We developed automatic methods to extract the AIF for lung first pass perfusion DCE and compared their performance with manual selection of an ROI in the pulmonary artery (PA). Dynamic enhancement was calculated by simple subtraction and with a shuffle subtraction. Early enhancing voxels with the highest enhancing values were selected. We demonstrated that the shuffle subtraction is a more robust method, since it avoids including the early enhancing subclavian vein. Our results suggest obtaining the AIF automatically from pixels within the RV may be more robust against partial volume effects and removes inter-reader variability.