This study aimed to explore the feasibility of dynamic contrasted enhancement MRI (DCE-MRI) for predicting the response to gemcitabine in pancreatic ductal adenocarcinoma, as well as the influence of different region of interests (ROIs) on quantitative parameters. We compared the differences of DCE-MRI parameters between responders and non-responders. Kep based on periphery ROI was the best predictive marker, showed the highest areas under ROC curve (AUC) of 0.806. Quantitative DCE-MRI may be a feasible method, and the parameters are useful for the prediction of response to gemcitabine in patients with PAC. The positions of ROI influenced the DCE-MRI parameters.