GE Lunar enCORE™ software version 16.2. From the whole-body DXA scans, an android region of interest (ROI) was defined as the region with its lower border placed at the top of the iliac crest and upper border defined by 20% of the distance between the top of the iliac crest and immediately below the chin. The ROI extended laterally to include the entire torso. DXA-derived VAT was obtained from the ROI using the GE Lunar CoreScan application. This application uses a validated algorithm to estimate visceral fat mass from the android region (14 (link)). DXA-derived VAT has been shown to correlate well with magnetic resonance imaging (MRI)-derived VAT from the L4-L5 region in our sample (r = 0.79, p < 0.001) (13 (link)) and with computed tomography (CT)-derived estimates of VAT in adolescents classified as overweight or obese (r = 0.86; 95% CI: 0.08, 0.90) (15 (link)).