Phantom Design: In order to simulate the different fat fraction values of biological tissues, different materials embedded in sixteen (16) test tubes were used. Peanut oil was chosen as a fat component, in order to replicate liver triglycerides, due to its proton NMR spectrum resembling the protons seen in adipose tissue from triglycerides [4]. A set of 25 mL tubes, each contained varying fat content (0%, 3%, 5%, 7%, 10%,15%, 20%, 25%, 50%, 75%, and 100% w/v ). For different fat contents, distilled water, agarose, water-soluble surfactant, sodium benzoate, gadolinium-diethylenetriaminepentacetate (DTPA) contrast agent, peanut oil, and oil-soluble surfactant in Gadolinium-DTPA were used, to better simulate the MRI relaxation characteristics of the phantom. Different Gd-DTPA concentrations were tested, finally a concentration of 100 and 150 μl was used. For reference purposes, one (1) test tube filled with double distilled water and one (1) test tube filled with corn oil were also used.
Imaging: MRI was performed using a 1.5T clinical MRI system with two different local protocols.
(LP-1): PD-T2* weighted single slice (1 coronal slice, breath-hold MEGRE sequence) with 12 echoes (LP-2): PD-T2* weighted multiple slice (5 coronal slices, breath-hold MEGRE sequence) with 8 echoes
Post processing: In both protocols Fat Fraction (FF) was evaluated, and the mean %FF values were calculated by post-processing T2*-weighted images using qMRI Utilities-X, designed for this purpose by two of the authors (GK, TGM) and developed on the local PACS system (EVORAD® platform). Bland-Altman and Concordance Correlation Coefficient (CCC) analyses were conducted to assess the agreement between the two protocols.
