Hing-Chiu Chang1, Chun-Jung Juan2, Yin-Cheng Kris Huang3, Hsiao-Wen Chung3
1Applied Science Laboratory, GE Healthcare
Taiwan, Taipei, Taiwan; 2Department of Radiology, Tri-Service
General Hospital, Taipei, Taiwan; 3Department of Electrical
Engineering, National Taiwan University, Taipei, Taiwan
Several groups have used different MRI imaging techniques to estimate liver fat content including spectroscopic approach, opposed-phase imaging, RF saturation and three-point Dixon IDEAL. However, tissue relaxation, in particular T1, is a potential source of bias in fat quantification unless corrections are performed. In this experimental work using phantoms with long and short T1 values, we report the quantitative influences of the T1 effect on fat quantification when using RF saturation and IDEAL methods. Our results suggest fat-fraction measurement is affected by the T1 value of the tissue in a nonlinear manner, with larger bias encountered in long-T1 than short-T1 tissues. While T1 relaxation is an obvious factor that could influence accuracy in fat quantification, previous studies seldom considered the T1 effects. The results from our study show that the T1 bias in fat-fraction measurements depends on scan parameters (T1 or PD) and sequences (FSE vs. GRE), with PD-weighted images better immune to T1 biasing errors. In clinical situations where relaxation parameters may alter pathologically, therefore, T1 effects should be included in fat-fraction quantification using RF saturation and IDEAL.