T1 mapping for liver function evaluation in gadoxetic acid-enhanced MR imaging: comparison of look-locker inversion recovery and B₁ inhomogeneity-corrected variable flip angle method.

Abstract

To compare look-locker inversion recovery (LLIR) and B₁ inhomogeneity-corrected variable flip angle T1 mapping methods for estimation of liver function and prediction of hepatic insufficiency and decompensation on gadoxetic acid-enhanced MR imaging. In this retrospective study, 248 patients with normal liver function, chronic liver disease, or cirrhosis underwent gadoxetic acid-enhanced liver MR imaging, including T1 mapping at 10-min and 20-min hepatobiliary phase (HBP) by using both methods. T1 relaxation times of the liver (T1_Liver-pre, T1_Liver-post) and the spleen (T1_Spleen) were correlated between two methods. ΔT1_Liver ([T1_Liver-pre - T1_Liver-post]/T1_Liver-pre), adjusted T1_Liver ([T1_Spleen - T1_Liver-post]/T1_Spleen), and functional liver volume-to-weight ratio (liver volume on volumetric T1 map/[T1_Liver-post × patient's weight]) were calculated. The diagnostic performance of T1 parameters and the predictive performance of models (serum marker, serum marker plus T1 parameter) were compared. T1_Liver-post showed a strong correlation (r = 0.93, p < 0.001) between two methods but was significantly different. For depicting cirrhosis, LLIR-adjusted T1_Liver at 10-min HBP showed the highest performance (p < 0.025). For predicting hepatic insufficiency and decompensation, LLIR-adjusted T1_Liver (Akaike information criterion (AIC), 58.37; C-index, 0.867) and LLIR-T1_Liver-post (AIC, 48.82; C-index, 0.885) at 10-min HBP showed the best performance, respectively, when added to serum albumin level. T1_Liver-post showed a strong correlation between two methods but with significant differences. T1 mapping using LLIR at 10-min HBP with obtainment of adjusted T1_Liver and T1_Liver-post may be the best approach for estimation of liver function and prediction of hepatic insufficiency and decompensation. • T1 _Liver-post showed a strong correlation between LLIR and B ₁ inhomogeneity-corrected VFA methods, both at 10-min and 20-min HBP but with significant differences. • T1 _Liver-post at 10-min and 20-min HBP using LLIR and B ₁ inhomogeneity-corrected VFA methods could not be used interchangeably during the follow-up in patients with chronic liver disease (CLD) or cirrhosis. • T1 mapping using LLIR at 10-min HBP with obtainment of adjusted T1 _Liver and T1 _Liver-post may be the most suitable method and parameter for estimation of global liver function and prediction of clinical outcomes in patients with CLD or cirrhosis.