Reference Values for Water-Specific T1 of the Liver at 3 T: T2*-Compensation and the Confounding Effects of Fat

Background: T1 mapping of the liver is confounded by the presence of fat. Multiparametric T1 mapping combines fat-water separation with T1-weighting to enable imaging of water-specific T1 (T1(Water)), proton density fat fraction (PDFF), and T2* values. However, normative T1(Water) values in the liver and its dependence on age/sex is unknown. Purpose: Determine normative values for T1(Water) in the liver with comparison to MOLLI and evaluate a T2*-compensation approach to reduce T1 variability. Study Type: Prospective observational; phantoms. Populations: One hundred twenty-four controls (56 male, 18-75 years), 50 patients at-risk for liver disease (18 male, 30-76 years). Field Strength/Sequence: 2.89 T; Saturation-recovery chemical-shift encoded T1 Mapping (SR-CSE); MOLLI. Assessment: SR-CSE provided T1(Water) measurements, PDFF and T2* values in the liver across three slices in 6 seconds. These were compared with MOLLI T1 values. A new T2*-compensation approach to reduce T1 variability was evaluated test/re-test reproducibility. Statistical Tests: Linear regression, ANCOVA, t-test, Bland and Altman, intraclass correlation coefficient (ICC). P < 0.05 was considered statistically significant. Results: Liver T1 values were significantly higher in healthy females (F) than males (M) for both SR-CSE (F-973 +/- 78 msec, M-930 +/- 72 msec) and MOLLI (F-802 +/- 55 msec, M-759 +/- 69 msec). T1 values were negatively correlated with age, with similar sex- and age-dependencies observed in T2*. The T2*-compensation model reduced the variability of T1 values by half and removed sex- and age-differences (SR-CSE: F-946 +/- 36 msec, M-941 +/- 43 msec; MOLLI: F-775 +/- 35 msec, M-770 +/- 35 msec). At-risk participants had elevated PDFF and T1 values, which became more distinct from the healthy cohort after T2*-compensation. MOLLI systematically underestimated liver T1 values by similar to 170 msec with an additional positive T1-bias from fat content (similar to 11 msec/1% in PDFF). Reproducibility ICC values were >= 0.96 for all parameters. Data Conclusion: Liver T1(Water) values were lower in males and decreased with age, as observed for SR-CSE and MOLLI acquisitions. MOLLI underestimated liver T1 with an additional large positive fat-modulated T1 bias. T2*-compensation removed sex- and age-dependence in liver T1, reduced the range of healthy values and increased T1 group differences between healthy and at-risk groups.