Quantification of fat using chemical shift imaging and 1H-MR spectroscopy in phantom model

Abstract

Objective To evaluate the accuracy of chemical shift imaging (CSI) and MR spectroscopy (MRS) for fat quantification in phantom model. Methods Eleven phantoms were made according to the volume percentage of fat ranging from 0 to 100% with an interval of 10% . The fat concentration in the phantoms were measured respectively by CSI and MRS and compared using one-sample t test The correlation between the two methods was also analyzed. The concentration of saturated fatty acids ( FS), unsaturated fatty acids (FU) and the polyunsaturation degree (PUD) were calculated by using MRS. Results The fat concentration was (48.0 ±1.0)%, (57.0 ±0.5)%, (67.3±0.6)%, (77.3 ±0.6)%, (83.3±0.6)% and (91.0±1.0)% respectively with fat volume of 50% to 100% by CSI. The fat concentration was (8.3 ±0.6)%, (16.3 ±0.7)%, (27.7 ±0.6)%, (36.0 ±1.0)%, (43.5±0. 6)% and (56. 5 ±1. 0)% respectively with fat volume of 10% to 60% by MRS, the fat concentration were underestimated by CSI and MRS (P < 0.05 ), and had high linear correlation with the real concentration in phantoms ( CSI: r = 0. 998, MRS: r = 0.996, P < 0.01 ) . There was also a linear correlation between two methods (r = 0. 992, P < 0. 01 ) but no statistically significant difference ( pairedsamples t test, t = -0. 125 ,P = 0.903). By using MRS, the relative ratio of FS and FU in fat were 0. 15and 0. 85, the PUD was 0. 0325, respectively, and highly consistent with these in phantoms. Conclusion Both CSI and MRS are efficient and accurate methods in fat quantification at 7.0 T MR. Key words: Magnetic resonance imaging; Magnetic resonance spectroscopy; Fats; Models