A linear relationship between MR signal and contrast-agent concentration (CAC) of the arterial-input function (AIF) is crucial for MR lung-perfusion quantification. The aim was to determine the in-vivo real maximum CAC of the AIF, using cine CT measurements in a porcine model. A dilution series (Gd-DTPA, 0-20 mM) was examined by clinical time-resolved 3D-GRE MRI and by MDCT in cine CT mode. Using the CT setup, data were acquired in five pigs immediately after the injection of 0.05 mmol and 0.07 mmol/kg BW Gd-DTPA. For phantom measurements, mean signal values were determined using a region-of-interest (ROI) analysis and for animal measurements, a ROI was placed in the pulmonary trunk of the cine CT perfusion data sets. The CT phantom measurements were used to calculate the in-vivo maximum CAC corresponding to the HU values obtained in the pulmonary trunk by the cine CT study. Linearity of the AIF of the CT perfusion measurements was verified using the MR phantom measurement results. MR phantom measurements demonstrated linearity for concentrations of 0-4 mM. CT phantom measurements showed linear relation for the entire CAC range. Comparing in-vivo and in-vitro measurements, three of five CA injections at 0.05 mmol/kg and all 0.07 mmol/kg injections exceeded the range of linearity in MRI. The CA dose for quantification of lung perfusion with time-resolved MR studies must be chosen carefully since even with low doses (0.05 mmol/kg) the CAC may exceed the range of linearity in the AIF.