Figure 1.
Results of phantom experiment.
The bottle on the lower left side is filled with pure water (fat content: 0%), and the upper right one is filled with 100% peanut oil, while others have fat content ranging from 10% to 90% in sequence. In the fat phase image, the higher the fat content is, the higher signal intensity can be seen. But in water phase, the signal intensity of each bottle turned opposite to that of fat phase. The fat volume fraction (FVF) map demonstrates clear differences for different amounts of fat and water through the change of color.
Figure 2.
Irregular-shaped Regions of Interest (ROI) (1) covering the entire liver was placed in 21 consecutive slices (max-area centered) on the fat volume fraction (FVF) map of a typical patient.
The average FVF in the ROI was recorded.
Figure 3.
Correlation of actual fat volume fraction (FVF) of the water/fat phantoms against the mean fat fraction computed from Dixon technique.
The best-fit-line was displayed. (r2 = 0.994, p = 0.000).
Figure 4.
FVF maps of two typical subjects from IGT group (A) and NGTLean group (B).
A: a 58 years old female who was diagnosed as IGT (BMI = 31.85), with a mean FVF of 24.23%. The fat phase and FVF map demonstrated that the deposition of fat is homogeneous. B: a 50 years old female (BMI = 23.13) with a mean FVF of 4.15%. The color of the liver in case B is much darker than that in case A, indicating a lower fat content.
Table 1.
Correlations between the liver FVF values and clinical parameters in all subjects as determined by Spearman correlation test.a.b
Figure 5.
Boxplots of FVF values for the IFT, NGTFat and NGTLean groups.
Significant differences was found between IGT group and NGTLean group (p = 0.004), as well as the NGTFat group and NGTLean group (p = 0.006). But there was no significant difference between IGT group and NGTFat group (p = 0.359).