Abstract
PURPOSE: A fast T(1) mapping method of the whole pancreas remains a challenge, due to the complex anatomy of the organ. In addition, a technique for pancreas water T(1) mapping is needed, since the T(1) is biased in the presence of fat. The purpose of this work is to accelerate the acquisition of water selective T(1) (wT(1)) mapping for the whole pancreas at 3 T. METHODS: The proposed method combines a continuous inversion-recovery Look-Locker acquisition with a single-shot gradient echo spiral readout, water-fat separation and dictionary matching for wT(1) mapping of the whole pancreas at 3 T. The bias of T(1) in the presence of fat was evaluated in a phantom by comparing the modified Look-Locker inversion-recovery (MOLLI) and the proposed method to MRS measurements. The present method was validated in 11 volunteers by evaluating its pancreas coverage and repeatability and by comparing it to MOLLI. Four pancreatitis patients were evaluated using the proposed method and clinical scans. RESULTS: The phantom wT(1) results are in better agreement to MRS ( wT1 = 1.02*wT1MRS - 47.81, R2 = 0.99) than MOLLI ( T1MOLLI = 1.13 ∗ wT1MRS - 74.65, R2 = 0.98) . The volunteer wT(1) results demonstrate the whole pancreas coverage capability for different fat fractions, good repeatability ( wT1, 2∘ = 0.98 ∗ wT1, 1∘ + 17.40, R2 = 0.69) and lower T(1) values than MOLLI ( wT1 = 0.34*T1MOLLI + 383.65, R2 = 0.26) . The wT(1) maps in patients captured diverse pancreatitis regions with higher values (wT1Patients = (831 - 1696) ms) than in the volunteers (wT1Volunteers = (605 - 799) ms) , thus showing their potential clinical feasibility. CONCLUSION: The present work proposes a wT(1) mapping methodology of the whole pancreas at 3 T, where 24 slices ( 2 × 2 × 5 mm3 ) were acquired in three short breath-holds of 12 s each.