Abstract
OBJECTIVE: To explore "whole abdomen" MRI methods for quantifying adipose tissue volumes and to establish associations with body mass index (BMI) and measurement reproducibility-relative to existing "partial abdomen" methods. METHODS: 15 healthy volunteers were scanned on a 3T MRI scanner using a double-echo three-point-Dixon gradient echo sequence. Whole abdomen volumes were acquired via three separate scans ("supine 1", "supine 2" and "prone"). Segmentation was applied to derive (i) "whole abdomen" visceral (VAT) and subcutaneous adipose tissue (SCAT) volumes, and (ii) "partial abdomen" volumes at the lumbar spine (L3 to L5). Root-mean-square coefficients of variation (RMS CoV) were calculated to quantify the variability of each measurement. RESULTS: "Whole abdomen" measurements were found to correlate better with BMI (r(2)(max) = 0.74) than "partial abdomen" volumes (r(2)(max) = 0.66). Total adipose tissue (TAT) measurements correlated better with BMI (r(2)(max) = 0.74) than SCAT (r(2)(max) = 0.43) or VAT (r(2)(max) = 0.33) for both methods. Scan-to-scan RMS CoV's for "whole abdomen" VAT and SCAT measurements were 4.16 and 3.61% compared to 6.31 and 5.07% for "partial abdomen" measurements. CONCLUSION: "Whole abdomen" measures of abdominal adiposity are better correlated with BMI and demonstrate better scan-to-scan reproducibility than "partial abdomen" measures. It is recommended that "whole abdomen" measures be used in longitudinal MRI radiology investigations, where small volume changes may occur. Advances in knowledge: Whole abdomen adipose tissue volumes can be measured and quantified using commercial MRI sequences and post-processing software. These methods are better correlated with BMI and are more reproducible than partial abdomen measures.