Genetically engineered mouse (GEM) model of pancreatic ductal adenocarcinoma (PDA) recapitulates a dense stroma exhibited in human disease and is thus relevant for testing stroma-directed drugs. However, the lesion location makes it highly susceptible to motion. We present a multi-slice 2D saturation-recovery technique using golden-angle radial k-space sampling combined with KWIC reconstruction for DCE-MRI of PDA. This method minimizes respiration motion artefacts in PDA tumors and increases the effective temporal resolution of the AIF. Ktrans and kep maps derived from PK model fitting of DCE series exhibit adequate spatial resolution to reveal permeability and perfusion heterogeneity in the PDA tumor.