Johannes Tran-Gia1,
Thomas Troalen2, Herbert Kstler1, Frank Kober2
1Institute
of Radiology, University of Wrzburg, Wrzburg, Germany; 2Centre
de Rsonance Magntique Biologique et Mdicale (CRMBM), UMR CNRS N7339,
Facult de Mdecine, Aix-Marseille Universit, Marseille, France
A model-based algorithm for the reconstruction of Look-Locker FAIR Gradient Echo Cartesian ASL datasets is presented. By incorporating a signal model into the image reconstruction, parameter mapping can be performed from highly undersampled k-spaces acquired after FAIR magnetization preparation. In conjunction with a time log of the inversion times of all acquired phase encoding steps, this allows to take into account any variations in inversion time during the experiment, highly improving the temporal resolution of the obtained T1. The functionality of the algorithm and the accuracy of the quantified parameters in comparison to the conventional reconstruction method are demonstrated in-vivo.