The aim of this study was to evaluate angiogenic compositions and tumour response in the course of neoadjuvant chemotherapy in patients with locally advanced breast cancer (LABC) using dynamic contrast-enhanced (DCE) MRI. Thirteen patients with LABC underwent serial DCE MRI during the course of chemotherapy. DCE MRI was quantified using a two-compartment model on a pixel-by-pixel basis. Analysis of parametric histograms of amplitude, exchange rate k(out) and peak enhancement over the whole tumour was performed. The distribution patterns of histograms were correlated with the tumour response. Initial kurtosis and standard deviation of amplitude before chemotherapy correlated with tumour response, r = 0.63 and r = 0.61, respectively. Comparing the initial values with the values after the first course of chemotherapy, tumour response was associated with a decrease in standard deviation of amplitude (r = 0.79), and an increase in kurtosis and a decrease in standard deviation of k(out) (r = 0.57 and 0.57, respectively). Comparing the initial values with the values after completing the chemotherapy, tumours with better response were associated with an increase in kurtosis (r = 0.62), a decrease in mean (r = 0.84) and standard deviation (r = 0.77) of amplitude, and a decrease in mean of peak enhancement (r = 0.71). Our results suggested that tumours with better response tended to alter their internal compositions from heterogeneous to homogeneous distributions and a decrease in peak enhancement after chemotherapy. Serial analyses of parametric histograms of DCE MRI-derived angiogenic parameters are potentially useful to monitor the response of angiogenic compositions of a tumour throughout the course of chemotherapy, and might predict tumour response early in the course.