Tilman Johannes Sumpf1, Martin Uecker1, Susann Boretius1, Jens Frahm1
Quantitative evaluations of the T2 relaxation time are of high importance for diagnostic MRI. Standard T2 mapping procedures rely on the time-demanding acquisition of fully sampled k-space datasets at multiple echo times. Recently, a nonlinear inverse reconstruction method has been proposed which allows for reconstructions of spin-density and T2 maps from highly undersampled radial data. Here, we extended this concept for Cartesian acquisition schemes and even for settings without parallel imaging. An automatic scaling procedure as well as a dedicated undersampling pattern minimize residual artifacts. The approach is validated for T2 mapping of a numerical phantom and the human brain.