Reproducibility of M1-optimized Liver DWI Across Field Strengths and Gradient Performance

Liver DWI using M1-optimized diffusion imaging (MODI) minimizes motion-induced signal dropout while suppressing lesion-mimicking vessel signal. However, early MODI implementations have been confined to 3.0T MR systems with high-performance gradients. The feasibility and reproducibility of MODI DWI on 1.5T systems and systems with conventional gradient performance remains untested. In this work, liver DWI was acquired using conventional and MODI acquisitions in 8 healthy human subjects on 3 separate MR systems with different field strength and gradient performance. ADC estimation with MODI achieved smaller coefficients of variation and reproducibility coefficients across MR systems in comparison to conventional Stejskal–Tanner (monopolar) DWI.