Development and evaluation of emulsified reference materials for magnetic resonance imaging

Hepatic steatosis is characterized by an abnormal accumulation of lipids within hepatocytes. Magnetic resonance imaging (MRI) is a widely used noninvasive method that can accurately and objectively quantify liver fat. To evaluate the accuracy of the quantitatively measured fat fraction, stable and homogenous qualified material is needed as a reference. Surfactant-free micro-emulsions of three fat fractions I, II, and III, corresponding to (9.12 ± 0.02) %, (18.32 ± 0.04) %, and (27.86 ± 0.05) %, respectively, were prepared using a high-intensity focused ultrasonic emulsification technique. The targeted fat fraction of 10-30 % covers the range of grade I moderate non-alcoholic fatty liver disease (NAFLD), which occurs in the early stages that require early detection. Water contents as the main component of the emulsified reference materials (RMs) were determined using the Karl Fisher titration method to evaluate the stability and homogeneity of the RMs. The water contents of fat fraction I, II, and III were (89.12 ± 1.08) %, (79.87 ± 0.81) %, and (72.71 ± 1.29) %, respectively. The RMs were stable for six months and showed good homogeneity with both standard deviations between and within units in the range of 0.3 – 0.6 %. The physical phantom consisted of nine vials of RMs surrounded by agarose gel. The phantom was scanned on 3 T MRI (Siemens MAGNETOM Vida, Siemens Healthineers, Erlangen, Germany). The correlation between the measured proton density fat fraction (PDFF) values and the fabricated fat fraction values was evaluated using linear regression analysis. The slope of the linear fitting was 0.99, and the intercept was –0.88 %. These results show that the developed RMs can provide a reference value for the measured fat fraction from a medical imaging system to evaluate the effectiveness of a measurement procedure. It is also expected that the developed RMs can be utilized to harmonize the measured values across multi-site.