Quantification of visceral adipose tissue in polycystic ovary syndrome: dual-energy X-ray absorptiometry versus magnetic resonance imaging

Background: Polycystic ovary syndrome (PCOS) is associated with frequent overweight and abdominal obesity. Quantifying visceral adipose tissue (VAT) in PCOS patients can be a tool to assess metabolic risk and monitor effects of treatment. The latest dual-energy X-ray absorptiometry (DXA) technology can measure VAT and subcutaneous adipose tissue (SAT) in a clinical setting. Purpose: To compare DXA-measurements of VAT and SAT with the gold standard MRI in women with PCOS. Material and Methods: A cross-sectional study of 67 overweight women with PCOS was performed. Measurements of VAT and SAT were performed by DXA in a 5-cm thick transverse slice at the L4/L5 level and by MRI in a 1-cm thick transverse slice at the L3 level. Results: Mean (SD) DXA-VAT was 81 (34) cm(3), DXA-SAT was 498 (118) cm 3, MRI-VAT was 117 (48) cm 3, and MRI-SAT was 408 (122) cm 3. MRI and DXA measures of VAT (r = 0.82, P < 0.001) and SAT (r = 0.92, P < 0.001) correlated closely, and DXA-VAT was stronger correlated with MRI-VAT than BMI (r = 0.62, P < 0.001) and waist circumference (r = 0.60, P < 0.001). DXA-VAT coefficient of variance was 6.7% and inter correlation coefficient was 0.98. Bland-Altman analyses showed DXA to slightly underestimate VAT and SAT measurements compared with MRI. Conclusion: DXA and MRI measurements of VAT and SAT correlated closely despite different size of region of interest, and DXA-VAT was superior to waist circumference and BMI in estimating MRI-VAT. DXA showed high reproducibility making it is suitable for repeated measurements in the same individual over time.