Objectives: To determine the diagnostic accuracy of pharmacokinetic parameters measured by dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) in predicting the response of brain metastases to antineoplastic therapy in patients with lung cancer. Methods: Forty-four consecutive patients with lung cancer, harbouring 123 newly diagnosed brain metastases prospectively underwent conventional 3-T MRI at baseline (within 1 month before treatment), during the early (7-10 weeks) and midterm (5-7 months) post-treatment period. An additional DCE MRI sequence was performed during baseline and early post-treatment MRI to evaluate baseline pharmacokinetic parameters (K trans, k ep, v e, v p) and their early variation (∆K trans, ∆k ep, ∆v e, ∆v p). The objective response was judged by the volume variation of each metastasis from baseline to midterm MRI. ROC curve analysis determined the best DCE MRI parameter to predict the objective response. Results: Baseline DCE MRI parameters were not associated with the objective response. Early ∆K trans, ∆v e and ∆v p were significantly associated with the objective response (p = 0.02, p = 0.001 and p = 0.02, respectively). The best predictor of objective response was ∆v e with an area under the curve of 0.93 [95% CI = 0.87, 0.99]. Conclusions: DCE MRI and early ∆v e may be a useful tool to predict the objective response of brain metastases in patients with lung cancer. Key points: • DCE MRI could predict the response of brain metastases from lung cancer • ∆v e was the best predictor of response • DCE MRI could be used to individualize patients' follow-up.